[Federal Register Volume 61, Number 132 (Tuesday, July 9, 1996)]
[Proposed Rules]
[Pages 36154-36219]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-17058]


      

[[Page 36153]]


_______________________________________________________________________

Part III





Department of Health and Human Services





_______________________________________________________________________



Food and Drug Administration



_______________________________________________________________________



21 CFR Parts 106 and 107



Current Good Manufacturing Practice, Quality Control Procedures, 
Quality Factors, Notification Requirements, and Records and Reports, 
for the Production of Infant Formula; Proposed Rule

  Federal Register / Vol. 61, No. 132 / Tuesday, July 9, 1996 / 
Proposed Rules  

[[Page 36154]]



DEPARTMENT OF HEALTH AND HUMAN SERVICES

Food and Drug Administration

21 CFR Parts 106 and 107

[Docket No. 95N-0309]
RIN 0910-AA04


Current Good Manufacturing Practice, Quality Control Procedures, 
Quality Factors, Notification Requirements, and Records and Reports, 
for the Production of Infant Formula

AGENCY: Food and Drug Administration, HHS.

ACTION: Proposed rule.

-----------------------------------------------------------------------

SUMMARY: The Food and Drug Administration (FDA) is proposing to revise 
its infant formula regulations to establish requirements for quality 
factors and current good manufacturing practice (CGMP); to amend its 
quality control procedure, notification, and records and report 
requirements for infant formulas; to require that infant formulas 
contain, and be tested for, required nutrients and for any nutrient 
added by the manufacturer throughout their shelf life, and that they be 
produced under strict microbiological controls; and to require that 
manufacturers implement the CGMP and quality control procedure 
requirements by establishing a production and in-process control system 
of their own design. This action is being taken to improve the 
protection of infants that use infant formula products.

DATES: Comments by October 7, 1996, except that comments regarding 
information collection should be submitted by August 8, 1996. The 
agency proposes that any final rule that may issue based on this 
proposal become effective 120 days after its date of publication.

ADDRESSES: Submit written comments, data, or information to the Dockets 
Management Branch (HFA-305), Food and Drug Administration, rm. 1-23, 
12420 Parklawn Dr., Rockville, MD 20857. Comments regarding information 
collection to the Office of Information and Regulatory Affairs, OMB, 
New Executive Office Bldg., 725 17th St. NW., rm. 10235, Washington, DC 
20503, ATTN: Desk Officer for FDA.

FOR FURTHER INFORMATION CONTACT: Carolyn W. Miles, Center for Food 
Safety and Applied Nutrition (HFS-456), Food and Drug Administration, 
200 C St. SW., Washington, DC 20204, 202-401-9858.

SUPPLEMENTARY INFORMATION:

I. Background

A. The Infant Formula Act of 1980

    In 1978, a major manufacturer of infant formula reformulated two of 
its soy products by discontinuing the addition of salt. This 
reformulation resulted in infant formula products that contained an 
inadequate amount of chloride, an essential nutrient for growth and 
development in infants. By mid-1979, a substantial number of infants 
had been diagnosed with hypochloremic metabolic alkalosis, a syndrome 
associated with chloride deficiency. Development of this syndrome in 
these infants was found to be associated with prolonged exclusive use 
of chloride-deficient soy formulas.
    After reviewing the matter, Congress determined that, to improve 
protection of infants using infant formula products, greater regulatory 
control over the formulation and production of infant formula was 
needed, including modifications of industry's and FDA's recall 
procedures. Accordingly, Congress passed, and the President signed into 
law on September 26, 1980, the Infant Formula Act of 1980 (the 1980 
act) (Pub. L. 96-359). This law amended the act to include section 412 
(21 U.S.C. 350a).
    In 1982, FDA adopted infant formula recall procedures, establishing 
subpart D of part 107 of its regulations (21 CFR part 107) (47 FR 
18832, April 30, 1982), and infant formula quality control procedures 
(21 CFR part 106 (47 FR 17016, April 20, 1982)). In 1985, FDA further 
implemented the 1980 act by establishing subparts B, C, and D in 21 CFR 
part 107 regarding the labeling of infant formula, exempt infant 
formulas, and nutrient requirements for infant formula, respectively 
(50 FR 1833, January 14, 1985; 50 FR 48183, November 22, 1985; and 50 
FR 45106, October 30, 1985).

B. The 1986 Amendments to the Infant Formula Act

    In 1986, Congress, as part of the Drug Enforcement, Education, and 
Control Act of 1986 (the 1986 amendments) (Pub. L. 99-570) completely 
revamped section 412 of the act to address concerns that had been 
expressed by Congress and consumers about the 1980 act and FDA's 
implementation of that statute. These concerns included whether the 
quality control testing, CGMP, recordkeeping, and recall requirements 
that FDA had adopted would prevent children ``from ever again being 
threatened by defective baby formula'' (Ref. 1). The 1986 amendments: 
(1) State that an infant formula is deemed to be adulterated unless it 
provides certain required nutrients, meets the quality factor 
requirements established by the Secretary of Health and Human Services 
(the Secretary) (and, by delegation, FDA), and is manufactured in 
accordance with CGMP and quality control procedures established by the 
Secretary; (2) require that the Secretary issue regulations 
establishing requirements for quality factors and CGMP, including 
quality control procedures; (3) require that infant formula 
manufacturers regularly audit their operations to ensure that those 
operations comply with CGMP and quality control procedure regulations; 
(4) expand the circumstances in which manufacturers must make a 
submission to the agency to include when a manufacturer makes major 
changes in an infant formula, and when a manufacturer makes changes 
that may affect whether the formula is adulterated; (5) specify the 
nutrient quality control testing that must be done on each batch of 
infant formula; (6) modify the infant formula recall requirements; and 
(7) give the Secretary authority to establish requirements for 
retention of records, including records necessary to demonstrate 
compliance with CGMP and quality control procedures.
    In 1989, the agency responded to the provisions of the 1986 
amendments on recalls (sections 412(f) and (g) of the act) by 
establishing subpart E in part 107 (54 FR 4006, January 27, 1989). In 
1991, the agency adopted infant formula record and record retention 
requirements that implemented the 1986 amendments by revising 
Sec. 106.100 (56 FR 66566, December 24, 1991).
    Although the agency has adopted regulations that respond to a 
number of the provisions of the 1986 amendments, it has not issued 
regulations on infant formula CGMP and quality factors or revised the 
notification procedures and quality control procedures to reflect the 
1986 amendments. Since the passage of the 1986 amendments, agency 
representatives have visited infant formula plants to observe the 
manufacturing practice and quality control procedures that they employ, 
and the agency has solicited and received recommendations on CGMP from 
the Infant Formula Council. In addition, FDA has contracted with the 
Committee on Nutrition of the American Academy of Pediatrics (CON/AAP) 
to obtain expert advice on clinical testing of infant formulas with 
respect to the quality factor requirements. Moreover, both industry and 
the agency have

[[Page 36155]]

increased experience with the quantity and quality of information that 
should be submitted to meet the notification requirements of section 
412(c) and (d) of the act.
    This proposal addresses CGMP, quality control procedures, quality 
factors, and notification procedures and incorporates information 
resulting from the interactions between FDA and industry and between 
FDA and AAP. This proposal updates the language in part 107 to reflect 
the 1986 amendments and the November 1992 reorganization of the Center 
for Food Safety and Applied Nutrition (CFSAN).

C. FDA's Regulations on Nutrient Requirements

    Section 412(i) of the act includes a table that lists nutrients 
that every infant formula must contain. This section also establishes a 
minimum level for each of the listed nutrients and a maximum level for 
eight of the listed nutrients. In addition, section 412(i)(2) of the 
act grants the Secretary (and by delegation FDA) the authority to 
revise the list of nutrients in section 412(i), and the minimum and 
maximum levels of those nutrients, by regulation. In the Federal 
Register of October 30, 1995, FDA established the nutrient requirements 
for infant formulas in Sec. 107.100 (50 FR 45106). For the purpose of 
this document, the nutrients that are required to be in infant formula 
under Sec. 107.100 will be referred to as ``required nutrients,'' and 
the levels of these required nutrients established in Sec. 107.100 will 
be referred to as ``required levels.''

II. The Need for Regulation

    Relative to per unit of body weight, nutrient requirements are 
generally greater in infancy than at any other time during life. During 
the first year, the rate of growth is at its maximum, with birth weight 
typically doubling by 4 months of age and tripling by 1 year (Refs. 2 
and 3). Moreover, the metabolic rate in infants is greater, and the 
turnover of nutrients is more rapid, than in adults (Ref. 4). Thus, 
infants must ingest adequate nutrients to support a rapid rate of 
growth and of developmental changes and to supply maintenance needs. 
Without adequate nutrition, infants would be unable to achieve their 
genetic potential for growth and development.
    These nutritional needs must be met in early infancy by food in 
liquid form. Sucking and involuntary swallow reflexes are the 
mechanisms by which very young infants ingest food until teeth and 
motor coordination develop. Consequently, for infants who are not fed 
breast milk, infant formula often serves as the sole source, or the 
major source, of nutrition during this time of rapid growth and 
development.
    Therefore, the importance of proper infant formula manufacture, 
composition, and nutrient levels cannot be overstated. Senator 
Metzenbaum explained why infant formula needs more regulation than 
other foods when he stated ``there is simply no margin for error in the 
production of baby formula. An infant relies on the formula to sustain 
life and provide the proper nourishment at a time of rapid physical and 
mental development'' (Ref. 1). The requirements contained in this 
proposal are designed to ensure that the formula fed to American 
infants fulfills its important function.
    The CGMP and quality control procedures that FDA is proposing are 
designed to prevent the production of an adulterated infant formula. 
Defining CGMP will help to ensure that all of the required nutrients 
are included at appropriate levels in the formula, and that the formula 
is not contaminated with microorganisms or other materials that may be 
harmful to the infant.
    Quality control procedures are designed to ensure that an infant 
formula contains the nutrients that are necessary to support growth and 
development, at the appropriate levels, not only when it enters into 
commerce but throughout its shelf life. FDA is proposing that each 
batch of infant formula be tested for all required nutrients and any 
nutrient added by the manufacturer, and that finished batches be 
periodically sampled and tested for nutrients throughout the shelf life 
of the product.
    Quality factors are designed to ensure that the required nutrients 
and any nutrient added by the manufacturer actually reach the infant in 
a useable form. Quality factors ``pertain to the bioavailability of a 
nutrient and the maintenance of level or potency of nutrients during 
the expected shelf life of the product'' (Ref. 5). The 1986 amendments 
directed that the Secretary, by regulation, ``establish requirements 
for quality factors for infant formulas to the extent possible 
consistent with current scientific knowledge, including quality factor 
requirements for the nutrients required by (section 412(i) of the 
act).''
    In 1986, FDA advised Congress that the technology and science with 
respect to quality factors was still evolving, and that it was only 
possible to establish a quality factor for one nutrient. The agency 
said that it had already done so. However, in the 1986 Congressional 
Record (Ref. 1), Senator Metzenbaum stated that ``the legislation 
contemplates that the Secretary will move to promptly develop and issue 
appropriate quality factor standards for different nutrients as the 
state of the science progresses.'' Since that time, as stated above, 
FDA has contracted with CON/AAP to obtain expert advice on quality 
factors; i.e., on the clinical testing of infant formula with respect 
to its nutritional safety and suitability for term infants.
    In 1988, CON/AAP submitted a report (Ref. 6) under the contract 
that identified and discussed the types of clinical studies that might 
be considered for evaluation of the nutritional suitability of a 
formula for normal term infants. FDA has reviewed this report and the 
available scientific literature and has identified quality factors for 
protein and for complete infant formulas. The agency is proposing to 
adopt these quality factors as part of these regulations.
    FDA has received numerous inquiries from industry for specific 
guidance on what information must be submitted to meet the requirements 
of sections 412(c) and (d) of the act, which state when a manufacturer 
must register with, submit to, or notify the agency about a new or 
changed infant formula, and what must be in the registration, 
submission, or notification. The agency is responding to these requests 
in this proposal. The agency is providing this information not only in 
response to these inquiries but also to facilitate more consistent 
registrations, submissions, and notifications. The lack of consistency 
in the format and content of registrations, submissions, and 
notifications has caused inefficiencies and delays in the agency's 
review. Accordingly, the agency is proposing to establish a consistent 
format and content for infant formula registrations, submissions, and 
notifications.
    Within the past year, FDA has investigated a number of instances in 
which infant formula manufactured in the United States has been 
diverted from normal distribution channels and relabeled, sometimes 
with counterfeit labels for the same brand of infant formula but in 
other instances with counterfeit labels for different formulations. 
Infant formula bearing counterfeit labels is a potentially serious 
public health problem. It could cause infant formula that is past the 
use by date to enter the marketplace if the counterfeit label bears an 
incorrect use by date. The more serious consequence of this practice, 
however, is that it could cause infants that are intolerant to certain 
infant formula ingredients to be fed an incorrect formula, with serious 
consequences to the health of the infant,

[[Page 36156]]

if an infant formula has been relabeled with an incorrect label (e.g., 
a milk-based infant formula relabeled to indicate that it is a soy-
based infant formula). Therefore, as part of this proposed regulation, 
the agency is requesting comments on new or modified procedures or 
controls that could be instituted during the labeling, packaging, or 
distribution of infant formula and that would be effective in 
preventing or reducing the potential for the diversion of infant 
formula from normal distribution channels and its relabeling with 
counterfeit labels.

III. Scope of this Document

    To implement the 1986 amendments, the agency is proposing to amend 
its regulations by adding new subparts B, D, and E to part 106 and by 
redesignating existing subparts B, C, and D as subparts C, F, and G. 
Table 1 sets out the current and proposed subpart designations.

                                                     Table 1                                                    
----------------------------------------------------------------------------------------------------------------
     Subparts                     Current regulation                            Proposed regulation             
----------------------------------------------------------------------------------------------------------------
A.................  General Provisions...........................  General Provisions.                          
B.................  Quality Control Procedures for Assuring        Current Good Manufacturing Practice.         
                     Nutrient Content of Infant Formulas.                                                       
C.................  Records and Reports..........................  Quality Control Procedures.                  
D.................  Notification Requirements....................  Conduct of Audits.                           
E.................  None.........................................  Quality Factors for Infant Formulas.         
F.................  None.........................................  Records and Reports.                         
G.................  None.........................................  Registration, Submission, and Notification   
                                                                    Requirements.                               
----------------------------------------------------------------------------------------------------------------

    The proposed regulation adds a new Sec. 107.1 and will amend 
Sec. 107.10(a)(2) by requiring that ``any nutrient added by the 
manufacturer'' be listed on the label. The proposed regulation amends 
Secs. 107.240 and 107.250 by changing the reference to the Division of 
Regulatory Guidance to the Division of Enforcement to reflect the 
November 1992 reorganization of CFSAN.

IV. The Proposed Regulations

A. General Provisions

    To reflect the expanded scope of the proposed regulations, FDA is 
revising the heading of part 106 to read, ``Infant Formula-Requirements 
Pertaining to Current Good Manufacturing Practice, Quality Control 
Procedures, Quality Factors, Records and Reports, and Notifications.''
1. Status and Applicability of the Regulations in Part 106
    Proposed Sec. 106.1 sets out the authority for each of the proposed 
subparts and the consequences under the act of failure to comply with 
any of the regulations in the proposed subparts. FDA is including 
proposed Sec. 106.1 because it is important for manufacturers to be 
aware of the legal consequences of failure to comply with these 
regulations, which are being issued to implement specific sections of 
the act.
2. Definitions
    The agency is proposing to amend Sec. 106.3 by adding several 
definitions that are needed to explain activities that specifically 
concern the infant formula industry. It is important whenever possible 
to maintain consistent terminology throughout the agency's regulations. 
Therefore, as described in detail below, FDA has relied, where 
possible, on existing definitions in 21 CFR parts 105, 110, and 210 in 
arriving at these proposed definitions. Other definitions were derived 
from specific provisions in the act.
    Proposed Sec. 106.3(a), (g), (h), and (p) incorporate into part 106 
the definitions for ``batch,'' ``lot,'' ``lot number, control number, 
or batch number,'' and ``representative sample'' derived from 21 CFR 
210.3(b)(2), (b)(10), (b)(11), and (b)(21), respectively. In addition 
to promoting consistency in the agency's regulations, FDA has 
tentatively determined that use of these definitions in part 106 is 
appropriate because they permit the agency to refer to the product in 
terms that reflect the fact that it is produced in bulk rather than on 
a unit-by-unit basis.
    Proposed Sec. 106.3(k), (q), and (r) incorporate into part 106 the 
definitions for ``microorganisms,'' ``shall,'' and ``should'' from 21 
CFR 110.3(i), (p), and (q), respectively. In addition to promoting 
consistency, these definitions reflect the generally recognized 
scientific or legal meaning of these terms.
    Proposed Sec. 106.3(c), (f), (j), and (n) incorporate into part 106 
the definitions for ``indicator nutrient,'' ``in-process batch,'' 
``manufacturer,'' and ``nutrient premix'' from current Sec. 106.3. The 
definition of ``manufacturer'' in proposed Sec. 106.3(j) warrants 
particular note. In the past there has been some confusion about who is 
and who is not a manufacturer of infant formula. This definition makes 
clear that a manufacturer is not only a person who combines raw 
ingredients together to produce an infant formula but also is a person 
who reconstitutes or otherwise changes the physical or chemical 
characteristics of an infant formula or who packages or labels the 
product in a container for distribution. For example, the agency is 
aware of a firm that reconstitutes powdered infant formulas and puts 
the reconstituted formula in bottles to sell to hospitals. This 
definition makes clear that this firm is a ``manufacturer.''
    Proposed Sec. 106.3(d) incorporates into part 106 the definition 
for ``infant'' from 21 CFR 105.3(e).
    In addition to the definitions derived from FDA's existing 
regulations, the agency is proposing to amend Sec. 106.3 by adding 
definitions that are derived from the definitions provided by Congress 
in the act.
    Proposed Sec. 106.3(e) and (l) incorporate into part 106 the 
definitions for ``infant formula'' and ``new infant formula'' from 
sections 201(aa) (21 U.S.C. 321(aa)) and 412(c)(2), respectively.
    Proposed Sec. 106.3(e) defines ``infant formula'' as a food that 
purports to be or is represented for special dietary use solely as a 
food for infants by reason of its simulation of human milk or its 
suitability as a complete or partial substitute for human milk. The 
phrase ``solely as a food for infants'' is somewhat ambiguous. Where 
there is an ambiguity in a statutory provision, it is appropriate to 
look to the legislative history to determine the appropriate 
interpretation. In the legislative history of the Infant Formula Act, 
whenever the words ``sole'' or ``solely'' are used, they appear in the 
context of describing infant formula as the ``sole'' or primary source 
of nutrition for infants or babies. For example, in explaining how the

[[Page 36157]]

1980 act would change existing laws, then-Congressman Gore stated: 
``First it would require that any infant formula marketed in the United 
States as the sole source of nutrition for normal babies include 
minimum amounts of all essential nutrients.'' (Ref. 7.) Congressman 
Mottl stated that the 1980 act ``is concerned with human lives at their 
most vulnerable stage. We are talking about food that may be the sole 
source of nourishment for infants.'' (Ref. 7.) This language and other 
similar language in the legislative history evidence that Congress 
intended the act to apply to any food that purports to be or that is 
represented as an infant formula, regardless of whether other possible 
uses of the product are suggested in its labeling. If the law only 
applied to foods that are represented only for use as infant formula, 
then manufacturers could easily evade the requirements of the act for 
infant formula by representing their products for a second purpose. 
Such an interpretation would be inconsistent with the remedial purposes 
of the infant formula provisions of the act.
    Proposed Sec. 106.3(b) incorporates into part 106 the definition 
for ``final-product-stage'' derived from section 412(b)(3)(E) of the 
act. FDA has modified the definition, however, by adding the phrase 
``due to processing'' at the end of the definition to clarify that the 
final-product-stage is when the infant formula ``is homogeneous and is 
not subject to further degradation due to processing'' and to 
distinguish the point in time after which the formula is subject to 
further degradation during the shelf life of the product.
    Proposed Sec. 106.3(i) incorporates into part 106 a definition of 
``major change'' that is derived from section 412(c)(2)(B) of the act, 
which states that ``* * * the term `major change' has the meaning given 
to such term in section 106.30(c)(2) of title 21, Code of Federal 
Regulations (as in effect on August 1, 1986), and guidelines issued 
thereunder'' (Ref. 8). Proposed Sec. 106.3(i) defines ``major change'' 
as it is defined in current Sec. 106.30(c)(2). It also provides a 
number of examples of infant formulas deemed to differ fundamentally in 
processing or in composition. These examples are derived from the 
guidelines that were issued by the agency and were incorporated into 
the definition of ``major change'' in section 412(c) of the act by the 
1986 amendments.
    Proposed Sec. 106.3(m) revises the definition for ``nutrient'' in 
current Sec. 106.3(d) to reflect changes to the act made by the 1986 
amendments. As stated above, the 1986 amendments moved the nutrient 
table from section 412(g) to section 412(i)(1) and moved the provision 
on promulgation of standards for nutrients from section 412(a)(2)(A) to 
section 412(i)(2). The proposed regulation references the new section 
numbers. Proposed Sec. 106.3(m) also includes the statement that 
nutrients are substances determined to be essential by the Food and 
Nutrition Board of the National Research Council or by FDA. The agency 
is including this statement in the proposed definition to provide 
consistency with Sec. 107.10(b)(5) on labeling nutrient information. 
This paragraph allows such information to include any vitamin or 
mineral in the formula, provided that the nutrient has been identified 
as essential by the National Academy of Sciences through its 
development of a recommended dietary allowance or an estimated safe and 
adequate daily dietary intake range, or the nutrient has been 
identified as essential by FDA through a Federal Register publication.
    Proposed Sec. 106.3(o) defines ``quality factors.'' The definition 
that FDA is proposing derives from the language of the act and its 
legislative history. Section 412(b)(1) of the act states that the 
Secretary shall ``establish requirements for quality factors for infant 
formulas * * *, including quality factor requirements for the nutrients 
required by subsection (i).'' House Report 96-936 (Ref. 5) states that 
quality factors ``pertain to the bioavailability of a nutrient and the 
maintenance of level or potency of nutrients during the expected shelf 
life of the product.'' The language of the act and the House report 
show that Congress intended that infant formulas marketed in the United 
States should not only be safe, and contain all of the nutrients 
required to support infant growth and health, but should provide those 
nutrients in a bioavailable form that will mean that, throughout its 
shelf life, the formula will support optimal infant growth and health.
    Thus, quality factors encompass something different than the 
analyzable nutrient content of the finished infant formula. Quality 
factor requirements not only ensure that the nutrient potency and 
biological effectiveness of a formula, as formulated, are adequate to 
support healthy growth, but also that subsequent processing, ingredient 
interactions, and time do not reduce the biological effectiveness of a 
formula. Quality factor requirements also ensure that unsafe nutrient 
``super potencies'' or by-products are not created from ingredient 
breakdowns or interactions caused by processing or time.

B. CGMP

1. Introduction
    The agency is proposing to adopt a new subpart B to implement the 
CGMP requirements in section 412(b)(2) of the act. Proposed Sec. 106.5 
is introductory. It reflects FDA's tentative view that the CGMP 
requirements set out in subpart B are the minimum necessary to ensure 
that the infant formula that is produced contains all the requisite 
nutrients and is not otherwise adulterated.
    To develop the proposed CGMP regulations, as stated above, agency 
representatives visited infant formula plants to observe the 
manufacturing practice that they employ, and the agency has solicited 
and received recommendations on CGMP from the infant formula industry 
through the Infant Formula Council (Ref. 9). The agency also is relying 
on its knowledge of industry manufacturing practices gained through 
inspections of infant formula manufacturing establishments, review of 
infant formula submissions received from industry since 1986, and 
monitoring of infant formula recalls.
    The proposed CGMP regulations also are based in part on FDA's 
existing regulations concerning CGMP for foods (21 CFR part 110) and 
for drugs (21 CFR part 211). Because infant formulas are foods, they 
should, at a minimum, be manufactured in a manner that is consistent 
with CGMP for all foods under section 402(a)(4) of the act (21 U.S.C. 
342(a)(4)). Moreover, infant formulas are often the sole source of 
nutrition for infants during a period of rapid growth and development 
and, hence, are used during a period of nutritional vulnerability. 
Thus, if the formula is to promote optimal infant health and growth, 
each batch of infant formula must provide the nutrients prescribed 
under section 412(i) of the act at the levels specified in that 
section, much like each batch of drugs must meet compositional 
requirements for active ingredients if they are to have their intended 
effect. Therefore, FDA has tentatively concluded that some of the 
manufacturing practices required of drug manufacturers are relevant to 
infant formula manufacturers.
2. Production and In-Process Control System
    Section 412(b)(2)(B)(iii) of the act states that CGMP and quality 
control procedures shall include requirements for ``in-process controls 
including, where necessary, testing required by CGMP designed to 
prevent adulteration of each batch of infant formula.'' In the past, 
manufacturers of infant formula have referred to production and in-

[[Page 36158]]

process control systems intended to ensure that required nutrients are 
included in the formula and to prevent adulteration by such terms as 
``quality control plans,'' ``standard operating procedures,'' or 
``master manufacturing procedures.'' Infant formula manufacturers also 
have investigated adopting a system, known as the ISO.9000 series, 
developed by the International Organization for Standardization (ISO).
    The agency is proposing to establish a framework in which decisions 
about the production of infant formula are left to the manufacturer but 
that charges the manufacturer with incorporating into its production 
process measures that are designed to ensure the safety and nutritional 
quality of the formula.
    For example, proposed Sec. 106.10(a) requires that there be 
sufficient personnel, qualified by training and experience, to perform 
all operations, including all required recordkeeping, in the 
manufacture, processing, packing, and holding of each infant formula 
and to supervise such operations to ensure that they are correctly and 
fully performed. This provision is a performance standard for 
determining how many employees are necessary, i.e., that there be 
enough to achieve, maintain, and document CGMP. FDA is not proposing to 
provide the specific number of employees required, the specific type of 
training that they must have, the specific task they are to perform, or 
the specific method by which records are to be kept.
    In another example, proposed Sec. 106.35(b)(4) requires that infant 
formula manufacturers ensure that automatic (mechanical or electronic) 
systems are validated before their first use to manufacture commercial 
product. However, in this provision, the agency is not stipulating any 
standards or specifications for the validation process because the 
extent of the validation that is necessary is related to the level of 
risk that each component of the system presents. These decisions about 
the validation necessary are left to the infant formula manufacturer to 
make.
    As a third example, proposed Sec. 106.91(b) requires that the 
manufacturer conduct nutrient stability testing at the beginning, 
midpoint, and end of the shelf life of the infant formula and with 
sufficient frequency to ensure that the formula complies with 
Sec. 107.100 throughout its shelf life. Because manufacturers have 
experience with the nutrient stability of the infant formula matrices 
that they produce and are in a position to determine how frequently 
testing is necessary, the agency is proposing only to require testing 
``with sufficient frequency,'' instead of specifying what frequency is 
required.
    Proposed Sec. 106.6(a) requires that infant formula manufacturers 
comply with the requirements of subpart B of part 106 by implementing a 
system of production and in-process controls that covers all stages of 
processing, from receipt and acceptance of raw materials, ingredients, 
and components through storage and distribution of finished product, 
and that is designed to ensure that all requirements of subpart B of 
part 106 are met.
    Infant formula manufacturing requires a degree of sophistication 
(e.g., in research and development, production equipment and 
procedures, and analytical equipment and methodology) that a vast 
majority of companies in the food processing industry do not have. A 
manufacturer must maintain constant control because a seemingly 
innocuous change in formulation or in a preparation method, or exposure 
to an unanticipated environmental condition, could create a health 
hazard. Moreover, infant formula manufacturers must be concerned not 
only that something is present in the formula that may adulterate that 
formula, such as a contaminant or a level of a required nutrient that 
exceeds the maximum level allowed by Sec. 107.100, but also that 
something is absent from the formula, such as the lack or 
unavailability of a required nutrient. For example, the lack of a 
nutrient or the unavailability of an added nutrient has been 
responsible for a number of documented problems that have occurred in 
infant formulas (Ref. 1). Thus, FDA has tentatively concluded that the 
use of a production and in-process control system covering all stages 
of processing is necessary to ensure that the infant formula is 
manufactured in a manner that will prevent adulteration of the infant 
formula.
    Proposed Sec. 106.6(b) requires that the production and in-process 
control system be set out in a written plan, or set of procedures, that 
is designed to ensure that the infant formula is manufactured in a 
manner that will prevent adulteration of the formula. FDA has 
tentatively concluded that requiring that the production and in-process 
control system be set out in a written plan or a set of procedures is 
necessary to provide consistency in production of different batches of 
infant formula and to facilitate the preparation of each batch of 
infant formula. Consistency is provided because the plan means that 
there is a single set of procedures established that are to be followed 
in producing the formula. The plan also facilitates preparation of the 
formula because, given the sophistication of the infant formula 
manufacturing process, a written plan to which ready and easy reference 
can be had is essential. The importance of a written plan is well-
recognized by industry. The use of a written plan or set of procedures 
for production of a batch of infant formula is already a wide-spread 
practice.
    The agency has sought to develop a basic list of items that a firm 
would need to consider in developing its plan or procedures, but the 
agency is reluctant to offer such a list at this stage of the 
rulemaking, before it has received comments on the proposed good 
manufacturing practice regulations. The agency requests comments on 
whether such a basic list, over and above the provisions of Subpart B 
itself, is possible or desirable, and if it is, what such a list should 
include.
    The agency would conceive of such a list, at a minimum, as 
consisting of a number of items. It would need to direct the 
manufacturer to establish the safeguards that it will rely upon to 
protect against the foreseeable sources of adulteration in the 
production of infant formula. It would also need to direct the 
manufacturer to establish procedures for ensuring that the 
manufacturing process functions properly. Several of the procedures 
that would have to be established to do so are defined in the proposed 
regulations, including: (1) Procedures, in accordance with proposed 
Sec. 106.35(b)(2), to calibrate, inspect, and check hardware; (2) 
specifications, in accordance with proposed Sec. 106.40(d), for the 
acceptance or rejection of ingredients, containers, and closures used 
in infant formula manufacture; (3) the master manufacturing orders in 
accordance with proposed Sec. 106.50(a)(1); and (4) testing procedures, 
under proposed Sec. 106.55(b), to ensure that powdered infant formula 
complies with the microbiological quality standards. Other items that 
would also seem to be appropriately included on such a list would be 
procedures for controlling the release of product, for ensuring its 
traceability, and for conducting GMP audits. However, FDA requests 
comments on whether these items provide an adequate checklist for the 
development of the type of written plan that is necessary under these 
proposed regulations.
    For now, FDA is leaving the specific content of the procedures that 
are in the written plan to the manufacturer's discretion. FDA requests 
comment on whether the agency should develop guidance on the content of 
any of the

[[Page 36159]]

procedures that are part of the written plan.
    Proposed Sec. 106.6(c) specifies requirements for a manufacturer's 
handling of any point, step, or stage in its production process where 
control of the process is necessary to prevent adulteration of the 
formula. These in-process control points, steps, or stages may include 
retorting or other heating steps, cooling steps, points where specific 
sanitation procedures are needed, product formulation control steps, 
points where cross contamination may occur, and steps where employee 
and environmental hygiene are necessary to prevent adulteration of the 
product.
    Proposed Sec. 106.6(c)(1) requires that infant formula 
manufacturers establish standards or specifications to be met at such 
points, steps, or stages. These standards or specifications establish 
the boundaries of safety at the point, step, or stage. Such standards 
or specifications may include, for example, upper and lower limits for 
parameters such as temperature, time, pH, visual appearance, and 
moisture level as well as chemical, nutrient, and microbiological 
specifications for raw materials. These standards or specifications can 
be set based on published or unpublished studies, on regulatory levels 
established by FDA, or on consultation with experts in infant formula 
production. As discussed in more detail below, FDA is proposing (see 
proposed Sec. 106.100(e)(3)(i)) that manufacturers make and retain a 
list of the standards and specifications that they establish under 
proposed Sec. 106.6(c)(1) including documentation of the scientific 
basis for each standard or specification. Maintaining such a list will 
mean that these standards and specifications are readily available for 
comparison to the actual values obtained in monitoring (i.e., making a 
planned sequence of observations or measurements) the production and 
in-process control system.
    Proposed Sec. 106.6(c)(2) requires that infant formula 
manufacturers monitor the points, steps, or stages in their production 
process where control is necessary to prevent adulteration of the 
infant formula. Regular monitoring of these points is necessary to 
ensure that the product meets the standards and specifications set 
under proposed Sec. 106.6(c)(1) and to ensure that any trend toward 
loss of control is quickly identified. Quick identification will mean 
that adjustments can be made to prevent a deviation from occurring, or, 
in the event that a deviation does occur, that effective corrective 
actions can be taken to remove adulterated product from the system.
    For many standards or specifications, continuous monitoring is 
possible. For example, temperature and time for a scheduled thermal 
process can be recorded continuously on temperature- recording charts. 
When it is not possible to monitor a particular point, step, or stage 
on a continuous basis, monitoring intervals need to be reliable enough 
to permit the manufacturer to determine whether the production control 
point is under control.
    Monitoring involves not only making observations at an appropriate 
frequency but also ensuring that the instruments and equipment, such as 
thermometers, temperature-recording devices, and computer software, 
that the manufacturer relies on to make its observations are accurate 
and reliable (see proposed Sec. 106.30(d)).
    Proposed Sec. 106.6(c)(3) requires that infant formula 
manufacturers establish corrective action plans for use when a standard 
or specification established in accordance with proposed 
Sec. 106.6(c)(1) is not met. FDA has tentatively concluded that this 
requirement is necessary because a manufacturer will often need to take 
corrective action quickly, and the best way to ensure that a corrective 
action is appropriate is to determine the action in advance. The 
corrective action plans should provide, for example, for the 
disposition of any infant formula or of any partially manufactured 
infant formula that was produced when a deviation was occurring.
    Proposed Sec. 106.6(c)(4) requires that infant formula 
manufacturers review the results of the monitoring required under 
proposed Sec. 106.6(c)(2). This review will reveal whether the 
monitoring is actually being done and being done correctly, and whether 
standards and specifications are being met.
    Proposed Sec. 106.6(c)(4) further requires that infant formula 
manufacturers review, and evaluate the public health significance of, 
any deviations from standards or specifications established in 
accordance with proposed Sec. 106.6(c)(1). This proposed requirement is 
necessary to ensure that products that may have been affected by a 
deviation do not enter commerce if they are likely to be unsafe. It 
also will ensure that the disruption of a manufacturer's business is 
minimized when a deviation does occur. For example, if review of 
monitoring records reveals that an ingredient premix does not contain 
the required nutrients at the required levels, the manufacturer can 
take steps to dispose of the premix before it is used in the 
manufacture of an infant formula. If the monitoring records are not 
reviewed, a product made with a deficient premix may be placed on the 
market, and a costly and embarrassing recall may be required.
    Proposed Sec. 106.6(c)(4) also requires that this review be 
conducted by an individual qualified by training and experience to 
conduct such reviews. This proposed requirement is necessary to ensure 
that the review is conducted by a person who understands the production 
and in-process control system, understands the significance of a 
processing deviation, and knows how to respond to a deviation. Such 
understanding and knowledge will ensure that the review is 
appropriately conducted, and that the response to any deviation is 
measured and appropriate.
    Proposed Sec. 106.6(c)(5) requires that infant formula 
manufacturers establish recordkeeping procedures, in accordance with 
proposed Sec. 106.100(e)(3), that ensure that compliance with the 
requirements of proposed Sec. 106.6(c) is documented. As discussed 
below in the description of the proposed revisions to subpart F of part 
106, FDA has authority to require that these records be made and 
retained under section 412(b)(4)(A)(i) of the act. FDA is proposing to 
provide a complete description of all recordkeeping requirements in 
subpart F. When applicable, FDA is including cross-references to these 
recordkeeping requirements in the regulations in subparts B, C, and D. 
These records will allow manufacturers to discern trends or to pinpoint 
the onset of a problem if a standard or specification is not being met 
at a point where control is deemed necessary to prevent adulteration, 
or if a batch of infant formula is associated with an adverse event.
3. Controls to Prevent Adulteration by Workers
    Proposed Sec. 106.10(a) requires that there be sufficient 
personnel, qualified by training and experience, to perform all 
operations, including all required recordkeeping, in the manufacture, 
processing, packing, and holding of each infant formula and to 
supervise such operations to ensure that they are correctly and fully 
performed. Proposed Sec. 106.10(a) is consistent with existing 
regulations concerning CGMP for foods (Sec. 110.10(c)) and drugs 
(Sec. 211.25). In this provision, FDA is proposing a general standard 
for determining how many employees are necessary, i.e., that there be 
enough to achieve, maintain, and document CGMP. However, FDA is leaving 
the determination of the actual number of employees necessary to the 
manufacturer's discretion.

[[Page 36160]]

    Proposed Sec. 106.10(a) also requires that such personnel be 
qualified by training and experience. Training is necessary to ensure 
that employees know how to correctly and fully perform the operations 
in question and to ensure that employees are competent to produce a 
safe and clean infant formula. The extent and frequency of training is 
left to the manufacturer's discretion.
    Proposed Sec. 106.10(b) requires that personnel working directly 
with infant formula, infant formula raw materials, infant formula 
packaging, or infant formula equipment or utensil contact surfaces 
practice good personal hygiene to protect the product against 
contamination. Proposed Sec. 106.10(b) is consistent with existing 
regulations concerning CGMP for foods (Sec. 110.10(b)) and drugs 
(Sec. 211.28(a) and (b)). FDA has tentatively concluded that it is 
necessary that these employees practice good hygiene so that they will 
not transmit disease to others in the workforce, and so that they will 
not transmit filth or pathogenic microorganisms to the infant formula.
    In addition, proposed Sec. 106.10(b) enumerates the basic elements 
of good personal hygiene. Proposed Sec. 106.10(b)(1) lists clean outer 
garments and protective apparel as one element. To be ``clean,'' 
clothing must be free of filth or microorganisms that may contaminate 
the infant formula. Protective apparel, such as head, face, hand, and 
arm coverings, will help to ensure that the infant formula is protected 
from contaminants such as hair.
    Proposed Sec. 106.10(b)(2) states that good personal hygiene 
includes workers washing their hands thoroughly in a hand washing 
facility with soap and running water at a suitable temperature before 
starting work, after each absence from the work station, and at any 
other time when hands may become soiled or contaminated. Filth and 
pathogenic microorganisms can be brought into the processing 
environment on the employee's hands from outside areas, restrooms, 
contaminated raw materials, waste or waste receptacles, and other 
insanitary objects (Refs. 10, 11, and 12). FDA has tentatively 
concluded that requiring workers to practice good personal hygiene by 
washing their hands at the times specified will help to prevent the 
introduction of this type of contamination into infant formula.
    Proposed Sec. 106.10(c) requires that any person who reports that 
he or she has, or appears by medical examination or supervisory 
observation to have, an illness, open lesion, including boils, sores, 
or infected wounds, or any other source of microbial contamination that 
creates a reasonable possibility that the safety of the formula may be 
adversely affected, be excluded from direct contact with ingredients, 
containers, closures, in-process materials, equipment, utensils, and 
infant formula product until the condition is corrected or determined 
by competent medical personnel not to jeopardize the safety of the 
infant formula. Proposed Sec. 106.10(c) is consistent with existing 
regulations concerning CGMP for foods (Sec. 110.10(a)) and drugs 
(Sec. 211.28(d)). Employees can transmit the organisms responsible for 
diseases, such as salmonellosis, shigellosis, and hepatitis, to the 
infant formula. Additionally, open sores, boils, or infected wounds 
present the potential for contamination of the infant formula with such 
pathogenic microorganisms as Staphylococcus aureus (Refs. 14 and 15). 
Thus, proposed Sec. 106.10(c) will exclude employees who carry 
potential microbial contamination that may adversely affect the safety 
of the formula from direct contact with the infant formula and from 
direct contact with materials and surfaces that come in contact with 
the infant formula and thus will minimize the potential for employees 
to transmit microorganisms to the infant formula that may cause the 
infant formula to pose a health hazard to the infant.
4. Controls to Prevent Adulteration Caused by Facilities
    Proposed Sec. 106.20(a) requires that buildings used in the 
manufacture, processing, packing, or holding of infant formula be 
maintained in a clean and sanitary condition. This proposed requirement 
is necessary to prevent contamination of the infant formula. It is 
consistent with FDA's existing regulations concerning CGMP for foods 
(Secs. 110.20(b) and 110.35(a)) and drugs (Sec. 211.42). Trash, litter, 
and waste must be disposed of to avoid creating conditions that attract 
and harbor potentially pathogenic microorganisms and attract and harbor 
pests, such as rodents or insects. Such pests can carry a variety of 
human disease agents, including microorganisms that are potentially 
pathogenic in infants, and introduce them into the manufacturing 
environment (Refs. 10 and 12). They are also sources of feces and hair 
that can contaminate infant formula.
    Proposed Sec. 106.20(a) also requires that buildings used in the 
manufacture of infant formula have space for the separation of 
incompatible operations, such as the handling of raw materials, the 
manufacture of the product, and packaging and labeling operations. If 
raw materials are not separated from the site of product manufacture, 
there is a significant possibility that they will be used in infant 
formula manufacture before they have been tested and found acceptable 
for use in infant formula. Therefore, FDA has tentatively concluded 
that the separation of incompatible operations is necessary to ensure 
that infant formula is manufactured in a manner designed to prevent 
adulteration. The proposed requirement that incompatible operations be 
separated is consistent with FDA's existing regulations concerning CGMP 
for foods (Sec. 110.20(b)(2)) and drugs (Sec. 211.42(c)) and is 
consistent with the recommendations made to FDA by the Infant Formula 
Council (Ref. 9).
    Proposed Sec. 106.20(b) requires separate holding areas to protect 
against mixups that could lead to contamination of infant formula. 
Failure to separate raw materials or in-process materials that have not 
been released, or that have been rejected but not disposed of, from 
those that have been released creates the potential for the use of 
ingredients that do not meet the applicable specifications and thereby 
can lead to the production of finished infant formula that is 
adulterated. Similar types of problems can develop if final product 
that has not been released, or that has been rejected but not disposed 
of, is not separated from final product that has been released. 
Proposed Sec. 106.20(b) is consistent with FDA's existing regulations 
concerning CGMP for drugs (Sec. 211.42(c)).
    Proposed Sec. 106.20(c) defines a standard for adequate lighting 
and allows the manufacturer to exercise discretion in determining the 
precise level of lighting that is sufficient to meet that standard. 
Adequate lighting is important. Inadequate lighting may make it 
difficult to read a label or an instrument, and as a result incorrect 
ingredients may be used in infant formula production, or instruments 
may be read incorrectly, which increases the risk of producing an 
adulterated infant formula.
    Proposed Sec. 106.20(c) also requires that any lighting fixtures 
directly over or adjacent to exposed raw materials, in- process 
materials, or bulk (unpackaged) finished product be protected to 
prevent glass from contaminating the product in the event of breakage. 
Glass in an infant formula may be a safety hazard and would render the 
formula adulterated (Ref. 14). Proposed Sec. 106.20(c) is consistent 
with FDA's existing regulations concerning CGMP's for food 
(Sec. 110.20(b)(5)) and drugs (Sec. 211.44).
    FDA is proposing a requirement in Sec. 106.20(d) for air filtration 
systems to improve air quality in production areas

[[Page 36161]]

and thus reduce the potential for contamination by air-borne sources 
(Ref. 15). This proposed requirement is consistent with FDA's existing 
regulations concerning CGMP for drugs (Sec. 211.46(c)).
    Proposed new requirements in Sec. 106.20(e) protect against the 
contamination of infant formula by pest control agents and cleaning 
agents. The agency recognizes that these agents are needed in infant 
formula facilities. However, because many of them are toxic, they must 
be handled and stored in a manner that prevents contamination of the 
infant formula. Proposed Sec. 106.20(e) is consistent with FDA's 
existing regulations concerning CGMP for food (Sec. 110.35(b)(2)) and 
drugs (Sec. 211.56(c)).
    Proposed Sec. 106.20(f)(1) states that potable water used in the 
manufacturer of infant formula must meet the Environmental Protection 
Agency's (EPA's) Primary Drinking Water Regulations (40 CFR part 141) 
(with the one exception that the fluoride level be as low as possible, 
as discussed below). This proposed regulation is consistent with FDA's 
existing regulations concerning CGMP for drugs (Sec. 211.48(a)).
    The Safe Drinking Water Act gives EPA the responsibility for 
establishing standards for public drinking water. Therefore, FDA is 
proposing to use EPA's standards for water used in the production of 
infant formulas. Application of these standards will ensure that the 
water used in infant formula is safe. The agency is proposing to 
require that water from both municipal sources and the firm's own wells 
meet these standards.
    The safety and sanitary quality of water from public water systems 
is generally ensured through public water treatment, chlorination, or 
monitoring and control by local health authorities. Private sources of 
water, however, particularly surface waters or water from shallow 
wells, may be subject to microbiological, chemical, or radiological 
contamination attributable to the source itself or to surface 
contamination at the well head or intake. Private sources are also 
frequently untreated or minimally treated. Thus, under the proposed 
regulation, when a manufacturer uses a private source of water, it will 
need to take steps to ensure that the water is safe and sanitary. These 
steps may include ensuring that the well design has been approved by 
the local health authority, ensuring that the well meets coliform test 
standards, performing periodic inspections of the sanitary condition of 
the well head and source intake, and performing and monitoring 
appropriate water treatment procedures, including filtration, 
sedimentation, and chlorination. The type and frequency of controls 
exercised by the manufacturer will be based upon the type of source 
water and its historic safety and sanitary quality.
    Proposed Sec. 106.20(f)(1) makes one exception to the use of EPA 
standards for drinking water. On April 2, 1986, EPA issued a maximum 
contaminant level (MCL) for fluoride in drinking water of 4 milligrams 
per liter (mg/L) (51 FR 11396) and reaffirmed this level on December 
29, 1993 (58 FR 68826). The National Academy of Sciences (NAS) 
recommends 0.1 to 0.5 mg/day as the safe and adequate intake for 
infants from 0 to 6 months of age. Mottling of teeth in children has 
been observed at 2 to 8 milligrams/kilogram (mg/kg) concentration of 
fluoride in diet and drinking water (Ref. 16). Thus, if 4 mg of 
fluoride/L of water was allowed in the water used in infant formula 
manufacture, infants consuming ready-to-feed infant formula could 
receive enough fluoride to adversely affect their teeth. Currently, no 
infant formulas are manufactured with fluoridated water (Ref. 17), so 
that the pediatrician or other health care provider is able to decide 
whether a fluoride supplement is appropriate for formula-fed infants, 
principally by considering whether the formula was diluted with 
fluoridated water (Ref. 18).
    NAS has established a safe and adequate daily dietary intake of 
fluoride for infants (Ref. 19). The agency is considering proposing to 
revise the infant formula nutrient requirements in Sec. 107.100 to 
include fluoride and other nutrients that NAS has determined are 
essential for infants. FDA will consider fluoride levels for infant 
formulas at that time. FDA has tentatively concluded that, until it has 
revised the levels of required nutrients, manufacturers should continue 
their practice of not using fluoridated water in the manufacture of 
infant formula.
    Proposed Sec. 106.20(f)(1) also requires that the water be supplied 
under continuous positive pressure in a plumbing system that is free of 
defects that could contaminate an infant formula. FDA has tentatively 
concluded that this requirement is necessary to ensure that all potable 
water coming into the plant is not adversely affected by the in-plant 
plumbing. Contaminated water can serve as a vehicle for contamination 
of infant formula, both when used as an ingredient in the infant 
formula and when allowed to enter the product indirectly, as can occur, 
for example, when water is used to cool the product after retorting. 
Thus, FDA tentatively concludes that it is appropriate to include this 
positive requirement in this regulation.
    Proposed Sec. 106.20(f)(2), which sets forth requirements for 
testing representative samples of potable water used in infant formula 
manufacturing, is necessary to provide assurance that the water used in 
infant formula manufacturing meets EPA's standards. Proposed 
Sec. 106.20(f)(3) requires that manufacturers conduct these tests with 
appropriate frequency. The regulation allows manufacturers some 
discretion in determining the testing frequency necessary to ensure 
that EPA standards are met, but it requires a minimum frequency of 
testing for certain contaminants (i.e., chemical contaminants, 
radiological contaminants, and bacteriological contaminants). FDA is 
basing these proposed minimum frequencies on those adopted by EPA for 
primary drinking water. This frequency of testing is consistent with 
FDA's own regulations concerning processing and bottling of bottled 
drinking water (Sec. 129.35(a)(3)).
    Proposed Sec. 106.20(f)(4) requires that manufacturers make and 
retain records of the frequency and the results of the testing that 
they do on the water used in the production of infant formula. These 
records will document that the manufacturer is complying with the 
potable water testing requirements of Sec. 106.20(f)(2) and (f)(3), and 
that the water complies with EPA standards. They will identify any 
trend toward loss of compliance with these standards, so that the 
manufacturer can take corrective actions before the water becomes 
inappropriate for use in infant formula. As discussed below in the 
description of the proposed revisions to subpart F, FDA has authority 
to require the creation and retention of these records under section 
412(b)(4)(A)(i) of the act.
    In proposed Sec. 106.20(g), FDA sets out requirements regarding 
piping systems to prevent a source of contamination (i.e., waste water) 
from coming in contact with the infant formula. Cross connections could 
allow back siphonage into a potable system from a nonpotable system 
under negative pressure conditions and thus could result in the 
chemical or microbiological contamination of the potable water system 
(Ref. 20). Proposed Sec. 106.20(g) is consistent with FDA's regulations 
concerning CGMP for food (Sec. 110.37(b)(5)) and drugs 
(Sec. 211.48(b)).
    Proposed Sec. 106.20(h) requires that steam that comes in direct 
contact with infant formula be safe and free of rust

