[Federal Register Volume 81, Number 70 (Tuesday, April 12, 2016)]
[Notices]
[Pages 21559-21573]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2016-08327]


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

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Food and Drug Administration

[Docket No. FDA-2016-N-0832]


Phibro Animal Health Corp.; Carbadox in Medicated Swine Feed; 
Opportunity for Hearing

AGENCY: Food and Drug Administration, HHS.

ACTION: Notice of opportunity for hearing.

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

SUMMARY: The Food and Drug Administration (FDA), Center for Veterinary 
Medicine (CVM), is proposing to withdraw approval of all new animal 
drug applications (NADAs) providing for use of carbadox in medicated 
swine feed. This action is based on CVM's determination that the use of 
carbadox under the approved conditions of use results in residues of 
carcinogenic concern in the edible tissues of the treated swine.

DATES: Phibro Animal Health Corp. may submit a request for a hearing by 
May 12, 2016. Submit all data and analysis upon which the request for a 
hearing relies by July 11, 2016.

ADDRESSES: The request for a hearing may be submitted by Phibro Animal 
Health Corp. by either of the following methods:

Electronic Submission

     Federal eRulemaking Portal: http://www.regulations.gov. 
Follow the instructions for submitting comments to submit your request 
for hearing. Your request for a hearing submitted electronically, 
including any attachments to the request for hearing, to http://www.regulations.gov will be posted to the docket unchanged.

Written/Paper Submission

     Mail/Hand delivery/Courier (for written/paper request for 
a hearing): Division of Dockets Management (HFA-305), Food and Drug 
Administration, 5630 Fishers Lane, Rm. 1061, Rockville, MD 20852.
    Because your request for a hearing will be made public, you are 
solely responsible for ensuring that your request does not include any 
confidential information that you may not wish to be publicly posted, 
such as confidential business information, e.g., a manufacturing 
process. The request for a hearing must include the Docket No. FDA-
2016-N-0832 for ``Phibro Animal Health Corp.; Carbadox in Medicated 
Swine Feed; Opportunity for Hearing.'' The request for a hearing will 
be placed in the docket and publicly viewable at http://www.regulations.gov or at the Division of Dockets Management between 9 
a.m. and 4 p.m., Monday through Friday.
    Phibro Animal Health Corp. may submit all data and analysis upon 
which the request for a hearing relies in the same manner as the 
request for a hearing except as follows:
     Confidential Submissions--To submit any data and analyses 
with confidential information that you do not wish to be made publicly 
available, submit your data and analyses only as a written/paper 
submission. You should submit two copies total of all data and 
analysis. One copy will include the information you claim to be 
confidential with a heading or cover note that states ``THIS DOCUMENT 
CONTAINS CONFIDENTIAL INFORMATION.'' The Agency will review this copy, 
including the claimed confidential information, in its consideration of 
any decisions on this matter. The second copy, which will have the 
claimed confidential

[[Page 21560]]

information redacted/blacked out, will be available for public viewing 
and posted on http://www.regulations.gov or available at the Division 
of Dockets Management between 9 a.m. and 4 p.m., Monday through Friday. 
Submit both copies to the Division of Dockets Management. Any 
information marked as ``confidential'' will not be disclosed except in 
accordance with 21 CFR 10.20 and other applicable disclosure law.
    Comments Submitted by Other Interested Parties: For all comments 
submitted by other interested parties you may submit comments as 
follows:

Electronic Submissions

    Submit electronic comments in the following way:
     Federal eRulemaking Portal: http://www.regulations.gov. 
Follow the instructions for submitting comments. Comments submitted 
electronically, including attachments, to http://www.regulations.gov 
will be posted to the docket unchanged. Because your comment will be 
made public, you are solely responsible for ensuring that your comment 
does not include any confidential information that you or a third party 
may not wish to be posted, such as medical information, your or anyone 
else's Social Security number, or confidential business information, 
such as a manufacturing process. Please note that if you include your 
name, contact information, or other information that identifies you in 
the body of your comments, that information will be posted on http://www.regulations.gov.
     If you want to submit a comment with confidential 
information that you do not wish to be made available to the public, 
submit the comment as a written/paper submission and in the manner 
detailed (see ``Written/Paper Submissions'' and ``Instructions'').

Written/Paper Submissions

    Submit written/paper submissions as follows:
     Mail/Hand delivery/Courier (for written/paper 
submissions): Division of Dockets Management (HFA-305), Food and Drug 
Administration, 5630 Fishers Lane, Rm. 1061, Rockville, MD 20852.
     For written/paper comments submitted to the Division of 
Dockets Management, FDA will post your comment, as well as any 
attachments, except for information submitted, marked and identified, 
as confidential, if submitted as detailed in ``Instructions.''
    Instructions: All submissions received must include the Docket No. 
FDA-2016-N-0832 for ``Phibro Animal Health Corp.; Carbadox in Medicated 
Swine Feed; Opportunity for Hearing.'' Received comments will be placed 
in the docket and, except for those submitted as ``Confidential 
Submissions,'' publicly viewable at http://www.regulations.gov or at 
the Division of Dockets Management between 9 a.m. and 4 p.m., Monday 
through Friday.
     Confidential Submissions--To submit a comment with 
confidential information that you do not wish to be made publicly 
available, submit your comments only as a written/paper submission. You 
should submit two copies total. One copy will include the information 
you claim to be confidential with a heading or cover note that states 
``THIS DOCUMENT CONTAINS CONFIDENTIAL INFORMATION.'' The Agency will 
review this copy, including the claimed confidential information, in 
its consideration of comments. The second copy, which will have the 
claimed confidential information redacted/blacked out, will be 
available for public viewing and posted on http://www.regulations.gov. 
Submit both copies to the Division of Dockets Management. If you do not 
wish your name and contact information to be made publicly available, 
you can provide this information on the cover sheet and not in the body 
of your comments and you must identify this information as 
``confidential.'' Any information marked as ``confidential'' will not 
be disclosed except in accordance with 21 CFR 10.20 and other 
applicable disclosure law. For more information about FDA's posting of 
comments to public dockets, see 80 FR 56469, September 18, 2015, or 
access the information at: http://www.fda.gov/regulatoryinformation/dockets/default.htm.
    Docket: For access to the docket to read background documents or 
the electronic and written/paper comments received, go to http://www.regulations.gov and insert the docket number, found in brackets in 
the heading of this document, into the ``Search'' box and follow the 
prompts and/or go to the Division of Dockets Management, 5630 Fishers 
Lane, Rm. 1061, Rockville, MD 20852.

FOR FURTHER INFORMATION CONTACT: Vernon Toelle, Center for Veterinary 
Medicine (HFV-230), 7519 Standish Pl., Rockville, MD 20855, 240-276-
9200.

SUPPLEMENTARY INFORMATION: 

I. Approved NADAs for Use of Carbadox in Swine Feed

    Carbadox, a quinoxaline derivative, is a synthetic organic acid 
antimicrobial. Currently, there are three approved NADAs for use of 
carbadox in medicated swine feed, either by itself or in combination 
with other approved new animal drugs. Phibro Animal Health Corp. 
(Phibro), 65 Challenger Rd., Ridgefield Park, NJ 07660, is currently 
the sponsor of all three approved NADAs.
    Carbadox is marketed as a Type A medicated article used to 
manufacture complete Type C medicated feeds that are administered ad 
libitum to swine. Carbadox is indicated for the control of dysentery 
and bacterial enteritis, and for growth promotion. A tolerance of 30 
parts per billion (ppb) \1\ has been established for residues of 
quinoxaline-2-carboxylic acid (QCA), the marker residue, in liver of 
swine (21 CFR 556.100).
---------------------------------------------------------------------------

    \1\ For consistency and readability throughout this document, 
concentrations are reported as parts per billion even though 
original references may report some concentrations as parts per 
trillion (ppt).
---------------------------------------------------------------------------

    The following three NADAs are approved for the use of carbadox:
    NADA 041-061, originally approved in 1972 (37 FR 20683, October 3, 
1972), provides for the use of MECADOX 10 (carbadox) Type A medicated 
article to manufacture single-ingredient Type C medicated swine feeds 
for the following conditions of use:
     Carbadox at 10 to 25 grams per ton (g/ton) of feed for 
increased rate of weight gain and improved feed efficiency; and
     Carbadox at 50 g/ton of feed for control of swine 
dysentery (vibrionic dysentery, bloody scours, or hemorrhagic 
dysentery); for control of bacterial swine enteritis (salmonellosis or 
necrotic enteritis caused by Salmonella choleraesuis); and for 
increased rate of weight gain and improved feed efficiency.
    Currently, the withdrawal period for these uses of carbadox is 42 
days (Sec.  558.115(d)(1)(ii) and (d)(2)(ii) (21 CFR 558.115(d)(1)(ii) 
and (d)(2)(ii))).
    NADA 092-955, originally approved in 1975 (40 FR 45164, October 1, 
1975), provides for the use of MECADOX 10 (carbadox) Type A medicated 
article with BANMINTH (pyrantel tartrate) Type A medicated article to 
manufacture two-way, combination drug Type C medicated swine feeds for 
the following conditions of use:
     Carbadox at 50 g/ton of feed plus pyrantel tartrate at 96 
g/ton of feed for control of swine dysentery (vibrionic dysentery, 
bloody scours, or hemorrhagic dysentery); for control of bacterial 
swine enteritis (salmonellosis or necrotic enteritis caused by 
Salmonella choleraesuis); as an aid in the prevention of migration and 
establishment of large roundworm

[[Page 21561]]

(Ascaris suum) infections; and as an aid in the prevention of 
establishment of nodular worm (Oesophagostomum) infections.
    The withdrawal period for the use of this drug combination is 70 
days (Sec.  558.115(d)(3)(ii)).
    NADA 141-211, originally approved in 2004 (69 FR 51173, August 18, 
2004), provides for the use of MECADOX 10 (carbadox) Type A medicated 
article with TERRAMYCIN 50, TERRAMYCIN 100, or TERRAMYCIN 200 
(oxytetracycline) Type A medicated articles to manufacture two-way, 
combination drug Type C medicated swine feeds for the following 
conditions of use:
     Carbadox at 10 to 25 g/ton of feed plus oxytetracycline at 
levels in feed to deliver 10 mg carbadox per pound of body weight for 
treatment of bacterial enteritis caused by Escherichia coli and S. 
choleraesuis susceptible to oxytetracycline; for treatment of bacterial 
pneumonia caused by Pasteurella multocida susceptible to 
oxytetracycline; and for increased rate of weight gain and improved 
feed efficiency.
    The withdrawal period for the use of this animal drug combination 
is 42 days (Sec.  558.115(d)(4)(ii)).

