Blood Supply: FDA Oversight and Remaining Issues of Safety (Chapter
Report, 02/25/97, GAO/PEMD-97-1).

Pursuant to a congressional request, GAO evaluated the Food and Drug
Administration's (FDA) "layers of safety" that provide the framework for
regulating and monitoring the U.S. blood industry, focusing on the
actual and potential vulnerabilities in the layers of safety that may
present a threat to the public health.

GAO found that: (1) the transmission of human immunodeficiency virus
(HIV) by transfusion decreased dramatically after HIV testing for donors
was introduced in 1985, and more and better tests for other diseases
also have reduced the the risks from blood transfusions; (2) while the
blood supply is very safe, no amount of federal regulation can entirely
eliminate blood-transfusion risks because of human error, technological
limitations of state-of-the-art tests, and the biological nature of the
product itself; (3) within the overlapping layers of safety, GAO found
areas where FDA can take action that would further improve the safety of
the blood supply: (a) the lack of a uniform donor questionnaire allows
variability in donor screening; (b) the lack of mandatory deferral
notification allows some donors who have tested positive for viruses to
unwittingly attempt donation again; (c) untested units donated for
self-use may inadvertently be used for unintended recipients; and (d)
FDA has been slow to investigate error and accident reports that may
warrant a recall; (4) FDA does not require unlicensed facilities, those
that do not engage in the sale, barter, or exchange of blood products
across state lines, to report errors and accidents; (5) because
unlicensed facilities constitute more than two thirds of all blood
facilities that, together, produce 10 percent of the nation's blood, FDA
has not fully monitored the quality of this portion of blood products;
(6) FDA's inspections for both licensed and unlicensed blood facilities
appear to be inconsistent in focus, scope, and documentation; (7) in
addition, these inspections are often not conducted within time periods
set by FDA's own guidelines; (8) FDA does not maintain a central
repository for inspection reports and, thus, does not examine national
trends; and (9) GAO's survey results also indicated confusion within the
blood industry regarding the interpretation for FDA policy guidance and
regulations.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  PEMD-97-1
     TITLE:  Blood Supply: FDA Oversight and Remaining Issues of Safety
      DATE:  02/25/97
   SUBJECT:  Acquired immunodeficiency syndrome
             Infectious diseases
             Safety regulation
             Health hazards
             Inspection
             Testing
             Reporting requirements
             Product safety
             Medical supplies
             Quality control
IDENTIFIER:  Medicare Program
             Medicaid Program
             AIDS
             FDA Program Oriented Data System
             FDA Error and Accident Reporting System
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Ranking Minority Member, Committee on Commerce, House
of Representatives

February 1997

BLOOD SUPPLY - FDA OVERSIGHT AND
REMAINING ISSUES OF SAFETY

GAO/PEMD-97-1

Blood Supply:  Oversight and Safety Issues

(973418)


Abbreviations
=============================================================== ABBREV

  AABB - American Association of Blood Banks
  ABC - America's Blood Centers
  ABRA - American Blood Resources Association
  ALT - Alanine aminotransferase
  ARC - American Red Cross
  BSI - Blood Systems Incorporated
  CBER - Center for Biologics and Evaluation Review
  CCBC - Council of Community Blood Centers
  CDC - Centers for Disease Control and Prevention
  CJD - Creutzfeldt-Jakob disease
  CMV - Cytomegalovirus
  CUE - Confidential unit exclusion
  DDR - Donor deferral registry
  EAR - Error and accident report
  EARS - Error and Accident Reporting System
  EIR - Establishment inspection report
  FDA - Food and Drug Administration
  HAV - Hepatitis A virus
  HBV - Hepatitis B virus
  HBc - Hepatitis B core
  HBsAg - Hepatitis B surface antigen
  HCFA - Health Care Financing Administration
  HCV - Hepatitis C virus
  HDV - Hepatitis D virus
  HEV - Hepatitis E virus
  HGV - Hepatitis G virus
  HHS - Department of Health and Human Services
  HIV - Human immunodeficiency virus
  HTLV - Human T-lymphotropic virus
  IMIG - Intramuscular immune globulin
  IPPIA - International Plasma Products Industry Association
  IVIG - Intravenous immune globulin
  NHLBI - National Heart, Lung, and Blood Institute
  NIH - National Institutes of Health
  PODS - Program-oriented data system
  SOP - Standard operating procedure
  UBS - United Blood Services

Letter
=============================================================== LETTER


B-271101

February 25, 1997

The Honorable John D.  Dingell
Ranking Minority Member
Committee on Commerce
House of Representatives

Dear Mr.  Dingell: 

You asked us to evaluate the Food and Drug Administration's "layers
of safety" that provide the framework for regulating and monitoring
the U.S.  blood industry.  Specifically, you asked us to examine the
actual and potential vulnerabilities in the layers of safety that may
present a threat to the public health.  In this report, we address
these potential vulnerabilities in light of changes in the blood
industry that have occurred since the mid-l980s, when there was
widespread concern about the safety of the nation's blood supply. 

You also asked us to examine the disparate estimates of
transfusion-associated AIDS and hepatitis cases and asked that we
determine the current risks of these viruses in the blood supply. 
This information, as well as information on other risks known to
occur as a result of blood transfusions, is contained in our 1997
report entitled Blood Supply:  Transfusion-Associated Risks
(GAO/PEMD-97-2). 

As we arranged with your office, unless you publicly announce the
report's contents earlier, we plan no further distribution until 15
days after the date of this letter.  We will then send copies of this
report to the Secretary of Health and Human Services, the
Commissioner of the Food and Drug Administration, and others who are
interested.  If you have any questions or would like additional
information, please call me at (202-512-3652).  Major contributors to
this report are listed in appendix V. 

Sincerely,

Kwai-Cheung Chan
Director of Program Evaluation in Physical
 Systems Areas


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

Approximately 4 million patients annually receive life-saving
transfusions of blood donated by 8 million donors around the nation. 
AIDS and the possibility of contracting HIV through blood
transfusions have nonetheless focused public attention on the safety
of this blood.  Representative
John D.  Dingell, the ranking minority member of the House Committee
on Commerce, asked the General Accounting Office (GAO) to identify
issues that might threaten the nation's blood supply.  Therefore,
this report answers the question, What are the elements of the Food
and Drug Administration's (FDA's) layers of blood safety and do they
ensure that the blood supply is safe? 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

In testimony on July 28, 1993, before the Subcommittee on Oversight
and Investigations of the House Committee on Energy and Commerce, the
Commissioner of FDA outlined five overlapping "layers of safety" that
provided a framework to regulate and monitor the blood industry:  (1)
donor screening, (2) donor deferral registries, (3) viral testing,
(4) quarantining blood until tests and control procedures have
established its safety, and (5) monitoring and investigating adverse
incidents to ensure that deficiencies are corrected. 

Since the mid-1980s, the blood industry, with the assistance of FDA,
has instituted standard operating procedures, quality assurance
programs, and good manufacturing procedures that have improved donor
screening, blood collection, viral testing, and how blood is stored
and distributed.  These actions have improved the overall safety of
the blood supply, as discussed in a companion GAO report, Blood
Supply:  Transfusion-Associated Risks (GAO/PEMD-97-2), that examined
the risks of contracting AIDS and hepatitis from blood as well as
other known hazards of blood transfusion, comparing these to other
health-related risks. 

In this report, GAO examined the five layers to identify areas of
potential improvement that would further improve blood safety.  GAO
reviewed FDA's regulations and guidelines issued between 1989 and the
present, interviewed FDA officials and blood industry
representatives, visited blood facilities, and attended technical
conferences and FDA workshops.  GAO also assessed 1990-94 FDA error
and accident reports to assess lapses in quality control and
collected FDA inspection reports from a nationally representative
sample of blood facilities.  GAO's analysis of these data is the
first and only source of this information on a national level. 
Finally, GAO queried quality-control directors about the focus and
scope of FDA's inspections and possible changes in FDA's policy to
enhance compliance and overall safety. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

The transmission of HIV by transfusion decreased dramatically after
HIV testing for donors was introduced in 1985, and more and better
tests for other diseases also have reduced the risks from blood
transfusions.  While the blood supply is very safe, no amount of
federal regulation can entirely eliminate blood-transfusion risks
because of human error, technological limitations of state-of-the-art
tests, and the biological nature of the product itself. 

Within the overlapping layers of safety, GAO found areas where FDA
can take action that would further improve the safety of the blood
supply.  For example: 

  -- lack of a uniform donor questionnaire allows variability in
     donor screening,

  -- lack of mandatory deferral notification allows some donors who
     have tested positive for viruses to unwittingly attempt donation
     again,

  -- untested units donated for self-use may inadvertently be used
     for unintended recipients, and

  -- FDA has been slow to investigate error and accident reports that
     may warrant a recall. 

FDA does not require unlicensed facilities--those that do not engage
in the sale, barter, or exchange of blood products across state
lines--to report errors and accidents.  Because unlicensed facilities
constitute more than two thirds of all blood facilities that,
together, produce 10 percent of the nation's blood, FDA has not fully
monitored the quality of this portion of blood products. 

FDA's inspections for both licensed and unlicensed blood facilities
appear to be inconsistent in focus, scope, and documentation.  In
addition, these inspections are often not conducted within time
periods set by FDA's own guidelines.  Furthermore, FDA does not
maintain a central repository for inspection reports and, thus, does
not examine national trends.  GAO's survey results also indicated
confusion within the blood industry regarding the interpretation of
FDA policy guidance and regulations. 


   PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4

The blood industry has made many positive changes in collecting and
processing blood in response to FDA initiatives.  Facilities have
standard operating procedures and good manufacturing practices that
detail how to ensure high-quality products.  Donor education and
screening exclude donors with known risk factors or diseases. 
Deferral registries of donors whose blood is unsuitable are
maintained and consulted.  Viral testing with powerful screening
tests eliminates most infectious products, and products are
quarantined from the general supply until they have been found to
meet current requirements. 

Nevertheless, some facilities do not use uniform donor
questionnaires, do not adequately ensure privacy during donor
screening, or do not notify donors who have been permanently
deferred.  Bacterial contamination of platelets is increasingly
recognized but FDA does not require blood facilities'
quality-assurance programs to include processes that monitor for
bacterial contamination. 

Seven tests are routinely used to screen blood, and others are
available that reduce the risk of transmitting diseases through blood
transfusions.  However, FDA does not require additional, confirmatory
testing on units that test positive for viral markers except for HIV. 
FDA requires that blood facilities notify consignees (that is,
transfusion services) that receive blood from donors who subsequently
test positive for HIV, and these consignees are required to attempt
to notify recipients of the units.  However, there are no
requirements for notifying consignees or recipients of blood that
subsequently test positive for other viruses, even though
confirmatory tests and treatments are available for some of these
viruses and patients who might be notified could take steps to
prevent transmission of infection to others. 

FDA requires that blood that donors give for their own use proceed
through elaborate systems to ensure that it is transfused to the
correct patient.  However, FDA does not require facilities to test
such units for viruses, and some do not.  Studies have indicated that
untested units can make their way into the blood supply system and
can be transfused to unintended recipients. 

GAO identified no major safety problems in quarantining blood, but
the data indicate that there are problems in inventory management in
that many units are unaccounted for or lost before they can be
transfused.  This is not directly a safety issue but could contribute
to instances of blood supply shortages. 

Unlicensed facilities are not required to report errors and
accidents, and in 1994 they submitted only 1 percent of all error and
accident reports, although they collected 10 percent of the U.S. 
blood supply.  Without full reporting of errors and accidents, FDA is
unable to monitor the quality control of the entire industry. 
Further, in a nationally representative sample of establishment
inspection reports, GAO found that more than half of all observations
of problems by FDA inspectors were issued to unlicensed facilities. 
The discrepancy between the proportions of problems observed and the
voluntarily reported errors and accidents by unlicensed facilities
underscores the need for better FDA oversight. 

FDA publishes its positions on some important industry issues as
guidelines and memoranda, but they are often ambiguous in content and
intent, and no public comment is required.  Additionally, although
inspections are the primary means by which FDA ensures the safety of
the blood supply, it does not perform statistical analyses of
inspection reports to identify trends in deviations or variability in
the implementation of inspection policies.  GAO also found problems
relating to FDA's ability to discriminate between facilities that are
in and out of compliance and to inspect them in a timely manner. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5

GAO recommends that the Secretary of Health and Human Services (HHS)
require blood facilities to

  -- notify all donors who are permanently deferred that they have
     been deferred and the medical reasons for their deferral. 

  -- require blood facilities' quality-assurance programs to include
     processes that monitor for bacterial contamination. 

  -- require viral testing for all self-donated blood units in order
     to minimize the potential vulnerability of untested autologous
     units entering the blood supply. 

  -- require confirmatory testing of all repeatedly reactive viral
     test results for which there is a licensed confirmatory test. 

  -- require that transfused patients be notified when they have been
     transfused with blood from a donor whose subsequent donations
     were found to be positive by confirmatory testing.  The
     reasonable time period for tracing back units to recipients
     varies with each virus, and decisions should be made in
     consultation with the blood industry. 

  -- require the identification of implicated units that have not
     been transfused or further manufactured. 

  -- require unlicensed facilities to report all errors and
     accidents. 

Additionally, GAO recommends that the Secretary

  -- publish in the form of regulations the guidelines that FDA
     believes are essential to ensure the safety of the nation's
     blood supply.  FDA should clarify its position on the extent to
     which facilities should adopt its guidelines and memoranda in
     order to remain in compliance with the agency's regulations. 

  -- correct problems GAO identified in FDA inspection processes--FDA
     should perform statistical analyses of inspection reports,
     develop policies for FDA inspectors to list on inspection
     reports the activities they observe, publish better guidance on
     the types of activities that warrant reports on deviations and
     warning letters, and ensure that all blood facilities are
     inspected in a timely fashion. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6

In a written response to a draft of this report, HHS generally
concurred with GAO's findings and recommendations regarding donor
deferral notification, quality assurance for bacterial contamination,
viral marker testing of self-donated units, error and accident
reporting by unlicensed facilities, and clarification of FDA guidance
to blood establishments. 

HHS did not fully concur with GAO's recommendation on requiring
confirmatory testing and consignee and recipient notification for
diseases other than HIV.  HHS concurred that confirmatory testing is
important and pointed out that it has recommended such testing for
hepatitis B and hepatitis C.  However, this procedure is only
recommended by FDA; it is not a required activity.  HHS disagreed
that there should be procedures in place to notify recipients of
units from donors who subsequently test positive for viruses other
than HIV ("lookback").  However, hepatitis, like HIV, can be
transmitted to others; recent studies suggest that there are
effective therapies for some patients with hepatitis; and informed
patients can curtail certain behaviors (such as consuming alcohol)
that could cause more progressive harm after being infected with
hepatitis. 

HHS also disagreed with GAO's recommendation regarding problems
identified in FDA's inspection processes by stating that FDA already
reviews and analyzes inspection reports and has several manuals and
compliance programs to guide its inspectors.  However, GAO found that
FDA does not perform statistical analyses of inspection reports that
would result in information whereby FDA could determine compliance
rates among blood facilities.  Also, GAO found differences in the
number and kind of observations of problems across FDA districts as
well as inconsistencies in the application of official observations
and warning letters. 

HHS also provided a number of technical comments, which GAO
incorporated into the report as appropriate. 


INTRODUCTION
============================================================ Chapter 1

Since the human immunodeficiency virus (HIV) was introduced into the
U.S.  blood supply in the early 1980s, the benefits of a potentially
life-saving transfusion have had to be weighed against the risks
posed by the most deadly disease known to be transmitted through
blood.  The risks posed by HIV have spurred many changes in how blood
is collected and processed.  Also, the blood industry is concerned
about bacterial contamination of the blood supply as well as viral
and nonviral agents known to be transmissible through blood such as
Chagas' disease, cytomegalovirus (CMV), hepatitis A-G, human T-cell
leukemia and lymphoma viruses (HTLV-I and HTLV-II), parvovirus, and
syphilis. 

In testimony on July 28, 1993, before the Subcommittee on Oversight
and Investigations of the House Committee on Energy and Commerce, the
Commissioner of the Food and Drug Administration (FDA), the agency
that has main responsibility for regulating the safety of blood
products, described "five layers of safety" that were present
throughout the blood industry to help ensure safe blood: 

1.  screening donors,

2.  maintaining donor deferral registries to eliminate unsuitable
donors from the rolls,

3.  testing blood,

4.  quarantining blood until tests and control procedures establish
its safety, and

5.  monitoring and investigating adverse incidents to ensure that
deficiencies are corrected. 

Subsequently, Congressman John D.  Dingell asked us to examine these
layers and FDA's implementation of programs and policies to ensure
the safety of the nation's blood products.  To do this, we answered
the following question:  What are the elements of FDA's layers of
blood safety and do they ensure that the blood supply is safe?\1


--------------------
\1 Congressman Dingell made this request when he was chairman of the
Energy and Commerce Committee of the U.S.  House of Representatives. 
He is now ranking minority member of the renamed House Committee on
Commerce.  Mr.  Dingell asked us at the same time to assess the risk
estimates of diseases transmitted through transfusion.  We have done
this in Blood Supply:  Transfusion-Associated Risks, GAO/PEMD-97-2
(Washington, D.C.:  1997), noting there that the blood supply is
safer than it has ever been and that, in terms of threats to life,
receiving a blood transfusion is much safer than many other
activities. 


   DONATED BLOOD AND ITS PRODUCTS
---------------------------------------------------------- Chapter 1:1

About 8 million volunteers donate approximately 14 million units of
whole blood each year.  This whole blood is rarely transfused into
patients.  Instead, blood services in the blood industry separate
each unit of whole blood into an average of 1.8 specialized
components that, in blood-banking terminology, are "products"
consisting of various types of blood cells, plasma, and special
preparations of plasma.  Health care facilities transfuse the
resulting 23 million components--4 to 5 units at a time, on
average--into as many as 4 million patients to treat specific
conditions such as anemia and hemophilia.  Donors give an additional
12 million units of plasma each year, for a total of approximately 26
million annual blood donations. 

Fewer than 5 percent of the Americans who are eligible to donate
blood each year actually do.  Most people donate at a blood drive
where they work.  The average blood donor is a college-educated white
male 30 to 50 years old, married, with an above-average income. 
These statistics are changing, however, as more white women and
minority men and women are entering the workforce. 

To be eligible to donate blood, a person should be at least 17 years
old, weigh at least 110 pounds, be in good physical health, and pass
a physical and medical history examination.\2 Men have about 12 pints
of blood in their circulatory system, women about 9.  At any one
time, donors give about 1 pint of blood each.  Interestingly, their
bodies replace this fluid in about 24-72 hours, although it may take
up to 2 weeks to replace the plasma proteins.  It normally takes 6-8
weeks to replenish the lost red blood cells from one unit of whole
blood.  Thus, those who donate whole blood may do so only once every
8 weeks.  Some states limit the number and frequency of donations a
person can make in a 12-month period.  In apheresis, specific
components of the blood are removed and the unremoved portions of the
blood are returned to the donor.  Because this preserves the donor's
red blood cells, apheresis donors usually can donate once every 48
hours but no more than twice a week.  (Apheresis is limited to 20
times a year.)

Red blood cells, commonly used to treat anemia, may be preserved as a
liquid for up to 42 days but they may also be frozen for up to 10
years.  Plasma can be kept frozen for up to 1 year and may be used to
control bleeding.  Cryoprecipitate contains clotting factors, useful
in controlling bleeding.  It is made from fresh frozen plasma and may
be kept for 1 year.  Platelets are important in controlling bleeding
and are used to treat patients with leukemia and other cancers; they
should be stored at room temperature for a maximum of 5 days.  White
blood cells are sometimes used to fight infections but should be
transfused as soon as possible after collection and must be
transfused within 24 hours of donation. 

In addition to separating blood into component products, plasma
facilities manufacture "derivative products" by fractionating plasma
chemically into concentrated proteins.  These include albumin, used
to treat shock; immune globulin, used to prevent certain infectious
diseases and to treat deficiencies of protein; clotting factor
concentrates, used to control bleeding in patients with clotting
factor deficiencies; and specific immune globulins, prepared from
plasmas collected from donors with antibodies to specific diseases
and then used to prevent those diseases in others.  Derivatives are
commonly made by commercial manufacturers.  Depending on the product,
they may pool plasma from as many as 60,000 donors for fractionation
in order to produce sufficient amounts of the final concentrated
material cost-effectively.  These therapies processed from plasma
also undergo viral and bacterial removal and inactivation procedures
that are effective in destroying most of these agents. 


--------------------
\2 There is no FDA minimum age requirement although some facilities
voluntarily implement an age requirement.  Donors weighing less than
110 pounds may donate provided that a proportionately smaller volume
of blood is drawn. 


   THE BLOOD SERVICES INDUSTRY
---------------------------------------------------------- Chapter 1:2

The blood services industry has a volunteer and a commercial sector. 
Voluntary donors are unpaid and usually donate whole blood. 
Commercial facilities collect plasma from paid donors for
manufacturing various derivatives.  Table 1.1 outlines the different
types of blood collection services and the amount of blood they
collect annually. 



                               Table 1.1
                
                U.S. Blood Collection Facilities and the
                        Blood Units They Collect

                                                           Commercial
                                    Volunteer sector         sector
                                ------------------------  ------------
                                                                Plasma
Type of facility                  Licensed    Unlicensed      center\a
------------------------------  ----------  ------------  ------------
Number of facilities                   308         2,274           463
Number of units collected             12.6           1.4            12
 (millions)
----------------------------------------------------------------------
\a All plasma centers are licensed

The three types of facilities in the volunteer sector are (1)
regional and community blood centers, which usually collect and
distribute blood and blood components to hospitals within
circumscribed geographical areas; (2) hospital blood facilities,
which collect and transfuse whole blood and blood components; and (3)
hospitals, which primarily store and transfuse blood but do not
collect it. 

Regional and community blood centers provide a full range of blood
services to a surrounding geographical area.  They generally collect,
test, and label blood, as well as distribute blood and blood products
to hospitals, physicians, and hemophilia care centers.  Hospital
blood facilities usually provide a smaller range of services, limited
to collecting and storing whole blood and its components.  Some
hospitals conduct their own viral testing, while others send blood
and blood products to outside laboratories for viral testing. 

The volunteer sector is represented by three organizations:  the
American Association of Blood Banks (AABB), the American Red Cross
(ARC), and America's Blood Centers (ABC), formerly known as the
Council of Community Blood Centers (CCBC).  ABC member centers
collect approximately 45 percent of all blood, ARC collects another
45 percent, and independent facilities collect the remaining 10
percent.  The members of the AABB include both ARC and the majority
of ABC member centers. 

AABB is the professional society of blood facilities and transfusion
services and it also includes individual members such as physicians,
scientists, nurses, and administrators, among others.  ABC is a
council of community based blood-collection facilities.  ARC is a
single corporation consisting of all ARC blood centers.  Until 1994,
ARC served as an organizational framework for its centers, each
operating somewhat independently and self-sufficiently.  In an
organizational change that began in 1994 and was completed in 1995,
ARC centralized and standardized its operations, reducing the number
of regions and limiting testing to a few centralized laboratories. 

The commercial sector, which is generally called the "source plasma
sector" and receives plasma from paid donors, has three main
components:  (1) collectors, or plasmapheresis centers; (2)
fractionators; and (3) brokers.  (Brokers do not collect source
plasma.) The plasmapheresis centers collect plasma that they either
sell to U.S.  fractionators (who manufacture derivatives such as
albumin from it) or export to fractionators in Europe, Japan, and
South America.  Some fractionators also operate their own source
plasma collection centers. 

Plasma brokers purchase and market recovered plasma from whole-blood
facilities (that is, the volunteer sector) and sell this directly to
fractionators.  Plasma is "recovered" after components have been
removed from whole blood or after whole blood has become outdated. 

The commercial sector is represented by the American Blood Resources
Association (ABRA), a nonprofit trade association that represents the
interests of businesses that collect certain biological products (in
particular, plasma) for further manufacturing.  This sector is also
represented by the International Plasma Products Industry Association
(IPPIA), which represents all the commercial processors of
plasma-based therapies in the United States. 


   THE FIVE LAYERS OF SAFETY
---------------------------------------------------------- Chapter 1:3

The five layers of safety are designed to overlap so that they will
prevent the distribution of contaminated blood and blood products. 
The layers' overlapping safeguards start where the blood is collected
and extend to the manufacturers and distributors of blood products. 


      DONOR SCREENING
-------------------------------------------------------- Chapter 1:3.1

The first layer is designed to prevent the donation of blood by
persons who have known risk factors or other conditions such as low
blood pressure.  High-risk donors, those whose blood may pose a
health hazard, are encouraged to exclude themselves.  Everyone who
seeks to donate blood must answer a series of behavioral and medical
questions.  If the answers indicate high risk, the prospective donor
is deferred.  These requirements are completed before the donor is
allowed to give blood.  If the questions are answered truthfully,
they isolate about 90 percent of all persons whose risk of having HIV
is too recent for their bodies to have produced sufficient antibodies
or antigen to be detected by viral screening tests. 


      DONOR DEFERRAL REGISTRIES
-------------------------------------------------------- Chapter 1:3.2

The safeguard of this layer is the constant updating of lists, known
as "donor deferral registries," of unsuitable donors and the checking
of names of donors with the names in the donor deferral registry to
prevent blood being used from donors previously determined to be
unsuitable.  Individuals who were entered into a deferral registry
are those who were found not to meet donor suitability requirements
during screening or who have had a positive test for any of the
diseases checked at a previous donation.  Services that collect blood
must check the donor deferral registry for each donor, and if they
find a donor listed, they do not distribute that person's blood.  The
deferral registry includes the names of donors who have donated in
the past 8 weeks and are, thus, ineligible to donate until this
8-week period has expired.  The deferral registry may be checked
either before or after blood is donated. 


      TESTING BLOOD
-------------------------------------------------------- Chapter 1:3.3

After a donor's blood has been drawn in a donation, it is tested for
an ABO group and Rh type.  Additionally, viral testing, the third
safety layer, and perhaps the most widely recognized layer, may be
the most critical link in protecting the public from the risk of
receiving contaminated blood transfusions.  Screening tests are
performed for hepatitis B surface antigen (HBsAg), hepatitis B core
(HBc) hepatitis C (HCV), human immunodeficiency virus (antibody for
HIV-1 and HIV-2 and antigen for HIV-1), human T-lymphotropic virus
type I (HTLV-I), and syphilis.\3

Blood facilities also notify the consignee (the facility that
receives the product) if the product is from a donor who may have
been in the "window period" at the time of his or her last
donation--that is, repeat donors who subsequently test positive for
HIV.\4 Even though the previous donations may have met all test
requirements at the time of donation, recipients of blood from such
donors may need to be tested to determine whether a disease has been
transmitted to them.  Additionally, consignees may be notified if
they have received blood from donors who subsequent to their donation
disclose historical information that would have compromised their
eligibility as donors. 

Two tests--one for alanine aminotransferase (ALT) and one for
hepatitis B core (HBc)--were introduced as "markers" for the major
viruses noted above.  That is, donors with elevated ALT counts or
those found to be positive for HBc have, at times, been found
positive for viruses such as HCV and HIV.  These two tests were
introduced when more specific tests for hepatitis C and HIV had not
yet been developed.  A positive result on the syphilis test is
considered by some to be a surrogate marker for high-risk behavior,
since it may be a sign of behavior that increases the risk of
infection from HIV.  However, more specific tests for hepatitis C
have since been developed, and a 1995 National Institutes of Health
(NIH) consensus development conference recommended discontinuing the
use of ALT as a surrogate.\5 AABB also recommended that the ALT test
be dropped for donated blood, and FDA has stated that it will not
object if it is dropped. 

Among the many other infections, viral and nonviral agents that have
garnered public attention because of their prevalence in the U.S. 
blood supply include B-19 parvovirus, Chagas' disease,
cytomegalovirus, and hepatitis D-G.  For various reasons, however,
tests are not routinely conducted for them.  Additionally, different
components of blood do not harbor all these infectious agents, and
much remains to be learned about the location of different viruses in
blood components.\6 Table 1.2 lists the viral and nonviral infectious
agents that we discuss in this report. 



                                Table 1
                
                2: Viral and Nonviral Infectious Agents
                        Discussed in this Report

Agent                               Disease
----------------------------------  ----------------------------------
Parasite: T. cruzi                  Chagas'

Prion, protein (may be a virus)     Creutzfeldt-Jakob

Spirochete: T. pallidum             Syphilis

Virus                               Cytomegalovirus

                                    Hepatitis A-G

                                    Human immunodeficiency

                                    Human T-lymphotropic disease
----------------------------------------------------------------------

--------------------
\3 HIV antibody tests detect antibodies that the human body produces
as an immune response to HIV, whereas HIV antigen tests detect the
actual presence of HIV.  HTLV is a retrovirus that can lead to
neurologic disease or adult T-cell leukemia and lymphoma.  The test
for human lymphotropic virus type II (HTLV-II) uses the HTLV-I test;
although the HTLV-I test is not specific for HTLV-II, it is the
closest test now available for this virus. 

\4 The window period is the time from infectivity to the point at
which currently licensed test kits can ascertain antibodies or
antigens to certain viruses tested for by blood facilities. 

\5 This consensus development conference, "Infectious Disease Testing
for Blood Transfusions," was held on January 9-11, 1995.  The
conference also examined the utility of HBc testing and determined
that this test should still be used to assist in reducing the risk of
HBV and as a surrogate marker for HIV.  It was also recommended that
syphilis testing continue because it may contribute to the prevention
of transfusion-transmitted syphilis. 

\6 For example, HIV-1 appears in plasma and platelets, but it is not
known whether HIV-1 resides in red cells.  Leukocytes do contain HIV
and HTLV-I, but HTLV-I is not found in plasma and red cells, and
whether or not it is located in platelets is not known. 


      BLOOD QUARANTINING
-------------------------------------------------------- Chapter 1:3.4

The fourth safety layer that FDA enforces is the quarantine of all
donated blood until tests and other controls have established its
safety.  This means that blood units cannot be used, except in
emergencies, until all the requirements of the three preceding layers
have been satisfied.  At the fourth layer, blood facilities maintain
separate storage for untested units of blood and for units that are
suitable and units that are unsuitable for use.  "Autologous" units
are also stored separately from "allogeneic" units.  That is,
donations a person makes in order to receive his or her own
blood--autologous units--are stored separately from donations made
allogeneically, by individuals for other people.  Autologous donation
is often made when a person plans for elective surgery. 


      MONITORING AND INVESTIGATING
      PROBLEMS
-------------------------------------------------------- Chapter 1:3.5

Blood facilities are obligated to monitor and investigate errors and
accidents in their procedures, to audit their systems, and to correct
deficiencies.  Licensed blood facilities--those that may engage in
the sale, barter, or exchange of blood products across state
lines--must file "error and accident reports" (EARs) with FDA in
order to notify it of problems.  Unlicensed blood facilities--those
that do not ship blood products across state lines--are not required
to report EARs to FDA but may do so voluntarily.  However, unlicensed
blood facilities must follow the same safety procedures as licensed
facilities. 

All members of the blood industry are also obligated to determine the
causes of errors and accidents and to institute changes to make sure
such problems do not recur.  Finally, this layer includes FDA
inspections of blood facilities to monitor compliance with federal
requirements. 