[[Page 36162]]

and other particulate matter that could contaminate the formula. Steam 
comes in direct contact with infant formula when the steam is injected 
into the head space of a can of infant formula to create a vacuum. 
Thus, this proposed requirement is necessary to ensure that the steam 
does not adulterate the infant formula.
    Proposed Sec. 106.20(h) also requires that boiler water additives 
in the steam meet safety standards set forth in FDA regulations at 21 
CFR 173.310 which lists boiler water additives that may be safely used 
in the preparation of steam that will contact food and the conditions 
for the safe use of those boiler water additives. This proposed 
requirement is necessary because boiler water additives dissolve in 
water and can be carried over as a residue in the steam. A proposed 
requirement that boiler water additives in the steam comply with 
Sec. 173.310 will ensure that any residue is safe to come in contact 
with the infant formula.
    Proposed Sec. 106.20(i) requires that each infant formula 
manufacturing site provide its employees with readily accessible toilet 
and hand washing facilities. This proposed requirement is consistent 
with good sanitary practice common to all food-processing facilities 
and is consistent with FDA's CGMP regulations for foods (Sec. 110.37(d) 
and (e)) and drugs (Sec. 211.52). The requirement is also a necessary 
adjunct to the requirement in proposed Sec. 106.10(b)(2) that employees 
wash their hands before starting work, after each absence from the work 
station, and at any other time when the hands may become soiled or 
contaminated. Hand-washing facilities are not likely to be used in an 
appropriate manner by employees if the facilities are not conveniently 
located.
    Proposed Sec. 106.20(i) also requires that these facilities be 
equipped with hot and cold water, ordinary soap or detergent, and 
single-service towels to ensure that microbiological contamination does 
not occur through the repeated use of the same towel by several 
individuals.
    In addition, proposed Sec. 106.20(i) requires that toilet 
facilities be maintained in good repair and in a sanitary condition at 
all times, and that these facilities provide for proper disposal of 
sewage, so that the processing environment is protected against 
pathogenic microorganisms shed in fecal material. Restroom floors and 
the grounds around the processing facility can become contaminated with 
pathogens if fecal material is not removed by an adequate sewage 
system. Foot traffic over the affected areas can introduce pathogens 
into the processing room and cause product contamination. Insanitary 
toilet facilities can also increase the potential for contamination of 
employees' hands and, ultimately, of the product itself (Refs. 10 and 
11). Proposed Sec. 106.20(i) further protects against potential 
microbiological contamination by setting forth requirements for the 
positioning of toilet facility doors.
5. Controls to Prevent Adulteration Caused by Equipment or Utensils
    Equipment used in infant formula manufacture, packaging, or holding 
that is of an inappropriate design or an inadequate size, or that is 
installed improperly, can result in a variety of problems. For example, 
a mixer for the blending of powdered ingredients will not properly 
perform its function if the blade is too small relative to the size of 
the mixer, or if the mixer blade or auger is not properly positioned in 
the inside of the mixer. Such a mixer may produce infant formula that 
is not uniform in composition throughout a batch and that is, 
consequently, adulterated because the required nutrients are not 
provided at the required levels throughout the batch.
    Installing equipment in a manner that will facilitate its cleaning 
and maintenance is also important in preventing adulteration. Equipment 
that is not properly cleaned can be the source of contaminants that 
adulterate the infant formula. Equipment that is not properly 
maintained can result in a variety of problems. For example, improper 
maintenance of equipment such as a mixer may result in inadequate 
compositional uniformity in a batch of formula. Improper maintenance of 
equipment used to measure a parameter such as temperature may result in 
the processing of the infant formula at a temperature that can 
adversely affect the product. In either case, the product would be 
adulterated. Design and installation of equipment also needs to be 
checked when the equipment is modified or repaired to ensure that the 
equipment is still designed and installed to function as intended as 
part of the manufacturing process. Thus, proposed Sec. 106.30(a) 
requires that equipment be appropriately designed and installed. This 
proposed requirement is consistent with FDA's CGMP regulations for 
foods (Sec. 110.40(a)) and drugs (Sec. 211.63).
    If a food-contact surface is constructed of toxic material, the 
product may be directly contaminated with that material (Ref. 11). 
Therefore, FDA is proposing to require in Sec. 106.30(b) that equipment 
and utensils be made of materials that are not reactive or absorptive, 
so that the equipment and utensil materials do not contaminate the 
infant formula and cause it to be adulterated. Proposed Sec. 106.30(b) 
also requires that such equipment and utensils be designed to be easily 
cleanable because they can be vehicles for microbial contamination of 
both raw and finished products. Utensils, equipment, and other food-
contact surfaces that are made of corrosive material, or that contain 
breaks, pits, cuts, or grooves, are difficult to clean because the 
pores and crevices shield the microorganisms from the action of 
cleaning and sanitizing agents (Ref. 21). In addition proposed 
Sec. 106.30(b) requires that equipment and utensils be designed to 
withstand the environment in which they are used. This requirement will 
ensure that equipment and utensils are constructed of materials that 
will not corrode or undergo other types of chemical or physical 
degeneration resulting from their use in infant formula production. 
Degeneration of the equipment and utensils may introduce contaminants 
into the formula and thereby lead to adulteration. Surfaces that are 
not adequately cleaned and sanitized can be a source of filth, an 
attractant for vermin, and a reservoir for microorganisms.
    Proposed Sec. 106.30(b) requires regular, effective cleaning and 
sanitizing of all food-contact surfaces to minimize the probability of 
contamination of the infant formula (Ref. 21) and prescribes 
requirements for effective sanitizing agents. An effective sanitizing 
agent is one that has a good bactericidal effect on the types of 
microorganisms normally present in the plant environment and that is 
safe, stable, and convenient for use (Ref. 22). Sanitizing agents are 
indirect food additives and must be used in accordance with 21 CFR 
178.1010, which prescribes their conditions of safe use. Examples of 
sanitizing agents that comply with Sec. 178.1010 include hypochlorites, 
iodophors, and quaternary ammonium compounds. However, sanitizers can 
achieve their intended effect only if they are applied to a surface 
that has been thoroughly cleaned, and if they are applied at a proper 
concentration (Ref. 22).
    Thus, it is important that effective cleaning compounds be used. An 
effective cleaning compound is one that will lower the surface tension 
of water so that spills can be lifted and flushed away (Ref. 23). 
Ordinary soap has a limited ability to solubilize fats, oils, and 
proteins, and inorganic alkaline

[[Page 36163]]

detergents can dissolve food solids such as fats and proteins, but 
mineral deposits will frequently require the use of acid cleaners (Ref. 
23).
    In order to ensure that infant formula is not contaminated with 
unsafe substances that are a part of the manufacturing process, FDA is 
proposing requirements in Sec. 106.30(c) regarding substances necessary 
for the operation of equipment, such as lubricants or coolants.
    Proposed Sec. 106.30(d)(1) sets forth requirements for maintaining 
the accuracy of instruments, since an instrument that is not easily 
read, or that is not properly calibrated, may not provide accurate 
measurements. If an instrument is not properly maintained, it may not 
be reliable over time, and the readings obtained from it may lead to 
adulteration of the infant formula during processing. This proposed 
regulation also requires that such instruments be sufficient in number 
for their intended use. For example, if the temperature of a large 
piece of equipment needs to be monitored, several temperature-
indicating devices may be needed to accurately monitor the temperature 
in all parts of the equipment. Also, instruments and controls must be 
tested for accuracy (i.e., calibrated) against a known reference 
standard before first use and at routine intervals thereafter, as 
specified in writing by the manufacturer of the instrument or control, 
or as otherwise deemed necessary to ensure the accuracy of the 
instrument. FDA has tentatively concluded that this requirement is 
necessary because equipment used to manufacture infant formula must 
operate properly to ensure production of a safe, uniform product with a 
consistent nutrient content throughout a lot or a batch.
    The accuracy of an instrument is the degree to which it produces a 
correct result. The instruments used to measure parameters such as 
temperature or pressure at points where control is deemed necessary to 
prevent adulteration must reflect the true measurement so that, for 
example, a manufacturer can have confidence that when a thermometer 
indicates that the temperature is 240  deg.F, the temperature really is 
240  deg.F. FDA's experience is that calibration of the instrument 
using a reference standard is the most reliable method to ensure 
accuracy. FDA is proposing to require that this test for accuracy be 
done before first use to provide assurance that the instruments and 
controls will perform as intended and at routine intervals afterward to 
ensure that the instruments and controls continue to perform as 
intended.
    Reliability is the instrument's accuracy over time. The reliability 
of the instrument will determine the length of time that it can be used 
before it begins to lose accuracy. The manufacturer of the instrument 
is in the best position to establish how frequently recalibration is 
needed because that manufacturer is responsible for putting together 
the technology by which the instrument operates. However, if the infant 
formula manufacturer's experience with the instrument demonstrates that 
the instrument needs to be calibrated more frequently than the 
instrument manufacturer suggests, FDA has tentatively concluded that 
the infant formula manufacturer must act on its own experience with the 
instrument and calibrate it as often as necessary to ensure the 
accuracy of the instrument.
    Proposed Sec. 106.30(d)(1) further requires that the known 
reference standard be certified for accuracy at routine intervals 
specified in writing by the manufacturer of the instrument, or as 
otherwise deemed necessary. Known reference standard devices are 
accompanied by certificates of accuracy, but these certificates do not 
preclude the possibility that these instruments will go out of 
calibration. Just as a calibration routine needs to be established for 
the process instrumentation, a recertification of the known reference 
standard needs to be established in accordance with the equipment 
manufacturer's recommendations. For example, the length of time that a 
certified thermometer can be considered reliable will depend on the 
materials used in its manufacture, the degree of control exercised in 
its manufacture, and its use, as would be the case for the indicating 
thermometer used in the production line. The accuracy of a calibrated 
thermometer is only going to be as good as the accuracy of the known 
reference standard that is used during its calibration.
    Proposed Sec. 106.30(d)(1) also requires that manufacturers make 
and retain records of accuracy checks in accordance with the provisions 
of proposed Sec. 106.100(f)(2). As discussed below in the description 
of the proposed revisions to subpart F of part 106, FDA has authority 
to require these records under section 412(b)(4)(A)(i) of the act. 
These records will enable the manufacturer to establish the historical 
performance of the instrument to determine whether the calibration 
schedule is sufficient to ensure the accuracy of the instrument and 
will provide information on when and how the instruments were 
calibrated to assist the manufacturer in identifying the cause of a 
problem that may arise with a batch of infant formula.
    Proposed Sec. 106.30(d)(2) requires that instruments and controls 
that cannot be adjusted to agree with the reference standard be 
repaired or replaced. FDA is proposing this requirement because an 
instrument or control cannot be trusted for use in infant formula 
production if it cannot be adjusted to agree with the reference 
standard. Adjustments made to reach agreement with a known accurate or 
reference standard must also be done in accordance with any adjustment 
range limitations specified by the vender of the instrument.
    Proposed Sec. 106.30(d)(3) provides that if calibration of an 
instrument (testing for accuracy against a known reference standard) 
shows that a specification or standard has not been met at a point 
where control is deemed necessary to prevent adulteration, a written 
evaluation must be made of all affected product and of any actions that 
need to be taken. FDA has tentatively concluded that this written 
evaluation is necessary because if an instrument has been giving 
inaccurate readings, all infant formula produced subject to such 
inaccuracies must be identified and evaluated for the possibility that 
the inaccuracies resulted in the production of adulterated formula. If 
the manufacturer determines that adulterated formula has been produced, 
the firm must decide what actions, if any, need to be taken to prevent 
such formula from reaching infants.
    FDA is also requiring that this written evaluation needs to be 
maintained in the firm's records. FDA tentatively concludes that this 
record is necessary to demonstrate that the firm has complied with 
CGMP. As discussed below in the description of the proposed revisions 
to subpart F of part 106, FDA has authority to require that these 
records be retained under section 412(b)(4)(A)(i) of the act.
    Proposed Sec. 106.30(e)(1) requires that the temperature in cold 
storage compartments used to store raw materials, in-process materials, 
or final product, as well as the temperature of thermal processing 
equipment used at points where temperature control is necessary to 
prevent adulteration, be monitored with such frequency as is necessary 
to ensure that temperature control is maintained. The frequency of the 
monitoring is left to the manufacturer to determine. Growth of 
microorganisms can occur and cause spoilage if materials that should be 
kept in cold storage compartments are not maintained at the proper 
temperature. Infant formula may also be adulterated if thermal 
processing equipment is not

[[Page 36164]]

operated at the proper temperature, and the final liquid infant formula 
product is not commercially sterile. Therefore, FDA tentatively 
concludes that their requirement is appropriate.
    In addition, FDA is proposing that a temperature of 40  deg.F (4.4 
deg.C) is appropriate in cold storage compartments to minimize the 
growth of pathogens (Ref. 24) and the deterioration of liquid 
ingredients, nutrients, and the formulated product before canning 
(proposed Sec. 106.30(e)(2)).
    Proposed Sec. 106.30(e)(3)(i) requires that cold storage 
compartments and thermal processing equipment be equipped with easily 
readable, accurate temperature-indicating devices. These devices are 
necessary to ensure that the manufacturer can monitor the temperatures 
where materials are stored or where product is processed. Proposed 
Sec. 106.30(e)(3)(ii) requires that thermal processing equipment be 
equipped with temperature-recording devices that reflect the true 
temperature on a continuing basis, so that the manufacturer will be 
able to determine whether the product was thermally processed at a 
minimum temperature for an appropriate period of time. Two factors, 
temperature and time, are relevant in ensuring that thermal processing 
is conducted in a manner that will produce commercially sterile infant 
formula after retorting. Thus, recording the temperature that is 
maintained during the time period used will show whether the thermal 
process is conducted properly.
    Proposed Sec. 106.30(e)(3)(ii) also requires that cold storage 
compartments be equipped with either a temperature-recording device 
that will reflect the true temperature within the compartment on a 
continuing basis, or a high-temperature alarm or a maximum-indicating 
thermometer that has been verified to function properly. These 
temperature records will show whether the materials were stored at an 
appropriate temperature to minimize the growth of pathogens and the 
deterioration of ingredients and formulated product. If the 
manufacturer does not wish to equip cold storage compartments with such 
temperature- recording devices, FDA is proposing to require that it 
maintain a temperature log in which the temperature in the compartment 
is noted with such frequency as is necessary to achieve control. The 
agency is leaving it to the manufacturer's discretion to determine what 
frequency of temperature notation is necessary to achieve control.
    The agency has tentatively concluded that it is not necessary for 
the manufacturer to record the temperature of the cold storage 
compartment on a continuous basis as long as the manufacturer can 
determine that the temperature of the cold storage compartment has gone 
above 40  deg.F. A high-temperature alarm set to go off when the cold 
storage compartment goes above 40  deg.F will allow the manufacturer to 
make this determination. Likewise, a maximum-indicating thermometer 
will remain at the highest temperature that it ever reaches. If the 
maximum indicating thermometer indicates a temperature above 40  deg.F, 
the infant formula manufacturer must assume that the temperature has 
been above 40  deg.F since the last check of the thermometer. Thus, FDA 
has tentatively concluded that either a high-temperature alarm or a 
maximum-indicating thermometer are acceptable alternatives for 
determining whether the cold storage compartment has gone above 40 
deg.F.
    In some cases, the actual location of the sensors may be an 
important factor in ensuring the accurate representation of 
temperature. For example, one sensor located at the end of a large 
piece of thermal processing equipment may not accurately represent the 
temperature in the whole piece of equipment. In addition, these 
temperature devices must often be read under less than ideal plant 
conditions, so they should be installed in a location that facilitates 
easy reading. Temperature-recording devices can be easily jarred and 
rendered inaccurate. They can be recalibrated against a reference 
temperature-indicating device (e.g., a thermometer) quite easily, 
however. Manufacturers should do so at least at the beginning and end 
of each production day in order to determine whether the instrument was 
accurate throughout the day's production. For thermal processing 
equipment used to produce commercially sterile liquid infant formula, 
the mandatory and recommended procedures of 21 CFR part 113 apply.
    FDA is also proposing that manufacturers make and retain records, 
in accordance with the provisions of proposed Sec. 106.100(f)(3), of 
the temperatures indicated or recorded by these devices (see 
Sec. 106.30(e)(3)). As discussed below in the description of the 
proposed revisions to subpart F of part 106, FDA has authority to 
require these records under section 412(b)(4)(A)(i) of the act. They 
are needed to show that the thermal processing equipment or cold 
storage compartments are being maintained at the correct temperatures 
to prevent adulteration of the product. They also will enable the 
manufacturer to identify trends in temperature fluctuations that can 
signal the need to perform nonscheduled maintenance.
    Proposed Sec. 106.30(e)(4) requires that for thermal processing, 
the temperature-recording device not read higher than the calibrated 
temperature-indicating device because it is important to ensure that 
the infant formula is processed at a minimum temperature for a 
continual period of time. A temperature-recording device reading higher 
than the reference temperature-indicating device for thermal processing 
equipment would show that the product had been processed at a 
temperature higher than the true processing temperature. Because 
thermal processing is used to destroy microorganisms, a temperature- 
recording device reading higher than the true processing temperature 
may mean that the product has not been processed at a temperature that 
is high enough to destroy all microorganisms.
    For cold storage compartments, the temperature-recording device 
must not read lower than the temperature-indicating device because when 
raw materials, in-process materials, or finished product must be stored 
at a cold temperature, it is important to ensure that the infant 
formula was not exposed to a temperature above the maximum temperature. 
A temperature-recording device reading lower than the reference 
temperature-indicating device for cold storage equipment would show the 
materials in the compartment as having been held at a lower temperature 
than the true temperature. Because cold storage is used to prevent 
microbiological growth, a temperature-recording device reading lower 
than the reference temperature-indicating device would mean that the 
material was actually being stored at a higher temperature than the 
recorded temperature, and that, as a result, microbial growth may have 
occurred.
    Proposed Sec. 106.30(f) requires that all equipment and utensils 
used in the manufacture of infant formula be cleaned, sanitized, and 
maintained at regular intervals to prevent adulteration of the infant 
formula. Any equipment or utensil that is not cleaned and maintained 
properly can be a source of contamination. FDA is therefore proposing 
to require that cleaning, sanitizing, and maintaining be done at 
regular intervals. The details of sanitation procedures e.g., equipment 
cleaning, can differ from plant to plant depending upon the type of 
operation and other conditions. In one plant, it may be necessary to 
disassemble all or part of the equipment to clean it. In other plants, 
breaking down the

[[Page 36165]]

equipment may not be necessary. Likewise, different cleaning compounds 
may be needed from one plant to another to solve specialized problems 
such as buildups of mineral deposits. Each manufacturer should study 
its own plant and develop a procedure that is tailored to that plant's 
needs and circumstances.
    FDA considers that cleaning, sanitizing, and maintaining equipment 
and utensils is so important for ensuring that adulterated infant 
formula is not produced that it is proposing to require that the 
cleaning, sanitizing, and maintenance be checked for satisfactory 
completion by an individual qualified to conduct such a review. Such an 
individual will understand the importance of ensuring that cleaning, 
sanitizing, and maintenance is properly done, so that equipment and 
utensils do not contribute to the adulteration of the infant formula. 
Also, the agency has tentatively concluded that this requirement will 
ensure that there is accountability for proper performance of this 
function.
    In addition, proposed Sec. 106.30(f) requires that manufacturers 
make and retain records on equipment cleaning, sanitizing, and 
maintenance in accordance with proposed Sec. 106.100(f)(4). As 
discussed below in the description of the proposed revisions to subpart 
F, FDA has authority to require these records under section 
412(b)(4)(A)(i) of the act. These records will document when the 
cleaning, sanitizing, and maintenance of equipment occurs and will 
allow the manufacturer to trace all formula that may be affected if 
cleaning, sanitizing, or maintenance is not properly performed.
    In order to ensure that compressed air or other gases will not 
contaminate the infant formula with unlawful indirect food additives or 
other chemical, physical, or microbiological contaminants, FDA is 
proposing to require in Sec. 106.30(g) that they be appropriately 
treated. Air or other gases that are not properly treated and filtered, 
or air that is not of the proper purity, can introduce contaminants 
into the infant formula that may render it adulterated. Also, 
compressed gases can be contaminated with oil from the compressor or 
with filth or microbiological contaminants from the compression, 
storage, or distribution equipment. Filtration at the air intake and 
after compression, storage, and distribution is an effective means of 
reducing the risk that such contaminants will enter the gases and, 
thereby, the food. Therefore, FDA is also proposing in Sec. 106.30(g) 
to require the use of a filter when compressed gases are used at 
product filling machines to replace air removed from the headspace of 
containers. The filter will prevent contaminants from entering the 
infant formula during that operation (Ref. 25).
6. Controls to Prevent Adulteration Due to Automatic (Mechanical or 
Electronic) Equipment
    Manufacturers of infant formula are increasingly relying on 
automatic equipment (including mechanical and electronic equipment) in 
production and quality control. In some cases, manufacturers are 
replacing manually initiated processing procedures with automated 
process control systems to ensure proper formulation (addition of 
ingredients and premixes), mixing, or processing of an infant formula 
or to test a batch of infant formula. Such automated process control 
systems frequently consist of a computer or system of computers that 
controls many or all stages of production, in-process sampling, and 
testing. In other cases, manufacturers are relying on programmable 
equipment (such as an autoanalyzer) to perform a critical function, 
such as testing a batch of infant formula to ensure that the batch 
meets the nutrient requirements of the act. In all cases, it is 
important that such systems and equipment function as expected to 
ensure that the infant formula contains the required nutrients at the 
required levels and is manufactured according to the CGMP and quality 
control procedures prescribed under section 412(b)(2) of the act and 
therefore is not adulterated under section 412(a)(1) or (a)(3) of the 
act.
    FDA is proposing to define ``hardware,'' ``software,'' ``system,'' 
and ``validation'' in Sec. 106.35 because the use of these terms will 
simplify the language of the proposed regulations and will clarify 
which sections of the proposed regulations apply to hardware only, to 
software only, or to systems consisting of both hardware and software.
    The definition of ``hardware'' in proposed Sec. 106.35(a)(1) is 
based on common usage of the term and makes clear that the regulations 
in proposed Sec. 106.35 apply to all automatic equipment, whether the 
equipment is mechanical or electronic in nature. Proposed 
Sec. 106.35(a)(1) also makes clear that electronic equipment includes, 
but is not limited to, computers. This definition of ``hardware'' 
distinguishes those elements of equipment that have a physical form 
from the elements considered to be intellectual property that may be 
encoded on a physical element such as a diskette, tape, or 
microprocessing chip.
    Software may be developed by an infant formula manufacturer, by a 
manufacturer of equipment purchased by the infant formula manufacturer, 
or by a third party vendor (such as the vendor of a computer operating 
system). The definition of ``software'' in proposed Sec. 106.35(a)(2) 
derives from the ISO International Guideline ISO-9000-3 1 (Ref. 
26) and the Institute for Electrical and Electronics Engineers, Inc. 
(IEEE) Standard 610-12-1990 2 (Ref. 27) and is consistent with the 
definition of software in FDA's ``Glossary of Computerized Systems and 
Software Development Terminology (Ref. 28). FDA is proposing to 
incorporate this definition into the agency's infant formula 
regulations because the definition is derived from internationally 
accepted definitions, includes documentation, applies to the operation 
of all types of hardware (rather than the narrowly defined ``data 
processing system'' or ``computer system'' included in the definitions 
from the ISO and IEEE, respectively), and is consistent with current 
FDA terminology. Software documentation consists of the instructions on 
how to use the software. FDA has tentatively concluded that such 
instructions need to be included in the definition of ``software'' to 
ensure the proper operation of the software.
---------------------------------------------------------------------------

    \1\ ISO is a world-wide federation of national standards bodies 
that set quality assurance guidelines for products that will enter 
international commerce. The ISO defines software as an 
``intellectual creation comprising the programs, procedures, rules 
and any associated documentation pertaining to the operation of a 
data processing system'' (Ref. 26).
    \2\ IEEE is a trade organization comprised of several societies. 
IEEE standards are developed within the technical committees of the 
IEEE societies and represent a consensus opinion of experts from 
within IEEE as well as experts who are not members of IEEE. IEEE 
defines software as ``computer programs, procedures, and possibly 
associated documentation and data pertaining to the operation of a 
computer system'' (Ref. 27).
---------------------------------------------------------------------------

    The definition of ``system'' in proposed Sec. 106.35(a)(3) derives 
from the IEEE Standard 610.12-1990 (Ref. 27). FDA is proposing to 
incorporate this definition because many of the requirements in 
proposed Sec. 106.35 cannot be related to software or hardware alone 
but rather to systems in which software is used in conjunction with 
hardware. For example, testing software under simulated conditions of 
use may be beneficial during the early and middle stages of software 
development, but validation of the software must be performed in 
conjunction with the relevant hardware in the operational environment 
it is

[[Page 36166]]

intended to be used in. Therefore in proposed Sec. 106.35(b)(4), FDA is 
proposing that all systems be validated ``before their first use to 
manufacture commercial product.''
    Proposed Sec. 106.35(a)(4) defines ``validation'' as establishing 
documented evidence that provides a high degree of assurance that a 
system will consistently produce a product meeting its predetermined 
specifications and quality characteristics. It is important that a 
process control system comply with specified requirements each time it 
operates. The proposed definition is derived from the ISO International 
Guideline ISO-9000-3, (which defines ``validation'' as ``the process of 
evaluating software to ensure compliance with specified requirements'' 
(Ref. 26)); the IEEE Standard 610.12-1990, which (defines it as ``the 
process of evaluating a system or component during or at the end of the 
development process to determine whether it satisfies specified 
requirements'' (Ref. 27)); and FDA's ``Glossary of Computerized System 
and Software Development Terminology,'' which defines it as 
``establishing documented evidence which provides a high degree of 
assurance that a specific process will consistently produce a product 
meeting its predetermined specifications and quality characteristics'' 
(Ref. 28). FDA is proposing to incorporate these definitions into its 
regulations because they are applicable to the types of systems used in 
infant formula manufacture, are derived from internationally accepted 
definitions, are consistent with existing FDA terminology, make clear 
that the process of evaluation includes the complete system (i.e., the 
hardware used in conjunction with the software), and include the 
concept of consistency.
    Proposed Sec. 106.35(b)(1) sets forth requirements for designing, 
installing, testing, and maintaining all systems so that they function 
as intended. Some systems may work properly only within a narrow range 
of environmental conditions, such as temperature and humidity, and some 
might be particularly sensitive to electromagnetic interference. The 
actual conditions of use of a system should be considered as early as 
possible in its design and development. Systems need to be installed in 
a manner that takes into account the inherent limitations of the 
system, tested under conditions that reflect actual conditions of use, 
and properly maintained to ensure that they continue to function as 
expected during their lifetime.
    Proposed Sec. 106.35(b)(2) requires that the manufacturer ensure 
that all hardware is routinely calibrated, inspected, and checked 
according to written procedures. FDA has tentatively concluded that 
this provision is necessary to ensure that any infant formula 
manufactured under the control of automatic equipment meets the 
requirements of the act and is manufactured in a manner designed to 
prevent adulteration. For example, a batch of infant formula may lack 
the required levels of nutrients if equipment used for the automatic 
dispensing of a nutrient premix is out of calibration or has a clogged 
delivery line. The routine calibration, inspection, and checking of 
hardware will ensure that it continues to perform as intended, and that 
its operation will not result in a process that deviates from 
established specifications. The establishment of written procedures for 
the calibration, inspection, and checking of hardware will ensure that 
these procedures are performed consistently and in an appropriate way.
    The incorporation of software into the operation of automatic 
equipment has not only increased the complexity of such equipment but 
also has resulted in a process that may operate differently for each 
execution because a software-based control system can be configured at 
will by the operator or by the system itself. Therefore, proposed 
Sec. 106.35(b)(3), (b)(4), and (b)(5) require that manufacturers 
exercise appropriate controls over systems and, in particular, over the 
software used in the systems.
    Proposed Sec. 106.35(b)(3) prescribes procedures for ensuring that 
systems are checked for input and output errors resulting from faulty 
data entry, faulty programming, or equipment malfunction. Such errors 
can result in serious production or quality control errors leading to a 
contaminated or adulterated infant formula. For example, a faulty 
position sensor on a downstream valve that improperly indicates that it 
is closed may result in a post-sterilization contamination. An 
improperly installed (or empty) ink cartridge in a color printer or 
multi-pen recorder may cause portions of a record to not be printed. 
FDA has tentatively concluded that the regulation is necessary to 
ensure that the infant formula produced or analyzed using the system is 
not adulterated. However, proposed Sec. 106.35(b)(3) also provides that 
the degree and frequency of input/output checks are to be based on the 
complexity and reliability of the system and the level of risk 
associated with the safe operation of the system.
    Proposed Sec. 106.35(b)(4) requires that manufacturers ensure that 
all systems are validated before their first use to manufacture 
commercial product. FDA has tentatively concluded that it is necessary 
that software programs that are used in a process control system to 
monitor and control established points deemed necessary to prevent 
adulteration (such as the speed of a pump, temperature of a heat 
exchanger, addition of vital nutrients, and air overpressure in an 
aseptic storage tank) be validated to ensure that use of the process 
control system will produce compliance with the specifications or 
standards at each control point. For example, if a continuous flow 
process is designed to heat an in-process batch of infant formula in a 
plate-to-plate heat exchanger to a specification of 271  deg.F, as 
indicated by the temperature at the end of the hold tube, and the 
system is mistakenly programmed to divert the product to the raw 
(unsterilized) surge tank only if the temperature drops below 261 
deg.F, an in-process batch of infant formula heated to 261  deg.F would 
not be diverted to the raw surge tank but rather would be handled by 
the computer as if it were adequately processed. Such an underprocessed 
batch of infant formula would likely pose a foodborne biological 
hazard. Thus, FDA has tentatively concluded that the validation 
required under proposed Sec. 106.35(b)(4) is necessary to ensure that 
infant formula that is produced or analyzed using the system is not 
adulterated.
    The validation of software ordinarily includes the following 
elements: Requirements development, design, coding, debugging, testing 
(with the hardware), and maintenance (Refs. 29, 30, and 31). Software 
validation also includes a review for correctness of the software 
documentation to ensure that the instructions prompt the input of the 
proper commands or data by the user. However, depending on the nature 
of the software and the hardware that it controls, some or all of these 
aspects of the validation process may be done by the infant formula 
manufacturer, by the manufacturer of equipment that is purchased by the 
infant formula manufacturer, or by a third party vendor.
    Proposed Sec. 106.35(b)(4) leaves the identity of the person that 
does the validation to the discretion of the infant formula 
manufacturer but makes clear that the infant formula manufacturer is 
responsible for ensuring that the system is validated. The proposal 
does not stipulate any standards or specifications for the validation 
process because the extent of the validation necessary is

[[Page 36167]]

related to the level of risk that each component of the system 
presents.
    More emphasis should be placed on validating portions of the system 
that represent major risk than on those that confer moderate or minor 
risk. A major risk is associated with systems that control or monitor a 
point where such control or monitoring is deemed necessary to prevent 
adulteration of the infant formula; for example, systems that control 
or monitor nutrient addition or processing temperature present a major 
risk. A moderate risk is associated with systems that influence, but 
that do not control or monitor, a point where control or monitoring is 
deemed necessary to prevent adulteration of the infant formula. For 
example, the speed of computer processing presents a moderate risk if 
software that is designed to be used on a high-speed computer is used 
on a slower computer. A minor risk is associated with systems that do 
not involve a point where control or monitoring is deemed necessary to 
prevent adulteration. For example, systems that control pallet stacking 
or product conveying present a low risk.
    Proposed Sec. 106.35(b)(5) requires that any system that is 
modified be revalidated after any modification and before use of the 
modified system to manufacture commercial product. FDA has tentatively 
concluded that revalidation is necessary to ensure that no errors are 
introduced into the system during the modification and to ensure that a 
modification in one aspect of a process control system does not, 
unknowingly but adversely, affect other aspects of the process control 
system, particularly those operations that follow the modified aspect 
of the system.
    Under Sec. 106.35(b)(5), FDA is also proposing that a specific 
individual (or group of individuals) is designated to modify software 
to prevent the indiscriminate modification of software and to ensure 
that all modifications are made consistently. The designated individual 
may be employed by the infant formula manufacturer, the manufacturer of 
equipment purchased by the infant formula manufacturer, or by a third 
party. The regulation states, however, that the infant formula 
manufacturer is responsible for ensuring that modified software is 
retested or revalidated regardless of who does the modification.
    Proposed Sec. 106.35(c)requires that infant formula manufacturers 
make and retain records concerning automatic (mechanical or electronic) 
equipment. FDA is proposing this requirement under the authority of 
section 412(b)(4)(A)(i) of the act, which requires the retention of all 
records necessary to demonstrate compliance with the CGMP and quality 
control procedures prescribed under section 412(b)(2) of the act, 
including the results of all testing required under section 
412(b)(2)(B) of the act. These records will allow manufacturers to 
readily determine whether this crucial equipment is being appropriately 
operated and maintained. They will allow manufacturers to troubleshoot 
and to operate these systems with a minimum of downtime when problems 
occur because the records will include a copy of all software used and 
a backup file of data entered into the computer or related system which 
can be used to reload the system. The records will also provide 
information that the manufacturer can use in trying to determine why a 
problem with the system is occurring or why the system is not producing 
an infant formula that complies with the manufacturer's specifications 
for the product.
7. Controls to Prevent Adulteration Caused by Ingredients, Containers, 
and Closures
    Proposed Sec. 106.40(a) specifies that the only substances that may 
be used in infant formulas are food ingredients that are generally 
recognized as safe (GRAS) for use in infant formula, that are used in 
accordance with the agency's food additive regulations, or that are 
authorized by a prior sanction issued by FDA. Under section 
412(b)(2)(A) of the act, FDA is to establish CGMP's that it determines 
are necessary to ensure that the infant formula is manufactured in a 
way that is designed to prevent adulteration of the formula. Unless the 
safety of the ingredients of an infant formula has been established, 
the formula is adulterated under section 402(a)(1) and (a)(2)(C) of the 
act. Thus, the agency has tentatively concluded that CGMP requires that 
the manufacturer ensure that the ingredients that it uses in its 
formula are safe and suitable.
    Proposed Sec. 106.40(b) requires that infant formula containers and 
closures not be reactive or absorptive so as to affect the safety of 
the infant formula, and that all packaging material that comes in 
contact with an infant formula be composed of authorized substances and 
be used in accordance with any prescribed limitations. Various 
regulations that authorize the use of a material in contact with the 
food product also set conditions and limitations on that use. Thus, the 
agency proposes to require that the manufacturer not only use only 
materials specified in proposed Sec. 106.40(b), but also that the 
materials be used as specified in the regulations authorizing their 
use. This provision will ensure that the food contact surface of 
containers and closures will not adulterate the infant formula.
    In order for the manufacturer to maintain a complete record of how 
each ingredient, container, or closure was used and to determine which 
lots of infant formula are adulterated if a problem is ultimately 
identified with a particular lot of ingredients, containers, or 
closures, FDA is proposing, in Sec. 106.40(c), that they be identified 
with batch or lot numbers. This batch or lot number can be used to 
identify ingredients, containers, or closures that have been released 
for use in infant formula or rejected for use in infant formula 
manufacture. It also can be used to track the ingredients, containers, 
or closures that were used in the manufacture of each batch of infant 
formula.
    Proposed Sec. 106.40(d) requires that infant formula manufacturers 
develop written specifications that stipulate the standards for 
acceptance or rejection of ingredients, containers, and closures. 
Stipulating the standards for acceptance or rejection of ingredients 
used to supply nutrients is important to ensure that all the required 
nutrients are present in the formula at the required levels. For 
example, the level of endogenous nutrients that a manufacturer expects 
will be supplied by an ingredient should be stipulated as a standard 
for acceptance or rejection of that ingredient. Endogenous nutrients 
are nutrients provided as a part of other nutrients, such as minerals 
provided as a part of the protein source. Sodium, for example, is 
frequently provided as part of the protein ingredient ``caseinate.''
    To ensure that the mineral is provided in the infant formula at at 
least the minimal level, and not above the maximum level, required by 
Sec. 107.100, the infant formula manufacturer must know what amount of 
a mineral is provided to the formula by all ingredients that are 
sources of the mineral. Thus, a standard for the level of the 
endogenous nutrient that is to be provided by an ingredient is an 
appropriate specification for the manufacturer to develop. If the level 
of the mineral is too high in the ingredient, it may cause the formula 
to exceed the maximum established in Sec. 107.100. Similarly, if the 
level is too low, the formula may not meet the required minimal level.
    Developing standards for acceptance or rejection of ingredients 
used in infant formula manufacture is also important to ensure that 
contaminants in the

[[Page 36168]]

ingredients that may lead to adulteration of the product are not 
present in the formula. Examples of contaminants that may lead to 
adulteration of an infant formula include certain heavy metals, such as 
lead. Infant formula manufacturers are currently setting standards for 
the lead in the ingredients that they use in infant formula to ensure 
that the lead level in infant formulas is at or below the 
quantification limit of the method used for lead determination (Ref. 
32).
    Stipulating the standards for acceptance or rejection of containers 
or closures used in infant formula manufacture is important to ensure 
that the integrity of the container and of the closure is maintained to 
prevent leakage of the formula and to prevent an infant formula from 
becoming adulterated, which can occur if the container or closure is 
not impenetrable to air (which can cause nutrient degradation), or if 
the container or closure allows outside contaminants to get into the 
infant formula.
    Proposed Sec. 106.40(d) also requires that manufacturers establish 
written specifications that stipulate the procedures for determining 
whether the ingredients, containers, and closures meet the standards. 
Examples of procedures manufacturers may use to determine whether they 
meet the standards are acceptance of a supplier's guarantee or 
certification and testing conducted by the infant formula manufacturer. 
In some cases, manufacturers must conduct their own testing to ensure 
that the standards for acceptance or rejection of the ingredient are 
met. For example, section 412(b)(3)(B) of the act requires that 
manufacturers test each nutrient premix for each relied-upon nutrient 
to ensure that the premix complies with its specifications or 
certifications by a premix supplier, but the act does not require 
testing of individual nutrient ingredients when such nutrients are not 
supplied as a nutrient premix. However, a manufacturer may find through 
experience that the best way to ensure that the final product will meet 
all specifications is to test certain nutrient ingredients for 
identity, purity, and potency before using them in the infant formula.
    In addition, manufacturers should have controls in place to ensure 
that any ingredients, containers, or closures that do not meet any of 
their specifications are not used in production of a batch of infant 
formula. However, if these controls fail, and any such ingredients, 
containers, or closures are used in a batch of formula, FDA is 
proposing under Sec. 106.40(d) that an individual qualified by training 
or experience conduct an investigation to ensure that the failure does 
not lead to release into the marketplace of an adulterated product.
    Proposed Sec. 106.40(e) requires that ingredients, containers, and 
closures be stored in areas clearly designated for materials pending 
release for use, materials released for use, or materials rejected for 
use in infant formula production in order to prevent mixups in using 
materials that are inappropriate for infant formula manufacturing. FDA 
is further proposing to require that any lot of ingredients, 
containers, or closures that does not meet the manufacturer's 
specifications be rejected and controlled under a quarantine system 
designed to prevent its use in the manufacture of infant formula. 
Failure to protect against the use of these materials would 
significantly increase the likelihood that an adulterated product will 
be produced.
    Some ingredients used in infant formula are vulnerable to 
degradation when they are exposed to heat or air. Moreover, containers 
or closures may be exposed to air containing dust and dirt and become 
contaminated. Thus, the ingredients, containers, and closures may need 
to be reexamined after they are exposed to air, heat, or other 
conditions that may adversely affect them to ensure that they still 
meet the manufacturer's specifications. Thus, FDA is proposing, in 
Sec. 106.40(f), to require retesting or reexamination after approved 
materials have been exposed to conditions that may adversely affect 
them.
    Proposed Sec. 106.40(g) requires that manufacturers make and retain 
records on ingredients, containers, and closures used in the 
manufacture of infant formula so that if adulteration of formula 
occurs, the manufacturer will be able to determine the source of the 
material, so that its use can be halted. In addition, the records will 
show the basis on which each ingredient, container, and closure was 
released for use in infant formula production, if questions about such 
release later arise. FDA has authority to require these records, under 
section 412(b)(4)(A)(i) of the act.
8. Controls to Prevent Adulteration During Manufacturing
    The infant formula manufacturing process involves a number of 
complicated processes that may cause adulterated formula to be produced 
if the processes are not properly conducted or monitored. Therefore, 
FDA is proposing, under section Sec. 106.50, to require that 
manufacturers establish controls to minimize the risk that 
manufacturing process errors will produce an adulterated or unsafe 
formula. The proposed requirements reflect many of the practices 
currently used by infant formula manufacturers and manufacturers of 
other commodities that require strict production controls to prevent 
product adulteration (e.g., Ref. 9 and 21 CFR 211.100 through 211.115).
    Proposed Sec. 106.50(a)(1) carries forward and amends the 
requirement in current Sec. 106.25(a) that a master manufacturing order 
be prepared and followed. A master manufacturing order is necessary to 
ensure that the manufacturer will produce each batch of a particular 
infant formula the same way. If the master manufacturing order is not 
followed, all necessary ingredients may not be added to the formula in 
the appropriate concentrations and in the appropriate manner.
    FDA is also proposing that manufacturers make and retain records 
that include complete information relating to the production and 
control of the batch at the time each manufacturing operation is 
performed (see proposed Sec. 106.50(a)(2)). This proposed requirement 
will ensure that the complete history of each batch of infant formula 
is available for review in the event that a problem arises with a 
particular batch.
    Proposed Sec. 106.50(a)(2) also requires that an individual 
qualified by training or experience conduct an investigation of any 
deviations from the master manufacturing order and any corrective 
actions taken. This investigation is necessary to ensure that any 
deviations from the master manufacturing order do not lead to an 
adulterated product.
    If any changes are made to the master manufacturing order, proposed 
Sec. 106.50(a)(3) requires that they be drafted, reviewed, and approved 
by a responsible official and include an evaluation of the effect of 
the change on the nutrient content and the suitability of the formula 
for infants. This process is necessary to prevent unintended adverse 
effects that could result from changes to the master manufacturing 
order made by persons not qualified to assess their impact. The 
production of infant formula is a sophisticated process, and all 
organizational units that are involved in critical formulation and 
production steps, such as production, engineering, research, and 
regulatory affairs, should review and approve changes to the master 
manufacturing order. FDA has tentatively concluded, however, that all 
changes to the master manufacturing order need to be

[[Page 36169]]

reviewed by at least one responsible official, and that this official 
will need to evaluate how the change will affect the nutrient content 
and the suitability of the product for infants, to ensure that the 
infant formula is not adulterated.
    A significant change in the master manufacturing order without 
proper approval may result in the production of an infant formula that 
lacks a required nutrient or that is not manufactured in an appropriate 
way. For example, homogenization of an infant formula is done to ensure 
a uniform dispersion throughout the formula of the lipid ingredients as 
well as the fat-soluble nutrients. If the master manufacturing order 
were changed, and the homogenization process done before the fat source 
was added, the fat-soluble nutrients would not be uniformly dispersed 
in the formula, and the formula would be adulterated. The system of 
review and approval required by proposed Sec. 106.50(a)(3) will 
minimize the possibility that a significant change could result in an 
adulterated product.
    In order to ensure that the appropriate ingredients are added 
during the manufacturing process, and that the formula contains all of 
the nutrients required by Sec. 107.100 and therefore is not 
adulterated, FDA is proposing in Sec. 106.50(b) that each raw or in-
process ingredient required by the master manufacturing order be 
examined by one person and checked by a second person or system. This 
requirement will ensure that there will be a check to prevent mixups in 
the use of ingredients and to prevent the use of unapproved 
ingredients. Confirmation that the master manufacturing order is being 
followed, and that ingredients are being properly added, is 
particularly important because these matters are fundamental to 
ensuring that the formula is manufactured correctly, and that it 
contains the nutrients required by Sec. 107.100 but not unapproved 
ingredients that might adulterate the formula.
    In proposed Sec. 106.50(c), FDA is requiring the identification of 
all compounding and storage containers, processing lines, and major 
equipment used during the production of a batch of infant formula. 
Identification of these items will enable the manufacturer to 
accurately determine the status of all batches of infant formula during 
all stages of the manufacturing process, will help to prevent mixups in 
the addition of ingredients to the formula, and will facilitate prompt 
action by the manufacturer if any problems in processing are 
identified. For example, identifying that a particular storage 
container contains a batch of formula that has not yet had all 
ingredients added to it will prevent a manufacturer from inadvertently 
final-stage packaging the product and thus will help to ensure that 
adulterated product is not introduced into interstate commerce. The 
presence of the lot or batch number will help to identify the product 
if a problem does occur.
    Proposed Sec. 106.50(d) requires that manufacturers establish 
controls to ensure that required nutrient levels are maintained in the 
formula, and that the formula is not contaminated with microorganisms 
or other contaminants and thereby adulterated. In addition, the agency 
is proposing to require establishment of controls for mixing time, 
speed, temperature, and flow rate of product and other critical 
parameters necessary to ensure the addition of required ingredients to, 
and the homogeneity of, the formula. These parameters are determined by 
the manufacturer according to its experience and knowledge of what will 
result in a homogeneous, safe, and uniform product. It is essential 
that controls be established for each of these parameters, or the 
likelihood that there will be inconsistencies in production from batch 
to batch will be greatly increased. For example, if processing 
temperatures are not specified, the formula could be processed at high 
temperatures that can destroy vitamins or other essential nutrients, 
resulting in a product that is adulterated because it does not meet the 
nutrient requirements specified in section 412(i) of the act. 
Similarly, without established procedures for mixing time and speed, 
the product may be produced using processing parameters that will not 
result in formula that is uniformly mixed and thus does not contain all 
nutrients at the required levels.
    FDA is proposing to require that manufacturers establish controls 
for the spray-drying process for powdered infant formula to prevent 
microbial and other contamination (Sec. 106.50(d)(2)). Although spray 
drying involves a heat treatment, the temperature is not sufficient to 
sterilize the formula. Consequently, powdered infant formulas are 
vulnerable to microbial contamination during the spray-drying process. 
Even if the equipment and the formula are free of microbial and other 
forms of contamination initially, the spray-drying process may permit 
contamination of the product as a result of dust or other air-borne 
gross particulates in the intake air. Thus, FDA has tentatively 
concluded that it is important that the manufacturer establish controls 
for the spray-drying process that will ensure that the powdered formula 
does not become contaminated with microorganisms or other contaminants.
    The controls that manufacturers should consider include: (1) Using 
equipment constructed to ensure that static accumulation of particulate 
matter is controlled; (2) using and maintaining equipment constructed 
to protect the product from dust and environmental contamination; (3) 
controlling condensation, moisture, and temperature conditions 
throughout the plant to prevent Salmonella and Listeria growth in 
static materials; (4) controlling condenser cooling water to prevent 
potential Salmonella and other bacterial contamination; (5) controlling 
sampling and cleanout ports on the evaporator for buildup of static 
material and avenues for airborne contaminants; and (6) controlling 
product flow through the plant to prevent unnecessary product movement 
between areas that may increase the likelihood of cross-contamination.
    As stated above, contaminants may enter the product in the air 
introduced into the spray-drying equipment during the spray- drying 
process. Air can contain free microorganisms or particulate material 
that is contaminated with microorganisms. Controls to prevent microbial 
contamination of the formula by airborne sources must address not only 
the presence of microorganisms themselves but also the sources of dust, 
moisture, and other airborne contaminants that may be sources of 
microbial contamination. Therefore, proposed Sec. 106.50(d)(2) requires 
that manufacturers filter the intake air before heating to remove dust 
or other air-borne gross particulates that can result in the production 
of adulterated formula.
    FDA is proposing to require that manufacturers control the removal 
of air from finished product containers (proposed Sec. 106.50(d)(3)) 
and ensure that containers of finished products are properly sealed 
(proposed Sec. 106.50(d)(4)), that visible closure and seal defects are 
detected (proposed Sec. 106.50(d)(4)(i)), and that destructive tests 
are performed to determine closure strength (proposed 
Sec. 106.50(d)(4)(ii)). These requirements are necessary to prevent 
oxidation and deterioration of nutrients in the formula caused by air 
or contaminants during the product's shelf life. FDA is also proposing 
that equipment that is used to prevent adulteration be monitored, 
either by personnel or monitoring equipment, to alert the manufacturer 
to malfunctions (see Sec. 106.50(e)). As a result of such monitoring, 
the manufacturer will be

[[Page 36170]]

able to minimize the amount of product produced subject to a 
malfunction that may develop and to take prompt corrective actions.
    In order to prevent rejected in-process materials from being 
inadvertently commingled with acceptable materials, FDA is proposing 
that manufacturers establish controls that ensure that the rejected 
materials are clearly identified and quarantined, and that reprocessed 
materials will not produce adulterated formula (see Sec. 106.50(f)).
9. Controls to Prevent Adulteration from Microorganisms
    An infant formula that is contaminated with microorganisms may, 
depending on the characteristics of the microorganisms, raise a safety 
concern that would cause the infant formula to be adulterated under 
section 402(a)(1) of the act. For example, all serotypes of the genus 
Salmonella can cause illness (often gastrointestinal) in infants and 
adults (Refs. 33 and 34) and the infectious dose is low (Ref. 35). 
Moreover, microorganisms that are generally harmless in older children 
and adults can cause serious bacterial infections in infants because 
the immune systems of infants are still developing (Ref. 36). For 
example, newborns and infants are susceptible to infection with 
Listeria monocytogenes that may cause severe illness or death (Ref. 37) 
and, as in the case of Salmonella, the infectious dose is believed to 
be low (Ref. 38).
    Likewise, Staphylococcus aureus is harmful to infants because some 
strains of this microorganism produce an enterotoxin that causes acute 
gastrointestinal illness (nausea, vomiting, cramps) soon after the food 
is ingested (Ref. 39). Bacillus cereus can produce diarrhea and 
vomiting in adult humans (Ref. 40) when food contaminated with at least 
10\5\ B. cereus cells is consumed. The infectious dose of B. cereus for 
infants is not known; however, as already noted, infants are more 
susceptible to bacterial infections than are healthy adults and older 
children because the immune systems of infants are not fully developed.
    FDA has long held that health concerns may arise due to the 
presence of any detectable Salmonella, Listeria, or S. aureus bacteria 
in infant formula or due to levels of B. cereus that exceed 1,000 
``colony forming units'' (CFU's) per gram (g) of a powdered infant 
formula. Such health concerns would cause the agency to consider an 
infant formula that is so contaminated to be adulterated under section 
402(a)(1) of the act (see 54 FR 3783, Jan. 26, 1989, and 56 FR 66566, 
Dec. 24, 1991).
    Moreover, the presence of microorganisms in an infant formula 
reflects that the formula was prepared, packed, or held under 
insanitary conditions whereby it may have been rendered injurious to 
health and therefore is adulterated under sections 402(a)(4) and 412 of 
the act. For example, the presence of Escherichia coli in a sample of 
infant formula is an indicator of fecal contamination, implying that 
the infant formula has been contaminated by manufacturing practices 
conducted under insanitary conditions and therefore is adulterated 
under sections 402(a)(4) and 412 of the act. In addition, consistent 
with the standard adopted by the International Commission on 
Microbiological Specifications for Foods (ICMSF) of the Food and 
Agricultural Organization of the United Nations and the World Health 
Organization (WHO) and based on the results from FDA and Canadian 
Surveys (Refs. 41, 42, and 43), an aerobic plate count (APC) (i.e., the 
number of microorganisms that will grow under certain specified 
conditions) that is greater than 10,000 CFU's per g of a powdered 
infant formula evidences that the formula has been prepared, packed, or 
held under insanitary conditions.
    Illnesses from the use of microbiologically contaminated infant 
formulas have occurred (Ref. 33). Moreover, as recently as May 1993, 
infant formula contaminated with Salmonella bacteria was the subject of 
a recall (Ref. 44). Thus, contamination of infant formula with 
microorganisms of public health significance is more than a theoretical 
possibility. Therefore, FDA has tentatively concluded that 
manufacturers need to have in place controls to ensure that formulas 
are not microbiologically contaminated at levels of public health 
significance, and that, if they are, those formulas do not enter 
interstate commerce. Proposed Sec. 106.55 requires manufacturers to 
establish such controls.
    Proposed Sec. 106.55(a) requires that manufacturers of liquid 
infant formula comply with the procedures specified in part 113. These 
products are thermally-processed low-acid foods that are packaged in 
hermetically sealed containers that are heated to achieve commercial 
sterility. Therefore, they are appropriately subject to the 
requirements of part 113.
    Proposed Sec. 106.55(b) requires that manufacturers of powdered 
infant formula test representative samples of every batch of the 
formula at the final product stage, before distribution, to ensure that 
the infant formula meets the microbiological quality standards 
specified in proposed Sec. 106.55(c). This proposed requirement is 
necessary because although powdered infant formulas are heat treated 
during processing, they are not thermally processed to achieve 
commercial sterility. Proposed Sec. 106.55(b) requires testing at the 
final product stage because microbiological contamination can be 
inadvertently introduced by ingredients at any time during production 
or through improper processing or holding procedures (Ref. 45).
    Proposed Sec. 106.55(c) establishes that any powdered infant 
formula that contains any microorganism at levels that exceed the 
microbiological quality standards for that microorganism as listed in 
this section will be deemed to be adulterated under sections 402 and 
412 of the act. Proposed Sec. 106.55(c) defines microbiological quality 
standards as the maximum allowable number of microorganisms present in 
1 g of dry formula, expressed as CFU/g or ``most probable number'' 
(MPN)/g, and herein designated the ``M value'' for the specific 
microorganism.
    The microorganisms for which FDA is proposing M values are those 
that are of known public health significance or that are indicators 
that the formula have been prepared, packed, or held under insanitary 
conditions. The microorganisms and each proposed M value listed in 
proposed Sec. 106.55(c) are adapted from guidelines previously 
published and discussed in the proposed and final rules on infant 
formula record and record retention requirements (see 54 FR 3783, Jan. 
26, 1989, and 56 FR 66566, Dec. 24, 1991, respectively). The agency 
notes, however, that microorganisms that must be tested for in infant 
formula and the proposed M values for each microorganism listed in this 
proposed rule represent minimum requirements for the microbiological 
quality of an infant formula based on standards and methods currently 
available.
    a. Aerobic plate count (APC). Proposed Sec. 106.55(c) establishes 
an APC M value of 10,000 CFU/g as the maximum level that is consistent 
with sanitary conditions in the facility in which a powdered infant 
formula is produced. An APC M value greater than the proposed standard 
indicates that the formula was produced under insanitary conditions 
whereby it may have been rendered injurious to health and thus is 
adulterated under sections 402(a)(4) and 412 of the act.
    The APC is the number of microorganisms that will grow on the APC 
nutrient medium, incubated at 35  deg.C for 24 hours in air (Ref. 46). 
``Microorganisms'' (as defined in