II. Basis for Withdrawal of Approval

    CVM is providing notice of an opportunity for a hearing (NOOH) on a 
proposal to withdraw approval of the NADAs providing for use of 
carbadox in medicated swine feeds. New evidence regarding carcinogenic 
residues in edible tissues of swine treated with carbadox raises 
serious questions about the human food safety of the drug. Grounds for 
withdrawing carbadox are twofold. First, new evidence demonstrates that 
the Delaney Clause in section 512(d) of the Federal Food, Drug, and 
Cosmetic Act (the FD&C Act) (21 U.S.C. 360b), which requires that no 
residue of a carcinogenic drug can be found in any edible portion of 
the animal after slaughter, applies because the Diethylstilbestrol 
(DES) Proviso exception is no longer met (see, Section III.C). Second, 
new evidence demonstrates that carbadox is not shown to be safe under 
the General Safety Clause (section 512(e)(1)(B) of the FD&C Act).
    During the review of a supplemental application to NADA 041-061 
approved in January 1998, CVM made the following conclusions about the 
drug: (1) The parent compound carbadox is rapidly metabolized and 
carcinogenic residues of the drug are not identifiable in any edible 
tissues beyond 72 hours post dosing; (2) remaining unextracted residues 
of carbadox are noncarcinogenic residues related to the noncarcinogenic 
metabolite QCA; and (3) QCA is a reliable marker residue for carbadox 
and its metabolites (Ref. 1).
    Since the evaluation of information submitted by the sponsor in 
that supplemental application, CVM has become aware of new information 
that calls into question the basis for its previous conclusions. As 
described more fully in Section V., this includes new residue depletion 
data presented to the Joint FAO/WHO Expert Committee on Food Additives 
(JECFA) \2\ in 2003 that shows that when the marker residue QCA reaches 
the approved tolerance of 30 ppb in liver, concentrations of the 
carcinogen desoxycarbadox (DCBX) in the liver would be approximately 4 
times higher than the concentration that would be considered safe (Ref. 
2 at pp. 16-17). In addition, the new residue depletion data presented 
to JECFA in 2003 call into question CVM's previously held conclusion 
that the unextracted residues of carbadox at the withdrawal period are 
noncarcinogenic compounds related to the QCA metabolite (Ref. 1). The 
Agency treats the unidentified residues--metabolites of a carcinogenic 
parent drug with demonstrated carcinogenic metabolites--as 
carcinogenic. Therefore, the drug is not shown to be safe under the 
General Safety Clause and the Delaney Clause applies to the drug, 
because the DES Proviso exception is no longer met.
---------------------------------------------------------------------------

    \2\ JECFA is an independent committee of international 
scientific experts administered jointly by the United Nations Food 
and Agriculture Organization (FAO) and the World Health Organization 
(WHO) for the purpose of providing independent scientific advice to 
the FAO, WHO, and member countries. It has been meeting since 1956 
specifically to evaluate the safety of food additives, including the 
animal drug residues in edible tissues. See http://www.codexalimentarius.org/scientific-basis-for-codex/jecfa/en/ and 
http://www.who.int/foodsafety/areas_work/chemical-risks/jecfa/en/.
---------------------------------------------------------------------------

    Continued approval of carbadox would expose humans to 
concentrations of total residues of carcinogenic concern that are 
approximately 30 times higher (for the approved 42-day withdrawal 
period) or 11 times higher (for the approved 70-day withdrawal period) 
than the 0.915 ppb concentration of total residues of carcinogenic 
concern in liver that would be considered safe (Ref. 3 at p. 17, Table 
8). Moreover, the sponsor has not identified an appropriate marker and 
analytical method to assure that residues of carcinogenic concern are 
below the level at which the residues present in the total human diet 
present no significant increase in the risk of cancer to people (the 
So).
    In addition to the new information presented to JECFA (Ref. 2), 
publications by Boison, et al., in 2009 (Ref. 4) and Baars, et al., in 
1990 (Ref. 5) that were recently provided to CVM by the sponsor call 
into question the previous conclusion that QCA is an appropriate marker 
and that all residues of carcinogenic concern deplete within 72 hours 
after dosing.
    The new evidence from the 2003 JECFA report (Ref. 2) in conjunction 
with the publications by Boison, et al., in 2009 (Ref. 4) and Baars, et 
al., in 1991 (Ref. 6), erode the scientific justification for, and 
validity of, conclusions previously made about the drug in 1998. Based 
on this new information, evaluated together with the information 
available at the time of the approvals, CVM has determined that the 
drug is not shown to be safe under the General Safety Clause and that 
the Delaney Clause applies to the drug, because the DES Proviso 
exception is no longer met. Therefore, CVM proposes to withdraw 
approval of all NADAs for new animal drugs containing carbadox.

III. Legal Context of the Proposed Action and Grounds for Withdrawal

A. The Determination of Safety in Section 512

    Carbadox, for each of its uses in swine, is a new animal drug as 
defined in section 201(v) of the FD&C Act (21 U.S.C. 321(v)). As such, 
under sections 301, 501, 512, 571, and 572 of the FD&C Act (21 U.S.C. 
331, 351, 360b, 360ccc, 360ccc-1), the drug cannot be legally 
introduced or delivered for introduction into interstate commerce in 
the absence of an NADA approval, a conditional approval, or an animal 
drug indexing. The requirements for approval of an NADA are set out in 
section 512(d)(2)(A) of the FD&C Act. Section 512(b)(1)(A) of the FD&C 
Act requires that a new animal drug must be shown to be safe and 
effective for its intended uses. Section 201(u) of the FD&C Act 
provides that ``safe'' as used in section 512 of the FD&C Act ``has 
reference to the health of man or animal.'' The determination of safety 
requires CVM to consider, among other relevant factors, ``the probable 
consumption of such drug and any substance formed in or on food because 
of the use of such drug . . .'' (section 512(d)(2)(A) of the FD&C Act). 
Accordingly, CVM must consider not only safety of the new animal drug 
to the target animal, but also the safety to humans of substances 
formed in or on food as a result of the use of the new animal drug.

[[Page 21562]]

    ``Safe,'' in the context of human food safety, means a ``reasonable 
certainty of no harm.'' The definition is derived from language in H. 
Rep. No. 85-2284, at 4-5 (1958), defining the term ``safe'' as it 
appears in section 409 of the FD&C Act, which governs food additives 
(21 U.S.C. 348). Until passage of the Animal Drug Amendments of 1968 
(Pub. L. 90-399) (the 1968 amendments), substances formed in or on food 
due to the use of animal drugs in food-producing animals were regulated 
under the food additive provisions in section 409 of the FD&C Act. The 
1968 amendments consolidated all of the existing statutory authorities 
related to animal drugs into section 512 of the FD&C Act, and the 
legislative history shows that the consolidation in no way changed the 
authorities with respect to the regulation of new animal drugs (S. Rep. 
No. 90-1308, at 1 (1968)). During the new animal drug application 
review process, CVM has consistently applied the ``reasonable certainty 
of no harm'' standard in determining the safety of substances formed in 
or on food as a result of the use of a new animal drug in a food-
producing animal.
    In order to determine whether a new animal drug meets this 
standard, section 512(b)(1)(G)-(H) of the FD&C Act requires that 
whenever a drug may result in residues of the drug or its metabolites 
in food, an application must include not only full reports of 
investigations to show that the use of the drug is safe, but also a 
description of practicable methods for monitoring food to assure that 
there are no unsafe residues in human food attributable to the drug 
use, and a demonstration that the conditions of use are adequate to 
assure there are no unsafe residues.
    In sum, under section 512(d)(2) of the FD&C Act, the Agency is 
required, in the evaluation of the supporting safety data, among other 
things, to consider:
     The probable consumption of such drug and of any substance 
formed in or on food because of the use of such drug (i.e., probable 
human consumption of residues including the parent drug and its 
metabolites);
     The cumulative effect on man or animal of such drug, 
taking into account any chemically or pharmacologically related 
substance, i.e., toxicological effects of the compounds comprising the 
residues; and
     Safety factors which, in the opinion of experts qualified 
by scientific training and experience to evaluate the safety of such 
drugs, are appropriate for the use of animal experimentation data 
(i.e., establishing ``safe'' levels of residues using appropriate 
safety factors to extrapolate animal data on cumulative effects to 
humans).
    When establishing the human food safety of a noncarcinogenic new 
animal drug used in food-producing animals, CVM establishes a no 
observed effect level (NOEL) for the residues of that drug in edible 
tissues--namely, the highest dose of the drug that does not produce the 
most sensitive treatment-related toxic endpoint in test animals (Ref. 
7). From the NOEL, CVM uses safety factors to calculate an acceptable 
daily intake, and consumption factors to calculate the safe 
concentration of residues in a particular edible tissue (Ref. 7 at p. 
15; section 512(b)(1)(H) of the FD&C Act).
    Carbadox is both a genotoxic \3\ and mutagenic carcinogen in 
animals. In the case of a genotoxic carcinogenic drug, establishing the 
human food safety of the compound via a NOEL is not feasible, therefore 
human food safety of carcinogenic compounds is ordinarily evaluated by 
using linear, low-dose extrapolation to evaluate the maximum 
concentration of total residues of carcinogenic concern that can be 
present in the total human diet without a significant increase in the 
risk of cancer to the human consumer (section 512(d)(1)(I) of the FD&C 
Act; 21 CFR 500.82 and 500.84). In both cases, the safe residue level 
of the drug is determined through an evaluation of the relevant data 
relating to the three factors listed above; viz., the probable 
consumption of the drug residue and its cumulative effect as determined 
through all relevant safety factors (section 512(d)(2) of the FD&C 
Act).
---------------------------------------------------------------------------

    \3\ Genotoxic refers to chemicals that react with DNA or 
chromosomes to cause damage. When the damage is not repaired and the 
effect is a heritable change (cell to cell or parent to offspring), 
it is also termed mutagenic. Thus not all genotoxic chemicals are 
mutagenic, but all mutagenic chemicals are genotoxic. Uncorrected 
mutagenesis is thought to be a key step in the development of 
cancer. ``Mechanisms of Toxicity,'' in Casarett & Doull's 
Toxicology: The Basic Science of Poisons, edited by Klassen, C.D., 
8th Ed., pp. 49-123, 2013.
---------------------------------------------------------------------------

B. Grounds for Withdrawal Under the FD&C Act

    Section 512(e)(1)(B) of the FD&C Act provides grounds for 
withdrawal of approval of an NADA if new evidence not contained in an 
approved application or not available to the Secretary of Health and 
Human Services until after such application was approved, or tests by 
new methods, or tests by methods not deemed reasonably applicable when 
such application was approved, evaluated together with the evidence 
available to the Secretary when the application was approved, shows 
that such drug is not shown to be safe for use under the conditions of 
use upon the basis of which the application was approved or that 
subparagraph (I) of paragraph (1) of subsection (d) applies to such 
drug. The Secretary of Health and Human Services has delegated this 
authority to the Commissioner of Food and Drugs. See FDA Staff Manual 
Guide 1410.10 (April 11, 2014).
    In other words, grounds for withdrawal exist where new evidence 
shows either that the Delaney Clause applies to the drug 
(``subparagraph (I) of paragraph (1) of subsection (d)'') or that the 
drug is not shown to be safe under the approved conditions of use (the 
General Safety Clause). As explained further, new evidence demonstrates 
that carbadox meets both grounds for withdrawal.
    In a proceeding to withdraw the approval of an NADA, the sponsor 
has the burden of proof to demonstrate that the product is safe and 
therefore that the NADA approval should remain in effect (21 CFR 
12.87(d): (``At a hearing involving issuing, amending, or revoking a 
regulation or order relating to the safety or effectiveness of a drug . 
. . the participant who is contending that the product is safe or 
effective or both and who is . . . contesting withdrawal of approval 
has the burden of proof in establishing safety or effectiveness or both 
and thus the right to approval.''); (see also Rhone-Poulenc, Inc. v. 
FDA, 636 F.2d 750, 752 (D.C. Cir. 1980); Hess & Clark v. FDA, 495 F.2d 
975, 992 (D.C. Cir. 1974)). Nevertheless, CVM bears an initial burden 
of showing that new evidence regarding the new animal drug raises 
serious questions about the safety of the new animal drug. See Rhone-
Poulenc, 636 F.2d at 752. Once CVM has satisfied the initial burden, 
the burden shifts to the sponsor to establish the safety of the drug:

    In the Hess & Clark case we held that the ``new evidence'' 
requirement of the safety clause ``plainly places on the [CVM] an 
initial burden to adduce the `new evidence' and what that evidence 
`shows'. Only when the [CVM] has met this initial burden of coming 
forward with the new evidence is there a burden on the manufacturer 
to show that the drug is safe.'' Rhone-Poulenc, 636 F.2d at 752 
(quoting Hess & Clark, 495 F.2d at 992).

    To meet its initial burden of proof to withdraw approval of a new 
animal drug that is ``not shown to be safe,'' CVM must provide ``a 
reasonable basis from which serious questions about the ultimate safety 
of [the drug] and the residues that may result from its use may be 
inferred.'' See Diethylstilbestrol: Withdrawal of Approval of New 
Animal Drug Applications; Commissioner's Decision (44 FR 54852 at 
54861,

[[Page 21563]]

September 21, 1979) (hereinafter DES Commissioner Decision) (quoting 
Proposal to Withdraw Approval of New Animal Drug Applications for 
Diethylstilbestrol, ALJ Initial Decision, Docket No. FDA-1976-N-0028 
(formerly 76N-0002), I.D. at 8 (September 21, 1978)), aff'd Rhone-
Poulenc, 636 F.2d 750; see also Nitrofurans Commissioner Decision (56 
FR 41902 at 41902, August 23, 1991). Serious questions can be raised 
where the evidence is not conclusive but merely suggestive of an 
adverse effect. See DES Commissioner Decision.