   FEDERAL OVERSIGHT AND
   RESPONSIBILITY
---------------------------------------------------------- Chapter 1:4

The four federal agencies outlined in table 1.3 have some of the
major oversight authority related to blood safety in the United
States:  FDA, the Centers for Disease Control and Prevention (CDC),
the Health Care Financing Administration (HCFA), and NIH's National
Heart, Lung, and Blood Institute (NHLBI).  Additionally, the table
shows that the Department of Health and Human Services (HHS) has
recently organized a national blood safety committee whose director
and advisory council help ensure that the government's response to
future bloodborne infectious agents is coordinated.\7 Although the
advisory council was announced in October 1995 and formally approved
by HHS in October 1996, HHS has only recently asked for nominations
to the council, and council meetings have yet to take place.\8



                                    Table 1.3
                     
                      Federal Organizations Responsible for
                                U.S. Blood Safety

                           Responsibility
Organization               -----------------------------------------------------
Centers for Disease        Collects data on the incidence of infectious diseases
Control and Prevention     (including those affecting hemophiliacs) and on
                           state-reported clinical AIDS cases

                           Provides guidance and recommendation for preventing
                           disease\a

Food and Drug              Inspects facilities, compiles and summarizes EARs,
Administration             has regulatory authority, promulgates and distributes
                           memoranda and guidelines, and can recommend product
                           recalls

Health Care Financing      Inspects blood facilities that perform viral testing
Administration             procedures and blood transfusion services that are
                           reimbursed through Medicare and Medicaid\b

National Heart, Lung, and  Conducts clinical studies on the effects of blood
Blood Institute            transfusions in patients with cytomegalovirus and HIV

                           Awards research grants for assessing the risks of
                           transfusion-transmitted diseases, developing virus-
                           screening tests, and assessing new infection agents\c

                           Funds genetic testing technologies to close the
                           period between donors' giving blood and the
                           subsequent discovery of their infection

                           Sponsors educational conferences and workshops

U.S. Department of Health
and Human Services

Advisory Council on Blood  Examines broad issues of public health and the social
Safety                     implications of blood safety; serves the Blood Safety
                           Committee\d

Blood Safety Committee     The FDA commissioner and the directors of CDC and NIH
                           report to the Blood Safety Director

Blood Safety Director      Coordinates and oversees Public Health Service blood
                           safety programs
--------------------------------------------------------------------------------
\a As with FDA's guidance documents, these recommendations are not
binding on members of the blood industry. 

\b A memorandum of understanding between FDA and HCFA delineates that
FDA will inspect manufacturers of blood products, but FDA can also
inspect transfusion services that are HCFA's responsibility if there
are indications of noncompliance with good manufacturing practices. 

\c Includes the Transfusion Safety Study that tracks the natural
history of transfusions associated with HIV and the Retrovirus
Epidemiology in Donors Study that has, among other topics,
investigated the clinical course of blood donors infected with HTLV-I
and HTLV-II. 

\d Issues include social choice, informed consent, the allocation of
research resources, the availability of blood, and the effect of
economic factors on its availability. 

The regulations governing oversight of most aspects of blood banking
are found in the Code of Federal Regulations (CFR).\9 FDA also issues
memoranda and guidelines as guidance on specific topics to blood
facilities.  These guidance documents are not binding on the blood
facility and, thus, blood facilities may follow the guidance or
choose to use appropriate alternative procedures not provided in the
guidance.\10

The memoranda topics range widely.  Fifty-two that still represent
current guidance were issued between August 1982 and August 1994; an
additional 22 issued during this period are no longer current. 
Topics include recommendations for the management of donors who are
found to be positive for hepatitis, equivalent methods for
compatibility-testing, deferral of blood donors who have received the
drug Accutane, and revised recommendations for preventing the
transmission of HIV through blood and blood products. 

In regard to FDA's responsibility for inspecting blood facilities, a
detailed checklist for inspectors was recently abandoned for a more
systems-oriented approach in conducting its inspections.  Its new
"Guide to Inspection of Blood Banks" outlines major areas that an
inspection should examine:  (1) errors, accidents, and fatalities;
(2) facilities, equipment, and personnel; (3) quality assurance; (4)
the disposal of infectious waste; (5) whole blood and donor
suitability; (6) laboratory operations; (7) uniform blood labeling;
(8) compatibility-testing and transfusion reactions; (9) storage and
distribution; (10) platelets and pheresis; (11) computerization; (12)
red blood cells, plasma, platelets, and cryoprecipitate; (13)
records; and (14) operations. 


--------------------
\7 This entity was organized as a result of recommendations in an
Institute of Medicine report, "HIV and the Blood Supply," Washington,
D.C., July 1995, that examined the federal government's response to
the discovery of HIV and the protection of the blood supply in the
early 1980s. 

\8 The formation of a blood safety director, blood safety committee,
and advisory council on blood safety and availability was announced
by the HHS Secretary in testimony before the House Committee on
Government Reform and Oversight, Subcommittee on Human Resources and
Intergovernmental Relations, on October 12, 1995. 

\9 21 C.F.R.  parts 210, 211, 606, 607, 610, and 640. 

\10 FDA's recent "Guideline for Quality Assurance in Blood
Establishments" is one example.  It is intended to assist blood
facilities in developing quality-assurance programs that "are
consistent with recognized principles of QA [quality assurance] and
current good manufacturing practices .  .  .  ."


   SCOPE AND METHODOLOGY
---------------------------------------------------------- Chapter 1:5

We limited the scope of this report to policies and procedures that
became current in 1994.  We did not examine problems of the
mid-1980s, when HIV was first recognized as a bloodborne disease, or
the sequence of changes intended to address HIV.  We examined FDA's
oversight of licensed and unlicensed blood facilities in the United
States, including plasma centers. 

The focus of the work is the general policies and procedures in place
to help ensure the safety of the blood supply.  We did not examine
patterns of violations of these policies and procedures by individual
blood facilities.  While many of the recurrent problems in the
industry relate to failures to comply with safety requirements, our
review considers whether there are proper safeguards in place to
identify such occurrences, not which specific blood facilities may
have problems in this regard. 

We reviewed pertinent documents, interviewed relevant officials, and
surveyed and visited blood facilities.  The documents we reviewed
included FDA statutes, regulations, compliance manuals and compliance
program, and memoranda.  We supplemented our interviews of various
government officials by interviewing other officials of the blood
industry as well as interest groups such as AABB, ABC, ARC, and
IPPIA.  We accompanied FDA officials during an inspection and visited
various types of blood facilities.  Among the FDA data sources that
we analyzed were error and accident reports (EARs) and establishment
inspection reports (EIRs), including Form 483 reports of inspection
observations.  We conducted our review from October 1994 to May 1996
in accordance with generally accepted government auditing standards. 


      FDA STATUTES, REGULATIONS,
      AND MEMORANDA
-------------------------------------------------------- Chapter 1:5.1

We examined FDA's statutes, regulations, and more than 70 memoranda
to determine what is required of and recommended to blood facilities
to help ensure a safe blood supply.  When we reviewed the memoranda,
we categorized them by topic, which ranged in scope and specificity
from a guideline for deferring donors who have received Accutane to a
guideline for the validation of computer systems.  We also used these
documents to ascertain potential vulnerabilities in the layers of
safety. 


      INTERVIEWS
-------------------------------------------------------- Chapter 1:5.2

When we interviewed FDA personnel, we asked them about their
operations, inspection procedures, and databases.  The personnel in
the blood facilities additionally gave us important details about
FDA's oversight and interactions.  The information we gathered from
AABB, ABC, ARC, and IPPIA told us about overall blood industry
practices and potential safety issues. 


      SITE VISITS
-------------------------------------------------------- Chapter 1:5.3

We visited seven sites to cover the range of facilities:  licensed
and unlicensed, ARC and non-ARC, source plasma centers and
fractionation companies.  At each site, we examined the physical
operations of the blood facility and interviewed the staff who were
responsible for its daily operations:  directors of compliance and
quality assurance, medical directors, vice presidents of research and
scientific services, directors of component production and of
operations, and executive officers. 


      ERROR AND ACCIDENT REPORTS
-------------------------------------------------------- Chapter 1:5.4

FDA requires licensed blood facilities to report errors and accidents
that resulted in an unsuitable unit of blood being made available for
distribution.  In March 1991, FDA asked unlicensed blood facilities
to submit EARs voluntarily.  We obtained FDA's annual summary reports
of the EARs submitted by licensed and unlicensed facilities for 1990
through 1994, which constitutes data on the universe of EARs in that
period.\11

FDA's summary EAR data are reported by facility type (licensed,
unlicensed, ARC, non-ARC, plasma center, transfusion service) and
include the total number of reports received, the type of error or
accident (whether in viral testing, labeling, quarantining, or other
procedures), the number of events attributable to computer or data
entry errors in 1994, and the number of EARs resulting in potential
recall of a blood unit.  In addition to analyzing these data, we
identified changes in rules and regulations that might have affected
reporting criteria, analyzed the differences between types of blood
facilities, and highlighted the EAR information that shed light on
specific blood-banking processes. 

In appendix II, we outline these data as FDA compiled them for fiscal
year 1994 (in appendix I, we discuss issues relating to viral and
nonviral agents).  However, we based our report's analysis on the
reporting rate per type of blood facility and on the rate of
reporting per 100,000 units each type of blood facility collected. 
We did this because FDA's analysis does not take into account the
interdependence of reporting for the different processes by the
different facilities used. 


--------------------
\11 In fiscal year 1991, FDA received 3,836 EARs; in 1992, the number
was 10,456; the numbers for fiscal years 1993 and 1994 were 8,991 and
11,298. 


      ESTABLISHMENT INSPECTION
      REPORTS AND FORM 483
-------------------------------------------------------- Chapter 1:5.5

FDA's annual inspections of blood facilities result in establishment
inspection reports that descriptively narrate the activities covered
in the inspection and any problems found during the inspection.\12 An
inspector who identifies significant infractions that could affect
blood safety files a Form 483.  We analyzed the most recent EIRs and
Form 483s from a nationally representative sample of licensed and
unlicensed blood facilities, including plasma centers.  We randomly
sampled eight FDA inspection districts and, from these districts, a
total of 373 EIRs (representing reports from the total of 2,980 U.S. 
blood facilities).\13

For the 373 blood facilities in our study, we were able to analyze
information on 325:  48 licensed centers, 114 unlicensed centers, 91
transfusion services, and 72 plasma centers.\14 The remaining 48
blood facilities either were plasma brokers, viral testing or reagent
manufacturers, testing laboratories, or depot sites or had been
inspected for specific purposes that were not part of the annual
inspection process. 

We analyzed the EIRs in a manner similar to FDA's analysis of EARs. 
That is, we applied FDA's coding scheme of blood-banking processes to
our analysis.\15 By using the same coding scheme, we were able to
outline information on EARs and EIRs that highlighted potential
safety concerns for specific blood-banking processes. 


--------------------
\12 Beginning in 1995, blood facilities that have complied with FDA
requirements for 2 years become eligible for biennial rather than
annual inspections.  FDA inspectors need to list the activities they
observe only if it is a limited inspection.  In all other cases,
inspectors need only list the compliance program under which the
inspection is taking place. 

\13 The districts were Boston, Chicago, Cincinnati, Dallas, Los
Angeles, New Orleans, Philadelphia, and Seattle. 

\14 Licensed facilities may engage in the sale, barter, or exchange
of blood products across state lines.  They often collect autologous
and allogeneic blood.  Unlicensed facilities do not ship blood
products across state lines but can collect both types of blood. 
Transfusion services routinely collect only autologous blood.  Plasma
centers collect source plasma for processing into plasma-based
therapies.  All of these types of facilities should be registered
with FDA. 

\15 In our analysis of EIRs, we used the same categories of blood-
banking processes that are defined in FDA's EARs:  (1) donor
screening, (2) donor deferral, (3) collection and processing, (4)
routine testing, (5) viral testing, (6) post-donation information,
(7) product quarantine, (8) labeling, and (9) storage and
distribution.  FDA used a tenth category, "miscellaneous," that
captured errors and accidents related to transfusion-transmitted
viruses, recipient reactions, lookback, and emergency release of
products.  We incorporated these issues into the 9 other categories
by their specific topic.  We added an eleventh category for our
analysis of EIRs, which we called "machines," in order to identify
problems related to computer hardware and software issues and quality
control of machines (recordkeeping) used in blood-banking.  We have
not outlined these issues in our report because they were often
related to specific topics that we subsumed under FDA's 9 categories
noted above. 


      SURVEY OF BLOOD CENTERS
-------------------------------------------------------- Chapter 1:5.6

We surveyed all the full-service blood facilities in our sample of
inspection reports.\16 This survey gave us additional information on
most of the processes we studied in our analysis of EARs and EIRs. 
One hundred percent of the 45 blood facilities we surveyed responded
to our questionnaire.\17 Appendix III contains the questionnaire used
in our survey. 


--------------------
\16 By "full-service facility," we mean one that carries out the full
range of activities covered by the five layers of safety:  collecting
(screening and deferral), testing, processing (quarantine and
control), and distributing blood products.  Therefore, we excluded,
for example, donor-collection centers that send their blood elsewhere
for testing. 

\17 Our original sample contained 47 full-service blood facilities,
but 2 had closed before we began our survey. 


   THE STRENGTHS AND LIMITATIONS
   OF OUR STUDY
---------------------------------------------------------- Chapter 1:6

By examining EIR and Form 483 information with FDA's EAR coding
scheme, we were able to present analyses from both data sources for
individual blood-banking processes.  Furthermore, our sample of blood
facilities represents blood facilities in the United States, and our
findings can therefore be generalized to the blood-banking industry
at large. 

However, our analysis of EIRs was predicated on the accuracy of the
information contained in them.\18 We did not collect primary data
from the blood facilities.  Furthermore, our information on EARs was
based on FDA's annual summaries and did not involve original data
analysis. 

The organization of this report reflects the five layers of safety. 
In chapter 2, we cover issues related to the first two layers, donor
screening and deferral, as well as collection processes.  In chapter
3, we focus on the third layer, testing; in chapter 4, on the fourth
layer, the quarantine of blood and other processes.  We discuss the
fifth layer, monitoring and investigations, in chapter 5.  Finally,
in chapter 6, we present a summary of our findings, our conclusions,
and our recommendations. 


--------------------
\18 Thus, much of our analysis is directed at Form 483 observations
because information contained in the EIRS was not a reliable
indicator of activities observed by FDA inspectors.  See chapter 5
for a discussion on the content of EIRS and the ramifications for our
analyses provided in that chapter. 


SCREENING, DEFERRAL, AND
COLLECTION
============================================================ Chapter 2

Donor screening and deferral are the first two layers of safety. 
Screening prospective donors by asking them about high-risk behavior
and their medical history enables the blood-banking community to
exclude unsafe blood.  Donor deferral registries, if checked before
donation, can help ensure that those who have been deferred do not
donate.  Collection and processing of blood is another area of blood
banking that takes place prior to the testing of blood.  Only
screening and deferral eliminate blood hazards such as malarial and
Chagas' infection, but the redundancy of the three remaining safety
layers--testing, quarantining, and monitoring--mitigates many other
consequences that would follow without these layers of safety. 

We found, however, that (1) questionnaires for screening out
high-risk donors are not uniform throughout the blood industry, and
accurate responses may be difficult to obtain where respondents are
not assured of privacy.  Moreover, (2) donating blood before the
donor deferral registry (DDR) is checked can cause problems, DDRs can
yield false checks where they have not been computerized, and lack of
donor deferral notifications may lead to unsuitable donors'
continuing to donate blood.  Finally, (3) the blood industry's
collection processes appear to cause few safety problems but
bacterial contamination is a leading cause of blood-transfusion
fatalities. 


   DONOR SCREENING
---------------------------------------------------------- Chapter 2:1

The blood industry practices several methods for selecting donors of
safe blood.  One is to exclude particular donor groups; for example,
blood is not collected at prisons or mental hospitals where the risk
of hepatitis and other diseases is high.\1 Another is to eliminate
cash incentives for making whole-blood donations:  data show that
paid donors have a higher likelihood of being infected with HIV and
other diseases than volunteer donors.\2 Plasma centers still pay
donors because a cash incentive is deemed necessary if they are to
sit through the 2-hour procedure (whole-blood donations often take
less than 1 hour). 

Another way of ensuring safe blood donations is to conduct health
history interviews designed to defer donors who might transmit
infectious disease.  Table 2.1 shows the focus of some of the
questions blood facilities ask prospective blood donors in order to
ascertain risk. 



                               Table 2.1
                
                 Donor Screening Questions and Targeted
                                Diseases

Question focus                      Targeted disease
----------------------------------  ----------------------------------
Country of birth                    AIDS (HIV-2), malaria, Chagas'

Travel history                      Malaria

Medical history of a specific       AIDS, babesiosis\, Chagas',
disease                             hepatitis, malaria\a

Medical symptoms compatible with a  AIDS, bacteremia, viremia
specific disease

Exposure through transfusion or     AIDS, hepatitis
occupation

Medical treatment                   Creutzfeldt-Jakob\b

Sexual contact or drug use of       AIDS, HTLV-I and HTLV-II,
donor or donor's partner            hepatitis
----------------------------------------------------------------------
\a Babesiosis, like Chagas' disease, is caused by a parasite. 

\b Some researchers believe that Creutzfeldt-Jakob disease is caused
by a prion, a small protein particle.  Others suggest it may be
caused by a virus.  Persons who have been infected can remain
asymptomatic for decades but then progress rapidly to dementia and
death.  Although no scientific evidence supports the notion that it
is transmitted through blood products, it has been transmitted
through cornea transplants and brain tissue transplants as well as
through the administration of the human pituitary-derived growth
hormone. 

A brief medical examination of all donors is performed, records are
maintained, and the donors sign an informed-consent form that
outlines the possible consequences of donation deferral.\3 The donors
medical record and history is intended to determine the time of the
last donation; the physical examination is intended to help ensure
that the donor is in good health by assessing the temperature, blood
pressure, and hemoglobin levels.  Donors are also checked to see if
there is evidence of respiratory disease or diseases transmissible by
blood transfusion and have neither infectious diseases at the site
where blood is drawn nor scars that indicate abusive self-injection
of drugs. 

Blood facilities impose additional requirements on persons who donate
source plasma:  acceptable levels of total protein,
syphilis-screening every 4 months, and a more detailed annual
physical examination that includes urinalysis and may include
toxicology screening.  This physical examination also includes
observations of heart and lung sounds; lymph nodes, mouth, and skin;
and abdominal and neurological conditions. 

Another screening method is to give prospective donors a chance to
exclude themselves.  This method may include confidential unit
exclusion (CUE) and telephone callback.  CUEs require donors to place
one of two bar code stickers ("transfuse" or "do not transfuse") on
their donation record before they donate.  The CUE is intended to
help donors who may feel pressured to donate by peers, for example. 
(A survey published in 1989 found that almost a third of the 304
seropositive donors responded that their colleagues had pressed them
to donate.\4 ) In a telephone callback, persons who have donated
blood call the blood center to report additional information
pertinent to their medical history.  Often this pertains to
post-donation headaches and acute illness, but it may also relate to
risky behavior prior to the donation that would have precluded the
donation had it been known at the time. 

Some fractionation companies have also instituted programs to
increase the safety of the blood supply by instituting stringent
screening processes for their donors.  For example, one plasma
company has developed an inventory-hold program in which the company
collects all units of plasma that have been screened as safe and
usable for production and holds them for 3 months.  If during this
time one of the company's donors is found to be reactive to viral
screening or surrogate tests, the company has the ability to identify
and destroy all plasma units previously obtained from that donor
during this 3-month hold period. 

This process is used because the company's data have shown that
approximately 96 percent of its plasma collections are followed by at
least one additional donation by the same donor.  The inventory-hold
program thus attempts to identify unsuitable blood during the window
period.  The company also destroys all plasma from first-time donors
who do not return to make a second donation within 3 months. 
Ninety-five percent of the blood units that test positive for
hepatitis B virus (HBV), HCV, or HIV at this company's facilities are
from first-time donors. 


--------------------
\1 Patients from mental hospitals can donate at a blood facility, and
FDA has recently promulgated guidance on deferring inmates of
correctional institutions.  New prisoners and those who have been
incarcerated for more than 72 consecutive hours during the previous
12 months are deferred for 12 months. 

\2 For example, the California Department of Health Services found
that plasma centers, where donors were paid, had a confirmed HIV rate
of 0.016 percent (16 per 100,000 units tested) while the rate at
blood facilities, where donors were not paid, was 0.002 percent. 
These were second-quarter 1994 data from 98 percent of all California
facilities required to report HIV test results. 

\3 See 21 C.F.R.  606.160(b)1.  Blood facilities must keep donor
records that contain the medical interview and examination record and
the informed-consent form.  A donor consent form describes to each
donor that his or her acceptability will be determined by a medical
interview, examination, and laboratory testing.  Donors should be
informed of all the laboratory tests that are performed on samples of
their blood and of the consequences of an unacceptable, or positive,
test.  These include the possible detection of infectious agents,
temporary or permanent deferral, the listing of their names in
deferral registries, reporting to the public health agencies, and
governmental inspection of the registries and the donors' test
records. 

\4 Susan Leitman et al., "Clinical Implications of Positive Tests for
Antibodies to Human Immunodeficiency Virus Type-I in Asymptomatic
Blood Donors," New England Journal of Medicine, 321 (1989), 917-24. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 2:1.1

Thirteen percent of all error and accident reports submitted to FDA
in fiscal year 1994 were for screening errors (see appendix II). 
These included the facilities' not performing donor deferral
screening, their use of incorrect names during a deferral search, and
donors' giving a medical history that warranted but did not result in
a deferral.\5 Tables 2.2 and 2.3 provide data from EARs and our
analysis of EIRs that highlight the need for continued vigilance in
the area of donor screening. 



                               Table 2.2
                
                 Screening EAR Rates by Facility Type,
                                 1994\a

                                           Unlicensed\
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c\              3.8         0.01    0.53    0.48
EAR rate per 100,000 units            9.3          2.1     2.0     5.6
 collected\\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 2.3
                     
                         Screening Problems and Form 483
                         Observations by Facility Type\a

                        Unlicensed\\  Transfusion      Plasma
             Licensed        b          service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities  14 of  37%    12 of  15%  9 of 36  25%    22 of  42%    57 of    27%
 with          38            83                          52           209
 problems\
 c
Facilities  11 of   29    10 of   12  7 of 36   19    15 of   29    43 of     21
 receiving     38            83                          52           209
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined donor screening during its inspection.  Problems
were those that were characterized by the inspector on the inspection
report whereas Form 483 observations were problems deemed serious
enough to be noted on a Form 483. 

Licensed facilities reported EARs for screening at a rate more than
380 times that of unlicensed facilities and 7 times that of plasma
centers.  Per 100,000 units collected, the rates of EARs for
screening at licensed facilities were 4 and 5 times higher than
unlicensed facilities and plasma centers, respectively.  However,
reporting problems we discuss in chapter 5 make it impossible to draw
any conclusions about these rates--that is, neither FDA nor we can
say whether the differences stem from licensed facilities' having
more errors and accidents in donor screening or from licensed
facilities' reporting their errors and accidents more readily than
unlicensed facilities and plasma centers.\6

Interestingly, at plasma centers, 15 percent of all EARs were related
to donor screening in that screening was not performed but donors
were later deferred because of HBsAg or HIV reactivity or a history
of hepatitis.  Seventy-five percent of screening errors at plasma
centers were related to computer malfunctions, suggesting a possible
technological reason for these problems. 

In our analysis of EIRs, we found that FDA inspectors found many
facilities with problems relating to donor screening.  In fact, about
40 percent of licensed facilities and plasma centers for which we
could determine that donor screening was observed by the FDA
inspector had problems in this area.  Similarly, among facilities for
which an EIR indicated an FDA review of this process, 29 percent (11
of 38 licensed facilities; 14 of 52 plasma centers) received Form 483
observations in donor-screening processes.  We were unable to draw
any firm conclusions or comparisons from these data.  Differences in
the likelihood of receiving an inspection observation may reflect
compliance problems in different facility types or inconsistencies in
FDA's inspection criteria for establishing noncompliance among
different facility types.\7 (We discuss this problem further in
chapter 5 in relation to FDA's monitoring activities.)


--------------------
\5 Appendix II shows FDA's summary report of the actual number of
screening EARS.  It also gives the percentage of EARs different types
of blood facilities submitted for each blood-banking process we
report in chapters 2-4 and the percentage of submissions as they
relate to the total number of EARs. 

\6 In fiscal year 1994, most of the reports from plasma centers were
submitted by one facility (723/856 = 84 percent).  The majority of
their reports were related to donor screening (206/723 = 28 percent)
and donor deferral (514/723 = 71 percent).  However, EARs submitted
by plasma facilities in fiscal year 1993 resulted in 48 percent of
EARs in the areas of donor screening and deferral.  Licensed
facilities reported EARs in these two areas at a much higher rate
than plasma centers. 

\7 The same interpretive difficulty holds for all the EIR data we
present in chapters 2-4. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 2:1.2

Two areas of safety that are of concern regarding screening are the
lack of a uniform questionnaire and the lack of privacy for donors. 


         QUESTIONNAIRE
------------------------------------------------------ Chapter 2:1.2.1

The types of medical history questions asked and the manner in which
they are asked differ from facility to facility and can affect
donors' responses and thus, the potential that blood could be drawn
from a donor who should have been deferred.  Research indicates that
asking donors blunt and direct questions about drug abuse and sexual
behavior screens out significantly more high-risk donors than
less-direct questions; moreover, donors are not offended by explicit
questioning.\8 However, questions must be sensitive to different
terminology and the perspectives that respondents may have about
high-risk behavior. 

For example, the AABB questionnaire asks men about their past sexual
activity with other men without asking specific questions about
homosexuality.  Research has shown that such questioning elicits more
accurate responses, since some men might not consider themselves
homosexuals although they may have had sex with men. 

Other research has found that asking direct oral questions about
sexual behavior is associated with a significant increase in HIV
deferrals, but the study did not find any evidence of an increase in
blood safety as measured by HIV seroprevalence.  That is, direct
questioning probably resulted in the deferral of at-risk but
predominately nonpositive HIV donors.\9

California has recently instituted a uniform donor history
questionnaire.  FDA and AABB have also recommended general guidelines
on questions to be asked.  However, FDA does not require that a
uniform donor questionnaire be followed although ARC uses a uniform
questionnaire.  It is not known how many blood facilities follow the
AABB questionnaire. 


--------------------
\8 Donna J.  Mayo, "Screening Potential Blood Donors at Risk for
HIV," Transfusion, 31 (1991), 466-74. 

\9 E.  Johnson et al., "The Impact of Direct Oral Questions on Blood
Donor Screening for Human Immunodeficiency Virus," Transfusion, 34
(1994), 769-74. 


         PRIVACY
------------------------------------------------------ Chapter 2:1.2.2

According to AABB's 1994 accreditation requirements, verbal privacy
is mandatory during medical history questioning in order to elicit
honest answers.  However, when we visited blood facilities, we found
that some have not met this requirement.  Studies have indicated that
from 14 percent to 30 percent of donors feel that screening areas
provide inadequate privacy and that 20 percent of donors would have
given different answers had they been in a more private situation.\10

Although FDA regulations do not specifically require private
interviews, FDA guidance to inspectors states that "interview areas
have to offer the donor a degree of privacy so that the donor will be
comfortable answering the questions without fear of being
overheard."\11


--------------------
\10 L.  S.  Doll et al., "Human Immunodeficiency Virus Type
1-infected Blood Donors:  Behavioral Characteristics and Reasons for
Donation," Transfusion, 31 (1991), 704-9, and M.  A.  Popovsky et
al., "Privacy of Donor Screening:  Perception vs.  Reality,"
Transfusion, 31 supp.  (1991), 67S. 

\11 See Food and Drug Administration, Guide to Inspections of Blood
Banks (Washington, D.C.:  September 1994), p.  3.  FDA regulations do
require that a facility provide space for a private and accurate
examination of individuals to determine their suitability as blood
donors. 


   DONOR DEFERRAL
---------------------------------------------------------- Chapter 2:2

Blood facilities have several guidelines for deferring donors.  Each
facility must have a DDR to identify prospective donors who have
previously been deferred.  Facilities screen prospective donors
through physical examinations and medical history questioning, and
blood facilities are required to have records available from which
unsuitable donors may be identified.  FDA prescribes several periods
of deferral, defined by the perceived risk of a particular donor's
donating unsafe blood.  (See table 2.4.)



                               Table 2.4
                
                  Four FDA-Recommended or FDA-Required
                 Deferral Periods and Some Reasons for
                                  Them

Deferral
period      Reason
----------  ----------------------------------------------------------
8 weeks     Having made a prior donation of whole blood

1 month     Taking Accutane and Proscar\a

12 months   Traveling in areas where malaria is endemic\b

            Coming into close contact with a person who has viral
            hepatitis

            Paying for sex with drugs or money

            Having sex with
            --anyone who has AIDS or has had a positive test for HIV
            --anyone who has ever taken illegal drugs by injection
            --anyone who has taken clotting-factor concentrates for a
            bleeding disorder
            --a man who has had sex with another man even once since
            1977





            Having received blood or blood products

            Having been tattooed or having had body parts pierced with
            nonsterile techniques

            Receiving a positive test for syphilis or treatment for
            syphilis or gonorrhea

            Coming into contact with blood or body fluids from
            inoculations through the skin, an open wound, nonintact
            skin, or mucous membranes

            Being a victim of rape

Permanent   Using Tegison\c

            Having had viral hepatitis after age 11

            Receiving clotting-factor concentrate for a bleeding
            disorder or human pituitary growth hormone\d

            Having clinical or laboratory evidence of AIDS or HIV

            Being a man who has had sex with another man even once
            since 1977

            Being an intravenous drug user

            Testing positive for hepatitis B or C, HIV, or HTLV\e

            Selling sex for money or drugs since 1977
----------------------------------------------------------------------
\a Accutane, a drug prescribed for the treatment of acne, has been
shown to cause developmental malformations in children.  When
transfused through blood to a pregnant woman, it may increase risks
to the developing fetus.  Proscar, prescribed for the treatment of
enlarged prostate glands, has been shown to cause developmental
malformations in male offspring. 

\b Deferral is for 3 years if the donor has had malaria and has since
been asymptomatic or was an immigrant, refugee, or citizen of an area
where malaria is endemic.  Donations to be used for preparing plasma,
plasma components, or derivatives devoid of intact red blood cells
are not recommended for deferral because the malarial parasite is
found only in cellular components. 

\c Tegison is used to treat severe psoriasis but is not to be used
during pregnancy because major fetal abnormalities have been
reported.  Because of this and the possibility that Tegison may
remain in the blood for long periods, FDA has recommended permanent
deferral of donors who take this drug. 

\d Pituitary-derived human growth hormone is used in the long-term
treatment of children who fail to grow because they secrete normal
growth hormones inadequately.  Some of its recipients, however, have
been reported to have Creutzfeldt-Jakob disease, and animal studies
suggest that this disease may be transmitted through blood.  FDA has
recommended permanent deferral of persons who have received
injections of pituitary-derived human growth hormone, although
deferral is not necessary for those who have received recombinant
human growth hormone, because this product is manufactured with DNA
technology. 