[[Page 36171]]

proposed Sec. 106.3(k)) include yeasts, molds, bacteria, and viruses. 
The APC medium supports the growth of most microorganisms, including 
yeasts, molds, and all bacteria required to be tested for under 
proposed Sec. 106.55(c); however, the APC medium does not support the 
growth of viruses. The APC count is expressed in CFU's because multiple 
microorganisms may adhere together or attach to the same location on an 
agar plate, and microbiologists cannot determine whether one or several 
individual microorganisms initiated the colony that they detect growing 
on the plate.
    This M value for the APC proposed in Sec. 106.55(c) is consistent 
with the standard adopted by the ICMSF and the WHO and the results from 
FDA and Canadian Surveys (Refs. 41, 42, and 43). The ICMSF based its 
standards on the degree of health hazard the microorganisms present and 
conditions of use of the product (Ref. 41).
    FDA has tentatively arrived at this APC M value because the 
microbial quality of products consumed by infants is of primary concern 
(Ref. 43). When infant formulas are produced under good commercial 
processing, the available evidence shows that the APC will be below 
this M value (Refs. 42 and 43). The agency is notaware of adverse 
events occurring in infants who consumed products with an APC below 
this M value.
    b. Coliforms, fecal coliforms, and E. coli. E. colli are bacteria, 
including some strains that are pathogenic for infants, that thrive in 
the human intestinal tract. The presence of E. coli in a sample of 
powdered infant formula is an indicator that the infant formula has 
been contaminated by manufacturing practices conducted under insanitary 
conditions and therefore is adulterated under sections 402(a)(4) and 
412 of the act.
    E. coli bacteria are a subset of a more diverse group of bacteria 
known collectively as fecal coliforms, which also thrive in the human 
intestinal tract and therefore are also indicators of fecal 
contamination. Fecal coliforms are destroyed by pasteurization, and the 
presence of these microorganisms in a pasteurized product evidences 
that there has been post-process contamination of the formula (Ref. 
47). Fecal coliforms in turn are a subset of a still further diverse 
group of bacteria known as coliforms, which include bacteria that may 
or may not be indicators of fecal contamination. However, contamination 
with coliforms is a reliable indicator of post-process contamination of 
the formula, even if the source of the contamination is not fecal.
    In previously issued guidelines, the agency recommended that 
powdered infant formulas be tested for the presence of E. coli (54 FR 
3783); however, one comment on this recommendation suggested that, to 
allow greater flexibility and reduce the cost for manufacturers, the 
manufacturer should be given the option of testing for coliforms, fecal 
coliforms, or E. coli. Specific tests for contamination with E. coli 
provide the most definitive evidence of fecal contamination, but tests 
for specific bacteria are more cumbersome than general tests for a 
group of bacteria such as fecal coliforms. Similarly, general tests for 
fecal coliforms are more cumbersome than universal tests for an even 
more diverse group of bacteria such as coliforms.
    The agency is proposing in Sec. 106.55(c) that manufacturers screen 
their samples of powdered infant formula for evidence of contamination 
with E. coli using sequential tests for detecting and enumerating 
coliforms and fecal coliforms. Under the proposal, manufacturers 
ordinarily would only perform the simplest test (i.e., the test for 
coliforms) using a test sample of the infant formula. The results of 
the coliform test determine whether the manufacturer needs to followup 
with a more specific test for fecal coliforms using as the test sample 
cultured bacteria prepared during the coliform test. As discussed 
below, the agency is not proposing that manufacturers followup a 
positive result in the fecal coliform test with a more specific test 
for E. coli but rather is proposing that a violative sample in the 
fecal coliforms test will represent conclusive evidence that the infant 
formula is adulterated.
    The general test for coliforms is an example of an MPN test. MPN 
counts are estimates of the number of organisms present in a sample. 
Methods resulting in an MPN require inoculation of multiple tubes of 
liquid culture medium with multiple dilutions of the sample. The method 
specified in FDA's Bacteriological Analytical Manual (BAM) (Ref. 46) 
requires inoculation of 3 replicate tubes of culture medium with each 
of 3 sample dilutions, for a total of 9 tubes. The tubes contain 
culture medium selective for the microorganism of interest. After 
appropriate incubation (time, temperature, and atmosphere), each tube 
is scored as positive or negative for the presence of the organism. 
Examples of a positive result include the presence of growth, a 
biochemical color change, and the production of gas.
    A mathematical formula is used to calculate the MPN of 
microorganisms present based on the number of positive tubes in each of 
the three separate dilutions. Since the calculation in question 
involves a repetitious process, the mathematical formula used to 
calculate the MPN has been employed to create easy-to-use tables that 
are available in the BAM and in other books of statistical tables. Most 
tables present both a value for the MPN and confidence limits for that 
value. The calculated table values for the MPN, using BAM methods, are 
dependent on the level of the dilution in which a positive result is 
found. The following table values are based on an inoculation series of 
0.1, 0.01 g, and 0.001 g (or mL) of the infant formula. When no tubes 
in any dilution produce a positive result, the calculated MPN value is 
zero.3 When a single tube in the greatest dilution (least 
concentrated) produces a positive result, the calculated MPN value is 
equal to 3.01.4 When a single tube in the middle dilution produces 
a positive result, the calculated MPN value is equal to 3.05.5 In 
all other situations in which there is a positive result in at least 
one tube (including a single positive tube in the lowest dilution 
(greatest concentration)), the calculated MPN value is greater than 
3.05.
---------------------------------------------------------------------------

    \3\ The calculated MPN value of zero when no tubes in any 
dilution produce a positive result is a recent change that appears 
in the MPN tables of the 8th ed. of the BAM. In previous editions of 
the BAM, the calculated MPN value when no tubes in any dilution 
produce a positive result was ``less than 3.''
    \4\ The calculated MPN value of 3.01 when a single tube in the 
greatest dilution produces a positive result is a recent change that 
appears in the MPN tables of the 8th ed. of the BAM. In previous 
editions of the BAM, the calculated MPN value when a single tube in 
the greatest dilution produces a positive result was 3.
    \5\ The calculated MPN value of 3.05 when a single tube in the 
middle dilution produces a positive result is a recent change that 
appears in the MPN tables of the 8th ed. of the BAM. In previous 
editions of the BAM, the calculated MPN value when a single tube in 
the middle dilution produces a positive result was 3.
---------------------------------------------------------------------------

    If no tubes in any dilution produce a positive result in a test for 
bacterial contamination of a powdered infant formula (i.e., if the MPN 
is zero), such contamination is unlikely. If a single tube in any 
dilution produces a positive result in a test for bacterial 
contamination of the product, such contamination is a possibility. 
However, there are two situations in which a single positive tube is 
generally considered to reflect a false positive test result: (1) When 
no tube in the lowest dilution (greatest concentration) produces a 
positive result, but a single tube in the middle dilution produces a

[[Page 36172]]

positive result (i.e., the calculated MPN value is equal to 3.01); or 
(2) when no tube in the lowest dilution produces a positive result, but 
a single tube in the greatest dilution (least concentration) produces a 
positive result (i.e., the calculated MPN value is equal to 3.05). FDA 
considers that if a sample of a powdered infant formula produces 
positive test results that reflect one of these two situations, 
bacterial contamination also is unlikely.
    However, in all other situations (e.g., if a single tube in the 
lowest dilution (greatest concentration) produces a positive result, or 
if two or more tubes in any dilution produce a positive result), 
bacterial contamination of a powdered infant formula is likely. 
Therefore, when the calculated MPN value in a test for bacterial 
contamination is greater than 3.05, that is if a sample of powdered 
infant formula produces positive test results in which a single tube in 
the lowest dilution produces a positive result or in which two or more 
tubes in any dilution produce a positive result, the powdered infant 
formula likely is contaminated with bacteria.
    FDA is proposing to use the calculated MPN values in the BAM as a 
means of setting a numerical specification because these tables are 
generally available, represent standard practice in the industry, and 
provide a simple way to classify samples as violative or nonviolative. 
Based on the above discussion of calculated MPN values, FDA is 
proposing in Sec. 106.55(c) that powdered infant formula be classified 
as nonviolative for coliforms in all situations in which the calculated 
MPN value is less than or equal to 3.05 and classified as presumptively 
violative for coliforms in all situations in which the calculated MPN 
value is greater than 3.05. In other words, FDA is proposing that an 
MPN value of 3.05 represents the maximum allowable number of coliforms 
present in 1 g of dry infant formula. This proposal is consistent with 
current FDA infant formula microbiological guidelines. The agency 
requests comment on the specification of 3.05 MPN/g as the maximum 
allowable number of coliforms in dry infant formula.
    FDA has stated that infant formula with a calculated MPN value of 
greater than 3.05 in the coliform test is presumptively violative 
because, under proposed Sec. 106.55(c), the manufacturer may either 
consider the sample violative without further testing or may conduct an 
additional test, the fecal coliform test. Although an MPN value of 
greater than 3.05 MPN/g is a valid quality indicator of microbial 
contamination, coliform contamination may not be fecal in origin, and 
it may not reflect the presence of infant pathogenic microorganisms. 
Therefore, FDA has tentatively concluded that an infant formula for 
which an MPN value of greater than 3.05 MPN/g is found in the coliform 
test need not be considered violative if a negative result is found in 
a more specific test for fecal coliforms.
    If the coliform test using powdered infant formula samples results 
in an M value greater than 3.05 MPN/g, the manufacturer may use the 
cultured bacteria from one or more of the tubes producing the positive 
result as a sample inoculum for the fecal coliform test. A sample 
inoculum producing an MPN value in the fecal coliform test of less than 
or equal to 3.05 would indicate that the coliform contamination is not 
fecal in origin, because under incubation conditions that are specific 
for fecal coliforms, the bacteria were not detected. The testing would 
effectively screen out coliforms that are not of concern, which is not 
possible with the more general test. Therefore, FDA has tentatively 
concluded that an MPN value less than or equal to 3.05 in the fecal 
coliform test be classified as nonviolative. FDA also has tentatively 
concluded that an MPN value greater than 3.05 in the fecal coliform 
test is a valid quality indicator demonstrating that the formula 
contains fecal coliforms such as E. coli and, therefore, is adulterated 
under sections 402(a)(4) and 412 of the act. The agency is proposing 
that powdered infant formula that results in an MPN value greater than 
3.05 in the fecal coliform test be classified as violative.
    If the E. coli test was performed, the sample inoculum would be the 
cultured bacteria from positive tubes in the fecal coliforms test. 
However, the agency is not proposing to require specific testing for 
the presence of E. coli, or to set a specification for an M value for 
E. coli, because the specification of less than or equal to 3.05 MPN/g 
in the fecal coliforms test is sufficient to ensure that nonviolative 
samples do not contain E. coli since E. coli is a type of fecal 
coliform. Moreover, FDA has tentatively concluded that an MPN value 
greater than 3.05 in the fecal coliform test is a sufficient quality 
indicator of fecal contamination that the agency need not propose, as 
an option, that a manufacturer may conduct an additional specific test 
for the presence of E. coli. The agency requests comments on the 
proposed requirements for sequential testing for coliforms and fecal 
coliforms, with no testing for E. coli.
    c. Salmonella. Tests for the presence of Salmonella involve the 
enrichment in a broth of the entire analytical unit followed by plating 
onto culture plates rather than the culture of a series of dilutions 
that is performed in tests for coliforms. A positive result in a test 
for Salmonella is based on the detectable presence of the microorganism 
on the culture plate rather than on the mathematical calculations that 
result in a MPN.
    Proposed Sec. 106.55(c) requires that powdered infant formula be 
tested for Salmonella and provides that the formula is adulterated if 
any Salmonella is found. All serotypes of this genus of bacteria can 
cause illness (often gastrointestinal) in infants and adults (Refs. 33 
and 34). The presence of any Salmonella in infant formula could render 
it injurious to an infant who consumes it because the infectious dose 
of these bacteria is low (Ref. 35). Therefore, FDA has tentatively 
concluded that the risk from Salmonella is of such significance that an 
M value of zero (i.e., none detectable) for Salmonella in infant 
formula is necessary to protect the health of infants.
    d. Listeria monocytogenes. Tests for the presence of L. 
monocytogenes are similar to those for Salmonella and a positive result 
is based on the detectable presence of the microorganism on the culture 
plate rather than on the mathematical calculations that result in a 
MPN.
    Proposed Sec. 106.55(c) requires that powdered infant formula be 
tested for L. monocytogenes and provides that the formula is 
adulterated if any L. monocytogenes is found. Individuals with immune 
systems that make them susceptible to infections, such as newborns and 
infants with incompletely developed immune systems, are susceptible to 
infection with L. monocytogenes which may cause severe illness or death 
(Ref. 37). The infectious dose of this bacterium is believed to be low 
(Ref. 38). Because the specific dose of this bacterium that may cause 
illness is not known but is believed to be low, FDA has tentatively 
concluded that the risk from L. moncytogenes is of such significance 
that an M value of zero (i.e., none detectable) for L. monocytogenes in 
powdered infant formula is necessary to protect the health of infants. 
The agency requests comment on this proposed specification for L. 
monocytogenes.
    e. Staphylococcus aureus. S. aureus is harmful to infants because 
some strains of this microorganism produce an enterotoxin that causes 
acute gastrointestinal illness (nausea,

[[Page 36173]]

vomiting, cramps) soon after the food is ingested (Ref. 39). Tests for 
S. aureus involve liquid culture of series of dilutions as was 
discussed previously in reference to coliform and fecal coliform 
testing and results are calculated as MPN based on tables in the BAM. 
Proposed Sec. 106.55(c) requires that powdered infant formula be tested 
for S. aureus and establishes an M value of 3.05 for this 
microorganism. FDA has tentatively concluded that the risk from S. 
aureus is of such significance that an M value of 3.05 is necessary to 
protect the health of infants.
    f. Bacillus cereus. Tests for B. cereus involve liquid culture of a 
series of dilutions as was discussed previously in reference to 
coliform and fecal coliform testing and results are calculated as MPN 
based on tables in the BAM. Proposed Sec. 106.55(c) requires that 
powdered infant formula be tested for B. cereus when the APC exceeds 
100 CFU/g and establishes an M value for B. cereus of 100 MPN/g or 100 
CFU/g. This proposed M value for B. cereus is lower than the M value of 
1,000 MPN/g or 1,000 CFU/g in the current recommended infant formula 
microbiological guidelines (54 FR 3783). B. cereus can produce diarrhea 
and vomiting in adult humans (Ref. 40) when food contaminated with at 
least 105 B. cereus cells is consumed. The infectious dose of B. 
cereus for infants is not known; however, because the immune systems of 
infants are not fully developed, infants are more susceptible to 
bacterial infections than are healthy adults and older children. In the 
absence of data on the dose of B. cereus capable of causing disease in 
infants, the agency is concerned that a safety standard of 1,000 MPN/g 
or 1,000 CFU/g poses a potential risk to infants who consume rehydrated 
formula because B. cereus in rehydrated powdered infant formula is 
capable of rapid growth and can reach 4.9 x 106 cells/g within 24 
hours at 26  deg.C (Ref. 48), a level sufficient to cause disease. 
Therefore, FDA has tentatively concluded that the risk from B. cereus 
is of such significance that an M valve that is lower than the current 
standard of 1,000 MPN/g or 1,000 CFU/g is necessary to protect the 
health of infants.
    Powdered infant formulas and similar products (e.g., powdered milk) 
produced under CGMP contain less than 100 MPN/g or 100 CFU/g of B. 
cereus (Refs. 43 and 48). Additionally, an FDA survey of different 
production lots of milk-, soy-, and protein hydrolysate-based powdered 
infant formulas (Ref. 49) showed that the maximum APC was 103 CFU/g, 
and that the proportion of B. cereus in the samples ranged from 1.2 to 
63.9 percent of the APC. Therefore, FDA has tentatively concluded that 
an M value of 100 MPN/g or 100 CFU/g for B. cereus will adequately 
protect the health of infants. Moreover, because this M value is higher 
than the B. cereus levels typically found in infant formula currently 
being produced (Refs. 43, 48, and 49), the proposed M value of 100 MPN/
g or 100 CFU/g will not be overly burdensome.
    g. Methods. Proposed Sec. 106.55(c) states that the agency intends 
to determine compliance with the proposed M values using the methods in 
the BAM. These methods provide reproducible, consistent, and accurate 
results at different laboratories. The agency proposes to incorporate 
the BAM by reference in Sec. 106.55(c) in accordance with 5 U.S.C. 
552(a) and 1 CFR part 51. While manufacturers may use other equivalent 
methods, a manufacturer who uses methods that do not provide results 
that are consistent with the results obtained by methods approved by 
FDA will bear the risk that the firm's product is not in compliance 
with the law.
    The agency intends to test for Salmonella using the method 
described in Chapter 5, BAM, including the sample preparation 
procedures described in section C, paragraph 1 and the sampling plan 
described in Chapter 1, BAM; for L. monocytogenes using the method 
described in Chapter 10, BAM and the sampling plan described in Chapter 
1, BAM; for coliforms, fecal coliforms, and E. coli using the MPN 
method described in Chapter 4, BAM; for S. aureus using the MPN method 
described in Chapter 12, BAM; for B. cereus using the MPN or plate 
count method described in Chapter 14, BAM. The agency intends to 
determine the APC using the method described in Chapter 3, BAM. All 
chapter references are to the 8th ed. BAM. FDA intends to update the 
reference to reflect the most recent edition of the BAM at the time the 
final rule based on this proposed rule is issued.
    h. Records. Proposed Sec. 106.55(d) requires that manufacturers 
make and retain records, in accordance with proposed Sec. 106.100 
(e)(5)(ii) and (f)(7) on the testing of infant formula for 
microorganisms. As discussed in the description of the revisions to 
proposed subpart F of part 106, FDA has the authority to require such 
records under section 412(b)(4)(A)(i) of the act. These records will 
document whether the batch of powdered infant formula meets the 
microbiological quality standards of proposed Sec. 106.55(c) and is 
therefore not adulterated. Records that describe the full methodology 
for testing powdered infant formula for microbiological quality will 
provide consistency in the testing of the microbiological quality of 
the formula, even if different laboratory personnel conduct the tests. 
The accuracy and reproducibility of microbiological quality testing 
depend on the procedure used to conduct the test. In addition, the 
records will provide the manufacturer with data to evaluate any 
complaints received associated with a particular batch of infant 
formula by showing whether microbiological contamination could have 
contributed to the adverse event.
10. Controls to Prevent Adulteration During Packaging and Labeling of 
Infant Formula
    Because consumers rely on correct labels to select a formula to 
meet their childrens' individual needs and to have proper instructions 
for the use of the formula, FDA is proposing Sec. 106.60(a) which 
requires manufacturers examine packaged and labeled infant formula to 
ensure that containers and packages bear the correct labels, use-by 
dates, and traceability codes. The proposal also requires that labels 
be designed, printed, and applied so that they remain attached and 
legible during processing, handling, storage, and use (proposed 
Sec. 106.60(b)), and that all formula held in a single package be the 
same product bearing the same traceability code, and that the package 
carry the product name, name of the manufacturer, and the code 
(proposed Sec. 106.60(c)).
    These proposed requirements will ensure that infants who have 
allergies will not be placed at risk by consuming formula containing 
ingredients to which they are allergic, and that consumers will be 
aware of the date when the product may no longer be appropriate for 
use. In addition, the traceability codes will show the origin of the 
product if there were a recall, and the packaging requirements will 
make it more difficult for counterfeit formula, or formula with 
counterfeit labels, to be shipped in interstate commerce. There have 
been cases of counterfeit shipments in which a single package held more 
than one product, or held a single product which bore more than one 
code. The proposed regulations are not only intended to reduce the 
incidence of counterfeit activities, but to ensure that firms that 
receive the formula are aware that only one product should be in the 
packaging, and that all containers should be identified with the code 
shown on the package. This requirement will not impose an additional 
burden on industry because manufacturers routinely package a

[[Page 36174]]

single infant formula product bearing the same code.
11. Controls on the Release of Finished Infant Formula
    Proposed Sec. 106.70(a) requires that the manufacturer determine 
that each batch of formula meets all of the manufacturer's 
specifications before releasing the batch for distribution. 
Specifically, each batch must meet the requirements of Sec. 106.55 on 
microbiological contamination to ensure that the infant formula does 
not contain microorganisms at levels that may be injurious to the 
health of infants and render the formula adulterated and must meet the 
requirements of Sec. 106.91(a) on quality control procedures to ensure 
that the infant formula provides the required nutrients at the required 
levels, and that it provides any nutrient added by the manufacturer. 
Proposed Sec. 106.70(a) is designed to ensure that any infant formula 
that fails to meet the manufacturer's specifications, or that is 
adulterated for any reason, will not be introduced into interstate 
commerce.
    Proposed Sec. 106.70(b) requires that each batch of infant formula 
that fails to meet the manufacturer's specifications be rejected. 
Although proposed Sec. 106.70(b) recognizes that the formula may be 
reprocessed, it requires that the reprocessed product be shown to meet 
the requirements of Sec. 106.70(a) before the product is released. FDA 
has tentatively concluded that this proposed requirement is necessary 
to ensure that any defect that caused a batch of infant formula to be 
rejected is corrected before the formula is released into commerce.
    Proposed Sec. 106.70(c) requires that an individual qualified by 
training or experience conduct an investigation of a finding that a 
batch of infant formula fails to meet any manufacturer's 
specifications. This investigation is necessary to determine why such a 
failure occurred and to assist the manufacturer in developing controls 
to ensure that such a failure does not reoccur. FDA has proposed to 
require that the individual who conducts the investigation be qualified 
to ensure that the investigation is properly conducted.
12. Traceability
    Section 412(g)(1) of the act requires that each manufacturer make 
and retain such distribution records as may be necessary to effect and 
monitor recalls of the formula, and section 412(b)(4)(A)(vi) requires 
that each manufacturer retain all complaints concerning infant formulas 
that may reveal the possible existence of a hazard to health. 
Therefore, infant formulas must be traceable to permit identification 
of the product that is the subject of a complaint and to make it 
possible to determine whether that batch of infant formula presents a 
possible hazard to health. Traceability of an infant formula is also 
necessary so that the recall requirements of the act can be met.
    The agency's view, based on its experience, is that coding is the 
most effective method for ensuring traceability. It provides a uniform 
system that is able to identify large numbers of batches of infant 
formula with a distinctive code that is easily understood and that can 
be used by manufacturers, retailers, and consumers. A code also allows 
a large amount of information to be presented on the container of 
infant formula in a very small space. Therefore, the agency is 
proposing, under sections 412 (b)(4)(A)(vi) and (g)(1) and 701(a) of 
the act that batches of infant formula be identified with a distinctive 
code that will allow the traceability of an infant formula.
    Current Sec. 106.90 requires that manufacturers ensure traceability 
by coding all infant formulas in conformity with the coding 
requirements in Sec. 113.60(c) for thermally processed low- acid foods 
packaged in hermetically sealed containers. Section 113.60(c) requires 
that the code identify the establishment where the product is packed, 
the product contained therein, the year packed, the day packed, and the 
period during which packed, and that the packing period code be changed 
with sufficient frequency to permit ready identification of lots during 
their sale and distribution. FDA is proposing to carry the requirement 
that manufacturers code their product in accordance with Sec. 113.60(c) 
forward in proposed Sec. 106.80(a).
    FDA has tentatively determined that it is appropriate to code 
liquid infant formulas in this manner because they are thermally 
processed low-acid foods, and a batch is produced in a relatively short 
period of time, usually a day. It also may be appropriate for coding 
some powdered infant formulas in this manner if they are processed in a 
short enough time to make the day packed and the period during which 
packed meaningful information.
    Proposed Sec. 106.80(b) allows for alternative coding of batches of 
powdered infant formula. Powdered infant formula is usually 
manufactured in stages over a longer period of time than liquid infant 
formula. Some powdered infant formulas are dry mixed in a number of 
stages over an extended period of time. In other cases, powdered infant 
formula is mixed in liquid form at one manufacturing facility and 
shipped to a second site for spray drying and packaging. Powdered 
infant formula manufacturing is often not completed in a short enough 
period of time for coding based on the date packed or the period of 
time in which it was packed to be meaningful information. Therefore, 
under the alternate method that FDA is proposing, a sequential code 
would be assigned so that all the essential information needed to track 
any problems with the infant formula could be determined.
13. Audits of CGMP
    Proposed Sec. 106.90 requires that manufacturers (or their agents) 
conduct regularly scheduled audits to determine whether they are 
complying with CGMP. This provision derives from section 
412(b)(2)(B)(iv) of the act, which requires that the CGMP include ``the 
conduct by the manufacturer of an infant formula or an agent of such 
manufacturer of regularly scheduled audits to determine that such 
manufacturer has complied with the regulations prescribed under'' 
section 412(b)(2)(A) of the act. Section 412(b)(2)(A) requires that the 
Secretary (and by delegation FDA) establish CGMP's by regulation.
    FDA is proposing to require that regularly scheduled audits be part 
of CGMP because such audits are the best way to ensure overall 
compliance with CGMP and to identify recurring problems that may 
dictate an alteration in the master manufacturing order. For example, 
regularly scheduled audits of all deviations from the manufacturer's 
specifications or procedures will accentuate deviations that occur 
repeatedly and will enable the manufacturer to identify specifications 
or procedures that should be reassessed.
    Section 412(b)(2)(B)(iv) of the act also specifies that such audits 
are to ``be conducted by appropriately trained individuals who do not 
have any direct responsibility for the manufacture or production of 
infant formula.'' FDA is therefore proposing that an individual be 
knowledgeable in all aspects of infant formula production perform the 
audit. Without such broad knowledge, the individual conducting the 
audit will not be able to adequately evaluate the manufacturer's 
production and in-process control procedures. In addition, because the 
purpose of the audit is to determine whether the manufacturer is 
complying with the CGMP regulations issued under section 412(b)(2)(A) 
of the act, the agency has tentatively concluded that the person 
conducting the audit needs to be knowledgeable in

[[Page 36175]]

these regulations. Without such knowledge, the person would be unable 
to make the determinations that are the very purpose of the audit.
    The requirement that the audit be performed by an individual who 
has no direct responsibility for the matters being audited is one way 
to ensure the objectiveness of the audit process. The person should be 
free of any past involvement in the activities being audited because 
the audit is intended to uncover any problems or shortcomings in the 
manufacturer's procedures. A person who has been involved may feel that 
finding problems will reflect poorly on his or her work. Therefore, FDA 
has tentatively concluded that the audit must be conducted by someone 
who has no direct interest in the outcome of the audit.

C. Quality Control Procedures

1. Introduction
    FDA is proposing to redesignate and revise subpart B of part 106 as 
subpart C of part 106. Under this proposal, several sections of the 
current regulations will be revoked, and several sections will be 
redesignated without change. The latter sections are being recodified, 
however, to fit the organization of the proposed regulations. Table II 
describes the current and proposed regulations as follows:

                                Table II                                
------------------------------------------------------------------------
            Current regulation                   Proposed regulation    
------------------------------------------------------------------------
                           INGREDIENT CONTROL                           
------------------------------------------------------------------------
Sec.  106.20(a), Sec.  106.20(b)(1), Sec.   Changed by Secs.            
 106.20(b)(2).                               106.91(a)(1) and 106.40(d).
------------------------------------------------------------------------
                           IN-PROCESS CONTROL                           
------------------------------------------------------------------------
Sec.  106.25(a)...........................  Sec.  106.50(a)(1).         
Sec.  106.25(b)(1)........................  Omitted.                    
Sec.  106.25(b)(2)........................  Sec.  106.91(a)(4).         
Sec.  106.25(b)(3)........................  Sec.  106.91(a)(2).         
Sec.  106.25(b)(4)........................  Sec.  106.91(a)(4) with     
                                             modification.              
Sec.  106.25(b)(5)........................  Sec.  106.91(a)(3) with     
                                             modification.              
------------------------------------------------------------------------
                       FINISHED PRODUCT EVALUATION                      
------------------------------------------------------------------------
Sec.  106.30(a)...........................  Sec.  106.91(a).            
Sec.  106.30(b)(1)(i).....................  Sec.  106.91(a)(3).         
Sec.  106.30(b)(1)(ii)....................  Sec.  106.91(a) with        
                                             modification.              
Sec.  106.30(b)(2), Sec.  106.30(b)(3)....  Sec.  106.91(b) with        
                                             modification.              
Sec.  106.30(c)(1)........................  Omitted.                    
Sec.  106.30(c)(2)........................  Sec.  106.3(i)              
                                            Secs.  106.91(b)(1) and     
                                             106.97(b)(1) with          
                                             elimination of the         
                                             osmolality and vitamin D   
                                             assay.                     
Sec.  106.30(d)...........................  Omitted.                    
------------------------------------------------------------------------

    FDA is proposing quality control procedures under the authority 
granted by section 412(b)(2), (b)(3), and (b)(4) of the act, which 
direct the Secretary (and by delegation, FDA) to establish by 
regulation the quality control procedures that he or she determines are 
necessary to ensure that an infant formula provides the required 
nutrients at the required levels. In the Congressional Record of 
September 27, 1986, Senator Metzenbaum stated: ``The most important 
provision of this amendment is the simple requirement that each batch 
of formula must be tested for each essential nutrient that must be 
contained in the formula'' (Ref. 1). The quality control procedures in 
proposed subpart C of part 106 are the minimum practices that 
manufacturers must implement to ensure that the infant formula that 
they produce contains the required nutrients at the required levels 
throughout the shelf life of the product. Under section 412(a)(3) of 
the act, an infant formula is deemed to be adulterated if the 
processing of the formula does not comply with quality control 
procedures prescribed by the Secretary.
2. Nutrient Testing
    Proposed Sec. 106.91(a) describes the testing that FDA has 
tentatively concluded each manufacturer must conduct on each batch of 
infant formula to ensure that it provides the required nutrients at the 
required levels and provides any nutrient added by the manufacturer. 
FDA is proposing these requirements under the authority of two sections 
of the act. Section 412(b)(2)(B)(i) of the act provides that the 
quality control procedures shall include requirements for testing, in 
accordance with section 412(b)(3), of each batch of infant formula for 
each required nutrient, before distribution of such batch. Section 
412(b)(3)(D) of the act states that if the Secretary adds a required 
nutrient, the Secretary must require that the manufacturer of the 
infant formula test each batch of such formula for that nutrient in 
accordance with section 412(b)(3)(A), (b)(3)(B), and (b)(3)(C) of the 
act.
    Current Sec. 106.20(a) and (b)(2), which FDA is proposing to 
replace with Sec. 106.91(a)(1), do not require that manufacturers 
analyze nutrient premixes if the premixes come with a supplier's 
guarantee or certification. Proposed Sec. 106.91(a)(1), however, 
requires that each nutrient premix used in the manufacture of an infant 
formula be tested by the formula manufacturer for each nutrient that 
the manufacturer is relying on the premix to provide to ensure that the 
premix complies with the manufacturer's specification. This change is 
required by section 412(b)(3)(B) of the act. Section 412(b)(3)(B) was 
included in the 1986 amendments because infant formula manufacturers 
were increasingly relying on the use of formula premixes, and Congress 
felt that relying on a premix supplier's written assurance that its 
premix product was properly tested was inadequate (Ref. 1). In 1985, 
the Department of Justice sought an injunction against a premix 
supplier because, ``as a result of inadequate quality control, numerous 
* * * vitamin and mineral mixes--used in infant formula--have been 
misbranded and adulterated'' (Ref. 3). The premix supplier entered into 
a consent decree of permanent injunction that enjoined it from shipping 
any of its vitamin/mineral premixes for use in infant formulas until it 
completed a number of specific acts that were designed to improve its 
quality control (Ref. 50).
    FDA is proposing to redesignate current Sec. 106.25(b)(3) as 
Sec. 106.91(a)(2), which requires that after the addition of the 
premix, or at the final-product stage but before distribution, each 
batch of infant formula be tested to confirm that the nutrients 
contained in any nutrient premix used in such infant formula are 
present in each batch of infant formula in the proper concentration. 
This requirement implements section 412(b)(3)(C)(ii) of the act, which 
requires that infant formula be tested to ensure that any nutrient 
premixes used by the manufacturer are actually included in the batch of 
infant formula in the proper amount. Without this check, inadvertent 
failure to include the premix could go undetected, and infant formula 
that is deficient in the nutrients that were to be provided by the 
premix would be introduced into the market.
    Current Sec. 106.30(b)(1)(i) requires that the manufacturer analyze 
representative samples of each batch of finished infant formula for 
specific nutrients to assess process degradation. FDA is carrying 
forward a modified version of this requirement in proposed 
Sec. 106.91(a)(3), which requires that each batch of infant formula be 
tested for vitamins A, C, and

[[Page 36176]]

E and thiamin at the final-product stage, before distribution. This 
regulation is proposed under section 412(b)(3)(A) of the act, which 
states: ``At the final product stage, each batch of infant formula 
shall be tested for vitamin A, vitamin B1, vitamin C, and vitamin E * * 
*.'' In the Congressional Record, Senator Metzenbaum stated that 
testing for these vitamins is required at the final-product stage 
because they are vulnerable to degradation (Ref. 1). Testing at the 
final-product stage will ensure that these nutrients are present in the 
infant formula at the end of all the processing steps that may destroy 
them.
    Proposed Sec. 106.91(a)(4) requires that, before distribution, each 
batch of infant formula be tested for all nutrients required to be 
included, and any others that have been included, but for which testing 
to comply with Sec. 106.91(a)(1) or (a)(3) was not conducted. This 
proposed provision takes a markedly different tack than current 
Sec. 106.30(b)(1)(ii), which states that no analyses are needed for 
linoleic acid, vitamin D, vitamin K, choline, inositol, and biotin 
before release of a batch of infant formula for commercial or 
charitable distribution. This change in approach is necessary because 
section 412(b)(3)(C) of the act, which was added by the 1986 
amendments, states that each batch of formula must be tested for each 
nutrient required by the law to be present in an infant formula. Also, 
manufacturers are adding nutrients not required by Sec. 107.100, such 
as selenium, to infant formulas. These nutrients meet the definition 
for ``nutrient'' in proposed Sec. 106.3(m) because they have been 
identified as essential for infants by NAS through its development of a 
Recommended Dietary Allowance or an Estimated Safe and Adequate Daily 
Dietary Intake range. The agency has not objected to the addition of 
nutrients not required by Sec. 107.100 to infant formulas. However, it 
is important that the level of these added nutrients be controlled, and 
that the level of the added nutrient be consistent from batch to batch 
and be uniform throughout the batch of infant formula.
    The level of a nutrient needs to be controlled because some 
nutrients can be toxic to an infant if given at too high a level. 
Controlling the level of the added nutrient for consistency from batch 
to batch and in a particular batch of infant formula will ensure that 
the infant receives the essential nutrient on a consistent basis and 
will also ensure that the infant does not receive too high, or too low, 
a level of the nutrient because the nutrient was not uniform throughout 
the batch of infant formula.
3. Stability Testing
    Current Sec. 106.30(c) requires that the manufacturer, using 
representative samples collected from finished product batches, conduct 
stability analysis for selected nutrients with sufficient frequency to 
substantiate the maintenance of nutrient content throughout the shelf 
life of the product. The 1986 amendments added subsection 
412(b)(2)(B)(ii) to the act, which requires ``regularly scheduled 
testing, by the manufacturer of an infant formula or an agent of such 
manufacturer, of samples of infant formula during the shelf life of 
such formula to ensure that such formulas are in compliance with'' 
section 412 of the act. To implement this section of the act, the 
agency is redesignating and revising current Sec. 106.30(b)(3) as 
proposed Sec. 106.91(b), which requires quarterly collection of samples 
of infant formula for stability testing to provide a check on nutrient 
stability. This periodic check will alert the manufacturer if nutrient 
stability has changed in some unpredicted way so that the formula no 
longer complies with section 412 of the act. Quarterly testing of 
infant formulas for nutrient stability is currently conducted by the 
industry (Refs. 51 and 52), and the agency is not aware of any problems 
that have resulted from this frequency of testing. The agency requests 
comment on whether this proposed frequency of sample collection for 
stability testing is appropriate.
    The agency has tentatively concluded that this periodic sample 
collection to check on nutrient stability must be performed on a batch 
of each physical form (powder, ready-to-feed, or concentrate) of each 
infant formula, at each different manufacturing facility, because 
different forms of the product may contain different ingredients, and 
different forms of infant formula are subjected to different processing 
procedures. Therefore, ensuring the nutrient stability of one form of 
the product, such as the powder, will not answer questions about the 
nutrient stability of other forms of the product. Thus, the agency has 
tentatively concluded that each form of the infant formula must be 
sampled on a periodic basis for nutrient stability. Also, the agency 
has tentatively concluded that the sampling of one batch of each 
physical form of each infant formula must be conducted at each 
manufacturing facility. This proposed requirement is necessary because 
manufacturers may produce the same infant formula at more than one 
facility, and the manufacturing conditions at one facility may not be 
the same as the conditions at another facility. The differences in 
conditions cannot be allowed to affect the quality of the formula.
    Proposed Sec. 106.91(b) further requires testing at the beginning, 
midpoint, and end of the shelf life of the infant formula. Testing at 
the beginning of the shelf life shows that the formula is in compliance 
with the nutrient requirements of the act when it is released for 
distribution. Testing at the midpoint of the shelf life will alert the 
manufacturer if any nutrient is deteriorating at a rate different from 
that predicted, so that the nutrient may not be in the formula at a 
level to comply with the act throughout the formula's shelf life. 
Testing at the end of shelf life will ensure that the formula contained 
all the nutrients needed to comply with the act throughout its shelf 
life and will provide continued justification for the predicted shelf 
life.
    Additional testing may be necessary to ensure that a formula 
complies with section 412 of the act throughout its shelf life. Such 
testing is likely to focus on a particular nutrient and its stability 
within the matrix of the formulation. This additional testing will 
ensure that, if there is a significant deterioration in the level of 
the nutrient in the formula, the manufacturer will be aware of this 
fact and will be able to take steps promptly to have the product 
removed from the market, before a significant number of infants are 
exposed to a deficient product.
    The agency is not proposing to specify what frequency is required 
because manufacturers have experience with the nutrient stability of 
the infant formula matrices that they produce and are thus in a 
position to determine how frequently testing is necessary. For example, 
the manufacturer is in a position to know whether the nutrient levels 
of a milk-based infant formula need to be tested on a different basis 
than that of a soy-based product, or whether the nutrient levels of an 
infant formula that contains hydrolyzed protein needs to be tested more 
frequently than that of an infant formula that contains non-hydrolyzed 
protein. Manufacturers will be able to comply with section 
412(b)(2)(B)(ii) of the act by testing different nutrients at different 
frequencies. For example, unstable nutrients, such as vitamins, may 
require testing on a more frequent basis than more stable nutrients, 
such as minerals. Proposed Sec. 106.91(b) allows the manufacturers the 
discretion to determine the necessary frequency of testing to ensure 
that their infant formula complies with the nutrient

[[Page 36177]]

requirements of the act, as long as the minimum testing (i.e., at the 
beginning, middle, and end of the shelf life) required by proposed 
Sec. 106.91(b) is accomplished.
    Proposed Sec. 106.91(b)(1) provides for an addition to the 
stability testing required under Sec. 106.91(b). FDA is proposing that 
the first batch of each form of a new infant formula be subjected to 
such testing to ensure that the product complies with the nutrient 
requirements of section 412 of the act throughout its shelf life.
    Proposed Sec. 106.91(b)(2) requires the sampling of the first batch 
of an infant formula in which there has been a change in formulation or 
in processing that could affect whether the formula is adulterated 
under section 412(a) of the act and requires testing of these samples 
for each nutrient that has been, or may have been, affected by the 
change. The change in formulation or processing referred to here would 
not be a ``major change'' because a ``major change'' would mean that 
the formula is a ``new infant formula.'' Examples of the types of 
changes that are subject to proposed Sec. 106.91(b)(2) are: (1) 
Reducing a ``required nutrient'' in a minor way or increasing a 
``required nutrient'' that is subject to maximum limits in Sec. 107.100 
in a minor way; (2) replacing one nutrient form with another form, such 
as replacing vitamin A acetate with vitamin A palmitate or replacing 
calcium carbonate with tricalcium phosphate; (3) changing a time-
temperature condition of preheating, handling, mixing, or sterilizing 
an in-process product; or (4) changing the oxygen content of a packaged 
product that might have a minimal effect on the level of nutrients. 
Requiring sample collection for stability testing when a manufacturer 
makes changes such as these in the manufacture of the product will 
ensure that the manufacturer can verify the predicted shelf life of the 
changed formula.
    Proposed Sec. 106.91(b)(2) requires that the manufacturer ensure 
that the infant formula meets all the nutrient requirements of section 
412 of the act. This provision is proposed under the authority of 
section 412(b)(2)(A) of the act, which provides for the establishment 
of CGMP's for infant formulas, including quality control procedures 
that are necessary to assure that the infant formula provides nutrients 
in accordance with section 412 (b) and (i) of the act, as well as 
section 412(b)(2)(B)(ii). If the formulation or processing of the 
infant formula has been changed, the manufacturer must consider what 
nutrients may have been affected by the change and test for each of 
these nutrients in the final-product stage of the first batch of the 
changed formula. For example, if the manufacturer makes a change in the 
amount of a protein source used in the infant formula, the firm must 
test the formula for protein content and for any nutrients provided 
endogenously to the formula by the protein, such as minerals like 
calcium and phosphorus. The manufacturer is aware of how much of each 
mineral it is relying on the protein source to provide to the formula. 
When the amount of the protein source used in the formula is changed, 
the manufacturer must test for the level of all nutrients it relies on 
the protein source to provide to the formula to ensure that all 
nutrients in the formula meet the requirement of Sec. 107.100.
4. Quality Control Records
    Proposed Sec. 106.91(c) requires that manufacturers make and retain 
records of the results of all testing performed on the batch of infant 
formula in accordance with proposed Sec. 106.100(e)(5)(i) and a full 
description of the methodology used in accordance with proposed 
Sec. 106.100(f)(7). As discussed in the description of the proposed 
revisions to subpart F of part 106, FDA has authority to require these 
records under section 412(b)(4)(A)(i) of the act. Providing a record of 
the results of quality control testing will verify that each nutrient 
required by Sec. 107.100 is present in each batch of infant formula at 
the required level, and that any nutrients added by the manufacturer 
are present at the appropriate level. These records will show the 
levels of nutrients in the formula and will provide data needed to 
evaluate a batch of infant formula if problems, such as adverse events 
in infants, occur later with that particular batch. Records that 
describe the full methodology used to conduct the quality control 
testing will provide consistency in the procedure that the manufacturer 
is using to test for the nutrients in each batch of infant formula, 
even when different laboratory personnel are conducting the testing. 
The accuracy and reproducibility of quality control testing depend on 
the procedure used to conduct the test.
5. Audits of Quality Control Procedures
    Proposed Sec. 106.92 requires that the manufacturer of an infant 
formula, or an agent of such a manufacturer, conduct regularly 
scheduled quality control audits to ensure that an infant formula 
provides required nutrients and has been manufactured in a manner 
designed to prevent adulteration. Proposed Sec. 106.92 derives from 
section 412(b)(2)(B)(iv) of the act, which requires that the quality 
control procedures prescribed by the Secretary include ``the conduct by 
the manufacturer of an infant formula or an agent of such manufacturer 
of regularly scheduled audits to determine that such manufacturer has 
complied with the regulations prescribed under'' section 412(b)(2)(A) 
of the act (stating that the Secretary (and FDA by delegation) 
establish by regulation ``quality control procedures that the Secretary 
determines are necessary to assure that an infant formula provides 
nutrients in accordance with'' section 412 (b) and (i) and ``is 
manufactured in a manner designed to prevent adulteration of the 
formula''. FDA is proposing to require that regularly scheduled audits 
be part of quality control procedures because such audits will document 
compliance with the quality control procedures and will identify 
recurring problems that may dictate an alteration in the master 
manufacturing order. For example, regularly scheduled audits of the 
results of tests of nutrient levels in infant formulas and of any 
deviations from the manufacturer's specifications or procedures for 
acceptable nutrient levels will reveal deviations that occur on a 
repeated basis and will enable the manufacturer to identify 
specifications or procedures that should be reassessed.
    Proposed Sec. 106.92 further requires that the audits be performed 
by an individual who, as a result of education, training, and 
experience, is knowledgeable in all aspects of infant formula 
production and of the agency's regulations concerning quality control 
procedures, but who has no direct responsibilities for the matters 
being audited. The legal authority for this provision, the importance 
of the responsible individual's knowledge in all aspects of infant 
formula production and the agency's regulations, and the need for the 
audit to be performed by an individual who has no direct responsibility 
for the matters being audited were discussed previously under the 
proposed CGMP regulations in Sec. 106.90.
    By proposing different regulations (proposed Secs. 106.90 and 
106.92) that require audits of CGMP and of quality control procedures, 
the agency is not suggesting that it will require that separate audits 
be conducted. These regulations are being proposed separately to make 
clear that the regularly scheduled audits required by section 
412(b)(2)(B)(iv) of the act are an aspect both of CGMP and of quality 
control procedures. The agency would have no objection to a combined 
audit

[[Page 36178]]

of CGMP and of quality control procedures.
6. Revocation of the Requirement for Determination of Vitamin D by the 
Rat Bioassay Method
    FDA is proposing to revoke the requirement in current 
Sec. 106.30(c)(2) for the determination of vitamin D by a rat bioassay 
method. This rat bioassay for vitamin D is no longer a viable assay 
because appropriate animals for conducting this test are difficult to 
acquire (Ref. 53), and an alternate analytical method for the 
determination of vitamin D in infant formulas has been approved by the 
Association of Official Analytical Chemists (Ref. 54).