C. Withdrawal Under the Delaney Clause and the DES Proviso

    Section 512(e)(1)(B) of the FD&C Act provides grounds for 
withdrawal of approval of an NADA if new evidence, tests by new 
methods, or tests by methods not deemed reasonably applicable when such 
application was approved, evaluated together with the evidence 
available when the application was approved shows that the Delaney 
Clause, section 512(d)(1)(I) of the FD&C Act, applies to the drug. 
Under the Delaney Clause, the Secretary may not approve a new animal 
drug application if ``such drug induces cancer when ingested by man or 
animal or, after tests which are appropriate for the evaluation of the 
safety of such drug, induces cancer in man or animal'' (section 512 
(d)(1)(I) of the FD&C Act). An exception to this general rule, referred 
to as the DES Proviso, allows for the approval of a carcinogenic new 
animal drug where FDA finds that, under the approved conditions of use: 
(1) The drug will not adversely affect the animals treated with the 
drug and (2) no residues of the drug will be found by an approved 
regulatory method in any edible tissues of or in any foods yielded by 
the animal (section 512(d)(1)(I)(i)-(ii) of the FD&C Act).
    FDA has issued implementing regulations that set the requirements 
for demonstrating that no residues of the drug will be found by an 
approved regulatory method in any edible tissues of or in any foods 
yielded from the animal (21 CFR part 500, subpart E). These 
regulations, referred to as the sensitivity of the method regulations 
(SOM regulations), describe how FDA determines whether the regulatory 
method proposed by a sponsor to detect no residues of the carcinogenic 
drug is sufficiently sensitive to ensure that residues of carcinogenic 
concern in edible tissues will not exceed concentrations that represent 
no significant increase in the risk of cancer to humans.
    Pursuant to these regulations, CVM determines for each drug and 
each drug metabolite (on the basis of the results of chronic bioassays 
and other information) whether the drug or any of its metabolites 
should be regulated as a carcinogen (Sec.  500.84(a)). For the drug and 
each metabolite determined to be carcinogenic, CVM calculates, based 
upon submitted assays, the concentration of the test compound in the 
total diet of the test animal that corresponds to a maximum lifetime 
risk of cancer in the test animal of 1 in 1 million (Sec.  
500.84(c)(1)). CVM designates the lowest value thus calculated as the 
So (Sec.  500.84(c)(1)). The So corresponds to a 
concentration of residue of carcinogenic concern in the total human 
diet that represents no significant increase in the risk of cancer to 
people (Sec.  500.82(b). Residue of carcinogenic concern includes all 
compounds in the total residue of a demonstrated carcinogen excluding 
any compound judged by CVM not to present a carcinogenic risk (Sec.  
500.82(b)). The total residues of carcinogenic concern (the drug and 
all of its metabolites less metabolites shown to be noncarcinogenic) 
are regulated based on the most potent carcinogenic residue (Sec.  
500.84(c)(1)). This approach ensures that use of the drug does not 
present a significant increase in the risk of cancer when considering 
all residues in edible tissues.
    Because the total diet is not derived only from food-producing 
animals, the SOM regulations make adjustments for human food intake of 
edible tissues, and determine the concentration of residues of 
carcinogenic concern in a specific edible tissue that corresponds to no 
significant increase in the risk of cancer to the human consumer. CVM 
assumes for purposes of these regulations that this value will 
correspond to the concentration of residues in a specific edible tissue 
that corresponds to a maximum lifetime risk of cancer in test animals 
of 1 in 1 million. This value is termed the Sm (Sec. Sec.  
500.82(b) and 500.84(c)(1)).
    Based upon residue depletion data submitted by a sponsor, CVM 
selects a target tissue (the edible tissue selected to monitor for 
residues in the target animals) and a marker residue (a residue whose 
concentration is in a known relationship to the concentration of the 
residues of carcinogenic concern in the last tissue to deplete to the 
Sm) and designates the concentration of the marker residue 
that the regulatory method must be capable of detecting in the target 
tissue (Sec.  500.86(a)-(c)). This value, termed the Rm, is 
the concentration of a marker residue in the target tissue when the 
residue of carcinogenic concern is equal to Sm, such that 
the absence of the marker residue in the target tissue above 
Rm can be taken as confirmation that the residue of 
carcinogenic concern does not exceed Sm in each of the 
edible tissues (Sec. Sec.  500.82(b) and 500.86(c)). When the marker 
residue is at or below the Rm, the residue of carcinogenic 
concern in the diet of people does not exceed So (Sec.  
500.86(c)).
    A sponsor must submit a regulatory method that is able to detect 
the marker residue at or below the Rm ((Sec. Sec.  500.88(b) 
and 500.84(c)(2)) (``The LOD [Limit of Detection for the regulatory 
method] must be less than or equal to Rm.'')). If a method 
cannot be developed that can detect the marker residue at or below the 
Rm, the requirements of the SOM regulations are not 
satisfied, and FDA cannot approve the drug. The DES Proviso and FDA's 
implementing regulations are satisfied where no marker residue is 
detectable using the approved regulatory method under the proposed 
conditions of use of the drug, including the proposed preslaughter 
withdrawal period (Sec.  500.84(c)(3)).
    As stated above, pursuant to section 512(e)(1)(B) of the FD&C Act, 
the Secretary shall, after due notice and an opportunity for a hearing, 
withdraw approval of an NADA if the Secretary finds that new evidence, 
tests by new methods, or tests by methods not deemed reasonably 
applicable when such application was approved, evaluated together with 
the evidence available when the application was approved shows that the 
Delaney Clause applies to the drug. Evidence that the Delaney Clause 
applies to a drug exists where the drug has previously been determined 
to be a carcinogen and the new evidence shows CVM's prior establishment 
of an analytical method and residue tolerance under the DES proviso 
exception to the Delaney Clause is inadequate. An analytical method is 
inadequate where new evidence demonstrates that the method does not 
accurately detect the marker residue or where new evidence demonstrates 
that not all residues of carcinogenic concern have depleted at the 
approved tolerance level of the marker residue (see, e.g., Rhone-
Poulenc, 636 F.2d at 752-53.)
    In establishing that grounds for withdrawal of approval exist under 
this clause, CVM carries an initial burden to demonstrate that the new 
animal drug and/or any of its metabolites induces cancer when ingested 
by man or animals. Proposal to Withdraw New Animal Drug Applications 
for Furazolidone (NF-180) and Nitrofurazone (NF-7), ALJ Decision, FDA 
Docket No. FDA-1976-N-0511, at 73 (formerly 76N-0172; November 12,

[[Page 21564]]

1986) (hereinafter ALJ Decision, November 12, 1986). Once CVM has 
satisfied its initial burden, the sponsor bears the burden of showing 
that the drug satisfies the DES Proviso exception to the Delaney Clause 
and FDA's implementing regulations. ALJ Decision, November 12, 1986, at 
73. (``Since furazolidone is also being challenged under the Delaney 
Clause, an additional issue . . . is whether new evidence put forth by 
the Center shows that furazolidone and/or its metabolites induces 
cancer when ingested by man or animal. If this burden is met, the 
sponsors must show [that the drug satisfies the DES proviso and FDA's 
implementing regulations]''); see also 21 CFR 500.92(b) (providing that 
for those compounds that FDA determines have been shown to induce 
cancer when ingested by man or animals, Sec. Sec.  500.82 through 
500.90 apply).
    In this case, CVM had previously determined, in the approval and 
supplemental approvals of new animal drugs containing carbadox, that 
carbadox and its metabolites, including DCBX, induce cancer in animals, 
but that the drug could be approved under the DES Proviso exception to 
the Delaney Clause. See Section IV. However, new evidence raises 
questions about whether the drug is properly approved under the DES 
Proviso to the Delaney Clause and FDA's implementing regulations. See 
Criteria and Procedures for Evaluating Assays for Carcinogenic Residues 
(44 FR 17070 at 17104, March 20, 1979) (reproposal of rules revoked in 
accordance with court order). (``[The FD&C Act] defines the new 
evidence that the Commissioner can consider in determining whether a 
previously approved compound is safe. [Proper analytical methods 
establishing residue levels] are necessary to show that a sponsored 
compound is safe under the FD&C Act. For that reason, the absence of 
data satisfying the [criteria in 512(e)(1)(B) of the FD&C Act], in 
conjunction with the evidence already available about a compound, 
clearly can support the withdrawal of approval of an application.''). 
In particular, new evidence indicating that an approved regulatory 
method can no longer be relied upon is sufficient to satisfy the 
Agency's burden to support withdrawal of approval under section 
512(e)(1)(B) of the FD&C Act and the Delaney Clause:

    In the case of an approved NADA for a carcinogenic compound, if 
FDA determines based on new information that the approved analytical 
method for detecting residues is inadequate . . . FDA could withdraw 
the approval on the basis of the Delaney Clause. Faced with evidence 
that an approved method was inadequate, FDA could not make a finding 
that ``no residue'' of the sponsored compound would be found in the 
edible products of treated animals. The DES Proviso cannot begin to 
operate without that finding, and, accordingly, the Delaney Clause 
would preclude continued approval. See Sponsored Compounds in Food 
Producing Animals; Criteria and Procedures for Evaluating Safety of 
Carcinogenic Residues; Proposed Rule (50 FR 45530 at 45550, October 
31, 1985); \4\ see DES Commissioners' Decision (44 FR 54852 at 
54859, September 21, 1979).
---------------------------------------------------------------------------

    \4\ Under FDA's regulations implementing the Delaney Clause for 
animal drugs, part 500, subpart E, a carcinogenic drug may not be 
approved if the regulatory method to test for the compound is not 
sufficiently sensitive. Sec. Sec.  500.84(c)(2) and 500.88(b). A 
carcinogenic drug will be withdrawn if new evidence shows that an 
approved regulatory method is not sufficiently sensitive.

    In this case, new evidence raises serious questions both about the 
acceptability of the current method in determining levels of known 
carcinogenic residues of carbadox, and, further, demonstrates that 
previously unidentified carcinogenic metabolites exist that are 
entirely unaccounted for in current approved testing methodology. 
Because the current analytic method is inadequate to identify the level 
of known carcinogens and does not identify the residue level of 
unidentified metabolites of carcinogenic concern, the current method 
and tolerance are inadequate to satisfy the DES Proviso.

D. Withdrawal Under the General Safety Clause

    The General Safety Clause in section 512(e) of the FD&C Act 
provides grounds for withdrawal of approval of an NADA if new evidence, 
tests by new methods, or tests by methods not deemed reasonably 
applicable when such application was approved, evaluated together with 
the evidence available when the application was approved shows that the 
drug is ``not shown to be safe for use under the conditions of use upon 
the basis of which the application was approved'' (section 512(e)(1)(B) 
of the FD&C Act). CVM has the initial burden to present new evidence 
that raises serious questions about the safety of the drug. Only upon 
that showing is there a burden on the manufacturer to demonstrate that 
the drug is safe. See Rhone-Poulenc, 636 F.2d at 752-53; Hess & Clark, 
495 F.2d 975, 992 (D.C. Cir. 1974).
    When evaluating a drug for withdrawal under the General Safety 
Clause, for CVM to satisfy its initial burden that new evidence raises 
serious human food safety questions, it must demonstrate a relationship 
between the drug residues found in edible tissues and risk to human 
health.

    [Without using] the Delaney Clause, it is not enough for the 
Commissioner merely to show that animal carcasses contain residues 
and that [the drug] is a carcinogen. Instead, the FDA must show that 
two different issues are resolved in its favor before it can shift 
to petitioners the burden of showing safety: (1) whether the 
detected residues are related to the use of [the drug]; (2) if so, 
whether the residues, because of their composition, and in the 
amounts present in the tissue, present some potential hazard to the 
public health. See Hess & Clark, 495 F.2d at 992 (D.C. Cir. 1974).