\e Blood facilities must test prospective donors for hepatitis B
(both surface antigen and core), hepatitis C, HIV, and HTLV.  Source
plasma centers must test for hepatitis B (surface antigen), HCV, and
HIV but not hepatitis B (core) or HTLV.  FDA has outlined procedures
(specific "confirmatory" tests) through which a donor's deferral for
hepatitis B and C and HIV (but not HTLV) can be lifted (known as
re-entry algorithms).  Blood facilities may use these procedures when
they can determine that the original positive test results were
"false positives," meaning that the donor actually did not have viral
infections. 

The FDA Guide to Inspections of Blood Banks notes that "records must
be maintained to prevent the distribution of subsequent units of
blood drawn from unsuitable donors."\12

Federal regulations also require blood facilities to maintain records
of permanent and temporary deferrals and the reasons for them. 
Source plasma centers must also establish a system to identify donor
participation in other plasmapheresis programs in the surrounding
area, in order to ensure that individual plasma collections do not
exceed recommended volumes. 

Some blood facilities, such as ARC, combine their local registries
into wider ones.\13 Data from 1993 show that ARC's DDR comprised some
300,000 entries.  If all ARC DDRs were collated into one file,
national and local, its registry would contain approximately 1.6
million entries.  Adding non-ARC facilities to this list would raise
this number to approximately 3 million entries, representing about 1
percent of the U.S.  population.\14

These numbers are one reason why some have suggested that a national
DDR would be cumbersome to develop, validate, and maintain. 


--------------------
\12 Food and Drug Administration, Guide to Inspections of Blood
Banks, p.  2. 

\13 ARC collects approximately 45 percent of all blood collected in
the United States.  California has a statewide DDR.  United Blood
Services' (UBS) facilities, which annually collect some 700,000 units
of blood, or about 6 percent of the national total, have their own
registry that serves communities in 19 states.  Source plasma centers
have a national DDR that is checked for first-time but not repeat
donors. 

\14 William Sherwood, "Donor Deferral Registries," in Morris
Blajchman (ed.), Transfusion Medicine Reviews, 7:2 (April 1993),
121-28. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 2:2.1

Errors and accidents related to such issues as donors being
incorrectly identified, deleted, or missing from deferral lists
accounted for 8 percent of all EARs in fiscal year 1994 (see appendix
II).  Tables 2.5 and 2.6 outline EARs reported by different types of
blood facilities and data from our analysis of EIRs. 



                               Table 2.5
                
                  Deferral EAR Rates by Facility Type,
                                 1994\a

                                           Unlicensed\
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c              1.26        0.001     1.1    0.30
EAR rate per 100,000 units            3.1          0.2     4.3     3.5
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 2.6
                     
                          Deferral Problems and Form 483
                      Observations by Facility Type, 1994\a

                        Unlicensed\\  Transfusion      Plasma
             Licensed        b          service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities  15 of  37%  8 of 49  16%  0 of 27   0%    23 of  47%    46 of    28%
 with          41                                        49           166
 problems\
 c
Facilities  10 of   24  6 of 49   12  0 of 27    0    20 of   41    36 of     22
 receiving     41                                        49           166
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined donor deferral during its inspection.  Problems
were those that were characterized by the inspector on the inspection
report whereas Form 483 observations were those problems deemed
serious enough to be denoted on a Form 483. 

Licensed facilities reported deferral EARs at a rate that was about
equal to that of plasma centers but more than 1,000 times that of
unlicensed facilities.  Their rates per 100,000 units collected were
about equal but 15 and 20 times higher, respectively, than the rate
for unlicensed facilities.  Interestingly, 21 percent of all EARs
reported by plasma centers related to missing or incorrectly
identified donors on the deferral list who were later deferred
because of HBsAg or HIV reactivity or a history of hepatitis. 
Combined with the screening data, 36 percent of plasma center EARs
were associated with inadequate screening or deferral of donors who
were later deferred for HBsAg or HIV reactivity.\15

Our analyses of screening and deferral EIRs and Form 483s proved
similar in that the facilities most likely to have had problems found
during an FDA inspection and to have received Form 483 observations
were licensed facilities and plasma centers.  Furthermore, our
analysis of EIRs mirrors FDA's information on EAR submissions in that
plasma centers seem especially vulnerable to problems in this area. 


--------------------
\15 In fiscal year 1994, most of the reports from plasma centers were
submitted by one facility (723/856 = 84 percent).  The majority of
their reports were related to donor screening (206/723 = 28 percent)
and donor deferral (514/723 = 71 percent).  However, EARs submitted
by plasma facilities in fiscal year 1993 resulted in 48 percent of
EARs in the areas of donor screening and deferral.  Licensed
facilities reported EARs in these two areas at a much higher rate
than plasma centers. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 2:2.2

Three areas of safety that are of concern regarding donor deferral
are the timing of donor deferral registry checks, lack of
computerization for these registries, and varied practices for donor
deferral notification. 


         DDR CHECKS
------------------------------------------------------ Chapter 2:2.2.1

Blood facilities are not required to query their donor deferral
registries before accepting blood from a donor.  This is a special
problem at mobile sites, from which blood is typically shipped to the
main facility where DDR checking occurs after it has been collected. 
The representatives of blood facilities whom we interviewed cited two
reasons for this practice:  (1) mobile sites customarily have no
computer hookup to the central registry and (2) many computerized
registries do not allow blood from a donor who is in the deferral
system to be shipped to hospitals, giving the collection facilities
confidence that unsuitable blood will not leave the central blood
facility. 

Such confidence may be misplaced, however, if donors are not
"flagged" correctly and unsafe blood passes undetected from the blood
facility.  Indeed, some blood facilities use portable computers so
that their mobile sites can access a main, computerized DDR registry
before blood is collected.  However, some facilities do not have
computerized DDRs or cannot afford the portable systems. 
Nevertheless, such practices may needlessly subject deferred donors
to a blood collection procedure and incur needless costs to the blood
facility if viral testing is performed on such units. 


         MANUAL DDRS
------------------------------------------------------ Chapter 2:2.2.2

Regarding the lack of computerization, we found that DDRs are
sometimes compilations of alphabetized index cards similar to those
of a traditional library card catalog.  The potential for error is
enhanced in this type of system.  In fact, during one of our visits,
a blood facility representative found it very difficult to locate a
known donor deferral card because the cards had been used but not
placed back in alphabetical order.  Such problems open up the
possibility that a deferred donor's blood would be collected. 


         DONOR DEFERRAL
         NOTIFICATION
------------------------------------------------------ Chapter 2:2.2.3

When donors have been notified that they have been deferred, they are
usually told the reasons for the deferral and whether a confirmatory
test based on positive viral marker results was performed.  However,
the information that blood facilities offer differs from one facility
to another.  Moreover, FDA has recommendations in its memoranda only
on notifying donors who test positive for HIV.  FDA memoranda on
hepatitis B and C do not include language recommending such
notification.  While many facilities notify deferred donors for
ethical and public health reasons, some do not.  Those that do not
raise the risk that donors of unsuitable blood will unknowingly
continue to donate blood or transmit a disease within the community. 


   COLLECTION AND PROCESSING
---------------------------------------------------------- Chapter 2:3

The normal unit of blood that is drawn is 415 to 495 milliliters in
volume (about 1 pint).  Units containing a lower volume of red blood
cells can be transfused if they are properly prepared with
anticoagulant, but other blood components cannot be made from them. 
Federal regulations require blood facilities to collect this blood in
sterile containers and to include it in laboratory testing. 
Additionally, they are required to prepare a donor's skin where the
blood is to be drawn in a way that maximally ensures the container's
sterility, and they must identify each unit of blood by its donor. 

Every unit of blood and plasma is also to be refrigerated unless the
product is to be used as a source of platelets.  For source plasma,
regulations require that the plasma is to be removed and the cells
returned to the donor by sterile and aseptic means. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 2:3.1

The EARs suggest that reported errors and accidents in collection and
processing are rare.  This would include such issues as bacterial
contamination, blood being drawn into outdated bags, and incorrect
preparation of components.  For fiscal year 1994, blood collection
and processing accounted for only 3 percent (362 of 11,292) of EARs
submitted by licensed and unlicensed blood facilities, transfusion
services, and plasma centers.  (See appendix II.) Tables 2.7 and 2.8
outline EARs reported by different types of blood facilities and data
from our analysis of EIRs. 



                               Table 2.7
                
                 Collection and Processing EAR Rates by
                         Facility Type, 1994\a

                                           Unlicensed\
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c\              1.1        0.004    0.02    0.12
EAR rate per 100,000 units            2.7         0.71    0.07    1.39
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 2.8
                     
                      Collection and Processing Problems and
                     Form 483 Observations by Facility Type,
                                      1994\a

                        Unlicensed\\  Transfusion      Plasma
             Licensed        b          service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities  13 of  34%    11 of  12%    16 of  36%    18 of  35%    58 of    25%
 with          38            95            45            51           229
 problems\
 c
Facilities  12 of   32  9 of 95   10    11 of   24    12 of   24    44 of     19
 receiving     38                          45            51           229
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined collection and processing during its inspection. 
Problems were those that were characterized by the inspector on the
inspection report whereas Form 483 observations were problems deemed
serious enough to be noted on a Form 483. 

As with screening and deferral data, our analysis of EAR data found
that licensed facilities reported collection and processing EARs at
much higher rates than unlicensed facilities.  Also, even though
collection and processing EARs made up only a small percentage of the
total EARs reported to FDA, our analysis of EIRs found that 25
percent of the facilities in our sample for which we could determine
that collection and processing were observed by the FDA inspector had
problems, while 19 percent had Form 483 observations.  Thus, even
though few EARs are submitted, FDA inspectors regularly find problems
serious enough to warrant a Form 483.\16


--------------------
\16 There are, of course, many situations that could warrant a Form
483 observation that may not be required to be reported as an error
or accident.  Nevertheless, our analysis of EIRs suggests that FDA
regularly finds problems in collection and processing procedures. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 2:3.2

Below we summarize bacterial contamination, the safety issue that we
identified in the area of collection processes. 


         BACTERIAL CONTAMINATION
------------------------------------------------------ Chapter 2:3.2.1

Bacterial contamination is a serious concern, even though disposable
plastic containers and closed systems for blood collection have been
used for many years, improving the aseptic preparation of blood and
blood components.  Data the Canadian Red Cross collected for 1987-91
indicate positive bacterial cultures in approximately 0.4 percent (or
1 in 250) of all units of blood.\17

The incidence of bacterial contamination increases when patients
receive platelet transfusions, because these are often concentrated
from pools of 5 to 10 different donors and stored at room
temperature.  In the Canadian data, the risk of transfusing
bacterially contaminated units into such patients rose to
approximately 2 percent (1 in 50).  Some have pointed out that if
only 5 percent of those bacterially contaminated units could cause a
significant reaction, 1 in 1,000 recipients of pooled platelets would
be exposed to septic reactions.\18 Recognizing this problem, blood
facilities are increasingly using single-donor platelet preparations
in place of pooled platelets because they are thought to offer less
risk of contamination.  However, there are few data to support a
conclusion that the single donor preparations offer a significant
reduction in the risk of bacterial contamination. 

During the past decade, the number of bacterial sepsis episodes (one
in which toxins from bacteria are spread) associated with the use of
blood components has risen dramatically.  The increase mirrors the
increase in the use of platelet concentrate transfusions, but
reactions to bacterially contaminated red cells have also been
reported.  Most of the increase in septic episodes stems from the
room temperature required for storing platelet concentrates.  Twenty
to 24 degrees Celsius is ideal for platelet viability and function,
but it facilitates bacterial proliferation, as does prolonged
storage.\19

With regard to red cells, septic episodes are most likely associated
with bacteria that can proliferate at the recommended refrigeration
temperature of 4 degrees Celsius.  In fact, one risk estimate of
infectious complications from blood transfusions points to bacteria
as the leading cause of death, compared to viruses, parasites,
hemolytic reactions, lung disease, and anaphylaxis.\20

Yet another safety concern is that it has been postulated that a
small core of skin can enter the needle--and, thus, the blood--at the
time of donation.  Available data appear to indicate that the vast
majority of bacteria isolated from platelet concentrates come from
this source.\21

Data also suggest an increasing number of fatalities associated with
bacterial contamination (which is often a result of improper
collection and processing of blood products).  In 1975, FDA
established a registry to compile information on
transfusion-associated deaths.  From 1976 to 1978, 4 percent of such
deaths were attributed to bacterial contamination, a figure that rose
to 10 percent in 1986-88. 


--------------------
\17 According to FDA, Canadian standards for blood collection and
processing differ from U.S.  standards.  Bacterial contamination is
seen as a problem by most experts in the field of blood safety. 

\18 M.  Blajchman and A.  Ali, "Bacteria in the Blood Supply:  An
Overlooked Issue in Transfusion Medicine," in S.  J.  Nance (ed.),
Blood Safety:  Current Challenges (Bethesda, Md.:  American
Association of Blood Banks, 1992). 

\19 M.  Goldman and M.  A.  Blajchman, "Blood Product-Associated
Bacterial Sepsis," Transfusion Medicine Review, 5 (1991), 73-83. 

\20 R.  Dodd, "Adverse Consequences of Blood Transfusion: 
Quantitative Risk Estimates," in S.  T.  Nance, Blood Supply:  Risks,
Perceptions, and Prospects for the Future (Bethesda, Md:  American
Association of Blood Banks, 1994). 

\21 Blajchman and Ali, pp.  220-21. 


      CORRECTIVE MEASURES
-------------------------------------------------------- Chapter 2:3.3

Measures that might eliminate transfusion-associated bacterial sepsis
include improving or instituting quality-control programs, extending
donor screening, modifying blood collecting and processing
techniques, shortening blood-component storage times, testing, and
removing or eliminating the bacteria. 

In an attempt at quality control, some blood facilities (including
all ARC facilities) ask screening questions (such as recent dental
and medical procedures) to determine whether a prospective donor's
blood may be contaminated with bacteria.  However, others have
pointed out that even a 3-day deferral for such events would lose
many potential, healthy donors. 

Most organisms introduced into platelet concentrate units show a
growth lag of about 1-2 days, followed by rapid proliferation.  This
suggests that with longer storage times, the frequency of significant
levels of bacteria would increase.  However, the results of
bacteriologic surveys examining this effect of storage time and
bacterial contamination are inconsistent.\22

Bacterial testing would help catch contaminated blood units, but
traditional culture techniques often require incubation periods of
several days and false-positive and false-negative results are often
a problem.  With this in mind, researchers are developing more rapid
and reliable detection techniques.  Additionally, recent studies have
indicated that bacteria can be filtered from blood by removing white
cells. 


--------------------
\22 Goldman and Blajchman, pp.  72-83. 


TESTING
============================================================ Chapter 3

Testing blood is the third layer of safety.  Routine testing helps
ensure that the right blood type is transfused.  Viral testing and
inactivation procedures help ensure that transfused units of blood
carry no viruses.  As we report in Blood Supply: 
Transfusion-Associated Risks, the risks of viral and nonviral
complications from blood transfusions are quite small in relationship
to risks from other life activities.\1

In routine testing, both blood facilities and hospital transfusion
services make blood-typing errors that can be fatal.  We found
several problems in viral testing, too (all discussed in this
chapter):  improvements in testing to close the window period will be
increasingly costly with fewer cases of positive units being caught;
lack of a requirement to test autologous units for viral markers
could lead to the transfusion of infected blood; lack of confirmatory
testing of repeatedly reactive blood units could hamper a blood
facility's ability to communicate specific information to implicated
donors; lack of lookback procedures for viruses other than HIV could
mean that recipients of infected units might not be informed,
resulting in their failure to seek treatment. 

Further, divergent strains of viruses that blood facilities do not
test for are rarely found in the United States, although some cases
have recently arisen.  However, the viral tests currently in use have
different levels of sensitivity and, thus, do not catch all blood
units that are positive for viral markers.  Viral inactivation
procedures that are used in plasma fractionation rarely remove
nonenveloped viruses (such as hepatitis A and parvovirus).  Plasma
manufacturers do not always employ inactivation procedures for every
plasma product.  And emerging viruses that are not being tested for
could affect the U.S.  blood supply and public health. 


--------------------
\1 U.S.  General Accounting Office, Blood Supply: 
Transfusion-Associated Risks, GAO/PEMD-97-2 (Washington, D.C.: 
1997). 


   ROUTINE TESTING
---------------------------------------------------------- Chapter 3:1

Federal regulations require blood facilities to test each unit of
blood they collect to determine the donor's blood type within the ABO
system.  Discovered in 1900, this system remains the most widely
known.  Next to it in importance is the Rh system, which designates a
person's blood as being either "Rh positive" or "Rh negative." Among
the many other blood typing systems, the ABO and Rh groups are the
most familiar and the most important in determining which blood can
be transfused to which patients. 

Type testing is required also for blood from which plasma is
recovered but not for source plasma.  Additionally, AABB standards
stipulate that a donor's previous ABO and Rh record not be used to
identify his or her blood type in subsequent donations.  This means
that when discrepancies arise, typing is to be determined by
additional direct testing. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 3:1.1

Tables 3.1 and 3.2 summarize our EAR and EIR information for such ABO
blood typing issues as misinterpreted test results, incorrect test
procedures, and products being released prior to testing.  The EAR
data show that routine testing represents 5.7 percent (646 of 11,292)
of all EARs reported to FDA (see appendix II). 



                               Table 3.1
                
                 Routine Testing EAR Rates by Facility
                              Type, 1994\a

                                           Unlicensed\
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c                 1         0.01       0    0.21
EAR rate per 100,000 units            4.8          2.2       0     2.5
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 3.2
                     
                      Routine Testing Problems and Form 483
                      Observations by Facility Type, 1994\a

                                      Transfusion      Plasma
             Licensed   Unlicensed\b    service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %      No.    %      No.      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities   2 of   9%  1 of 19  5.3  6 of 54  11%   0 of 3    0  9 of 98     9%
 with          22                  %
 problems\
 c
Facilities   2 of    9  1 of 19    5  3 of 54    6   0 of 3    0  6 of 98      6
 receiving     22
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined routine testing during its inspection.  Problems
were those that were characterized by the inspector on the inspection
report whereas Form 483 observations were problems deemed serious
enough to be denoted on a Form 483. 

Licensed facilities reported routine testing EARs at a rate more than
100 times that of unlicensed facilities.  (Plasma centers do not
conduct routine testing.) But the EAR rate for licensed facilities
per 100,000 blood units collected was only 2 times greater than the
rate for unlicensed facilities.  In our analysis of EIRs, we found
that FDA inspectors found occasional problems in routine testing
procedures and made few Form 483 observations in this area. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 3:1.2

We describe below the issue of blood typing, a safety concern in the
area of routine testing processes. 


         BLOOD TYPING
------------------------------------------------------ Chapter 3:1.2.1

Routine testing does not appear to have any inherent weaknesses
provided that blood typing is done properly and that correctly typed
units are transfused to the intended patient.  The frequency of
errors is low; however, the consequences of error can be serious. 

A study of errors reported in New York State in 1990-91 found 104
erroneous red cell transfusions out of 1,784,641 (0.006 percent), 54
of which were related to ABO incompatibility.\2 Most of the 50 other
errors were related to the transfusion of an incorrect ABO blood type
that was fortuitously compatible with the recipient's blood type or
to the transfusion of ABO-incompatible fresh-frozen plasma. 

Fifty-eight percent, or 61, of the 104 erroneous transfusions were
solely the result of errors outside the blood facility.  The majority
were caused by the person administering the transfusion failing to
verify the identity of the recipient of the blood unit.  Nearly 25
percent, or 25 incidents, were attributable to the blood facility; 17
percent, or 18 incidents, to both the blood bank and hospital
service.  The authors of the New York study calculated that the
incidence rate of ABO-incompatible errors was 0.003 percent, or 1 in
every 33,000 transfusions.  They also concluded that 3 patients died
from acute transfusion reactions, for a death rate of 1 per 600,000
red cell transfusions. 

Although the error of transfusing ABO-incompatible units can lead to
serious complications for patients, such error occurs most often at
the hospital rather them stemming from the misapplication of
regulations or procedures at the blood facilities.  However, the New
York study outlined blood-facility release, clerical, and technical
errors that accounted for one fourth of all errors in the study.  No
data are available that would allow us to assess the magnitude of
this problem on a national scale. 


--------------------
\2 This is important because transfusing ABO-incompatible blood is a
major noninfectious risk.  J.  Linden, B.  Paul, and K.  P. 
Dressler, "A Report of 104 Transfusion Errors in New York State,"
Transfusion, 32 (1992), 601-6. 


   VIRAL TESTING
---------------------------------------------------------- Chapter 3:2

Viral testing has received the most attention in terms of the safety
of the nation's blood supply.  Many people perceive this to be the
"layer" at which most of the unsafe blood can be caught if it has
worked its way through screening, deferral, and collection.\3

As recently as 1984, blood facilities had to test blood only for HBV
antigen and syphilis.  Since then, further tests have been protecting
the nation's blood supply from infectious diseases.  Blood facilities
presently conduct seven such tests for viruses:  hepatitis B (core
antibody), hepatitis B (surface antigen), hepatitis C antibody, HIV-1
and HIV-2 (antibody), HIV-1 (antigen), HTLV-I and HTLV-II, and
syphilis.\4

FDA has licensed a new HIV-1 test to detect the p24 antigen, a
protein that is part of the virus itself, rather than merely the
virus's antibodies.  Because it detects infections before the HIV
antibody tests, it will close the window period from approximately
22-25 days to about 16-19 days.  It is projected to prevent up to 25
percent of the window-period cases, or about 5 to 10 cases, of
transfusion-transmitted HIV infection per year.  FDA recommended that
blood facilities begin using this test by June 14, 1996. 

FDA's protocols for viral testing stipulate that if the initial test
for viruses is reactive, then two duplicate tests should be made to
determine whether the blood unit has antibodies to a particular
virus.  If either duplicate test is also reactive, the blood
facilities may perform a more specific, confirmatory test to
determine whether the reactivity is false or true.\5

Deciding whether a donation is or is not positive is affected also by
the sensitivity and specificity of the viral tests.\6

Initial tests are fast and usually automated and screen large numbers
of samples.  They are extremely sensitive in order to minimize the
number of false-negative outcomes.  Confirmatory tests are more
time-consuming, usually less sensitive than initial tests, but very
specific.  Table 3.3 outlines the different types of viral test
results and the consequent actions. 



                               Table 3.3
                
                  Results From and Actions After Viral
                                Testing

Result        Definition                   Action
------------  ---------------------------  ---------------------------
Initially     Initial test is reactive     Two duplicate tests are
reactive                                   performed

Repeatedly    One or both duplicate tests  A confirmatory test is
reactive      are reactive                 performed (this test is not
                                           always required); the
                                           prospective donor is
                                           deferred and the collected
                                           unit is discarded

Indeterminat  Duplicate tests are          The donor is deferred and
e             repeatedly reactive and      the collected unit is
              confirmatory test is         discarded
              neither positive nor
              negative

Positive      Duplicate tests are          The donor is deferred and
              repeatedly reactive and      the collected unit is
              confirmatory test is         discarded
              positive

Negative      Initial test is negative     None; the donor is not
              or, if reactive, both        deferred
              duplicate tests are
              negative
----------------------------------------------------------------------
Thus, any unit that is repeatedly reactive is considered positive
even if a confirmatory test determines that the testing procedure
produced a false-positive result.  Such results require that the
donor be deferred.  FDA recommends but does not require that donors
who are repeatedly reactive but indeterminate or negative by a
confirmatory test should be notified and placed on donor deferral
registries. 

FDA has also outlined procedures by which donors who have repeatedly
tested reactive for HBsAg, HCV, and HIV can be brought back as
donors.  There are no such procedures for HBc and HTLV because
licensed confirmatory tests do not exist for them. 

FDA requires all blood facilities to maintain quality-assurance
programs and to test their laboratory devices and personnel for
proficiency in order to keep testing errors to a minimum.  FDA also
issues quality-assurance guidance that includes quality-control
procedures for standard operating procedures, competency evaluations
of personnel training and education, and laboratory proficiency
tests.  Additionally, laboratories that perform viral testing are
inspected by HCFA (through a memorandum of understanding with FDA)
and state health departments.  Table 3.4 shows key features of viral
and nonviral testing. 



                                    Table 3.4
                     
                        Key Features of Viral and Nonviral
                                     Testing

                            Date licensed or
                            recommended by    Formal            Reentry
Disease   Test              FDA               requirements      procedure\a
--------  ----------------  ----------------  ----------------  ----------------
Chagas'   None licensed                       None              None

CJD       None licensed                       None              None

CMV       None licensed                       None              None

HAV       None licensed                       None              None

HBV       Core              Sept. 1991        All units must    For HBsAg
                                              be tested

          Surface antigen   1972

          3rd generation    Dec. 1987
          (HBsAg)

HCV       1st generation    Nov. 1990         All units must    Yes
                                              be tested

          2nd generation    March 1992

HIV-1     1 antibody        March 1985        All units must    Yes
and HIV-                                      be tested
2

          1/2 antibody      June 1992

          p24 antigen       March 1996

HTLV-I    Antibody          Nov. 1988         All units must    None
                                              be tested

HTLV-II   None licensed                       Tested through    None
                                              HTLV-I tests

Parvovir  None licensed                       None              None
us

Syphilis                    Approximately     All units must    Yes
                            1960              be tested
--------------------------------------------------------------------------------
\a Procedures can be followed by blood facilities to allow previously
deferred donors to donate again if certain protocols are followed. 
These protocols are outlined in memoranda to blood facilities
relating to specific viruses. 

In addition to testing procedures, a series of manufacturing steps
remove or inactivate viruses that are in plasma pools from source and
recovered plasma donations.\7 Two main techniques decrease viral
ability to infect plasma products:  partitioning, or removal of a
virus, is the physical separation of the virus or viral particles
from the therapeutic component.  Inactivation of a virus destroys it
so that the remaining viral fragments lack the structure and
components needed to infect the blood.\8

Removal processes include filtration, affinity chromatography, ion
exchange chromatography, and polyethylene glycol fractionation. 
Heating and solvent detergent treatments are examples of processes
that inactivate viruses.  Additionally, some processes, such as
ethanol fractionation, both remove and inactivate viruses. 

In order to be effective, viral removal or inactivation techniques
must destroy at least one of the essential elements of viral
replication.\9 These techniques work in different ways to accomplish
this task.  Photosensitizing techniques use light- activated dyes
that are irradiated, causing the dyes to convert to molecules that
can destroy DNA or membrane lipoproteins.  Heat treatment denatures
viral proteins and nucleic acids, rendering them incapable of viral
replication.  Irradiation processes inhibit viral DNA by inducing
breaks and linkages.  Solvent detergent techniques destroy the viral
envelope in lipid-enveloped viruses. 


--------------------
\3 In appendix I, we characterize some viral and nonviral agents that
are transmissible in blood and highlight key federal guidance and
industry practice as they relate to these agents. 

\4 In response to a January 9-11, 1995, NIH consensus development
conference, AABB dropped a test to measure alanine aminotransferase
(ALT), a surrogate marker for hepatitis.  The conference had
concluded that ALT testing was not needed as a surrogate marker for
non-A, non-B, hepatitis because of the increased sensitivity of HCV
tests.  FDA has stated that it does not recommend either for or
against ALT testing.  The CFRs require tests for hepatitis B surface
antigen (HBsAg), HIV, and syphilis but not HTLV or HCV.  This
conference recommended that syphilis testing continue. 

\5 False-negative blood units are truly positive for a virus that is
undetected by the initial test.  False-positive units test positive
for a virus that proves in a confirmatory test not to be present. 
Confirmatory tests can also be "indeterminate," meaning that it is
not possible to tell for sure whether a virus is or is not present. 
Some studies have suggested that most indeterminate confirmatory
tests are probably negative.  However, FDA considers indeterminacy to
be a positive reading because of the chance that the blood unit does
indeed contain a virus. 

\6 "Sensitivity" is the probability of a unit's testing positive if a
virus is truly present.  As sensitivity increases, the number of
persons whose blood contains the virus but who are missed (false
negatives) by being incorrectly classified decreases.  In other
words, sensitivity = true positives / (true positives + false
negatives).  "Specificity" is the probability of a unit's testing
negative if a virus is truly absent.  A highly specific test is
rarely positive when a virus is not present and therefore results in
fewer persons without the virus being incorrectly classified (false
positives).  In other words, specificity = true negatives / (true
negatives + false positives). 

\7 Cytomegalovirus (CMV)is not present in plasma or plasma products. 
Nonenveloped viruses such as hepatitis A virus (HAV) and parvovirus
are not affected by some inactivation procedures.  FDA has not
recommended the exclusion of repeatedly reactive HBc plasma because
exclusion might decrease the safety of plasma derivatives through the
likely reduction of an antibody to HBsAg.  Plasma donors are tested
for HBsAg, HCV, HIV, and syphilis.  Testing of plasma donors for
HTLV-I and HTLV-II is not required because of their cell association. 

\8 These techniques are not used to remove or inactivate viruses in
red cells or platelets because the techniques are usually accompanied
by red cell damage. 

\9 Viral replication requires cell attachment by the virus to a cell
receptor, penetration of the cell, replication and translation of
viral nucleic acids, and exit from the cell with integrated viral
particles. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 3:2.1

Only 2 percent (274 of 11,292) of EARs in 1994 related to viral
testing, probably a result of the increasing automation of viral
testing procedures.  Errors in viral testing included misinterpreting
the results, releasing products before testing, and testing
incorrectly.  Table 3.5 shows that licensed facilities reported viral
testing EARs nearly 300 times more than unlicensed facilities and 30
times more than plasma centers.  Table 3.6 shows that a large
percentage of all types of blood facilities for which we found
evidence that viral testing had been observed by an FDA inspector
were found to have problems relating to viral testing procedures. 
Also, 24 percent (9 of 37) of licensed facilities and 50 percent (6
of 12) of unlicensed facilities received Form 483 observations
associated with viral testing. 



                               Table 3.5
                
                  Viral Testing EAR Rates by Facility
                              Type, 1994\a

                                           Unlicensed\
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c              0.83        0.003    0.03    0.09
EAR rate per 100,000 units            2.0          0.5     0.1     1.1
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 3.6
                     
                       Viral Testing Problems and Form 483
                         Observations by Facility Type\a

                                      Transfusion      Plasma
             Licensed   Unlicensed\b    service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities  10 of  29%  7 of 12  58%  4 of 11  37%  3 of 12  25%    24 of    34%
 with          35                                                      70
 problems\
 c
Facilities   9 of   26  6 of 12   50  2 of 11   18  2 of 12   17    19 of     27
 receiving     35                                                      70
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined viral testing during its inspection.  Problems were
those that were characterized by the inspector on the inspection
report whereas Form 483 observations were problems deemed serious
enough to be noted on a Form 483. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 3:2.2

Most of the safety issues related to viral testing result in a very
remote chance of transfusion-transmitted infections.  This is because
of the low incidence of infectious disease in the U.S.  blood supply
and other factors such as transmission rates through blood products. 


         THE WINDOW PERIOD
------------------------------------------------------ Chapter 3:2.2.1

The window period of undetectability differs from test to test,
ranging from 16-19 days for the p24 antigen HIV test to approximately
70 days for the HCV test.  Other testing procedures can reduce the
window period, but the tests are expensive and are not yet automated. 
For example, a test that incorporates a technology known as
"polymerase chain reaction" may reduce the window period for HIV
testing from 16-19 days to approximately 11 days.  While the cost of
implementing it is roughly $200 million, it would catch an estimated
additional 5-10 HIV transmissions through blood products.  Efforts
continue to develop more effective tests, but important cost-benefit
trade-offs are often part of the discussion as to the merits of such
tests. 