D. Conduct of Audits

    Section 412(b)(2)(B)(iv) of the act provides that CGMP and quality 
control procedures include regularly scheduled audits to determine 
whether the manufacturer is complying with CGMP, including following 
the quality control procedures that are necessary to ensure that an 
infant formula provides the required nutrients at the required levels, 
and whether it is operating in a manner designed to prevent 
adulteration of the formula. FDA is proposing to require in 
Sec. 106.94(a) that manufacturers develop and follow a written audit 
plan that is available at the manufacturing facility for FDA 
inspection. A written audit plan is necessary to provide consistency in 
how audits are conducted and to ensure that the auditor can determine 
whether the facility is operating in compliance with the applicable 
procedures.
    Proposed Sec. 106.94(b) requires that the audit plan include the 
procedures that the manufacturer uses to determine whether the facility 
is operating in accordance with CGMP, with the applicable quality 
control procedures, and in a manner designed to prevent adulteration of 
the infant formula it produces. This proposed requirement derives from 
current Sec. 106.100(j), which defines audit procedures as the methods 
used to review the manufacturing and quality control procedures and is 
intended to direct the manufacturer's attention to the fundamental 
goals of the manufacturing process in formulating its audit plan.
    Proposed Sec. 106.94(c) sets out the minimum requirements for the 
audit procedures that are to be employed by manufacturers. Under 
proposed Sec. 106.94(c)(1) these procedures are to include a review of 
how the production and in-process control system established under 
Sec. 106.6(b) is operating. In particular, proposed 
Sec. 106.94(c)(1)(i) specifies that the evaluation of the production 
and in-process control system include observation of the production of 
infant formula and a comparison of the observed process to the written 
production and in-process control plan required under proposed 
Sec. 106.6(b). FDA has tentatively concluded that such observations 
will show whether the production and in-process control system is being 
followed appropriately, and, if not, they will identify any deviations 
from the production and in-process control system, so that the 
manufacturer can take corrective actions to ensure that infant formula 
is produced in compliance with the production and in-process control 
system.
    Proposed Sec. 106.94(c)(1)(ii) requires that the evaluation of the 
production and in-process control system include a review of records of 
the monitoring of points, steps, or stages where control is deemed 
necessary to prevent adulteration. As discussed below, proposed 
Sec. 106.100(e)(3) requires that the batch production and control 
records document the monitoring of all points where control is deemed 
necessary to prevent adulteration in the manufacturing of the batch. 
FDA has tentatively concluded that proposed Sec. 106.94(c)(1)(ii) is 
necessary because the auditor can observe the production of only a 
limited number of batches of infant formula. A review of the production 
and in-process control records of all batches produced in a given 
period of time will ensure that the production and in-process control 
system is working appropriately on a continuous basis, will identify 
any point that monitoring reveals is out of control on a recurring 
basis, and will identify where the production and in-process control 
system needs improvement.
    Proposed Sec. 106.94(c)(1)(iii) requires that the evaluation of the 
production and in-process control system include a review of records of 
how deviations from any standard or specification at points, steps, or 
stages where control is deemed necessary to prevent adulteration were 
handled. As discussed below, proposed Sec. 106.100(e)(4)(iii) requires 
that the batch records include the conclusions and followup of an 
investigation of the failure to meet any specification or standard at 
any point where control is deemed necessary to prevent adulteration. A 
review of these records as a part of the audit will identify failures 
that occur on a repeated basis and will show how these failures are 
handled by the manufacturer. The auditor will be able to evaluate 
whether the conclusions and followup of these investigations are 
appropriate for each failure to meet the specification or standard.
    Proposed Sec. 106.94(c)(2) requires that the audit procedures 
include a review of a representative sample of all records maintained 
in accordance with proposed Sec. 106.100 (e) and (f). As discussed 
below, proposed Sec. 106.100(e) sets out the requirements for the batch 
production and control records, and proposed Sec. 106.100(f) sets out 
the requirements for records related to observance of CGMP. A review of 
a representative sample of these records will show the auditor whether 
there has been compliance with the appropriate regulations in producing 
the batches of product so that the formula is not adulterated. Section 
412(b)(2)(B)(iv) of the act states that the audit is conducted to 
determine whether the manufacturer has complied with the regulations 
establishing CGMP for infant formulas, including quality control 
procedures. FDA has tentatively concluded that review of a 
representative sample of the records maintained in accordance with 
Sec. 106.100 (e) and (f) is necessary to determine whether the 
manufacturer is complying with these regulations.

E. Quality Factors for Infant Formulas

1. What Are Quality Factors?
    The agency is proposing to create a new subpart E to implement the 
quality factor requirements of sections 412 (a)(2) and (b)(1) of the 
act. Section 412(a)(2) of the act states that an infant formula is 
adulterated unless it meets the quality factor requirements that are 
established under section 412(b)(1). Section 412(b)(1) of the act 
states that the Secretary shall by regulation establish requirements 
for quality factors, including quality factor requirements for required 
nutrients for infant formulas to the extent possible consistent with 
current scientific knowledge. Therefore, it is incumbent on 
manufacturers to establish that the infant formula that they produce 
meets the minimum quality factor requirements that FDA adopts.
    What Congress meant by ``quality factors'' is discussed in the 
report of the House Committee on Interstate and Foreign Commerce that 
accompanied the 1980 act. The report states that quality factors 
``pertain to the bioavailability of a nutrient and the maintenance of 
levels or potency of nutrients during the expected shelf life of the 
product'' (Ref. 5). FDA, in proposed Sec. 106.3(o), has defined 
``quality factors'' in a manner that encompasses several basic 
concepts, including the concepts of

[[Page 36179]]

``bioavailability'' and of ``healthy growth.''
    The concept of ``healthy growth'' was discussed in the report of 
the House Committee on Interstate and Foreign Commerce that accompanied 
the 1980 act. The report states that infant formulas are often the sole 
source of nutrition for infants, and that ``the growth of infants 
during the first few months of life often determines the pattern of 
development and quality of health in adult life'' (Ref. 5). FDA 
considers the concept of ``healthy growth'' to be broad, encompassing 
all aspects of physical growth and normal maturational development, 
including maturation of organ systems and achievement of normal 
functional development of motor, neurocognitive, and immune systems. 
All of these growth and maturational developmental processes are major 
determinants of an infant's ability to achieve his/her biological 
potential, and all can be affected by the nutritional status of an 
infant.
    ``Bioavailability'' of a nutrient for an infant means that the 
nutrient is physiologically available in sufficient quantities to 
perform its metabolic functions (Ref. 55). In a formula product, 
bioavailability of individual nutrients is affected by the net effect 
of the formulation and processing of the product on the chemical form 
of the nutrient. These processes are influenced by such factors as the 
chemical form of the nutrient in the ingredient source, the chemical 
form of the nutrient after processing, and the net effect of various 
inhibitors and enhancers in a food or meal on the chemical form of the 
nutrient and its ability to be absorbed and utilized by the infant. In 
the infant, the bioavailability of a nutrient is determined by the net 
effect of the amount of nutrient that is converted during digestion to 
an absorbable form, the proportion of the nutrient that is absorbed 
into the bloodstream, the proportion of the absorbed nutrient that is 
converted to its biologically useful form, and the proportion that is 
lost through excretory processes (Ref. 55). Bioavailability varies 
among nutrients within a given food product and, for a given nutrient, 
among foods. The factors affecting nutrient bioavailability are complex 
and can be difficult to predict based on analyzed nutrient values 
alone.
    Bioavailability issues are particularly critical for infants during 
the first few months of life, where a single food (infant formula) 
serves as the sole source of all nutrients at a period when rapid 
physical growth and development and maturation of various organ systems 
makes the infant particularly vulnerable to harm by nutritional 
insults. Unlike the mixed diet of persons beyond infancy where poor 
bioavailability in one food can be compensated for by other foods in 
the diet, a problem with bioavailability in an infant formula affects 
the total amount of nutrient available to that infant for several 
months after birth. Furthermore, requirements for nutrients are higher 
per kilogram body weight during early infancy than at any other time 
during the life cycle. Because numerous critical developmental 
milestones (e.g., neurocognitive or immune functions) must be achieved 
by young infants, a nutrient insufficiency during infancy can quickly 
develop into serious, and in some cases, permanent adverse effects on a 
range of developmental processes, including physical growth and organ 
maturation. Thus, a problem with bioavailability is far more critical 
for a food such as infant formula than it is for foods that are used as 
part of a mixed diet by the general population.
    Furthermore, the rapidly changing and increasingly complex 
physical, chemical, and biologically significant characteristics of 
ingredients used in new and reformulated infant formulas make it 
important to continually ensure that quality factor requirements are 
met. Changes in formulation of infant formulas are made by 
manufacturers for a variety of reasons, including enhancing the 
functional characteristics of the formula (e.g., to prevent separation 
of ingredients or to prevent clumping that will plug nipples on 
bottles), to enhance digestibility of the formula (e.g., different 
sources or blends of fats), or to improve the nutritional quality 
(e.g., a different source of protein or of a vitamin or mineral, or 
adding a nonrequired nutrient such as selenium). For example, in some 
formulas, novel sources of vegetable oils (e.g., fractions of plant 
oils that are particularly rich in certain types of fatty acids) have 
partially or fully replaced cow's milk fat as the fat source (Refs. 56 
and 57). Whey proteins or highly processed proteins (e.g., hydrolyzed 
proteins) are now frequently used as partial or complete replacements 
for more traditional cow's milk protein sources. In other cases, 
nutrient/nutrient interactions (e.g., high iron inhibiting absorption 
of zinc) or nutrient/ingredient interactions (e.g., phytates from soy 
protein isolates inhibiting absorption of zinc, or the replacing of the 
milk sugar (lactose) that enhances absorption of calcium with a sugar 
source that does not have this ability) can adversely affect nutrient 
availability.
    New processing methods may also have unintended consequences when 
used with established ingredients or formulations. For example, a new 
processing method that subjects the formula to conditions that are less 
denaturing to cow's milk proteins than traditional heat treatments 
could produce a formula that is less digestible and that causes 
reactivity of the gastrointestinal wall, such as has been seen with 
whole cow's milk (Ref. 58).
    In summary, consideration of quality factors goes beyond analytical 
measures of the presence or absence of a nutrient in the formula 
product and is needed to provide assurance that adverse effects on the 
nutritional value of the formula for the infant do not unintentionally 
or unknowingly occur as a result of the formulation or the processing 
of an infant formula. Chemical analysis of the formula product to 
define its nutrient composition often overestimates the amount of 
nutrient that is bioavailable for physiological use by the infant. The 
quality factors, therefore, provide a means of evaluating whether a 
nutrient has become less bioavailable than would be expected, so that 
it is not sufficiently effective to meet its normal nutritive 
functions, or whether its bioavailability has been enhanced to a level 
that raises safety concerns.
    Quality factor requirements are distinctly different from quality 
control procedures. While ``quality control procedures are intended to 
insure that the safety and nutritional potency of a formula is built 
into the manufacturing process'' (Ref. 5), quality factors are intended 
to ensure that an infant formula contains an adequate amount of each 
nutrient in a form that can be digested, absorbed, and utilized so that 
the infant's physiological needs for these nutrients will be met (Ref. 
5). Changes in ingredient sources and processing can affect the 
chemical forms of nutrients in the formula product. Such changes can 
affect the digestion and absorption of food nutrients such that: (1) 
Absorption is incomplete, (2) absorbed nutrients are not in a form that 
allows use by metabolic pathways, or (3) the nutrient may interact with 
other dietary substances to cause excessive excretion. Thus, the amount 
of nutrients (i.e., the analyzable amounts) in formulas must generally 
be higher than the physiological requirements of infants (i.e., the 
amounts of nutrients needed by the body to meet metabolic and growth 
needs of infants). Although these inefficiencies are generally taken 
into account when recommending nutrient levels for infant formulas, 
there

[[Page 36180]]

is always the potential for affecting nutrient bioavailabilities in 
unexpected ways.
    In summary, a demonstration that both the quantitative and quality 
factor requirements for essential nutrients in an infant formula are 
met is necessary to ensure that the infant formula is likely to meet 
all of the known physiological nutritional needs of infants and to 
ensure that healthy growth and nutritional well-being will be achieved 
by an infant consuming the infant formula as the sole source of 
nutrition.
2. Identification of Quality Factors
    In testimony before the passage of the 1986 amendments, the agency 
informed the Senate that the state of knowledge and science with 
respect to quality factors was still evolving, and that, therefore, 
there was a basis for only one quality factor for a nutrient. (Although 
the testimony to the Senate does not specify the identity of the 
nutrient for which there was a basis for a quality factor, the quality 
factor was the protein efficiency ratio used for assessing protein 
quality (Ref. 1).) Senator Metzenbaum stressed that the amendments 
contemplated that additional quality factors would emerge, and that the 
Secretary should implement requirements for such factors as quickly as 
scientific advances would allow.
    The agency subsequently took a major step toward establishing 
quality factors through a contract in 1986 with the CON/AAP. The AAP 
earlier had published recommendations regarding the quantities of 
nutrients needed in infant formulas (Ref. 59). These recommendations 
were relied upon during the development of the nutrient specifications 
of the act (Ref. 60). In its report to FDA, ``Clinical Testing of 
Infant Formulas with Respect to Nutritional Suitability for Term 
Infants'' (Ref. 6), the CON/AAP identified those conditions in which 
changes in formula composition warranted clinical testing. The CON/AAP 
stated that ``clinical testing is primarily useful for determining (1) 
acceptability of the formula, (2) ability of the formula to support 
normal growth, and (3) availability of selected nutrients.'' The CON/
AAP also discussed the limitations of the available measurements, 
providing an assessment of the limits of scientific knowledge.
    The agency has considered the CON/AAP report carefully and has also 
considered new scientific information published since the release of 
that report to determine what quality factors are appropriate for 
nutrients in infant formula. Based on its consideration, FDA is 
proposing to adopt Sec. 106.96. This section, if adopted, will require 
that all infant formula be of sufficient quality that it meets the 
nutritional requirements of infants for healthy growth when fed as the 
sole source of nutrition, as indicated by a general quality factor for 
physical growth, assessed using anthropometric measures of infants 
consuming the formula, and by a nutrient-specific quality factor for 
protein biological quality, assessed by an animal bioassay using the 
formula.
    The agency is not proposing to require that manufacturers measure, 
individually, the absorption, metabolism, metabolic transformation, or 
utilization of any of the other essential nutrients. These measures are 
often technically difficult or unavailable, difficult to interpret, or 
invasive, thus causing unnecessary testing of infants without potential 
for providing meaningful results. Rather, the agency has tentatively 
concluded that current scientific knowledge and ethical and practical 
considerations are supportive only of requiring two quality factor 
measures: (1) Physical growth of infants consuming the formula as an 
integrative indicator of the net effect of the overall nutritional 
quality of the formula, and (2) a rat bioassay of protein quality in 
the formula product to ensure that the infant's needs for individual 
amino acids will be met.
    The agency has tentatively determined that these are minimum 
requirements. The agency recognizes that, on a case-by-case basis as 
warranted by the formulation and intended use of a particular infant 
formula, demonstration of additional quality factors may be necessary. 
For example, a formula intended for use by premature infants who are at 
a particularly vulnerable developmental stage relative to nutritional 
needs to support neurocognitive development may need to be subject to 
testing that includes measurement of this endpoint to ensure that the 
formula supports healthy growth. In addition, a formula in which a 
novel fatty acid has been added to enhance the formula's ability to 
meet nutritional needs for supporting visual development may need to be 
evaluated to determine whether it has adverse nutritional effects on 
other aspects of healthy growth (e.g., on development of immune 
function).
3. The Regulation
    Proposed Sec. 106.96(a) sets forth quality factor requirements that 
reflect the minimum measures needed to evaluate the nutritional quality 
of an infant formula product, taking into account current scientific 
knowledge and the usefulness of the outcome measures for evaluating 
quality factors, while minimizing unnecessary testing of infants 
serving as subjects in clinical trials. Infant formula is defined in 
the act as a complete or partial substitute for human milk (section 
201(aa) of the act). Obviously, the greatest need for a nutritionally 
complete formula that meets all quality factors is when the formula is 
used as a complete substitute for human milk. When no other form of 
nutriture is available to the infant, the formula must provide all of 
the nutrients needed for the healthy growth of the infant. There is no 
room for error or miscalculation. The absence or an inadequate level of 
an essential nutrient will be evidenced by growth failure and other 
signs or symptoms resulting from nutritional insufficiencies. FDA has 
tentatively concluded, therefore, that an evaluation of the ability of 
a formula to support healthy growth must be made under its most 
demanding conditions of use, i.e., when it is used as the sole source 
of nutrition, because other foods may mask or compensate for 
deficiencies in the formula that would occur if the formula were used 
as a complete substitute for human milk, which would produce results 
that cannot be meaningfully interpreted.
    Proposed Sec. 106.96(b) identifies ``normal physical growth'' as a 
quality factor. This quality factor reflects the CON/AAP recommendation 
that the determination of physical growth rate is the most valuable 
component of the clinical evaluation of an infant formula (Ref. 6). 
Physical measures of growth such as weight gain are the most widely 
accepted and used markers of a young infant's overall ability to digest 
and utilize those nutrients provided by the formula. The very rapid 
rate of growth in early infancy means that abnormalities in growth rate 
can be detected in a few months, providing an easily measured and 
sensitive, although nonspecific, indication of nutritional 
insufficiencies (Ref. 4). Physical measures of growth rate are easily 
done, are familiar to both parents and health professionals, and are a 
normal part of routine office visits. They are noninvasive and pose 
little or no risk to infants and provide meaningful results for 
evaluating the ability of an infant formula to support physical growth 
in very young infants. Thus, the agency has tentatively concluded that 
the ability of the formula, when fed as a sole source of nutrition, to 
meet the nutritional requirements of young infants for normal physical 
growth is a

[[Page 36181]]

necessary indicator of the overall nutritional quality of the formula.
    Proposed Sec. 106.96(c) requires that the protein in infant 
formulas be of sufficient biological quality to meet the protein 
nutritional requirements of infants. Protein, while generally discussed 
as a single nutrient, depends for its nutritive value on the inclusion 
of all essential amino acids at levels and relative proportions needed 
to support healthy growth. The protein requirement is really the sum of 
different requirements for 10 essential amino acids that occur at 
different levels and proportions in various food protein sources. 
Protein quality is also affected by differences in digestibility of 
different protein sources, by factors that modify digestion, and by 
chemical reactions that affect the ability of enzymes in the infant's 
gastrointestinal tract to digest and absorb the amino acids in the 
protein source. Once absorbed, the relative proportions of the amino 
acids can affect their uptake by body tissues because of competition 
for receptors and transport systems. Thus, protein quality depends on a 
number of complex interactions and conditions that can be difficult to 
predict.
    Chemical analysis of foods generally only measures the amount of 
total protein present and does not identify specific amino acids or 
their ability to meet the physiological needs of infants for the 
essential amino acids. Chemical analysis alone, therefore, is not 
capable of predicting whether adequate amounts of all essential amino 
acids are present, or whether the amino acids present are able to 
support healthy growth in infants. Yet ensuring that the protein in an 
infant formula's is of high biological value is critical to an infant's 
health. For example, during the first year of life, the protein content 
of an infant's body increases from 11 to 15 percent at the same time 
that the infant's body weight increases by 7 kg. The average increase 
in body protein is about 3.5 g/day for the first 4 months of life and 
about 3.1 g/day for the next 8 months. These protein requirements must 
be met by a formula that not only contains adequate protein but also 
contains protein of high biological quality in a form that can be 
utilized by the infant. Because biological quality varies among protein 
sources and may be adversely affected by processing methods and other 
constituents present in the formula, the agency has tentatively 
concluded that the biological quality of the protein in an infant 
formula is a necessary quality factor. This quality factor will require 
an evaluation of whether the formula contains the essential amino acids 
and total nitrogen in the amounts and proportions necessary to permit 
normal tissue and organ growth and development. As discussed later in 
this document, the agency is proposing in Sec. 106.97(b) that the 
biological quality of the test protein be measured by the Protein 
Efficiency Ratio (PER) rat bioassay and be comparable to the biological 
quality of the milk protein casein.
    Proposed Sec. 106.96 does not include quality factor requirements 
for all nutrients required by infants because methods to determine 
whether these requirements are met are not available or are not 
practical for most nutrients (e.g., results cannot be meaningfully 
interpreted, or methods are invasive, thus causing unnecessary testing 
of infants). Nonetheless, FDA has tentatively concluded that, as the 
science evolves, establishing quality factor requirements for other 
nutrients needed by infants would provide assurance, beyond that 
provided by the general quality factor of physical growth in proposed 
Sec. 106.96(b) and the specific protein quality factor in 
Sec. 106.96(c), that a formula will meet the overall nutritional needs 
of infants. As the science evolves, FDA anticipates being able to 
progress beyond generalized, nonspecific indicators of overall 
nutritional intakes (e.g., measures of physical growth), to more 
specific and sensitive measures of biochemical and functional 
nutritional status. FDA also has tentatively concluded that, on a case-
by-case basis, additional quality factors may be needed for a specific 
formula product if formulation or processing concerns raise sufficient 
quality factor questions such that additional measures are necessary to 
adequately ensure that the nutritional quality of the formula supports 
healthy growth. FDA asks for comment on criteria as to when such 
measures are required.
4. Request for Comment on Need for Establishing Requirements for Other 
Quality Factors
    Proposed Sec. 106.96(b) and (c) set forth minimum requirements for 
quality factors (physical growth and protein quality) that all infant 
formulas should meet. FDA has tentatively concluded that these quality 
factors are consistent with current state-of- the-art science and 
provide significant information on the nutritional quality of the 
infant formula without requiring unnecessary or meaningless testing of 
infant enrollees in studies.
    As discussed above, the 1986 amendments contemplated that when 
scientific research identified criteria that could be used to establish 
quality factors for specific nutrients in infant formula, the agency 
would establish quality factor requirements for those nutrients. 
Proposed Sec. 106.96 will establish two quality factors (physical 
growth and protein quality) because the agency has tentatively 
concluded that there is sufficient scientific evidence of the 
importance of these quality factors, and because adequate methods exist 
to meaningfully and ethically measure these factors.
    However, the CON/AAP report discussed other nutrients necessary for 
healthy growth of infants and for which the report recommended 
establishing quality factor requirements (Ref. 6). The agency has 
studied the evidence supporting the establishment of quality factor 
requirements for these other nutrients, and the methods available for 
determining whether an infant formula meets quality factor requirements 
for these nutrients. FDA has tentatively concluded that establishing 
quality factor requirements for the three additional nutrients 
recommended by CON/AAP (i.e., (a) fat, as measured by fat balance; (b) 
calcium and phosphorus, as measured by calcium and phosphorus balance; 
and (c) iron as measured by iron bioavailability) is not warranted at 
this time. FDA, however, solicits additional information that it will 
consider before reaching a final decision on whether the scientific 
evidence and usefulness of results are sufficient to support 
establishing these additional quality factor requirements. Therefore, 
the agency requests comments and information on: (1) The scientific 
evidence on the importance of the amount, type, and sources of fat, 
calcium and phosphorus, and iron in infant formula, and (2) the 
appropriate methods and interpretative criteria to determine whether an 
infant formula meets the nutritional requirements for fat, calcium and 
phosphorus, and iron of infants consuming the formula as the sole 
source of nutrition. The basis upon which the agency is considering 
establishing quality factor requirements for these nutrients is 
discussed below.
    a. Fat. The agency requests comment on a quality factor for fat 
balance that would require that all infant formulas be formulated and 
manufactured to provide fat in a manner that allows the fat to be 
absorbed and retained by infants at a level that the energy and other 
nutritional requirements of the infant are not adversely affected (Ref. 
6). Normal, healthy, full-term infants fed various mixtures of the fats 
traditionally used in infant formulas in the United States rarely 
excrete more than 15 percent of their fat intake (Ref. 6). This level 
of fat excretion is an indication

[[Page 36182]]

that the fat is highly digestible. The use of a fat with lower 
digestibility would adversely affect energy balance, could reduce the 
absorption of fat-soluble vitamins and other nutrients, and could have 
a negative impact on healthy growth of the infants.
    b. Iron. The agency solicits comment on a quality factor that would 
require that all infant formula be formulated and manufactured such 
that the iron used is bioavailable and meets the iron requirements of 
the growing infant. The maintenance of adequate iron status in the 
infant is important because iron is required to transport oxygen in the 
red blood cells to body tissues (as a component of hemoglobin), to 
supply oxygen to muscle tissue (as a component of myoglobin), and to 
support normal mental development. Full-term infants are generally born 
with adequate iron stores to meet their iron needs for the first few 
months of life, but the iron needs of premature infants and older 
infants must be met by the diet.
    Iron bioavailability from infant formulas is low compared to the 
iron bioavailability from human milk (Refs. 61 and 62).
    Nutrient sources and other ingredients, such as protein sources, 
can affect the chemical form of iron, thus interfering with its 
potential for absorption (Ref. 63). Furthermore, factors that enhance 
iron bioavailability from human breast milk are poorly understood and 
currently are not present in commercial formulas. Consequently, infant 
formulas are fortified with up to 10 times the amount of iron found in 
human milk. If, however, the bioavailability of the iron in the infant 
formula is substantially improved by a change in the formulation or 
processing of the formula, then reductions in the amounts of iron added 
to the infant formula may be necessary to prevent the infant from 
absorbing excessive amounts of iron which could be unsafe because high 
dietary intakes of iron can adversely interfere with the 
bioavailabilities of other nutrients (59 FR 51030, October 6, 1994). 
If, however, the iron was bound to another ingredient such that it 
interfered with absorption, the infant's physiological needs for iron 
might not be met. Infant formula iron levels and iron bioavailability, 
thus, represent a delicate balance between effectiveness and safety 
that cannot be adequately predicted by chemical analysis of the iron 
content of the formula, but can best be assessed by measurement of 
clinical indicators of iron status.
    Early changes in iron nutritional status are not likely to be 
detected by the general quality factor of physical growth. Therefore, a 
quality factor requirement for an infant formula to meet the iron 
requirements of infants, and to contain sufficient bioavailable iron 
for this purpose, may be needed. The agency, however, is concerned that 
clinical studies, as described in proposed Sec. 106.97(a), in which 
selection criteria include requirements that enrollees be healthy, 
full-term infants aged 0 to 4 and 5 months, may not be sensitive enough 
to detect significant differences in iron bioavailability of a formula 
product. Healthy, full-term infants are usually born with adequate iron 
stores to maintain normal iron status for the first 3 to 4 months of 
life--the period of time that a clinical trial would be conducted. 
Without assurance that the test results are meaningful, the agency has 
tentatively decided not to require a specific quality factor for iron 
bioavailability.
    c. Calcium and phosphorus. The agency also requests comment on a 
quality factor that would require that all infant formulas be 
formulated and manufactured such that the calcium and phosphorus are 
bioavailable and meet the calcium and phosphorus needs of infants. 
Calcium and phosphorus are essential for healthy bone mineralization 
and growth in infants. Calcium bioavailability is of particular concern 
because inadequate intakes of calcium impair bone mineralization and 
can cause rickets in severe cases (Refs. 64 and 65).
    Interactions with other ingredients and manufacturing processes can 
reduce calcium and phosphorus bioavailability. High concentrations of 
calcium and phosphorus can interact to form insoluble complexes that 
may be unavailable (Ref. 66). Calcium can interact with free fatty 
acids and form soaps that are not absorbed (Ref. 66). Lactose-free 
formulas have been found to have lower calcium absorption than formulas 
containing this sugar (Refs. 67 and 68).
    Some phosphorus compounds, such as the phytates found in plant 
protein sources, may not be readily digested and absorbed by infants 
(Ref. 69). Inadequate dietary phosphorus can cause a loss of calcium 
from the body as a result of bone resorption (i.e., loss of bone mass) 
(Ref. 70). Formulation or processing changes that affect other formula 
ingredients that influence calcium and phosphorus absorption require 
careful consideration of their potential effects on calcium and 
phosphorus bioavailability and the calcium and phosphorus status of the 
infant.
    A dietary insufficiency of calcium and phosphorus of a magnitude 
that decreases bone formation may not be detected by physical measures 
of growth (Ref. 71). Therefore, a quality factor requirement for an 
infant formula to ensure that it meets the calcium and phosphorous 
requirements of infants, and to ensure that it contains sufficient 
bioavailable calcium and phosphorus for this purpose, may be needed. 
FDA is concerned, however, that meaningful measures for assessing the 
bioavailability of calcium and phosphorus may not be available.
    d. Summary. FDA has tentatively concluded that the clinical and 
nutritional sciences have not reached a state where specific tests are 
available that would permit manufacturers to establish that they meet 
quality factors for each of the essential nutrients listed in 
Sec. 107.100, except for protein. Therefore, except for the quality 
factor requirements for physical growth and protein quality discussed 
above and set forth in proposed Sec. 106.96 (b) and (c), the agency has 
tentatively concluded that it is not useful to propose quality factor 
requirements for specific nutrients at this time.
    Thus, to meet the nutritional needs of infants consuming formula, 
manufacturers must use forms or sources of essential nutrients that are 
bioavailable. The agency is concerned that manufacturers could 
unintentionally or unknowingly use forms of nutrients that have a 
relatively low bioavailability or ingredients or processing methods 
that will produce interactions that adversely affect the 
bioavailability of nutrients, thereby adulterating the formula because 
it no longer meets the nutritional needs of the infant. However, at 
this time, FDA is not aware of a means to systematically identify those 
circumstances that could adversely affect all nutrient 
bioavailabilities. FDA does not believe that it is ethical to 
unnecessarily subject infants to testing protocols when meaningful 
results cannot be assured. However, because of the potential 
seriousness of the public health impact of not meeting quality factors, 
FDA also believes that it is desirable to establish additional quality 
factors, as soon as they are warranted by evolving scientific 
knowledge, to ensure adequate nutrient bioavailability.
    FDA, therefore, requests comment on the: (a) Need for routine 
testing of quality factors, in addition to measures of physical growth 
and protein quality; (b) criteria to be used in determining that such a 
need can be meaningfully implemented, and (c) if a need is established, 
the type of qualitative and quantitative measurements that could be 
used by manufacturers to demonstrate

[[Page 36183]]

that an infant formula meets with those quality factors. If FDA 
receives information demonstrating the need for additional quality 
factors, it will consider including them in any final rule that results 
from this proceeding.
5. Assurances for Quality Factors
    a. Quality factor--physical growth of infants. Proposed 
Sec. 106.97(a)(1) requires that the manufacturer conduct an adequate 
and well-controlled clinical study to determine whether the formula 
supports normal physical growth in infants when it is fed as the sole 
source of nutrition. The CON/AAP Task Force on Clinical Testing of 
Infant Formulas (Ref. 6) concluded that the capability to support 
physical growth is the most widely accepted and used measurement 
available of the nutritional adequacy of an infant formula. Gains in 
weight and length of young infants reflect the long-term, integrative 
physiological processes that can only be achieved if the infant's 
nutritional needs are met.
    A randomized, controlled study represents the most sensitive type 
of study to measure the nutritional adequacy of infant formula. The use 
of concurrent treatment and control groups is in agreement with the 
CON/AAP Task Force recommendations (Ref. 6) and with the agency's 
recommendations for human bioavailability studies of drugs (21 CFR 
320.25). Although comparisons to historical controls (e.g., population 
reference standards) have been used by some investigators to evaluate 
growth of infants consuming a particular formula product, this type of 
study lends itself to misleading results because population reference 
standards are generally for the total population of infants (regardless 
of birth weight, health status, socioeconomic status, or other factors 
that can affect growth unrelated to nutritional components). In a study 
to evaluate the nutritional adequacy of a formula, on the other hand, 
selection criteria are usually used to limit enrollment to healthy, 
full-term infants. Thus, differences or similarities in growth between 
study infants and population reference standards cannot be meaningfully 
interpreted. Therefore, the agency is proposing to require that 
adequate and well-controlled clinical studies be conducted to collect 
the data needed to determine whether a formula satisfies the quality 
factor requirements for physical growth. To assist manufacturers in 
understanding the general principles for adequate and well-controlled 
clinical studies, FDA has prepared the ``Guideline for the Format and 
Content of the Clinical and Statistical Sections of New Drug 
Applications,'' U.S. Department of Health and Human Services, July, 
1988 (Ref. 72).
    FDA has tentatively concluded that it is necessary to enroll 
infants into a clinical study shortly after birth, and that the studies 
be at least 4 months in duration (see proposed 
Sec. 106.97(a)(1)(i)(A)), to ensure that the study focuses on the 
period during which infant formula generally serves as the sole source 
of nutrition, and, thus, the infant is most vulnerable to a problem 
with a formula since the infant is not consuming other foods that could 
mask or compensate for a deficiency in the formula. Also, the 
sensitivity of growth studies for identifying nutritional problems with 
an infant formula is highest during early infancy. Young infants, those 
less than 4 to 5 months, allocate a substantially higher percentage of 
the intakes of energy, protein, and other nutrients for growth than do 
older infants. After this early period of rapid growth, the rate of 
physical growth slows, and the allocation of nutrient intakes for 
growth is lower. Thus, early infancy is the period of greatest 
nutritional risk and is the age associated with the most sensitive 
growth phase.
    Because of the rapid rate of growth in infants less than 4 months 
of age, adverse nutritional impacts that affect growth rate can be 
detected within a few months (Ref. 4). Growth studies in older infants, 
where growth rates are of smaller magnitude and where solid foods are 
also consumed, are not sensitive enough to provide a meaningful 
evaluation of the ability of the formula to support healthy growth.
    The CON/AAP Task Force (Ref. 6) also recommended that clinical 
studies be conducted for a period of 3 to 4 months, and that growth be 
examined at least during the first 8 weeks of life, because nutrient 
requirements per kg body weight are greatest during this period. It 
also pointed out that such a study will cover a period when the infant 
is not consuming solid foods, and the infant formula is fed as a sole 
source of nutrition.
    Therefore, FDA has tentatively concluded that a clinical trial that 
lasts at least 4 months will be long enough to detect adverse effects 
of nutritional inadequacies on growth rate. FDA also has tentatively 
concluded that a clinical trial must be conducted with infants less 
than 1 month of age at the time of their entry into the study (see 
proposed Sec. 106.97(a)(1)(i)(A)) to ensure that the formula is tested 
during the period of time when growth rate and nutrient requirements 
are proportionately greatest, and when the infant formula serves as the 
sole source of nutrition. These requirements are intended to ensure 
that the study assesses the nutritional adequacy of the formula for 
supporting normal physical growth in the young infant.
    Under proposed Sec. 106.97(a)(1)(i)(B), the manufacturer will be 
required to collect and maintain individual and group summary data on 
anthropometric measures of physical growth and plot the data on 
National Center for Health Statistics (NCHS) reference percentile body 
weight and body length curves, which are standard measurements of 
infant physical growth (Refs. 73, 74, and 75) and provide the most 
widely accepted assessment of infant growth (Ref. 6).
    Plotting each infant's anthropometric data on NCHS reference 
percentile body weight and body length curves, and providing individual 
data on increments of weight gain, provide a means to make a 
quantitative assessment of the growth pattern over the 4 months 
duration of the study for individual infants. There is normally wide 
variation in body weights and lengths among healthy infants, with some 
being smaller than average and others average or above average. Single 
point measures of weight or length are difficult to interpret relative 
to a given infant because one does not know whether, for example, a 
smaller than average weight is attributable to inadequate nutrition or 
to a healthy and thriving infant whose body size is smaller than 
average.
    Over time, young infants tend to individualize their track within a 
given percentile on population reference growth standards. An infant at 
the 25th percentile level for weight shortly after birth tends to stay 
at or near the 25th percentile for weight throughout the first few 
months of life. When multiple longitudinal measures of weight (or 
length) of an infant are plotted on a weight-for-age reference chart, a 
reviewer can make a quick assessment as to whether an infant's pattern 
of weight or length gain is similar to that expected for healthy 
infants of the same age, taking into account the range of normal 
individual variation in body weights and lengths and that infant's 
percentile track. Similar comparisons can be made with a given infant's 
weight or length incremental gain data relative to population reference 
standards. These data allow for identification of infants with 
unusually slow or rapid growth, an observation that is masked by 
grouped data.
    Thus, plots of changes in individual infant's weight and length in 
conjunction with comparisons of increments per unit time of weight or 
length gains against population

[[Page 36184]]

reference standards allow researchers and reviewers to identify those 
infants whose growth is not following expected longitudinal patterns 
and, therefore, for whom a more thorough review of their medical and 
dietary histories is necessary to assess the possibility that the 
infant formula is responsible for reduced growth rates in a subgroup of 
infants. This careful review of individual infant growth patterns in 
addition to group summary data is particularly important because 
studies, while adequate to evaluate differences in group means between 
test and control formulas, often lack the statistical power to detect 
subgroups of infants whose growth patterns deviate from normal. These 
data will also provide useful information on possible trends towards 
failure to thrive or obesity, or on catchup growth in infants who 
experienced transient adverse effects relative to expected growth 
rates.
    FDA has tentatively concluded that a comparison of a manufacturer's 
data to well-established population reference standards can provide the 
basis for an evaluation of the growth patterns of individual infants to 
identify, and to provide the basis for an investigation of, possible 
causes of unusually slow or fast rates of gain. Thus, the agency is 
proposing that the NCHS growth charts for individuals and for grouped 
data be incorporated by reference into the regulation (proposed 
Sec. 106.97(a)(1)(i)(B)).
    Proposed Sec. 106.97(a)(1)(i)(C) requires that the manufacturer 
collect the anthropometric measurements at the beginning of the 
clinical study, at 2 weeks and at 4 weeks of the study, at least 
monthly thereafter, and at the conclusion of the study. These 
measurements will permit the calculation of incremental gains in the 
different measurements. Incremental gains, such as weight gain per unit 
of time, are generally considered the most sensitive indicator of the 
ability of a formula to support the physical growth of individual 
infants over time (Ref. 4). Also, because growth rate and nutritional 
requirements are curvilinear rather than linear during early infancy, 
multiple measurements help in assessing whether the formula meets the 
nutritional needs throughout the period of the clinical study and aids 
in more accurately placing infants in their ``correct'' reference 
percentile tract, particularly since age of enrollment varies somewhat 
among infants (although, if adopted, this regulation should serve to 
minimize that variation). Additionally, measures of an infant's body 
weight, the most critical anthropometric measure, are subject to a 
number of measurement errors unrelated to the nutritional value of the 
formula (e.g., timing of weighing of infant relative to feeding or 
defecation or urination).
    For these reasons, multiple measures over a relatively long period 
(e.g., 4 months) provide a more accurate picture of the pattern of 
growth of infants than do one or two point measures. The agency has 
tentatively concluded that the requirement of four measurements taken 1 
month apart will provide a sufficient number of measurements to permit 
evaluation of whether the formula meets the nutritional needs for 
physical growth of the infant throughout the study period. However, the 
agency requests comment, supported by data, on which measurements are 
needed to provide evidence that the formula meets the nutritional needs 
for physical growth of infants.
    FDA has tentatively concluded that more frequent measurements are 
needed during the early stages of the study because variations in 
measured body weight that are a result of factors unrelated to the 
nutritional quality of the formula can be particularly serious in early 
infancy. For example, during the first week of life, there is a normal 
loss of body weight by the infant because of fluid loss that may reach 
6 to 10 percent of body weight (Ref. 76). This weight loss will reduce 
the apparent growth of the infant as measured by body weight. This 
reduction may affect the ability to evaluate and interpret the weight 
gain data collected early in the study. FDA has tentatively concluded 
that requiring more frequent anthropometric measurements, especially 
for weight, early in the study, increases the ability to accurately 
place individual infants in the correct percentile track for monitoring 
their growth patterns in relation to the population reference curves 
and for monitoring physical growth during the most sensitive part of 
their growth phase.
    To minimize the burdens of this regulation, FDA has not proposed to 
require that blood samples obtained from infants during the time period 
of their enrollment in the clinical study, or at completion of the 
study, be analyzed for biochemical and clinical indicators of 
nutritional and growth status. However, the CON/AAP Task Force (Ref. 6) 
recommended that some blood tests be conducted at the conclusion of 
required clinical studies to provide a more comprehensive evaluation of 
the nutritional adequacy of a formula. Thus, the agency requests 
comments on whether it would be useful for the manufacturer to collect 
and maintain data on standard laboratory measures, including complete 
blood count (white blood cell count and red blood cell count), 
hemoglobin concentration or hematocrit percentage, and serum or plasma 
concentrations of albumin, urea nitrogen, electrolytes (sodium, 
potassium, and chloride), alkaline phosphatase, and creatinine. These 
measurements are standard practice when infants are seen clinically and 
can be made with very small quantities of blood. The maintenance of 
these indicators within normal limits at the end of the study provides 
additional assurance over and above measures of physical growth that 
the infant's general state of well-being is healthy and ``normal,'' 
particularly because changes in biochemical measures may occur before 
detectable differences in physical growth are identified or may not be 
detected by measures of physical growth. General anthropometric 
measurements of physical growth provide indirect, although very 
important, evidence that the formula is able to help the infant 
maintain overall good health, but they are not as specific, and may not 
be as sensitive, as are biochemical indicators of health.
    FDA also requests comment on whether requiring some, or all, of the 
biochemical and clinical tests described above would provide useful and 
necessary information for determining whether a formula causes adverse 
consequences that may not be reflected in the quality factor 
requirements for measurements of physical growth in proposed 
Sec. 106.97(a)(1)(i).
    The identification of deviations from expected values for these 
biochemical and clinical measurements, throughout the duration of the 
clinical study, could serve as an early warning of unexpected risk to 
infants enrolled in the study and, therefore, result in early actions 
to prevent undue risk to infant enrollees in the study. Conversely, 
collection of blood samples throughout the study could discourage 
parents from continuing their infants in the study, thus causing a high 
attrition rate and producing final study results that are difficult to 
interpret.
    Proposed Sec. 106.97(a)(1)(ii) sets forth guidelines for the design 
of clinical study protocols. A comprehensive clinical study protocol 
will ensure that individual investigators understand and follow 
generally accepted scientific principles for the design and conduct of 
clinical trials, thus enhancing the likelihood of interpretable results 
while maintaining minimal or no risk to infants enrolled in the 
studies. In the conduct of all studies, manufacturers should use the 
general principles,

[[Page 36185]]

described in Sec. 314.126 (21 CFR 314.126) for adequate and well-
controlled clinical studies to ensure that the design and conduct of 
the study are adequate to permit scientific review and interpretation 
of the study's results. Studies that cannot produce meaningful results 
because of poor or inadequate study design and conduct mean that 
infants will be subjected to unnecessary testing. Such a situation 
places infant enrollees at undue risk and is clearly unethical.
    In this section, FDA is not establishing mandatory elements for 
inclusion in a protocol, nor requiring that manufacturers provide the 
agency with the protocol used for a study intended to provide data to 
show that an infant formula meets the quality factor requirements. 
However, as discussed above, a protocol is an essential part of the 
design and execution of a well- controlled scientific study. 
Furthermore, a protocol often provides invaluable information that 
assists in the analysis and interpretation of the study data. 
Consequently, the agency strongly encourages manufacturers to develop 
and use protocols that incorporate the specific elements in proposed 
Sec. 106.97(a)(1)(ii) in all research studies using infants because 
these elements will ensure that the study is designed and conducted in 
a manner that will produce results that will permit meaningful 
evaluation of the usefulness of the infant formula.
    The steps outlined in proposed Sec. 106.97(a)(1)(ii)(A) represent 
standard practice in the design and conduct of clinical studies (Ref. 
72). Proposed Sec. 106.97(a)(1)(ii)(B) states that the clinical study 
protocol should describe the necessary qualifications and experience of 
the investigators. It is essential that clinical studies be conducted 
by personnel with sufficient experience and training to ensure that 
their work will yield interpretable and meaningful results. If a study 
is conducted by an investigator who is not qualified, it increases the 
likelihood that the study will have to be redone, and that more infants 
will be exposed to risk. Therefore, it is important that in the 
protocol, the manufacturer define the requisite qualifications to 
conduct the study it is designing.
    Proposed Sec. 106.97(a)(1)(ii)(C) states that the protocol should 
be reviewed and approved by an Institutional Review Board (IRB) in 
accordance with part 56 (21 CFR part 56), and that the manufacturer 
should establish procedures to obtain written informed consent from the 
parents or legal representatives of the infants enrolled in the study 
in accordance with part 50 (21 CFR part 50). These steps are necessary 
to protect the rights and safety of subjects involved in the studies.
    Proposed Sec. 106.97(a)(1)(ii)(D) states that the clinical study 
protocol should explain how the study population represents the 
population for which the new infant formula is intended. FDA has 
tentatively concluded that such an explanation is necessary so that if 
questions about the relevance of the study population arise, the answer 
is readily available and free of any taint that it is a post hoc 
rationalization. For example, FDA has recently had questions about a 
study that involved hospitalized infants that were offered to support 
use of the product on post-discharge infants. If there had been the 
type of explanation available that FDA is proposing in this guideline, 
it would have greatly minimized the questions about this product.
    Proposed Sec. 106.97(a)(1)(ii)(D) also states that the clinical 
study protocol should explain how the study addresses the intended 
conditions of use of the formula. FDA has tentatively concluded that, 
by having manufacturers consider this question before the study is 
conducted, this guideline will prevent clinical studies that are 
conducted under conditions of use that do not accurately reflect the 
proposed conditions of use. For example, a clinical study protocol for 
testing a formula designed to be used by premature infants throughout 
infancy should explain how the study design will provide information to 
support the claim that the formula supports healthy growth under these 
conditions.
    Proposed Sec. 106.97(a)(1)(ii)(E) states that the clinical study 
protocol should describe the sample size calculations and the power 
calculations and the basis for selecting the sample size and study 
design. This information is necessary to establish the likelihood that 
the study will not fail to detect a real difference, should there be a 
difference for the measurements of interest, between the infant formula 
being tested and the control. For example, a study might not find a 
difference in incremental rate of weight gain between infants consuming 
two formulas because too few infants were enrolled in the study to 
provide sufficient statistical power to detect this difference. 
Inadequate statistical power could mask the nutritional inferiority of 
a product and could result in the marketing of a formula that does not 
meet the quality factor requirements and, therefore, is not safe for 
its intended use. Therefore, FDA has tentatively concluded that this 
guideline is needed to ensure that manufacturers design their growth 
studies to be capable of detecting biologically meaningful differences 
for the endpoints of interest between the two formulas. Identification 
of differences would raise safety concerns or serious questions of 
nutritional quality of the test formula product.
    Proposed Sec. 106.97(a)(1)(ii)(F) states that the clinical study 
protocol should include a plan to identify and evaluate any adverse 
events. This proposed guideline is necessary to document that 
appropriate attention is given to the systematic evaluation and 
recording of any adverse events that may occur during the course of the 
study. Inadequate planning for and conduct of the monitoring of adverse 
events may result in an erroneous conclusion that the formula is safe 
and suitable, when in fact the formula is not safe and suitable for 
infants under intended conditions of use.
    Proposed Sec. 106.97(a)(1)(ii)(G) states that the clinical study 
protocol should describe the quality control procedures that the 
investigator will use to ensure the validity and reliability of the 
measurements collected. This proposed guideline represents standard 
practice in the design and conduct of clinical studies and is necessary 
to allow a meaningful interpretation of study results. Data obtained 
with unreliable measures, or with indicators that do not accurately or 
meaningfully measure identified endpoints, may produce misleading study 
results that are uninterpretable and that suggest that a formula is 
safe and suitable, when more valid or reliable measures would not have 
supported this conclusion. The institution of adequate quality control 
procedures before beginning a study provides a mechanism for 
manufacturers to ensure that the data collected are reliable, and that 
the study provides interpretable results.
    Proposed Sec. 106.97(a)(1)(ii)(H) states that the clinical study 
protocol should describe and compare the composition of the control and 
test formulas. These descriptions of the control and the test formulas 
are necessary to establish that the formula used as the control 
provides an adequate comparison for evaluating the quality factors of 
the test formula. If the control formula is not comparable to (i.e., 
bioequivalent to) formulas in current use, differences between the test 
and control formulas have no meaning. They cannot be generalized to 
projected conditions of use. For example, comparable or enhanced 
physical growth in infants consuming a test formula as compared to 
infants consuming a control formula when the control formula does not 
meet

[[Page 36186]]

requirements in Sec. 107.100 for nutrients, or is not bioequivalent 
relative to quality factors to currently marketed formulas in the 
United States, cannot be interpreted as supporting healthy growth 
because it is not possible to determine whether the apparent ``equal'' 
or ``enhanced'' physical growth is attributable to the fact that the 
formula is nutritionally adequate, or whether the formula looks 
adequate because it is being compared to a nutritionally inadequate 
formula. The nature of the differences between control and test 
formulas will also affect sample size and measurement (endpoint) 
considerations.
    FDA's experience in reviewing clinical data submitted with 90-day 
notifications has been that the absence of information on control 
formulas is not uncommon. Thus, FDA has tentatively concluded that a 
guideline on the information that needs to be considered in selecting a 
control formula is necessary to ensure that study results are 
meaningful and interpretable.
    If the test formula used in a study is not identical to the formula 
that is intended to be marketed in the United States, proposed 
Sec. 106.97(a)(1)(ii)(I) states that the clinical protocol should 
describe the basis upon which the manufacturer has decided that the 
test formula is appropriate for use in the study. This proposed 
guideline is necessary to ensure that the manufacturer considers such 
factors as the bioequivalence of the studied (test) formula relative to 
the formula that is to be marketed in this country and can document why 
its choice of test formulas is appropriate. Without this documentation, 
it would not be possible to determine whether the marketed formula 
meets the quality factor requirement in proposed Sec. 106.96(b).
    FDA has had experience under the 1986 amendments in which 
manufacturers have submitted data on test formulas that were 
significantly different (e.g., in calorie levels) from the formula that 
they intended to market as evidence of the safety and suitability of 
the latter formula. In these instances, the agency has had considerable 
difficulty in interpreting study results. Therefore, if the guidance in 
proposed Sec. 106.97(a)(1)(ii)(I) is followed, this significant study 
design issue will be critically reviewed by manufacturers before they 
initiate their studies, and, as a result, they will be more likely to 
design and conduct a study that will produce data that can be 
meaningfully interpreted as evidence that an infant formula is safe, 
and that it supports healthy growth.
    As provided in proposed Sec. 106.97(a)(2), however, FDA recognizes, 
that while changes in ingredients or in the processes used in the 
manufacture of infant formulas can have a significant adverse impact on 
the levels or availability of nutrients that affect healthy growth of 
infants, other changes may not be likely to do so. In the latter 
circumstances, it may be possible to demonstrate that the quality 
factor requirements are met by means of measures or data that do not 
involve the use of clinical trials. If such assurances can be provided 
without clinical trials, then infants will not be subjected to 
unnecessary testing. Therefore, FDA sets out in proposed 
Sec. 106.97(a)(2), the circumstances in which a manufacturer can 
request an exemption from the clinical study requirement.
    Proposed Sec. 106.97(a)(2)(i) provides for an exemption if the 
manufacturer can cite experience that shows that the ingredient, 
ingredient mixture, or processing method has been used to make an 
infant formula that meets the quality factor requirements in proposed 
Sec. 106.96(a). For example, if the manufacturer has previously 
submitted information to FDA in response to the quality factor 
requirements of the act that showed that an infant formula that 
contains the ingredient or ingredient mixture, or that was produced by 
the processing method, in question supported adequate physical growth, 
this information could form the basis on which the new infant formula 
could qualify for an exemption from this quality factor requirement. 
Under this provision, FDA will evaluate the experience cited in support 
of an exemption on a case-by-case basis. FDA requests comment on this 
proposed provision.
    Proposed Sec. 106.97(a)(2)(ii) provides for an exemption if a 
manufacturer that markets a formulation in more than one form (such as 
liquid and powdered forms) can demonstrate that the quality factor 
requirements are met by the form of the formula that is processed using 
the method that has the greater potential for adversely affecting the 
formula's nutrient content and bioavailability. For example, the 
temperatures used to retort liquid formulas during processing can cause 
a loss of protein quality compared to powdered forms processed at lower 
temperatures (Refs. 77 and 78). Thus, if the liquid formula is tested 
and shown to meet the quality factors requirements, it will provide 
reasonable assurance that the powdered form of the formula, that is, 
the less processed form is of appropriate nutritional quality. Thus, 
FDA tentatively concludes that it would be unnecessary to test the less 
processed form.
    Proposed Sec. 106.97(a)(2)(iii) provides for an exemption if the 
manufacturer can demonstrate that the requirements of proposed 
Sec. 106.97(a)(1) are not appropriate for the formula, and an 
alternative method or study design for showing that the formula 
supports healthy growth in infants fed the formula as a sole source of 
nutrition is available. As stated above, double- blind, well-
controlled, clinical studies are generally the most powerful and 
sensitive method for demonstrating that an infant formula will support 
physical growth. Nonetheless, the agency anticipates that there will be 
circumstances in which a clinical study of a new infant formula would 
not be appropriate. For example, double-blind clinical studies would 
not be appropriate in situations such as those involving some exempt 
infant formulas in which they would cause withholding of conventional 
treatment and, therefore, would be unethical. Other situations that may 
not be amenable to double-blind clinical trials are those in which it 
would be difficult to enroll an adequate number of infants (e.g., for 
exempt infant formulas where the formula is intended for a rare 
disease). Alternative study designs may also be appropriate in 
situations in which a manufacturer has access to extensive reference 
data, such as a database on many similarly conducted clinical studies 
using infants from the same potential study population, provided that 
the manufacturer can demonstrate that the reference data apply to the 
new infant formula, its intended use, and its study population. FDA has 
tentatively concluded that such an exemption will permit flexibility in 
the design of suitable experimental protocols but still provide 
reasonable and documentable assurance that the study design can 
demonstrate the safety and suitability of the infant formula.
    b. Specific quality factors. Proposed Sec. 106.97(b) establishes 
requirements for demonstrating that a formula meets the protein quality 
factor requirement in proposed Sec. 106.96(c) and requires that the 
manufacturer collect and maintain data that establish that the 
biological quality of protein in an infant formula is sufficient to 
meet the protein requirements of infants by demonstrating that the 
protein source supports adequate growth using the PER rat bioassay, 
which the agency proposes to incorporate by reference. The PER provides 
an estimate of the bioavailability and relative proportion of the 
essential amino acids in the protein-containing ingredient.