    Applying this test, the D.C. Circuit Court of Appeals has held that 
new evidence of drug residues in edible tissues in conjunction with 
evidence that any drug residues of the drug in question present safety 
concerns is sufficient to satisfy CVM's burden of raising serious 
questions regarding the safety of the drug. See Rhone-Poulenc, 636 F.2d 
at 752-53. CVM, acknowledging the Hess & Clark standard and its 
subsequent application, has withdrawn approval of a new animal drug 
under the General Safety Clause where new evidence showed that: (1) The 
new animal drug was carcinogenic; (2) some drug metabolites were 
mutagenic; and (3) residues left in edible tissues at the withdrawal 
time were unidentified. See Nitrofurans Commissioners' Decision, 56 FR 
41902 at 41910, August 23, 1991 (``Since the nature of these residues 
and their toxicity were not evaluated, they cannot be regarded as safe 
. . . Contrary to the sponsors' assertions, the evidence fails to 
demonstrate that furazolidone's metabolites pose no health risk to the 
human consumers. Given all the other evidence in the record 
demonstrating that furazolidone is a carcinogen and that its 
metabolites are mutagens, I find that, contrary to the sponsors' 
assertions, the metabolites of furazolidone pose a potential health 
risk to human consumers.'') see also DES Commissioners' Decision, 44 FR 
54852 at 54868 (explaining that, ``[w]here new evidence shows that use 
of the drug results in residues of unidentified substances,'' CVM must 
decide whether, despite this lack of knowledge, ``the drug may be 
considered to be 'shown to be safe[,]' '' as the General Safety Clause 
requires). In other words, because residues of a mutagenic carcinogen 
are presumptively carcinogenic, and therefore presumptively unsafe, 
where new evidence demonstrates that unidentified residues of a 
mutagenic carcinogen remain at the time of withdrawal, the drug meets 
the standard set forth in Hess & Clark.

[[Page 21565]]

    Applying the Hess & Clark standard here, the new evidence regarding 
carbadox clearly meets both prongs of that test. New evidence 
demonstrates that previously unidentified mutagenic residues of 
carbadox, a known carcinogen, remain present well after the established 
withdrawal period. As discussed further in Section V.D., because 
carbadox is a mutagenic carcinogen and QCA is the only known quantified 
noncarcinogenic residue of carbadox, all other residues are of 
carcinogenic concern. The new evidence demonstrates that the total 
residues of carcinogenic concern at the established 42-day withdrawal 
period are much higher than previously thought because the residues are 
no longer shown to be residues related to a noncarcinogenic compound, 
QCA, as previously believed. See, infra, Section V.D. Thus, the new 
evidence demonstrates that: (1) The unidentified residues are related 
to the use of carbadox and (2) the residues pose a potential hazard to 
public health because of the amount present and because they are 
residues of carcinogenic concern.

IV. Regulation of Residues of Carbadox

A. 1972 and 1975 Approvals

    Carbadox is a carcinogen and was approved as a new animal drug 
pursuant to the DES Proviso exception to the Delaney Clause. At the 
time of the initial approval of carbadox in 1972, CVM (then the Bureau 
of Veterinary Medicine) recognized that carbadox is a carcinogen and 
therefore required that no residues of carbadox or its metabolite QCA 
be found in uncooked edible tissues of swine at the time of slaughter, 
as determined by the approved method of analysis. See 37 FR 20683, 
October 3, 1972, as amended by 37 FR 23906, November 10, 1972. This 
approval occurred prior to FDA's 1987 initial issue of regulations 
implementing the DES Proviso and therefore did not involve the 
development of a regulatory method sensitive enough to detect a marker 
residue that corresponded to a lifetime risk of cancer to test animals 
of 1 in 1 million (as described in Section III.C).
    In this initial approval, based upon the submission of studies 
showing the depletion of carbadox residues in edible tissues, CVM 
determined that ``[a]ll tissues except the liver [were] free of all 
residues'' of unchanged carbadox at 24 hours after withdrawal of 
treatment and that unchanged carbadox ``ha[d] disappeared from the 
liver after 24 hours'' (Ref. 8). CVM also determined from submitted 
studies that the carcinogenic parent drug was undetectable in liver at 
24 hours (Id.). CVM further determined that a ``restriction of use in 
the labeling provides a withdrawal period long enough [70 days] to 
assure no hazard to humans consuming residues in meat. In proper use 
there would be virtually no residues'' of carbadox in tissues at 
slaughter (Ref. 9). The conclusions CVM made in 1972 regarding the 
rapid depletion of carcinogenic residues were later independently 
corroborated by a 1990 evaluation of carbadox by JECFA (Ref. 10 at p. 
30).
    Labeled use restrictions, as the drug was approved in 1972, 
included an upper weight limit of 75 pounds body weight and a 
prohibition on mixing into complete feeds containing less than 15 
percent crude protein, thus limiting the drug's use to young pigs. 
These use restrictions provided assurances that the 70-day withdrawal 
period would likely be followed in practice (Ref. 11).
    Similarly in 1975, FDA approved NADA 092-955 for the use of 
carbadox with pyrantel tartrate in Type C medicated swine feed (40 FR 
45164, October 1, 1975). At that time, CVM reviewed drug residue 
studies of carbadox and pyrantel tartrate used in combination. The 
studies showed that, at 45 and 60 days withdrawal, concentrations of 
residues of carbadox in all tissues tested were undetectable using the 
previously approved analytical method with a 30 ppb limit of detection 
(Ref. 12 at p. 2).

B. 1986 Citizen Petition

    On May 9, 1986, the Center for Science in the Public Interest 
submitted a citizen petition requesting that FDA withdraw approval of 
new animal drug applications for ipronidazole, dimetridazole, and 
carbadox (Ref. 13). The petition asserted that FDA must withdraw the 
approval of carbadox because carbadox and its metabolites DCBX and 
hydrazine were found to be carcinogenic, and the approved test method 
for carbadox residues is ``unsuitable'' (Ref. 13 at p. 20). The 
asserted unsuitability of the approved test method was based upon the 
fact that only a small portion of total residues had been positively 
identified and that the analytical method for carbadox residues was not 
sensitive enough to ensure that all residues had depleted.
    FDA responded to the 1986 citizen petition in 1995 after a review 
of new residue depletion data submitted by (the then sponsor) Pfizer as 
well as data previously submitted to the Agency as part of the carbadox 
NADAs. Based upon this review, FDA denied the petition as it related to 
carbadox because it determined that ``if used according to label 
directions, residues of carbadox remaining in edible tissues of swine 
do not pose a human food safety risk to consumers'' (Ref. 14 at p. 2). 
FDA based this safety determination on the following findings:

    1. At 70 days withdrawal, the drug-related residue in swine 
liver measured 13 ppb. 2. Ten percent of the drug-related residue 
was extractable and identified to be a noncarcinogenic metabolite, 
quinoxaline-2-carboxylic acid. 3. The remaining 90% of the drug-
related residue was unextractable or bound residues. 4. The bound 
residues were related to quinoxaline-2-carbodoxaldehyde and 
quinoxaline-2-carboxylic acid, both of which are of no carcinogenic 
concern. (Ref. 14 at p. 1).

C. Approval of 1998 Supplemental NADAs

    In 1998, FDA approved two supplemental applications to NADA 041-
061. The first supplement, approved in January 1998, assigned the 
noncarcinogenic metabolite QCA as the marker residue and set a 
tolerance of 30 ppb QCA in swine liver (Ref. 1).
    Toxicology studies, including carcinogenicity bioassays with 
carbadox, DCBX (a primary metabolite of carbadox), and hydrazine were 
submitted as part of that supplemental application (Ref. 1 at pp. 1-5). 
The studies demonstrated the carcinogenicity of carbadox, DCBX, and 
hydrazine, and indicated that DCBX was the most potent of the three 
carcinogenic compounds (id.). Consequently, based on DCBX, CVM 
calculated an So of 0.061 ppb for total residues of 
carcinogenic concern for carbadox in the total diet (Ref. 1 at p. 5). 
CVM calculated an Sm value for total residues of 
carcinogenic concern in muscle at 0.305 ppb, in liver at 0.915 ppb, and 
in kidney and fat at 1.830 ppb (Ref. 1 at pp. 8-9).
    The SOM regulations, as they existed in 1998, directed CVM to 
establish an Rm for carcinogenic compounds used in food-
producing animals. CVM did not establish an Rm because CVM 
concluded the parent carbadox was rapidly metabolized, carcinogenic 
residues were not detectable beyond 72 hours post dosing, and 
unextracted residues \5\ were

[[Page 21566]]

related to noncarcinogenic QCA and not of carcinogenic concern. Because 
the noncarcinogen QCA was the only detectable metabolite persisting 
beyond 72 hours post dosing, CVM assigned it as the marker residue 
(id.).
---------------------------------------------------------------------------

    \5\ Unextracted residues are residues of the drug that are not 
released when tissues are exposed to mild aqueous or organic 
extraction conditions. Guidance on analysis of unextracted total 
radiolabeled residue is provided in ``Guidance for Industry: General 
Principles for Evaluating the Safety of Compounds Used in Food-
Producing Animals (GFI #3),'' 2006. Unextracted or bound residues 
can be either: (1) Endogenous components resulting from fragments of 
the radiolabeled compound being incorporated into naturally 
occurring molecules such as amino or nucleic acids or (2) covalently 
bound residues. Covalently bound residues are considered to be of 
toxicological concern and their availability for absorption into the 
human gastrointestinal tract is considered during an evaluation of 
human food safety. Residues incorporated into endogenous molecules 
are not considered bioavailable or to be of toxicological concern. 
However, CVM has determined that establishing a potentially 
carcinogenic compound is bound and not of carcinogenic concern can 
be complicated by the possibility of gastrointestinal binding and 
gastrointestinal carcinogenesis and consequently can involve a more 
comprehensive assessment of the bound compounds as described in GFI 
#3. Note that while CVM has recognized that carbadox residues have 
not been fully extracted and characterized, CVM has not made an 
assessment that the compounds are not carcinogenic because they are 
bound to endogenous molecules (Ref. 15 at pp. 3-4). Moreover, 
residue studies presented to JECFA in 2003 suggest that carcinogenic 
residues that had not been extracted when exposed to organic 
extraction were released by simulated digestive enzymes (Ref. 2 at 
pp. 7-8, Table 5).
---------------------------------------------------------------------------

    At the time it approved the supplement in January 1998, CVM said:

    The sponsor and academic researchers have conducted numerous 
studies evaluating the fate of carbadox in animals. These residue 
depletion data are summarized in FAO Food and Nutrition Paper 41/3 
(Food and Agriculture Organization (FAO) of the United Nations, 
1991) and show that carbadox, desoxycarbadox and hydrazine do not 
persist in edible tissue as detectable residues beyond 72 hours. The 
agency's evaluation of these data, and the new information provided 
by the sponsor, demonstrate that following administration, parent 
carbadox is rapidly metabolized; that the metabolism of carbadox is 
similar among species; that the in vivo metabolism of the compounds 
of carcinogenic concern is also rapid and irreversible such that the 
resulting metabolic products cannot regenerate compounds of 
carcinogenic concern; that the unextractable residues are related to 
non-carcinogenic compounds, quinoxaline-2-carboxylic acid [QCA] and 
quinoxaline-2-carboxaldehyde; and that quinoxaline-2-carboxylic acid 
[QCA] is the only residue detectable in the edible tissues beyond 72 
hours post dosing. Thus, the agency concludes that the unextractable 
bound residue is not of carcinogenic concern and that QCA is a 
reliable marker residue for carbadox. (Ref. 1 at p. 9).

    CVM established a tolerance of 30 ppb for residues of QCA in liver, 
the tissue in which residues persist for the longest time. CVM 
concluded that the concentration of residues of carcinogenic concern in 
edible tissues was below the Sm when the concentration of 
QCA in liver had depleted to 30 ppb.\6\
---------------------------------------------------------------------------

    \6\ The SOM regulations, as they existed in 1998, permitted 
approval of a regulatory method that could detect the marker residue 
of the drug, as long as the marker residue would only be detected at 
or below the Rm under the proposed conditions of use. See 
Sec.  500.86(c) (1998).