         AUTOLOGOUS DONATIONS
------------------------------------------------------ Chapter 3:2.2.2

There is no requirement that all autologous blood be tested for viral
markers, but recent information on errors involving such blood raises
some questions.  A 1995 AABB survey of its institutional members
found that 1.2 percent of the 1,829 respondents reported giving one
or more autologous blood units to an unintended transfusion
recipient.  Of the 22 who did this, 5 did not test autologous
collections for viral markers.  Additionally, 3.7 percent of the
respondents reported that untested, recovered plasma from autologous
donors was shipped for further manufacture; 12.3 percent reported
that autologous units had been lost in transit.  Lastly, the survey
found that approximately half of the respondents did not test for
viral markers on autologous collections.  This information points to
a potential vulnerability of viral testing in allowing the
possibility for untested units to be transfused to other recipients. 
FDA is currently developing a recommendation regarding testing
autologous units of blood. 


         CONFIRMATORY TESTING
------------------------------------------------------ Chapter 3:2.2.3

No FDA guidance requires confirmatory testing of all units that test
positive for viral markers, although repeatedly reactive donations
are discarded and such donors are permanently deferred.  A recent
final rule published on September 9, 1996, does require blood
facilities to perform more specific tests when a donor who previously
donated blood is tested on a later donation and has repeatedly
reactive test results for HIV.\10 However, this requirement is only
for HIV.\11

Confirmatory tests do not in and of themselves improve the safety of
the blood supply.  However, without such tests, blood facilities
cannot know what specific information they should provide to a donor
or whether the donor is infected.  This could prove problematic if,
for example, a blood facility notified a donor of a repeatedly
reactive result but stated that it might be a false-positive finding
and counseled the donor that he or she might want to obtain a
confirmatory test from a physician.  If the donor chose not to do
this, public health might suffer. 


--------------------
\10 The final rule amended the current good manufacturing practices
for blood and blood products by requiring blood facilities to notify
consignees who had received blood and blood components at increased
risk for transmitting HIV infection.  A companion HCFA final rule,
"Medicare and Medicaid programs:  Hospital Standard for Potentially
HIV Infectious Blood and Blood Products," requires all transfusion
services subject to HCFA's conditions of Medicare participation for
hospitals to notify transfusion recipients who have received blood or
blood components from a donor whose subsequent donation test results
were positive for antibody to HIV.  FDA is requiring transfusion
services that do not participate in Medicare, and are therefore not
subject to HCFA's final rule, to notify transfusion recipients. 
Transfusion services are also required to notify the physician of
patients who receive units that may be positive for HIV; if the
physician refuses to notify the patient, the transfusion service is
required to make attempts at notification. 

\11 Not all screening tests have a licensed confirmatory test (for
example, HTLV), but such tests are currently available for HCV and
HBV, in addition to HIV. 


         LOOKBACK PROCEDURES
------------------------------------------------------ Chapter 3:2.2.4

Lookback procedures have been established by FDA to notify consignees
(that is, transfusion services) of blood from donors who subsequently
test positive for HIV.  These transfusion services are responsible
for notifying the physicians of recipients who receive blood from
donors.  If the physician is unavailable or declines to notify the
recipient, the transfusion service is to notify the recipient and
inform him or her of the need for HIV testing and counseling. 
However, these requirements pertain only to repeatedly reactive HIV
donations.\12 The result is that patients who are transfused with
units that are repeatedly reactive for HBV and HCV may never be told
that they may be infected, with potentially adverse consequences for
their sexual partners as well as the general public. 

Although HCV is the virus most often transfused in blood, lookback
procedures for HCV are only now being considered.  The reasons given
for this are that, first, there was until recently no confirmatory
test for HCV, so that false-positive units could not be identified. 
This is no longer the case since FDA has licensed an HCV confirmatory
test. 

A second argument put forth in the past for not having lookback for
HCV was that there was no treatment for persons infected with HCV. 
Thus, a lookback procedure would not assist a patient in treating
conditions resulting from transfusions tainted with HCV-positive
blood.  However, recent studies of treatment with interferon suggest
that it may control HCV and lead to complete or nearly complete
recovery in some patients.\13 Also, some recipients might benefit
from being notified so that they might curtail behavior that could
cause more progressive harm after being infected with such viruses as
HBV and HCV (for example, consumption of alcohol).  Furthermore,
lookback is recommended for HIV even though no treatment for this
virus results in complete recovery. 

Third, some point out that the way in which HCV is transmitted is not
precisely known.  Thus, it would be difficult to tell people how to
protect themselves.  However, Centers for Disease Control and
Prevention (CDC) surveillance data from 1992 note that non-A, non-B,
hepatitis (most often HCV) is transmitted through blood transfusions,
intravenous drug use, and sexual and household contact.\14 Even
though the exact means of transmission have not been defined, it is
well understood that certain activities increase the likelihood of
acquiring HCV.\15 In a related argument, some have noted that most
HCV transmissions are not associated with blood transfusions.  This
is also true for HIV--most transmissions of HIV are not related to
blood or blood products--yet FDA now requires lookback for
HIV-implicated blood products. 

An internal public health service study, "Public Health Service
Options for Identification of Hepatitis C Virus Infection Among
Transfusion Recipients," dated March 28, 1996, pointed out that a
decision to conduct lookback should be based on several
considerations.  One of these was "the cost of case-finding,
including diagnosis and treatment, should be reasonably comparable
with respect to other medical care and preventive services." This
argument, based on cost considerations, has also been used to argue
against lookback for HCV.  However, a recent study suggests that the
cost-effectiveness of lookback for HCV may be comparable to that of
many common public health interventions.\16

As with confirmatory testing, lookback procedures do not increase the
safety of the blood supply.  However, they do allow the provision of
more accurate information to donors and recipients.  With such
information, a donor who has been identified as having given blood
that tests positive and a recipient who receives such blood could
alter their behavior to ensure that they did not infect others. 
Additionally, recipients might be more likely to seek treatment if
they knew that they had received blood which was likely to have been
infectious. 


--------------------
\12 This notification process is to include a minimum of three
attempts to notify the recipient and to be completed within a maximum
8 weeks of the receipt of the result of a licensed confirmatory test
for HIV.  Additionally, the transfusion service is required to
document the notification or attempts to notify the recipient's
physician or the recipient. 

\13 According to G.  Davis et al., "Treatment of Chronic Hepatitis C
With Recombinant Interferon Alfa," New England Journal of Medicine,
321 (1989), 1501-6, after 6 months of treatment with interferon, 46
percent of patients had complete or nearly complete recovery with 3
million units of interferon versus 28 percent for those receiving 1
million units and 8 percent in untreated patients.  However, relapse
of high ALT levels 6 months after the completion of treatment
occurred in 47 percent of the patients.  This study followed these
patients for only 6 months after the treatment ended, and the
researchers noted that further follow-up might find a late recurrence
in the form of elevated ALT levels.  More recent data have shown that
approximately 50 percent of patients with chronic HCV respond to
alpha-interferon, with 10 to 20 percent achieving long-term response. 

\14 Centers for Disease Control and Prevention, Hepatitis
Surveillance, report 55 (Atlanta:  June 1994). 

\15 A recent presentation at the 1996 AABB National Meeting outlined
a case of sexual transmission of HCV.  See C.  Capelli et al., "A
Case of Transmission of Hepatitis C Virus Between Sexual Partners,"
Transfusion, 36 supp.  (1996), 51S. 

\16 J.  P.  Aubuchon, J.  D.  Birkmeier, and M.  S.  Alter,
"Cost-Effectiveness of HCV Lookback," Transfusion, 36 supp.  (1996),
51S. 


         DIVERGENT VIRAL STRAINS
------------------------------------------------------ Chapter 3:2.2.5

A potential problem for HIV testing is the inability to detect
divergent viral strains.  Recent CDC work found that 6 of 10 licensed
HIV antibody screening tests failed to detect one or more samples of
a rare, divergent strain of HIV-1, of which almost all the
approximately 100 cases had been identified in West and Central
Africa.\17

Additionally, in July 1996 the first documented case of one of these
divergent strains (HIV group O) was recognized in the United
States.\18 Viral testing of this individual throughout 1995 showed
both negative and positive tests for HIV and indeterminate results
with confirmatory tests (this individual had emigrated to the United
States in 1994).  CDC investigators also evaluated five licensed HIV
tests using blood samples from this individual in April 1996.  At
that time, four of the five tests were positive while one test was
nonreactive.\19 Current data suggest that, overall, FDA-approved HIV
tests now in use detect group O HIV infections approximately 80
percent of the time.\20

The CDC investigators noted that the risk to the U.S.  blood supply
was remote because most persons infected with this HIV-1 strain are
excluded before donating blood by current malaria screening
guidelines.  Additionally, of the more than 590,788 AIDS and HIV
cases reported to CDC through December 1995, 106 have been from
persons whose country of origin was in West Africa or Central Africa
where group O infections have been reported.  CDC has pointed out
that divergent strains could infect persons living in the United
States and that these often remain undetected by current HIV antibody
tests.  CDC has also noted that this should be a concern to public
health officials and blood facilities.  In response, FDA has
recommended three additional screening questions relating to birth
and travel to several West African countries. 

Additionally, FDA has mandated that any new HIV tests being submitted
for licensure in the U.S.  be capable of detecting this HIV strain. 
FDA has also directed manufacturers of all currently-licensed tests
to modify the test kits to ensure that this strain could be
identified in U.S.  blood donors. 


--------------------
\17 C.  Schable et al., "Sensitivity of United States HIV Antibody
Tests for Detection of HIV I Group O Infections," Lancet, 344 (1994),
1333-34. 

\18 Almost all the cases of HIV in the United States are from the
HIV-M group. 

\19 Centers for Disease Control and Prevention, "Identification of
HIV I Group O Infection-Los Angeles County, California, 1996,"
Morbidity and Mortality Weekly Report, 45 (1996), 561-65. 

\20 A second documented case of HIV-1 group O infections was
identified in the U.S.  as part of CDC's surveillance activities for
unusual HIV-1 variants.  Both of these individuals have never donated
blood or plasma. 


         TEST SENSITIVITY
------------------------------------------------------ Chapter 3:2.2.6

Most units of infected blood are caught by testing before
transfusion.\21 However, some are not.  A recent case of an
individual who had AIDS but tested negative on the HIV test
illustrates that the tests presently used are not perfect in
detecting all donations that have positive viral markers.\22 This
case, although extremely rare, involved an individual who had a rare
immune reaction that interfered with the development of HIV
antibodies.  Information from CDC indicated that this is one of only
a handful of isolated reports of HIV-infected persons who do not
produce enough antibodies to be detected.  Furthermore, DNA analysis
of this individual's blood ruled out an atypical HIV viral strain. 
This individual was a regular plasma donor and, to date, no HIV
infections have been identified among recipients of products from
this donor. 

HBV is a virus that seems to be at times difficult to detect with
available testing procedures.  A recent study that examined
open-heart-surgery patients who had unexplained posttransfusion
hepatitis found that 20 percent of them (4 of 20) had no
immunological indications for HBV but were, in fact, HBV positive as
determined by polymerase chain reaction testing.\23 The results
showed that HBV may be transmitted despite rigorous testing of donors
for HBc and HBsAg. 

The present HCV test can identify most persons infected with the
virus, but this test may not to capture 10 percent of those who are
positive for HCV.  This inability stems from the sensitivity of the
present HCV test and the potential for a chronic carrier state for
HCV that goes undetected by antibody testing.  The uniformly high
rate of chronic hepatitis after HCV infection suggests HCV may be a
major cause of chronic liver disease in the United States. 

Recent advances in the sensitivity of the HTLV-I tests to detect
HTLV-II have increased the efficacy of this test.  The currently
licensed HTLV-I tests still do not detect about 3 to 4 percent of
HTLV-II positive units.\24 The most prevalent strain of HTLV in the
United States is HTLV-II, and the results from the Gallo study point
out that improvements still need to be made to increase test kit
sensitivities for HTLV-II.  Until recently, there has been little
evidence of a known disease condition associated with the presence of
HTLV-II antibodies.  However, some recent evidence suggests an
association with immunologic impairment with HTLV-II.\25


--------------------
\21 See U.S.  General Accounting Office, Blood Supply: 
Transfusion-Associated Risks, GAO/PEMD-97-2 (Washington, D.C.: 
1997). 

\22 Centers for Disease Control and Prevention, Morbidity and
Mortality Weekly Report, March 8, 1996. 

\23 J.  Rasenack, "Hepatitis B Virus Infection Without Immunological
Markers After Open-Heart Surgery," Lancet, 345 (1995), 355-56. 

\24 D.  Gallo et al., "Comparison of Four Enzyme Immunoassays for
Detection of Human T-Cell Lymphotropic Virus Type II Antibodies,"
Journal of Clinical Microbiology, 34:1 (1996), 213-15. 

\25 E.  L.  Murphy et al., "Medical Conditions Associated with Human
T-Lymphotropic Virus Types I and II (HTLV-I and II) Infection,"
Transfusion, 36 supp.  (1996), 43S. 


         VIRAL INACTIVATION
------------------------------------------------------ Chapter 3:2.2.7

Plasma fractionation companies have introduced several new steps to
inactivate viruses but they are not very successful against
nonenveloped viruses such as hepatitis A.\26 For example, in January
1996, U.S.  health officials reported the first documented
transmission of HAV through blood-clotting substances. 

Furthermore, FDA gives little guidance on the inactivation procedures
that manufacturers should use to inactivate specific products from
viruses.  Thus, a manufacturer may or may not be using inactivation
procedures to eliminate viruses from plasma pools.  In fact, this
problem arose in the fall of 1993 when some intravenous immune
globulin (IVIG) products--used to treat patients with lymphocytic
leukemia or immune disorders, including AIDS--were implicated in the
transmission of HCV to transfused patients.  These products were from
a fractionation company that did not have an inactivation procedure
in its manufacturing process for IVIG, although other manufacturers
did. 

As of January 1995, 5 of 6 manufacturers had incorporated a viral
inactivation step in their IVIG processes.  However, this is still a
problem because another product, intramuscular immune globulin
(IMIG), is not put through an inactivation step by most of the
manufacturers.\27 As one FDA official noted, "while there has been no
transmission of HCV by IMIG, this is a very scary situation." Some of
this problem may have been mitigated when FDA announced that it would
test all lots of immunoglobulin products for HCV that had not
undergone viral inactivation steps.  Nevertheless, this example
illustrates disparities among the fractionation companies and how
similar products may or may not be undergoing viral removal
procedures. 


--------------------
\26 Most inactivation procedures attack the physical envelope of the
virus, negating its ability to replicate.  By definition,
nonenveloped viruses do not have this envelope and are therefore
difficult to kill. 

\27 FDA has licensed to one manufacturer a viral inactivation
procedure for IMIG. 


         EMERGING VIRUSES
------------------------------------------------------ Chapter 3:2.2.8

Among a number of emerging viruses that could affect the U.S.  blood
supply are hepatitis E (HEV) and hepatitis G (HGV).  Tests to detect
these viruses are not currently available.\28 Other emerging viruses,
such as ebola, that have gained worldwide attention have not been
seen in the U.S.  blood supply. 

HEV, too, does not appear to be endogenously transmitted in the
United States.  It should be expected only very rarely in travelers
returning from overseas where it is endemic, such as in developing
countries where it is transmitted through the oral-fecal or drinking
water routes.  The major cause for concern with this virus is that,
although it mimics HAV in its course of infection, fulminant
hepatitis is much more common with HEV than HAV.  This is
particularly a concern for pregnant women, in whom the overall
mortality rate may be as high as 20 percent.  Severe complications
from infection with HEV may be avoidable in the near future, since
recent research has found that an HEV vaccine now going through
laboratory studies protects infected persons from developing
hepatitis. 

The discovery of HGV portends another safety issue in viral testing. 
This virus is associated with chronic hepatitis and is transmissible
through blood transfusions.  Preliminary donor studies have indicated
that between 1 percent and 2 percent of the U.S.  blood donor
population is infected with HGV and that HGV accounts for 0.3 percent
of all acute hepatitis in the United States.  The risk factors for
HGV appear to be similar to those for HCV (hemophiliacs, anemia
patients who have multiple transfusions, and intravenous drug users). 
Additionally, studies show that between 10 and 20 percent of patients
with chronic hepatitis that could not be attributed to other causes
were infected with the virus.\29

Because HGV is a newly discovered virus, there are no tests to detect
it.  Some have suggested that tests may not be needed because HGV
carriers are often infected with other hepatitis viruses.  In
contrast, the transmission of HGV by transfusion was documented in 3
of 13 open-heart surgery patients at NIH with posttransfusion
hepatitis and no evidence of hepatitis A-E.  In both cases, an
HGV-positive blood donor was identified. 


--------------------
\28 We do not discuss hepatitis D because it is an incomplete virus
that requires the helper function of HBV to replicate.  Thus, HDV is
acquired as either a co-infection with HBV or a superinfection of
chronic HBV. 

\29 J.  Linnen et al., "Molecular Cloning and Disease Association of
Hepatitis G Virus:  A Transfusion-Transmissible Agent," Science, 271
(1996), 505-8. 


QUARANTINING AND OTHER PROCESSING
STEPS
============================================================ Chapter 4

The fourth safety layer involves quarantining units of blood.  Other
procedures discussed in this chapter include gathering postdonation
information, labeling, and storage and distribution.  Recording
postdonation information allows blood facilities to flag units of
blood that may be unsuitable for use.  Labeling delineates a unit's
blood type (ABO and Rh) and product type (such as red cells and
platelets) and whether it is for autologous or allogeneic use. 
Quarantining, the actual safety layer, includes procedures that
separate blood that has been tested and found suitable for
transfusion from untested blood and from blood that has been tested
and found to be unsuitable for transfusion.  The storage and
distribution processes allow blood facilities to ensure that blood
products are stored at proper temperatures and sent to their proper
destinations. 

More than one third of all EARs submitted to FDA in 1994 were in the
area of postdonation information (see appendix II).  This could
indicate either that the blood safety system is working well or that
what relates to postdonation information in FDA's EAR guidance is
poorly understood.  Additionally, there is a wide disparity between
EARs reported by licensed blood facilities and plasma centers with
regard to postdonation information.  It is unknown why this disparity
exists, since these two types of blood facilities collect
approximately the same number of units of blood.  There are no
weaknesses inherent in the labeling and quarantining procedures when
they are carried out properly.  It should be noted, however, that
mislabeling, while not common, can have fatal consequences.  We found
that only inventory management is a safety issue in storage and
distribution. 


   POSTDONATION INFORMATION
---------------------------------------------------------- Chapter 4:1

Postdonation information from the donor or someone else-- whether a
blood facility receives it by telephone or by some other
means--alerts the facility as to whether or not the donation should
be used.  This might include a donor's alert that he or she became
ill after donating the blood or other information such as high-risk
behavior that would have deferred the donor had it been known
earlier.  Blood facilities establish and maintain procedures for
receiving, evaluating, investigating, and following up possible
errors and accidents relating to postdonation information. 

FDA recommends that facilities have processes in place to (1) receive
and document postdonation information that identifies the
information's source, (2) perform medical evaluations that assess and
investigate potential risks, (3) make timely investigations of EAR
reports to determine whether the quality of blood or blood products
has been compromised, (4) notify those to whom blood is distributed
about how to dispose of affected units, and (5) assess the donor's
suitability as a future donor. 

Blood facilities do not need to submit an EAR if the donor should not
have been deferred and if the medical evaluation indicates that the
blood product's quality was not compromised.  For example, subsequent
cold symptoms do not have to be reported.  However, FDA may evaluate
the situation as a potential recall. 

FDA also recommends how blood facilities should handle situations in
which donors call and report that their blood should not be used but
provide no further information.  In such cases, the facilities are to
retrieve the blood products donated by those donors. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 4:1.1

Postdonation information represented a large percentage of all EARs
submitted to FDA in fiscal year 1994 (3,815 of 11,292, or 34
percent).  Postdonation information includes a donor's informing a
facility of hepatitis, cold, or influenza symptoms or of sexual
partners who have tested positive for HIV.  Table 4.1 shows that
licensed facilities reported postdonation EARs at a rate more than
3,000 times higher than that of unlicensed facilities and 135 times
higher than that of plasma centers.  Their rate per 100,000 units
collected was 52 times higher than unlicensed facilities and 88 times
higher than plasma centers.  According to our analysis of EIRs,
postdonation information issues resulted in few problems being found
by FDA inspectors and rarely resulted in Form 483 observations.  In
fact, we found that only quarantining and routine testing resulted in
fewer Form 483 observations. 



                               Table 4.1
                
                Postdonation EAR Rates by Facility Type,
                                 1994\a

                                           Unlicensed\
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c              12.2        0.004    0.09    1.25
EAR rate per 100,000 units           29.8         0.57    0.34    14.7
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 4.2
                     
                        Postdonation Problems and Form 483
                         Observations by Facility Type\a

                        Unlicensed\\  Transfusion      Plasma
             Licensed        b          service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities   1 of   4%  7 of 74  10%  2 of 40   5%  0 of 30    0    10 of     7%
 with          28                                                     142
 problems\
 c
Facilities   0 of    0  5 of 74    7  2 of 40    5  0 of 30    0     7 of      4
 receiving     28                                                     172
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined postdonation information during its inspection. 
Problems were those that were characterized by the inspector on the
inspection report whereas Form 483 observations were those problems
deemed serious enough to be denoted on a Form 483. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 4:1.2

Below we provide information on discrepancies between the number of
EARs submitted by licensed, unlicensed and plasma facilities, the one
safety issue in the area of postdonation information. 


         EAR DISCREPANCIES
------------------------------------------------------ Chapter 4:1.2.1

The large number of EARs from licensed blood facilities is a concern. 
It could indicate that the system is working properly or that FDA
should more clearly define what is to be reported.  Since
postdonation processes in licensed, unlicensed, and plasma facilities
are similar, the large discrepancy in their numbers of postdonation
EARs is also of concern.  The source of the discrepancy might
indicate problems in the blood-banking system that require attention. 
According to one large blood organization, there are no complete
guidelines for postdonation EARs, which also may result in over- or
underreporting EARs. 

Furthermore, EARs associated with postdonation information appear to
point to potential problems in donor-screening practices.  For
example, in fiscal year 1995, 65 percent of all EARs relating to
postdonation information stemmed from information obtained at a
subsequent donation.  It is not known whether blood-facility
personnel had erred during the first screening or whether the donors
lied or had forgotten about certain activities.  Regardless of the
reason, the data indicate that information that might have been
obtained at earlier screenings was not collected and, therefore, did
not lead to warranted deferral.  Also, blood industry representatives
pointed out that some FDA guidelines do not clearly define the scope
of changes requested in a new guidance document.  This, they believe,
often results in unnecessary reporting of EARs that are not the
result of failure to elicit information. 


   LABELING
---------------------------------------------------------- Chapter 4:2

Carefully identifying and properly labeling blood units and the tubes
they are collected in for testing are essential safety steps.  AABB's
accreditation manual notes that the "original label and added
portions of the label shall be attached firmly to the container and
shall be in clear, eye-readable type, which also may be machine
readable."\1 Typewritten or computer-generated labels are most often
used; handwritten labels are acceptable but only for temporary
expedience.\2

Each laboratory that processes donor blood must ensure that the
unique number it assigns to a donor appears on the donor record, the
primary collection bag, all satellite collection bags, and all tubes
used for processing.  This allows the prompt identification of
specific blood units when and if tests reveal abnormal or discrepant
results. 

Recently, AABB's Committee on Commonality has been working with the
International Society for Blood Transfusion and an FDA liaison member
to develop a world standard for labeling blood and blood products
with a bar code system that by July 4, 1997, would replace the most
widely used bar code system in the United States. 


--------------------
\1 American Association of Blood Banks, Accreditation Requirements
Manual, 5th ed.  (Bethesda, Md.:  1994), p.  107. 

\2 Labels requiring information that is not standard must be
handwritten; for example, labels that require the specific volume of
the product such as a frozen plasma unit. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 4:2.1

We found that labeling errors were commonly reported in 1994 (1,503
of 11,292 EARs, or 13 percent), including missing or incorrect labels
for ABO and Rh typing, autologous units, and expiration and
collection dates.  Table 4.3 shows that licensed facilities reported
labeling EARs at a rate about 475 times more than that of unlicensed
facilities and nearly 300 times more than that of plasma centers. 
Their rate per 100,000 units collected was 5 times higher than
unlicensed facilities and nearly 300 times higher than plasma
centers.  We found from the EIR information from facilities where we
could determine that labeling activities were observed by an FDA
inspector that licensed blood facilities had more problems in
labeling (based on problems found by FDA inspectors and the
percentage of Form 483 observations) than unlicensed ones.  (See
table 4.4.)



                               Table 4.3
                
                  Labeling EAR Rates by Facility Type,
                                 1994\a

                                            Unlicensed
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c              4.74         0.01   0.016    0.49
EAR rate per 100,000 units           11.6          2.3    0.04     5.8
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 4.4
                     
                          Labeling Problems and Form 483
                         Observations by Facility Type\a

                                      Transfusion      Plasma
             Licensed   Unlicensed\b    service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities   9 of  27%  5 of 41  12%  9 of 53  17%  7 of 40  18%    30 of    18%
 with          33                                                     167
 problems\
 c
Facilities   8 of   24  5 of 41   12  5 of 53    9  6 of 40   15    24 of     14
 receiving     33                                                     167
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined labeling during its inspection.  Problems were
those that were characterized by the inspector on the inspection
report whereas Form 483 observations were problems deemed serious
enough to be denoted on a Form 483. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 4:2.2

Labeling practices do not appear to have any inherent weaknesses
provided labeling is done properly. 


   QUARANTINING
---------------------------------------------------------- Chapter 4:3

The fourth safety layer, quarantining, is very important in
preventing the distribution of unsuitable blood.  Blood facilities
maintain separate storage areas for units that have not yet been
tested, units that are to be retested or are repeatedly reactive, and
units that are suitable for distribution.  Blood intended for
autologous use is stored separately from blood for allogeneic use. 
However, FDA's guidance states that although products must be stored
separately, they do not have to be placed in different refrigerators. 
In addition to separating products, quarantining is often aided by
the use of computer systems to prevent the erroneous release of blood
or blood products. 

FDA requires that blood facilities promptly (within 72 hours if
possible) identify and quarantine units from prior collections dating
back 5 years or 12 months prior to the most recent negative screening
test, whenever a donor has a repeatedly reactive screening test for
antibodies to HIV.  For plasma for fractionation, this figure is
reduced to 6 months, provided it has not been pooled or further
processed.  Furthermore, consignees that have been sent such blood
products are to be notified so that they can hold them in quarantine. 
Releasing blood and plasma from quarantine requires that the donor
subsequently tests negative on a confirmatory test for antibodies to
HIV-1.\3 However, as noted previously these requirements are directed
only at units that might be positive for HIV.  No such requirements
are present for units that might be positive for other viruses. 


--------------------
\3 Pending availability of a licensed confirmatory test for HIV-2, a
second different antibody test for HIV-2 should be used along with a
licensed confirmatory test for HIV-1 when the donor's subsequent
donation is found to be HIV-2 positive. 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 4:3.1

EARs submitted in 1994 indicate that quarantining made up 10 percent
of EARs submitted to FDA (1,087 of 11,298).  This includes the
release of products other than those ordered, the release of outdated
products, and the failure to quarantine units that are reactive for
viral markers.  As table 4.5 shows, licensed facilities reported
quarantine EARs at a rate more than 300 times that of unlicensed
facilities and 85 times that of plasma centers.  Their rates per
100,000 units collected were 4.5 and 64 times higher, respectively. 
In contrast, table 4.6 shows that FDA found very few problems
relating to quarantine procedures during inspections, and facilities
received the fewest number of Form 483 observations in this area. 



                               Table 4.5
                
                Quarantining EAR Rates by Facility Type,
                                 1994\a

                                            Unlicensed
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c              3.39         0.01    0.04    0.36
EAR rate per 100,000 units            8.3          1.9    0.13     4.2
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 4.6
                     
                        Quarantining Problems and Form 483
                          Observations by Facility Type

                        Unlicensed\\  Transfusion      Plasma
             Licensed        b          service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities   1 of   3%  2 of 40   3%   2 of 6    2  0 of 30    0     5 of     3%
 with          30                               3%                    163
 problems\
 c
Facilities   1 of    3  2 of 40    5   1 of 6  2 2  0 of 30    0     4 of      2
 receiving     31                                                     163
 Form 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined quarantining during its inspection.  Problems were
those that were characterized by the inspector on the inspection
report whereas form 483 observations were problems deemed serious
enough to be denoted on a Form 483. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 4:3.2

Quarantining practices do not appear to have any inherent weaknesses
provided quarantining is done properly. 


   STORAGE AND DISTRIBUTION
---------------------------------------------------------- Chapter 4:4

The storage and distribution of products constitute the last step in
blood-banking.  Blood facilities should be able to follow every unit
of blood (including each component prepared from a unit) through
records obtained between screening and final transfusion or
destruction.  These steps include charting gauges in refrigerators,
freezers, and platelet incubation mechanisms and comparing their
readings to automated temperature recordings.  For example, there are
requirements that storage temperatures for source plasma be lower
than 20 degrees Celsius.  Units exposed to higher temperatures may be
issued but must be relabeled as "source plasma, salvaged."\4

Furthermore, the AABB technical manual states that when it is
necessary to destroy a product, the identification of each of the
components destroyed, the reasons for destruction, and the data and
methods of destruction must be recorded.\5

According to AABB's technical manual, blood facilities must, when
they ship units, record the name and address of the receiving
facility; the date and time of shipment; a list of all donor unit
numbers, blood types, and expiration dates; the names of all blood
components; the final inspection of whole blood or red blood cell
units; periodic tests to determine that the shipping containers have
maintained an acceptable range of storage temperatures; and the name
of the person filling the order.\6

According to federal regulations, "Distribution and receipt
procedures shall include a system by which distribution or receipt of
each unit can be readily determined to facilitate its recall, if
necessary."\7 Essentially, this means the name and address of the
facility receiving the blood products, the date and quantity
delivered, the lot number of each unit, and the date of expiration or
collection. 

Several FDA memoranda pertain to the disposition and retrieval of
units that have been tested for viral markers from donors who
subsequently tested positive or repeatedly test reactive.  Other FDA
information notes that manufacturers of plasma derivatives are
allowed to receive units of source plasma before they receive all
written test results (such as viral marker testing) if the collection
facility is owned by the manufacturer and has the same license
number.  Manufacturers that collect source leukocytes can ship them
before receiving the written infectious disease test results (because
leukocytes have a short shelf life) but they cannot use them except
in an emergency. 


--------------------
\4 A unit labeled "source plasma, salvaged" has exceeded its
expiration date or required storage temperature or has been subject
to other problems that prohibit its use in plasma pools.  Such units
can be used for research, however. 

\5 American Association of Blood Banks, Technical Manual, 11th ed.,
(Bethesda, Md.:  1993), p.  574. 

\6 American Association of Blood Banks, Technical Manual, p.  574. 

\7 21 C.F.R.  606.165(a). 