[[Page 36187]]

    A chemical analysis of the protein can identify the amino acids 
contained in a protein source but does not measure their 
bioavailability. A protein source may contain the necessary amino 
acids, but they may be in a form that the infant cannot digest and 
absorb. Furthermore, processing methods may alter the chemical nature 
of the protein source, possibly making the protein more resistant to 
digestion by the infant. FDA has tentatively concluded that the rat 
bioassay is necessary to establish that the amino acids in a protein 
source are present, and that adequate amounts and proportions of all 
essential amino acids are capable of being digested by an infant. Such 
a showing is particularly important when a manufacturer is using a 
novel protein source (e.g., a hydrolyzed protein), a new protein 
mixture, a new processing method that could affect the chemical form or 
bonding of amino acids, or a formulation that provides an amount of 
protein near the minimum required level (<2.0 g/100 kilocalorie (kcal)) 
specified in Sec. 107.100.
    Proposed Sec. 106.97(b)(1) also provides that if the manufacturer 
is unable to conduct a PER rat bioassay, it must demonstrate that the 
amino acid composition of the protein meets the known amino acid 
requirements of infants for whom the formula is intended. For example, 
FDA is aware that a PER would not provide useful data for an exempt 
infant formula intended for use in infants that cannot metabolize a 
specific amino acid and from which that amino acid has been 
purposefully omitted or is limited to a level inadequate to support 
healthy growth. The lack of that amino acid is necessary for the 
dietary management of the intended infant population but would result 
in an incomplete protein and would reduce the growth rate of the rat, 
invalidating the conditions upon which the PER rat bioassay is based. 
FDA is not aware of alternative methods for ensuring bioavailability of 
such a protein source. In these circumstances, proposed 
Sec. 106.97(b)(1) will provide an alternate means of evaluating whether 
the protein at least contains adequate amounts of essential amino acids 
to meet the known amino acid requirements of the infant, even though 
the bioavailability of these amino acids cannot be assured using 
available methods.
    Proposed Sec. 106.97(b)(2) establishes the circumstances in which a 
manufacturer may request an exemption from the requirements of proposed 
Sec. 106.97(b)(1). Proposed Sec. 106.97(b)(2)(i) provides that if the 
protein source (including the processing method used to produce it) is 
already used in another of the infant formulas marketed by its 
manufacturer in the United States, the manufacturer may request an 
exemption if it can demonstrate that such other infant formula meets 
the quality factor requirements prescribed in Sec. 106.96(b)(1). The 
purpose of the PER or amino acid analyses is to estimate the quality of 
the protein in the proposed formula. Once a manufacturer has 
established standard sources and processing of protein in a formula, 
and has demonstrated that the technology is effective, in its hands, in 
producing a formula that meets the quality factor requirement for 
protein, other formulation changes would not be expected to markedly 
affect protein quality. Thus, the quality of the processed protein 
would be retained in other formulas. However, under proposed 
106.97(b)(2)(i), it will be incumbent on the manufacturer to 
demonstrate that the quality of the protein is not affected.
    Proposed Sec. 106.97(b)(2)(ii) provides for an exemption if the 
protein source, or the processing method used to produce the protein 
source, in the infant formula does not constitute a major change from 
the infant formula that it replaces, and the manufacturer can 
demonstrate that the infant formula that it replaces meets the quality 
factor requirements prescribed in Sec. 106.96(b). FDA is proposing to 
allow this exemption because it is unlikely that the methods for 
assessing protein quality prescribed are sensitive enough to measure 
any change in protein quality that is not a major change.
    Because FDA has, as a matter of policy, been requesting that infant 
formula manufacturers submit data from a PER or amino acid analysis as 
part of their submission 90 days prior to marketing infant formula, 
many infant formulas that are on the market have been shown to meet the 
proposed quality factor requirement for protein. Therefore, if the 
proposed exemption criteria in Sec. 106.97(b)(2) are adopted, those 
formulas that contain protein sources, or proteins which were produced 
using processing methods, that were the subject of a submission to FDA 
in response to the quality factor requirements of the act may qualify 
for an exemption.
6. Request for Comment on Establishing Assurances for Other Quality 
Factors
    As discussed above, FDA has solicited comment on whether to 
establish quality factor requirements for fat, iron, and calcium and 
phosphorus. If such quality factors are adopted, appropriate methods 
will be needed to provide assurance that an infant formula meets these 
nutrient-specific quality factors. Therefore, FDA discusses below 
measurements of fat balance and of calcium and phosphorus balance, as 
well as measurements that reflect iron bioavailability. The agency 
requests comments and information on these or other methods for these 
three quality factors:
    a. Apparent fat absorption. Apparent digestibility and apparent 
absorption measure the amount of fat that was able to be digested and 
absorbed by the infant. Apparent digestibility is expressed as a 
percentage of intake, while apparent absorption is expressed in units 
of fat (e.g., g) absorbed per day. If a quality factor for fat were 
established, manufacturers would be required to collect and maintain 
data establishing that the apparent digestibility or apparent 
absorption by the infant of the fat in an infant formula is adequate to 
meet the infant's energy requirements. These data would be necessary 
because fat represents the major dietary source of energy for the 
infant and must be readily digested and absorbed if the formula is to 
support healthy growth.
    The CON/AAP Task Force (Ref. 6) recommended that studies that are 
conducted to determine whether a formula meets the quality factor for 
fat should use a cross-over experimental design. This type of study 
requires that the manufacturer compare apparent fat absorption of 
infants fed the test formula at one time and a currently marketed 
formula at another time. An experiment using this design would enable a 
manufacturer to make measurements of apparent fat absorption using a 
small number of infants, since the variance in fat excretion of infants 
fed most fat sources currently available is less than 5 percent. 
Furthermore, the method is noninvasive, is easily implemented, and does 
not require costly or sophisticated equipment to conduct. Other 
experimental designs could be used but would require larger numbers of 
infants and would be more expensive. Thus, FDA asks for comments on 
whether there should be a specific requirement that manufacturers 
measure apparent fat absorption using cross-over studies.
    The CON/AAP Task Force (Ref. 6) recommended that studies that are 
conducted to determine the apparent absorption of fat be conducted such 
that measurements are made using infants fed each formula for at least 
72 hours. The Task Force report suggested that measurements of apparent 
fat absorption for this length of time would accurately reflect the 
apparent

[[Page 36188]]

absorption of the fat in the formula being tested. FDA is considering 
requiring that a study of at least 72 hours for each formula tested be 
conducted and requests comment on what duration would be appropriate. 
FDA also is considering whether to require that the manufacturer 
document the method that it used to analyze for fat and explain the 
reason for choosing that method. The agency believes that this 
information is important because the method used to analyze the 
excreted fat must be appropriate for the specific type of fat in the 
formula.
    FDA also is considering whether circumstances exist that would 
justify establishing an exemption from the requirements to measure fat 
balance. FDA has tentatively concluded that the reasons and 
justification for such an exemption are essentially those set forth 
above in the discussion of proposed Sec. 106.97(b)(2). FDA requests 
comment on whether, if the agency adopts a quality factor for fat, it 
should provide for exemptions from testing, to show that the formula 
meets that quality factor, such as those set forth in proposed 
Sec. 106.97(b)(2), and to allow manufacturers to assure the agency that 
their products meet that quality factor requirement without subjecting 
infants to unnecessary testing.
    b. Calcium and phosphorus balance. If FDA were to establish a 
quality factor for calcium and phosphorus, manufacturers would be 
required to collect and maintain data from clinical studies conducted 
in infants to show that the calcium and phosphorus contained in the 
infant formula are sufficient to meet the infant's requirements. There 
are currently no satisfactory clinical laboratory measurements that are 
practical for directly assessing calcium and phosphorus nutritional 
status in infants (Ref. 79). Furthermore, there are no accurate 
indirect measurements that could be made on the infant formula itself 
that would be useful in predicting how effective the amount and the 
sources of calcium and phosphorus in the formula would be in meeting 
the needs of infants consuming that formula. Therefore, FDA is 
considering requiring that manufacturers implement the recommendations 
of the CON/AAP Task Force and make a measurement that provides a 
reasonable estimate of the amount of calcium and phosphorus that is 
capable of being absorbed and retained for use by infants (i.e., 
calcium and phosphorus balance) from the formula.
    FDA asks for comment concerning the appropriateness and usefulness 
of a measurement of calcium and phosphorus balance as one that reflects 
both the bioavailability of the calcium and phosphorus in the formula 
and how well the diet meets the metabolic requirements for these two 
minerals. As discussed above with regard to the conduct of trials to 
measure apparent fat absorption, FDA requests comment on whether it is 
necessary to require that a cross-over study design be used for 
clinical studies to measure calcium and phosphorus balance.
    FDA also requests comment on what would be an appropriate duration 
for studies to measure calcium and phosphorus balance. The CON/AAP task 
force suggested that calcium and phosphorus balance studies be 
conducted for a 72-hour balance period after an 11-day adaptation 
period. FDA requests comment on whether these time periods are 
appropriate, both to minimize the effects of previous dietary intake on 
the availability of calcium from the formula being tested (Ref. 6) and 
to ensure that the results of the balance study are reliable and 
interpretable, and on whether they provide a meaningful basis on which 
to determine that a formula meets the quality factor requirement for 
calcium and phosphorus.
    FDA is considering requiring that the formula used as the control 
in any clinical studies to measure calcium and phosphorus balance 
contain approximately the same calcium and phosphorus levels as the 
test formula because the absolute amounts of these nutrients absorbed 
and retained by infants may be different between formulas with 
different calcium and phosphorus levels. FDA is asking for comment on 
requirements for appropriate control formulas for calcium and 
phosphorus balance studies.
    Amounts of calcium and phosphorus in urine and feces, along with 
calculated amounts absorbed and retained expressed in milligrams per 
kilogram and as percentages of intake, provide evidence of the rates of 
absorption and retention of these nutrients but do not specifically 
measure the ability of the formula to provide adequate calcium and 
phosphorus for proper bone mineralization, the most important need for 
these minerals in the infant. FDA is considering requiring that serum 
alkaline phosphatase be measured in situations in which calcium and 
phosphorus balance studies are required in order to assess the adequacy 
of formula minerals to support normal bone mineralization. Alkaline 
phosphatase is an enzyme involved in bone remodeling and in maintaining 
serum calcium concentration (Ref. 64). Increased serum alkaline 
phosphatase activity may be a marker of reduced bone mineralization 
(Ref. 80) and therefore may be useful in determining whether a formula 
meets a quality factor requirement for calcium and phosphorus.
    Because of the limits of metabolic balance studies, including short 
duration, dependence on previous diet, and expense, the agency is 
considering the appropriateness of alternative methods for the 
assessment of bone mineral accretion. The agency is aware that 
sophisticated instruments, such as single-photon absorptiometry and 
dual-energy x-ray absorptiometry, have been tested for measuring bone 
mineral content in infants (Refs. 81 through 84), and that some 
authorities recommend them for determining bone mineralization in 
infants (Ref. 85). These types of measurements have the potential to 
provide an accurate measure of bone mineral accretion over the duration 
of use of the formula, while at the same time reducing many sources of 
variation inherent in balance studies. The agency is concerned, 
however, that these methods have not been adequately validated in 
infants, and that reference standards for mineralization in infants 
have not been established to support a requirement for manufacturers to 
measure bone mineralization in order to provide assurance that a 
formula satisfies a quality factor requirement for calcium and 
phosphorus. The agency asks for comment on the usefulness of these 
methods of analysis of bone mineral accretion in infants, and on 
whether they should be used in lieu of calcium and phosphorus balance 
studies as measurements of whether an infant formula meets the quality 
factor requirements for calcium and phosphorus assuming that the agency 
adopts such a quality factor. The agency also asks for comment on the 
criteria that it should use, on a case-by-case basis, in deciding 
whether to require these types of measures when there is particular 
reason to be concerned that calcium and phosphorus bioavailability may 
be problematic.
    FDA also is considering whether circumstances exist that would 
justify establishing an exemption from a requirement to measure calcium 
and phosphorus balance. FDA has tentatively concluded that the reasons 
and justification for such an exemption are essentially those set forth 
above in the discussion of proposed Sec. 106.97(b)(2), and requests 
comment on whether, if it adopts a quality factor for calcium and 
phosphorus, it should provide for exemptions from testing to show that 
the formula meets the quality

[[Page 36189]]

factor similar to those in proposed Sec. 106.97(b)(2) and allow 
manufacturers to assure the agency that their products meet that 
requirement without requiring redundant testing.
    c. Iron status. If FDA were to adopt a quality factor for iron, 
manufacturers would be required to collect and maintain data that 
establish that the iron in an infant formula is bioavailable and 
maintains the iron status of infants that consume the formula. These 
data would be necessary to demonstrate that an infant formula provides 
enough iron to prevent iron deficiency and anemia.
    Alterations in a number of biochemical measurements are useful 
signs associated with inadequate iron intake or the development of iron 
deficiency. Early signs of inadequate iron intake, which reflect the 
depletion of iron storage sites, are reductions in serum ferritin 
concentration and transferrin saturation (Ref. 86). If the dietary 
intake of iron remains inadequate, impaired erythropoiesis (i.e., the 
process whereby the body produces new red blood cells) may be reflected 
in alterations in erythrocyte maturation and increases in erythrocyte 
size, erythrocyte protoporphyrin concentration, or serum transferrin 
receptor levels. If the period of inadequate iron intake continues, 
erythropoiesis is further impaired, and hemoglobin concentration, 
hematocrit, and mean corpuscular volume decrease.
    Iron deficiency without anemia should be considered to be a risk 
factor for iron-deficiency anemia, which may be associated with long-
lasting, adverse effects in infants (Ref. 86). Therefore, FDA is 
considering requiring one measurement of iron status that is sensitive 
to each of the three stages of inadequate iron intake (stage 1, 
decreased stores, normal erythropoiesis; stage 2, decreased stores and 
early stage impaired erythropoiesis; and stage 3, decreased stores and 
late stage impaired erythropoiesis). For example, FDA is considering 
requiring that manufacturers measure: (1) Serum ferritin concentration, 
because such a measurement is sensitive to decreased iron stores and 
normal erythropoiesis; (2) transferrin saturation or erythrocyte 
protoporphyrin concentration, because such measures are sensitive to 
decreased iron stores and early stage impaired erythropoiesis; and (3) 
hematocrit percentage, hemoglobin concentration, or mean corpuscular 
volume, because such measurements are sensitive to decreased iron 
stores and late stage impaired erythropoiesis. This approach would be 
consistent with the recommendations of the CON/AAP Task Force (Ref. 6). 
It would also provide reasonable assurance that low iron availability 
in an infant formula would be detected, and that an infant formula that 
does not provide sufficient iron to meet the infant's requirement, and 
thereby does not meet the quality factor requirement for iron, will not 
be marketed.
    FDA also is considering whether circumstances exist that would 
justify establishing an exemption from the requirements to determine 
iron status. FDA has tentatively concluded that the reasons and 
justification for such an exemption are essentially those set forth 
above in the discussion of proposed Sec. 106.97(b)(2). FDA requests 
comment on whether, if it adopts a quality factor for iron, it should 
provide for exemptions from testing similar to those set forth in 
proposed Sec. 106.97(b)(2) to show that the formula meets that factor 
and allow manufacturers to assure the agency that their products meet 
that quality factor requirement without requiring redundant testing.

F. Records and Reports

1. Introduction
    Under subpart C of part 106, FDA is proposing to revise the 
requirements on the records that must be made and retained. FDA is 
proposing requirements on batch records; records on CGMP and quality 
control procedures; maintenance of distribution records on formulas for 
export only; audits; and notifications to FDA. These proposed changes 
to current Sec. 106.100 are outlined in Table III below:

                                Table III                               
------------------------------------------------------------------------
            Current Regulation                   Proposed Regulation    
------------------------------------------------------------------------
Sec.  106.100(a)..........................  No Change.                  
Sec.  106.100(b)..........................  No Change.                  
Sec.  106.100(c)..........................  No Change.                  
Sec.  106.100(d)..........................  No Change.                  
Sec.  106.100(e), (f), and (h)............  Current Sec.  106.100(e),   
                                             (f), and (h) will be       
                                             incorporated into proposed 
                                             Sec.  106.100(e).          
                                            New Sec.  106.100(f) will   
                                             codify the records required
                                             for the CGMP regulations   
                                             found in proposed subpart  
                                             B.                         
Sec.  106.100(g)..........................  Current Sec.  106.100(g)    
                                             with modification.         
Sec.  106.100(h)..........................  Current Sec.  106.100(h) is 
                                             incorporated into Sec.     
                                             106.100(e). Sec.           
                                             106.100(h) Reserved.       
Sec.  106.100(i)..........................  No Change.                  
Sec.  106.100(j)..........................  Current Sec.  106.100(j)    
                                             with modification.         
Sec.  106.100(k)..........................  Current Sec.  106.100(k)    
                                             with modification.         
Sec.  106.100(l)..........................  No Change.                  
Sec.  106.100(m)..........................  No Change.                  
Sec.  106.100(n)..........................  No Change.                  
Sec.  106.100(o)..........................  No Change.                  
------------------------------------------------------------------------

2. Batch Production and Control Records
    Proposed Sec. 106.100(e) requires that manufacturers make and 
retain records (hereafter referred to as ``batch records'') that 
include complete information relating to the production and control of 
each batch of infant formula. Section 412(b)(4)(A)(i) of the act 
requires the establishment, by regulation, of requirements for the 
retention of all records, including records containing the results of 
all testing required under section 412(b)(2)(B) of the act, necessary 
to demonstrate compliance with the CGMP requirements and quality 
control procedures prescribed under section 412(b)(2). In proposed 
Sec. 106.100(e) FDA is proposing to require that manufacturers prepare 
and maintain records that include complete information relating to the 
production and control of the batch to ensure that the complete history 
of each batch of infant formula is available for review in the event 
that a problem arises with a particular batch.
    Proposed Sec. 106.100(e)(1) requires that the batch records include 
the appropriate master manufacturing order. As discussed above, 
proposed Sec. 106.50(a) requires that manufacturers produce each infant 
formula in accordance with a master manufacturing order that has been 
approved by a responsible official of the company. The master 
manufacturing order thus provides fundamental information about the 
batch. Having all the information concerning the production of a batch 
of infant formula, including the master manufacturing order, in one 
place as a part of a batch record will ensure that there is a document 
available that makes readily apparent whether a batch was properly 
produced. It will also ensure that all the information needed to 
evaluate the cause of any problem that may develop with a batch of 
infant formula is readily available. Thus, FDA has tentatively 
concluded that the master manufacturing order is an essential part of 
the batch record.
    Proposed Sec. 106.100(e)(1)(i) requires that the master 
manufacturing order include the significant steps in the production of 
the batch of infant formula and the date on which each

[[Page 36190]]

significant step occurred. Thus, the master manufacturing order will 
include a list of the significant steps for the production of each 
infant formula and a space to write in the date the step was performed. 
Thus, it will provide both a check that the step was performed and a 
record of when it was performed. FDA has tentatively concluded that 
this information is necessary because all production activities for a 
specific batch of infant formula may not be accomplished in one day but 
may occur over a number of days, and people who begin work the second 
day will know what work has been completed, and what has not been. 
Moreover, each date is needed so that a batch of formula can be traced 
if, at a later date, a problem that may adversely affect an infant 
formula is identified at a specific production stage. Having the date 
available will allow the manufacturer to identify all batches that may 
have been affected by the problem.
    Proposed Sec. 106.100(e)(1)(ii) requires that, if the manufacturer 
has more than one line or set of equipment in the plant in which the 
formula is made, the master manufacturing order include the identity of 
equipment and processing lines used in producing the batch of infant 
formula. This information will allow the manufacturer to ensure that 
the equipment on which the formula was produced met the requirements of 
Sec. 106.30. This information also will facilitate the identification 
of all batches of formula that may be affected by equipment 
malfunctions or that were produced on the same equipment as a batch 
that is discovered to be microbiologically contaminated.
    Proposed Sec. 106.100(e)(1)(iii) requires that the master 
manufacturing order include the identity of each batch or lot of 
ingredients, containers, and closures used in producing the batch of 
infant formula. All materials used in infant formula will have to meet 
the specifications of proposed Sec. 106.40(d) and be identified by a 
batch or lot number as specified in proposed Sec. 106.40(c). FDA has 
tentatively concluded that it is necessary to propose that the identity 
of each batch or lot of ingredients, containers, and closures used in 
producing the batch of infant formula be recorded in the master 
manufacturing order to enable the manufacturer to ensure that all of 
those materials met the requirements of Sec. 106.40, particularly the 
standards for acceptance or rejection of the materials. Recording this 
information also will allow the manufacturer to evaluate the 
contribution of specific ingredients, containers, and closures to any 
problem with a batch of infant formula that may develop.
    FDA is not proposing to require that the batch records contain the 
results of any tests conducted on ingredients, containers, and closures 
in accordance with proposed Sec. 106.40(d) because the same lot of raw 
materials may be used in multiple batches. The identification of the 
batch or lot of all ingredients, containers, and closures in the master 
manufacturing order should be sufficient to allow the manufacturer to 
locate and review relevant test results if problems arise with a 
particular batch of infant formula.
    Proposed Sec. 106.100(e)(1)(iv) requires that the master 
manufacturing order include the amount of each ingredient to be added 
to the batch of infant formula and a check (verification) that the 
correct amount was added. As discussed above, proposed Sec. 106.50(b) 
requires that the manufacturer establish controls to ensure that raw 
and in-process ingredients required by the master manufacturing order 
are examined by one person and checked by a second person or system to 
ensure that the correct weight or measure of the ingredient is added to 
the batch. The agency has tentatively concluded that recording in the 
master manufacturing order the amount of each ingredient added to the 
batch of formula, and a check (verification) that the correct amount 
was added, are appropriate controls to ensure that the correct weight 
or measure of the ingredient is added to the batch. This proposed 
requirement is necessary to ensure that there is compliance with 
proposed Sec. 106.50(b), to provide a record that the batch of infant 
formula includes all of the ingredients in the amounts specified in the 
master manufacturing order, and to provide assurance that the product 
contains all of the required nutrients.
    Proposed Sec. 106.100(e)(1)(v) requires that the master 
manufacturing order include copies of all labeling used and the results 
of the examinations conducted during the finishing operations to ensure 
that containers and packages in the batch are correctly labeled. (The 
importance of ensuring that containers are correctly labeled was 
discussed in conjunction with proposed Sec. 106.60(b).) The inclusion 
in the batch records of copies of the labeling used on each batch of 
infant formula will provide a record of such labeling and will document 
that the finishing operation examinations, required by proposed 
Sec. 106.60(b), are conducted.
    Proposed Sec. 106.100(e)(2) requires that the batch record include 
any deviations from the master manufacturing order and any corrective 
actions taken. While the manufacturer's goal should be to produce the 
infant formula in accordance with the master manufacturing order, on 
occasion deviations may occur. On these occasions, the deviations, and 
any corrective actions taken because of the deviations, should become a 
part of the batch record. For example, if a batch of liquid infant 
formula was thermally processed at a different temperature than the 
temperature specified in the master manufacturing order, the batch 
record would state the actual processing temperature. The record would 
also state any corrective actions taken because of this processing 
temperature, such as a change in processing time. A record of 
deviations from the master manufacturing order and of the corrective 
actions taken by the manufacturer will allow the manufacturer to 
quickly determine whether all deviations have been appropriately 
addressed, and if they have not been, whether the actions needed to 
correct the deviations have been identified. It will also provide 
relevant information if a problem arises with that batch of infant 
formula.
    Proposed Sec. 106.100(e)(3) requires that the batch records include 
documentation of the monitoring at any production and in-process 
control point, step, or stage where control is deemed necessary to 
prevent adulteration. As discussed above, proposed Sec. 106.6(c)(2) 
requires this monitoring. FDA is proposing that the documentation that 
the monitoring required by proposed Sec. 106.6(c)(2) is occurring be 
included in the batch records to ensure that a measurement or 
observation made at one particular point in time can be related to a 
particular batch. The linkage of the record to the batch is especially 
important when a standard or specification is not met. It will enable 
the manufacturer to determine what batches may have been affected by a 
deviation and to take appropriate action, such as withholding a batch 
from distribution.
    Proposed Sec. 106.100(e)(3)(i) requires that the batch records 
include a list of the standards or specifications established at each 
point, step, or stage in the production process where control is deemed 
necessary to prevent adulteration, and that it include documentation of 
the scientific basis for each standard or specification. As discussed 
above, proposed Sec. 106.6(c)(1) requires the establishment of such 
standards or specifications. The agency has tentatively concluded that 
a list of these standards or specifications must be a part of the batch 
record so that the manufacturer will have them readily

[[Page 36191]]

available to compare to the actual values obtained during the 
monitoring operation of the production and in-process control system. 
Also, the documentation of the scientific basis for each standard or 
specification will verify that each was established by trained and 
experienced sources. Such documentation will summarize the work 
performed to establish the standard or specification and will establish 
the source used. If changes to the standard or specification become 
necessary, this documentation of the scientific basis for each standard 
or specification will assist the manufacturer in making such changes.
    Proposed Sec. 106.100(e)(3)(ii) requires that the batch records 
include the actual values obtained during the monitoring (such as the 
actual temperatures and actual times that the measurements were taken), 
any deviations from the established standards or specifications, and 
any corrective actions taken. For example, notations that refrigeration 
temperatures are satisfactory or unsatisfactory, without a record of 
the actual temperatures, are subject to varying interpretation and thus 
will not ensure that preventive controls are working. It is important 
that the actual values be recorded. In addition, actual values are 
necessary to discern trends or to pinpoint the onset of a problem. The 
record of any corrective actions taken will show what the manufacturer 
did when a standard or specification was violated, and how the 
manufacturer is ensuring that the infant formula is not adulterated. 
Entry of information on the records at the time of the monitoring 
ensures that the record does not rely on the memory of the observer and 
thus is as accurate and valid as possible.
    Proposed Sec. 106.100(e)(3)(iii) requires that the batch records 
identify the person monitoring each point where control is deemed 
necessary to prevent adulteration. FDA has tentatively concluded that 
it is important that the responsible individuals be identified in the 
batch record so that the manufacturer can check that a qualified person 
is actually monitoring the point, step, or stage where control is 
deemed necessary to prevent adulteration, and so that such individual 
can be contacted if a problem with a batch of infant formula is 
identified at a later date. These individuals are in the best position 
to know of any other information that may not have seemed pertinent at 
the time but, in retrospect, could be important in identifying the 
cause of the problem and initiating actions to prevent it from 
recurring.
    Proposed Sec. 106.100(e)(4) requires that the batch records include 
the conclusions and followup, along with the identity, of the qualified 
individual who investigated any deviations, or failures to meet 
specifications, that occurred during the production of the batch. Under 
these proposed regulations, individuals qualified by training or 
experience must conduct an investigation of any deviation from the 
master manufacturing order and of the corrective actions taken 
(Sec. 106.50(a)(2)); conduct an investigation of a finding that a batch 
or any of its ingredients failed to meet any manufacturer's 
specifications (Secs. 106.40(d) and 106.70(c)); and conduct an 
investigation of a failure to meet any specification or standard at any 
point where control is deemed necessary to prevent adulteration 
(Sec. 106.6(c)(4)).
    FDA has tentatively concluded that the record of the conclusions 
and followup of these investigations is necessary to enable the 
manufacturer to ensure that it has complied with proposed 
Secs. 106.6(c)(4), 106.40(d), 106.50(a)(2), and 106.70(c). Such records 
will provide information on how the production of the batch of infant 
formula deviated from established standards or specifications and on 
the cause of any problem with the formula, if infants are reported to 
have been adversely affected by the product at a later date. 
Identification of the qualified individual who conducted the 
investigations will ensure that there is responsibility and 
accountability for the investigation and will allow the responsible 
individuals to be contacted, if necessary. These individuals will be in 
the best position to provide information if additional details about 
the record are needed.
    Proposed Sec. 106.100(e)(5) requires that the batch records include 
the results of all testing performed on the batch of infant formula, 
including testing on the in-process batch, at the final-product stage, 
and on finished product throughout the shelf life of the product. 
Section 412(b)(2)(B) of the act requires that manufacturers conduct 
such testing. FDA has tentatively concluded that the assembly of such 
records in one place will enable the manufacturer to ensure that the 
batch of infant formula complies with proposed Secs. 106.55 and 106.91 
and will facilitate the review of the test results in the event that a 
problem arises with the batch.
    Proposed Sec. 106.100(e)(5)(i) states that the batch records are to 
include the results of any quality control testing conducted, in 
accordance with proposed Sec. 106.91(a) and (b), to verify that each 
nutrient required by Sec. 107.100 is present at the required level, and 
that any nutrient added by the manufacturer is present at the 
appropriate level. Including the results of this testing in the batch 
records will provide data needed to evaluate compliance of the batch of 
infant formula with proposed Sec. 106.91, and provide data needed to 
evaluate a batch of infant formula if problems, such as adverse events 
in infants, occur later with that particular batch. These records will 
show the levels of nutrients in the formula and will provide 
information to help the manufacturer determine whether any problems 
associated with the formula are attributable to the nutrient levels in 
the product.
    Proposed Sec. 106.100(e)(5)(i)(A) requires that manufacturers 
maintain a summary table in the batch record that identifies the stages 
of the manufacturing process at which the nutrient analysis is 
conducted for each nutrient, in accordance with proposed 
Sec. 106.91(a). As discussed above, proposed Sec. 106.91(a) provides 
flexibility in the stage at which many of the nutrients are tested. A 
summary table will facilitate the manufacturer's compliance with 
quality control procedures because it will allow a manufacturer to 
quickly verify that it has tested for all the nutrients required by 
Sec. 107.100 during the production of the infant formula.
    Proposed Sec. 106.100(e)(5)(i)(B) requires that the quality control 
records in the batch record include a summary table on the stability 
testing program, conducted in accordance with proposed Sec. 106.91(b), 
including the nutrients tested and the frequency of testing of 
nutrients throughout the shelf life of the product. As discussed above, 
proposed Sec. 106.91(b) requires that manufacturers test infant formula 
at the beginning, midpoint, and end of the shelf life, and with 
sufficient frequency to ensure that the manufacturer is aware if there 
is a significant deterioration in the required level of a nutrient. 
Therefore, proposed Sec. 106.91(b) provides flexibility in the testing 
frequency, depending on the shelf life and the characteristics of the 
product. A summary table will facilitate the manufacturer's compliance 
with quality control procedures because it will allow a manufacturer to 
quickly determine whether it has tested for all the nutrients required 
by Sec. 107.100 with sufficient frequency to verify that the ``use by'' 
date on the formula is appropriate.
    Proposed Sec. 106.100(e)(5)(ii) requires that the batch records for 
powdered infant formula include the results of any testing conducted in 
accordance with proposed Sec. 106.55(b) to document that the tests were 
done and to verify compliance with the microbiological

[[Page 36192]]

quality standards in proposed Sec. 106.55(c). As discussed above, 
proposed Sec. 106.55(b) requires that manufacturers test representative 
samples of each batch of powdered infant formula to ensure that the 
batch meets the microbiological quality standards in proposed 
Sec. 106.55(c) and therefore is not adulterated. This record will also 
provide the manufacturer with data to evaluate adverse events that 
infants may have experienced after consuming this batch of infant 
formula by showing whether microbiological contamination could have 
contributed to the adverse event.
3. CGMP Records
    Proposed Sec. 106.100(f) identifies the records that manufacturers 
must make and retain pertaining to CGMP described in proposed subpart B 
of part 106. Section 412(b)(4)(A)(i) of the act requires the 
establishment by regulation of requirements for the retention of all 
records necessary to demonstrate compliance with the CGMP, including 
testing designed to prevent the adulteration of infant formula. FDA has 
already discussed proposed regulations (proposed Sec. 106.100(e)) 
respecting the retention of records relating to each batch of infant 
formula. FDA also is proposing regulations respecting the retention of 
records relating to the overall operation of the plant and the 
maintenance of equipment, because these records are necessary to 
demonstrate that the infant formula was manufactured in a manner 
designed to prevent adulteration. Maintenance of these records will 
help manufacturers identify trends in the processing of the infant 
formula, in particular trends that show when the process is breaking 
down in a way that will lead to the production of adulterated product. 
These records also will provide information to assist the manufacturer 
in tracking the cause of adverse events to a formula, if such events 
are reported.
    Proposed Sec. 106.100(f)(1) requires that manufacturers make and 
retain records of the frequency and results of the testing of water 
used in the production of infant formula. These records will show if 
problems are starting to develop with the water supply so that 
manufacturers can take corrective actions before the water is 
inappropriate for use in infant formula.
    Proposed Sec. 106.100(f)(2) requires that manufacturers make and 
retain records, in accordance with Sec. 106.30(d), of accuracy checks 
on instruments and controls. Under this proposal, these records must 
include a certification of the accuracy of any known reference standard 
used and a history of its recertification. As discussed previously, the 
accuracy of the reference standard must be ensured before it can be 
used to ensure that the production instruments are properly calibrated. 
These records also will provide information to assist the manufacturer 
in tracing the source of a problem, if one arises, with a batch of 
infant formula. For example, if infants have adverse events to a batch 
of infant formula, records containing a certification of accuracy of 
the reference standards used and a history of their recertification 
would assist the manufacturer in determining whether the problem was 
created because a production instrument was calibrated with an 
inaccurate reference instrument.
    FDA is proposing to require that, at a minimum, the records specify 
the instrument or control being checked, the date of the accuracy 
check, the standard used, the calibration method used, the results 
found, any actions taken if the instrument is found to be out of 
calibration, and the initials or name of the individual performing the 
test. These records will enable the manufacturer to determine, based on 
the performance of the instrument, whether the calibration schedule is 
sufficient to ensure the accuracy of the instrument. These records also 
will provide information on when and how the instruments were 
calibrated to assist the manufacturer in identifying the cause of a 
problem, if one arises, with a batch of infant formula.
    Including the date of the accuracy check in the record will permit 
a determination of the accuracy of the instrument or control over time; 
including the standard used will allow the manufacturer to verify that 
the standard was properly calibrated; and including the calibration 
method used will ensure that the instrument is being calibrated free 
from the variability that can occur when different laboratory personnel 
perform the same calibration. The results of the accuracy check in the 
record will show whether the instrument or control is accurate, or 
whether a correction was necessary. Documenting the actions taken if 
the instrument is found to be out of calibration will enable the 
manufacturer to ensure that a correction was made. Requiring that the 
individual performing the test note his or her initials or name in the 
record will document who was last responsible for ensuring the accuracy 
of the instrument or control and will allow the manufacturer to discuss 
questions that may arise about the record with the person in the best 
position to know additional, but unrecorded, details about the record.
    If calibration of an instrument shows that a specification or 
standard, at a point, step, or stage in the production process where 
control is deemed necessary to prevent adulteration, has not been met, 
a written evaluation of all affected product, and of any actions that 
need to be taken with respect to that product, needs to be made. For 
example, if the manufacturer is monitoring temperature to ensure that a 
specification or standard of 250  deg.F is maintained as a minimum 
temperature, and calibration of the temperature indicating instruments 
against a reference standard reveals that it was reading a true 
temperature of 248  deg.F, an evaluation of the health hazard 
significance of this temperature deviation must be made. This proposed 
requirement is necessary because, if an instrument is found to have 
been giving inaccurate readings, all infant formula produced subject to 
such inaccuracies must be identified and evaluated for the possibility 
that the inaccuracies caused the formula to be adulterated. In 
identifying the affected product to ensure that the health of 
potentially affected infants is fully protected, in the absence of 
evidence to the contrary, such evaluation would cover all product 
manufactured since the last time the instrument was calibrated and 
found to be accurate.
    Proposed Sec. 106.100(f)(3) requires that manufacturers make and 
retain records, in accordance with proposed Sec. 106.30(e)(3)(ii), of 
the temperatures monitored for cold storage compartments and thermal 
processing equipment. These records are needed to show that the thermal 
processing equipment or cold storage compartments are being maintained 
at the correct temperatures to prevent adulteration of the product. The 
records of these temperatures will enable the manufacturer to identify 
trends in temperature fluctuations that can signal the need to perform 
nonscheduled maintenance.
    FDA is proposing in Sec. 106.100(f)(4) that equipment cleaning, 
sanitizing, and maintenance records, showing the date and time of 
maintenance, as well as the lot number of each batch of infant formula 
processed between equipment startup and shutdown for cleaning, 
sanitizing, and maintenance, be made and maintained. These records will 
allow the manufacturer to ensure that equipment and utensils are being 
cleaned and maintained regularly and to check that the frequency of 
such cleaning, sanitizing, and maintenance is appropriate in light of 
the actual, as

[[Page 36193]]

opposed to planned, use of the equipment. For example, a manufacturer 
may need to increase the frequency of cleaning, sanitizing, and 
maintenance if actual rate of production consistently exceeds the 
predicted rate of production. These records also will allow the 
manufacturer to trace all formula that may be affected if evidence 
becomes available that a particular cleaning, sanitizing, or 
maintenance was improperly performed.
    Proposed Sec. 106.100(f)(4) also requires that the person 
performing and checking the cleaning, sanitizing, or maintenance date 
and sign or initial the record indicating that the work was performed. 
Identification of the person performing and checking the cleaning, 
sanitizing, or maintenance will allow the manufacturer to ensure that a 
qualified person is doing these tasks and to discuss questions that may 
arise about the record with the person in the best position to know 
additional, but unrecorded, details about the record.
    Proposed Sec. 106.100(f)(5) requires that manufacturers make and 
retain records, in accordance with Sec. 106.35(c), on all automatic 
(mechanical or electronic) equipment used in the production or quality 
control of infant formula. Proposed Sec. 106.100(f)(5)(i) requires that 
the automatic equipment records include a list of all systems used, 
with a description of computer files and of the inherent limitations of 
each system. The manufacturer cannot effectively operate the system, 
and correct problems that arise, if it does not understand the system. 
It is not always possible for the individuals who developed and best 
understand the system to be present when the system is operating. 
Therefore, these records will enable the manufacturer to operate and 
troubleshoot the systems even when the individuals who best know the 
system are not available.
    Proposed Sec. 106.100(f)(5)(ii) requires that the automatic 
equipment records include a copy of all software used. Having a copy of 
all software used will minimize the manufacturer's down time if 
problems occur, and parts of the software are lost from the system. For 
example, if a computer virus is found in the software used to run the 
processing lines, having a copy of the software to reload into the 
hardware will minimize the time lost. Likewise, if there is a problem 
with the software used to perform quality control testing, having 
copies of this software will ensure that the testing can continue with 
a minimum amount of time lost.
    Proposed Sec. 106.100(f)(5)(iii) further requires that the 
automatic equipment records document installation, calibration, testing 
or validation, and maintenance of the systems used. These requirements 
are necessary for compliance with section 412(b)(4)(A)(i) of the act. 
As discussed more fully above with respect to proposed Sec. 106.35 
(b)(1), (b)(2), and (b)(4) CGMP requires that all systems be installed, 
calibrated, and maintained in a manner necessary to ensure that they 
are capable of performing their intended function and of producing or 
analyzing infant formula as intended, and that all systems be validated 
before their first use to manufacture commercial product. In addition 
to documenting that the manufacturer is complying with CGMP, records 
documenting installation, calibration, testing or validation, and 
maintenance of systems are necessary to provide information if the 
manufacturer later tries to determine why a problem with the system is 
occurring or tries to determine why the system is not producing an 
infant formula that complies with the manufacturer's specifications for 
the product.
    Proposed Sec. 106.100(f)(5)(iv) requires that the automatic 
equipment records include a list of all persons authorized to create or 
modify software. This record will help to minimize delays when the name 
of a person with those skills is needed quickly.
    Proposed Sec. 106.100(f)(5)(v) requires that the automatic 
equipment records document modifications to software, including the 
identity of the person who modified it. This documentation will ensure 
that the manufacturer is aware of any changes made to the software, and 
that it has a record of how the changed system works, so that it can 
continue to operate the system even in the absence of the responsible 
individual who made the modification to the system. A record of the 
identity of the person who modified the software will show who was 
responsible for modifying the software if problems arise with the 
operation of the system and will identify the person in the best 
position to know additional, but unrecorded, details about the software 
modification to help in troubleshooting the software problems.
    Proposed Sec. 106.100(f)(5)(vi) requires that the automatic 
equipment records include documentation of retesting or revalidation of 
modified systems. This proposed requirement is necessary for compliance 
with section 412(b)(4)(A)(i) of the act. As discussed more fully above 
in the section on proposed Sec. 106.35(b)(5), CGMP requires that all 
modifications to software be made by a designated individual, and that 
all systems be revalidated after any modification to ensure that infant 
formula produced or analyzed using the modified software complies with 
subparts B and C. FDA has tentatively concluded that records on 
retesting or revalidation of the modified systems, just like records on 
the initial testing or validation of the system 
(Sec. 106.100(f)(5)(iii)), are necessary to document that the work has 
been done properly and to provide information if the manufacturer later 
tries to determine why a problem with the system is occurring or tries 
to determine why the system isnot producing an infant formula that 
complies with the manufacturer's specifications for the product.
    Proposed Sec. 106.100(f)(5)(vii) requires that the manufacturer 
make and retain a backup file of data entered into a computer or 
related system. It also requires that this backup file consist of a 
hard copy or alternative system, such as duplicate diskettes, tapes, or 
microfilm, designed to ensure that backup data are exact and complete, 
and that they are secure from alteration, inadvertent erasures, or 
loss. This proposed requirement is necessary to ensure compliance with 
CGMP because computer files can be easily altered or erased. Backup 
files of data will allow the manufacturer to readily reload the files 
of data if problems occur in the operation of the computer or related 
system, so that the manufacturer's down time is minimized, and so that 
the data entered into the system will be an exact copy of the data 
previously used in the system.
    Proposed Sec. 106.100(f)(6) requires that manufacturers make and 
retain records on ingredients, containers, and closures, including the 
identity and quantity of each lot, the name of the supplier, the 
supplier's lot number, the name and location of the manufacturer (if 
different from the supplier), the date of receipt, and the receiving 
code as specified (proposed Sec. 106.100(f)(6) (i) through (vi)). These 
records will enable the manufacturer to document that it is complying 
with proposed Sec. 106.40(g). Moreover, this information is needed to 
enable the manufacturer to track the source of each ingredient, 
container, or closure used in infant formula if a problem arises. If an 
ingredient, container, or closure is found to cause adulteration of the 
formula, it is important to be able to determine the source of the 
material, so that use of such materials can be halted and prevented in 
the future.
    Proposed Sec. 106.100(f)(6)(vii) requires that the records on 
ingredients, containers, and closures include the results and 
conclusions of any test or examination, including retesting and

[[Page 36194]]

reexamination, performed on them and their disposition. These records 
will document that appropriate testing is being conducted to ensure 
that the ingredients will not adulterate the infant formula, and that 
the containers and closures will protect the infant formula against 
adulteration. Further, these records will show the basis on which each 
ingredient, container, and closure was released for use in infant 
formula production if questions about such release later arise. 
Individual lots of ingredients, containers, and closures are likely to 
be used in a number of different batches of infant formula; therefore, 
the agency is proposing that the records on ingredients, containers, 
and closures be a part of the records pertaining to CGMP. Retaining 
such records in the CGMP records, rather than in each batch record, 
will eliminate the duplication of records and simplify the 
recordkeeping. The disposition of the ingredients, containers, and 
closures will show which materials were destroyed because they did not 
meet the manufacturers specifications (and not used in manufacture in 
compliance with Sec. 106.40(d)), and which batches of infant formula 
were made using each lot of ingredients, containers, or closures. Thus, 
the manufacturer will know which lots of ingredients, containers, or 
closures were used in making infant formula and will be able to confirm 
that those lots complied with proposed Sec. 106.40(d). Moreover, if a 
batch of formula is shown to be adulterated, these records will help 
the manufacturer to identify the source of the adulteration.
    Proposed Sec. 106.100(f)(7) requires that manufacturers make and 
retain records that include a full description of the methodology used 
to test powdered infant formulas to verify compliance with proposed 
Sec. 106.55(c) and the methodology used to conduct quality control 
testing in accordance with Sec. 106.91 (a) and (b). The agency has not 
specified in these regulations the methodologies that must be used to 
conduct microbiological and quality control testing. Thus, FDA has 
tentatively concluded that a manufacturer needs to maintain a record 
that fully describes the methodology that it has decided to use to test 
powdered infant formula for microorganisms and for quality control 
testing. Such a record is necessary if there is to be consistency in 
the procedure that the manufacturer follows in testing each batch of 
infant formula, particularly in light of the fact that the laboratory 
personnel conducting the testing are likely to vary. The accuracy and 
reproducibility of microbiological and quality control testing depend 
on the procedure used to conduct the test.
    FDA is proposing that the full description of the methodology be 
retained as part of the CGMP records, rather than in the batch record 
provided for in proposed Sec. 106.100(e)(5), because these methods will 
be used to test multiple batches of infant formula. Retaining such 
records in the CGMP records, rather than in each batch record, will 
mean that the manufacturer has to maintain only one document, rather 
than having to reproduce it each time that it runs a batch of formula. 
Thus, the proposed approach will eliminate duplication of records and 
simplify recordkeeping.
4. Records on Distribution of Infant Formulas
    Proposed Sec. 106.100(g) adds to current Sec. 106.100(g) a 
requirement that records pertaining to distribution of the infant 
formula show that products intended for export only are in fact 
exported. It has recently come to the attention of the agency that 
infant formulas produced for export have been diverted and sold in the 
United States. All persons introducing any new infant formula into 
interstate commerce, which includes persons exporting an infant formula 
to a foreign country, are required by section 412(c) of the act to 
register and make a submission to the agency 90 days before marketing 
the formula. (See discussion of proposed Secs. 106.110 and 106.120.)
    As discussed in the section of this preamble on proposed 
Sec. 106.120(c), the agency has tentatively concluded that it will not 
require manufacturers who produce infant formula for export only to 
submit the same information that would be required for products 
intended or offered for sale in the United States. In lieu of the 
information required by Sec. 106.120(b), FDA is proposing to allow 
manufacturers of products for export only to give assurances that the 
infant formula will not be sold or offered for sale in domestic 
commerce. This provision is based, in part, on section 801(e) of the 
act, which states that a food will not be deemed to be adulterated or 
misbranded under the act if, among other things, it is not sold or 
offered for sale in domestic commerce. Thus, the agency has tentatively 
concluded that the additional recordkeeping requirement on distribution 
of infant formulas for export only in proposed Sec. 106.100(g) is 
necessary so that verification that the infant formula was not in fact 
sold or offered for sale in domestic commerce will be readily available 
in the manufacturer's records.
5. Audit Records
    Proposed Sec. 106.100(j) carries forward the requirement in current 
Sec. 106.100(j) that the manufacturer make and retain records, which 
include the audit plans and procedures, that pertain to regularly 
scheduled audits. As discussed above, the written audit plan, which 
includes audit procedures, is required under proposed Sec. 106.94(a) 
and (b). The proposed section further requires that records of audits 
include the findings of the audit and a listing of any changes made in 
response to these findings. This requirement is proposed under the 
authority of section 412(b)(4)(A)(v) of the act, which requires that 
manufacturers retain all records of the results of regularly scheduled 
audits conducted under the requirements prescribed by the Secretary 
(and by delegation, FDA) under the authority of section 
412(b)(2)(B)(iv).
    Proposed Sec. 106.100(j) also requires that the manufacturer make 
readily available for authorized inspection the audit plans and 
procedures and a statement of assurance that the regularly scheduled 
audits are being conducted. This provision implements section 
412(b)(4)(B)(ii) of the act, which requires that the manufacturer 
provide written assurance that the regularly scheduled audits are being 
conducted by the manufacturer. However, proposed Sec. 106.100(j) also 
provides that the findings of the audit and any changes made in 
response to these findings need not be made available to FDA. This 
provision is brought forward from current Sec. 106.100(j) and reflects 
section 412(b)(4)(B)(ii) of the act, which states that a ``manufacturer 
need only provide written assurances to the Secretary that the 
regularly scheduled audits required by'' section 412(b)(2)(B)(iv) of 
the act ``are being conducted by the manufacturer, and need not make 
available to the Secretary the actual written reports of such audits.''
6. Modification of Current Sec. 106.100(k)(3)
    The agency also is revising current Sec. 106.100(k)(3) to reflect 
the numbering changes in the regulations on notifying the agency of a 
causal relationship between the consumption of an infant formula and an 
infant's death. The agency is moving the requirements of current 
Sec. 106.120(b) to Sec. 106.150 to reflect the changes it is proposing 
in subpart G. Thus, the reference to Sec. 106.120 in Sec. 106.100 
(k)(3) will be changed to read ``Sec. 106.150,'' if the

[[Page 36195]]

agency adopts the relevant proposed changes.