    Under FDA's operational definition of ``no residue,'' a residue 
of carcinogenic concern, so long as it does not exceed the 
So, may be detectable by an approved method. The residue 
data show that carbadox, desoxycarbadox and hydrazine do not persist 
in edible tissue as detectable residues beyond 72 hours. The in vivo 
metabolism of the compounds of carcinogenic concern is irreversible. 
Therefore, in this case, no residue of carcinogenic concern, even 
below the So, is detectable by any method. The 
unextracted residues are related to a noncarcinogenic compound, 
quinoxaline-2-carboxylic acid (QCA), and extractable QCA is the only 
residue detectable in the edible tissues 72 hours postdosing. Thus, 
the agency concludes that QCA is a reliable marker residue for 
carbadox and its metabolites.
    From these data, FDA has selected liver as the target tissue and 
quinoxaline-2-carboxylic acid (QCA) as the marker residue. FDA has 
determined that when QCA, the marker, is at or below 30 ppb in the 
target tissue, liver, that no residue of carcinogenic concern, above 
the So, is detectable in each of the edible tissues by 
any method.
    The sponsor has submitted a regulatory method capable of 
measuring QCA at and below 30 ppb in the target tissue. (Ref. 1 at 
p. 14).

    As part of their application supporting the January 1998 
supplemental approval, the sponsor submitted a regulatory method for 
residues of QCA in swine liver. The regulatory method relies on a gas 
chromatograph assay with electron capture detection and has a limit of 
quantification of 5 ppb (Ref. 1 at p. 13), a 6-fold improvement of the 
sensitivity from the previously approved regulatory method (Ref 1.)
    In October 1998, FDA approved an additional supplement to NADA 041-
061 changing the withdrawal period for carbadox medicated feeds from 70 
days to 42 days. The supplement was approved based upon the previous 
approval of a tolerance of 30 ppb for QCA and a residue depletion study 
that showed that residues of QCA in liver depleted below 30 ppb by 42 
days (Ref. 16).
    To summarize, in 1998, when FDA approved supplements to NADA 041-
061 establishing a drug tolerance and shortening the withdrawal period, 
the evidence before CVM indicated:
     A 0.915 ppb concentration of total residues of 
carcinogenic concern in liver is the concentration that represents no 
significant increase in the risk of cancer to people--total residues of 
carcinogenic concern in liver above 0.915 ppb under the drug's approved 
conditions of use are unsafe. Such residues would preclude continued 
approval because the drug would not be shown to be safe and because the 
exception to the Delaney Clause would not apply (Ref. 1 at pp. 8-9, 10, 
14).
     The parent compound carbadox is rapidly metabolized and 
carcinogenic residues of the drug are not identifiable in any edible 
tissues beyond 72 hours post dosing (Ref. 1 at p. 9).
     Remaining unextracted residues of carbadox are 
noncarcinogenic residues related to the noncarcinogenic metabolite QCA 
(Ref. 1 at pp. 9, 14).
     QCA is a reliable marker residue for carbadox and its 
metabolites; that is, measuring QCA residues in swine liver is a valid 
method for demonstrating the absence of residues of carcinogenic 
concern in edible tissues (id.).
    Based upon these conclusions, CVM found that under the conditions 
of use the drug did not result in unsafe residues of carcinogenic 
concern in edible tissues and that the use of carbadox, as approved in 
the NADA supplements, satisfied the DES Proviso exception to the 
Delaney Clause prohibition on carcinogenic animal drugs (id.).

D. Approval of the 2004 Feed Use Combination

    In 2004, FDA approved a combination drug medicated feed containing 
carbadox and oxytetracycline under NADA 141-211 (Ref. 17). In 
accordance with section 512(d)(4)(A) of the FD&C Act, approval of a 
combination new animal drug, where the underlying new animal drugs have 
previously been separately approved for particular uses and conditions 
of use for which they are intended for use in the combination, will not 
be refused on human food safety grounds unless the application fails to 
establish that: (1) None of the animal drugs used in combination, at 
the longest withdrawal period for any of the drugs in the combination, 
exceeds its established tolerance or (2) none of the drugs in the 
combination interferes with the method of analysis for any of the other 
drugs in the combination (section 512(d)(4)(A)(i)-(ii) of the FD&C 
Act). In other words, in order to approve a combination new animal drug 
for a drug product that contains two previously approved new animal 
drugs, no new information needs to be supplied to establish the safety 
of either drug. Instead, the application need only demonstrate that use 
of the drugs in combination will not result in violative

[[Page 21567]]

residues of any component drug or in drug assay interference.
    Both carbadox and oxytetracycline had been previously and 
separately approved by FDA for the same conditions of use proposed for 
their use in combination. See 21 CFR 558.450 (Oxytetracycline); Sec.  
558.115 (Carbadox). The sponsor, Phibro, provided tissue residue 
depletion data demonstrating that QCA residues did not exceed the 
tolerance of 30 ppb when carbadox was administered in conjunction with 
oxytetracycline to swine (Ref. 17). A pharmacokinetic study comparing 
blood levels of oxytetracycline when administered alone and when 
administered in conjunction with carbadox satisfied the need to 
demonstrate that residues of oxytetracycline would not exceed the 
oxytetracycline tolerance at 42 days (id.).
    The sponsor further provided data demonstrating noninterference of 
oxytetracycline with the method of analysis of QCA in liver (id.). 
Having made the required human food safety demonstrations for 
combination animal drugs, there was no basis to refuse approval of the 
product on human food safety grounds. The combination new animal drug 
was subsequently approved (id.).

V. New Information Regarding Carcinogenic Residues in Edible Tissues

    Three sources provide new information regarding carcinogenic 
residues in edible tissues: Data submitted to the 2003 JECFA and the 
subsequent JECFA report (Ref. 2) and two publications in the peer-
reviewed literature (Refs. 4 and 6).
    JECFA is an internationally recognized expert body, providing the 
scientific evaluations that become the basis for international food 
standards established by the Codex Alimentarius Commission and 
supporting international treaties such as the Sanitary Phytosanitary 
Agreement. JECFA experts are chosen based on expertise, reputation, 
assurance of lack of conflict of interest, and familiarity with the 
subject of that particular evaluation.
    In addition, pursuant to section 512(l)(1) of the FD&C Act,\7\ FDA 
ordered Phibro to provide it with the same data provided to the 2003 
JECFA. CVM evaluated the submitted data and found that it raised 
questions regarding the safety of food resulting from swine treated 
with carbadox. Confidence in the information evaluated by the 2003 
JECFA that is the basis for CVM's concern about carbadox was increased 
by the independent findings reported in the two publications discussed 
further.
---------------------------------------------------------------------------

    \7\ An order issued pursuant to section 512(l) of the FD&C Act, 
requires a sponsor to submit such data and information as FDA may 
find necessary to determine or facilitate a determination whether 
grounds to withdraw approval of an NADA under section 512(e) of the 
FD&C Act exist.
---------------------------------------------------------------------------

A. New Information Provided to JECFA

    In 2003, at the request of the Codex Committee on Residues of 
Veterinary Drugs in Foods (CCRVDF), JECFA reevaluated the recommended 
Maximum Residue Limits (MRLs) for carbadox that were based upon a 1990 
JECFA evaluation of the new animal drug (Ref. 2). CCRVDF, which 
includes CVM as a participant, determines priorities for the 
consideration of residues of veterinary drugs in foods and recommends 
MRLs for veterinary drugs to the Codex Alimentarius Commission of the 
Food and Agriculture Organization and the World Health Organization of 
the United Nations. The Codex Alimentarius Commission develops 
harmonized international food standards, guidelines, and codes of 
practice to protect the health of the consumers and ensure fair 
practices in food trade (see footnote 2).
    Based on studies submitted to JECFA that showed the persistence of 
genotoxic, carcinogenic residues, JECFA could not determine an amount 
of residues of carbadox in human food that would have no adverse health 
effects in consumers. JECFA recommended that the Codex MRLs be 
withdrawn. CCRVDF concurred with JECFA's recommendation and proposed to 
the Commission that the MRLs be withdrawn. The Commission subsequently 
agreed and withdrew the Codex MRLs for carbadox (Ref. 18 at p. 120).
    As part of the JECFA reevaluation process, Phibro presented two new 
residue studies to JECFA in 2003. Only one of these studies involved 
measurement of the depletion of carcinogenic metabolites of carbadox in 
edible tissues. In that study, animals were fed for 14 days at the 
approved dose of 55 ppm carbadox in feed (Ref. 2 at pp. 6-10). Animals 
were euthanized at various time points between 0 hours and 15 days post 
treatment, and samples of swine muscle, liver, skin, and fat were 
collected (Ref. 2 at pp. 7-8, Table 5).
    Prior to analysis for residues, some of the tissue samples were 
exposed to human digestive enzymes \8\ (Ref. 2 at p. 7). This in vitro 
model of bioavailability was designed to mimic effects of gastric fluid 
and intestinal fluid incubation in human stomach and small intestine to 
evaluate whether residues potentially could be released in the human 
gastrointestinal tract. To allow comparison, some tissue samples were 
left untreated while other tissue samples were incubated in simulated 
gastric fluid (with pepsin) or in simulated intestinal fluid (with 
pancreatin). Residues of carbadox, DCBX, and QCA were measured in the 
untreated tissues, in tissues that were incubated with enzymes, and in 
the supernatant of those tissues that were incubated with enzymes 
(id.).
---------------------------------------------------------------------------

    \8\ The use of enzymic preparations to characterize residues is 
described in section 2.3.4.3.2 of CVM Guidance for Industry (GFI) 
#205 VICH GL 46, ``Studies to Evaluate the Metabolism and Residue 
Kinetics of Veterinary Drugs in Food-Producing Animals: Metabolism 
Study to Determine the Quantity and Identify the Nature of Residues 
(MRK),'' Sept. 15, 2011 (Ref. 19).
---------------------------------------------------------------------------

    Residues of carbadox, DCBX, and QCA were measured by liquid 
chromatography-atmospheric pressure chemical ionization tandem mass 
spectrometry (LC/APCI-MS/MS). The tissue samples that were not 
incubated with enzymes were extracted with acetonitrile prior to 
analysis. The tissue samples that were incubated with enzymes were 
extracted with ethyl acetate prior to analysis. Supernatants of the 
enzyme digestion were analyzed directly without extraction. The limits 
of quantification for LC/APCI-MS/MS were 0.050 ppb for carbadox 
residues and 0.030 ppb for DCBX residues (id.). The detection 
capabilities of this methodology were greatly enhanced compared to the 
previous method for carbadox and DCBX (i.e., the method used for the 
previous analytical work had a detection limit of 2 ppb) (Ref. 20).
    The study presented to JECFA showed that residue concentrations of 
carbadox and DCBX were higher and persisted for a longer period post 
dosing in liver than in the other sampled tissues. In liver without 
treatment with simulated digestive fluids, carbadox was detectable 
(0.050 ppb) as long as 48 hours post dosing and DCBX was detectable 
(0.138 ppb) at the last sampling time point, which was 15 days post 
treatment (Ref. 2 at pp. 7-8, Table 5). Treatment of tissues with 
simulated digestive fluids resulted in measurement of significantly 
higher concentrations of DCBX. ``Pretreatment of the samples with 
digestive fluids increased the amounts of carcinogenic residues found 
in all tissues. In liver the concentration of . . .[DCBX] increased by 
more than fourfold when the samples were treated with intestinal fluid, 
and large quantities were present 15 days after withdrawal . . .'' 
(Ref. 2 at p. 17).