      EAR AND EIR INFORMATION
-------------------------------------------------------- Chapter 4:4.1

EARs submitted to FDA in 1994 were rarely related to storage and
distribution issues.  Errors and accidents include shipping units to
an incorrect facility, losing or failing to receive units, and
storing at incorrect temperatures.  Less than 5 percent (553 of
11,292) of all EARs submitted were in this area.  Only issues related
to collection and viral testing had fewer EARs (362 and 274,
respectively).  Table 4.7 shows that licensed facilities reported
storage and distribution EARs at a rate nearly 1,800 times higher
than that of unlicensed facilities and nearly 900 times higher than
that of plasma centers.  Their rates per 100,000 units collected were
31 and nearly 4,400 times higher, respectively.  Table 4.8 shows, in
contrast to the EAR data noted above, that a large number of
facilities for which we could determine that storage and distribution
activities were observed by an FDA inspector were found to have
storage and distribution problems during FDA inspections.  This table
also illustrates a high percentage of Form 483s related to storage
and distribution.  In fact, this area received more Form 483
observations than any other layer or process we examined. 



                               Table 4.7
                
                 Storage and Distribution EAR Rates by
                         Facility Type, 1994\a

                                           Unlicensed\
                                                    or
                                           transfusion  Plasma
Source                           Licensed    service\b  center   Total
------------------------------  ---------  -----------  ------  ------
EAR rate per facility\c              1.79        0.001   0.002    0.18
EAR rate per 100,000 units           4.37         0.14   0.001     2.1
 collected or transfused\d
----------------------------------------------------------------------
\a There were 308 licensed blood facilities, 2,274 unlicensed blood
facilities and transfusion services, and 463 plasma centers in the
United States in 1994. 

\b FDA separates error and accident reports by unlicensed blood
facilities and transfusion services in its annual summaries of EARs. 
However, these establishments submit their EARs based on a
self-designation as either an unlicensed blood facility or
transfusion service and FDA does not check the accuracy of these
self-designations.  Therefore, we combined this information in our
analysis of EARs. 

\c We calculate rate per facility by dividing the total number of
EARs by the total number of facilities. 

\d We calculate rate per 100,000 units collected by dividing the
total number of EARs by the total number of units collected. 



                                    Table 4.8
                     
                      Storage and Distribution Problems and
                     Form 483 Observations by Facility Type\a

                        Unlicensed\\  Transfusion      Plasma
             Licensed        b          service        center         Total
            ----------  ------------  ------------  ------------  --------------
Source        No.    %      No.    %      No.    %       No    %       No      %
----------  -----  ---  -------  ---  -------  ---  -------  ---  -------  -----
Facilities  21 of  55%    12 of  26%    30 of  41%    17 of  34%    80 of    38%
 with          38            47            74            50           209
 problems\
 c
Facilities  14 of   37  8 of 47   17    26 of   35    14 of   28    62 of     30
 receiving     38                          74            50           209
 483
 observati
 ons
--------------------------------------------------------------------------------
\a There were 48 licensed facilities, 114 unlicensed facilities, 91
transfusion services, and 72 plasma centers in our sample (total =
325). 

\b In our analysis of EIRs and Form 483s we separated unlicensed
blood facilities and transfusion services based on information
contained in the EIRs. 

\c There were 38 licensed facilities, 83 unlicensed facilities, 36
transfusion services, and 52 plasma centers in our sample that
contained EIR information that allowed us to determine that FDA had,
in fact, examined storage and distribution during its inspection. 
Problems were those that were characterized by the inspector on the
inspection report whereas Form 483 observations were problems deemed
serious enough to be denoted on a Form 483. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 4:4.2

One area of safety that is of concern regarding storage and
distribution is the issue of inventory management. 


         INVENTORY MANAGEMENT
------------------------------------------------------ Chapter 4:4.2.1

The data indicate that blood facilities either cannot account for or
lose a large number of donated units, units that are never
transfused.  Data from AABB, ABC, ARC, and 3,600 independent
hospitals showed that 10.5 percent of the 1989 blood supply (nearly
1.5 million units) was not transfused.  Outdated or lost units
accounted for 7 percent (994,000 units) of the total number of units
collected.  Interestingly, 3.5 percent (501,000 units) of the blood
that was not used was not accounted for in any way.\8

Although units that are unaccounted for are not related directly to
safety, they highlight the storage and distribution problems at blood
facilities. 

Our earlier discussion of autologous donations and transfusion to
unintended recipients might be relevant here if we could determine
that units that were unaccounted for were transfused to the wrong
patient.  Of course, these data cannot exist because blood facilities
cannot account for them.  However, a recent AABB survey found that 48
of 491 respondents (9.8 percent) reported that one or more units were
associated with inventory management problems, inadvertent crossover
(giving a unit of blood to an unintended recipient), improper patient
identification, or discrepancies in blood typing.  Inadequate
processes for inventory control can therefore affect blood safety. 


--------------------
\8 E.  Wallace et al., "Collection and Transfusion of Blood and Blood
Components in the United States, 1989," Transfusion, 33 (1993),
139-44.  Recent ARC data indicate that lost units comprised only
0.0028 to 0.0043 percent of produced components in the first half of
1996. 


MONITORING AND INVESTIGATING
============================================================ Chapter 5

In the fifth safety layer, FDA monitors blood facilities for
compliance with federal good manufacturing practices and blood-
banking regulations by inspecting them.  It also requires licensed
blood facilities to notify FDA of errors and accidents in the
manufacturing of biological products.  EARs provide FDA with
information on potential problems within a blood facility and give it
a means with which to begin product recall procedures.\1

The Food, Drug, and Cosmetic Act and the Public Health Service Act
authorize FDA investigators to examine all pertinent parts of a blood
facility's operations and report their findings in an EIR; they note
objectionable conditions on the Form 483.  At the close of an
inspection, the investigators present the Form 483 to the head of the
facility to ensure that management is aware of their observations. 

A licensed facility that refuses to permit such inspections or
refuses to permit access to required records can have its license
revoked.  For unlicensed facilities, refusals can result in judicial
action to close a facility. 

FDA's annual summaries of EARs suggest that unlicensed blood
facilities are underreporting their errors and accidents.  (FDA
recommends that unlicensed facilities voluntarily report EARs.) We
found direct, if unconfirmed, evidence that unlicensed facilities are
significantly less likely than licensed ones to submit an EAR even in
the most serious cases, when product recalls occur.  Also, licensed
and unlicensed facilities are not submitting timely EARs and FDA is
not timely in confirming that recalls that have been initiated by
blood facilities have actually occurred. 

We found substantial confusion in the blood industry on the
distinction between FDA regulations and guidance in terms of what
practices were actually required and what were recommended.  Its
inspection procedures also have several deficiencies.  (1) FDA
conducts no statistical analyses of the information contained in EIRs
and their corresponding Form 483 observations.  (2) While FDA's
current list of licensed blood facilities is generally reliable, some
of the list's information is inaccurate.  (3) FDA fails to inspect
some blood facilities within the time periods set by its own
guidelines.  (4) FDA's present policy on completing EIRs creates
problems for determining what blood-banking processes have actually
been inspected.  (5) There were differences across districts in Form
483 observations given by FDA inspectors.  Also, we found
inconsistencies in what was considered an action that should result
in a Form 483 observation or warning letter. 

These problems may not directly jeopardize the safety of the blood
supply.  However, without adequate monitoring of the blood industry,
FDA cannot ensure that individual facilities conform to the federal
statutes and regulations that are designed to provide safe blood to
the nation. 


--------------------
\1 A recall is a blood facility's voluntary removal or correction of
a marketed blood product that violates laws administered by FDA.  The
Public Health Service Act authorizes FDA to require that a
manufacturer initiate a recall if there is an imminent hazard to the
public health. 


   ERROR AND ACCIDENT REPORTS
---------------------------------------------------------- Chapter 5:1

FDA's regulations require all blood facilities to maintain records of
errors, accidents, transfusion reactions, complaints, investigations,
and follow-up.  Licensed facilities are required to notify FDA of
errors and accidents that affect the safety, purity, or potency of
blood products, but unlicensed ones are not.  They are asked,
however, to notify FDA voluntarily.\2

FDA's guidance on what constitutes a reportable error or accident
includes, among others, the release of blood units (1) that are
repeatedly reactive to tests, indicating hepatitis or HIV; (2) in
which testing was performed incorrectly or misinterpreted; (3) from
donors who are, or should have been, permanently or temporarily
deferred; (4) that have not been completely tested or that are
incorrectly labeled; and (5) that are contaminated because of an
error in manufacturing.  A reportable error or accident also includes
incorrectly identifying samples used in routine testing, making
errors in routine testing that result in the wrong unit's being
released for transfusion, and issuing the wrong unit for transfusion. 
Errors and accidents should always be reported promptly when a
product has been made available for distribution.\3

EARs are submitted to the Center for Biologics and Evaluation Review
(CBER is the FDA center with main responsibility for regulating blood
and blood products), and if an EAR clearly does not require further
evaluation for a product recall it remains at CBER, where it is
entered into the error and accident reporting system (EARS) database. 
If CBER decides that further evaluation is warranted, it forwards the
EAR to the appropriate district office for follow-up as a potential
recall situation.  The district office determines if the situation
does warrant a recall and makes a recommendation to the office of
compliance within CBER.  This recommendation is evaluated for
completeness and to determine if the incident meets the definition of
a recall.\4 If the incident is determined to be a recall, a health
hazard assessment is performed and classified as to the severity of
the event.  A recall is confirmed when CBER notifies the district
that a recall should occur.  In fiscal year 1994, there were 427
blood recalls involving 8,529 units of blood or plasma, or about
0.003 percent of the approximately 26 million units collected
nationally that year.\5

FDA maintains a database of EARs and compiles annual summaries that
total them and categorize them by type of facility and type of error. 
From October 1991 to September 1994, FDA received more than 30,700
EARs.  Postdonation information errors and accidents accounted for by
far the greatest number.  For example, from October 1992 to September
1994, FDA received 20,289 EARs, of which 7,379 (36 percent) reported
postdonation errors and accidents.\6

Licensed blood facilities account for the vast majority of EARs,
reporting 10,283 in fiscal year 1994 while unlicensed facilities
reported 146.  (In April 1995, there were 739 licensed and 2,241
unlicensed blood facilities in the United States). 

Most EARs are not serious problems and do not represent immediate
danger.  In fact, EARs are an integral part of a system for catching
potentially dangerous units of blood before they enter the blood
supply.  For instance, when postdonation information from a donor
alerts a blood facility that a unit of blood should not be
transfused, the facility customarily reports this information as an
error or accident because of the way in which FDA has defined what is
to be reported through EARs.  In such cases, the layers of safety are
working effectively to protect the blood supply. 

Furthermore, few errors and accidents are egregious.  For example,
only 66 of the more than 10,000 fiscal year 1994 EARs were submitted
for HIV-1 and HIV-2 testing that resulted from incorrect testing,
misinterpretation, or product release prior to testing or before
testing was completed (0.006 percent).  Only 12 EARs reported a
failure to quarantine a product that was HIV reactive (0.001
percent).  In other words, HIV errors represent approximately 1 out
of every 307,692 blood donations. 


--------------------
\2 FDA is reviewing a proposed rule that would require unlicensed,
registered firms to submit error and accident report. 

\3 EARs are not required when a facility detects an error or accident
before a blood product has been made available for distribution. 

\4 FDA may request a firm to initiate a recall when it is determined
that a product has been distributed that presents a risk of illness
or injury or gross consumer deception, a firm has not initiated a
recall of the product, or the agency action is necessary to protect
the public health and welfare. 

\5 These figures do not include the recall of products used to
process blood, such as defective collection bags, nor does it include
any lots of intravenous immune globulin manufactured after February
1993--a plasma derivative recalled for potential transmission of
hepatitis C. 

\6 FDA did not use postdonation information as a category in its
fiscal year 1991 summary, so our numbers are based on 1992-94 data. 
In fiscal year 1991, FDA received 3,834 EARs; in 1992, more than
10,000.  The increase stemmed partly from the implementation of the
December 5, 1990, memorandum entitled "Revised Recommendations for
the Prevention of Human Immunodeficiency Virus (HIV) Transmission by
Blood and Blood Products" that recommended direct questioning about
high-risk behavior and the March 20, 1991, memorandum entitled
"Responsibilities of Blood Establishments Related to Errors and
Accidents in the Manufacture of Blood and Blood Components" regarding
the reporting of errors and accidents to FDA.  We confined our
analysis of EAR data to fiscal year 1994. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 5:1.1

The three issues related to errors and accidents that do merit
attention are that unlicensed blood facilities appear to underreport
them to FDA, many EARs are submitted to FDA long after the problem
has occurred, and FDA is not promptly investigating EARs that result
in product recalls. 


         UNDERREPORTED EARS
------------------------------------------------------ Chapter 5:1.1.1

Although there are more than three times as many unlicensed blood
facilities as licensed ones, the former account for only 1.3 percent
of reported EARs (146 of 11,298) whereas the latter (including ARC)
account for 91 percent of reported EARs (10,283 of 11,298).\7 If EARs
were related more to the number of units collected than to the number
of facilities, we might expect unlicensed facilities to report 10
percent of all EARs because they collect about 10 percent of the
nation's blood supply; this is still much higher than their current
proportion of EARs.  Similarly, plasma facilities collect 12 million
units of plasma, which is equal to the total number of whole blood
units collected by licensed and unlicensed blood facilities together,
yet plasma facilities report less than one tenth of all EARs. 

An additional cause for concern is that EARs from unlicensed
facilities are just as likely as EARs from licensed ones to result in
a potential recall (see table 5.1).  Thus, the failure to require
unlicensed facilities to report errors and accidents may result in
FDA's missing a number of potential product recall problems. 
Potential product recalls for plasma centers made up 39 percent of
all EARs that they submitted in fiscal year 1994. 



                               Table 5.1
                
                 Potential Recalls from Reported EARs,
                            Fiscal Year 1994

                                                               Percent
                                       Reports   Potential   recall to
Type of facility                      received     recalls     reports
----------------------------------  ----------  ----------  ----------
Licensed                                10,283         512          5%
Unlicensed                                 146          10           7
Plasma center                              856         333          39
Transfusion service                          7           0           0
----------------------------------------------------------------------
The commissioner of FDA in 1993 noted in testimony before the House
Subcommittee on Oversight and Investigations that the issue should be
looked at as FDA revises its error and accident reporting
procedures.\8 A May 1995 HHS Inspector General's report noted that
voluntary reporting by unlicensed blood facilities is a major
shortcoming in FDA's notification process and recommended that they
be required to submit EARs to FDA.\9

Unlicensed facilities underreport errors that end in product recalls. 
In 299 of the 468 recalls in 1994, an EAR was submitted before the
district office's recommendation for recall:  293 from licensed
facilities, including plasma centers, and 6 from unlicensed ones.\10
Our statistical analysis of this difference determined that it was
highly significant (t = -8.96; p < .0001).  More than 70 percent of
licensed facilities submitted an EAR before recall, but only 17
percent of unlicensed facilities did this.  Given that EARs are one
way of alerting FDA of the need for an immediate recall, we believe
that the underreporting by unlicensed facilities is a serious
problem. 


--------------------
\7 The remaining 7.8 percent of EARs (rounded) are reported by plasma
centers, vaccine manufacturers, and reagent manufacturers.  Our
interviews with representatives of licensed blood facilities revealed
that unlicensed blood facilities may have a competitive edge because
they are often not held to the same standards.  For example,
unlicensed blood facilities do not have to obtain FDA approval for
certain changes in their procedures, which, it is alleged, add costs
in personnel, salary, and time to licensed facilities that are not
borne by unlicensed ones. 

\8 David Kessler, Commissioner, Food and Drug Administration, "Blood
Supply Safety," Subcommittee on Oversight and Investigations,
Committee on Energy and Commerce, House of Representatives, July 28,
1993, p.  50. 

\9 FDA officials told us that FDA agreed with the HHS Inspector
General's report and that it is preparing a proposed rule that would
require unlicensed blood facilities to submit EARs. 

\10 Recalls do not always begin with an EAR.  In some cases, an FDA
inspection uncovers an error or accident that was not reported to FDA
and bases a recall recommendation on its severity.  Some facilities
then submit an EAR even though recall has begun.  We did not include
these cases in our analysis. 


         UNTIMELY EARS
------------------------------------------------------ Chapter 5:1.1.2

The HHS Inspector General's report noted that, for a random sample of
163 EARs from October 1992 to April 1993, the time between the date
when a blood facility detected an error or accident and the date when
it was reported to FDA ranged from less than 1 month to more than 1
year, the average being a little over 4 months.  The report also
found that about 14 percent of the sampled EARs were submitted within
1 month but that 13 percent were reported 6 months or more after the
error was detected.  This untimeliness may hamper FDA's ability to
investigate errors and accidents and to monitor blood facility
practices. 


         UNTIMELY FDA
         INVESTIGATION OF EARS
------------------------------------------------------ Chapter 5:1.1.3

Once a facility has reported an error or accident to CBER, depending
on the severity of the error or accident, the district office
evaluates it and may recommend a recall.  Our analysis of FDA's
recall database outlined in figure 5.1 shows that in 60.3 percent of
those cases, 7 months or more elapsed between the time an EAR was
submitted and the district recommended a recall to CBER.\11 The time
for FDA review (the time from a recommendation for a recall and when
a recall is confirmed) ranged from none to a year, with a mean of 9
weeks. 

   Figure 5.1:  Time From Error
   and Accident Detection to EAR
   to Recall Recommendation to
   Recall Confirmation

   (See figure in printed
   edition.)

Note:  Numbers may not sum to 100 percent because of rounding. 

Figure 5.1 also shows that in more than 70 percent of the cases, the
total time from EAR submission to when a recall is confirmed and
publicly announced is 7 months or more.  The total time ranged from a
little over a month to 2-1/2 years, with an average of nearly 9-1/2
months.  According to FDA, in about 25 percent of cases, a product
recall is not initiated by the facility by the time FDA recommends a
recall.  It is these cases that could compromise blood product safety
given the long time FDA takes to go through its formal recall
process. 

We also found no significant differences in the time it took for a
product recall to go through the process above, based on the severity
of the case.\12 That is, more serious cases were not processed faster
than less serious ones.  Since some of the products that are recalled
have been made available for transfusion, it is important that this
process be as timely as possible. 


--------------------
\11 In many cases, a recall has been initiated by a blood facility
before an EAR is submitted to FDA.  However, the time lag from the
submission of an EAR to when FDA completes its evaluation can be
lengthy. 

\12 If an incident is determined to be a recall, a health hazard
assessment is performed and classified as to the severity of the
case. 


   FDA'S REGULATIONS AND GUIDANCE
---------------------------------------------------------- Chapter 5:2

FDA communicates its requirements through CFR, title 21, and its
policies and recommendations through memoranda and letters,
compliance manuals and program, compliance policy guides, and a guide
for blood facility inspections.  The requirements in the Public
Health Service Act, Food, Drug, and Cosmetic Act and CFR are the only
mandatory requirements. 

According to FDA, inspectors do not cite relevant CFR provisions on
Form 483s when they find objectionable conditions because numerous
regulations may apply to any given situation.  However, FDA
inspectors are supposed to present their findings to the blood
facility immediately after an inspection, including any Form 483
observations.  After the inspection, and to ensure that inspectors
consider all relevant regulations in an investigation, other FDA
officials review EIRs and any Form 483 observations. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 5:2.1

Below we describe the one safety issue we found in regard to FDA's
use of regulations and guidance. 


         USE OF GUIDELINES AND
         RECOMMENDATIONS
------------------------------------------------------ Chapter 5:2.1.1

We found substantial confusion in the industry on the distinction
between FDA regulations and guidance, potentially leading to
different interpretations and applications of FDA's requirements and
recommendations.  Many of our survey respondents were unclear as to
which statements had to be followed and which were only FDA
recommendations.  Twenty-nine of the 45 full-service licensed
facilities we surveyed responded to an open-ended question on
possible areas for improvement within the blood industry:  10 (or 34
percent) of them answered that FDA's regulations and guidance are
ambiguous.  They noted that recommendations were sometimes used as
the basis for Form 483 observations, that the regulations should be
updated to incorporate current memoranda, and that the language in
the memoranda should be clarified as to whether actions to be taken
are required or recommended. 

An Institute of Medicine study on blood safety issues has recommended
that "when issuing instructions to regulated entities, FDA should
specify clearly whether it is demanding specific compliance with
legal requirements or is merely providing advice for careful
consideration." Responding to this study, AABB agreed that many
recommendations and guidance memoranda are often not clear as to
regulatory intent and even when ambiguities have been identified AABB
has also stated that they have not been successful in obtaining
clarification from FDA.\13

The issue has practical implications.  For instance, although FDA has
issued memoranda on procedures for HTLV testing, the regulations do
not refer to HTLV testing.  Thus, one could view this as only a
recommendation and not a requirement.  However, not testing for HTLV
would probably affect the purity, potency, and safety of blood
products, and a facility that failed to test for HTLV could be
considered in violation of the statutory legal standards, which
explicitly state that blood products are to be tested for purity,
potency, and safety, regardless of whether the regulations formally
require such specific tests. 

Our survey respondents indicated two other areas in which
improvements would enhance blood safety:  consistent regulation
between licensed and unlicensed blood facilities and better
regulation of transfusion procedures. 

FDA has to its credit historically issued memoranda to give the
industry immediate feedback on its position on new issues.  This is
an important tool for quickly reacting to advances in medical
knowledge or technology.  However, guidelines and memoranda that have
been issued for expediently stating expectations to the blood
industry appear to move rarely into the formal regulatory process. 
For example, FDA has not codified requirements for testing blood for
either HCV or HTLV, even though testing for them clearly affects
safety and even though FDA has recommended testing since 1988 for
HTLV and 1990 for HCV.  Only regulations codified in the Code of
Federal Regulations benefit from formal public comment, and issuing
statements through the CFRs is one of the only ways to clarify FDA's
purpose. 

Blood facilities often adopt FDA recommendations and integrate them
into their standard operating procedures (SOPs).  Once these
recommendations are incorporated into SOPs, the blood facility can
receive Form 483 observations for not following its SOPs under good
manufacturing practices.  This, however, does not overcome the
problem of required practices and the issue of public comment
opportunities. 


--------------------
\13 Letter to Assistant Secretary of Health, Department of Health and
Human Services, Regarding the Task Force to Review Current Blood
Safety Program, Washington, D.C., October 3, 1995. 


   INSPECTIONS
---------------------------------------------------------- Chapter 5:3

FDA is required to perform biennial inspections.  Facilities that
have received warning letters or that have been found deficient in
inspections within the past 2 years may be inspected annually until
two consecutive inspections pass without significant observations.\14

Inspectors are FDA officers who have "special knowledge of the
methods used in the manufacture and control of products." Their job
is to, among other things,

     "investigate .  .  .  the methods of propagation, processing,
     testing, storing, dispensing, recording, or other details of
     manufacture and distribution of each licensed product, or
     product for which a license has been requested, including
     observation of these procedures in actual operation .  .  .  . 
     "\15

Suspension or revocation of licenses, injunctions, and prosecutions
may ultimately result from a process begun with an inspector's Form
483 observations of a continuing pattern of deviation.  For isolated
deviations, FDA acts only when they may jeopardize the safety of
donors or products.  While FDA views the Form 483 as an observation,
the blood industry often sees it as a citation or violation of
applicable FDA regulations and guidance. 

Currently, FDA uses three levels for classifying inspections; no
action indicated (NAI) for insignificant deviations or no identified
deviations, voluntary action indicated (VAI) for deviations that are
amenable to corrective action by the firm with no compromise to
public safety, and official action indicated (OAI) for deviations of
a serious nature that require some FDA intervention to ensure that
corrections are made.  FDA inspectors are directed to list on the EIR
the specific areas covered only when a limited or incomplete
inspection is done.  The inspectors are also instructed to list on
the EIR everything they see that is questionable and that could
therefore be a violation of the regulations. 

Most inspections are conducted in accordance with specific compliance
manuals that explicitly state what is to be observed during the
inspection.  FDA inspectors are also directed to list in the EIR the
specific compliance program under which the inspection is performed,
and they are not expected to suggest remedies to problems that are
found during an inspection, nor are inspectors expected to discuss
the regulations that pertain to the problems.  By listing the
compliance program, FDA officials told us, all directions included in
the compliance program were followed unless otherwise stated on the
EIR.  Further, FDA officials stated that they often have substantial
experience with each blood facility, allowing inspections to be
tailored to look at areas known to be sources of problems, thus
making maximum use of FDA's limited resources. 

To examine EIRs, we randomly sampled 8 district offices.  Within
these, we selected a representative sample of 373 blood facilities,
including licensed and unlicensed blood facilities, blood donor
centers, plasma fractionators, plasma collection centers, testing
laboratories, transfusion services, and viral testing and reagent
manufacturers.  We looked at their last recorded inspections,
separating EIRs into those that should have contained a blood
facility inspection checklist and those that did not require one.\16
We also mailed a questionnaire to the 45 full-service blood
facilities within our representative sample.  (See appendix III.)


--------------------
\14 Performing yearly inspections of firms previously in violation is
FDA's own requirement.  FDA's inspectors work in 21 district offices
in six regions:  Pacific, Southwest, Midwest, Northeast,
Mid-Atlantic, and Southeast. 

\15 21 C.F.R.  600.22(d). 

\16 Until October 1994, FDA inspectors were required to fill out an
inspection checklist that outlined all the areas of blood-banking
that an inspector could examine.  After October 1994, FDA adopted a
"systems approach":  the checklist is no longer required and
inspectors examine blood-banking processes with a view to
establishing that systems adequately address quality-assurance and
good manufacturing practices concerns. 


      SAFETY ISSUES
-------------------------------------------------------- Chapter 5:3.1

We found several problems in FDA's inspection process in five broad
categories:  the use of EIR information, the tracking of blood
facilities, the timing of inspections, the completeness of inspection
reports, and the consistency of inspection reporting. 


         USE OF EIR INFORMATION
------------------------------------------------------ Chapter 5:3.1.1

We were told by FDA that it analyzes EIRs and Form 483s.  According
to FDA, examples of such analysis were the program-oriented data
system (PODS) database; the 1992-93 task force on ARC, which
categorized all Form 483s issued to ARC from 1988 to 1992; work
performed by FDA that led to injunctions against ARC and BSI; and a
study FDA conducted on Form 483 observations.\17

Information FDA provided to us on PODs contained no information that
would allow FDA to perform systematic analyses of EIRs and Form 483s. 
PODs contains information on who did the inspection, where the
inspection occurred, how long the inspection took, what was covered
in the inspection, and the results of the operation.  However, what
was covered merely identifies products involved in the inspection
(for example, food) while the results simply identify whether the
firm is operating in or out of compliance. 

Furthermore, FDA noted that PODs is in place to provide information
on accomplishments by FDA field personnel to justify annual budget
requests.  It is, therefore, not a system that contains information
that would allow for a statistical analysis of blood facility EIRs
and Form 483s.  Likewise, the 1992-93 task force work is not an
analysis of EIRs and Form 483s.  It is a listing of Form 483s given
to ARC facilities from 1988 to 1992 by category (for example, donor
screening, testing, labeling, equipment).  No statistical analysis of
this list was performed.\18

In sum, without collating, synthesizing, analyzing, and evaluating
EIR and Form 483 information, FDA has no means of assessing overall
national compliance, assessing trends by type of blood facility,
identifying the problems of different types of blood facilities, or
evaluating the effect of policy changes on compliance rates. 

By performing these types of statistical analyses, FDA could obtain
information on different rates of Form 483 observations between
district offices, rates of observations by type of activity (for
example, donor screening, donor deferral, viral testing), and
differing rates between types of facilities.  For example, our
analysis of Form 483 observations found differences in the number and
kind of Form 483 observation given by different FDA districts. 
Although the reasons for these differences are unclear, such
information could provide FDA with important data on inspection
findings and FDA procedures for carrying out inspections. 


--------------------
\17 See the last section of this chapter, on disparities in
inspection reporting, for information pertaining to the study
conducted by FDA on Form 483 observations. 

\18 FDA also summarizes ARC's progress under the terms of a May 12,
1993 consent decree.  That is, FDA inspectors give ARC annual reports
of Form 483 observations.  Similar to the 1992-93 task force work,
these reports are listings of Form 483 observations given under
topical headings such as management control, quality assurance, and
records management.  No statistical analyses are performed on these
data. 


         TRACKING BLOOD FACILITIES
------------------------------------------------------ Chapter 5:3.1.2

FDA maintains a list of all registered blood facilities with their
registration numbers.\19 The vast majority of those that were in our
sample were accurately identified.  However, we did find problems
with FDA's list of registered blood facilities.  For example, when we
queried FDA about the EIR for a particular blood facility through its
registration number, FDA told us erroneously that the registration
number in question belonged to a different facility (this was based
on its list of registered blood facilities). 

We also found a small number of cases in which the last inspection of
a blood facility was held more than a decade ago but it was still on
the FDA list of active registered blood facilities.  In these cases,
it appeared that these facilities had closed and were not operating
as blood facilities, but the fact that they still had registration
numbers and were on FDA's active list highlights inadequacies in
FDA's recordkeeping.  We also found that FDA could not find 4 EIRs (1
percent of the 373 EIRs in our sample).\20 Unfortunately, we cannot
know the extent of such monitoring problems or their potential effect
on FDA's oversight responsibilities. 


--------------------
\19 All blood facilities are registered with FDA and are given a
unique registration number.  This is distinct from a license number
given to facilities that engage in the sale, barter, or exchange of
blood products across state lines. 

\20 We were able to analyze data on the tracking and timing of
inspections for all 373 blood facilities in our sample.  The EIR
information below was based on the 325 blood facilities in our sample
that were licensed and unlicensed blood facilities, transfusion
services, and plasma centers.  The 48 other facilities were plasma
brokers, viral testing and reagent manufacturers, testing
laboratories, and depot sites or had been inspected for specific
purposes that were not part of an annual inspection and thus we did
not include them in the analyses below. 


         TIMING OF INSPECTIONS
------------------------------------------------------ Chapter 5:3.1.3

Of the 373 blood facilities in our sample, 45 (12 percent) had not
been inspected in more than 2 years.\21 One donor center had not been
inspected in more than 3-1/2 years.  Since our sample represents all
blood facilities in the nation, 348 of the 2,900 registered blood
facilities may not have been inspected within the past 2 years. 


--------------------
\21 FDA's response to our query for a list of the blood facilities in
our survey was dated August 14, 1995.  Thus, 45 facilities had not
been inspected since September 1993. 


         COMPLETENESS OF
         INSPECTION REPORTS
------------------------------------------------------ Chapter 5:3.1.4

We examined each facility in our sample for whether the EIR indicated
that a particular function had been examined.  For the purpose of our
analysis, if it was mentioned at all in the EIR, we considered it to
have been examined.  If it was not mentioned at any time in the EIR,
we considered that one could not determine whether the area had been
examined.  We excluded functions that inspectors noted were not
performed. 

For the time period when checklists were required, we found that many
blood inspection checklists were not completed.  Forty of 224
inspections (18 percent) in our sample that should have included an
inspection checklist did not have one.  We found that the lack of a
completed checklist made it very difficult to determine what areas of
a blood facility's processes were actually covered during an
inspection.  Many of the EIRs for which the checklist was missing
also lacked narratives from which to obtain the pertinent
information.  Thus, we often could not determine whether the FDA
inspectors based their findings on an observation of certain
blood-banking operations or on an examination of written standards of
operation. 