G. Registration, Submission, and Notification Requirements

1. Introduction
    The act provides for three types of notices that manufacturers of 
infant formula must provide to FDA and sets forth the general 
information that must be included in each type of notice. First, 
manufacturers of a new infant formula must register with FDA, in 
accordance with section 412(c)(1)(A) of the act, providing the name and 
address of the firm and all establishments that will manufacture the 
new infant formula. Second, manufacturers must submit to FDA, in 
accordance with section 412(d) of the act, certain information 
concerning a new infant formula or an infant formula in which there is 
a change in formulation or processing that may affect whether the 
formula is adulterated under section 412(a) of the act. Third, 
manufacturers must notify FDA, in accordance with section 412(e) of the 
act, of any adulterated or misbranded infant formula that has left 
their control.
    The agency has not specified the information that must be included 
in an infant formula registration, submission, or notification. While 
firms have been able to function under these requirements since the 
1986 amendments were enacted with respect to the notice that 
manufacturers must provide to the agency under section 412(c) and (d) 
of the act, inquiries from industry suggest that manufacturers are 
uncertain about the information that they must provide. Some 
manufacturers have needed to make multiple submissions for a new infant 
formula because of deficiencies in the initial submission. For example, 
some submissions have contained information concerning more than one 
formula without clearly identifying which information applied to which 
formula. Some submissions have not contained the information required 
by section 412(d)(1) of the act. Therefore, FDA recognizes that it will 
be useful both to manufacturers and to the agency to issue regulations 
to ensure that registrations and submissions required by the act follow 
a consistent format and contain the necessary information for the 
agency to determine whether there is a basis to object to the marketing 
of a new infant formula. Such regulations will facilitate the 
manufacturer's preparation of the notice and also will facilitate the 
agency's review of the notice once FDA receives it.
    These proposed regulations also will make clear when a 
registration, notification, or submission to the agency is needed. For 
example, as stated above, it has recently come to the attention of the 
agency that some firms that manufacture infant formula intended only 
for export are not aware of their registration and submission 
responsibilities. Section 412(c)(1) of the act requires that a person 
introducing a new infant formula into interstate commerce (which 
includes export to a foreign country) must register the infant formula 
and make the proper submission 90 days before marketing it. These 
proposed regulations make clear that registration and submission 
requirements apply to infant formulas intended only for export as well 
as to infant formula intended for the domestic market.
    Finally, for completeness, FDA has decided that it would be useful 
to both manufacturers and the agency, to carry forward current 
Sec. 106.240, concerning notification of a violative infant formula, as 
Sec. 106.150. Doing so will consolidate in one place in the agency's 
regulations all requirements concerning notice to the agency to meet 
the requirements of section 412(c), (d), and (e) of the act.
2. New Infant Formula Registration
    Proposed Sec. 106.110(a) requires that a manufacturer of a new 
infant formula register with FDA before introducing the formula, or 
delivering it for introduction, into interstate commerce. Because 
``interstate commerce'' is defined in section 201(b) of the act as 
``(1) commerce between any State or Territory and any place outside 
thereof, and (2) commerce within the District of Columbia or within any 
other Territory not organized with a legislative body,'' under this 
provision, a manufacturer is required to register with FDA before 
introducing a new infant formula into the United States market or 
before beginning exporting the formula. Proposed Sec. 106.110(a) sets 
out how to comply with section 412(c)(1)(A) of the act. Failure to 
provide the notice required by section 412(c)(1)(A) of the act is a 
prohibited act under section 301(s).
    Under section 412(c)(1)(A) of the act, proposed Sec. 106.110(b) 
sets out the information required in a new infant formula registration. 
While manufacturers may register at any time before introducing a new 
formula into interstate commerce, FDA urges that they do so at the same 
time that they submit notice of their intent to market a new infant 
formula in accordance with section 412(c)(1)(B) and (d)(1) of the act. 
Receiving registration and the 90 day submission at the same time will 
facilitate the agency's review.
3. New Infant Formula Submission
    Section 412(c)(1)(B) of the act requires that manufacturers of a 
new infant formula submit to FDA a notice of their intent to market the 
new formula that complies with section 412(d)(1) of the act. The notice 
must be submitted at least 90 days before the infant formula is 
introduced or delivered for introduction into interstate commerce 
6. Proposed Sec. 106.120 implements this requirement.
---------------------------------------------------------------------------

    \6\ While section 412(c)(1) and (c)(1)(B) of the act state ``No 
person shall introduce or deliver for introduction into interstate 
commerce any new infant formula unless--such person has at least 90 
days before marketing such new infant formula, made the submission 
to the Secretary required by'' section 412(c)(1) of the act, FDA has 
recognized since 1986 that this citation is in error (see 
``Requirements for Infant Formulas'' published by FDA's Industry 
Programs Branch, CFSAN), and that the correct citation is section 
412(d)(1). This correction agrees with the language of section 
412(d)(1) of the act, which states what a submission about any 
infant formula subject to section 412(c) of the act should include. 
It is also consistent with the rules of statutory construction. See 
Colonial Life & Accident Insurance Co. v. American Family Life 
Assurance Co., 846 F. Supp. 454, 463 n. 14(D.S.C. 1994) (where the 
legislature has made a mistake in reference, and its intent is 
manifest, the statute may be read as corrected in order to give 
effect to the legislative intent).
---------------------------------------------------------------------------

    Proposed Sec. 106.120(a) sets out the requirement that a 
manufacturer submit a notice of its intent to market a new infant 
formula and provides the address to which such notices are to be 
submitted.
    Proposed Sec. 106.120(b) sets forth the information that 
manufacturers must include in their new infant formula submission. This 
proposed regulation implements and specifies the information called for 
in section 412(d)(1) of the act.
    a. General information required in a 90-day submission. Because the 
registration of a new infant formula (proposed Sec. 106.110) need not 
accompany the new infant formula submission (proposed Sec. 106.120), 
and because a third submission on a newinfant formula that verifies 
that the new infant formula, as produced, contains all required 
nutrients (see proposed Sec. 106.130) will be submitted separately, FDA 
has tentatively concluded that the name of the infant formula is needed 
to ensure that all information on a particular infant formula is filed 
together and is available to determine whether the agency should object 
to the marketing of the formula. Information on the form of the product 
is necessary for an accurate evaluation of the product because 
different

[[Page 36196]]

requirements may apply to different forms of a formula. For example, 
powdered infant formula must meet the microbiological quality 
requirements in proposed Sec. 106.55, whereas liquid forms of a formula 
do not. Therefore, FDA is proposing to require this information in 
Sec. 106.120(b)(1), under the authority of sections 412(d)(1) and 
701(a) of the act, even though it is not explicitly required in section 
412(d)(1).
    Proposed Sec. 106.120(b)(2) requires that the submission include an 
explanation of why the formula is a new infant formula to facilitate a 
determination by the agency as to the type of evaluation the new infant 
formula requires. For example, if the formula is a new infant formula 
because a new manufacturing plant will be used to produce it, but the 
formulation of the product is not changed, FDA will evaluate the 
processing and arrange to inspect the new facility but may conclude 
that testing to provide assurance that quality factor requirements have 
been met is not necessary. Thus, FDA is proposing to require the 
submission of this information, even though, like the information 
required under proposed Sec. 106.120(b)(1), submission of this 
information is not specifically provided for in the act. The agency 
tentatively concludes that this information is necessary for the 
efficient enforcement of sections 412(c)(1)(B) and (d)(1) of the act.
    b. Formulation and processing information required in a 90-day 
submission. Pursuant to section 412(d)(1)(A) of the act, proposed 
Sec. 106.120(b)(3) requires that the submission include the 
quantitative formulation of the infant formula. The agency is proposing 
that, if the notice concerns more than one form of the formula, the 
submission include quantitative information on each form of the formula 
that is the subject of the notice. FDA is proposing to require that 
manufacturers submit the formulation in units per volume (for liquid 
formulas) or units per dry weight (for powdered formulas) because 
formulations expressed in these units will facilitate agency 
understanding of the formula. Manufacturers already will have the 
formulation available in these units as a part of the master 
manufacturing order, and submitting the formulations in these units 
should not require additional calculations by the manufacturer.
    Proposed Sec. 106.120(b)(3) also requires, under section 
412(d)(1)(B) of the act, that the submission include a description of 
any reformulation of the infant formula, including a listing of each 
new or changed ingredient and a discussion of the effect of such 
changes on the nutrient levels in the formulation. For example, if the 
protein source in an infant formula is replaced with a protein source 
that contains a different amount of protein (e.g., from casein to a 
mixture of casein and whey), it is important to ensure that the amount 
of the new protein source used will provide the amount of protein 
required by Sec. 107.100. As another example, if an ingredient such as 
sodium selenite is added to the formula for the first time, it is 
important to ensure that the level of the ingredient provides selenium 
(in the form of selenite) at a level that is consistent with the 
infant's needs and yet within the safe range of selenium intake.
    Proposed Sec. 106.120(b)(4) requires that the submission include a 
description, when applicable, of any change in processing of the infant 
formula, and that such description identify the specific change in 
processing, including side-by-side, detailed schematic diagrams 
comparing the new processing to the previous processing (including 
processing times and temperatures). This proposed requirement 
implements section 412(d)(1)(B) of the act, which states that the 
submission must include a description of any change in the processing 
of an infant formula. FDA is proposing that the description of the 
change in processing include detailed schematic diagrams comparing the 
new processing to the previous processing because schematic diagrams 
are efficient tools for identifying the nature and significance of 
changes in processing.
    c. Assurance that the infant formula will not be marketed unless it 
meets quality factor and nutrient requirements of the act. Pursuant to 
section 412(d)(1)(C) of the act, proposed Sec. 106.120(b)(5) requires 
that the submission include an assurance that the infant formula will 
not be marketed unless it meets the quality factor requirements of 
section 412(b)(1) of the act and the nutrient content requirements of 
section 412(i) of the act.
    Proposed Sec. 106.120(b)(5)(i) requires that the assurance that the 
formula meets the quality factor requirements, which are set forth in 
subpart E of part 106, be provided by a submission that complies with 
Sec. 106.121. Section 412(d)(1)(C) of the act requires that, 90 days 
before marketing a new infant formula, a manufacturer submit assurances 
that the infant formula will not be marketed unless it meets the 
quality factor requirements established by regulations under section 
412(b)(1). Section 412(d)(2) of the act requires that, after the first 
production of a new infant formula and before introduction into 
interstate commerce of such formula, the manufacturer submit a written 
verification that summarizes test results and records demonstrating 
that such formula complies with the quality factor requirements. 
However, FDA has tentatively concluded that to implement sections 412 
(d)(1) and (d)(2) of the act in a way that ensures that the statutory 
goals are achieved--that is, to ensure that the agency has all the 
relevant information for a sufficient period of time to conduct a 
meaningful review of the nutritional adequacy of the formula while 
enabling the infant formula manufacturer to market its product as 
expeditiously as possible--it is appropriate to require that the 
assurances that the quality factors will be met be provided by means of 
data that would otherwise be required as part of the verification 
submission. FDA notes that such a requirement would only codify current 
practice. Since passage of the 1986 amendments, infant formula 
manufacturers have been providing data demonstrating that a new infant 
formula meets the quality factor requirements as a part of the 
submission made 90 days before marketing.
    Proposed Sec. 106.120(b)(5)(ii) requires that the assurance that 
the formula complies with the nutrient content requirements, which are 
set forth in Sec. 107.100, be provided by a statement assuring that the 
formula will not be marketed unless it meets the nutrient requirements 
of Sec. 107.100, as demonstrated by testing required under subpart C of 
part 106.
    The agency acknowledges that there is an apparent inconsistency in 
how it interprets the word ``assurance'' in section 412(d)(1)(C) of the 
act as it relates to assurance that the infant formula meets the 
quality factor requirements and assurance that the infant formula meets 
nutrient content requirements. FDA has tentatively concluded, however, 
that assurance that the formula will meet the quality factor 
requirements is a threshold question that must be answered 
affirmatively before the effort in setting up the line for first 
production of the infant formula would be justified. Therefore, the 
agency is proposing to require that the assurance that the infant 
formula will meet the quality factor requirements be provided by data 
submitted 90 days before marketing the formula.
    On the other hand, the agency is proposing that the assurance that 
the formula will not be marketed unless it meets the nutrient 
requirements of Sec. 107.100 can be provided by a statement to that 
effect (as opposed to data) submitted 90 days before marketing of the 
formula because the

[[Page 36197]]

data and records demonstrating that the formula complies with the 
nutrient requirements of Sec. 107.100 will not be available until the 
production line is set up, and the first production of the infant 
formula has occurred. FDA will receive verification that the formula 
meets the nutrient requirements as a part of the submission required by 
section 412(d)(2) of the act (see proposed Sec. 106.130(b)(3), below). 
Therefore, FDA has tentatively concluded that it is adequate to receive 
a commitment from the manufacturer, 90 days before marketing, that the 
infant formula will not be marketed unless it meets the nutrient 
requirements of Sec. 107.100.
    d. Assurance that the processing of the infant formula complies 
with the CGMP and quality control procedures of the act. Under section 
412(d)(1)(D) of the act, proposed Sec. 106.120(b)(6) requires that the 
submission include assurance that the processing of the infant formula 
complies with section 412(b)(2) of the act (CGMP, including quality 
control procedures).
    Proposed Sec. 106.120(b)(6)(i) requires that the assurance that the 
processing of the infant formula complies with section 412(b)(2) of the 
act include a statement that the formula will be produced in accordance 
with subparts B and C of part 106. This proposed requirement is a 
necessary element of the assurance required by section 412(d)(1)(D) of 
the act because the requirements for CGMP are set forth in subpart B 
and the requirements for quality control procedures are set forth in 
subpart C. In the Congressional Record (Ref. 1), Senator Metzenbaum 
stated that the amendments to the Infant Formula Act set up 
requirements ``which will prevent our Nation's Children from ever again 
being threatened by defective baby formula. The most important 
provision of this amendment is the simple requirement that each batch 
of formula must be tested for each essential nutrient that must be 
contained in the formula'' (Ref. 1).
    Proposed Sec. 106.120(b)(6)(ii) requires that the assurance that 
the processing of the infant formula complies with section 412(b)(2) of 
the act include the basis on which the manufacturer has concluded that 
each ingredient meets the requirement of proposed Sec. 106.40(a), i.e., 
that the ingredient is an approved food additive, is authorized by a 
prior sanction issued by the agency, or is GRAS for its intended use. 
The statute provides that the manufacturer submit, 90 days before 
marketing a new infant formula, assurance that the processing of the 
formula complies with the CGMP regulations, and that the formula is 
manufactured in a way that is designed to prevent its adulteration. FDA 
has tentatively concluded that, to implement the act in a way that will 
ensure that the statutory goals are achieved, that is, to ensure that 
the agency has all the relevant information for a sufficient period of 
time to conduct a meaningful review of the formula while enabling the 
manufacturer to market its product as expeditiously as possible, it is 
appropriate to require that the assurance that none of the ingredients 
will adulterate the formula be provided by an explanation of how each 
ingredient meets proposed Sec. 106.40(a). FDA has tentatively concluded 
that this approach is appropriate because, like the evidence that the 
formula meets the quality factors, evidence that all the ingredients in 
the infant formula are safe goes to a threshold question that must be 
answered affirmatively before the effort in setting up the production 
line for the first production of the infant formula would be justified. 
Moreover, an infant formula manufacturer would want to have information 
demonstrating that each of the ingredients in the formula is safe 
before marketing the formula, because without such information, a 
responsible manufacturer would not include the ingredient in its 
product.
    FDA will review the new infant formula submission to ensure that a 
safe product will be produced (sections 201(s), 402(a)(1) and (a)(2), 
and 409 of the act). If the agency is not presented with basis on which 
it can be satisfied that the use of an ingredient in an infant formula 
will be safe, FDA will not be able to acquiesce in the marketing of the 
formula. The legislative history of the 1986 amendments supports that 
Congress anticipated that FDA would provide this type of review. In the 
Congressional Record of September 27, 1986, Senator Metzenbaum stated:

    I continue to be concerned, however, that our food and drug laws 
do not differentiate between foods and infant formulas. But they are 
fundamentally different. An infant formula is designed as the sole 
source of nutrition for a baby. An infant formula is used daily. A 
baby must thrive from its content for the first and most formative 
months of his or her life. I expect the Secretary to look closely at 
whether or not our standards in this area for foods are adequate 
standards for infant formula. I have no reason at this time to 
suspect that there is a problem here. But I continue to urge the 
Secretary to give thorough consideration to the important 
distinctions between infant formula and other foods, as well as food 
additives which may be used with infant formulas. (Ref. 1)

    One way for a manufacturer to satisfy the agency that proposed 
Sec. 106.40(a) is met would be for the manufacturer to use only 
ingredients that are: (1) Listed as GRAS for such use in 21 CFR part 
182 or affirmed as GRAS for such use in 21 CFR part 184 or otherwise 
GRAS for such use under the regulations included in those parts; (2) 
approved for such use by a food additive regulation; or (3) authorized 
by a prior sanction issued by FDA.
    Alternatively, the requirements of proposed Sec. 106.40(a) can be 
met by a showing that the substance is GRAS within the meaning of 
Sec. 170.30 (21 CFR 170.30), which states that ``general recognition of 
safety may be based only on the view of experts qualified by scientific 
training and experience to evaluate the safety of substances directly 
or indirectly added to foods'' (Sec. 170.30(a)). To clarify this point, 
Sec. 170.30(a) states that ``[g]eneral recognition of safety requires 
common knowledge about the substance throughout the scientific 
community knowledgeable about the safety of substances directly or 
indirectly added to food.'' The qualified experts can base their views 
on either: (1) Scientific procedures, or (2) in the case of a substance 
used in food prior to January 1, 1958, through experience based on 
common use in food (section 201(s) of the act).
    Under Sec. 170.30(b), general recognition of safety based upon 
scientific procedures requires the same quantity and quality of 
scientific evidence as is required to obtain approval of the ingredient 
as a food additive, and it must ordinarily be based on published 
studies, which may be corroborated by unpublished studies and other 
data and information. If the manufacturer of an infant formula wishes 
to use an ingredient because there is general recognition of safety 
based upon scientific procedures, FDA is proposing to require in 
Sec. 106.120(b)(6)(ii) that the manufacturer include as a part of its 
new infant formula 90-day submission the rationale for why the 
ingredient is GRAS and the evidence that demonstrates that there is 
common knowledge about the safety of the substance throughout the 
scientific community knowledgeable about the safety of such substance. 
FDA is proposing that this evidence include a bibliography of published 
studies, copies of those scientific publications about the substance, 
and an explanation as to why, based on the published studies, the use 
of the substance in infant formula is GRAS.
    Under Sec. 170.30(c)(1), if a substance is GRAS based on common use 
in food prior to January 1, 1958, this determination must be based 
solely on

[[Page 36198]]

food use of the substance before January 1, 1958, and must ordinarily 
be based upon generally available data and information. Thus, GRAS 
based on common use in food prior to January 1, 1958, may be determined 
without the quantity or quality of scientific procedures required for 
approval of a food additive regulation. If the manufacturer of an 
infant formula wishes to use an ingredient based solely on food use of 
the substance prior to January 1, 1958, it should provide as a part of 
the new infant formula 90-day submission the evidence supporting that 
the ingredient was in common use in infant formula prior to January 1, 
1958, and an explanation of why that use provides the basis for general 
recognition of the safety of the substance.
    FDA has recognized that it is impractical to list all substances 
that are GRAS for their intended use based on their common use in food 
prior to 1958 (see 21 CFR 182.1(a)). The agency regards such common 
food ingredients as salt, pepper, vinegar, and baking powder as safe 
for their intended use. Also, current Sec. 170.30(d) provides that a 
``food ingredient of natural biological origin that has been widely 
consumed for its nutrient properties in the United States prior to 
January 1, 1958, without known detrimental effects, which is subject 
only to conventional processing as practiced prior to January 1, 1958, 
and for which no known safety hazard exists, will ordinarily be 
regarded as GRAS * * *.'' Some ingredients are used in infant formulas 
even though they are not listed or affirmed as GRAS by the agency for 
their intended use. Vitamin K, for example, is required to be a part of 
an infant formula under section 412(i) of the act and, in the form of 
phylloquinone, is considered to be safe and suitable for infant 
formulas when used in accordance with prescribed levels in 
Sec. 107.100, although no source of vitamin K, such as phytonadione or 
phylloquinone, has been listed or affirmed as GRAS by the agency. 
Likewise, sodium selenite has been added to infant formulas to provide 
the amount of selenium that has been determined to be essential for 
infants by NAS (Ref. 19). Published experimental and clinical data 
provide a basis upon which experts qualified by scientific training and 
experience could evaluate the safety of sodium selenite as a source of 
selenium for use in infant formula and could conclude that it is safe. 
The agency anticipates that other ingredients may be shown to be GRAS 
because they are generally accepted sources of substances that are 
established as essential for infants by an authoritative body such as 
NAS. However, manufacturers should not take this acknowledgment to mean 
that they are free to declare that the use of any ingredient they want 
to use is GRAS. Any ingredient that cannot meet the standard of 
Sec. 170.30 for a GRAS determination will be viewed by the agency as a 
food additive, and any infant formula that contains a food additive 
that the agency has not approved for use in infant formula is subject 
to being acted against by the agency.
    If the safety of an ingredient is not expressly recognized in an 
FDA regulation, the burden will rest on the manufacturer of the infant 
formula to include in its new infant formula submission an explanation 
of why the substance is GRAS under Sec. 170.30, along with the 
published and other information that provides the basis for that 
explanation, in accordance with proposed Sec. 106.120(b)(6)(ii). If the 
agency adopts this approach, a failure of the agency to object to a 
manufacturer's determination that an ingredient is GRAS in a new infant 
formula submission will not constitute a GRAS affirmation by the 
agency. However, if FDA knows of no reason to question the safety of an 
ingredient to be used in infant formula, the agency will not object to 
the manufacturer's relying on its own determination that use of the 
substance is GRAS.
    e. Submission 90 days before marketing a new infant formula 
intended only for export. When a new infant formula is intended only 
for export, proposed Sec. 106.120(c) provides that manufacturers may 
submit, in lieu of the information required under proposed 
Sec. 106.120(b), a statement that the infant formula meets the 
specifications of the foreign purchaser, does not conflict with the 
laws of the country to which it is to be exported, is labeled on the 
outside of the shipping package to indicate that it is intended for 
export only, and will not be sold or offered for sale in domestic 
commerce. This proposed requirement recognizes that under section 
801(e) of the act, in certain limited circumstances, manufacturers may 
lawfully export products that are adulterated or misbranded. The 
information required under proposed Sec. 106.120(c) will demonstrate 
that those limited circumstances exist. FDA has tentatively concluded 
that proposed Sec. 106.120(c) will provide manufacturers with the 
flexibility allowed under section 801(e) of the act while meeting the 
requirements of sections 412(c) and (d) of the act.
    f. Submission 90 days before marketing--administrative procedures. 
Proposed Sec. 106.120(d) states that the submission will not constitute 
notice under section 412 of the act unless it complies fully with 
Sec. 106.120(b), and the information that it contains is set forth in a 
manner that is readily understandable, so that FDA can complete its 
review in a timely manner and advise the manufacturer if it has any 
concerns about the marketing of the formula before the 90 days is up. 
Proposed Sec. 106.120(d) makes clear that the agency will notify the 
submitter if the notice is not adequate because it does not meet the 
requirements of sections 412(c) and (d) of the act.
    Proposed Sec. 106.120(e) provides that if a new infant formula 
submission contains all the information required by proposed 
Sec. 106.120(b), FDA will acknowledge its receipt and notify the 
manufacturer of the date of receipt, which will be the filing date for 
the submission (and the manufacturer will be able to plan those actions 
necessary to begin marketing the new formula in reliance on that date). 
Further, pursuant to section 412(c)(1)(B) of the act, proposed 
Sec. 106.120(e) also requires that the manufacturer not market the new 
infant formula until 90 days after the filing date. Congress provided 
for 90-day notice so that the agency would have sufficient time to 
examine all of the material submitted and decide whether there is any 
basis for concern about the marketing of the formula.
    Proposed Sec. 106.120(f) makes clear that if the manufacturer 
provides additional information in support of a new infant formula 
submission, FDA will determine whether it represents a substantive 
amendment to the submission, and that, if it does, FDA will assign the 
new infant formula submission a new filing date. FDA is proposing to 
adopt Sec. 106.120(f) to clarify how it will treat amendments to infant 
formula notifications. In the 9 years since the passage of the 1986 
amendments, the treatment of additional submissions has been the source 
of some confusion. FDA has tentatively concluded that it is necessary 
to give a new filing date to a new infant formula submission when a 
substantive amendment is made to it so that the agency has time to 
examine all of the material submitted and to determine whether there is 
any basis for concern about the marketing of the formula.
4. Quality Factor Submission
    Proposed Sec. 106.121 sets forth the requirements for specific 
information that a manufacturer must submit to

[[Page 36199]]

FDA, in accordance with proposed Sec. 106.120(b)(5), to provide 
assurance that the infant formula meets the quality factor requirements 
set forth in subpart E of part 106. FDA has tentatively concluded that 
agency access to study records and data are necessary so that it can 
ensure that study results are meaningfully interpretable, and that the 
manufacturer's conclusion that the infant formula meets the quality 
factor requirements withstands scientific scrutiny and evaluation. 
Failure to adequately document study results and interpretation raises 
questions as to the validity of conclusions and could mean that infants 
have been unnecessarily subjected to testing procedures.
    Proposed Sec. 106.121(a) requires that the manufacturer submit an 
explanation, in narrative form, setting forth its conclusions on how 
all quality factor requirements of subpart E of part 106 have been met. 
This narrative will facilitate the agency's review by summarizing the 
results, and their interpretation, that provide the basis on which the 
manufacturer has concluded that the quality factor requirements have 
been met, or that the subject infant formula is eligible for the 
exemptions described in proposed Sec. 106.97(a)(2) and (b)(2).
    Proposed Sec. 106.121(b) requires that the manufacturer submit 
records that contain the information collected during the study for 
each infant enrolled in the study. The measurements and information 
collected for each infant enrolled in the study are necessary to an 
evaluation of whether the infant formula supported healthy growth. 
Proper identification of the records is necessary for proper use and 
analysis of the records.
    Proposed Sec. 106.121(c)(1) requires that the manufacturer submit a 
statistical evaluation of the data from the clinical study, including 
group means, group standard deviations, and measures of statistical 
significance for all measurements for each feeding group at the 
beginning of the study and at every point where measurements were made 
throughout the study. This evaluation forms the basis for the 
manufacturer's conclusion as to whether the formula meets the quality 
factor requirements. Without knowledge of the statistical basis upon 
which the manufacturer drew its conclusions, FDA would not have 
sufficient information to evaluate the conclusions reported by the 
manufacturer.
    Proposed Sec. 106.121(c)(2) requires that the manufacturer submit a 
calculation of the statistical power of the study at its completion. 
Proposed Sec. 106.97(a)(1)(ii)(E) recommends that the power calculation 
used to design the study be included in the study protocol. FDA is 
aware that circumstances (e.g., attrition, difficulty in recruiting 
sufficient numbers of infants, unexpectedly high measurement error in a 
particular variable) may unintentionally result in sample sizes and 
feeding group assignments that lack adequate statistical power for 
detecting differences between treatment and control groups, regardless 
of the apparent adequacy in planning for the study protocol. Reviewers 
must be aware of changes in the statistical power of a study so that 
they do not inadvertently misinterpret the absence of differences that 
occur between different formulas as meaning there are no differences. 
Failure to detect differences, if they are real, could result in 
erroneously concluding that a formula is safe and suitable for its 
intended use when, in fact, it is not. The agency is proposing to 
require that the manufacturer submit this calculation to FDA so that 
the agency can meaningfully review and interpret the data and study 
results contained in the submission.
    Proposed Sec. 106.121(d) requires that the manufacturer submit 
reports on attrition and on all occurrences of adverse events during 
the study.
    FDA has tentatively found that information on the occurrence of 
adverse events is a critical element of the data that must be evaluated 
to determine whether a formula meets quality factor requirements and is 
safe and suitable for infants. Adverse events associated with the use 
of an infant formula, although unexpected, can be a sign or symptom of 
a nutritional inadequacy or of a safety problem with the infant 
formula, and failure to use these results could result in inadvertent 
release of an unsafe product. Conversely, adverse events can be 
unrelated to a formula product (e.g., flu), but their occurrence can 
affect the way in which results are interpreted and used. For example, 
illnesses can influence the interpretation of growth data and of the 
laboratory measurements collected to evaluate the infant formula.
    For these reasons, FDA has tentatively concluded that complete 
reports, including the results of followup investigations, on the 
occurrence of all adverse events during the study, regardless of 
whether the adverse events are attributable to the use of the new 
infant formula or to some other illnesses, are necessary to properly 
evaluate the conclusions drawn from a clinical study (proposed 
Sec. 106.121(d)(1)). FDA has tentatively concluded that a complete 
report on the occurrence of an adverse event must include 
identification of the infant by subject number to permit evaluation of 
infant growth measurements; identification of the feeding group to show 
whether there is a pattern of adverse events in one feeding group 
versus another; and a complete description of the adverse event, 
including comparisons of the frequency of occurrence in each feeding 
group and information on the health of the infants during the course of 
the study, including the occurrence and duration of any illness, that 
occurred during the trial, so that it is possible to evaluate the 
significance of the illness.
    As discussed above, it is very important to be able to evaluate 
whether the adverse event is a result of a nutritional quality factor 
problem with the formula product. The results and evaluation of the 
infant's clinical status are essential to make this evaluation, and the 
health of the infant is also relevant to interpreting study endpoints, 
for example, growth data. Therefore, knowledge of the infant's health 
status is an essential piece of information in evaluating the 
circumstances surrounding an observed adverse event associated with use 
of a formula product. Thus, FDA has tentatively concluded that 
evaluations by a health care professional are necessary to provide the 
agency with relevant information on the circumstances surrounding the 
adverse event (see Sec. 106.121(d)(2)) to assist the agency in 
evaluating the nutritional adequacy and safety of the formula product 
for supporting healthy growth in infants. In some cases, this clinical 
assessment may be carried out by the infant's health care provider, 
rather than the investigators conducting this clinical study, because 
some parents will contact the infant's health care provider if the 
infant experiences any adverse event during the course of the study. 
The agency expects that the study investigators will take sufficient 
measures to obtain all available information to enhance the likelihood 
of being able to meaningfully interpret the likely relationship of the 
adverse event to the formula product and its impact on study 
conclusions.
    Attrition of infants from a study can result not only from adverse 
events and illnesses but also from a variety of reasons having no 
bearing on whether the new infant formula meets the quality factor 
requirements. For example, an infant enrolled in the study may be 
withdrawn from the study because the parents moved from the area. The 
effect of attrition on study results, however, must be evaluated in 
order to be able to meaningfully

[[Page 36200]]

interpret those results. To properly evaluate the impact of attrition 
on study results, FDA must have information that permits it to evaluate 
the likely cause of the attrition and its relationship to product use 
and study measurements. Therefore, the agency is proposing to require 
the submission of this information on attrition under 
Sec. 106.121(d)(3).
    Proposed Sec. 106.121(e) requires that the manufacturer submit the 
results of the Protein Efficiency Ratio. This proposed submission 
requirement is necessary to provide assurance that the manufacturer has 
complied with proposed Sec. 106.97(b) and to provide assurance that the 
infant formula meets the specific quality factor for protein quality.
    Under proposed Sec. 106.121(f), the manufacturer is required to 
submit a statement certifying that it has collected and considered all 
information and data on the ability of the infant formula to meet the 
quality factor requirements, and that it is not aware of anything that 
would show that the formula does not meet the quality factors. This 
proposed requirement is necessary to provide assurance that the 
manufacturer has complied with the regulations and considered all 
information and data of which it is aware, and that it has not made a 
selective submission of information that gives a false impression of 
the degree or extent to which a formula meets the quality factor 
requirements.
5. Verification Submissions
    Proposed Sec. 106.130(a) requires that manufacturers, after the 
first production, but before the introduction into interstate commerce, 
of a new infant formula, verify in a written submission that the infant 
formula complies with the requirements of the act and is not 
adulterated. This proposed requirement implements section 412(d)(2) of 
the act, which requires the submission of a written verification that 
summarizes test results and records demonstrating that a formula meets 
the requirements of section 412(b)(1), (b)(2)(A), (b)(2)(B)(i), 
(b)(2)(B)(iii), (b)(3)(A), (b)(3)(C), and (i) of the act. The failure 
to provide the notice required by section 412(d) of the act is a 
prohibited act under section 301(s) of the act.
    Proposed Sec. 106.130(b)(1) requires that the verification 
submission include the name of the new infant formula, the filing date 
for the new infant formula submission required under proposed 
Sec. 106.120, and the identification number assigned by FDA to the new 
infant formula submission, so that FDA is able to match the 
verification submission with the appropriate new infant formula 
submission.
    Proposed Sec. 106.130(b)(2) requires that the verification 
submission include a statement that the infant formula to be introduced 
into interstate commerce is the same as that which was the subject of 
the new infant formula submission and for which the manufacturer 
provided assurances in accordance with the requirements of proposed 
Sec. 106.120. FDA has tentatively concluded that if this statement can 
be made by the manufacturer, it means that the assurances that the 
manufacturer provided in the new infant formula submission with respect 
to the quality factor requirements and the safety of the ingredients 
remain relevant and applicable to the product. Thus, no additional 
information need be included in the verification to demonstrate 
compliance, in accordance with section 412(d)(2) of the act, with 
section 412(b)(1) or with this aspect of section 412(b)(2)(A).
    Proposed Sec. 106.130(b)(3) requires a summary of test results that 
show the levels of each nutrient required by Sec. 107.100 in the 
formula and of any nutrient added by the manufacturer. This proposed 
requirement is necessary to demonstrate compliance with section 412(i) 
of the act. Section 412(i) of the act sets forth those nutrients that 
an infant formula must contain in order not to be adulterated, and the 
submission of a summary of test results as required by section 
412(d)(2), and implemented by Sec. 106.130(b)(3), is necessary to show 
that an infant formula, after the first production, contains all of the 
required nutrients at the required levels.
    FDA has tentatively concluded that it is not necessary to require 
that the verification submission summarize test results or records 
demonstrating compliance with sections 412(b)(2)(A) and (b)(2)(B)(iii) 
of the act because the underlying records will be available for 
inspection by FDA. FDA has tentatively concluded that to require the 
manufacturer to create a report based on these records would be to 
require an unnecessary expenditure of effort. However, the agency is 
proposing to require (under Sec. 106.130(b)(4)) that the manufacturer 
certify as a part of its verification submission that it has 
established procedures that comply with sections 412(b)(2)(A) and 
(b)(2)(B)(iii) of the act.
    FDA has tentatively concluded that requiring additional test 
results or records demonstrating compliance with section 
412(b)(2)(B)(i), (b)(3)(A), and (b)(3)(C) of the act would be 
unnecessary because such showings would be subsumed in the testing to 
show whether the formula meets the requirements of Sec. 107.100 (under 
Sec. 106.130(b)(3)).
    Proposed Sec. 106.130(c) makes clear the consequences of failing to 
comply with Sec. 106.130 and that in such circumstances, the agency 
will notify the submitter that the notice is not adequate, and that the 
manufacturer has not met the requirements of section 412(d)(2) of the 
act.
6. Submissions Concerning a Change in Infant Formula That May 
Adulterate the Product
    Proposed Sec. 106.140(a) provides that, when a manufacturer makes a 
change in the formulation or processing of the formula that may affect 
whether the formula is adulterated under section 412(a) of the act, it 
shall make a submission to FDA before the first processing of such 
formula. This proposed requirement implements section 412(d)(3) of the 
act, which requires that manufacturers make the submission to FDA 
required by section 412(d)(1) of the act before first processing when 
they determine that a change in formulation or in the processing of an 
infant formula may affect whether the formula is adulterated under 
section 412(a) of the act. Examples of changes that may affect whether 
a formula is adulterated under section 412(a) of the act include, but 
are not limited to:
    (1) A change in the level of an ingredient that does not constitute 
a major change but that may affect whether the formula meets the 
requirements of section 412(i) of the act (for example, decreasing the 
amount of an ingredient such as sodium chloride could affect whether 
the formula provides two nutrients required by Sec. 107.100);
    (2) A change in an ingredient in an infant formula that does not 
constitute a major change but that may affect whether the formula meets 
the quality factor requirements of subpart E of part 106 (for example, 
a change in the level of an emulsifier could result in a change in the 
bioavailability of fat because the emulsifier may interfere with fat 
digestion); or
    (3) A change in the processing of the infant formula that does not 
constitute a major change but that may affect whether the CGMP 
requirements or the quality control procedures of subparts B and C of 
part 106 are met (for example, a change in the processing of the infant 
formula may affect whether a specification or a standard for a 
particular point in the manufacturing process where control is deemed

[[Page 36201]]

necessary to prevent adulteration is met; a change in a processing 
temperature or holding time may allow microorganisms to develop in 
violation of Sec. 106.55; or a change in a processing temperature may 
affect the level of a labile (temperature sensitive) nutrient in the 
formula).
    Proposed Sec. 106.140(b)(1) requires that the submission include 
information on the name and physical form of the product, so that the 
change in the formula can be evaluated with other information that the 
agency has received on the formula, and so that an accurate evaluation 
of the product can be made because different requirements may apply to 
different forms of a formula.
    Proposed Sec. 106.140(b)(2) requires an explanation of why the 
change in formulation or processing may affect whether the formula is 
adulterated, so that the agency can determine what type of evaluation 
the submission requires. For example, if a change in formulation may 
affect nutrient levels, the agency needs to evaluate the nutrient 
content of the formula to be assured that this formulation change will 
not lead to production of a formula that will not provide a required 
nutrient at the required amount. Likewise, if a change in processing 
may affect whether the formula is adulterated, the agency will need to 
evaluate the formula's processing to be assured that the processing of 
the formula will still comply with the CGMP regulations in subpart B of 
part 106.
    Proposed Sec. 106.140(b)(3) requires that the submission comply 
with Sec. 106.120(b)(3), (b)(4), (b)(5), and (b)(6). This proposed 
requirement implements section 412(d)(3) of the act, which provides 
that manufacturers make the submission required by section 412(d)(1). 
FDA has tentatively concluded that requiring that the submission comply 
with these aspects of Sec. 106.120(b) will promote consistency in the 
form and substance of the information that industry must submit, and 
FDA must review. If the information required on processing by 
Sec. 106.120(b)(4) has already been provided in compliance with 
Sec. 106.140(b)(2) as a part of the explanation of why the change in 
processing may affect whether the formula is adulterated, the same 
information does not need to be repeated in the submission. To avoid 
redundant submissions, proposed Sec. 106.140(b)(3) further provides 
that if the information required by Sec. 106.120(b)(3), (b)(4), (b)(5), 
or (b)(6) has been provided to the agency previously, and that 
information is not affected by the change that is the subject of the 
submission, a statement to that effect, together with the 
identification number assigned by the agency to the relevant infant 
formula submission, can be provided in lieu of a new submission.
    Proposed Sec. 106.140(b)(3) requires inclusion of the 
identification number assigned by the agency to the infant formula 
submission so that the agency can have ready access to the relevant 
information that was previously submitted. For example, if the 
manufacturer makes a submission as a result of a change in processing, 
but the formulation will remain the same, the manufacturer need not 
provide the information required by Sec. 106.120(b)(3). Likewise, if 
the manufacturer makes a submission as a result of a change in 
formulation, but the processing of the formula remains the same, the 
manufacturer need not submit the information required by 
Sec. 106.120(b)(4).
    A determination of whether the assurances required by 
Sec. 106.120(b)(5) and (b)(6) need to be given is based on the 
manufacturer's reason for providing the submission. For example, if the 
submission is provided because a change in formulation or processing 
may affect whether the formula is adulterated because it does not meet 
the quality factors set forth in subpart E of part 106, the assurance 
required by Sec. 106.120(b)(5)(i) would have to be provided. Likewise, 
if the submission is provided because a change in formulation or 
processing may affect whether the formula is adulterated because it 
does not meet the nutrient requirements of Sec. 107.100, the assurance 
required by Sec. 106.120(b)(5)(ii) would have to be provided. Further, 
if the submission is provided because a change in processing may affect 
whether the formula is adulterated because the processing of such 
formula may no longer be in compliance with CGMP or with appropriate 
quality control, as set forth in subparts B and C of part 106, or 
whether the formula is manufactured in a manner designed to prevent 
adulteration, the assurance required by Sec. 106.120(b)(6) would have 
to be provided.
    In proposed Sec. 106.140(c), the agency sets forth requirements 
necessary to ensure that the data and other information provided in the 
submission are in a form that will allow FDA to complete its review in 
a timely manner and to advise the manufacturer if the agency has any 
concerns about the marketing of the formula. Proposed Sec. 106.140(c) 
also makes clear that the agency will notify the submitter if the 
notice is not adequate because it does not meet the requirements of 
section 412(d)(3) of the act.
7. Notification of an Adulterated or Misbranded Infant Formula
    If FDA adopts the other regulations that it has proposed, it will 
redesignate current Secs. 107.240(a) and (b) as Sec. 106.150 so that 
all notification requirements on infant formulas can be found in one 
place in the agency's regulations. In Sec. 106.150(b), FDA has revised 
the address to reflect the reorganization of CFSAN.

H. Conforming and Editorial Changes to Part 107--Infant Formula

    The agency is making conforming and editorial changes to part 107 
to reflect the changes made by the 1986 amendments and the regulations 
that FDA is proposing to adopt in part 106. The references in part 107 
to the Division of Regulatory Guidance are being changed to the 
Division of Enforcement to reflect the reorganization of CFSAN in 
November 1992.
1. Changes in Subpart A
    The agency is proposing to add a new Sec. 107.1 which will parallel 
proposed Sec. 106.1. Proposed Sec. 107.1 describes the authority for 
each of the proposed subparts and the consequences under the act of 
failure to comply with any of the regulations in the proposed subparts.
2. Changes in Subpart B of Part 107--Labeling
    The agency is proposing to amend Sec. 107.10 to require a statement 
of the amount, supplied by 100 kcal, of each of any nutrient added by 
the manufacturer as well as of the listed nutrients. As discussed 
previously in the quality control section of this document, infant 
formula manufacturers are adding ingredients to infant formula to 
provide nutrients, such as selenium, that are not required by 
Sec. 107.100 to be in infant formulas. The proposed change to 
Sec. 107.10 requires that the amount of the added nutrients supplied by 
100 kcal of the formula be declared on the label of the infant formula. 
This proposed requirement is necessary to inform the consumer on a 
consistent basis of the level of all nutrients included in an infant 
formula.
3. Subpart C of Part 107--Exempt Infant Formulas
    At this time the agency is not proposing to revise the regulations 
in Sec. 107.50 pertaining to infant formulas that are subject to 
section 412(h) of the act. These regulations were finalized in 1985 (50 
FR 48183), before passage of the 1986 amendments. In the near future, 
the agency intends to reevaluate

[[Page 36202]]

the exempt infant formula regulations and the effect of the 1986 
amendments on exempt infant formulas and to issue a proposed rule to 
reflect the results of this reevaluation. The agency also plans to 
evaluate the effect of the Nutrition Labeling and Education Act of 1990 
(Pub. L. 101-535) (the 1990 amendments) on the regulations for exempt 
infant formulas. Exempt infant formulas are specifically exempted from 
requirements for health claims and nutrient content claims by section 
403(r)(5)(A) of the act. The basis for being an exempt infant formula, 
according to section 412(h)(1) of the act, is how the product is 
represented and labeled for use. This category of infant formula 
recognizes that infants who suffer from special medical disorders, such 
as maldigestion and malabsorption, inborn errors of metabolism such as 
phenylketonuria or maple syrup urine disease, or severe kidney disease, 
require formulas tailored specifically to their medical needs. 
Therefore, it is important that any claims made for these products be 
truthful, not misleading, and adequately substantiated because these 
infants make up a vulnerable population and must receive the 
appropriate nutrients for their medical condition. Because these 
formulas are exempt from the regulations governing claims that were 
developed under the 1990 amendments, the agency plans to evaluate how 
claims for these products need to be substantiated to ensure that 
infants with special nutritional needs are receiving appropriate infant 
formulas.
4. Subpart E--Infant Formula Recalls
    Current Sec. 107.240(a) sets out the requirements for notification 
of a violative infant formula, and current Sec. 107.240(b) sets out the 
method of notification. As stated above, the agency is moving the 
provisions of current Sec. 107.240(a) and (b) to Sec. 106.150, so that 
all of the agency's notification requirements are in one place. The 
agency is renumbering current Sec. 107.240(c)(1), (c)(2), and (c)(3) as 
Sec. 107.240(a), (b), and (c).
    Section 107.250 gives directions on the termination of an infant 
formula recall. The agency is changing the reference to the Division of 
Regulatory Guidance to the Division of Enforcement in Sec. 107.250 to 
reflect the 1992 reorganization of CFSAN.

V. Environmental Impact

    The agency has carefully considered the potential environmental 
effects of this action. FDA has concluded that the action will not have 
a significant impact on the human environment, and that an 
environmental impact statement is not required. The agency's finding of 
no significant impact and the evidence supporting that finding, 
contained in an environmental assessment, may be seen in the Dockets 
Management Branch (address above) between 9 a.m. and 4 p.m., Monday 
through Friday.

VI. Analysis of Impacts

    FDA has examined the impacts of the proposed rule under Executive 
Order 12866 and the Regulatory Flexibility Act (Pub. L. 96-354). 
Executive Order 12866 directs agencies to assess all costs and benefits 
of available regulatory alternatives and, when regulation is necessary, 
to select regulatory approaches that maximize net benefits (including 
potential economic, environmental, public health and safety, and other 
advantages; distributive impacts; and equity). According to Executive 
Order 12866, a regulatory action is ``economically significant'' if it 
meets any one of a number of specified conditions, including having an 
annual effect on the economy of $100 million or adversely affecting in 
a material way a sector of the economy, competition, or jobs. A 
regulation is considered ``significant'' under Executive Order 12866 if 
it raises novel legal or policy issues.
    The Regulatory Flexibility Act requires Federal agencies to 
minimize the economic impact of their regulations on small businesses. 
FDA finds that this proposed rule is neither an economically 
significant nor a significant regulatory action as defined by Executive 
Order 12866. In compliance with the Regulatory Flexibility Act, the 
agency certifies that this proposed rule, if issued, will not have a 
significant impact on a substantial number of small businesses. 
Therefore, under the Regulatory Flexibility Act, no further analysis is 
required. The agency examined three options in determining the economic 
impact of this proposed regulations. A summary of the options follow:

A. Options

    FDA has three primary options: (1) Adopt regulations with more 
stringent requirements than the proposed regulations; (2) adopt the 
proposed regulations; or (3) adopt regulations with less stringent 
requirements than the proposed regulations.
1. Option 1--Adopt Regulations More Stringent Than the Proposed 
Regulations
    FDA believes infant formula manufacturers already comply with most 
of the requirements of this proposed rule. One option would be to add 
provisions to this proposed rule that would require activity beyond 
that which is currently engaged in by infant formula manufacturers or 
that is likely to be engaged in by manufacturers entering the infant 
formula industry. Potential requirements of this type include specific 
production and in- process control systems, specific equipment or types 
of personnel, and additional testing and recordkeeping.
    Under this option, incumbent manufacturers would face higher 
production costs and would pass most of the costs on to consumers of 
infant formula. In addition, the startup and operating costs would 
increase, and thus discourage entry into the infant formula industry. 
The ability of new firms to enter an industry is an important element 
in promoting price competition and innovation. These additional 
requirements would reduce price competition in the infant formula 
industry.
    The price of infant formula is probably linked to certain risky 
infant feeding practices. With very high infant formula prices, some 
consumers may increase risks to infants by improperly diluting formula 
with water or other substances; using inappropriate substitutes for 
formula or breast milk; or prematurely switching from formula to cow's 
milk. For example, preliminary results of an FDA study on infant 
formula feeding practices showed that approximately 20 percent of 
infants (younger than 2 months) had their formula diluted by cereal, 
which is cheaper than infant formula.
2. Option 2--Adopt the Proposed Regulations
    There are two types of costs associated with this option: precluded 
future cost cutting behavior and direct compliance costs.
    a. Future cost cutting behavior. This type of cost may arise 
because the proposed rule precludes cost cutting behavior by either 
incumbent firms or firms entering the infant formula industry. Infant 
formula manufacturers currently undertake a considerable amount of 
activity, such as infant growth studies, that is designed to ensure the 
safety of infant formula but is not explicitly required by either 
current law or regulation. In the absence of this regulation, which 
mandates this activity, either incumbent or future manufacturers may 
choose not to undertake this activity in the future. However, because 
of reputation effects and liability laws, these costs are likely to be 
low.