[[Page 21568]]

    In particular, the study showed that concentrations of 
approximately 35 ppb of DCBX at 0 hours post dosing and approximately 
2.7 ppb of DCBX at 15 days post dosing were measured in liver treated 
with pancreatin (Ref. 2 at p. 8, Table 5). The significantly increased 
residues found in liver after treatment with intestinal enzymes show 
that enzymatic treatment was able to release carcinogenic residues that 
were not extractable by organic solvents, such as those used in tissue 
residue studies to support the original and supplemental approval of 
NADAs for use of carbadox.
    JECFA evaluated the percent recoveries of the analytes. Percent 
recovery is a measurement of accuracy of the analytical procedure and 
expresses the closeness of agreement between the true value of the 
analyte concentration and the mean value obtained by applying the 
analytical procedure (Ref. 21). JECFA reported that when carbadox, 
DCBX, and QCA were incubated for 4 hours with digestive enzymes, 
carbadox and DCBX were unstable (percent recovery decreased) in the 
samples treated with pepsin, but were stable in pancreatin (Ref. 2 at 
p. 16). JECFA also reported that the recoveries of the analytes from 
the liver samples were generally variable and decreased to low levels 
when digestive enzymes were used prior to extraction (Ref. 2 at pp. 17-
18).
    After evaluating the residue study, JECFA concluded that the poor 
recoveries obtained with the enzyme experiments ``showed that the true 
concentrations of the carcinogenic metabolites in tissues cannot yet be 
estimated with certainty, since an unknown portion of the releasable 
residue [of carbadox and DCBX] is destroyed during incubation [of liver 
tissues] with the [digestive] enzymes'' (Ref. 2 at p. 18). JECFA 
therefore concluded that the measured values of DCBX and carbadox 
``represent[ed] a lower estimate of the total present in the tissue'' 
(id.).
    Presented with data demonstrating both the depletion of QCA and 
depletion of the carcinogenic residue DCBX, JECFA established a 
relationship between the concentrations of QCA and DCBX in liver (Ref. 
2 at p. 14). The statistical analysis of the data showed a linear 
relationship between the logarithms of the concentrations of QCA and 
DCBX (Ref. 2 at pp. 14, 18). This relationship allowed JECFA to use 
regression analysis to assess the concentrations of DCBX when QCA 
depleted to 30 ppb in liver (the Codex MRL and FDA approved tolerance 
for carbadox). JECFA determined that ``[a]t the MRL [of 30 ppb] for QCA 
in liver, the average concentrations of the carcinogenic residue 
desoxy-carbadox in liver estimated by regression analysis were about 4 
[ppb]'' (Ref. 2 at pp. 14, 16-17). JECFA recognized that ``tolerance 
limits for the concentration of desoxycarbadox were several times 
higher owing to the wide variation of the data'' and thereby concluded 
that ``QCA is not a suitable marker for monitoring carcinogenic 
metabolites of carbadox in liver . . . and QCA does not ensure the 
absence of carcinogenic residues'' (Ref. 2 at p. 17).
    In contrast to the previous findings of JECFA, these new data show 
that carcinogenic residues, in particular DCBX, are present in edible 
tissues for a significant time during the depletion of parent carbadox 
(Ref. 2 at p. 18). Moreover, the study shows that treatment with 
simulated digestive enzymes releases higher levels of the carcinogenic 
residues DCBX than were recovered using organic extractions in the 
study. These higher concentrations provide evidence that the carbadox 
residues that were not extractable or identified in previous studies 
submitted to the Agency could include carcinogenic residues of carbadox 
that are releasable with enzymatic treatment of tissues. This evidence 
calls into question the Agency's previous conclusions that all 
unextracted and unidentified residues were noncarcinogenic residues 
related to QCA.
    After reviewing the new residue data, and considering the 
previously evaluated genotoxicity and carcinogenicity data, JECFA 
recommended withdrawal of the previously established Codex MRLs (Ref. 2 
at p. 18). Codex subsequently agreed and withdrew the MRLs for carbadox 
(Ref. 18 at p. 120).
    In summary, the studies considered by JECFA during its 2003 review 
of the drug indicated that:
     Residues of the carcinogenic metabolite of carbadox, DCBX, 
were measured in edible tissues for 15 days, which was the last 
sampling time point. DCBX was measured in swine liver after treatment 
with simulated digestive enzymes at concentrations as high as 2.69 ppb 
at 15 days post treatment (Ref. 2 at p. 8, Table 5).
     Analysis of measured concentrations of QCA and DCBX in 
liver indicated that approximately 4 ppb of DCBX would be present in 
the liver of treated animals when QCA reached the Codex MRL and the FDA 
tolerance of 30 ppb in liver (Ref. 2 at pp. 14, 17). This concentration 
of DCBX alone is more than 4 times higher than the concentration of 
total residues of carcinogenic concern in liver that would present no 
significant increase in the risk of cancer to people.
     Residues of carbadox previously unextracted from edible 
tissues could be released by gastric and intestinal fluids that mimic 
the human digestive process (Ref. 2 at p. 16). The enzymatic treatment 
used in the study significantly increased the recoveries of 
concentrations of DCBX and carbadox from edible tissues, thereby 
indicating that some portion of the previously unextracted and 
unidentified total residues is composed of carcinogenic compounds.

B. Additional New Evidence

    Following the reports of the 2003 JECFA reevaluation of carbadox, 
CVM requested that Phibro also provide the carcinogenic residue 
depletion study to CVM. In 2005, in response to CVM's request for 
information, Phibro submitted a summary of the carcinogenic residue 
depletion study previously provided to JECFA. Upon review of the 
summary data, CVM asked Phibro to submit existing studies or provide 
new and complete studies that address the relationship of QCA at 30 ppb 
and carbadox and DCBX residues, and about the use of QCA as the marker 
residue for surveillance purposes. In 2006, CVM asked for and received 
from Phibro a timeline for submission of complete information that 
addresses concerns about the relationship of QCA at 30 ppb and carbadox 
and DCBX residues, and about the use of QCA as the marker residue for 
surveillance purposes. Between 2006 and 2011, interactions between CVM 
and Phibro continued, with protocols submitted and reviewed, method 
validation reports submitted and reviewed, informal communications by 
email, and informal discussions by telephone. The focus of the 
interactions was development and validation of methods to measure QCA 
and DCBX in a tissue residue depletion study. Despite the continued 
interaction between Phibro and CVM, Phibro has not submitted the 
requested information.
    In 2011, pursuant to section 512(l)(1) of the FD&C Act, FDA ordered 
Phibro to provide all information in its possession with respect to: 
(1) The persistence of DCBX in edible tissues; (2) the appropriateness 
of QCA as an analyte for residue monitoring and for establishing a 
withdrawal time for the use of carbadox in pigs; and (3) whether an 
analytical method for monitoring carbadox-related carcinogenic residues 
in edible tissues can be developed that would comply with part 500, 
subpart E.

[[Page 21569]]

    In response to the 2011 FDA order, Phibro provided CVM with the 
full study report and appendices, previously provided to JECFA in 2003.
    CVM has independently evaluated the data from the Phibro study of 
depletion of carcinogenic residues reviewed by JECFA in 2003, and in 
particular has reviewed the JECFA conclusion that when QCA reaches 30 
ppb in liver, residues of DCBX in liver are ``estimated by regression 
analysis to be about 4 [ppb]'' (Ref. 2 at p. 18). CVM's statistical 
analysis of the residue concentrations of DCBX in liver treated with 
pancreatin (a simulated intestinal fluid) shows that concentrations of 
DCBX in liver, when QCA reaches the 30 ppb approved tolerance, would 
average 4 ppb and, based on the data in the JECFA report, could 
reasonably range from 1.4 ppb to 11 ppb, using a 95 percent prediction 
range. Based upon this analysis, DCBX alone--leaving aside additional, 
unidentified residues of carcinogenic concern--significantly exceeds 
the approved Sm when QCA, the approved marker residue, 
reaches the approved tolerance. The new evidence from the 2003 JECFA 
re-evaluation of carbadox, along with studies that were later submitted 
to CVM, undermine the human food safety conclusions that CVM had 
previously reached when considering the approval of the new animal drug 
applications for carbadox for its various uses. CVM has engaged with 
Phibro to evaluate the carbadox-associated safety concerns raised by 
the new evidence and repeatedly has asked Phibro to submit information 
that would address these safety concerns. Information provided by 
Phibro in response to these requests has not resolved CVM's human food 
safety concerns.
1. Boison, et al., 2009
    In addition, a 2009 publication calls into question conclusions 
made by CVM when it approved the NADAs and supplemental NADAs for 
carbadox (Ref. 4). Boison, et al., 2009, demonstrates the availability 
of a sensitive analytical method for DCBX, and provides information 
from which serious questions about the safety of carbadox can be 
inferred, specifically whether DCBX may be present in edible tissues of 
treated swine above the Sm even when the marker residue 
(QCA) concentration is below the tolerance of 30 ppb (id.).
    Boison, et al., report: (1) QCA is not a suitable marker for the 
regulation of carbadox because while QCA is very stable under 
temperature conditions above 60 [deg]C (i.e., 105 [deg]C), DCBX is not 
(Ref. 4 at p. 133); (2) the existence of an analytical method capable 
of detecting DCBX below the Sm for porcine muscle and liver 
(Ref. 4 at p. 132, Table 5); and (3) detection of DCBX at a 
concentration greater than 0.050 ppb in the diaphragm (but not the 
liver) of 2 of 6 hogs fed carbadox, while QCA was not detected in the 
liver of those same hogs at a limit of quantitation (LOQ) of 0.500 ppb 
(Ref. 4 at pp. 132-33). The findings of Boison, et al., are significant 
for two reasons: (1) QCA appears not to be a reliable marker residue 
and (2) DCBX is reported to be sensitive to the processing temperature 
used in the analytical method.
2. Baars, et al., 1991
    In 2012, in response to FDA's 2011 order under section 512(l) of 
the FD&C Act, Phibro sent CVM a letter citing Baars, et al., 1990 (Ref. 
5), an abstract of a study not previously provided. CVM obtained the 
study report Baars, et al., 1991 (Ref. 6), which reports an analytical 
method with a limit of detection of 1 ppb that detects the presence of 
DCBX in edible tissues for greater than 72 hours after removal of feed 
containing carbadox. Specifically, Baars, et al., 1991, demonstrated 
the presence of DCBX for up to 7 days (~168 hours) in the kidney and 14 
days (~336 hours) in the liver of swine fed carbadox (Ref. 5 at p. 3, 
Fig. 3; Ref. 6 at p. 290, Fig. 2). This observation called into 
question CVM's previous conclusion that all residues of carcinogenic 
concern deplete within 72 hours.

C. New Evidence Calls Into Question Prior CVM Conclusions That Were the 
Basis of the 1998 Supplemental Approval

    CVM's prior conclusion that QCA is a reliable marker residue for 
carbadox and its metabolites was predicated on several underlying 
conclusions (Ref. 1 at pp. 13-14). These underlying conclusions are 
reviewed below in light of the new evidence presented above.
    1. Previous Conclusion 1: The residue data show that carbadox, 
DCBX, and hydrazine do not persist in edible tissues as detectable 
residues beyond 72 hours.\9\
---------------------------------------------------------------------------

    \9\ This underlying conclusion is described in the January 30, 
1998, summary basis of approval under the Freedom of Information Act 
(FOI Summary) for NADA 041-061 (Ref. 1 at p. 9) and in the report of 
the 1990 JECFA meeting (Ref. 10 at p. 30).
---------------------------------------------------------------------------

    Since the time CVM made this previous conclusion, we have become 
aware of information that undermines the previous conclusion that 
carbadox and its carcinogenic metabolites do not persist in edible 
tissues beyond 72 hours. JECFA, in 2003, reviewed a study detecting 
DCBX in livers of swine up to 15 days after cessation of carbadox 
exposure. The study JECFA reviewed was limited to 15 days. The data 
presented to JECFA in 2003 provide new scientific evidence that DCBX 
persists in edible tissues of swine as a detectable residue beyond 72 
hours (Ref. 2).
    Further, Baars, et al., 1991, reports detecting DCBX in liver up to 
Day 14 after cessation of exposure to carbadox using an analytical 
method with a detection limit of 1 ppb (Ref. 6). Baars, et al., 1991, 
provides new scientific evidence that DCBX persists as a detectable 
residue in edible tissues of swine for greater than 72 hours.
    Scientific evidence from JECFA's 2003 evaluation of submitted 
information and Baars, et al., 1991, demonstrate that DCBX, one residue 
of carcinogenic concern for carbadox, persists in edible tissues of 
swine beyond 72 hours. All of this evidence was first received by CVM 
after the 1998 approval of the supplemental application to NADA 041-
061. Based on this new scientific evidence, the previous conclusion 
that DCBX does not persist in edible tissues of swine as a detectable 
residue beyond 72 hours is no longer justified.
    2. Previous Conclusion 2: The unextracted residues are related to a 
noncarcinogenic compound, QCA, and extractable QCA is the only residue 
detectable in the edible tissues of swine 72 hours post dosing.\10\
---------------------------------------------------------------------------