In many instances, we were unable to determine whether procedures
relating to donor screening, deferral, collection, routine testing,
viral testing, postdonation information, labeling, quarantining,
storage, and "machine" issues were examined at all in individual
inspections.  In fact, for all the matters in our EIR analysis that
FDA could have inspected, we could not find coverage in 33 percent
(963 of 2,957).\22 Further, we were able to determine in only half of
all reviewed reports that inspections covered all activities
necessary to ensure compliance.  Thus, regardless of FDA's policy on
what information should be contained on an EIR, we could not
determine what had actually been observed and what practices had been
examined only by reviewing SOPs.  As table 5.2 indicates, there were
many instances in which a given process was not mentioned at all in
the EIR. 



                               Table 5.2
                
                 Percentage of Facilities for Which We
                  Could Determine that Processes Were
                     Checked in Inspection Reports

                              Blood bank
                    ------------------------------

                                                    Plasma  Transfusio
Process             Licensed            Unlicensed  center   n service
------------------  ------------------  ----------  ------  ----------
Screening           83%                        77%     74%         64%
Deferral            89                          64      70          63
Collection          83                          85      73          71
Routine testing     56                          50      \a          72
Viral testing       95                          67      38          42
Labeling            73                          53      56          71
Postdonation        61                          69      43          62
 information
Quarantine          67                          53      43          70
Storage             83                          63      70          82
Machines            96                          79      73          88
Miscellaneous       83                          62      58          81
----------------------------------------------------------------------
\a Does not apply. 

As noted previously, FDA's policy is for the inspectors only to list
areas on the EIR that were not covered.  Thus, when an inspector
notes on the EIR the specific compliance program under which the
inspection is taking place, this means that all blood-banking
practices covered in the compliance program have been examined
(unless specifically listed on the EIR).  However, we found that this
policy is unreliable in ensuring that activities not covered during
an inspection are, in fact, listed on the EIR.  For example, at a
blood facility inspected in 1994, an inspector found that no lookback
procedures had been followed in several cases of reported
HIV-positive donors identified since 1990. 

When we examined the EIR for this facility for the inspection that
took place in 1993, we found no mention that lookback procedures were
not being followed.  This means either that the 1993 inspection
examined lookback procedures and did not find the problem that had
been evident since 1992 (according to the 1994 inspection) or that
the activity was not observed in the 1993 inspection and was not
listed on the EIR according to FDA's own stated policy.  In either
case, FDA's policy of not listing all activities covered during an
inspection results in the agency's inability to determine what
practices have actually been examined by its inspectors and hampers
its ability to perform any meaningful analysis of EIRs and Form 483s. 
Without knowing what has been inspected, FDA cannot know where a
facility is in or out of compliance. 

FDA officials told us that they have substantial previous experience
with each facility, allowing them to tailor inspections to look at
areas known to be sources of problems, thus making maximum use of
limited resources.  However, because EIRs do not list the activities
covered in the previous inspection (and, as noted above, such a
policy may not, in fact ensure that some practices are examined even
though they were not listed on the EIR), such tailoring of
inspections may result in blood-banking practices not being examined
for long periods of time at individual facilities.  Additionally,
even if FDA emphasizes certain areas more than others based on
previous "experience," this could result in missing problems in areas
that had previously not been out of compliance.  Because of all these
problems relating to information contained on the EIRs, we limited
our analysis of possible compliance problems to those listed on the
Form 483. 

We also found that facilities whose EIRs did not have a checklist,
whether one was required or not, were significantly less likely to
have Form 483 observations than facilities that had checklists.\23
This could mean that checklists promote a more methodical approach to
an inspection, resulting in more Form 483 observations, or that
formal procedures such as the completion of a checklist focus an
inspection on minor details that may or may not be real problems.  As
we discuss below, this finding may be a result of a lack of clear and
concise FDA guidance on what should constitute a Form 483
observation. 

In order to focus a current inspection clearly, FDA inspectors are
expected to review past EIRs for previously identified problems. 
Without a checklist or more comprehensive narrative in the EIRs, we
often could not obtain such information.  Table 5.3 presents the
results of a survey question in which we asked facilities to what
extent FDA examined standard operating procedures in 12 separate
areas.\24 In every area except deferral, more than half the
respondents indicated that FDA examined standard operating procedures
only to a moderate extent or less. 



                                    Table 5.3
                     
                     The Extent to Which Inspectors Examined
                         Standard Operating Procedures\a

                                                                Very
            Little or    Some       Moderate    Great          great    Does not
Area        no extent   extent        extent    extent        extent       apply
---------  ----------  ---------  ----------  ----------  ----------  ----------
Screening         11%     16%            22%     33%             18%
Donor              11     11              22      40              16
 history
 and
 examinat
 ion
Deferral            7      9              22      29              33
Collectio          11     11              31      36              11
 n and
 phlebeto
 my
Routine            16      7              40      22              16
 laborato
 ry
Viral              11      4              27      27              31
 laborato
 ry
Labeling           16     13              33      20              18
Postdonat          22     13              24      18              22
 ion,
 recall,
 and
 lookback
Quarantin          11     13              18      27              31
 e and
 storage
Product             7     18              18      29              29
 disposit
 ion
Quality            18     18              22      29              13
 assurance
 and good
 manufact
 uring
 practice
 s
Computer           27     13              13      9               18         20%
 validati
 on
--------------------------------------------------------------------------------
\a Row totals may not sum to 100 percent because of rounding. 

Similarly, the respondents reported that FDA does not observe or
otherwise examine firsthand major activities in the many areas listed
in table 5.4.  More than 20 percent of our respondents reported that
FDA does little or no observation in six different areas. 



                                    Table 5.4
                     
                     The Extent to Which Inspectors Directly
                           Observed Major Activities\a

                                                                Very
            Little or    Some       Moderate    Great          great    Does not
Area        no extent   extent        extent    extent        extent       apply
---------  ----------  ---------  ----------  ----------  ----------  ----------
Screening         13%     20%            20%     22%             24%
Donor               9     20              22      27              22
 history
 and
 examinat
 ion
Deferral           25     30               9      16              21
Collectio          11     20              18      24              27
 n and
 phlebeto
 my
Routine            20     29              20      16              13          2%
 laborato
 ry
Viral              11     20              22      20              24           2
 laborato
 ry
Labeling           22     27              18      18              16
Postdonat          32     30               5      14              14           7
 ion,
 recall,
 and
 lookback
Quarantin          18     18              27      13              24
 e and
 storage
Product            18     18              23      11              30
 disposit
 ion
Quality            23     21              25      18               9           5
 assurance
 and good
 manufact
 uring
 practice
 s
Computer           32     16               5      9               11          27
 validati
 on
--------------------------------------------------------------------------------
\a Row totals may not sum to 100 percent because of rounding. 

Furthermore, 35 percent of the respondents indicated that FDA
evaluated the existence and suitability of only half or fewer of the
critical control points their institutions had in place to ensure
safety, purity, and potency.  Among the facilities in which FDA found
a problem, 56 percent reported that FDA did little more than identify
that a problem existed.  According to FDA, inspectors are not to
suggest solutions or discuss the regulations or guidance that
pertains to problems found during an inspection.  However,
contradicting this position is other information provided by FDA in
which it has noted that "investigators provide general guidance on
applicable documents, policy, regulations, etc.  which are the basis
for the objectionable condition."

We also presented respondents with a list of areas that might be
examined to assess compliance and asked them to order the list in
terms of the emphasis that inspection teams gave to each area during
the last inspection.  Their ordering shows that inspectors focus on
documentation and whether records and files can be traced as well as
on adherence and completeness of standard operating procedures.  They
indicated that quality-control management is not a major focus of
inspections.  Their ordering of areas was

1.  documentation of records and files;

2.  adherence to standard operating procedures;

3.  traceability of records and files;

4.  completeness of standard operating procedures;

5.  quality-control management and accountability;

6.  employee training;

7.  software technology;

8.  hardware technology;

9.  physical plant and facilities. 

About two thirds of the respondents had received a Form 483 or other
form of observation or citation.  Seventy percent of these indicated
that the inspection team was able to articulate the significance of
the violations it had identified, but 22 percent indicated that the
inspection team was able to do so only to some extent or less.  Also,
nearly 30 percent of the respondents reported that one or more of the
items on their Form 483 were for problems that they had already
identified through their own quality-control process and had already
corrected before the beginning of the inspection. 

To FDA's credit, most respondents thought the FDA inspectors were
generally knowledgeable or very knowledgeable about blood-banking
terminology, technology, and practices.  All respondents to the
survey noted that FDA inspectors appeared to follow a systematic
approach.  Sixty-four percent also noted that most or all critical
control points were evaluated. 

Just as FDA expects blood facilities to have complete records of
their processes and activities between inspections, it is appropriate
that FDA have complete information on blood banking operations for
every blood facility inspection.  Without such information, it is
impossible to know if, in fact, blood facilities are in compliance
with all federal rules and regulations. 


--------------------
\22 As we noted previously, our EIR sample was based on 325 blood
facilities.  We categorized blood-banking processes into 11 subjects,
or a total of 3,575 potential areas that FDA should have inspected. 
However, many blood facilities did not perform all the operations we
categorized, so that those we could analyze numbered 2,957. 

\23 t(372)=-2.67, p<.01

\24 The survey asked respondents to report whether the FDA inspection
team examined standard operating procedures and whether the team
actually observed or examined firsthand 12 major blood-banking
operations:  donor screening, donor history and examination,
phlebotomy and collection, routine laboratory procedures, viral
laboratory procedures, donor deferral, labeling, quarantine and
storage, product disposition, postdonation recall and lookback,
computer validation, and quality assurance and good manufacturing
practices. 


         DISPARITIES IN INSPECTION
         REPORTING
------------------------------------------------------ Chapter 5:3.1.5

Across the 8 FDA districts that we examined, we found disparities in
the information on Form 483s and the issuance of warning letters. 
For example, more than 27 percent of the Form 483 observations in one
district were related to storage issues but only 13 percent in
another.  Similarly, more than 21 percent of one district's Form 483
observations were related to labeling issues but only about 2 percent
in another district.  Table 5.5 outlines the variations across
districts. 



                                              Table 5.5
                               
                                Percentage District Variation in Form
                                           483 Observations

                                       Region
                  District   District       3   District   District   District   District   District
                         1          2    (n =          4          5          6          7          8
Area              (n = 35)   (n = 38)     40)   (n = 41)   (n = 33)   (n = 46)   (n = 44)   (n = 48)
---------------  ---------  ---------  ------  ---------  ---------  ---------  ---------  ---------
Screening             7.8%      12.8%   18.2%      16.2%       8.7%      10.3%      13.1%      14.6%
Deferral               7.8       14.9    13.6        8.1       13.0        6.9        9.8       17.1
Collection            15.7       14.9    15.9       13.5       13.0       10.3        9.8       19.5
Routine testing        3.9          0       0        2.7          0          0        3.3        2.4
Viral testing          9.8        8.5     2.3          0        8.7       10.3        6.6        4.9
Labeling               5.9       10.6     6.8        5.4       21.7       10.3        8.2        2.4
Postdonation           2.0        4.3       0        5.4          0        3.5        3.3        2.4
 information
Quarantine             5.9          0       0          0          0          0        1.6          0
Storage               21.6       12.8    15.9       27.0       17.4       20.7       16.4       22.0
Machines               9.8       10.6    22.7       18.9       17.4       24.1       23.0        9.8
Miscellaneous          9.8       10.6     4.6        2.7          0        3.5        4.9        4.9
----------------------------------------------------------------------------------------------------
We found statistically significant differences between districts in
blood facilities' receipt of Form 483 observations.  For example,
blood facilities in district 6 received significantly fewer
observations than those in districts 1-3, 7, and 8 (see table 5.6). 



                               Table 5.6
                
                Blood Facilities That Received Form 483
                     Observations in Districts 1-8

                                                 Received observation
                                                ----------------------
                                     Number in
District                          EIR analysis      Number     Percent
--------------------------------  ------------  ----------  ----------
1                                           35          17       48.6%
2                                           38          16        42.1
3                                           40          17        42.5
4                                           41          15        36.6
5                                           33          11        33.3
6                                           46           9        19.6
7                                           44          23        52.3
8                                           48          22        45.8
----------------------------------------------------------------------
We also found disparities in the types of activities that warrant
Form 483 observations.  Why observations are issued inconsistently is
not clear.  It could be that different districts have different
problems or that different inspectors and supervisors interpret the
guidelines differently.  FDA officials believe that different
districts do, in fact, have different problems.  However, they were
not able to document for us the information on which they base this
claim. 

While some activities cited on the Form 483 appeared to be only
tangentially related to the safety, purity, or potency of a product,
other activities were not cited even though they clearly had the
potential to affect safety, purity, or potency.  For example, one
blood facility was cited because its records did not reflect a
machine weld alignment inspection, but another facility was not cited
even though the FDA inspector found one donor who had mental
retardation and did not understand several donor-screening questions
on Chagas' disease, malaria, syphilis, or yellow jaundice (a possible
symptom of hepatitis).  This donor also told the FDA inspector that
she was incapable of filling out the donation record and that the
screener at the blood bank filled out all the information for her. 

In 1996, FDA conducted a study of Form 483 observations in order to
assist in providing clearer guidance in terms of the significance,
content, and format of observations.\25 The study's conclusions were
that the majority of Form 483 observations were valid; however,
complete assessments could not be made outside the context of the
EIR.  The panel that conducted the study determined that the most
appropriate manner in which to use these conclusions would be to
develop a specific section for writing Form 483s in the blood bank
training courses provided to blood bank inspectors.  That FDA
conducted this study suggests that it is aware of problems in Form
483 consistency and its conclusion about the need for additional
training supports this viewpoint. 

FDA also issues warning letters inconsistently.  For example, one
blood facility received a warning letter detailing several instances
in which it had no written procedures for several processes such as
determining donor suitability and preparing packed red cells. 
However, another blood facility that did not receive a warning letter
knew that some of its blood units had tested positive for syphilis
but was shipping them for further manufacture without labeling them
positive for syphilis.\26

In another case, a blood facility was given 33 Form 483 observations
that included problems in transfusion of three HTLV-I positive units
of blood to three different patients, transfusion to a patient of an
initially reactive HBsAg unit of blood that was not retested in
duplicate, failure to file EARs, and "numerous donor deferral
deviations, donor reentry deviations, computer entry deviations, lack
of internal error and accident investigations, and lack of written
SOPs." However, this facility was not given a warning letter for the
lack of written procedures (as well as the many other observations)
while the blood facility noted above was given a letter for its lack
of written procedures for donor suitability and preparing packed red
cells. 

Our survey respondents raised several issues that affect the
consistency of inspections.  Twenty-seven percent reported that they
do not know what to expect from one inspection to the next; what one
inspector finds acceptable another considers an observable event. 
And while respondents reported that their current inspection team was
generally knowledgeable, 45 percent reported a wide variation in
inspectors' knowledge and training in blood-banking terminology and
procedures. 

When we asked FDA about its inspector training programs and policies,
the agency reported that field investigators undergo a series of
formal training courses and receive on-the-job training in all
product and program areas.  Its investigators are therefore regarded
as generalists, particularly those with experience and advanced
training.  By the time investigators are assigned to conduct
inspections, they have mastered basic inspection techniques and have
had ample experience.  While FDA uses the more experienced
investigators for inspections as much as possible, the less
experienced investigators do inspect facilities, and the agency has
no readily accessible way of determining the frequency with which
this occurs. 

Such inconsistency in inspection activity has ramifications for FDA's
ability to determine whether a blood facility is, in fact, in
compliance with FDA rules and regulations.  FDA expects blood
facilities to have consistent practices that follow blood facility
standard operating procedures and FDA guidelines.  It is equally
appropriate for FDA to make sure that inspections demonstrate
consistent enforcement of FDA rules and guidelines as reflected in
Form 483 observations and warning letters. 


--------------------
\25 The study was conducted by a panel of regional and national
biologic expert investigators. 

\26 This facility had interpreted FDA's memorandum on donor deferral
and product distribution relating to syphilis testing as not
requiring such labeling because the memorandum reads "the regulations
do not require the labeling of each unit with the screening tests
results." It interpreted this memorandum as stating that source
plasma could be used for further manufacture before test results were
available because the memorandum reads "source plasma collected
before serologic test results are received may be used for further
manufacture." FDA, in contrast, noted that the memorandum was
intended to convey that once a plasma-collection facility had become
aware of a donor's positive results for syphilis, all units collected
from that donor and held for shipment would have to be labeled as
reactive.  It appeared that the facility read the memorandum as
meaning that as long as the blood was collected before the test
results were completed, it did not have to label the products,
regardless of the test results. 


SUMMARY, RECOMMENDATIONS, AND
AGENCY COMMENTS AND OUR RESPONSE
============================================================ Chapter 6

We have highlighted many safety issues throughout this report that
can be broadly categorized as technology barriers, human error,
variations in blood-banking practice, and deficiencies in FDA's
inspections and monitoring.  Some of the hazards identified in
chapters 2-5 are amenable to immediate steps to reduce risk, with
some associated costs, while other issues are dependent on further
research or actions by the blood industry.  FDA can address four
major areas:  (1) gaps in the layers of safety that could have
serious repercussions, (2) error and accident reporting, (3) the
agency's regulations, and (4) inspections.  Below we first summarize
and then make recommendations affecting all four areas. 


   SUMMARY
---------------------------------------------------------- Chapter 6:1

We answered the question, What are the elements of FDA's layer of
safety and do they ensure that the blood supply is safe?  We found 24
issues related to safety in the processes that blood facilities
perform, and we summarize them below in tables 6.1 through 6.8. 
Table 6.1 presents the two issues identified for donor screening
processes. 



                               Table 6.1
                
                    Safety Issues in Donor Screening
                               Processes

Safety issue    Summary
--------------  ------------------------------------------------------
Uniform donor   History-taking questionnaires are developed by
history         individual blood facilities. Style and content of
                history taking may influence the accuracy and
                completeness of donor's answers. AABB's version is
                most comprehensive and readily available

Screening       Privacy is required for the medical examination. The
privacy         amount of privacy for screening donors varies across
                blood facilities. A lack of privacy during donor
                screening inhibits forthright communication. FDA
                recommends privacy for screening and has begun to
                include this in Form 483 observations
----------------------------------------------------------------------
Table 6.2 presents the three issues identified in the area of donor
deferral processes. 



                               Table 6.2
                
                    Safety Issues in Donor Deferral
                               Processes

Safety issue    Summary
--------------  ------------------------------------------------------
Timing of       Facilities differ on whether deferral status is
donor deferral  checked before or after donation. Checking before
registry (DDR)  collecting clearly reduces the likelihood that suspect
checks          units enter the system and eliminates unnecessary
                burden on ineligible donors

Computerizing   Donor deferral registries vary in form and size from
donor deferral  ARC's and ABRA's national, computerized systems to
registries      single facility's hard copy filing systems. FDA
                requires a donor check in some form of registry. Every
                facility could benefit in efficiency and accuracy with
                increased use of validated computerized donor deferral
                systems. Hardware and software costs are cited as a
                barrier for some facilities. Inexpensive personal
                computers might serve this purpose better than hard
                copy systems. Continued verification and validation is
                important for any system that a blood facility chooses
                to implement

Donor deferral  Donor notification varies by facility practice. FDA
notification    recommends the notification of donors deferred for HIV
                only. Many facilities notify donors who are
                permanently deferred for other reasons. Some
                notification does not take place. Not all facilities
                perform available licensed confirmatory tests to
                provide adequate information to these donors. Not
                notifying these donors could create public health
                problems
----------------------------------------------------------------------
Table 6.3 provides a summary of bacterial contamination, the safety
issue that we identified in the area of collection processes. 



                               Table 6.3
                
                  Safety Issue in Collection Processes

Safety issue    Summary
--------------  ------------------------------------------------------
Bacterial       Bacterially contaminated blood products can cause
contamination   serious harm. An increase in the use of platelets has
                added to the number of cases of bacterial sepsis from
                blood transfusions. Data suggest that this may be the
                leading cause of fatalities resulting from
                transfusions. Also, red blood cells are recognized as
                harboring bacteria under some conditions.
                Technological limitations for identifying blood
                products that have been bacterially contaminated make
                it difficult to test blood and blood products for this
                problem. However, methods for detecting bacteria
                immediately prior to transfusions are under
                development
----------------------------------------------------------------------
A summary of the issue of blood typing, a safety concern in the area
of routine testing processes, is provided in table 6.4. 



                               Table 6.4
                
                    Safety Issue in Routine Testing
                               Processes

Safety issue    Summary
--------------  ------------------------------------------------------
Blood typing    Human error can lead to incorrect blood typing. The
                process has no inherent weaknesses if typing is done
                properly and correctly typed and labeled units are
                transfused to the intended recipient. Data illustrate
                that this does not always occur. Although such
                mistakes appear to be few, the consequences can be
                fatal
----------------------------------------------------------------------
We identified eight safety issues of concern in the area of viral
testing processes.  These are summarized in table 6.5. 



                               Table 6.5
                
                Safety Issues in Viral Testing Processes

Safety Issue    Summary
--------------  ------------------------------------------------------
Window period   Increasingly sophisticated tests are closing the
                window period for viral markers. Gains from additional
                tests will decrease because of the small window period
                presently found with current tests. This period will
                probably never be completely eliminated. Other
                mechanisms, such as improved donor screening, might
                eliminate more window period donations than improved
                viral testing

Autologous      Many blood facilities test autologous units for viral
testing         markers. Some do not perform these tests. Survey data
                illustrate that untested units can make their way into
                the general blood supply system and can be transfused
                to unintended recipients. This could result in serious
                patient harm

Confirmatory    Facilities vary in confirmatory testing practices. FDA
testing         requires confirmatory testing for units repeatedly
                reactive for HIV. Units repeatedly reactive for other
                viral markers do not always have confirmatory testing
                performed. Also, confirmatory tests for some viruses
                have not been developed by test kit manufacturers or
                licensed by FDA. Facilities thus cannot adequately
                inform donors of their disease status, a potential
                public health problem

Recipient       Facilities vary in their policies for recipient
notification    notification and lookback. FDA requires consignee and
and lookback    recipient notification and lookback for units that are
                from a donor implicated in subsequent donations that
                are positive for HIV. No requirements exist for other
                viral markers. Unnotified recipients of units that may
                be positive for other viruses could represent a public
                health hazard

Divergent       Technology barriers hamper the ability of current
viral strains   tests to detect divergent strains of viruses in blood.
                These are usually rare cases and are not often found
                in the U.S. blood supply. CDC conducts surveillance to
                determine the extent of divergent strains of existing
                viruses in the United States

Viral           Fractionation companies employ several inactivation
inactivation    and removal techniques to destroy viruses in plasma
                pools. However, different manufacturers producing
                similar products may or may not use these techniques

Test            Viral testing captures the vast majority of positive
sensitivity     units. Some tests are less sensitive than others and
                some individuals are positive for viral markers but
                carry low-titre antibody levels that are not caught by
                current tests

Emerging        Many viruses not present in the U.S. blood supply are
viruses         not tested by blood facilities. Some newly discovered
                viruses (such as HGV) may pose a problem, since
                preliminary data indicate that 1-2% of U.S. blood
                donors are infected with this virus, which can cause
                chronic hepatitis. CDC continues to monitor emerging
                viruses to determine the extent of problems in the
                United States
----------------------------------------------------------------------
Table 6.6 provides summary information on the one safety issue we
identified in the area of postdonation information. 



                               Table 6.6
                
                Safety Issue in Postdonation Information
                               Processes

Safety issue    Summary
--------------  ------------------------------------------------------
Errors and      Information given by a donor after donating that would
accidents       have excluded that person had it been known at the
                time of collection accounts for a large number of EARs
                submitted to FDA. This may indicate that the process
                is working to ensure a safe blood supply, or it may
                indicate that the guidance on what is to be included
                in an EAR that relates to postdonation information is
                poorly understood. The preponderance of these EARs
                calls into question the adequacy of screening
                processes. Also, there is a large discrepancy between
                EARs submitted by licensed facilities and plasma
                centers, even though they collect approximately the
                same number of units
----------------------------------------------------------------------
Table 6.7 provides summary information on the single issue we
identified in the area of storage and distribution processes. 



                               Table 6.7
                
                Safety Issue in Storage and Distribution
                               Processes

Safety issue    Summary
--------------  ------------------------------------------------------
Inventory       Data indicate that because of human error, many units
management      are unaccounted for or lost before the unit is to be
                transfused. Surveys of blood facilities corroborate
                this problem. Although not directly a safety issue, it
                results in many donated units not being used
----------------------------------------------------------------------
We identified seven safety issues related to FDA's monitoring
activities.  These are summarized in table 6.8. 



                               Table 6.8
                
                    Safety Issues in FDA Monitoring
                               Activities

Safety issue    Summary
--------------  ------------------------------------------------------
EARs            Only licensed facilities are required to submit EARs
                to FDA. FDA information from annual summaries of EARs
                suggests that unlicensed facilities are underreporting
                their EARs (they collect 10% of the blood but submit
                about 1% of EARs). Plasma centers reported at rates
                much lower than licensed blood facilities, despite
                collecting equivalent amounts of blood products. Also,
                the timeliness of reporting EARs to FDA has been
                called into question. FDA has also been slow to
                investigate EARs that may warrant a recall

Use of          FDA guidance to blood facilities is often ambiguous
guidelines and  and results in confusion within the blood industry as
recommendation  to what actions are required and what actions are
s               recommended

Use of EIR      FDA does not perform statistical analyses of
information     information contained in EIRs and corresponding Form
                483 observations

Tracking of     FDA's current list of active registered blood
facilities      facilities contains blood facilities that should not
                be on the list. Also, information on some blood
                facilities is inaccurate. The number of these types of
                cases is small

Timing of       Some blood facilities are not being inspected in the
inspections     time periods set by FDA's guidelines

Incomplete      Many EIRs do not contain pertinent information from
inspection      which FDA supervisors or subsequent inspectors can
reports         determine what blood banking processes have been
                inspected. Analysis of EIR information could provide
                FDA with pertinent data on trends in Form 483
                observations and other issues that arise during an
                inspection

Disparities in  Form 483 observations differ between districts and
inspection      include disparities in what is considered an action
reporting       that should result in a Form 483 observation or
                warning letter
----------------------------------------------------------------------
In summary, we found that there continue to be issues of safety that
FDA, the blood industry, and the research community need to address. 
As we have indicated in another report, the nation's blood supply is
safer than ever before, and the risks associated with blood
transfusions are relatively small compared to many other medical
procedures and life activities.\1 Yet, some areas can be improved by
agency action that would further increase safety. 


--------------------
\1 See U.S.  General Accounting Office, Blood Safety: 
Transfusion-Associated Risks, GAO/PEMD-97-2 (Washington, D.C.: 
1997). 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 6:2

We have nine recommendations by which HHS could improve the safety of
the nation's blood supply.  Six concern gaps in the layers of safety,
one has to do with error and accident reporting, and two relate to
HHS's regulations and FDA inspection processes. 


      DONOR NOTIFICATION
-------------------------------------------------------- Chapter 6:2.1

We recommend that the Secretary of HHS require that blood facilities
notify donors who have been permanently deferred.  This notification
should be based on positive confirmatory results for viral markers
(for the viruses that have licensed confirmatory tests) and all other
medical reasons that result in permanent deferral (for example, the
intake of human pituitary growth hormone).  Notification should
include the reason for the permanent deferral, possibilities for
re-entry as a donor, and counseling or referral to the donor's
physician (including, when pertinent, actions to be taken to minimize
transmission of viruses to others).  We recommend such notification
because of the public health consequences of not informing donors. 


      COLLECTION
-------------------------------------------------------- Chapter 6:2.2

We recommend that the Secretary of HHS require blood facilities
quality-assurance programs to include processes that monitor for
bacterial contamination.  Bacteria can enter blood products during
collection through a donor's skin contamination or illness.  Bacteria
can also be introduced during manufacturing, as in the water baths in
the making of certain blood components.  Both collection and
manufacturing processes are within the control of blood facilities
and could be modified if quality-control information suggested that
products were bacterially contaminated. 


      VIRAL TESTING
-------------------------------------------------------- Chapter 6:2.3

We recommend that the Secretary of HHS require viral testing for all
autologous units.  Since the practice of viral testing for autologous
units varies and since mislabeling and transfusion errors do occur
with some frequency, HHS should require that the blood industry
minimize this vulnerability in the system by testing all units,
whether autologous or allogeneic. 

We recommend further that the Secretary of HHS require confirmatory
testing of all repeatedly reactive viral test results for which there
is a licensed confirmatory test.  We recommend this requirement in
order that the blood facility be given as much information as
possible when it considers whether to conduct lookback and how to
counsel donors and recipients who have a positive confirmatory test. 
However, the information that should be provided if confirmatory
tests are negative or indeterminate should be left to the discretion
of the blood facilities and the recipients' physicians. 


      RECIPIENT NOTIFICATION AND
      LOOKBACK
-------------------------------------------------------- Chapter 6:2.4

We recommend that the Secretary of HHS require that transfusion
recipients be notified when they have been transfused with blood from
a donor whose subsequent donations were found positive in
confirmatory testing.  Notifying recipients of blood that is negative
or indeterminate on a confirmatory test should be left to the
discretion of their physicians.  This recommendation is intended to
reduce the potentially adverse public health consequences of not
informing recipients. 

We also recommend that the Secretary of HHS require lookback in such
situations to find implicated blood units that have not been
transfused or further manufactured into blood components or plasma
derivatives.  The reasonable time period for lookback varies with
each virus, and decisions should be made in consultation with the
blood industry.  Thus, it might be determined that lookback
procedures should be implemented beginning at a specific date when a
memorandum to blood facilities is made final.  We believe that such a
recommendation should be a required practice as soon as possible. 


      ERROR AND ACCIDENT REPORTS
-------------------------------------------------------- Chapter 6:2.5

We recommend that the Secretary of HHS require unlicensed blood
facilities to report all EARs to the agency.  Our information,
analysis, and conclusions highlight the need for such a requirement. 
Such information will provide FDA with additional data from which to
direct inspections of particular blood facilities as well as the
blood industry as a whole. 


      REGULATIONS
-------------------------------------------------------- Chapter 6:2.6

We recommend that the Secretary of HHS publish in the form of
regulations the guidelines that the Secretary believes are essential
to ensure the safety of the nation's blood supply and that it clarify
its position on the extent to which facilities should adopt the
agency's guidelines and memoranda in order to remain in compliance
with HHS regulations.  The blood industry has consistently identified
this ambiguity as a source of confusion and frustration and has
raised concerns about the practice of setting standards through
inspection observations and warning letters.  Policy in the form of
guidelines does not have the enforcement power or public input of
formal regulations, whereas the use of regulations may increase
compliance and decrease the likelihood that guidelines will be
misinterpreted or applied inconsistently. 


      INSPECTION PROCESSES
-------------------------------------------------------- Chapter 6:2.7

Finally, we recommend that the Secretary of HHS correct the problems
we have identified in FDA inspection processes.  FDA needs to perform
statistical analyses of inspection reports, develop policies to FDA
inspectors that would require them to list on the inspection reports
what they had observed during an inspection, publish better guidance
to inspectors and district offices on the types of activities that
warrant observation reports and warning letters, and ensure that all
blood facilities are inspected in a timely fashion.  We believe that
these changes are necessary to improve FDA's ability to discriminate
between facilities that comply and those that do not. 