[[Page 36203]]

    b. Direct compliance costs. (i). CGMP. FDA believes that infant 
formula manufacturers already comply with most of the proposed CGMP's. 
These CGMP's include those dealing with: (1) Production and in-process 
control systems, including the evaluation of any deviation from these 
procedures or from established standards or specifications; (2) 
controls designed to prevent adulteration of infant formula by workers, 
by facilities, and during packaging and labeling; (3) controls to 
prevent adulteration during manufacturing, including recording and 
justifying deviations from the master manufacturing order and 
evaluating deviations from processing times; (4) controls on the 
release and storage of finished infant formula; (5) all requirements 
relating to batch production and control records, and to coding; and 
(6) all requirements dealing with general quality control procedures, 
including the testing of one batch of each physical form of infant 
formula at least once every 3 months.
    If all manufacturers already comply with these proposed CGMP's, 
then no compliance costs will result from them. FDA requests comments 
on whether all infant formula manufacturers are already in compliance 
with the proposed CGMP's listed above.
    FDA believes that all infant formula manufacturers already comply 
with the proposed CGMP's dealing with controls to prevent adulteration 
caused by ingredients, containers, and closures. The provision that FDA 
may object to the use of a particular substance in an infant formula 
during its prenotification review of ingredients used in a formula 
because it believes that the substance is not safe and suitable for 
that use does not represent a change in the way FDA reviews infant 
formula ingredients. This provision recognizes the fact that 
manufacturers may make independent GRAS determinations about 
ingredients. When a manufacturer makes such a determination, that 
manufacturer is not necessarily required to have the relevant 
ingredient affirmed as GRAS by FDA. However, FDA is reserving the right 
to review infant formula ingredient lists and documentation concerning 
whether particular ingredients are safe and suitable for use in infant 
formula. Theoretically, this provision could lead to a reduction in the 
number of ingredients that are independently determined to be GRAS and 
a corresponding increase in the number of ingredients for which food 
additive petitions are required. Petitions for direct food additives 
can take between 1 to 6 years to complete and cost approximately $1 
million per year. However, because manufacturers of infant formula 
generally obtain FDA concurrence on the safety and suitability of 
ingredients used in infant formula before making these determinations, 
FDA believes no additional compliance costs will be generated by this 
provision.
    FDA also believes that infant formula manufacturers already comply 
with many of the other proposed CGMP's. Provisions of CGMP's that some 
infant formula manufacturers may not currently be in compliance include 
the following:
    (1) Controls to prevent adulteration caused by equipment or 
utensils. Some manufacturers may not repair or replace instruments and 
controls when those instruments and controls cannot be adjusted to 
within essential agreement with the reference standard. In addition, 
most manufacturers probably do not perform a written evaluation of all 
affected product, or of actions taken when calibration results indicate 
that a specification or standard for a point where control is deemed 
necessary to prevent adulteration has not been met. FDA cannot estimate 
the repair or replacement costs of instruments and controls at this 
time. Written evaluations will take a supervising technician an 
estimated 2 hours to complete, which will generate some small 
compliance costs.
    (2) Controls to prevent adulteration because of automatic, 
mechanical, and electronic equipment. Most manufacturers will probably 
have to perform additional analysis of software modifications. FDA 
preliminarily estimates this analysis will add approximately 1 month to 
the time required to analyze programming and software modifications. 
One or two software modifications are probably made each year at each 
of the fifteen plants that produce infant formula. Assuming that a 
single computer scientist works on the additional activity required, 
compliance costs are estimated to be about $100,000 per year.
    ii. Audits, Quality factors, registration and notification 
requirements, and infant formula recalls. FDA believes that infant 
formula manufacturers already comply with the following provisions: (1) 
Regularly scheduled audits to determine compliance with CGMP's and 
Quality Control Procedures (QCP's), (2) growth and development studies 
to be submitted under certain conditions and new notification 
requirements (FDA already requests and receives these quality factor 
growth and development studies and notification material based on FDA's 
interpretation of the language of the 1986 amendments), and (3) all 
provisions involving registration and notification requirements.
    If infant formula manufacturers are already complying with these 
provisions, then no compliance costs will be generated by these 
provisions.
    FDA requests information on whether all infant formula 
manufacturers already comply with all provisions listed above, 
particularly those provisions dealing with quality factors.
    iii. Records. Under the current proposal, the records produced and 
maintained by infant formula manufacturers to establish compliance with 
FDA regulations will have to be expanded to include all new CGMP's and 
QCP's. FDA believes most of the specified records are already being 
kept by all firms; however, some records may not be. A plausible 
assumption is that current annual industry expenditures on 
recordkeeping may increase by about 10 percent, or $450,000 per year 
based on information received from industry on current recordkeeping 
costs. FDA requests information on the cost of increased recordkeeping.
    iv. Administrative costs. Interpreting and implementing changes in 
CGMP and QCP regulations generate administrative costs even when all 
activity required in those CGMP's and QCP's is already being done. FDA 
does not have information on the administrative costs involved in 
interpreting and implementing changes in CGMP and QCP regulations; 
however, it is plausible to suppose that 20 percent of the total 
compliance costs other than administrative costs may be used to reflect 
administrative costs.
    Administrative costs under this assumption would be approximately 
$100,000 and would accrue in the first year only. FDA requests 
information on administrative costs.
3. Option 3--Adopt Regulations Less Stringent Than the Proposed 
Regulations
    Another option is to limit the activity required by this proposed 
rule to activity already engaged in by all incumbent infant formula 
manufacturers. In this case, there would be no compliance costs based 
on current behavior. However, in the absence of this proposed rule, 
incumbent or new manufacturers might choose not to undertake all 
activity specified in this proposed rule. Therefore, the only costs 
associated with this option are the costs associated with precluded 
potential future behavior on the part of incumbent or new 
manufacturers.

[[Page 36204]]

B. Benefits

1. Option 1--Adopt Regulations More Stringent Than the Proposed 
Regulations
    More stringent regulations for infant formula would cause infant 
formula manufacturers to undertake further activity to ensure the 
safety of infant formula. If there were identifiable risks from infant 
formula that were not addressed by this proposal, then this additional 
activity might decrease those health risks. However, FDA is not aware 
of identifiable health risks from infant formula that are not addressed 
by this proposal.
2. Option 2--Adopt the Proposed Regulations
    The proposed regulation has two primary benefits: A potential 
direct reduction in the health risks posed by infant formula, and a 
potential reduction in the cost of entering the infant formula 
industry. The latter effect could lead to an increase in the 
competitiveness of the infant formula industry, resulting in lower 
infant formula prices and a reduction in the incidence of risky infant 
feeding practices linked to high infant formula prices.
    One example of a current activity that can be linked to a direct 
reduction in health risks but that is not explicitly required by 
current law or regulation is the performance of growth studies for new 
infant formulas. FDA currently requests and receives these studies to 
demonstrate that the infant formula meets the quality factor 
requirements of section 412(b)(1) of the act. However, because section 
412(b)(1) of the act does not list specific quality factors that infant 
formulas must meet, a quality factor for healthy growth currently is 
not expressly stipulated. In the absence of this proposed rule, 
manufacturers could decline to perform these growth studies in the 
future with a potential consequence that products that do not support 
normal growth would be marketed. Low growth rates would not be detected 
by existing regulatory and legal requirements that measure only the 
levels of required nutrients because the required nutrients may be 
present but not be bioavailable, and there is no mechanism for testing 
bioavailability other than the proposed studies.
    An example of a formula associated with low growth rates that would 
not have been detected in the absence of growth studies was an 
experimental formula that contained a source of fatty acids not 
previously used in infant formula. Because only a small amount of the 
new fat source was added to a commercial formula, it is reasonable to 
assume that all required nutrients were present within legal 
specifications. Consequently, it would likely have met all current 
regulations. Nonetheless, this formula was found to result in low 
infant growth rates (Ref. 87). In this case, the manufacturer undertook 
the necessary growth studies and detected the problem on its own. 
However, manufacturers might not undertake these studies on their own 
in the future. In addition, even if manufacturers continue to undertake 
these studies in the absence of this regulation, they may not do these 
studies correctly.
    In general, low rates of infant growth are associated with higher 
than normal levels of infant morbidity. If a problem of this type were 
to occur, a large number of infants could potentially be affected.
    Other types of current activity can also be linked to a direct 
reduction in health risks and also are not explicitly required by 
current law or regulation. In the absence of this regulation, incumbent 
or new manufacturers may not undertake this activity in the future. 
However, as explained earlier, because of reputation effects and legal 
liability, such a refusal seems unlikely.
    An example of a health risk from infant formula is the 1978 
incident, discussed elsewhere in this document, in which a required 
nutrient was missing from an infant formula. Recurrence of this 
particular problem is unlikely because section 412(d)(1)(A) of the act 
already explicitly requires the submission of the quantitative 
formulation of an infant formula as part of the mandatory FDA 
notification of a new infant formula. Recurrence of this problem is 
also made unlikely because section 412(b)(2) of the act already 
explicitly requires the testing of infant formula for all required 
nutrients. However, the risk of a formula being sold without a required 
nutrient is minimized to the extent possible by specifically clarifying 
this part of the infant formula law in the regulation.
    Another example of a health risk associated with infant formula is 
an incident in which infant formula was found to contain Salmonella. It 
appears that the manufacturer was testing for Salmonella in a manner 
consistent with the testing requirements of this proposed rule, and 
therefore it is not clear that this particular incident would have been 
avoided if the proposed rule had been in effect. This proposed rule 
will reduce the risk of microbiological contamination, however, because 
it requires manufactures to institute a production and in-process 
control system. The production and in-process control system 
establishes standards or specifications to be met throughout the 
production of their product. Other provisions of the proposed 
regulation that will also help to prevent microbiological contamination 
of infant formulas are controls to prevent adulteration by workers 
(proposed Sec. 106.10), controls on the required temperature of cold 
storage compartments used for storing ingredients and uncanned infant 
formula (proposed Sec. 106.30(e)(2)), controls on the monitoring of the 
temperature of both cold storage and thermal processing equipment 
(proposed Sec. 106.30(e)), controls on the spray-drying process for 
powdered infant formula including the filtering of the intake air 
before heating to prevent microbial growth (proposed 
Sec. 106.50(d)(2)), and controls to ensure that each container of 
finished product is properly sealed (proposed Sec. 106.50(d)(4)).
    The incident in which infant formula was found to contain 
Salmonella resulted in two reported cases of salmonellosis in infants. 
The average value of preventing a single case of salmonellosis is 
estimated to be about $2,000 (Ref. 88). If an incident like this is 
avoided in the future because of this proposed rule, the value of the 
adverse health effects avoided would be a benefit of this proposed 
rule.
    This incident also resulted in two recalls. FDA estimates a 
combined cost, including costs that accrued to both the manufacturer 
and FDA of approximately $0.7 million per recall. If an incident like 
this is avoided in the future because of this proposed rule, the recall 
costs that would otherwise have been associated with this incident 
would also be a benefit of this proposed rule.
    Another benefit of the proposed regulations is a potential 
reduction in the administrative and time costs of entering the infant 
formula industry. Currently, infant formula manufacturers must analyze 
and interpret the relevant laws to determine the legal requirements 
involved in the manufacture of infant formula. Incumbent firms have 
tended to accept FDA's interpretations of these laws and have received 
information on this interpretation incrementally over time, chiefly 
through direct contact with FDA on various issues.
    It is reasonable to expect that potential entrants into the infant 
formula industry would also prefer to rely on FDA's interpretations of 
the relevant laws. However, considerable time and administrative costs 
are involved in obtaining this information because there is no 
established

[[Page 36205]]

mechanism by which manufacturers can obtain this information other than 
direct communication with FDA on various particular issues. By 
providing an explicit specification of the activities that are required 
by the relevant laws, the proposed regulations, if adopted, will reduce 
the time and administrative costs involved in entering this industry.
    In order to determine the net effect of the proposed rule on the 
cost of entering the infant formula industry, the reduction in time and 
administrative costs must be weighed against the additional compliance 
costs imposed by this proposed rule on new firms. These countervailing 
compliance costs are probably low because new firms will probably 
undertake voluntarily the same activity that is currently undertaken 
voluntarily by incumbent manufacturers. Therefore, the net effect of 
this proposed rule is likely to be the reduction in the cost of 
entering the infant formula industry. Publication of the proposed and 
final regulations will provide a means of expedited entry for new firms 
into the infant formula market.
    A reduction in the cost of entering the infant formula industry 
will promote both price competition and innovation in this industry. 
Increased price competition may lead to health benefits because, as 
stated above, high infant formula prices may encourage some consumers 
to: (1) Improperly dilute infant formula to reduce the cost per 
serving; (2) prematurely switch from infant formula to cow's milk; or 
(3) use inappropriate substitutes for breast milk and infant formula.
    A final benefit of this proposed rule is the cost savings generated 
by the elimination of the current FDA requirement that a vitamin D rat 
bioassay be performed for all major changes in infant formula. In 1992, 
there were approximately 50 major changes. The cost of a rat bioassay 
for vitamin D for infant formula at a private lab is about $1,070 (Ref. 
89). Infant formula manufacturers should therefore save approximately 
$54,000 in testing costs per year.
3. Option 3--Adopt Regulations Less Stringent than the Proposed 
Regulations
    Except for the value of the risk reductions resulting from 
requirements that go beyond activity currently undertaken by infant 
formula manufacturers the benefits of this option are identical to 
those of Option 2.

C. Conclusions

    In accordance with Executive Order 12286, FDA has analyzed the 
economic effects of this proposed rule and has determined that this 
rule, if issued, will not be a significant rule as defined by that 
order. In accordance with the Regulatory Flexibility Act, FDA certifies 
that the proposed rule will not have a significant impact on a 
substantial number of small businesses.
    The primary compliance costs of Option 2 include both direct costs 
of new requirements and precluded production cost reductions which may 
occur without this regulation. FDA has estimated direct costs to 
incumbent manufacturers to be approximately $0.7 million in the first 
year and $0.6 million each additional year. An additional cost to 
incumbent manufacturers is the cost of repairing or replacing 
instruments and controls when those instruments and controls cannot be 
adjusted to agreement with the reference standard. FDA has insufficient 
information to estimate this cost. FDA does not expect compliance with 
the proposed regulations to cause any significant increase in the price 
of infant formula products. However, the agency requests comments about 
any potential effects of the proposed regulations on the price of 
infant formula products.
    The primary benefit of Option 2 is the reduction in the risk that 
defective infant formula will be produced, go undetected, and reach the 
market. FDA has insufficient information to estimate this potential 
benefit. In addition, this proposed rule is also expected to reduce the 
time and administrative costs of entering the infant formula industry. 
This benefit may increase price competition in the infant formula 
industry and reduce the health risks associated with high infant 
formula prices. FDA also has insufficient information to estimate these 
benefits.
    Except for the costs and benefits associated with activity required 
by this proposed rule that some incumbent manufacturers do not 
currently undertake, the costs and benefits of Option 3 are identical 
to those of Option 2. FDA has insufficient information to estimate 
either the costs or benefits of this option.
    Option 1 is expected to have higher costs and lower benefits than 
either Option 2 or Option 3.

VII. Paperwork Reduction Act of 1995

    This proposed rule contains information collection provisions that 
are subject to review by the Office of Management and Budget (OMB) 
under the Paperwork Reduction Act of 1995 (44 U.S.C. 3501-3520). The 
title, description, and respondent description for the proposed 
collection of information are shown below, along with an estimate of 
the annual recordkeeping and periodic reporting burden. Included in the 
estimate is the time for reviewing instructions, searching existing 
data sources, gathering and maintaining the necessary information, and 
completing and submitting the registrations, notifications, and other 
submissions that would be required under the proposed regulations.
    FDA solicits public comment in order to: (1) Evaluate whether the 
proposed collection of information is necessary for the proper 
performance of the functions of the agency, including whether the 
information will have practical utility; (2) evaluate the accuracy of 
the agency's estimate of the burden of the proposed collection of 
information, including the validity of the methodology and assumptions 
used; (3) enhance the quality, utility, and clarity of the information 
to be collected; and (4) minimize the burden of the collection of 
information on those who are to respond, including through the use of 
automated collection techniques, where appropriate or other forms of 
information technology.
    Title: Current Good Manufacturing Practice, Quality Control 
Procedures, Quality Factors, Notification Requirements, and Records and 
Reports, for the Production of Infant Formula.
    Description: FDA is proposing regulations on recordkeeping 
requirements that include: (1) Records pertaining to batch production 
and control; (2) records pertaining to current good manufacturing 
practice and quality control; (3) records pertaining to distribution of 
the infant formula; and (4) records pertaining to regularly scheduled 
audits. FDA is also proposing regulations on reporting requirements 
pertaining to: (1) Registration of a new infant formula; (2) submission 
requirements for a new infant formula; (3) submission requirements to 
provide assurance that an infant formula meets the quality factor 
requirements; (4) submission requirements when there is a change in the 
formulation or processing of the formula that may affect whether the 
formula is adulterated; and (5) submission requirements to provide 
assurance that the infant formula complies with the requirements of the 
Federal Food, Drug, and Cosmetic Act and is not adulterated.
    Description of Respondents: Infant Formula Manufacturers.

[[Page 36206]]



                                      Estimated Annual Recordkeeping Burden                                     
----------------------------------------------------------------------------------------------------------------
                                                              Annual        Total                               
                  21 CFR                       No. of      frequency of     annual      Hours per    Total hours
                                           recordkeepers  recordkeeping    records    recordkeeping             
----------------------------------------------------------------------------------------------------------------
106.6....................................             5              1             5           200         1,000
106.20(f)(4) and 106.100(f)(1)...........             5             52           260             3           780
106.30(d) and 106.100(f)(2)..............             5             25           125             4           500
106.30(e)(3)(ii) and 106.100(f)(3).......             5            365         1,825             2         3,650
106.30(f) and 106.100(f)(4)..............             5            365         1,825             3         5,475
106.35(c) and 106.100(f)(5)..............             5              2            10           500         5,000
106.40(d)................................             5             20           100            30         3,000
106.40(g) and 106.100(f)(6)..............             5            122           610             4         2,440
106.50...................................             5              1             5           200         1,000
106.55(d) 106.100(e)(5)(ii), and                                                                                
 106.100(f)(7)...........................             5            182           910             3         2,730
106.60(c)................................             5              1             5            40           200
106.91(c), 106.100(e)(5)(i), and                                                                                
 106.100(f)(7)...........................             5            365         1,825             4         7,300
106.94...................................             5              1             5            88           440
106.97...................................             5            0.6             3           225           675
106.100(e)...............................             5            365         1,825             9        16,425
                                                                                                    ------------
    Total................................  .............  .............  ...........  .............       50,615
----------------------------------------------------------------------------------------------------------------


                                        Estimated Annual Reporting Burden                                       
----------------------------------------------------------------------------------------------------------------
                                                                 Annual                                         
                                                    No of      frequency      Total      Hours per              
                     21 CFR                      respondents      per         annual      response   Total hours
                                                                response    responses                           
----------------------------------------------------------------------------------------------------------------
106.110........................................            3           NA           20            1           20
106.120........................................            3           NA           20           49          980
106.121........................................            3           NA           10           50          500
106.130........................................            3           NA           20            2           40
106.140........................................            3           NA           25         5-10      125-250
                                                                                                    ------------
    Total......................................  ...........  ...........  ...........  ...........        1,790
                                                                                                    ============
    Total Recordkeeping and Reporting Burden...       52,405                                                    
----------------------------------------------------------------------------------------------------------------

    FDA tentatively concludes that there are no capital costs or 
operating and maintenance costs associated with the reporting and 
recordkeeping provisions of this proposed rule. However, the agency 
welcomes comments on any such anticipated costs.
    As required by section 3507(d) of the Paperwork Reduction Act of 
1995, FDA has submitted a copy of this proposed rule to OMB for its 
review of the information collection requirements. Other organizations 
and individuals interested in submitting comments regarding this burden 
estimate or any aspect of these information collection requirements, 
including suggestions for reducing the burden, should direct them to 
the Office of Information and Regulatory Affairs, OMB, New Executive 
Office Bldg., 725 17th St. NW., rm. 10235, Washington, DC 20503, ATTN: 
Desk Officer for FDA. Written comments on the information collection 
should be submitted by August 8, 1996.
VIII. Requests for Comments
    Interested persons may, on or before October 7, 1996, submit to the 
Dockets Management Branch (address above) written comments regarding 
this proposal. Two copies of any comments are to be submitted, except 
that individuals may submit one copy. Comments are to be identified 
with the docket number found in brackets in the heading of this 
document. Received comments may be seen in the office above between 9 
a.m. and 4 p.m., Monday through Friday.
IX. References
    The following references have been placed on display in the Dockets 
Management Branch (address above) and may be seen by interested persons 
between 9 a.m. and 4 p.m., Monday through Friday.
1. Congressional Record--Senate S 14042-14047, September 27, 1986.
2. Hamill, P. V. V., T. A. Drizd, C. L. Johnson, R. B. Reed, A. F. 
Roche, W. M. Moore, ``Physical Growth: National Center for Health 
Statistics Percentiles,'' American Journal of Clinical Nutrition, 
32:607-629, 1979.
3. Chumlea, W. C. ``Growth and Development'' Chapter 1 in Handbook 
of Pediatric Nutrition, edited by Queen, P. M., and C. E. Lang, 
ASPEN Publication Gaithersburg, MD, 1993.
4. Fomon, S. J., and S. E. Nelson, Chapter 4, Size and growth in 
Nutrition of Normal Infants, edited by Fomon, S. J., Mosby-Year 
Book, Inc., St. Louis, MO, 1993.
5. House of Representatives, House Report 96-936 ``Infant Formula 
Act of 1980 (H.R. 6940),'' May 12, 1980.
6. Committee on Nutrition, American Academy of Pediatrics, 
``Clinical Testing of Infant Formulas With Respect to Nutritional 
Suitability for Term Infants,'' Elk Grove Village, IL, June 1988 
(under contract to FDA, Contract No. 223-86-2117, June 1988).
7. Congressional Record, vol. 126, House H11787-11793, May 20, 1980.
8. FDA, Guidelines Concerning Notification and Testing of Infant 
Formulas, 1987.
9. Infant Formula Council, ``Recommended Current Good Manufacturing 
Practice for Infant Formula,'' July 1992.
10. FDA, DHHS, ``FDA Fact Sheet: Shigella in Food,'' December 1969.
11. Bryan, F. L., ``Emerging Foodborne Diseases: II. Factors that 
Contribute to Outbreaks and Their Control,'' Journal of Milk and 
Food Technology, 35(11):632-638, November 1972.
12. Bryan, F. L., ``Emerging Foodborne Diseases: I. Their 
Surveillance and Epidemiology,'' Journal of Milk and Food 
Technology, 35(10):618-625, October 1972.
13. Bryan, F. L., ``What the Sanitarian Should Know About 
Staphylococci and Salmonellae in Non-Dairy Products: I. 
Staphylococci,'' Journal of Milk and Food Technology, 31(4); 110-
116, April 1968.
14. William F. Hooten, Warning Letter, November 9, 1993.

[[Page 36207]]

15. Placencia, A. M., and G. S. Oxborrow, ``Use of the Reuter 
Centrifugal Air Sampler in Good Manufacturing Practices 
Inspections,'' U.S. Food and Drug Administration, Sterility Research 
Center, Minneapolis Center of Microbiological Investigations, 
Minneapolis, MN, December 1984.
16. Committee on Biologic Effects of Atmospheric Pollutants, 
Division of Medical Sciences, National Research Council, Chapter 9 
``Effects of Fluoride on Human Health'' in Fluorides. National 
Academy of Sciences, Washington, DC, pp. 163-221, 1971.
17. Committee on Nutrition of American Academy of Pediatrics, 
Fluoride Supplementation: Revised Dosage Schedule, Pediatrics, 
63:150-152, 1979.
18. Barness, L. A., ``Fluoride in Infant Formulas and Fluoride 
Supplementation,'' Letter to the Editor, Pediatrics, 67:582-583, 
1981.
19. Subcommittee on the 10th Edition of the Recommended Dietary 
Allowances, Food and Nutrition Board, Commission on Life Sciences, 
National Research Council. ``Recommended Dietary Allowances, 10th 
ed.,'' Washington, DC, National Academy Press, 1989.
20. Cross Connection Control Committee, Pacific Northwest Section, 
American Water Works Associations, With the Assistance of EPA, 
``Cross Connection Control Manual Accepted Procedures and 
Practice,'' 5th ed., May 1990.
21. Food Service Sanitation Manual including A Model Food Service 
Sanitation Ordinance, U.S. Department of Health, Education, and 
Welfare, Washington, DC 20204.
22. Lentsch, S., ``Sanitizers for an Effective Cleaning Program,'' 
Klenzade Division, Economics Laboratories, St. Paul, MN.
23. Moody, M. W., ``How Cleaning Compounds Do the Job,'' Seafood 
Technology Cooperative Extension Service, Louisiana State 
University, undated.
24. Refrigerated Foods and Microbiological Criteria Committee, 
National Food Processors Association, ``Factors to be Considered in 
Establishing Good Manufacturing Practices for the Production of 
Refrigerated Foods,'' Dairy and Food Sanitation, 8(6):288-291, June 
1988.
25. Center for Drugs and Biologics and Office of Regulatory Affairs, 
Food and Drug Administration ``Guideline on Sterile Drug Products 
Produced by Aseptic Processing'' Division of Manufacturing and 
Product Quality, Office of Compliance, Center for Drugs and 
Biologics, Food and Drug Administration, Rockville, MD 20857, June 
1987.
26. The American National Standards Institute, International 
Standard, Quality Management and Quality Assurance Standards--Part 3 
Guidelines for the Application of ISO 9001 to the Development, 
Supply and Maintenance of Software, pp. 1-2.
27. Institute for Electrical and Electronics Engineers, Inc. (IEEE), 
IEEE Standard Glossary of Software Engineering Terminology IEEE Std 
610.12--1990.
28. Glossary of Computerized Systems and Software Development 
Terminology. Division of Field Investigations, Office of Regional 
Operations, Office of Regulatory Affairs, U.S. Food and Drug 
Administration, August 1995.
29. Tetzlaff, R. F., ``GMP Documentation Requirements for Automated 
Systems: Part I,'' Pharmaceutical Technology, March 1992.
30. Tetzlaff, R. F., ``GMP Documentation Requirements for Automated 
Systems: Part II,'' Pharmaceutical Technology, April 1992. 31. 
Tetzlaff, R. F., ``GMP Documentation Requirements for Automated 
Systems: Part III,'' Pharmaceutical Technology, May 1992.
32. Letters from the Infant Formula Council on the Lead Level in 
Infant Formulas.
33. Rowe, B., D. H. Hutchinson, R. J. Gilbert, B. H. Hales, N. T. 
Begg, H. C. Dawkins, M. Jacob, F. A. Rae, and M. Jepson, 
``Salmonella ealing Infections Associated with the Consumption of 
Infant Dried Milk,'' The Lancet, 2(8564):900-903, 1987.
34. Poelma, P. L., W. H. Andrews, J. H. Silliker, ``Salmonella'' in 
Compendium of Methods for the Microbiological Examination of Foods, 
2d ed., edited by Marvim L. Speck, American Public Health 
Association, Washington, DC, 1984.
35. D'Aoust, J. Y., Salmonella, Chapter 9 in Foodborne Bacterial 
Pathogens, edited by Doyle, M. P., Marcel Dekker, Inc., New York, 
NY, 1989.
36. Quie, P. G. ``Antimicrobial Defenses in the Neonate,'' Seminars 
in Perinatology, 14:2-9, 1990.
37. Farber, J. M., P. I. Peterkin, ``Listeria Monocytogenes, a Food-
borne Patogen,'' Microbiological Reviews, 55:476-511, 1991.
38. National Advisory Committee on Microbiological Criteria for 
Foods, ``Listeria monocytogenes,'' International Journal of Food 
Microbiology, 14:185-246, 1991.
39. Benenson, A. S., ed., ``Control of Communicable Diseases in 
Man,'' 15th ed., American Public Health Association, pp. 170-177, 
1991.
40. Bennett, R. W., and S. M. Harmon, ``Bacillus cereus Food 
Poisoning'' in Laboratory Diagnosis of Infectious Diseases: 
Principles and Practice; vol. 1, Bacterial, Mycotic, and Parasitic 
Diseases; edited by Balows, A., W. J. Hausler, Jr., M. Ohashi, and 
A. Turano, vol. 1, pp. 83-93, Springer-Verlag, New York, 1988.
41. The International Commission on Microbiological Specifications 
for Foods (ICMSF) of the International Association of 
Microbiological Societies, Chapter 12: Sampling Plans for Milk and 
Milk Products in Microorganisms in Foods 2. Sampling for 
Microbiological Analysis: Principles and Specific Applications, 
University of Toronto Press, Toronto, Canada; 1974.
42. Schwab, A. H., A. Swartzentruber, B. A. Wentz, R. B. Read, Jr., 
``Microbiological Quality of Dry-milk Mixes and Milk Substitute 
Infant Formulas,'' Applied and Environmental Microbiology, 43:389-
391, 1982.
43. Collins-Thompson, D. L., K. F. Weiss, G. W. Riedel, and S. 
Charbonneau, ``Microbiological Guidelines and Sampling Plans for 
Dried Infant Cereals and Powdered Infant Formula from a Canadian 
National Microbiological Survey,'' Journal of Food Protection, 
43(8):613-616, 1980.
44. FDA Enforcement Report, July 14, 1993.
45. Jackson, G. J., C. F. Langford, and D. L. Archer, ``Control of 
Salmonellosis and Similar Foodborne Infections,'' Food Control, 
2(1), pp. 26-34, 1991.
46. Food and Drug Administration Bacteriological Analytical Manual, 
7th ed., AOAC International, Gaithersburg, MD, 1992.
47. Flowers, R. S., W. Andrews, C. W. Donnelly, and E. Koenig, 
``Pathogens in Milk and Milk Products'' in Standard Methods for the 
Examination of Dairy Products, 16th ed., edited by Marshall, R. T., 
American Public Health Association, Washington, DC 20005, 1992.
48. Harmon, S. M., D. A. Kautter, ``Incidence and Growth Potential 
of Bacillus cereus in Ready-to-Serve Foods,'' Journal of Food 
Protection, 54(5):372-374, 1991.
49. Golden, D. A., ``Microbiological Evaluation of Powdered Infant 
Formula,'' Report submitted to the U.S. Food and Drug 
Administration, Center for Food Safety and Applied Nutrition.
50. FDA Consumer, Defendants: Watson Foods Co., Inc., Civil No. CV-
85-4659, p. 40, October 1988.
51. Memorandum of Records from Nicholas Dvy, FDA, dated June 25, 
1996.
52. Establishment Inspection Report, 1992.
53. Memo from Alan J. Sheppard to Carolyn W. Miles, on Status of 
Vitamin D Rat Bioassay, May 12, 1994.
54. Tanner, J. T., S. A. Barnett, and M. K. Mountford, ``Analysis of 
Milk-based Infant Formula,'' Phase IV. Iodine, Linoleic Acid, and 
Vitamins D and K: U.S. Food and Drug Administration-Infant Formula 
Council: Collaborative Study, Journal of the Association of Official 
Analytical Chemists International, 76:1042-1056, 1993.
55. Southgate, D. A. T., ``Conceptual Issues Concerning the 
Assessment of Nutrient Bioavailability,'' in Nutrient Availability: 
Chemical and Biological Aspects, edited by Southgate, D., I. 
Johnson, and G. R. Fenwick, Royal Society of Chemistry, 1989.
56. Jensen, R. G., M. M. Hagerty, and K.E. McMahon, ``Lipids of 
Human Milk and Infant Formulas: A Review,'' American Journal of 
Clinical Nutrition, 31:990-1016, 1978.
57. Jenson, R. G., and G. L. Jensen, ``Specialty Lipids for Infant 
Nutrition, I. Milks and Formulas,'' Journal of Pediatric 
Gastroenterology and Nutrition, 15:232-245, 1992.

[[Page 36208]]

58. Ziegler, E. E., S. J. Fomon, S. E. Nelson et al., ``Cow Milk 
Feeding in Infancy: Further Observations on Blood Loss From the 
Gastrointestinal Tract,'' Journal of Pediatrics, 116:11-18, 1990.
59. American Academy of Pediatrics, ``Commentary on Breast-Feeding 
and Infant Formulas, with Proposed Standards for Formulas,'' 
Pediatrics, 57:278-285, 1976.
60. Committee on Labor and Human Resources Report (Senate) August 
26, 1980.
61. Saarinen, U. M., M. A. Siimes, P. R. Dallman, ``Iron Absorption 
in Infants: High Bioavailability of Breast Milk Iron as Indicated by 
the Extrinsic Tag Method of Iron Absorption and by the Concentration 
of Serum Ferritin,'' Journal of Pediatrics, 91:36-39, 1977.
62. Hallberg, L., L. Rossander-Hulten, M. Brune, A. Gleerup, 
``Bioavailability in Man of Iron in Human Milk and Cow's Milk in 
Relation to Their Calcium Contents,'' Pediatric Research, 31:524-
527, 1992.
63. Derman, D. P., D. Ballot, T. H. Bothwell, B. J. MacFarlane, R. 
D. Baynes, A. P. MacPhail, M. Gillooly, J. E. Bothwell, and W. R. 
Bezwoda, ``Factors Influencing the Absorption of Iron From Soya-Bean 
Protein Products,'' British Journal of Nutrition, 57:345-353. 1987.
64. Koo, W. W. K., and R. G. Tsang, ``Calcium, Magnesium, 
Phosphorus, and Vitamin D,'' In Nutrition Needs of the Preterm 
Infant, edited by Tsang, R. C., et al., Academic Press, 1993
65. Steichen, J. J., T. L. Gratton, and R. C. Tsang, ``Osteopenia of 
Prematurity: The Cause and Possible Treatment,'' Journal of 
Pediatrics, 96:528-34, 1980.
66. Koletzko, B., R. Tangerman, R. Von Kries, et al. ``Intestinal 
Milk-Bolus Obstruction in Formula-Fed Premature Infants Given High 
Doses of Calcium,'' Journal of Pediatric Gastroenterology and 
Nutrition 7:548-553, 1988.
67. Ziegler, E. E., and S. J. Fomon, ``Lactose Enhances Mineral 
Absorption in Infancy,'' Journal of Pediatric Gastroenterology and 
Nutrition, 2:288-294, 1983.
68. Moya, M., E. Cortes, M. I. Ballester et al., ``Short-term 
Polycose Substitution for Lactose Reduces Calcium Absorption in 
Healthy Term Babies,'' Journal of Pediatric Gastroenterology and 
Nutrition, 14:57-61, 1992.
69. Shenai, J. P., B. M. Jhaveri, J. W. Reynolds et al., 
``Nutritional Balance Studies in Very Low Birth Weight Infants: Role 
of Soy Formula,'' Pediatrics, 67:631-637, 1981.
70. Lindroth, M., U. Westgren, and S. Laurin, ``Rickets in Very Low 
Birth Weight Infants; Influence of Supplementation with Vitamin D, 
Phosphorus, and Calcium,'' Acta Pediatrica Scandinavica, 75:927-931, 
1986.
71. Chan, G. M., and L. J. Mileur, ``Posthospitalization Growth and 
Bone Mineral Status of Normal Preterm Infants,'' American Journal of 
Diseases in Childhood, 139:896-8, 1985.
72. Food and Drug Administration, Public Health Service, U.S. 
Department of Health and Human Services, Guidelines for the Format 
and Content of the Clinical and Statistical Sections of New Drug 
Applications July, 1988.
73. American Academy of Pediatrics. Pediatric Nutrition Handbook, 3d 
ed., 1993.
74. Roche, A. F., S. Guo, and W. M. Moore, ``Weight and Recumbent 
Length From 1 to 12 Mo of Age: Reference Data for 1-Mo Increments,'' 
American Journal of Clinical Nutrition, 49:599-607, 1989.
75. Moore, W. M., and A. F. Roche, ``Pediatric Anthropometry,'' Ross 
Laboratories.
76. Behrman, R. E., Nelson Textbook of Pediatrics, 14th ed., W. B. 
Saunders Co., Philadelphia, 1992.
77. Sarwar, G., R. W. Peace, and H. G. Botting, ``Differences in 
Protein Digestibility and Quality of Liquid Concentrate and Powder 
Forms of Milk-Based Infant Formulas Fed to Rats,'' American Journal 
of Clinical Nutrition, 49:806-13, 1989.
78. Rudloff, S., and B. Lonnerdal, ``Solubility and Digestibility of 
Milk Proteins in Infant Formulas Exposed to Different Heat 
Treatments,'' Journal of Pediatric and Gastroenterology and 
Nutrition, 15:25-32, 1992.
79. Fomom, S. J., and S. E. Nelson, Chapter 11, Calcium, Phosphorus, 
Magnesium, and Sulfur in Nutrition of Normal Infants, edited by 
Fomon, S. J., Mosby-Year Book, Inc., St. Louis, MO, 1993.
80. Ryan, S. A., J. Truscott, M. Simpson, and J. James, ``Phosphate, 
Alkaline Phosphatase and Bone Mineralization in Preterm Neonates,'' 
Acta Pediatrica, 82:518-521, 1993.
81. Minton, S. D., J. J. Steichen, and R. C. Tsang, ``Bone Mineral 
Content in Term and Preterm Appropriate-for-Gestational-Age 
Infants,'' Journal of Pediatrics, 95:1037-1042, 1979.
82. Steichen, J. J., P. A. Steichen Asch, and R. C. Tsang, ``Bone 
Mineral Content Measurement in Small Infants by Single-photon 
Absorptiometry: Current Methodologic Issues,'' Journal of 
Pediatrics, 113 (suppl): 181-187, 1988.
83. Chan, G. M., ``Performance of Dual- Energy X-ray Absorptiometry 
in Evaluating Bone, Lean Body Mass, and Fat in Pediatric Subjects,'' 
Journal of Bone and Mineral Research, 7:369-374, 1992.
84. Braillon, P. M., B. L. Salle, J. Brunet, F. H. Glorieux, P. D. 
Delmas, and P. J. Meunier, ``Dual Energy X-ray Absorptiometry 
Measurement of Bone Mineral Content in Newborns: Validation of the 
Technique,'' Pediatric Research, 32:77-80, 1992.
85. Salle, B. L., and F. H. Glorieux, ``Assessment of Bone Mineral 
Content in Infants: The New Age,'' Acta Pediatrica, 82:709-10, 1993.
86. Fomon, S. J., Chapter 14, Iron in Nutrition of Normal Infants, 
edited by Fomon, S. J., Mosby-Year Book, Inc., St. Louis, MO, 1993.
87. Carlson, S. E., R. J. Cooke, S. H. Werkman, and E. A. Tolley, 
``First Year Growth of Preterm Infants Fed Standard Compared to 
Marine Oil n-3 Supplemented Formulas,'' Lipids, 27:901-907, 1992.
88. Memorandum of telephone conversation between Marty Hudtloff, 
FDA, and Ed Puro, FDA, March 15, 1994; cost of Maple Island recalls 
note, memorandum from Robert Fish, March 7, 1994; memorandum from 
Martin Hudtloff, February 28, 1994; Establishment History Report on 
B2110076.
89. Memorandum of telephone conversation between Hazleton Labs and 
Ed Puro, FDA, March 17, 1993.

List of Subjects

21 CFR Part 106

    Food grades and standards, Infants and children, Nutrition, 
Reporting and recordkeeping requirements, Incorporation by reference.

21 CFR Part 107

    Food labeling, Infants and children, Nutrition, Reporting and 
recordkeeping requirements, Signs and symbols.
    Therefore, under the Federal Food, Drug, and Cosmetic Act and under 
authority delegated to Commissioner of Food and Drugs, it is proposed 
that 21 CFR parts 106 and 107 be amended as follows:

PART 106--INFANT FORMULA--REQUIREMENTS PERTAINING TO CURRENT GOOD 
MANUFACTURING PRACTICE, QUALITY CONTROL PROCEDURES, QUALITY 
FACTORS, RECORDS AND REPORTS, AND NOTIFICATIONS

    1. The authority citation for 21 CFR part 106 continues to read as 
follows:

    Authority: Secs. 201, 412, 701 of the Federal Food, Drug, and 
Cosmetic Act (21 U.S.C. 321, 350a, 371).

    2. The heading for part 106 is revised to read as set forth above.
    3. Section 106.1 is revised to read as follows:


Sec. 106.1  Status and applicability of the regulations in part 106.

    (a) The criteria set forth in subparts B, C, and D of this part 
prescribe the steps that manufacturers must take under section 
412(b)(2) and (b)(3) of the Federal Food, Drug, and Cosmetic Act (the 
act) in processing infant formula. If the processing of the formula 
does not comply with any regulation in subparts B, C, or D of this 
part, the formula will be deemed to be adulterated under section 
412(a)(3) of the act.
    (b) The criteria set forth in subpart E of this part prescribe the 
quality factor requirements that infant formula must

[[Page 36209]]

meet under section 412(b)(1) of the act. If the formula fails to comply 
with any regulation in subpart E of this part, it will be deemed to be 
adulterated under section 412(a)(2) of the act.
    (c) The criteria set forth in subpart F of this part implement the 
record retention requirements established in section 412(b)(4) of the 
act. Failure to comply with any regulation in subpart F of this part is 
a violation of section 301(e) of the act.
    (d) The criteria set forth in subpart G of this part describe the 
circumstances in which infant formula manufacturers are required to 
register with, submit to, or notify the Food and Drug Administration, 
and the content of those registrations, submissions, or notifications, 
under section 412(c), (d), and (e) of the act. Failure to comply with 
any regulation in subpart G of this part is a violation of section 
301(s) of the act.
    4. Section 106.3 is revised to read as follows:


Sec. 106.3  Definitions.

    The definitions in this section and the definitions contained in 
section 201 of the Federal, Food, Drug, and Cosmetic Act (the act) 
shall apply to infant formula requirements in 21 CFR part 106 and part 
107 of this chapter.
    (a) Batch means a specific quantity of an infant formula or other 
material that is intended to have uniform character and quality, within 
specified limits, and is produced according to a single manufacturing 
order during the same cycle of manufacture.
    (b) Final-product-stage means the point in the manufacturing 
process, before distribution of an infant formula, at which the infant 
formula is homogeneous and is not subject to further degradation due to 
processing.
    (c) Indicator nutrient means a nutrient whose concentration is 
measured during the manufacture of an infant formula to confirm 
complete addition and uniform distribution of a premix or other 
substance of which the indicator nutrient is a part.
    (d) Infant means a person not more than 12 months of age.
    (e) Infant formula means a food which purports to be or is 
represented for special dietary use solely as a food for infants by 
reason of its simulation of human milk or its suitability as a complete 
or partial substitute for human milk.
    (f) In-process batch means a combination of ingredients at any 
point in the manufacturing process before packaging.
    (g) Lot means a batch, or a specifically identified portion of a 
batch, having uniform character and quality within specified limits; 
or, in the case of an infant formula produced by continuous process, it 
is a specific identified amount produced in a unit of time or quantity 
in a manner that assures its having uniform character and quality 
within specified limits.
    (h) Lot number, control number, or batch number means any 
distinctive combination of letters, numbers, symbols, or any 
combination of them, from which the complete history of the 
manufacture, processing, packing, holding, and distribution of a batch 
or lot of infant formula or other material can be determined.
    (i) Major change in an infant formula means any new formulation, or 
any change of ingredients or processes where experience or theory would 
predict a possible significant adverse impact on levels of nutrients or 
bioavailability of nutrients, or any change that causes an infant 
formula to differ fundamentally in processing or in composition from 
any previous formulation produced by the manufacturer. Examples of 
infant formulas deemed to differ fundamentally in processing or in 
composition include:
    (1) Any infant formula produced by a manufacturer who is entering 
the U.S. market;
    (2) Any infant formula powder processed and introduced for 
commercial or charitable distribution by a manufacturer who previously 
only produced liquids (or vice versa);
    (3) Any infant formula having a significant revision, addition, or 
substitution of a macronutrient (i.e., protein, fat, or carbohydrate), 
with which the manufacturer has not had previous experience;
    (4) Any infant formula manufactured on a new processing line or in 
a new plant;
    (5) Any infant formula manufactured containing a new constituent 
not listed in section 412(i) of the act, such as taurine or L-
carnitine;
    (6) Any infant formula processed by a manufacturer on new equipment 
that utilizes a new technology or principle (e.g., a change from 
terminal sterilization to aseptic processing); and
    (7) An infant formula for which there has been a fundamental change 
in the type of packaging used (e.g., changing from metal cans to 
plastic pouches).
    (j) Manufacturer means a person who prepares, reconstitutes, or 
otherwise changes the physical or chemical characteristics of an infant 
formula or packages or labels the product in a container for 
distribution.
    (k) Microorganisms means yeasts, molds, bacteria, and viruses and 
includes, but is not limited to, species having public health 
significance.
    (l) New infant formula means:
    (1) An infant formula manufactured by a person that has not 
previously manufactured an infant formula for the U.S. market, and
    (2) An infant formula manufactured by a person that has previously 
manufactured infant formula and in which there is a major change in 
processing or formulation from a current or any previous formulation 
produced by such manufacturer.
    (m) Nutrient means any vitamin, mineral, or other substance or 
ingredient that is required in accordance with the table set out in 
section 412(i)(1) of the act or by regulations issued under section 
412(i)(2) or that is identified as essential for infants by the Food 
and Nutrition Board of the National Research Council through its 
development of a Recommended Dietary Allowance or an Estimated Safe and 
Adequate Daily Dietary Intake range, or that has been identified as 
essential for infants by the Food and Drug Administration through a 
Federal Register publication.
    (n) Nutrient premix means a combination of ingredients containing 
two or more nutrients received from a supplier or prepared by an infant 
formula manufacturer.
    (o) Quality factors mean those factors necessary to demonstrate 
that the infant formula, as prepared for market, provides nutrients in 
a form that is bioavailable and safe as shown by evidence that 
demonstrates that the formula supports healthy growth when fed as a 
sole source of nutrition.
    (p) Representative sample means a sample that consists of a number 
of units that are drawn based on rational criteria, such as random 
sampling, and intended to ensure that the sample accurately portrays 
the material being sampled.
    (q) Shall is used to state mandatory requirements.
    (r) Should is used to state recommended or advisory procedures or 
to identify recommended equipment.
    5. Part 106 is amended by redesignating subparts B, C, and D as 
subparts C, F, and G, respectively, and adding new subparts B, D, and 
E; and by revising newly redesignated subparts C and G to read as 
follows:
* * * * *

[[Page 36210]]

Subpart B--Current Good Manufacturing Practice

Sec.
106.5  Current good manufacturing practice.
106.6  Production and in-process control system.
106.10  Controls to prevent adulteration by workers.
106.20  Controls to prevent adulteration caused by facilities.
106.30  Controls to prevent adulteration caused by equipment or 
utensils.
106.35  Controls to prevent adulteration due to automatic 
(mechanical or electronic) equipment.
106.40  Controls to prevent adulteration caused by ingredients 
containers, and closures.
106.50  Controls to prevent adulteration during manufacturing.
106.55  Controls to prevent adulteration from microorganisms.
106.60  Controls to prevent adulteration during packaging and 
labeling of infant formula.
106.70  Controls on the release of finished infant formula.
106.80  Traceability.
106.90  Audits of current good manufacturing practice.

Subpart C--Quality Control Procedures

106.91  General quality control.
106.92  Audits of quality control procedures.

Subpart D--Conduct of Audits

106.94  Audit plans and procedures.

Subpart E--Quality Factors for Infant Formulas

106.96  Quality factors in infant formulas.
106.97  Assurances for quality factors.
* * * * *

Subpart G--Registration, Submission, and Notification Requirements

106.110  New infant formula registration.
106.120  New infant formula submission.
106.121  Quality factor submission.
106.130  Verification submission.
106.140  Submission concerning a change in infant formula that may 
adulterated the product.
106.150  Notification of an adulterated or misbranded infant 
formula.
* * * * *

Subpart B--Current Good Manufacturing Practice


Sec. 106.5  Current good manufacturing practice.

    (a) The regulations set forth in this subpart and, for liquid 
infant formulas, in part 113 of this chapter define the minimum current 
good manufacturing practices that are to be used in, and the facilities 
or controls that are to be used for, the manufacture, processing, 
packing, or holding of an infant formula. Compliance with these 
provisions is necessary to ensure that such infant formula provides the 
nutrients required under Sec. 107.100 of this chapter and is 
manufactured in a manner designed to prevent its adulteration.
    (b) The failure to comply with any regulation set forth in this 
subpart or, for liquid infant formulas, in part 113 of this chapter in 
the manufacture, processing, packing, or holding of an infant formula 
shall render such infant formula adulterated under section 412(a)(3) of 
the Federal Food, Drug, and Cosmetic Act (the act).


Sec. 106.6  Production and in-process control system.

    (a) Manufacturers shall conform to the requirements of this subpart 
by implementing a system of production and in-process controls. This 
production and in-process control system shall cover all stages of 
processing, from the receipt and acceptance of the raw materials, 
ingredients, and components through the storage and distribution of the 
finished product and shall be designed to ensure that all the 
requirements of this subpart are met.
    (b) The production and in-process control system shall be set out 
in a written plan, or set of procedures, that is designed to ensure 
that an infant formula is manufactured in a manner that will prevent 
adulteration of the infant formula.
    (c) At any point, step, or stage in the production process where 
control is necessary to prevent adulteration, the manufacturer shall:
    (1) Establish standards or specifications to be met;
    (2) Monitor the production and in-process control point, step, or 
stage;
    (3) Establish corrective action plans for use when a standard or 
specification established in accordance with paragraph (b)(1) of this 
section is not met;
    (4) Review the results of the monitoring required by paragraph 
(c)(2) of this section, and review and evaluate the public health 
significance of any deviations from standards or specifications that 
have been established in accordance with paragraph (c)(1) of this 
section. This review shall be conducted by an individual qualified by 
training and experience to conduct such reviews; and
    (5) Establish recordkeeping procedures, in accordance with 
Sec. 106.100(e)(3), that ensure that compliance with the requirements 
of this section is documented.


Sec. 106.10  Controls to prevent adulteration by workers.

    (a) There shall be sufficient personnel, qualified by training and 
experience, to perform all operations, including all required 
recordkeeping, in the manufacture, processing, packing, and holding of 
each infant formula and to supervise such operations to ensure that 
they are correctly and fully performed.
    (b) Personnel working directly with infant formula, infant formula 
raw materials, infant formula packaging, or infant formula equipment or 
utensil contact surfaces shall practice good personal hygiene to 
protect the infant formula against contamination. Good personal hygiene 
includes, but is not limited to:
    (1) Wearing clean outer garments and, as necessary, protective 
apparel such as head, face, hands, and arm coverings; and
    (2) Washing hands thoroughly in a hand washing facility with soap 
and running water at a suitable temperature before starting work, after 
each absence from the work station, and at any other time when the 
hands may become soiled or contaminated.
    (c) Any person who reports that he or she has, or appears by 
medical examination or supervisory observation to have, an illness, 
open lesion, including boils, sores, or infected wounds, or any other 
source of microbial contamination that creates a reasonable possibility 
that the safety of an infant formula may be adversely affected, shall 
be excluded from direct contact with ingredients, containers, closures, 
in-process materials, equipment, utensils, and infant formula product 
until the condition is corrected or determined by competent medical 
personnel not to jeopardize the safety of the infant formula.


Sec. 106.20  Controls to prevent adulteration caused by facilities.

    (a) Buildings used in the manufacture, processing, packing, or 
holding of infant formula shall be maintained in a clean and sanitary 
condition and shall have space for the separation of incompatible 
operations, such as the handling of raw materials, the manufacture of 
the product, and packaging and labeling operations.
    (b) Separate areas shall be designated for holding raw materials, 
in-processing materials, and final product infant formula:
    (1) Pending release for use in infant formula production or pending 
release of the final product,
    (2) After rejection for use in infant formula and before 
disposition, and
    (3) After release for use in infant formula production or after 
release of the final product.
    (c) Lighting shall allow easy identification of raw materials,

[[Page 36211]]

packaging, labeling, in-process materials, and finished products that 
have been released for use in infant formula production and shall 
permit the easy reading of instruments and controls necessary in 
processing, packaging, and laboratory analysis. Any lighting fixtures 
directly over or adjacent to exposed raw materials, in-process 
materials, or bulk (unpackaged) finished product shall be protected to 
prevent glass from contaminating the product in the event of breakage.
    (d) Air filtration systems, including prefilters and particulate 
matter air filters, shall be used on air supplies to production areas 
where ingredients or infant formula are directly exposed to the 
atmosphere.
    (e) All rodenticides, insecticides, fungicides, fumigating agents, 
and cleaning and sanitizing agents shall be stored and used in a manner 
that protects against contamination of infant formula.
    (f)(1) Potable water used in the manufacture of infant formula 
shall meet the standards prescribed in the Environmental Protection 
Agency's (EPA's) Primary Drinking Water Regulations set forth in 40 CFR 
part 141, except that the fluoride level of the water used in infant 
formula manufacturing shall be as low as possible. The water shall be 
supplied under continuous positive pressure in a plumbing system that 
is free of defects that could contaminate an infant formula.
    (2) Manufacturers shall test representative samples of the potable 
water drawn at a point in the system at which the water is in the same 
condition that it will be when it is used in infant formula 
manufacturing.
    (3) Manufacturers shall conduct the tests required by paragraph 
(f)(2) of this section with sufficient frequency to ensure that the 
water meets the EPA's Primary Drinking Water Regulations but shall not 
conduct these tests less frequently than annually for chemical 
contaminants, every 4 years for radiological contaminants, and weekly 
for bacteriological contaminants.
    (4) Manufacturers shall make and retain records, in accordance with 
Sec. 106.100(f)(1), of the frequency and results of testing of the 
water used in the production of infant formula.
    (g) There shall be no backflow from, or cross-connection between, 
piping systems that discharge waste water or sewage and piping systems 
that carry water for infant formula manufacturing.
    (h) When steam comes in direct contact with infant formula, it 
shall be safe and free of rust and other particulate matter that may 
contaminate the formula. Boiler water additives in the steam shall be 
used in accordance with Sec. 173.310 of this chapter.
    (i) Each infant formula manufacturing site shall provide its 
employees with readily accessible toilet facilities and hand washing 
facilities that include hot and cold water, soap or detergent, and 
single-service towels and that are maintained in good repair and in a 
sanitary condition at all times, and that these facilities provide for 
proper disposal of the sewage. Doors to the toilet facility shall not 
open into areas where infant formula ingredients, containers, or 
closures are stored, or where infant formula is processed or stored.