    \10\ This underlying conclusion is described in the January 30, 
1998, summary basis of approval under the Freedom of Information Act 
(FOI Summary) for NADA 041-061 (Ref. 1 at p. 9) and in the report of 
the 1990 JECFA meeting (Ref. 10 at p. 30).
---------------------------------------------------------------------------

    At the time of the 1998 supplemental approval, CVM concluded that 
that unextracted residues were related to the noncarcinogenic compound, 
QCA, and that extractable QCA was the only residue detectable in the 
edible tissues after 72 hours post dosing. However, CVM is now aware of 
reports of extraction of residues being enhanced by pepsin or 
pancreatin digestion prior to organic extraction, making non-QCA 
residues previously thought to be unextractable currently extractable 
(Ref. 2). JECFA reports that some residues of carbadox previously 
identified as unextractable can now be extracted (id.). DCBX was found 
in the newly extractable residues. This scientific evidence 
demonstrates that some residues previously found to be unextractable 
are extractable and that the unextractable residues are not all related 
to QCA.
    As discussed above, residues of DCBX, a residue of carcinogenic 
concern, have been detected in edible tissues longer than 72 hours post 
dosing

[[Page 21570]]

(Refs. 2, 5, and 6). The previous underlying conclusions that 
unextracted residues are related to noncarcinogenic compound, QCA, and 
extractable QCA is the only residue detectable in the edible tissues 72 
hours post dosing is no longer justified based on new scientific 
evidence.
    3. Previous Conclusion 3: No residue of carcinogenic concern even 
below the S0, is detectable by any method beyond 72 
hours.\11\
---------------------------------------------------------------------------

    \11\ This underlying conclusion is part of the basis of the 
January 1998 supplemental approval (FOI Summary) (Ref. 1 at pp. 13-
14).
---------------------------------------------------------------------------

    Boison, et al., 2009, reports a method capable of detecting DCBX at 
0.05 ppb, which is below the 0.061 ppb So and below the 
Sm of 0.305 ppb in muscle, 0.915 ppb in liver, and 1.83 ppb 
in kidney and fat. The method is also capable of measuring QCA at 0.500 
ppb, below the current tolerance of 30 ppb (Ref. 4 at p. 132, Table 5). 
Consequently, measurement of the relationship of QCA to at least one 
residue of carcinogenic concern, DCBX, is now scientifically feasible 
at the time the last tissue depletes to its Sm. In fact, 
Boison, et al., 2009, reports the presence of DCBX at a concentration 
greater than 0.050 ppb in the diaphragm (muscle) of 2 of 6 market-
weight hogs fed carbadox, when QCA was not detected, at a limit of 
quantitation of 0.50 ppb, in the livers of those same hogs (Ref. 4 at 
pp. 132-133). This evidence raises a serious question about whether QCA 
at 30 ppb is an appropriate marker residue for carbadox residues of 
carcinogenic concern. Based on this new scientific evidence, the 
previous underlying conclusion that no residue of carcinogenic concern, 
even below the SO, is detectable by any method beyond 72 
hours is no longer justified.
    4. Previous Conclusion 4: QCA is a reliable marker residue for 
carbadox and its metabolites.\12\
---------------------------------------------------------------------------

    \12\ This underlying conclusion is part of the basis of the 
January 1998 supplemental approval (FOI Summary) (Ref. 1 at pp. 13-
14).
---------------------------------------------------------------------------

    In light of the new evidence presented above, the conclusion that 
QCA is a reliable marker residue for carbadox and its metabolites is no 
longer justified because: (1) Previous conclusions made by the Agency 
are no longer scientifically justified and (2) the relationship of QCA 
to a carbadox residue of carcinogenic concern, DCBX, in the last tissue 
to deplete to its Sm is not known.

D. CVM's Reanalysis of the Human Health Risk From Previously Submitted 
Residue Data

    CVM reevaluated the existing carbadox residue data as a result of 
discussions that took place during meetings in 2011 with Phibro about 
the composition of total residues of carbadox (Refs. 3 and 22). CVM 
also reexamined the residue data submitted in support of the 1998 NADA 
supplements in light of the new understanding from the 2003 JECFA 
report that carcinogenic residues of carbadox persisted in edible 
tissues for 15 days, which was the last sampling time point, and that 
the previously unextractable residues are not necessarily 
noncarcinogenic residues related to QCA (Ref. 2).
    Using data in the FOI Summary for the January 30, 1998, 
supplemental approval, CVM reviewed information on total residue 
concentrations (measured from total radioactivity present in tissue 
from swine administered the radiolabeled drug), as well as the percent 
of total residues represented by QCA--the only noncarcinogenic 
metabolite of carbadox identified and quantified in the total residues 
of carbadox (Ref. 1). CVM used the total residue data and the percent 
of total residues represented by QCA to calculate the total residue of 
carcinogenic concern present in liver. Under the SOM regulations, 
``residues of carcinogenic concern'' in edible tissues are total 
residues of a carcinogenic drug minus identified residues that are 
judged by CVM to be noncarcinogenic (Sec.  500.82(b)). CVM previously 
excluded the unextracted portions of total residues from carcinogenic 
concern because it believed they were noncarcinogenic, QCA-related 
residues. The data presented to JECFA in 2003 now refute that 
conclusion, and CVM has no information, from Phibro or otherwise, that 
identifies or measures noncarcinogenic residues other than QCA in total 
residues of carbadox at the withdrawal period. As such, CVM now 
identifies the total residue of carcinogenic concern by subtracting QCA 
(identified residues that are confirmed to be noncarcinogenic) from 
total residues of carbadox. Determining the concentration of residues 
of carcinogenic concern present in the liver allowed CVM to compare 
that value with the Sm established for residues of 
carcinogenic concern in liver.
    CVM reviewed data regarding concentrations of total residues in 
swine tissues following 5 days of feeding \14\C-carbadox contained in a 
residue depletion study (the same study submitted to JECFA for its 1990 
evaluation of carbadox (Ref. 10 at p. 31)) submitted by the sponsor in 
support of the supplemental application to NADA 041-061 approved in 
January 1998 (Ref. 1, Study No. 1525N-60-87-005). The study measured 
concentrations of total residues of \14\C-carbadox and residues of QCA. 
Using these data, the study reported QCA as a mean percentage of the 
total residues of carbadox. QCA represented 24.4 percent of the total 
residues at 30 days, 27.5 percent at 45 days, and 9.9 percent at 70 
days post dosing (Ref. 1 at p. 13, Table 9).
    Table 1 presents total carbadox residues and total carbadox 
residues minus the noncarcinogenic QCA. Column 1 lists the sampling 
time point when swine were slaughtered following administration of the 
last dose of carbadox. Column 2 presents mean total residues measured 
in livers collected from swine slaughtered at each time point. Column 3 
lists the mean QCA percentage of total residues at each time point. 
Column 4 lists the calculated mean total residues of carcinogenic 
concern based on a subtraction of QCA from the mean total residue 
values in Column 2.

   Table 1--Mean Total Residues Measured as \14\C-Carbadox Equivalents, the Mean Percentage of Total Residues
Represented by QCA, and Mean Total Residue of Carcinogenic Concern in Liver of Swine (n=3 or 4) Following 5 Days
                                       of Feeding \14\C-Carbadox at 55 PPM
----------------------------------------------------------------------------------------------------------------
                                                                                                   Total residue
                                                                                                        of
                        Days post dosing                          Total residues    Percent QCA    carcinogenic
                                                                       (ppb)                       concern (ppb)
                                                                                                        \1\
----------------------------------------------------------------------------------------------------------------
30..............................................................            74.5            24.4            56.3
45..............................................................            20.0            27.5            14.5

[[Page 21571]]

 
70..............................................................            13.3             9.9           11.98
----------------------------------------------------------------------------------------------------------------
\1\ Values calculated by subtracting noncarcinogenic QCA portion from total residues.

    FDA first approved the use of carbadox in 1972 prior to the 
issuance of the Agency's SOM regulations. CVM did not make a 
calculation comparing total residues less QCA to the Sm in 
approving the January 1998 NADA supplement because the data available 
at the time indicated that DCBX was not detectable beyond 72 hours post 
dosing (by the analytical method used at the time) and because CVM 
believed all unextractable residues were noncarcinogenic residues 
related to QCA (Ref. 1). No residue depletion data presented to the 
Agency in original or supplemental NADAs showed that carcinogenic 
residues persisted beyond 72 hours or that the unextractable residues 
were carcinogenic. As a result, CVM did not, at that time, ask for data 
regarding the composition of total residues beyond establishing QCA as 
an appropriate marker residue. New evidence presented to JECFA in 2003 
and reported by Boison, et al., 2009, and Baars, et al., 1991, calls 
CVM's prior conclusions into question and places new significance on 
the concentrations of total residues of carcinogenic concern for 
carbadox (Refs. 2, 4, and 6).
    The individual data shown as mean values in Table 1 were used to 
predict total residues of carcinogenic concern at the approved 42-day 
withdrawal period for carbadox in NADAs 041-061 and 141-211, and the 
approved 70-day withdrawal period for carbadox in NADA 092-955. CVM 
analyzed the data using the logarithm of the dependent variable 
(carbadox-equivalents in liver). The logarithmic transformation or 
``exponential model'' is consistent with the published JECFA analyses 
of carbadox and commonly observed elimination behavior of 
pharmaceuticals (Ref. 22). Using this modeling procedure, the total 
residues of carcinogenic concern at 42 days are estimated to be 27 ppb 
with a 95 percent prediction interval of 9 ppb to 80 ppb (Ref. 3 at p. 
17, Table 8). These predictions can be compared with the Sm 
for swine liver of 0.915 ppb. The regression model predicts that swine 
liver concentrations of total carcinogenic residues will be 
significantly in excess of the Sm--approximately 30-fold (27 
ppb / 0.915 ppb = 29.51) greater residues of carcinogenic concern than 
the Sm at the approved 42-day withdrawal period for NADAs 
041-061 and 141-211 (Ref. 3 at p. 16). Total residues of carcinogenic 
concern at 70 days are estimated to be 10 ppb with a 95 percent 
prediction interval of 3 ppb to 32 ppb (Ref. 3 at p. 17, Table 8). The 
analysis predicts that swine liver concentrations of total carcinogenic 
residues will be significantly in excess of the Sm--
approximately 11-fold greater residues of carcinogenic concern than the 
Sm at the approved 70-day withdrawal period for NADA 092-
955.
    Approval of a carcinogenic new animal drug under the DES Proviso to 
the Delaney Clause requires development of a sufficiently sensitive 
regulatory method that detects no residues of carcinogenic concern in 
the edible tissues of food-producing animals from the use of the animal 
drug. New evidence raises serious questions about whether the currently 
approved tolerance for uses of carbadox is adequate under the SOM 
regulations, and raises serious questions about the continued approval 
of the compound under the DES Proviso exception to the Delaney Clause 
due to the lack of a sufficiently sensitive regulatory method.
    Carbadox is currently approved based upon CVM's previous conclusion 
that unextractable residues were QCA related and noncarcinogenic. Given 
this conclusion and the fact that no residues of carcinogenic compounds 
were detectable by any method beyond 72 hours, CVM determined that QCA 
was an acceptable marker residue and established the tolerance at 30 
ppb. New evidence presented to JECFA in 2003 undermines the conclusion 
that all unextractable residues at the withdrawal period are QCA 
related. As a result, under FDA's SOM regulations, all unextractable 
residues except for measured residues of QCA must be considered 
residues of carcinogenic concern (Sec.  500.82(b)). Under CVM's 
analysis (Table 1), concentrations of total residues of carcinogenic 
concern in liver are approximately 30 times higher than the 
Sm at the approved 42-day withdrawal period and 11 times 
higher at the approved 70-day withdrawal period (Ref. 3 at pp. 16-17). 
CVM would expect that total residues of carcinogenic concern would also 
exceed the Sm when QCA reaches the approved tolerance of 30 
ppb in liver. CVM can no longer conclude that when QCA is at or below 
30 ppb, the residues of carcinogenic concern are present at or below a 
concentration that would present no significant increase in the risk of 
cancer to humans (Sec.  500.86(c)).
    The new evidence indicates that QCA is not an appropriate marker 
residue for residues of carcinogenic concern and that QCA at 30 ppb in 
swine liver is not an appropriate tolerance. The new evidence also 
shows that the approved regulatory method for all approved carbadox 
NADAs is inadequate under the SOM regulations (part 500, subpart E). 
The inadequacy of the regulatory method is a basis for withdrawal of 
approval of all carbadox NADAs under section 512(e)(1)(B) of the FD&C 
Act. See Sponsored Compounds in Food-Producing Animals; Criteria and 
Procedures for Evaluating the Safety of Carcinogenic Residues, Proposed 
Rule, preamble to the proposed SOM regulations II (50 FR 45530 at 
45550).
    Similarly, these findings demonstrate that carbadox is no longer 
shown to be safe under the General Safety Clause because residues of 
carcinogenic concern remain in swine tissue well past the established 
withdrawal period. Under the General Safety Clause, drug residues must 
be determined to be safe based on all available evidence. Where a drug 
is a known mutagenic carcinogen and new evidence shows that 
unidentified residues of carcinogenic concern are present at the 
established withdrawal time, the drug is no longer shown to be safe. 
See Section III.D.
    As stated previously, the new evidence presented to JECFA 
undermines the previously held conclusion that all unextracted residues 
are QCA related and noncarcinogenic. Because carbadox is a mutagenic 
carcinogen, all otherwise unidentified