   AGENCY COMMENTS AND OUR
   RESPONSE
---------------------------------------------------------- Chapter 6:3

In a written response to a draft of this report, the Department of
Health and Human Services (HHS) generally concurred with our findings
and recommendations.  Points of disagreement were primarily related
to our findings and recommendations on recipient notification and
lookback procedures for viruses other than HIV and FDA's inspection
process and knowledge of the compliance status of individual blood
facilities and the overall blood industry.  HHS also provided a
number of technical and editorial comments, which we have
incorporated into the report as appropriate. 

HHS agreed that notifying donors of their deferral status and the
medical reason for deferral could enhance public health.  However,
HHS pointed out that FDA has historically considered its jurisdiction
to apply primarily to product safety, purity, and potency.  It agreed
to explore regulatory options within its existing authority for
requiring notification. 

HHS agreed that a reduction in bacterial contamination of blood
products is an important safety issue.  HHS noted that this issue is
not easily resolved because of the limits of technology, and a study
is currently under way to estimate the incidence of, and identify
risk factors for, bacterial contamination of blood and blood
products.  We understand the technical limits in identifying
bacterial contamination and have recommended that there be a
requirement that blood facilities have a quality- assurance program
that includes processes to monitor for bacterial contamination. 

HHS agreed that testing autologous units for viral markers is an
important issue and is working on a recommendation to blood
facilities regarding testing of such units.  However, we believe that
such practices should be required in order to further reduce the risk
of transfusion-associated disease transmission. 

HHS agreed that units implicated from subsequent donations that are
found to be positive for viral markers should be identified and that
consignees of such products should be notified.  However, HHS
requires such action only for HIV-implicated units.  We believe
consignee notification and identification of blood and blood products
should be required for all subsequent donations that are found to be
repeatedly reactive for any viral markers currently tested for by
blood facilities and for which a positive result on a licensed
confirmatory test has occurred.  In regard to confirmatory testing,
FDA has recommended these tests be performed for HCV and HbsAg.  HHS
has recently issued a final rule that requires confirmatory testing
on units that are repeatedly reactive for HIV.  We believe that
confirmatory testing should be required for all units that test
repeatedly reactive and have a licensed confirmatory test.  HHS
presently requires notification of recipients of units that are from
a donor who subsequently tests repeatedly reactive and is positive by
a licensed confirmatory test for HIV.  We believe that such
procedures should be required for all recipients who received blood
or blood products that are from a donor who subsequently tests
repeatedly reactive and positive by a licensed confirmatory test. 

HHS pointed out several reasons why lookback procedures that include
notification of consignees and identification of implicated units,
confirmatory testing, and notification of recipients should not be
performed for non-HIV viruses.  We have outlined in the report
reasons that run counter to HHS's arguments.  We believe that such
lookback procedures should be required for all viruses currently
tested for by blood facilities for which there is a licensed
confirmatory test in order to further reduce the risk of viral
transmission through blood and blood products and to decrease the
risk of secondary transmission of these viruses to the public. 

HHS agreed that error and accident reporting requirements should be
applicable to all blood facilities and is currently working on a
proposed rule to require submission of error and accident reports by
unlicensed, registered blood facilities. 

HHS agreed that clarification of the nature of FDA's guidance
documents is an important issue and recognizes the need to have more
uniformity in its development and use of guidance documents.  To this
end, public comments have been solicited on this issue through a
notice published in the Federal Register on March 7, 1996 (61 Fed. 
Reg.  9181).  We believe the use of guidance documents is an
important tool that FDA can use to react quickly to emerging public
health threats and advances in medical knowledge and technology.  We
also believe that some recommendations in these guidance documents
are important enough that they should be codified in federal
regulations.  Through this process, such recommendations can also be
opened up for public comment for review and possible revision. 

HHS disagreed with much of our recommendation that FDA should perform
statistical analysis of inspection reports, require FDA inspectors to
list on the inspection reports what had been observed during blood
facility inspections, provide better guidance on the types of
activities that warrant reports on deviations and warning letters,
and ensure that all blood facilities are inspected in a timely
fashion.  HHS pointed out that FDA already reviews and analyzes
inspection reports, both for identification of conditions warranting
immediate action and for longer-term trends.  Furthermore, HHS noted
that the compliance program, investigations operations manual,
regulatory procedures manual, and other FDA directives to
investigators state the information that should be included in EIRs. 

Our analysis of EIRs and Form 483 observations was performed to
examine compliance rates among a nationally representative sample of
blood facilities.  After examining the EIRs in our sample, we
concluded that compliance rates could not be determined because many
of the EIRs had very little information as to what activities had
been inspected and observed by the FDA investigator.  We were aware
that FDA's policy was to allow investigators to list the compliance
program under which the blood facility was being inspected.  By doing
this, FDA assumes that all directions included in the compliance
program are followed unless otherwise stated on the EIR.  However, as
we have pointed out in this report, such a blanket assumption cannot
be made, since we found instances in which this policy was not
followed by FDA inspectors.  We do not believe that it poses a great
burden to ask that inspectors write a sentence or two listing the
areas they examined, and we found instances in which inspectors made
such notations. 

As a result of our initial conclusions regarding the robustness of
information contained in the EIRs, we performed statistical analyses
on Form 483 observations.  We found differences in the number and
kind of Form 483 observations across FDA districts as well as
examples of inconsistent application of Form 483 observations and
warning letters.  HHS noted that FDA has performed similar analyses
and points to the 1992-93 FDA task force on ARC as an example. 
However, when we reviewed these analyses, we found them to be simply
a compilation of Form 483 observations separated into different
categories.  No statistical analysis was performed on these data. 

Furthermore, FDA conducted a study of form 483 observations made by
inspectors.  The study's conclusions were that the majority of Form
483 observations were valid, but complete assessments could not be
made outside the context of the EIR (of course, with little
information in many EIRs, this might be problematic).  Those
conducting the study determined that the most appropriate manner in
which to use these conclusions was to develop a specific section for
writing Form 483 observations in the blood banking training courses
provided to blood bank inspectors.  We believe that conducting this
study suggests that FDA was aware of problems in Form 483
consistency, and its conclusion for additional training supports this
viewpoint. 

We believe that FDA's oversight of the blood industry could benefit
from the types of analyses we have recommended.  HHS noted that such
analysis would be difficult and costly to perform.  We disagree with
this assessment because we performed analyses on Form 483
observations that provided a wealth of information on the number and
kind of observations being handed out by FDA inspectors. 
Furthermore, such analyses could be similarly performed on EIR
information.  Of course, this would be worthwhile only if FDA changed
its present policy and required its inspectors to specifically note
on the EIR the areas of a blood facility that they had inspected. 
Such a change would provide FDA with needed information on compliance
rates between different types of blood facilities, areas of blood
banking that might require more or less investigative oversight,
possible inconsistent application of FDA guidance by inspectors, and
changes in compliance rates as a result of the institution of new
recommendations to blood facilities. 


VIRAL AND NONVIRAL AGENTS
DESCRIBED
=========================================================== Appendix I

In this appendix, we describe viral and nonviral agents that may
affect the U.S.  blood supply.  We provide information on the
characteristics of each agent, on how it is transmitted to humans,
and on some of the clinical outcomes from infection.  We also
highlight guidelines and recommendations to illustrate the federal
government's role in ensuring that these agents are eliminated from
the blood supply. 


   VIRAL AGENTS
--------------------------------------------------------- Appendix I:1

Among others, the agents described below are transmissible by blood
transfusions and therefore can pose a risk to transfusion recipients: 
CMV, HAV, HBV, HCV, HIV-1 and HIV-2, HTLV-I and HTLV-II, and
parvovirus.\1


--------------------
\1 Among the many other viruses transmissible through blood are
tropical viruses such as yellow fever, Dengue fever, ebola virus, and
malarial infections; others include parasitic infections such as
filariasis, toxoplasmosis, babesiosis, and Lyme disease (the latter
caused by a spirochete).  HDV, HEV, and HGV, discussed in chapter 3,
are recently discovered hepatitis viruses that are transmissible
through transfusion. 


      CYTOMEGALOVIRUS
------------------------------------------------------- Appendix I:1.1

CMV is a DNA virus that belongs to the herpes virus group and becomes
latent after primary infection.  It is acquired by respiratory or
sexual contact or from blood components or organ allografts.  It is a
cell-associated virus and does not reside in plasma or serum in
appreciable amounts.  Once a person has been infected with CMV, the
host develops a lifelong persistence of CMV antibodies. 

CMV is widespread in the general population.  While it is
asymptomatic in approximately 80 percent of the population--healthy
individuals--it is a major cause of morbidity and mortality in
immunocompromised individuals, such as newborns, bone marrow or organ
transplant patients, AIDS patients, and some oncology patients. 
People who are at highest risk for CMV infection and disease are
those who are seropositive and become infected from reactivation of
latent CMV. 

There are no regulatory requirements nor does FDA have
recommendations pertaining to CMV because it is ubiquitous in the
general population and has little effect on immunocompetent
individuals.  Because between 40 percent and 100 percent of the adult
population is infected with CMV (depending on geographic
variability), FDA has decided that testing for this virus is not
warranted.  Recommendations regarding CMV are found in the AABB
technical manual, which notes that

     "where transfusion-associated CMV disease is a problem, cellular
     components should be selected or processed to reduce the risk to
     infant recipients weighing less than 1,200 grams at birth, when
     either the infant or the mother is CMV antibody negative, or
     that information is unknown."\2

However, there is some evidence that using CMV-negative blood could
actually increase susceptibility to infection in infants whose
mothers are seropositive, and some studies do not support the need
for specialized components for neonates. 


--------------------
\2 American Association of Blood Banks, AABB Technical Manual, 11th
ed.  (Bethesda, Md.:  1993), pp.  104-5. 


      HEPATITIS A VIRUS
------------------------------------------------------- Appendix I:1.2

HAV is a nonenveloped RNA virus that is very stable and retains its
physical integrity and activity at high temperatures.\3

It has an incubation period of 2 to 6 weeks and is typically shed in
the stool during the final week of incubation, at which time there is
transient viremia.  It is almost always transmitted through the
fecal-oral route or through contaminated water.  Transmission through
blood products is rare because of the short viremic stage and because
no chronic carrier state exists. 

Since no viral persistence exists, liver-associated injury is
transient.  The clinical severity of HAV is directly related to an
individual's age.  Jaundice is unusual in children younger than 2
years old, while fulminant hepatitis and death are much more likely
in persons older than 50.  Approximately 100 deaths are reported each
year in the United States. 

Approximately 25 cases of transfusion-transmitted HAV had been
reported by 1989, representing an overall risk of less than one per
million blood units.  This is probably because HAV is transmitted
through the collection of blood during a short viremic phase during
acute infection. 

Neither regulations nor memoranda contain information pertaining to
HAV because it is rarely transmitted through blood and blood
products.  However, a recent report noted an outbreak of HAV
infection among hemophiliacs who had received pooled plasma products. 
These products had been inactivated with a solvent-detergent
treatment, but this would have had little effect on a nonenveloped
virus such as HAV.  Some have suggested that the addition of a second
virus inactivation procedure (such as heat inactivation) aimed at
nonenveloped viruses might eliminate this risk. 


--------------------
\3 Viruses are frequently characterized by the presence of an
envelope around them.  Viruses consist of a nucleic acid core
surrounded by a capsid, which protects the nucleic acid from enzymes
in a host organism.  Capsids, in turn, can be surrounded by an
envelope.  This envelope is important in the adsorption of the virus
into cell surfaces for infectivity. 


      HEPATITIS B VIRUS
------------------------------------------------------- Appendix I:1.3

HBV is a small DNA virus.  Its replication involves DNA molecules
that lead to the formation of RNA intermediate molecules.  This, in
turn, starts the production of viral DNA by reverse transcription
and, eventually, the complete viral genome.  HBV's mutation rate is
quite high but, because of its small genome, it is often incapable of
forming infectious viruses. 

The discovery in 1965 of Australia antigen, now known as hepatitis B
surface antigen (HBsAg), and its subsequent association with HBV led
to the development of sensitive, specific markers of HBV infection. 
HBsAg can be detected in serum 30 to 60 days after exposure to HBV
and persists for varying periods, depending on the severity of the
infection.  Donor screening for HbsAg began in 1969 and became
mandatory in 1972. 

HBV is a major cause of acute and chronic hepatitis, cirrhosis, and
hepatocellular carcinoma.  The most serious consequences stem from
chronic HBV infection, which occurs in 6 to 10 percent of infected
adults, 25 percent of infected children, and 70 to 90 percent of
infected infants.  In the United States, approximately 300,000
persons are infected with HBV annually.  Of these, 50 percent become
ill with symptoms of hepatitis, 10,000 require hospitalization, and
350 die of fulminant disease.  Furthermore, about 15 to 25 percent of
carriers of HBV develop chronic active hepatitis, which often
progresses to cirrhosis.  An estimated 6,000 persons die each year
from HBV-related chronic liver disease.  Approximately 80 to 90
percent of patients who receive a component of blood from a donor
infected with HBV will acquire the infection. 

Several studies have concluded that some persons infected with HBV
might transmit it despite being HBsAg negative.  A second hepatitis B
test was instituted in 1986-87 (anti-HBc) as a surrogate marker for
non-A, non-B, hepatitis, but it was also seen as a way of catching
some negative HBsAg donations that were, in fact, positive for HBV. 
However, recent information has shown that HBV may be transmitted
despite rigorous testing of donors for HBsAg and HBc antibodies. 
These cases may be caused by low-titre HBV infections from HBV
variants that have mutated. 

HBV can be transmitted through percutaneous or permucosal routes, and
infective blood or body fluids can be introduced at birth, through
sexual contact, or by personal contact.  According to CDC, other
groups at increased risk include injecting drug users, heterosexual
men and women and homosexual men who have multiple partners, infants
born to HBV-infected mothers, recipients of certain plasma-derived
products (including hemophiliacs), hemodialysis patients, and health
workers who have contact with blood. 

A plan that CDC developed in 1989 to eliminate HBV transmission in
the United States called for screening all pregnant women for HBsAg
and immunizing infants of HBsAg-positive women, integrating HBV
vaccines into routine childhood vaccination schedules, and
vaccinating high-risk individuals in selected settings.  CDC
estimated that this would eliminate HBV as a "significant health
problem" by 2015.  Immunization of infants began in 1993 with the
goal of vaccinating 90 percent of them by 1996. 

Title 21 CFR section 610.40 stipulates that each donation of blood,
plasma, or serum should be tested for the presence of HBsAg, while
section 610.41 notes that persons known to have previously tested
positive for HBsAg cannot serve as donors of blood, plasma, or serum
except for vaccine and laboratory purposes.  This also applies to
source plasma.  FDA's December 2, 1987, "Recommendations for the
Management of Donors and Units That Are Initially Reactive for
Hepatitis B Surface Antigen (HBsAg)" outlines several issues
pertaining to HBV:  all donations should be tested by a third
generation test, HBc antibody testing can be used to further evaluate
the status of donors, and, following a flow chart for HBV testing,
donors who had previously tested positive for HBsAg could be retested
for reentry into the donor pool.\4

As noted previously, FDA recommended the anti-HBc test in 1986-87 as
a surrogate marker for non-A, non-B, hepatitis.  In 1991, FDA
recommended the test's use to detect products repeatedly reactive for
HBV.  Additionally, an FDA compliance manual outlined the reentry
algorithm for HBsAg, although it did not include a reentry algorithm
for anti-HBc because no confirmatory test is available.  Source
plasma centers must test for HBsAg but do not have to test for
anti-HBc because the exclusion of repeatedly reactive HBc plasma from
pools processed into derivatives might result in decreased safety of
the derivatives as a result of a reduction in antibody to HBsAg. 


--------------------
\4 CDC's April 19, 1991, Morbidity and Mortality Weekly Report
outlines HBV and HCV guidelines for notifying donors and for medical
evaluation and counseling. 


      HEPATITIS C VIRUS
------------------------------------------------------- Appendix I:1.4

Non-A, non-B, hepatitis was first recognized in 1974.  In 1989, HCV
was isolated and determined to be the major cause of most
transfusion-associated non-A, non-B, hepatitis.  Replication of HCV
occurs primarily in the liver; however, the mechanism of cell
destruction in acute and chronic infection is largely unknown. 

Acute hepatitis C is characterized by mild or asymptomatic infection
in most patients with a gradual onset that may include vague
abdominal discomfort, nausea, vomiting, malaise, and absence of
appetite.  Acute HCV infection results in clinically apparent illness
in 20 to 30 percent of cases and rarely leads to fulminating fatal
disease.  Chronic hepatitis develops in an average of 70 percent of
infected persons.  Even in the absence of biochemical evidence of
chronic liver disease, persistent infection develops in at least 85
percent of infected persons. 

No effective neutralizing immune response to HCV has been identified. 
The genetic heterogeneity of HCV and its ability to undergo rapid
mutation probably represents the mechanism by which HCV evades host
immune surveillance and establishes and maintains persistent
infection.  Parenteral transmission for HCV includes blood
transfusions and recipients of plasma derivatives, hemodialysis and
organ transplant recipients, IV drug users, and health care
personnel. 

HCV is transmitted efficiently by large or repeated percutaneous
exposures to blood such as through transfusion of blood or blood
products from infectious donors or injection drug use.  While overt
percutaneous exposures to HCV (for example, accidental needle sticks)
have been documented as means of HCV transmission, the role of mucous
membrane and inapparent parenteral exposures is not well defined. 

With regard to plasma derivatives, hemophiliacs transfused solely
with untreated or incompletely inactivated clotting factor
concentrates have HCV prevalence of 80 to 90 percent; hemophiliacs
who receive appropriately inactived components or single-donor
cryoprecipitate are generally HCV negative.  Studies have found that
whole-blood recipients who receive a component of HCV-infected blood
are 80 to 90 percent likely to acquire the infection. 

The natural history of HCV infection is not well understood.  An
estimated 20 percent of patients ultimately develop cirrhosis, and
HCV infection has been associated with hepatocellular carcinoma. 
Chronic HCV infection may be symptomatic or asymptomatic, and
patients with HCV infection commonly have fluctuating levels of
aminotransferase.  There is no correlation between aminotransferase
level and disease severity based on liver biopsy findings, and up to
one third of patients with normal aminotransferase levels have
evidence of chronic hepatitis on biopsy. 

Population-based studies of patients with chronic liver disease
suggest that HCV may be as important as or more important than
alcohol as a cause.  In one study conducted in Jefferson county,
Alabama, 40 percent of identified patients with chronic liver disease
had evidence of HCV infection, 25 percent had HCV infection alone,
and 14 percent had both HCV infection and a history of excessive
alcohol intake.  Applying these proportions to the estimated 32,000
deaths each year in the United States from chronic liver disease
would find that approximately 8,000 to 10,000 deaths each year may be
related to chronic HCV infection. 

Title 21 CFR section 640.3(c) states that no person should be allowed
to be a source of whole blood who has a history of hepatitis, a
history of close contact within 6 months of donation with an
individual with viral hepatitis, or a history of having received
within 6 months human blood or a derivative of human blood that FDA
had advised blood facilities was a possible source of viral
hepatitis.\5 However, there is no specific mention of testing for HCV
for whole blood (sections 610.40 and 610.41) or source plasma
(section 640.67). 

FDA has issued several memoranda regarding HCV since the introduction
of testing in 1990.  Its April 23, 1992, "Revised Recommendations for
Testing Whole Blood, Blood Components, Source Plasma, and Source
Leukocytes for Antibody to Hepatitis C Virus Encoded Antigen"
outlined the major guidance for HCV.  It recommended that any
repeatedly reactive blood or plasma unit not be used and that a donor
reentry protocol could not be followed because of the lack of a more
specific licensed test.  An August 1993 revision outlined a reentry
protocol for donors who were positive for HCV because of the
introduction of such a test.  However, the recommendation did not
recommend any lookback procedures for previously collected products
from donors who subsequently tested positive for HCV. 


--------------------
\5 This 6-month deferral was changed from a 1-year deferral that had
been outlined in an April 23, 1992, FDA memorandum. 


      HUMAN IMMUNODEFICIENCY VIRUS
------------------------------------------------------- Appendix I:1.5

HIV-1 and HIV-2 are retroviruses that are unique in their replication
cycle:  following entry into a host cell, typically by fusion of the
virus and host-cell membrane, a reverse transcriptase enzyme copies
viral RNA from the virus into complementary DNA.  A virus-associated
integrase then mediates the integration of this complementary DNA
into random sites in the host's chromosomes.  Replication ensues and
is followed by a "budding" from the plasma membrane such that the
virus can infect other cells and, if shed into body fluids, other
organisms. 

Although experts now know that the HIV epidemic began to spread in
the late 1970s, it was not until 1981 that clusters of Karposi's
sarcoma and pneumocystis pneumonia were recognized in homosexual men
in New York and Los Angeles.  It was also in late 1982 that AIDS-like
illnesses were reported in hemophiliacs and recipients of blood
components.  Less than a year later, the HIV-1 virus was discovered,
and FDA required an anti-HIV-1 test by March 1985. 

Several studies have examined factors that affect the transmission
rates of HIV.  Studies have identified that the rate of progression
to AIDS is more rapid for transfusion recipients and for those who
receive transfusions from donors who are subsequently diagnosed with
AIDS within 2 years of donation.  However, subsequent studies have
refuted these findings. 

Studies have suggested that variables that correlate with the
likelihood of HIV transmission include type of blood component and
duration of storage.  Washed red cells stored more than 21 days had
significantly lower transmission rates than other components.  Thus,
manipulation of blood through the reduction of viable leukocytes or
free virus (through leukocyte filtration) in plasma may help reduce
infectivity.  Furthermore, studies have noted that the age of both
the donor and the recipient correlates with the disease's progression
rate, older patients showing symptoms of AIDS earlier than younger
ones. 

According to federal regulations, each donation of blood or blood
component is to be tested for antibody to HIV by a test approved by
FDA.  Additionally, FDA recommends that blood is to be tested for p24
antigen for HIV-1.  In dire emergencies, a blood facility can issue
blood products before the results of tests for antibody to HIV have
been performed.  However, such tests must be conducted as soon as
possible after the blood products have been issued.  These
regulations apply also to source plasma. 

An April 1992 FDA memorandum entitled "Recommendations for the
Prevention of Human Immunodeficiency Virus (HIV) Transmission by
Blood and Blood Products" outlines several steps blood facilities are
to take to protect the blood supply from HIV.  It recommends
education to permit a prospective donor's self-exclusion before
giving blood and criteria for permanent deferral based on risk
behavior, reentry algorithms, retrieval and quarantine of prior
collections, recalls of HIV positive blood, and "second exclusion
opportunities" such as telephone callbacks, or CUEs.  Additionally,
both AABB and FDA have issued a series of questions that donors are
to be asked to determine whether they manifest high-risk behavior.\6


--------------------
\6 As noted previously, FDA now requires consignee notification, more
specific testing of units repeatedly reactive for HIV, and
notification of patients transfused with blood from donors who
subsequently test positive for HIV. 


      HUMAN T-CELL
      LEUKEMIA-LYMPHOMA VIRUS
------------------------------------------------------- Appendix I:1.6

HTLV-I is similar to HIV in the manner in which it replicates itself
(that is, retroviruses).  It has been associated with two main
diseases:  adult T-cell leukemia (ATL) and tropical spastic
paraparesis or HTLV-I-associated myelopathy (TSP/HAM).  HTLV-II has
been associated with certain neurological diseases similar to
TSP/HAM.\7 It is believed that only about 4 percent of persons who
are infected with HTLV-I in childhood develop leukemia-lymphoma, and
no cases of ATL have been reported among U.S.  transfusion
recipients.  Estimates vary widely on the rate of infection of HTLV-I
with a subsequent diagnosis of TSP/HAM. 

In the 1980s, research performed in Japan and the Caribbean, where
HTLV was endemic, documented that HTLV could be transmitted through
transfusions.  As a result, ARC conducted a study in 1986-87 to
examine whether HTLV was prevalent in the U.S.  blood supply.  The
study concluded that there would be about 2,800 new HTLV-I infections
annually in the United States through blood transfusions.  Therefore,
U.S.  blood facilities began screening for HTLV-I when FDA-licensed
test kits became available in November 1988. 

Several studies have examined factors affecting HTLV transmission
rates.  The Transfusion Safety Study found that there was no
transmission of HTLV-I or HTLV-II in recipients of seropositive
donations from acellular components (such as fresh frozen plasma and
cryoprecipitate).  This is because of the required cell association
of the virus.  The study also found that there was no "probable
transmission" by components that had been stored more than 14 days. 

Another study found that 26 percent of recipients of seropositive
donations became infected with HTLV (26 out of 95 seropositive
donations).  This rate compares favorably with rates reported in
Japan and the Caribbean that showed cellular component transmission
rates at 63 percent and 45 percent, respectively.  One possible
reason for this difference is that blood in the United States is
often stored longer than in Japan and the Caribbean.  Estimates vary
widely on the rate of infection of HTLV-I and subsequent diagnosis of
TSP/HAM (0.068 percent to 2.4 percent). 

There are no specific federal regulations on testing for HTLV for
either whole-blood collections or source plasma.  A November 1988 FDA
memorandum entitled "HTLV-I Antibody Testing" outlines several
recommendations regarding HTLV:  handling of donations that are
repeatedly reactive; donor deferral, notification, and counseling;
blood product labeling; and education and informed consent.  The
memorandum also includes background information on HTLV-I and
HTLV-II, a summary of recommended actions on repeatedly reactive
units, and medical and biological aspects of HTLV-I presented by CDC
in its Morbidity and Mortality Weekly Report of December 9, 1988. 

Although there are no requirements regarding HTLV, an FDA compliance
manual recommends the testing of donations of whole blood and
cellular components for HTLV-I.  Additionally, firms that have
licenses for source leukocytes or red-blood-cell immunization
programs must test cells for HTLV-I.  However, source plasma centers
do not have to test for HTLV because of its cell association.  As
noted above, there is no reentry algorithm for HTLV because there is
no confirmatory test. 


--------------------
\7 A recent study suggests an increased prevalence for a variety of
infections in HTLV-II positive donors, which suggests immunologic
impairment.  See E.  L.  Murphy et al., "Medical Conditions
Associated with Human T-Lymphotropic Virus Types I and II (HTLV-I and
-II) Infection," Transfusion, 36 supp.  (1996), 43S. 


      PARVOVIRUS
------------------------------------------------------- Appendix I:1.7

Parvovirus is similar to HAV in that it is a nonenveloped virus.  It
is a single-stranded DNA virus discovered in 1975 in the serum of
normal blood donors; in most surveys, 50 percent of adults show
evidence of past infection.  The incubation period may vary from 6 to
16 days and illness begins with fever, malaise, and the development
of a skin rash on the face, trunk, and extremities.  It can also be
severely detrimental to fetuses. 

In healthy persons, antibodies develop in about 1 week and the
infection is cleared fairly rapidly.  It is believed that in most
healthy persons, the virus does not persist in the circulation but
some evidence suggests infected persons remain chronic carriers. 
Additionally, because it is a nonenveloped virus, hemophiliacs have a
90-percent seropositivity rate.  Solvent-detergent methods of
inactivation of plasma products are ineffective, and even heated
factor concentrates have a transmission rate of 30 to 60 percent. 
However, parvovirus is similar to CMV in that it appears to affect
only small subsets of the population such as immunocompromised
individuals. 

Blood facilities do not test for parvovirus because of the ubiquitous
nature of the virus in the general donor population; the side effects
of infection, which are mild for most individuals; the nonexistence
of a licensed test to detect parvovirus; and a short viremic phase
that results in only rare transmission of parvovirus through
transfusions.  As a result, the CFRs contains no requirements nor
does FDA have recommendations or guidelines. 


   NONVIRAL AGENTS
--------------------------------------------------------- Appendix I:2

Several nonviral agents are transmissible by blood transfusions and
therefore can pose a risk to transfusion recipients.  Those discussed
below are caused by a parasite (Chagas' disease), a prion
(Creutzfeld-Jacob disease), and a bacterial spirochete (syphilis).\8


--------------------
\8 Although no scientific information supports the notion that CJD is
transmitted through blood or blood products, it has been transmitted
through cornea transplants and brain tissue transplants as well as
through the administration of human pituitary-derived growth hormone. 
There is disagreement in the scientific community as to whether
prions are the vehicle by which CJD is transmitted. 


      CHAGAS' DISEASE
------------------------------------------------------- Appendix I:2.1

Trypanosoma cruzi is the causative agent for Chagas' disease.  It is
a protozoan parasite that upon human infection proceeds to an acute
parasitemic phase that lasts a few weeks and a chronic phase that is
lifelong.  Recent attention to this disease in the United States
stems from the growing Hispanic population from Central America and
South America, where it is endemic. 

Chagas' disease has a 10-to-14-day incubation period after which
follow fever and enlargement of the lymph nodes and liver. 
Approximately 10 percent of persons who are infected show signs of
damage to the heart, colon, esophagus, myocardial cells, and cells of
these organs.  The primary mode of transmission is skin contact with
the feces of the reduvid bug.  Infections in children can carry a
mortality rate of 10 percent in endemic areas, while older persons
are more likely to develop a chronic illness with no signs of
infection.  Two thirds of infected persons have no initial symptoms. 

It often does not manifest symptoms for 20 years.  The classic form
of Chagas' disease usually occurs decades after infection.  Estimates
are that more than 100,000 individuals are infected with T.  cruzi in
the United States.  It is also estimated that at least in South
America, the transmission rate is between 14 and 49 percent for
patients who receive transfusions from donors who are positive for
the parasite.  Some have estimated that there are probably more than
100 transfusion-associated T.  cruzi infections each year in the
United States.  However, the actual incidence is hard to estimate
because of the difficulty of diagnosing Chagas' disease and the
frequency of asymptomatic infection. 

There are no federal regulations pertaining to Chagas' disease and
FDA has no requirements because, until recently, T.  cruzi has rarely
been found in the U.S.  blood supply.  However, the AABB uniform
donor questionnaire, which complies with current FDA regulations and
recommendations for donor suitability, has a question on whether the
donor has ever had Chagas' disease.  Additionally, blood facilities
in geographic areas with a high proportion of Hispanic immigrants
include more detailed questions in their donor history interviews. 
Prospective donors who do have a history of Chagas' are permanently
deferred. 


      CREUTZFELDT-JAKOB DISEASE
------------------------------------------------------- Appendix I:2.2

There is some disagreement on the cause of CJD, although efforts by
some scientists point to a prion, a small protein particle that
resists inactivation by procedures that modify nucleic acids.  These
prion proteins can be found in the brain tissue of patients dying of
CJD.  The prion is infectious but does not invoke an immune
response.\9 Infection with this agent leads to a degenerative
neurologic disease that manifests as progressive dementia with memory
loss and poor judgment and intellectual function.  The infected
person can remain asymptomatic for decades after infection but then
progresses rapidly to dementia and death. 

Evidence has been found of CJD transmission through human pituitary
growth hormone and cornea and brain tissue transplants.  In fact, a
cluster of cases of CJD, reported to CDC several years ago from
patients who had received human pituitary growth hormone, resulted in
FDA's recommending that blood facilities defer donors who had
received this treatment.  Although no cases of
transfusion-transmitted CJD have been reported, blood from patients
with the disease have infected animals when inoculated directly in
the brain.  There is no test to detect this disease. 

Although there are no federal regulations pertaining to CJD, FDA
promulgated two memoranda outlining precautionary measures after
blood facilities acted to protect the nation's blood supply from
products that might transmit CJD.  This included a November 1994
market withdrawal by ARC and several plasma manufacturers of products
that had been manufactured from a donor who was later diagnosed with
CJD. 