Sec. 106.30  Controls to prevent adulteration caused by equipment or 
utensils.

    (a) Equipment used in the manufacture, processing, packing or 
holding of an infant formula shall be of appropriate design and shall 
be installed to facilitate its intended function and its cleaning and 
maintenance.
    (b) Equipment and utensils used in the manufacture, processing, 
packing, or holding of an infant formula shall be constructed so that 
surfaces that contact ingredients, in-process materials, or infant 
formula are made of nontoxic materials and are not reactive or 
absorptive. Such equipment and utensils shall be designed to be easily 
cleanable and to withstand the environment of their intended use. All 
surfaces that contact ingredients, in-process materials, or infant 
formula shall be cleaned, sanitized, and maintained to protect infant 
formula from being contaminated by any source. Sanitizing agents used 
on food-contact surfaces must comply with Sec. 178.1010 of this 
chapter.
    (c) Manufacturers shall ensure that substances, such as lubricants 
or coolants, that are required for operation of infant formula 
manufacturing equipment, but that would render the infant formula 
adulterated if they contaminated the formula, do not come in contact 
with formula ingredients, containers, closures, or in-process materials 
or with infant formula itself.
    (d)(1) Manufacturers shall ensure that instruments used for 
measuring, regulating, or controlling mixing time and speed, 
temperature, pressure, moisture, water activity, or other parameters at 
points where control is deemed necessary to prevent adulteration in the 
processing of an infant formula are accurate, easily read, properly 
maintained, and present in sufficient number for their intended use. 
The instruments and controls shall be tested for accuracy (calibrated) 
against a known reference standard before first use and thereafter at 
routine intervals, as specified in writing by the manufacturer of the 
instrument or control, or as otherwise deemed necessary to ensure the 
accuracy of the instrument. The known reference standard shall be 
certified for accuracy at routine intervals specified in writing by the 
manufacturer of the instrument, or as otherwise deemed necessary to 
ensure the accuracy of the instrument. Manufacturers shall make and 
retain records of the accuracy checks in accordance with 
Sec. 106.100(f)(2).
    (2) Instruments and controls that cannot be adjusted to agree with 
the reference standard shall be repaired or replaced.
    (3) If calibration of an instrument (testing for accuracy against a 
known reference standard) shows that a specification or standard for a 
point where control is deemed necessary to prevent adulteration has not 
been met, a written evaluation of all affected product, and of any 
actions that need to be taken with respect to that product, shall be 
made, in accordance with Sec. 106.100(f)(2).
    (e)(1) The temperature in cold storage compartments that are used 
to store raw materials, in-process materials, or final product, and in 
thermal processing equipment used at points where temperature control 
is necessary to prevent adulteration, shall be monitored with such 
frequency as is necessary to ensure that temperature control is 
maintained.
    (2) Cold storage compartments shall be maintained at a temperature 
of 40  deg.F (4.4  deg.C) or below.
    (3)(i) Cold storage compartments and thermal processing equipment 
shall be equipped with easily readable, accurate temperature-indicating 
devices.
    (ii) Thermal processing equipment shall be equipped with 
temperature-recording devices that will reflect the true temperature on 
a continuing basis. Cold storage compartments shall be equipped with 
either temperature-recording devices that will reflect the true 
temperature, on a continuing basis, within the compartment or, in lieu 
of a temperature-recording device, a high temperature alarm or a 
maximum-indicating thermometer that has been verified to function 
properly. If the manufacturer uses either of the latter options, it 
shall maintain a temperature log in which it notes temperature with 
such frequency as is necessary to achieve control. Manufacturers shall 
make and retain records, in accordance with Sec. 106.100(f)(3), of the 
temperatures indicated or recorded by these devices.

[[Page 36212]]

    (4) When a temperature-recording device is used, such device shall 
not read higher than the calibrated temperature-indicating device for 
thermal processing equipment or lower than the reference temperature-
indicating device for cold storage compartments.
    (f) Equipment and utensils used in the manufacture of infant 
formula shall be cleaned, sanitized, and maintained at regular 
intervals to prevent adulteration of the infant formula. An individual 
qualified by training or experience to conduct such a review shall 
check all cleaning, sanitizing, and maintenance to ensure that it has 
been satisfactorily completed. Manufacturers shall make and retain 
records on equipment cleaning, sanitizing, and maintenance, in 
accordance with Sec. 106.100(f)(4).
    (g) Compressed air or other gases that are mechanically introduced 
into infant formula, that are used to clean any equipment, or that come 
into contact with any other surface that contacts ingredients, in-
process materials, or infant formula shall be treated in such a way 
that their use will not contaminate the infant formula with unlawful 
indirect food additives or other chemical, physical, or microbiological 
contaminants. When compressed gases are used at product filling 
machines to replace air removed from the headspace of containers, the 
manufacturer shall install a 0.5 micrometer or smaller filter as close 
to the end of the gas line that feeds gas into the space, as practical.


Sec. 106.35  Controls to prevent adulteration due to automatic 
(mechanical or electronic) equipment.

    (a)(1) For the purposes of this section, ``hardware'' means all 
automatic equipment, including mechanical and electronic equipment 
(including computers), that is used in production or quality control of 
a infant formula.
    (2) For the purposes of this section, ``software'' means any 
programs, procedures, rules, and associated documentation used in the 
operation of a system.
    (3) For the purposes of this section, ``system'' means a collection 
of components (including software and hardware) organized to accomplish 
a specific function or set of functions in a specified environment.
    (4) For the purposes of this section, ``validation'' means 
establishing documented evidence that provides a high degree of 
assurance that a system will consistently produce a product meeting its 
predetermined specifications and quality characteristics.
    (b)(1) All systems shall be designed, installed, tested, and 
maintained in a manner that will ensure that they are capable of 
performing their intended function and of producing or analyzing infant 
formula in accordance with this subpart and subpart C of this part.
    (2) The infant formula manufacturer shall ensure that hardware is 
routinely calibrated, inspected, and checked according to written 
procedures.
    (3) The infant formula manufacturer shall check and document the 
accuracy of input into, and output generated by, any system used in the 
production or quality control of an infant formula. The degree and 
frequency of input/output verification shall be based on the complexity 
and reliability of the system and the level of risk associated with the 
safe operation of the system.
    (4) The infant formula manufacturer shall ensure that all systems 
are validated before their first use to manufacture commercial product.
    (5) The infant formula manufacturer shall ensure that any system 
that is modified is revalidated after the modification and before use 
of the modified system to manufacture commercial product. All 
modifications to software shall be made by a designated individual and 
shall be checked by the infant formula manufacturer to ensure that 
infant formula that is produced or analyzed using the modified software 
complies with this subpart and with subpart C of this part.
    (c) The infant formula manufacturer shall make and retain records, 
in accordance with Sec. 106.100(f)(5), concerning automatic (mechanical 
or electronic) equipment.


Sec. 106.40  Controls to prevent adulteration caused by ingredients, 
containers, and closures.

    (a) The only substances that may be used in infant formulas are 
food ingredients whose use in infant formula is safe and suitable under 
the applicable food safety provisions of the Federal Food, Drug, and 
Cosmetic Act; that is, the substance is generally recognized as safe 
(GRAS) for such use, is used in accordance with the agency's food 
additive regulations, or is authorized by a prior sanction.
    (b) Infant formula containers and closures shall not be reactive or 
absorptive so as to affect the safety of the infant formula, and all 
packaging material that comes in contact with infant formula shall be 
composed of substances that are GRAS for use in or on food, GRAS for 
their intended use in food packaging, authorized by a prior sanction 
issued by the agency, or authorized for use as an indirect food 
additive. Any packaging material that comes in contact with infant 
formula shall be used in accordance with any prescribed limitations.
    (c) Ingredients, containers, and closures used in the manufacture 
of infant formula shall be identified with a batch or lot number to be 
used in recording their disposition.
    (d) Infant formula manufacturers shall develop written 
specifications for their acceptance or rejection of ingredients, 
containers, and closures used in infant formula manufacture. These 
specifications shall stipulate the standards for acceptance or 
rejection of such ingredients, containers, and closures as well as the 
procedures for determining whether the ingredients, containers, and 
closures meet that standard. An individual qualified by training or 
experience shall conduct an investigation of a finding that any 
ingredients, containers, or closures used in a batch of infant formula 
failed to meet any of the manufacturer's specifications.
    (e) Ingredients, containers and closures shall be stored in areas 
clearly designated for:
    (1) Materials pending release for use,
    (2) Materials released for use, or
    (3) Materials rejected for use in infant formula production. Any 
lot of ingredients, containers, or closures that does not meet the 
manufacturer's specifications shall be rejected and controlled under a 
quarantine system designed to prevent its use in the manufacture of 
infant formula.
    (f) If an ingredient, a container, or a closure that has been 
tested and examined is exposed to air, heat, or other conditions that 
may adversely affect it, the ingredient, container, or closure shall be 
retested or reexamined to ensure that it still meets the manufacturer's 
specifications.
    (g) Manufacturers shall make and retain records, in accordance with 
Sec. 106.100(f)(6), on the ingredients, containers, and closures used 
in the manufacture of infant formula.


Sec. 106.50  Controls to prevent adulteration during manufacturing.

    (a)(1) Manufacturers shall prepare and follow a written master 
manufacturing order that establishes controls and procedures for the 
production of an infant formula.
    (2) The manufacturer shall make and retain records, in accordance 
with Sec. 106.100(e), that include complete information relating to the 
production and control of the batch. An individual qualified by 
training or experience shall conduct an investigation of any

[[Page 36213]]

deviations from the master manufacturing order and any corrective 
actions taken.
    (3) Changes made to the master manufacturing order shall be 
drafted, reviewed, and approved by a responsible official and include 
an evaluation of the effect of the change on the nutrient content and 
the suitability of the formula for infants.
    (b) The manufacturer shall establish controls to ensure that each 
raw or in-process ingredient required by the master manufacturing order 
is examined by one person and checked by a second person or system. 
This checking will ensure that the correct ingredient is added during 
the manufacturing process, that the ingredient has been released for 
use in infant formula, and that the correct weight or measure of the 
ingredient is added to the batch.
    (c) The manufacturer shall identify the contents, including the 
processing stage and the lot or batch number of a batch of infant 
formula, of all compounding and storage containers, processing lines, 
and major equipment used during the production of a batch of an infant 
formula.
    (d) The manufacturer shall establish controls to ensure that the 
nutrient levels required by Sec. 107.100 of this chapter are maintained 
in the formula, and that the formula is not contaminated with 
microorganisms or other contaminants. Such controls shall include but 
not be limited to:
    (1) The mixing time; the speed, temperature, and flow rate of 
product; and other critical parameters necessary to ensure the addition 
of required ingredients to, and the homogeneity of, the formula;
    (2) The spray-drying process for powdered infant formula, including 
the filtering of the intake air before heating, to prevent microbial 
and other contamination;
    (3) The removal of air from the finished product to ensure that 
nutrient deterioration does not occur;
    (4) Ensuring that each container of finished product is properly 
sealed. Such controls shall involve use of established procedures, 
specifications, and intervals of examination that are designed by 
qualified individuals and are sufficient to:
    (i) Detect visible closure or seal defects, and
    (ii) Determine closure strength through destructive testing. 
Manufacturers of liquid infant formulas, which are thermally processed 
low-acid foods packaged in hermetically sealed containers, shall 
perform such closure integrity testing in accordance with 
Sec. 113.60(a) of this chapter.
    (e) The manufacturer shall establish controls that ensure that the 
equipment used at points where control is deemed necessary to prevent 
adulteration is monitored, so that personnel will be alerted to 
malfunctions.
    (f) The manufacturer shall establish controls that ensure that 
rejected in-process materials:
    (1) Are clearly identified as having been rejected for use in an 
infant formula;
    (2) Are controlled under a quarantine system designed to prevent 
their use in manufacturing or processing operations for which they are 
unsuitable;
    (3) Meet the appropriate specifications, if reprocessed, before 
being released for use in infant formula.


Sec. 106.55  Controls to prevent adulteration from microorganisms.

    (a) Manufacturers of liquid infant formula shall comply with the 
procedures specified in part 113 of this chapter for liquid infant 
formula.
    (b) Manufacturers of powdered infant formula shall test 
representative samples of every batch of the formula at the final 
product stage, before distribution, to ensure that the infant formula 
meets the microbiological quality standards listed in paragraph (c) of 
this section.
    (c) Any powdered infant formula that contains any microorganism 
that exceeds the M value listed for that microorganism in Table 1 of 
this section will be deemed to be adulterated under sections 402 and 
412 of the Federal Food, Drug, and Cosmetic Act (the act). FDA will 
determine compliance with the M values listed below using the 
Bacteriological Analytical Manual (BAM), 8th ed. (1995), published by 
the AOAC International Association of Official Analytical Chemists, 
which is incorporated by reference in accordance with 5 U.S.C. 552(a) 
and 1 CFR part 51. Copies are available from the Association of 
Official Analytical Chemists, 481 North Frederick Ave., suite 500, 
Gaithersburg, MD 20877, or may be examined at the Center for Food 
Safety and Applied Nutrition's Library, 200 C St. SW., rm. 3321, 
Washington, DC, or may be examined at the Office of the Federal 
Register, 800 North Capitol St. NW., suite 700, Washington, DC.

------------------------------------------------------------------------
               Microorganism                         M value \1\        
------------------------------------------------------------------------
Aerobic Plate Count (APC).................  10,000 CFU/gram (g). \2\    
Coliforms \3\.............................  3.05 MPN/g. 4,5             
Fecal coliforms \6\.......................  3.05 MPN/g.                 
Salmonella................................  0. \7\                      
Listeria monocytogenes....................  0. \7\                      
Staphylococcus aureus.....................  3.05 MPN/g.                 
Bacillus cereus \8\.......................  100 MPN/g or CFU/g.         
------------------------------------------------------------------------
\1\ The M value is the maximum allowable number of microorganisms       
  present in 1 g of dry infant formula.                                 
\2\ CFU/g, colony forming units per g.                                  
\3\ M values for coliforms greater than 3.05 are not violative if       
  testing for fecal coliforms results in an M value equal to or less    
  than 3.05.                                                            
\4\ MPN/g, most probable number per g.                                  
\5\ The MPN value of 3.05 in this table is derived from the tables of   
  calculated MPN values that appear in the 8th ed. of the BAM when using
  an inoculation series of 0.1, 0.01, and 0.001g (or ml) of the infant  
  formula sample.                                                       
\6\ No testing for fecal coliforms is required when the M value for     
  coliforms is less than or equal to 3.05.                              
\7\ None detected.                                                      
\8\B. cereus testing must be performed only if the APC exceeds 100 CFU/ 
  g.                                                                    

    (d) Manufacturers shall make and retain records, in accordance with 
Sec. 106.100(e)(5)(ii) and (f)(7), on the testing of infant formulas 
for microorganisms.


Sec. 106.60  Controls to prevent adulteration during packaging and 
labeling of infant formula.

    (a) Manufacturers shall examine packaged and labeled infant formula 
during finishing operations to ensure that containers and packages in 
the lot have the correct label, the correct use-by date, and the 
correct code established under Sec. 106.80.
    (b) Labels shall be designed, printed, and applied so that the 
labels remain legible and attached during the conditions of processing, 
storage, handling, distribution, and use.
    (c) All infant formula held in a single package shall be the same 
product bearing the same code, established under Sec. 106.80. Packaging 
used to hold multiple containers of infant formula shall be labeled 
with the product name, the name of the manufacturer or shipper, and the 
code.


Sec. 106.70  Controls on the release of finished infant formula.

    (a) The manufacturer shall hold, or maintain under its control, 
each batch of infant formula until it determines that the batch meets 
all of its specifications, including those adopted to meet the 
requirements of Sec. 106.55 on microbiological contamination and 
Sec. 106.91(a) on quality control procedures, and releases the batch 
for distribution.
    (b) Each batch of infant formula that fails to meet the 
manufacturer's specifications shall be rejected. Although the batch may 
be reprocessed, any batch of infant formula that is reprocessed shall 
be shown to meet the

[[Page 36214]]

requirements of Sec. 106.70(a) before it is released.
    (c) An individual qualified by training or experience shall conduct 
an investigation of a finding that a batch of infant formula fails to 
meet any manufacturer's specifications.


Sec. 106.80  Traceability.

    (a) Manufacturers shall ensure traceability by coding infant 
formulas in conformity with the coding requirements prescribed in 
Sec. 113.60(c) of this chapter for thermally processed low-acid foods 
packaged in hermetically-sealed containers, except as provided in 
paragraph (b) of this section.
    (b) Batches of powdered infant formula that are manufactured in 
stages over more than 1 day, in lieu of being coded in accordance with 
Sec. 113.60(c) of this chapter, may be coded with a sequential number 
that identifies the product and the establishment where the product was 
packed and that permits tracing of all stages of manufacture of that 
batch, including the year, the days of the year, and the period during 
those days that the product was packed, and the receipt and handling of 
raw materials used.


Sec. 106.90  Audits of current good manufacturing practice.

    Manufacturers of an infant formula, or an agent of such 
manufacturers, shall conduct regularly scheduled audits to determine 
whether the manufacturer has complied with the current good 
manufacturing practice regulations in this subpart. These audits shall 
be performed by an individual who, as a result of education, training, 
and experience, is knowledgeable in all aspects of infant formula 
production and of the agency's regulations concerning current good 
manufacturing practice but who has no direct responsibility for the 
matters being audited.

Subpart C--Quality Control Procedures


Sec. 106.91  General quality control.

    (a) Nutrient testing to ensure that each batch of infant formula 
provides nutrients in accordance with Sec. 107.100. Manufacturers shall 
test each batch as follows:
    (1) Each nutrient premix used in the manufacture of an infant 
formula shall be tested for each nutrient that the manufacturer is 
relying on the premix to provide to ensure that the premix is in 
compliance with the manufacturer's specifications;
    (2) During the manufacturing process, after the addition of the 
premix, or at the final-product-stage but before distribution, each 
batch of infant formula shall be tested for at least one indicator 
nutrient for each of the nutrient premixes used in the infant formula 
to confirm that the nutrients supplied by each of the premixes are 
present, in the proper concentration, in the batch of infant formula.
    (3) At the final-product-stage, before distribution of an infant 
formula, each batch shall be tested for vitamins A, C, E, and thiamin.
    (4) During the manufacturing process or at the final-product-stage, 
before distribution, each batch shall be tested for all nutrients 
required to be included in such formula under Sec. 107.100 of this 
chapter and for any nutrient added by the manufacturer for which 
testing is not conducted for compliance with paragraphs (a)(1) or 
(a)(3) of this section.
    (b) Stability testing. Every 3 months, manufacturers shall collect 
representative samples from the final-product-stage of one batch of 
each physical form (powder, ready-to-feed, or concentrate) of each 
infant formula, at each manufacturing facility. The manufacturer shall 
test these samples for each nutrient required under Sec. 107.100 of 
this chapter and for any nutrient added by the manufacturer. The 
frequency of such testing shall be at the beginning, midpoint, and end 
of the shelf life of the infant formula and, depending on the nutrient 
and its stability within the matrix of the formulation, with additional 
frequency as is necessary to ensure that such formula complies with 
section 412 of the Federal Food, Drug, and Cosmetic Act (the act) 
throughout the shelf life of the infant formula; except that:
    (1) If the infant formula is a new infant formula, manufacturers 
shall collect a representative sample from the final-product-stage of 
each physical form (powder, ready-to-feed, or concentrate) of the first 
batch of the new infant formula and test these samples according to the 
requirements of this section; and
    (2) If an infant formula has been changed in formulation or in 
processing in a way that does not make it a new infant formula but that 
may affect whether it is adulterated under section 412(a) of the act, 
the manufacturer shall collect a representative sample from the final-
product-stage of each physical form (powder, ready-to-feed, or 
concentrate) of the first batch of the infant formula and shall test 
these samples according to the frequency required by this section for 
each nutrient that has been or may have been affected by the change.
    (c) Quality control records. Manufacturers shall make and retain 
quality control records in accordance with Sec. 106.100(e)(5)(i) and 
(f)(7).


Sec. 106.92  Audits of quality control procedures.

    A manufacturer of an infant formula, or an agent of such a 
manufacturer, shall conduct regularly scheduled audits to determine 
whether the manufacturer has complied with the quality control 
procedures that are necessary to ensure that an infant formula provides 
nutrients in accordance with section 412(b) and (i) of the Federal 
Food, Drug, and Cosmetic Act and is manufactured in a manner designed 
to prevent adulteration of the infant formula under section 412(a)(1) 
and (a)(3) of the Federal Food, Drug, and Cosmetic Act. These audits 
shall be performed by an individual who, as a result of education, 
training, and experience, is knowledgeable in all aspects of infant 
formula production and of the agency's regulations concerning quality 
control procedures but who has no direct responsibility for the matters 
being audited.

Subpart D--Conduct of Audits


Sec. 106.94  Audit plans and procedures.

    (a) Manufacturers shall develop and follow a written audit plan 
that is available at the manufacturing facility for FDA inspection.
    (b) The audit plan shall include audit procedures that set out the 
methods the manufacturer uses to determine whether the facility is 
operating in accordance with current good manufacturing practice, with 
the quality control procedures that are necessary to assure that an 
infant formula provides nutrients in accordance with section 412(b) and 
(i) of the Federal Food, Drug, and Cosmetic Act, and in a manner 
designed to prevent adulteration of the infant formula.
    (c) The audit procedures shall include, but not be limited to:
    (1) An evaluation of the production and in-process control system 
established under Sec. 106.6(b) by:
    (i) Observing the production of infant formula and comparing the 
observed process to the written production and in-process control plan 
required under Sec. 106.6(b);
    (ii) Reviewing records of the monitoring of points, steps, or 
stages where control is deemed necessary to prevent adulteration; and
    (iii) Reviewing records of how deviations from any standard or 
specification at points, steps, or stages where control is deemed 
necessary to prevent adulteration were handled; and

[[Page 36215]]

    (2) A review of a representative sample of all records maintained 
in accordance with Sec. 106.100(e) and (f).

Subpart E--Quality Factors for Infant Formulas


Sec. 106.96  Quality factors in infant formulas.

    (a) All infant formulas shall, when fed to infants as a sole source 
of nutrition, be of sufficient quality to meet the nutritional 
requirements for healthy growth. The regulations set forth in this 
subpart define the minimum quality factors for infant formulas.
    (b) All infant formulas shall be capable of supporting normal 
physical growth of infants.
    (c) All infant formulas shall be formulated and manufactured such 
that the protein is of sufficient biological quality to meet the 
protein requirements of infants.


Sec. 106.97  Assurances for quality factors.

    (a) General quality factor of normal physical growth. (1) The 
manufacturer shall conduct an adequate and well-controlled clinical 
study, in accordance with good clinical practice, to determine whether 
an infant formula supports normal physical growth in infants when the 
formula is fed as the sole source of nutrition.
    (i) The manufacturer shall:
    (A) Conduct a clinical study that is no less than 4 months in 
duration, enrolling infants no more than 1 month old at time of entry 
into the study.
    (B) Collect and maintain data in the study on anthropometric 
measures of physical growth, including body weight, recumbent length, 
head circumference, and average daily weight increment, and plot the 
data on National Center for Health Statistics (NCHS) reference 
percentile body weight and body length curves. The NCHS growth charts 
are incorporated by reference in accordance with 5 U.S.C. 552(a) and 1 
CFR part 51. Copies are available from the Office of Constituent 
Operations (HFS-565), Center for Food Safety and Applied Nutrition, 
Food and Drug Administration, 200 C St. SW., Washington, DC 20204, may 
be examined at the Office of Special Nutritionals (HFS-456), Center for 
Food Safety and Applied Nutrition, Food and Drug Administration, 200 C 
St. SW., Washington, DC 20204, or the Office of the Federal Register, 
800 North Capitol St. NW., suite 700, Washington, DC.
    (C) Collect anthropometric measurements at the beginning of the 
clinical study, at 2 weeks, at 4 weeks, at least monthly thereafter, 
and at the conclusion of the study.
    (ii) The clinical study protocol should:
    (A) Describe the scientific basis and objectives of the study, the 
planned control and treatment feeding regimens, the entrance criteria 
used to enroll infants in the study, the method of randomization used 
for the assignment of infants to feeding groups, the collection of 
specific measurements and other data, the methods used to limit sources 
of bias, and the planned methods of statistical analysis;
    (B) Describe the necessary qualifications and experience of 
investigators;
    (C) Be reviewed and approved by an Institutional Review Board (IRB) 
in accordance with part 56 of this chapter. The manufacturer shall 
establish procedures to obtain written informed consent from parents or 
legal representatives of the infants enrolled in the study in 
accordance with part 50 of this chapter;
    (D) Explain how the study population represents the population for 
which the new infant formula is intended and how the study addresses 
the intended conditions of use of the formula.
    (E) Describe the sample size calculations and the power 
calculations and the basis for selecting the sample size and study 
design;
    (F) Describe the plan to identify and evaluate any adverse effects;
    (G) Describe the quality control procedures used to ensure the 
validity and reliability of the measurements collected.
    (H) Describe and compare the composition of the test and control 
formulas.
    (I) Describe the basis upon which the test formula is appropriate 
for use in evaluating the formula that the manufacturer intends to 
market, if the test formula used in a study is not identical to the 
formula that is intended to be marketed in the United States.
    (2) The manufacturer may request an exemption from the requirements 
of paragraph (a)(1) of this section if:
    (i) The manufacturer has similar experience using an ingredient, an 
ingredient mixture, or a processing method in the production of an 
infant formula marketed in the United States and can demonstrate that 
infant formula made with that ingredient, ingredient mixture, or 
processing method meets the quality factor requirements in Sec. 106.96;
    (ii) The manufacturer markets a formulation in more than one form 
(e.g., liquid and powdered forms) and can demonstrate that the quality 
factor requirements are met by the form of the formula that is 
processed using the method that has the greatest potential for 
adversely affecting nutrient content and bioavailability;
    (iii) The manufacturer can demonstrate that the requirements of 
paragraph (a)(1) of this section are not appropriate for evaluation of 
a specific infant formula, and that an alternative method or study 
design for showing that the formula supports healthy growth in infants 
fed it as their sole source of nutrition is available.
    (b) Specific quality factor for protein quality of infant formula. 
(1) The manufacturer shall collect and maintain data that establish 
that the biological quality of protein in an infant formula is 
sufficient to meet the protein requirements of infants. The 
manufacturer shall establish the biological quality of the protein in 
its infant formula by demonstrating that the protein source supports 
adequate growth using the Protein Efficiency Ratio (PER) rat bioassay 
described in the ``Official Methods of Analysis of the Association of 
Official Analytical Chemists,'' 16th ed., sections 43.3.04 and 43.3.05, 
``AOAC Official Method 960.48 Protein Efficiency Ratio Rat Bioassay'' 
which is incorporated by reference in accordance with 5 U.S.C. 552(a) 
and 1 CFR part 51. Copies are available from the Association of 
Official Analytical Chemists, 481 North Frederick Ave., suite 500, 
Gaithersburg, MD 20857, or the Office of Special Nutritionals (HFS-
456), Center for Food Safety and Applied Nutrition, Food and Drug 
Administration, 200 C St. SW., Washington, DC 20204, or may be examined 
at the Office of the Federal Register, 800 North Capitol St. NW., 
Washington, DC. If the manufacturer is unable to conduct a PER rat 
bioassay because of the composition of the protein in the formula, then 
it shall demonstrate that the amino acid composition of the protein 
meets the known amino acid requirements of infants for whom the formula 
is intended.
    (2) The manufacturer may request an exemption from the requirements 
of paragraph (b)(1) of this section if:
    (i) The protein source, including any processing method used to 
produce the protein source, is already used in another infant formula 
marketed in the United States, manufactured by the same manufacturer, 
and the manufacturer can demonstrate that such infant formula meets the 
quality factor requirements prescribed in Sec. 106.96;
    (ii) The protein source, including any processing methods used to 
produce the protein source, is not a major change from the infant 
formula it replaces, and the manufacturer can demonstrate that the 
infant formula it replaces meets the

[[Page 36216]]

quality factor requirements prescribed in Sec. 106.96.
    6. In newly redesignated subpart F, Sec. 106.100 is amended by 
revising paragraphs (e), (f), (g), (j), and (k)(3), and by removing and 
reserving paragraph (h) to read as follows:


Sec. 106.100  Records.

* * * * *
    (e) Batch production and control records. For each batch of infant 
formula, manufacturers shall prepare and maintain records that include 
complete information relating to the production and control of the 
batch. These records shall include but are not limited to:
    (1) The master manufacturing order. The master manufacturing order 
shall include but is not limited to:
    (i) The significant steps in the production of the batch and the 
date on which each significant step occurred;
    (ii) The identity of equipment and processing lines used in 
producing the batch, if the plant in which the formula is made includes 
more than one set of equipment or more than one processing line;
    (iii) The identity of each batch or lot of ingredients, containers, 
and closures used in producing the batch of formula;
    (iv) The amount of each ingredient to be added to the batch of 
infant formula and a check (verification) that the correct amount was 
added; and
    (v) Copies of all labeling used and the results of examinations 
conducted during the finishing operations to provide assurance that 
containers and packages in the lot have the correct label.
    (2) Any deviations from the master manufacturing order and any 
corrective actions taken because of the deviations.
    (3) Documentation, in accordance with Sec. 106.6(c), of the 
monitoring at any point, step, or stage in their production process 
where control is deemed necessary to prevent adulteration. These 
records shall include, but not be limited to:
    (i) A list of the standards or specifications established at each 
point, step, or stage in their production process where control is 
deemed necessary to prevent adulteration including documentation of the 
scientific basis for each standard or specification;
    (ii) The actual values obtained during the monitoring operation, 
any deviations from established standards or specifications, and any 
corrective actions taken;
    (iii) Identification of the person monitoring each point, step, or 
stage in their production process where control is deemed necessary to 
prevent adulteration.
    (4) The conclusions and followup, along with the identity, of the 
individual qualified by training or experience who investigated:
    (i) Any deviation from the master manufacturing order and any 
corrective actions taken;
    (ii) A finding that a batch or any of its ingredients failed to 
meet the infant formula manufacturer's specifications; and
    (iii) A failure to meet any specification or standard at any point, 
step, or stage in the production process where control is deemed 
necessary to prevent adulteration.
    (5) The results of all testing performed on the batch of infant 
formula, including testing on the in-process batch, at the final-
product stage, and on finished product throughout the shelf life of the 
product. The results recorded shall include but are not limited to:
    (i) The results of all quality control testing conducted, in 
accordance with Sec. 106.91(a) and (b), to verify that each nutrient 
required by Sec. 107.100 of this chapter is present in each batch of 
infant formula at the level required by Sec. 107.100, and that any 
nutrient added by the manufacturer is present at the appropriate level 
with:
    (A) A summary table identifying the stages of the manufacturing 
process at which the nutrient analysis for each required nutrient under 
Sec. 106.91(a) is conducted, and
    (B) A summary table on the stability testing program, including the 
nutrients tested and the frequency of testing of nutrients throughout 
the shelf life of the product under Sec. 106.91(b); and
    (ii) For powdered infant formula, the results of any testing 
conducted in accordance with Sec. 106.55(b) to verify compliance with 
the microbiological quality standards in Sec. 106.55(c).
    (f) Manufacturers shall make and retain all records pertaining to 
current good manufacturing practice as described in subpart B of this 
part, including but not limited to:
    (1) Records, in accordance with Sec. 106.20(f)(3), of the frequency 
and results of testing of the water used in the production of infant 
formula;
    (2) Records, in accordance with Sec. 106.30(d), of accuracy checks 
of instruments and controls. A certification of accuracy of any known 
reference standard used and a history of recertification shall be 
maintained. At a minimum, such records shall specify the instrument or 
control being checked, the date of the accuracy check, the standard 
used, the calibration method used, the results found, any actions taken 
if the instrument is found to be out of calibration, and the initials 
or name of the individual performing the test. If calibration of an 
instrument (testing for accuracy against a known reference standard) 
shows that a specification or standard at a point, step, or stage in 
the production process where control is deemed necessary to prevent 
adulteration has not been met, a written evaluation of all affected 
product, and any actions that need to be taken with respect to that 
product, shall be made.
    (3) Records, in accordance with Sec. 106.30(e)(3)(ii), of the 
temperatures monitored for cold storage compartments and thermal 
processing equipment.
    (4) Records, in accordance with Sec. 106.30(f), on equipment 
cleaning, sanitizing, and maintenance that show the date and time of 
such cleaning, sanitizing, and maintenance and the lot number of each 
batch of infant formula processed between equipment startup and 
shutdown for cleaning, sanitizing, and maintenance. The person 
performing and checking the cleaning, sanitizing, and maintenance shall 
date and sign or initial the record indicating that the work was 
performed.
    (5) Records, in accordance with Sec. 106.35(c), on all automatic 
(mechanical or electronic) equipment used in the production or quality 
control of infant formula. These records shall include but not be 
limited to:
    (i) A list of all systems used with a description of computer files 
and the inherent limitations of each system;
    (ii) A copy of all software used;
    (iii) Records that document installation, calibration, testing or 
validation, and maintenance of the systems used;
    (iv) A list of all persons authorized to create or modify software;
    (v) Records that document modifications to software, including the 
identity of the person who modified the software;
    (vi) Records that document retesting or revalidation of modified 
systems; and
    (vii) A backup file of data entered into a computer or related 
system. The backup file shall consist of a hard copy or alternative 
system, such as duplicate diskettes, tapes, or microfilm, designed to 
ensure that backup data are exact and complete, and that they are 
secure from alteration, inadvertent erasures, or loss.
    (6) Records, in accordance with Sec. 106.40(g), on ingredients, 
containers, and closures used in the manufacture of infant formula. 
These records shall include, but are not limited to:

[[Page 36217]]

    (i) The identity and quantity of each lot of ingredients, 
containers, and closures;
    (ii) The name of the supplier;
    (iii) The supplier's lot numbers;
    (iv) The name and location of the manufacturer of the ingredient, 
container, and closure, if different from the supplier;
    (v) The date of receipt;
    (vi) The receiving code as specified; and
    (vii) The results of any test or examination (including retesting 
and reexamination) performed on the ingredients, containers, and 
closures and the conclusions derived therefrom and the disposition of 
all ingredients, containers, or closures.
    (7) A full description of the methodology used to test powdered 
infant formula to verify compliance with the microbiological quality 
standards of Sec. 106.55(c) and the methodology used to do quality 
control testing, in accordance with Sec. 106.91(a) and (b).
    (g) The manufacturer shall maintain all records pertaining to 
distribution of the infant formula, including records that show that 
products produced for export only are exported. Such records shall 
include, but not be limited to, all information and data necessary to 
effect and monitor recalls of the manufacturer's infant formula 
products in accordance with subpart E of part 107 of this chapter.
    (h) [Reserved]
* * * * *
    (j) The manufacturer shall make and retain records pertaining to 
regularly scheduled audits, including the audit plans and procedures, 
the findings of the audit, and a listing of any changes made in 
response to these findings. The manufacturer shall make readily 
available for authorized inspection the audit plans and procedures and 
a statement of assurance that the regularly scheduled audits are being 
conducted. The findings of the audit and any changes made in response 
to these findings shall be maintained for the time period required 
under Sec. 106.100(n), but need not be made available to FDA.
    (k) * * *
    (3) When there is a reasonable possibility of a causal relationship 
between the consumption of an infant formula and an infant's death, the 
manufacturer shall, within 15 days of receiving such information, 
conduct an investigation and notify the agency as required in 
Sec. 106.150.
* * * * *

Subpart G--Registration, Submission, and Notification Requirements


Sec. 106.110  New infant formula registration.

    (a) Before a new infant formula may be introduced or delivered for 
introduction into interstate commerce, the manufacturer of such formula 
shall register with the Food and Drug Administration, Center for Food 
Safety and Applied Nutrition, Office of Special Nutritionals, Division 
of Programs and Policy Enforcement (HFS-455), Infant Formula 
Coordinator, 200 C St. SW., Washington, DC 20204. An original and two 
copies of this registration shall be submitted.
    (b) The new infant formula registration shall include:
    (1) The name of the new infant formula,
    (2) The name of the manufacturer,
    (3) The place of business of the manufacturer, and
    (4) All establishments at which the manufacturer intends to 
manufacture such new infant formula.


Sec. 106.120  New infant formula submission.

    (a) At least 90 days before a new infant formula is introduced or 
delivered for introduction into interstate commerce, a manufacturer 
shall submit notice of its intent to do so to the Food and Drug 
Administration at the address given in Sec. 106.110(a). An original and 
two copies of the notice of its intent to do so shall be submitted.
    (b) The new infant formula submission shall include:
    (1) The name and physical form (e.g., powder, ready-to feed, or 
concentrate) of the infant formula;
    (2) An explanation of why the formula is a new infant formula;
    (3) The quantitative formulation of each form of the infant formula 
that is the subject of the notice in units per volume (for liquid 
formulas) or units per dry weight (for powdered formulas). When 
applicable, the submission shall include a description of any 
reformulation of the infant formula, including a listing of each new or 
changed ingredient and a discussion of the effect of such changes on 
the nutrient levels in the formulation;
    (4) A description, when applicable, of any change in processing of 
the infant formula. Such description shall identify the specific change 
in processing, including side-by-side, detailed schematic diagrams 
comparing the new processing to the previous processing (including 
processing times and temperatures);
    (5) Assurance that the infant formula will not be marketed unless 
the formula meets the quality factor requirements of section 412(b)(1) 
of the Federal Food, Drug, and Cosmetic Act (the act) and the nutrient 
content requirements of section 412(i) of the act.
    (i) Assurance that the formula meets the quality factor 
requirements, which are set forth in subpart E of this part, shall be 
provided by a submission that complies with Sec. 106.121.
    (ii) Assurance that the formula complies with the nutrient content 
requirements, which are set forth in Sec. 107.100 of this chapter, 
shall be provided by a statement assuring that the formula will not be 
marketed unless it meets the nutrient requirements of Sec. 107.100 of 
this chapter, as demonstrated by testing required under subpart C of 
this part;
    (6) Assurance that the processing of the infant formula complies 
with section 412(b)(2) of the act. Such assurance shall include but not 
be limited to:
    (i) A statement that the formula will be produced in accordance 
with subparts B and C of this part;
    (ii) The basis on which each ingredient meets the requirements of 
Sec. 106.40(a), e.g., that it is an approved food additive, that it is 
authorized by a prior sanction issued by the agency, or that it is GRAS 
for its intended use. Any claim that an ingredient is GRAS shall be 
supported by a citation to the agency's regulations or by an 
explanation, including a list of published studies and a copy of those 
publications, for why, based on the published studies, there is general 
recognition of the safety of the use of the ingredient in infant 
formula.
    (c) For products for export only, a manufacturer may submit, in 
lieu of the information required under paragraph (b) of this section, a 
statement that the infant formula meets the specifications of the 
foreign purchaser, does not conflict with the laws of the country to 
which it is intended for export, is labeled on the outside of the 
shipping package to indicate that it is intended for export only, and 
will not be sold or offered for sale in domestic commerce.
    (d) The submission will not constitute notice under section 412 of 
the act unless it complies fully with paragraph (b) of this section, 
and the information that it contains is set forth in a manner that is 
readily understandable. The agency will notify the submitter if the 
notice is not adequate because it does not meet the requirements of 
section 412(c) and (d) of the act.
    (e) If a new infant formula submission is adequate, FDA will 
acknowledge its receipt and notify the manufacturer of the date of 
receipt. The date that the agency receives the new infant formula

[[Page 36218]]

submission is the filing date for the submission. The manufacturer 
shall not market the new infant formula before the date that is 90 days 
after the filing date.
    (f) If the manufacturer provides additional information in support 
of a new infant formula submission, the agency will determine whether 
the additional information is a substantive amendment to the new infant 
formula submission. If the agency determines that the new submission is 
a substantive amendment, FDA will assign the new infant formula 
submission a new filing date. FDA will acknowledge receipt of the 
additional information and, when applicable, notify the manufacturer of 
the new filing date, which is the date of receipt by FDA of the 
information that constitutes the substantive amendment to the new 
infant formula submission.


Sec. 106.121  Quality factor submission.

    To provide assurance that an infant formula meets the quality 
factor requirements set forth in subpart E of this part, the 
manufacturer shall submit the following data and information:
    (a) An explanation, in narrative form, setting forth how all 
quality factor requirements of subpart E of this part have been met.
    (b) Records that contain the information required by proposed 
Sec. 106.97 (a)(1)(i) and (a)(1)(ii) collected during the study for 
each infant enrolled in the study. The records shall be identified by 
subject number, age, feeding group, gender, and study day of 
collection.
    (c)(1) Statistical evaluation for all measurements, including: 
Group means, group standard deviations, and measures of statistical 
significance for all measurements for each feeding group at the 
beginning of the study and at every point where measurements were made 
throughout the study.
    (2) Calculation of the statistical power of the study at its 
completion.
    (d) A report on attrition and on all occurrences of adverse events 
during the study, which shall include:
    (1) Identification of the infant by subject number and feeding 
group and a complete description of the adverse event, including 
comparisons of the frequency and nature of occurrence in each feeding 
group and information on the health of the infant during the course of 
the study, including the occurrence and duration of any illness;
    (2) A clinical assessment, by a health care provider, of the 
infant's health during each suspected adverse event;
    (3) A complete listing of all infants who did not complete the 
study, including the infant's subject number and the reason that each 
infant left the study.
    (e) The results of the Protein Efficiency Ratio, in accordance with 
Sec. 106.97(b).
    (f) A statement certifying that the manufacturer has collected and 
considered all information and data concerning the ability of the 
infant formula to meet the quality factor requirements, and that the 
manufacturer is not aware of any information or data that would show 
that the formula does not meet the quality factors requirements.


Sec. 106.130  Verification submission.

    (a) Manufacturers shall, after the first production and before the 
introduction into interstate commerce of the new infant formula, verify 
in a written submission to FDA at the address given in Sec. 106.110(a), 
that the infant formula complies with the requirements of the Federal 
Food, Drug, and Cosmetic Act (the act) and is not adulterated. An 
original and two copies of this verification shall be submitted.
    (b) The verification submission shall include the following 
information:
    (1) The name of the new infant formula; the filing date for the new 
infant formula submission, in accordance with Sec. 106.120, for the 
subject formula; and the identification number assigned by the agency 
to the new infant formula submission;
    (2) A statement that the infant formula to be introduced into 
interstate commerce is the same as the infant formula that was the 
subject of the new infant formula notification and for which the 
manufacturer provided assurances in accordance with the requirements of 
Sec. 106.120;
    (3) A summary of test results of the level of each nutrient 
required by Sec. 107.100 of this chapter and any nutrient added by the 
manufacturer in the formula, presented in units per 100 kilocalories at 
the final-product-stage.
    (4) A certification that the manufacturer has established current 
good manufacturing practices including quality control procedures and 
in-process controls, including testing required by current good 
manufacturing practice, designed to prevent adulteration of this 
formula in accordance with subparts B and C of this part.
    (c) The submission will not constitute written verification under 
section 412(d)(2) of the act when any data prescribed in paragraph (b) 
of this section are lacking or are not set forth so as to be readily 
understood. In such circumstances the agency will notify the submitter 
that the notice is not adequate.


Sec. 106.140  Submission concerning a change in infant formula that may 
adulterate the product.

    (a) When a manufacturer makes a change in the formulation or 
processing of the formula that may affect whether the formula is 
adulterated under section 412(a) of the Federal Food, Drug, and 
Cosmetic Act (the act), it shall, before the first processing of such 
formula, make a submission to the Food and Drug Administration at the 
address given in Sec. 106.110(a). An original and two copies shall be 
submitted.
    (b) The submission shall include:
    (1) The name and physical form of the infant formula (i.e., powder, 
ready-to-feed, or concentrate);
    (2) An explanation of why the change in formulation or processing 
may affect whether the formula is adulterated; and
    (3) A submission that complies with Sec. 106.120(b)(3), (b)(4), 
(b)(5), and (b)(6). When appropriate, a statement to the effect that 
the information required by Sec. 106.120(b)(3), (b)(4), (b)(5), or 
(b)(6) has been provided to the agency previously and has not been 
affected by the changes that is the subject of this submission, 
together with the identification number assigned by the agency to the 
relevant infant formula submission, may be provided in lieu of such 
submission.
    (c) The submission will not constitute notice under section 412 of 
the act unless it complies fully with paragraph (b) of this section, 
and the information that it contains is set forth in a manner that is 
readily understandable. The agency will notify the submitter if the 
notice is not adequate because it does not meet the requirements of 
section 412(d)(3) of the act.


Sec. 106.150  Notification of an adulterated or misbranded infant 
formula.

    (a) A manufacturer shall promptly notify FDA in accordance with 
paragraph (b) of this section, when the manufacturer has knowledge 
(that is, the actual knowledge that the manufacturer had, or the 
knowledge which a reasonable person would have had under like 
circumstances or which would have been obtained upon the exercise of 
due care) that reasonably supports the conclusion that an infant 
formula that has been processed by the manufacturer and that has left 
an establishment subject to the control of the manufacturer:
    (1) May not provide the nutrients required by section 412(i) of the 
act or by regulations issued under section 412(i)(2); or

[[Page 36219]]

    (2) May be otherwise adulterated or misbranded.
    (b) The notification made according to paragraph (a) of this 
section shall be made by telephone, to the Director of the appropriate 
Food and Drug Administration district office. After normal business 
hours (8 a.m. to 4:30 p.m.), FDA's emergency number, 202-857-8400, 
shall be used. The manufacturer shall send written confirmation of the 
notification to the Food and Drug Administration, Center for Food 
Safety and Applied Nutrition, Office of Special Nutritionals, Division 
of Programs and Policy Enforcement (HFS-455), Infant Formula 
Coordinator, 200 C St. SW., Washington, DC 20204, and to the 
appropriate Food and Drug Administration district office specified in 
Sec. 5.115 of this chapter.

PART 107--INFANT FORMULA

    7. The authority citation for 21 CFR part 107 continues to read as 
follows:

    Authority: Secs. 201, 403, 412, 701 of the Federal Food, Drug, 
and Cosmetic Act (21 U.S.C. 321, 343, 350a, 371).

    8. Section 107.1 is added to subpart A to read as follows:


Sec. 107.1  Status and applicability of the regulations in part 107.

    (a) The criteria set forth in subpart B of this part describes the 
labeling requirements applicable to infant formula under section 403 of 
the Federal Food, Drug, and Cosmetic Act (the act). Failure to comply 
with any regulation in subpart B of this part will render an infant 
formula misbranded under that section of the act.
    (b) The criteria set forth in subpart C of this part describes the 
terms and conditions for the exemption of an infant formula from the 
requirements of section 412(a), (b), and (c) of the act. Failure to 
comply with any regulations in subpart C of this part will result in 
the withdrawal of the exemption given under section 412(h)(1) of the 
act.
    (c) Subpart D of this part sets forth the nutrient requirements for 
infant formula under section 412(i) of the act. Failure to comply with 
any regulation in subpart D of this part will render an infant formula 
adulterated under section 412(a)(1) of the act.
    9. Section 107.10 is amended by revising the introductory text of 
paragraph (a)(2) to read as follows:


Sec. 107.10   Nutrient information.

    (a) * * *
    (2) A statement of the amount, supplied by 100 kilocalories, of 
each of the following nutrients and of any nutrient added by the 
manufacturer:
* * * * *
    10. Section 107.240 is revised to read as follows:


Sec. 107.240   Notification requirements.

    (a) Telephone report. When a determination is made that an infant 
formula is to be recalled, the recalling firm shall telephone within 24 
hours the appropriate Food and Drug Administration district office 
listed in Sec. 5.115 of this chapter and shall provide relevant 
information about the infant formula that is to be recalled.
    (b) Initial written report. Within 14 days after the recall has 
begun, the recalling firm shall provide a written report to the 
appropriate Food and Drug Administration district office. The report 
shall contain relevant information, including the following cumulative 
information concerning the infant formula that is being recalled:
    (1) Number of consignees notified of the recall and date and method 
of notification, including recalls required by Sec. 107.200, 
information about the notice provided for retail display and the 
request for its display.
    (2) Number of consignees responding to the recall communication and 
quantity of recalled infant formula on hand at the time it was 
received.
    (3) Quantity of recalled infant formula returned or corrected by 
each consignee contacted and the quantity of recalled infant formula 
accounted for.
    (4) Number and results of effectiveness checks that were made.
    (5) Estimated timeframes for completion of the recall.
    (c) Status reports. The recalling firm shall submit to the 
appropriate Food and Drug Administration district office a written 
status report on the recall at least every 14 days until the recall is 
terminated. The status report shall describe the steps taken by the 
recalling firm to carry out the recall since the last report and the 
results of these steps.
    11. Section 107.250 is amended by revising the introductory text to 
read as follows:


Sec. 107.250   Termination of an infant formula recall.

    The recalling firm may submit a recommendation for termination of 
the recall to the appropriate Food and Drug Administration district 
office listed in Sec. 5.115 of this chapter for transmittal to the 
Division of Enforcement (HFS-605), Office of Field Programs, Center for 
Food Safety and Applied Nutrition, for action. Any such recommendation 
shall contain information supporting a conclusion that the recall 
strategy has been effective. The agency will respond within 15 days of 
receipt by the Division of Enforcement (HFS-605), Office of Field 
Programs, Center for Food Safety and Applied Nutrition, of the request 
for termination. The recalling firm shall continue to implement the 
recall strategy until it receives final written notification from the 
agency that the recall has been terminated. The agency will send such 
notification, unless it has information, from FDA's own audits or from 
other sources demonstrating the recall has not been effective. The 
agency may conclude that a recall has not been effective if:

* * * * *
    Dated: April 19, 1996.
David A. Kessler,
Commissioner of Food and Drugs.
Donna E. Shalala,
Secretary of Health and Human Services.
[FR Doc. 96-17058 Filed 7-8-96; 8:45 am]
BILLING CODE 4160-01-P