[[Page 21572]]

residues are treated as carcinogenic. No evidence has been presented to 
CVM by Phibro or any other source to show that the unidentified 
residues are noncarcinogenic or that the residues do not otherwise 
present a threat to public health. As a result, carbadox is not shown 
to be safe under the General Safety Clause.

VI. Notice of Opportunity for a Hearing

    New evidence regarding carcinogenic residues in edible tissues of 
swine treated with carbadox raises serious questions about the human 
food safety of the drug. Therefore, CVM is proposing to withdraw 
approval of the three NADAs that provide for use of carbadox in swine 
feed because new evidence demonstrates that the drug no longer meets 
the DES Proviso exception to the Delaney Clause and because new 
evidence demonstrates that carbadox is not shown to be safe under the 
General Safety Clause.
    Therefore, notice is given to Phibro Animal Health Corp., 65 
Challenger Rd., Ridgefield Park, NJ 07660, and to all other interested 
persons, that the Director of CVM proposes to issue an order under 
section 512(e) of the FD&C Act withdrawing approval of all NADAs 
providing for use of carbadox in medicated swine feed.
    In accordance with section 512 of the FD&C Act and part 514 (21 CFR 
part 514) and under the authority delegated to the Director of CVM, 
Phibro Animal Health Corp., the sponsor, is hereby given an opportunity 
for hearing to show why approval of NADAs 041-061, 092-955, and 141-211 
should not be withdrawn.
    If the sponsor, Phibro Animal Health Corp., wishes to request a 
hearing the sponsor must file: (1) On or before [see DATES], a written 
notice of appearance and request for a hearing and (2) on or before 
[see DATES], the data, information, and analyses relied on to 
demonstrate that there is a genuine and substantial issue of fact to 
justify a hearing as specified in Sec.  514.200. Any other interested 
person may also submit comments on this notice (see, ADDRESSES). 
Procedures and requirements governing this NOOH, a notice of appearance 
and request for a hearing, submission of data, information, and 
analyses to justify a hearing, other comments, and a grant of denial of 
a hearing, are contained in Sec.  514.200 and 21 CFR part 12.
    The failure of a holder of an approval to file timely a written 
appearance and request for hearing as required by Sec.  514.200 
constitutes an election not to avail himself or herself of the 
opportunity for a hearing and a waiver of any contentions concerning 
the legal status of any such drug product, and the Director of CVM will 
summarily enter a final order withdrawing the approvals. Any new animal 
drug product marketed without an approved NADA is subject to regulatory 
action at any time.
    A request for a hearing may not rest upon mere allegations of 
denials, but must set forth specific facts showing that there is a 
genuine and substantial issue of fact that requires a hearing. If it 
conclusively appears from the face of the data, information, and 
factual analyses in the request for hearing that there is no genuine 
and substantial issue of fact that precludes the withdrawal of approval 
of the applications, or when a request for hearing is not made in the 
required format or with the required analyses, the Commissioner of Food 
and Drugs will enter summary judgment against the person who requests a 
hearing, making findings and conclusions, and denying a hearing.
    If a hearing is requested and is justified by the sponsor's 
response to this NOOH, the issues will be defined, a presiding officer 
will be assigned, and a written notice of the time and place at which 
the hearing will commence will be issued as soon as practicable.
    This notice is issued under section 512 of the FD&C Act and under 
the authority delegated to the Director of CVM.

VII. Environmental Impact

    The Agency has determined under 21 CFR 25.33(g) that this action is 
of a type that does not individually or cumulatively have a significant 
impact on the human environment. Therefore, neither an environmental 
assessment nor an environmental impact statement is required.

VIII. Paperwork Reduction Act of 1995

    The collections of information requirements for this document are 
covered under OMB control numbers 0910-0032 and 0910-0184.

IX. References

    The following references have been placed on display in the 
Division of Dockets Management (see ADDRESSES) and may be seen by 
interested persons between 9 a.m. and 4 p.m., Monday through Friday, 
and are available electronically at http://www.regulations.gov. (FDA 
has verified the Web site addresses, but FDA is not responsible for any 
subsequent changes to the Web sites after this document publishes in 
the Federal Register.)

1. FDA, Freedom of Information (FOI) Summary, NADA 041-061, MECADOX 
10 (carbadox) Type A medicated article, supplemental approval 
January 30, 1998. Available at http://www.fda.gov/downloads/AnimalVeterinary/Products/ApprovedAnimalDrugProducts/FOIADrugSummaries/UCM429999.pdf (accessed on March 19, 2016).
2. JECFA, Report on Carbadox, 2003. Available at ftp://ftp.fao.org/ag/agn/jecfa/vetdrug/41-15-carbadox.pdf (accessed on March 19, 
2016).
3. FDA, Memorandum to the File, Claycamp, H.G., ``Preliminary Risk 
Characterization: Cancer Risk Estimation from Carbadox Residues in 
Pork from Swine Treated with Carbadox,'' December 16, 2014.
4. Boison, J.O., S.C. Lee, and R.G. Gedir, ``A Determinative and 
Confirmatory Method for Residues of the Metabolites of Carbadox and 
Olaquindox in Porcine Tissues,'' Analytica Chimica Acta, 637:128-
134, 2009.
5. Baars, A.J., L.A. van Ginkel, M.M.L. Aerts, et al., ``Kinetics of 
Carbadox Residues in Pigs,'' In: Proceedings of the EuroResidue 
Conference, Noorwijkerhout (Haagsma, N., A. Ruiter, and P.B. Czedik-
Eysenberg, eds., May 21-23, 1990.
6. Baars, A.J., L.P. Jager, T.J. Spierenberg, et al., ``Residues of 
Carbadox Metabolites in Edible Pork Products,'' Archives of 
Toxicology Supplement, 14:288-92, 1991.
7. FDA, CVM Guidance for Industry (GFI) #3, ``General Principles for 
Evaluating the Safety of Compounds Used in Food-Producing Animals,'' 
July 25, 2006. Available at http://www.fda.gov/downloads/animalveterinary/guidancecomplianceenforcement/guidanceforindustry/ucm052180.pdf (accessed on March 19, 2016).
8. FDA, Memorandum to the File, from Director, Division of New 
Animal Drugs to Director, Bureau of Veterinary Medicine regarding 
NADA 41-061--Carbadox for Swine (September 22, 1972).
9. FDA, Memorandum to the File, S.H. Frazier, Jr., Division of 
Toxicology, to Director, Bureau of Veterinary Medicine, regarding 
Carbadox for Swine, August 27, 1970.
10. JECFA, Report on Carbadox, 1990. Available at ftp://ftp.fao.org/ag/agn/jecfa/vetdrug/41-3-carbadox.pdf (accessed on March 19, 2016).
11. FDA, Memorandum to the File, from Division of New Animal Drugs 
to Director, Bureau of Veterinary Medicine, regarding NADA 41-061, 
Carbadox for Swine (July 7, 1972).
12. FDA, Memorandum to the File, Approval of Original New Animal 
Drug Application NADA 92-955 (July 29, 1975).
13. Citizen Petition, Center for Science in the Public Interest, 
Docket No. FDA-1986-P-0299 (formerly 86P-0212), May 9, 1986.
14. FDA, Response to Citizen Petition, Center for Science in the 
Public Interest, Docket No. FDA-1986-P-0299 (formerly 86P-0212), May 
30, 1995.
15. FDA, Memorandum to the File, from Residue Evaluation Branch, 
Division of Chemistry to Director, Division of

[[Page 21573]]

Chemistry, regarding Review of Carbadox Metabolism (September 7, 
1994).
16. FDA, Freedom of Information (FOI) Summary, NADA 041-061, MECADOX 
10 (carbadox) Type A medicated article, supplemental approval 
October 5, 1998. Available at http://www.fda.gov/AnimalVeterinary/Products/ApprovedAnimalDrugProducts/FOIADrugSummaries/ucm064223.htm 
(accessed on March 19, 2016).
17. FDA, Freedom of Information (FOI) Summary, NADA 141-211, MECADOX 
10 (carbadox) and TERRAMYCIN 50, 100, or 200 (oxytetracycline) in 
Type C medicated feed, original approval July 21, 2004. Available at 
http://www.fda.gov/downloads/AnimalVeterinary/Products/ApprovedAnimalDrugProducts/FOIADrugSummaries/ucm118005.pdf (accessed 
on March 19, 2016).
18. Codex Alimentarius Commission, Twenty-Eighth Session, 
Headquarters, Food and Agriculture Organization, Rome, Italy, 2005.
19. FDA, CVM Guidance for Industry (GFI) #205, VICH GL 46, ``Studies 
to Evaluate the Metabolism and Residue Kinetics of Veterinary Drugs 
in Food-Producing Animals: Metabolism Study to Determine the 
Quantity and Identify the Nature of Residues (MRK),'' September 15, 
2011. Available at http://www.fda.gov/downloads/AnimalVeterinary/GuidanceComplianceEnforcement/GuidanceforIndustry/UCM207939.pdf 
(accessed on March 19, 2016).
20. MacIntosh, A.I., G. Lauriault, and G.A. Neville, ``Liquid 
Chromatographic Monitoring of the Depletion of Carbadox and its 
Metabolite Desoxycarbadox in Swine Tissues,'' Journal--Association 
of Official Analytical Chemists, 68:665-71, 1985.
21. FDA, CVM Guidance for Industry (GFI) #208, VICH GL 49, ``Studies 
to Evaluate the Metabolism and Residue Kinetics of Veterinary Drugs 
in Food-Producing Animals: Validation of Analytical Methods Used in 
Residue Depletion Studies,'' September 15, 2011. Available at http://www.fda.gov/downloads/AnimalVeterinary/GuidanceComplianceEnforcement/GuidanceforIndustry/UCM207942.pdf 
(accessed on March 19, 2016).
22. FDA, Memorandum to the File, Claycamp, H. G., Verification and 
Extension of the 2003 JECFA Carbadox Monograph Analyses, July 29, 
2012.

    Dated: April 6, 2016.
Tracey Forfa,
Acting Director, Center for Veterinary Medicine.
[FR Doc. 2016-08327 Filed 4-8-16; 11:15 am]
 BILLING CODE 4164-01-P