Similar withdrawals had been made four other times between 1983 and
1992, with another in March 1995.  As a result, a Special Blood
Products Advisory Committee meeting on June 22, 1995, led to
recommendations to further develop policies for CJD because of the
theoretical risk of its transmission through blood.  Despite their
recommendation, the committee emphasized that no scientific evidence
suggested that CJD was transmitted through blood or blood products. 
However, CDC is collecting data and pathologic specimens on AIDS and
hemophilia patients who have been diagnosed with dementia in order to
examine these patients for any evidence of CJD. 

The FDA memoranda recommended permanent deferral of donors who had a
family history of CJD or who received dura mater transplant
grafts.\10 The memoranda also recommended quarantining products and
notifying consignees for products from donors who were subsequently
diagnosed with CJD, had a family history of CJD, had received human
pituitary growth hormone, or had received dura mater transplants. 
Furthermore, FDA has issued revised guidelines for deferring donors
who have a family history of CJD. 


--------------------
\9 However, findings from a recent study suggest that CJD may not be
caused by prions.  Instead, the researchers hypothesize that CJD may
be caused by a tiny virus or a piece of genetic material. 

\10 Dura mater is the tough, fibrous, outer membrane covering the
brain and spinal cord. 


      SYPHILIS
------------------------------------------------------- Appendix I:2.3

Syphilis is caused by the spirochete Treponema pallidum as it
penetrates small abrasions in epithelium or mucosal membranes.  It
has an incubation period of 10 to 90 days (usually 21 days), and in
its primary stage it is seen as a lesion at the point of entry.  The
lesion persists for 2 to 6 weeks, which is also the period of
infectivity.  Tests for syphilis usually become reactive about a week
after lesions appear.  About 50 percent of persons with syphilis are,
however, seronegative during this stage. 

The second stage of infection is characterized by fever, malaise,
headache, and inflamed lymph nodes.  The last stage can take three
forms:  neurosyphilis, cardiovascular syphilis, or a form that
involves skin and bones.  Treatment with penicillin in the first,
second, or early third stage can result in an absolute noninfectious
cure with complete healing of lesions and no development of any of
the late manifestations of the disease. 

From the 1950s, syphilis was thought to have been brought under
control with antibiotics after an intensive national education
campaign.  In the mid-1980s to early 1990s, there was a geometric
rise in the number of cases of syphilis reported to state health
departments.  Since then, syphilis has declined sharply.  One of the
main risk factors is drug use, including the exchange of sex for
drugs.  Transmission through blood is possible but it requires that
blood be drawn during the brief period of spirochetemia.  The
spirochete that causes syphilis rarely survives more than 72 hours at
4 degrees Celsius, so it is usually components stored at room
temperature (largely platelet concentrates) or transfused promptly
after donation that transmit syphilis. 

Most states require reporting of reactive screening results to the
department of health, and they rather than blood facilities do most
of the confirmatory testing.  Whole blood and red blood cells with
reactive screening tests and negative confirmatory tests are usually
discarded, although FDA has stated that use is acceptable if units
are labeled appropriately.  Also, source plasma collected before
screening-test results have been received has been considered
acceptable for further manufacturing.  FDA has not recommended
product retrieval when repeat donors test positive for syphilis
because it does not consider the transmission of syphilis a health
risk for plasma derivatives. 

The test for syphilis is often negative in the incubation phase of
the disease and during much of the first stage.  It is also negative
during many of the late manifestations, such as cardiovascular
symptoms and neurosyphilis.  Conversely, most persons whose serum is
STS-reactive do not have circulating spirochetes.  Thus, syphilis is
more likely to be present in the blood during the seronegative phase
and absent during the seropositive phase.  As a result, the routine
STS test does not ensure protection against transfusion-transmitted
syphilis. 

Federal regulations require that whole blood and plasma are to be
tested for syphilis, and FDA has recommended that donors who have
been diagnosed with or treated for syphilis in the past 12 months be
deferred.  Donors with a positive confirmatory test should be
deferred 12 months.  After 12 months, deferred donors may donate
blood if they have a negative screening test.  FDA also encourages
blood facilities to obtain a letter from a physician documenting
evidence of adequate treatment for syphilis. 


ERRORS AND ACCIDENTS REPORTED TO
FDA BY FACILITY TYPE, FISCAL YEAR
1994
========================================================== Appendix II

                                       Transfusion      Plasma
            Licensed      Unlicensed     service       centers         Total
          -------------  ------------  ------------  ------------  -------------
Process      No.      %    No.      %    No.      %    No.      %     No.      %
--------  ------  -----  -----  -----  -----  -----  -----  -----  ------  -----
Donor      1,178    10%     30    21%      0      0    244    28%   1,452    13%
 screeni
 ng
\Donor       387    3.8      3    2.1      0      0    513     60     903      8
 deferral
Collecti     343    3.3      9    6.2      1  14.3%      8    0.9     361      3
 on and
 process
 ing
Routine      615      6     27     18      4     57      0      0     646      6
 testing
Viral        255    2.5      7    4.8      0      0     12    1.4     274      2
 testing
Postdona   3,766     36      8    5.6      0      0     41    4.8   3,815     34
 tion
 informa
 tion
Labeling   1,461     14     32     22      0      0      5    0.6   1,498     13
Product    1,044     10     25     17      2     29     16      2   1,087     10
 quarant
 ine
Storage      550    5.3      2    1.4      0      0      1    0.1     553      5
 and
 distrib
 ution
================================================================================
Total     10,283   90.9    146   98.4      7  100.3    856    624  11,292     94
--------------------------------------------------------------------------------



(See figure in printed edition.)Appendix III
BLOOD SUPPLY SAFETY QUESTIONNAIRE
========================================================== Appendix II



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)




(See figure in printed edition.)Appendix IV
COMMENTS FROM THE DEPARTMENT OF
HEALTH AND HUMAN SERVICES
========================================================== Appendix II



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


The following are GAO's comments on the HHS October 23, 1996, letter. 


   GAO COMMENTS
-------------------------------------------------------- Appendix II:1

1.  We recognize the fundamental concept of producer responsibility
in FDA's legislative mandate and have changed the wording of the
report to reflect that FDA helps ensure blood safety. 

2.  We understand that EARs are sent directly to CBER.  We have
clarified the language in the report to convey this point. 

3.  We understand that most recalls are initiated by the blood
facility and have stated so in our report, where we note that "a
recall is a blood facility's voluntary removal or correction of a
marketed blood product.
.  .  ." (See page 71, footnote 1.) We also note that "recalls do not
always begin with an EAR.  In some cases, an FDA inspection uncovers
an error or accident that was not reported to FDA and bases a recall
recommendation on its severity.  Some facilities then submit an EAR
even though recall has begun .  .  .  ." (See page 75, footnote 10.)

4.  We understand that this is the case and have clarified the report
to note that

     "a recall is a blood facility's voluntary removal or correction
     of a marketed blood product that violates laws administered by
     FDA.  The Public Health Service Act authorizes FDA to require
     that a manufacturer initiate a recall if there is an imminent
     hazard to the public health." (See page 71, footnote 1.)

Nevertheless, an FDA official told us that 25 percent of the time FDA
must follow up on a recall, meaning that a blood facility had not
taken any actions until FDA had recommended to the facility that a
product recall was warranted.  Additionally, in discussions with a
representative of a large blood facility, we learned that facilities
often wait for FDA's decision on a product recall before initiating
action. 

5.  See comment 6. 

6.  Our report does not state that EIRs are not reviewed.  However,
we do question the ability of FDA to perform an analysis of
inspection activities and findings.  We understand that EIRs are
reviewed, and we stated so in our report:  "after the inspection and
to ensure that inspectors consider all relevant regulations in an
investigation, other FDA officials review EIRs and any Form 483
observations." (See page 78.) This information was obtained from a
written FDA response to our inquiry.  In that response, FDA noted
that to ensure that all relevant regulations are considered, the
inspection reports involving Form 483 observations that indicate a
potential violation are reviewed by FDA officers.  After this review,
the findings are sent to the firm in violation for its corrective
action. 

However, we have added language to the report noting that this
characterization by FDA is inaccurate.  In fact, representatives
within the blood industry have stated that blood facilities do not
receive the review performed on the inspector's Form 483 observations
by FDA officials except when a warning letter or other regulatory
action may arise.  Furthermore, comments made by the inspected
facility regarding the Form 483 observations are not acknowledged by
FDA, nor is any indication given as to the acceptability of any
proposed or completed corrective action.  Also, blood facilities do
not know what classification has been given to their inspection (that
is, no action indicated, voluntary action indicated, or official
action indicated). 

It is only through a Freedom of Information Act request that the
blood facility can obtain the actual EIR of its facility.  As a
result, the blood facility is unaware of the degree to which its
practices have not complied with federal regulations and would not
know the extent to which corrective actions should be taken.  Thus,
even if FDA determines the "compliance status and potential
corrective actions" during a review of the inspector's Form 483
observations, the blood facility would not receive such information
unless the Form 483 observations warranted a warning letter or
further regulatory action. 

7.  We sought information on any analyses that had been performed by
FDA on the content of EIRs and Form 483s at a meeting with FDA
officials.  At this meeting, FDA officials stated that there were no
databases that tracked information on EIRs or Form 483s.  When we
learned of the PODS database, we sought information on it from FDA. 
The information we received shows that this system does not allow for
the systematic analysis of compliance and noncompliance rates at a
national level.  The data elements contained in this system inform
management about the work (operations and resources) performed in the
district and regional offices.  This includes such information as who
performed the operation (employee name and position), where the
operation was accomplished (district), what was covered in the
operation (products inspected), and the results of specified
operations (classification of the inspection).  Thus, it is not a
system for tracking Form 483 observations or activities covered
during an inspection. 

8.  Our report states that

     "we examined each facility in our sample for whether the EIR
     indicated that a particular function had been examined.  If it
     was mentioned at all in the EIR, we considered it to have been
     examined.  If it was not mentioned at any time in the EIR, we
     considered that one could not determine whether the area had
     been examined." (See page 83.)

This was done for the purpose of our analysis of compliance rates
among a nationally representative sample of blood establishments.  We
have clarified that this methodology was used for the purpose of the
analysis in question. 

We were aware that since the use of checklists was discontinued in
October 1994, FDA inspectors only needed to list on the EIR the Form
483 observations and the compliance program under which the blood
facility was being inspected.  As a result of this policy, and after
examining the EIRs in our sample, we concluded that compliance rates
could not be determined.  Because the EIRs often had little
information on what operations had been observed by the FDA
inspector, we did not believe it was appropriate to analyze the
contents of the EIRs. 

FDA would have us make the assumption that if an operation was not
mentioned, that meant that it was checked and found to be in
compliance.  We understand that in many instances this would be the
case.  However, we could not make the assumption as to how often this
was the case for several reasons:  (1) We were told by FDA inspectors
that they focus on certain activities and do not check all practices
occurring at a blood facility.  FDA's instructions to its inspectors
are that, unless it is a limited inspection, inspectors should list
on the EIRs the areas that they did not inspect that are outlined in
the compliance program under which the inspection is taking place. 
Yet, very few EIRs noted areas that were not covered during an
inspection.  (2) We were told by FDA officials that an inspector
cannot check everything on any one inspection yet, again, few EIRs
delineated what was not covered. 

(3) Individual EIRs illustrate to us that FDA's stated policy is not
being followed.  For example, a blood facility inspected in 1994
resulted in a Form 483 observation that no lookback procedures had
been followed at the firm in 1992-94.  However, when we examined the
EIR for this facility for the inspection that took place in 1993,
there was no mention in the EIR that lookback procedures were not
being followed.  This means that either the 1993 inspection examined
lookback procedures and did not find any problem that had been
evident since 1992 (according to the 1994 inspection) or the activity
was not observed in the 1993 inspection and was not listed on the
EIR. 

(4) We were told by FDA that it often tailors its inspections because
it has substantial previous experience with each blood facility,
enabling an inspector to examine areas known to be sources of
problems.  However, when we examined the EIRs, we could not determine
which inspections were "tailored" and which inspections examined all
areas of a blood facility's practices.  In short, we know that there
are instances in which the inspector failed to note areas that were
not examined.  We were not able to determine the percentage of cases
this occurred in. 

Because of FDA's policy, it would be impossible for an FDA
supervisor, outside auditor, or blood bank facility to determine what
activities had been observed and what areas the blood facility had
and had not complied with by simply reviewing the EIR.  We believe
that there is no analytical basis from which one could determine that
the inspector is following the compliance program by simply listing
the program under which the inspection is being conducted. 
Therefore, we found that a meaningful analysis of compliance rates
among blood facilities based on EIRs could not be performed.  Thus,
we reported only Form 483 observation rates in chapters 2-4 because
this was the only meaningful information that one could analyze from
the EIRs. 

FDA officials also stated that the reason FDA does not have a policy
requiring inspectors to list all the practices at a blood facility
and whether they observed them or not was that such a practice was
found to add significant time and cost with no value added.  However,
our analysis of Form 483 observations shows, in fact, that a
statistically significant difference does occur when a checklist is
used.  This is not to suggest that a "checklist" approach is
necessarily a better method than a "systems approach" to inspecting
blood facilities.  However, we do not believe that listing what had
been observed during an inspection on the EIR would be a major burden
to FDA or individual inspectors.  In fact, in several examples,
inspectors did note on the EIR what areas had been observed. 

9.  We understand that this is the case and say so in our report: 
"suspensions or revocation of licenses, injunctions, and prosecutions
may ultimately result from a process begun with an inspector's Form
483 observations of a continuing pattern of deviation." (See page
80.)

10.  We have several pieces of information that illustrate that there
is inconsistency in inspection activity.  (1) We found a
statistically significant difference between the number of Form 483
observations when inspectors did and when they did not use a
checklist to inspect a blood facility.  (2) We found a statistically
significant difference between the kind and number of Form 483
observations between the eight FDA districts examined in our analysis
of Form 483s. 

(3) In our survey of blood facilities, we found that 27 percent of
the respondents did not know what to expect from one inspection to
the next, and 45 percent noted a wide variation in inspectors'
knowledge and training in blood banking terminology and procedures. 
(4) During a recent forum at an AABB national meeting, FDA officials
were asked to comment on a Form 483 observation received by an
audience member's facility.  The FDA officials stated that the
observation in question should not have been a Form 483 observation
and that that was why FDA inspectors were now being sent to auditing
training.  (5) Eighteen percent of all inspections in our sample that
were supposed to have a checklist did not have one.  (6) We found
instances of the inconsistent application of Form 483 observations
and warning letters, which we have outlined in the report.  From
these points, we conclude that there is not nationwide consistency in
the EIR process, contrary to FDA's comments. 

11.  We have added language to note that it is FDA's policy that
inspectors are not expected to suggest remedies to problems that are
found during an inspection nor are inspectors expected to discuss the
regulations that pertain to the problems.  Statements in our draft
report were based on FDA's written response to our inquiry regarding
requirements that FDA might have on delineating specific guidance to
its inspectors.  In its response, FDA noted that "investigators
provide general guidance (to the facility) on applicable documents,
policy, regulations, etc.  which are the basis for the objectionable
condition." Thus, there appears to be some confusion within FDA as to
the policy for its inspectors when it comes to discussing Form 483
observations with a blood facility. 

12.  The report does not take a position on whether a checklist
approach is a more useful method than a systems approach for
inspecting blood facilities.  We do note that there is a
statistically significant difference in the number of Form 483
observations for the inspections that use a checklist.  As we note on
page 86, we could not determine why this difference occurred.  In
regard to FDA's limiting inspections to areas where problems are
likely to be found, we believe FDA has not performed the statistical
analyses that would be necessary to determine these areas.  Also, one
would need to examine all areas intermittently in order to determine
those that are not likely to require extensive inspection oversight. 

13.  We are unaware of any nationwide analysis performed on the
content of EIRs, Form 483 observations, compliance and noncompliance
rates of blood facilities, or disparities in inspection activities
between inspectors.  We have added language acknowledging FDA's
injunctions against ARC and Blood Systems Incorporated (BSI). 
However, our discussion with ARC representatives indicates that the
uniformity mentioned above was only transient and that present
inspections have reverted back to a situation in which ARC finds
large disparities between inspection practices at its facilities. 
FDA has pointed to work performed by a 1992-93 task force that
categorized all Form 483 observations issued to ARC in 1988-92 as an
example of its ability to conduct evaluations that help shape
compliance policy.  However, when we examined this work, we found
that it was merely a list of Form 483 observations broken down by
categories.  No analysis had been performed on this information that
could assist FDA in determining compliance rates among ARC facilities
or trends in the types of problems found. 

FDA issued to ARC annual reports in 1994, 1995, and 1996 on its
progress under the terms of the May 12, 1993, consent decree.  These
annual reports list the Form 483 observations given to ARC facilities
in the preceding year and categorized these observations by topical
headings covered in the consent decree.  This work demonstrates that
FDA has the ability to perform analyses on Form 483 observations. 
However, this has only been done for ARC facilities and is still
merely a listing of the number of Form 483 observations by category. 

14.  We do not believe that a database that included a nationally
representative sample of blood facilities that contained information
on the type of facility, registration number, areas observed and not
observed by the inspector, date of inspection, areas where inspection
observations where found, and classification of the inspection (that
is, NAI, VAI, or OAI) would be costly or overly burdensome.  In fact,
we established such a database for our analysis of EIR content and
Form 483 observational differences. 

15.  We agree.  However, a national analysis of the types of problems
that are being found by FDA inspectors would provide valuable
information to FDA on the activities in blood banking that might need
more or less attention and oversight.  An analysis of such problems
might also provide information on areas where FDA has made
recommendations that might require further clarification in terms of
FDA's regulatory intent.  Lastly, such analyses would also provide
information as to the application of FDA inspection procedures across
different districts. 

16.  We believe that addressing policy questions with investigators
and industry representatives is a worthwhile practice and FDA should
continue such contacts.  However, the evidence presented in this
report regarding inconsistencies in the application of FDA's policies
and guidance illustrates that such activities are not preventing such
problems. 

17.  We stated in our report that "FDA maintains a list of all
registered blood facilities with their registration numbers.  The
vast majority of those that were in our sample were accurately
identified." (See page 83.) However, when we queried FDA for the
latest EIR for a representative sample of blood facilities, we were
forwarded some for which no inspections had occurred for several
years.  Our query to FDA was based on establishments that were
denoted as being active.  Those that were denoted as being "out of
business/no blood processing" were not part of our query.  Thus, our
findings regarding long periods between inspections was based on the
active list of blood facilities.  Furthermore, we found cases in
which an inspector visited a facility only to find that there was no
business in operation.  It is clear that the districts charged with
inspecting such establishments were not aware that the facilities
were not open.  This could mean that a blood facility did not notify
FDA of its intentions to close or that this information was not
conveyed to the district and appropriately noted on the active list
of blood facilities.  In either case, these examples were still
listed as "active" on FDA's list of registered blood establishments. 

18.  The report does not state that inspectors are not knowledgeable
or are not highly professional.  We do note in the report that, in
fact, all the survey respondents felt that the FDA inspectors
appeared to follow a systematic approach during the inspection.  (See
page 89.) Also, our report states that the survey respondents found
inspectors to be generally knowledgeable.  (See page 89.) However,
these same respondents noted that there was a wide variation in the
inspector's knowledge and training in blood-banking terminology and
procedures.  This may be a result of who inspects blood facilities. 
There are 321 field investigators who conduct inspections of blood
facilities.  Of these, 22 (7 percent) are dedicated to inspecting
blood facilities.  This may be one reason for the survey respondents'
noting inconsistencies between the level of knowledge of blood
inspectors.  The survey respondents also noted inconsistencies in how
inspections are conducted.  Additionally, as noted in comment 10
above, there are several pieces of information that call the
consistency of the actual inspections into question. 

19.  Notifying donors of positive and indeterminant test results is
not the same as requiring the notification of donors that they have
been permanently deferred.  Criteria that require that a donor be
permanently deferred (such as positive test results for viral
markers, being an intravenous drug user, or receiving human pituitary
growth hormone) should be in place to protect the safety, purity, and
potency of blood products by notifying such donors that they cannot
donate in the future.  FDA's recommendation to permanently defer
donors for positive HIV test results is in place not only to protect
the safety, purity, and potency of blood products but also to protect
the public health from transmissible diseases.  Other viruses, such
as HBV, have relatively high rates of transmissibility and should be
considered by FDA in a similar fashion as HIV in terms of protecting
the public health from secondary infection. 

20.  We are aware of the technological limitations of identifying
blood products that have been bacterially contaminated before
transfusion.  We are also aware that bacterial contamination is one
of the leading causes of adverse outcomes in blood transfusions.\1 We
have modified our recommendation to take note of these technological
limitations.  As a result, we recommend that FDA require a blood
facility's quality assurance program to include processes that
monitor for bacterial contamination.  This would permit the inclusion
of multiple procedures to recognize and manage transfusion-associated
sepsis and septic complications.  Further, we believe that the study
that is under way to estimate the incidence of, and identify risk
factors for, bacterial contamination of blood products is a good
first step in addressing this problem.  Results from this study
should be used to assist FDA and the blood industry in identifying
ways to overcome problems relating to the bacterial contamination of
blood products. 

21.  We have added language to the report indicating that work is
under way within FDA to examine this issue and that a recommendation
from FDA is currently being developed.  However, we are recommending
that testing of autologous units be required (not recommended). 
Again, we believe that this should be required to assist in
protecting the safety, purity, and potency of blood and blood
products. 

22.  We recommend that blood facilities be required to perform
confirmatory testing on all repeatedly reactive test results for
which there is a licensed confirmatory test (this would currently
include HBV, HCV, and HIV).  We recommend this because we became
aware that some facilities do not always perform confirmatory testing
on repeatedly reactive tests for which there are confirmatory tests
available.  Thus, we believe this should be a required, and not just
a recommended, practice.  This should be done to enhance the safety
of blood products as well as to notify donors of their deferral
status (see recommendation 1) and to have as complete information as
possible for retrospective notification of recipients.  We also
believe that consignee notification should be required for units that
have been shipped for further manufacture so that such units can be
pulled from inventory if they have not been transfused.  This should
also be done to assist in tracing recipients of the implicated units
that have been transfused. 

We have added language to the report noting that FDA has issued a
final rule that requires consignee notification for blood products
potentially contaminated for HIV.  We note, however, that this final
rule pertains only to HIV. 

We also recommend that there be a required lookback for patients who
have been transfused with units that are from donors who subsequently
test repeatedly reactive and confirmatory positive for viral markers. 
Several reasons have been presented in public forums regarding the
pros and cons of lookback.  FDA's comments to our report point out
four such issues that argue against lookback.  First, the present
policy regarding lookback for HIV is in place because it is almost
always fatal and there is a public health risk from secondary
transmission.  Thus, lookback might not be justified for other
viruses, given the high cost of doing a lookback.  However, as noted
above, other viruses are also known to have high secondary
transmission rates (such as HBV).  Furthermore, a recent study
presented at the 1996 AABB annual meeting suggests that, given
certain assumptions regarding the blood supply, lookback for HCV
could be as cost-effective as other common health-related
interventions.\2

Second, FDA points out that treatment for HCV is often not effective. 
However, some studies suggest that long-term recovery may, in fact,
occur with alpha interferon therapy for those diagnosed with HCV
(especially those diagnosed at the early stages of infection).\3
Additionally, FDA has recently approved an interferon therapy to
treat chronic HCV in adult patients.  In clinical trials for this
product, it was found that 23 percent of the patients had a complete
response at the end of their treatment.  Also, some recipients might
benefit from being notified so that they might curtail behavior (such
as consuming alcohol) that could cause more progressive harm after
being infected with such viruses as HBV and HCV. 

Third, FDA has argued that considerations for implementing a program
to identify HCV-infected persons should be based on certain
principles, one of which is that effective treatment and acceptable
guidelines or criteria should be available to determine which
patients should be treated.  However, other viruses such as HIV do
not have an effective treatment, yet FDA now requires lookback for
this virus. 

Fourth, FDA notes that secondary transmission of HCV and other agents
from blood products is minimal and is not generally seen as a public
health threat.  However, the transmission of HIV through blood
products also rarely occurs, yet FDA now requires lookback for HIV. 
Thus, the mere fact that transmission of a given virus rarely occurs
as a result of transfusions has not precluded FDA from requiring
lookback.  Also, secondary transmission does occur with HCV and other
viruses.\4

Fifth and finally, FDA notes that targeted testing of all recipients
of positive transfusions would include a high proportion of false
positive test results because of the high false positive rate of
early screening tests.  Experience with HIV lookback indicates that
the number of persons who can be recontacted after 6 to 12 months is
very low.  Thus, lookback testing is not cost effective.  Our report
does not outline how FDA might handle specific lookback procedures
for non-HIV viruses.  We do note, however, that "the reasonable time
period for lookback varies with each virus, and decisions should be
made in consultation with the blood industry." (See page 101.) Thus,
it might be determined that lookback procedures should be implemented
beginning at a specific date when a memorandum to blood
establishments is finalized (we do recommend that such a
recommendation be required in the future).  FDA should also note that
our recommendation relates only to units that are repeatedly reactive
and confirmatory positive. 

23.  We have added language to the report to indicate that a proposed
rule change is now under review. 

24.  We are aware of the need for guidance documents and state so
explicitly in the report where we noted that "FDA has to its credit
historically issued memoranda to give the industry immediate feedback
on its position on new issues.  This is an important tool for quickly
reacting to advances in medical knowledge or technology." (See page
79.) However, as the information in the report suggests, there is, in
fact, substantial confusion within the blood industry on the
different uses and practical implications of regulations, memoranda,
and guidance documents.  Furthermore, some activities within blood
banking should be required and not simply recommended.  For this
reason, we have recommended that FDA publish such activities in the
form of regulations in order to more thoroughly ensure blood product
safety. 

25.  FDA's reply to this recommendation has several points.  First,
FDA noted that it already reviews and analyzes inspection reports,
both for identification of conditions warranting immediate action and
for longer term trends.  Our use of the words "systematically
analyze" in our recommendation was meant to convey the notion that
FDA should perform statistical analyses on the contents of EIRs,
activities that have and have not been observed, compliance and
noncompliance rates, and Form 483 observations.  We know that FDA
does not presently perform these types of analyses. 

Second, FDA's comment notes that an example of trends analyses
performed by FDA is the 1992-93 FDA Task Force on ARC that
categorized all Form 483s issued to ARC in 1988-92.  At an interview
with FDA officials to discuss databases that were present within FDA,
we asked whether any databases existed that tracked information from
Form 483s.  We were told at that meeting that there were no databases
that had such information.  Regardless of this, we do not view the
task force work as the kind of nationally representative analysis
described above.  The analysis performed by the task force was merely
a list of all Form 483 observations given to ARC in 1988-92,
separated into different categories.  No further analysis was
performed on this information that could inform FDA of trends in
inspection findings or compliance rates. 

Third, the FDA reply outlines several manuals and other directives
that are available to FDA investigators that include what is to be
contained in an EIR and Form 483.  Our data suggest that FDA
investigators do not always follow such information.  One example of
this is our analysis of the checklists completed by FDA investigators
prior to fiscal year 1995.  We found that 18 percent of the EIRs did
not contain a checklist when they should have.  Thus, policy
directives to complete a checklist did not always result in the
checklists being completed by the investigators.  Also, at the 1996
AABB national meeting, FDA officials were asked to comment on a Form
483 observation received by an audience member's facility.  The FDA
officials stated that the observation in question should not have
been a Form 483 observation and that was why FDA inspectors were
being sent back for more training.  Furthermore, in our analysis of
Form 483 observations, we found a statistically significant
difference between the kind and number of Form 483 observations
between FDA districts. 

Information contained in some EIRs that we reviewed had such little
information that it would have been impossible for FDA reviewers,
outside auditors, or future investigators to determine what had and
had not been observed during the inspection.  Therefore, we believe
that FDA cannot determine compliance and noncompliance rates among
the blood facilities that it inspects.  We are aware that FDA has a
policy that allows inspectors to only list on the EIR the Form 483
observations and the compliance program under which the inspection is
taking place.  However, in comment 8 above, we illustrated that this
does not always occur. 

Our survey respondents noted that in many cases FDA inspectors do not
always observe several practices that take place at the blood
facilities.  Because FDA inspectors do not always write down on the
EIR what was not inspected, FDA would be unable to determine in which
areas a blood facility was in or out of compliance.  Thus, the
presence of manuals and directives to inspectors does not guarantee
correct implementation contained in these guidance documents or
consistency in what is to be considered an objectionable event. 

Fourth, FDA's comment mentions that regulation citations are not
included on the Form 483 because in many instances there are several
regulations that may relate to a specific observation.  We are aware
of this and have added language to the report on this topic.  It was
also noted that while FDA believes FDA investigators are very
familiar with how to write a Form 483, a group of FDA's regional and
national biologic expert investigators performed a study in July 1996
that was to assist in providing clearer guidance in terms of the
significance, contents, and format of observations.  When we asked
FDA for the results of this study, we were provided some information. 
The conclusions of the study were that the majority of Form 483
observations were valid but that complete assessments could not be
made outside the context of the EIR.  The panel determined that the
most appropriate manner in which to use the general conclusions drawn
would be to develop a specific module for writing Form 483s in the
blood-banking training courses provided to blood bank inspectors.\5

That FDA conducted this study suggests that the agency is aware of
problems in Form 483 consistency, and the conclusion on additional
training supports this viewpoint.  Additionally, our analysis of
regional differences in the kind and number of Form 483 observations
indicates that additional training is warranted.  Furthermore, FDA's
admission at the recent AABB national meeting regarding further
training of inspectors on what should be included on a Form 483 would
appear to be a good first step in resolving these problems. 

Fifth, FDA's reply described how the agency has changed its
inspection frequency so that blood establishments that are in
compliance may be inspected once every 2 years.  We actually noted
this in several places in our report, most conspicuously on page 23,
footnote 12.  Thus, we were aware that FDA is now using this less
frequent inspection time and we used this in our analysis of whether
the inspections were occurring within the required time periods. 
(See pages 80 and 83, footnote 21.)


--------------------
\1 See U.S.  General Accounting Office, Blood Supply: 
Transfusion-Associated Risks, GAO/PEMD-97-2 (Washington, D.C.: 
1997). 

\2 J.P.  Auchon, J.D.  Birkmeyer, and M.J.  Alter, "Cost
Effectiveness of HCV Lookback," Transfusion, 35 Supp.  (1996), 51S. 

\3 G.  Davis et al., "Treatment of Chronic Hepatitis C With
Recombinant Interferon Alpha," New England Journal of Medicine, 321
(1989), 1501-6. 

\4 A recent presentation at the 1996 AABB national meeting outlined a
case of sexual transmission of HCV.  See C.  Capelli, "A Case of
Transmission of Hepatitis C Virus Between Sexual Partners",
Transfusion, 36 Supp.  (1996) 51S. 

\5 We did find problems in the way in which this study was conducted,
although the conclusions drawn from the study support our findings on
inconsistent inspection activity as it relates to Form 483
observations.  Problems with this study included (1) a
nonrepresentative sample of Form 483s, (2) reviews of the Form 483s
by two investigators without determining interrater reliability, and
(3) no formal coding scheme for classifying the Form 483s. 


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix V


   PROGRAM EVALUATION AND
   METHODOLOGY DIVISION
--------------------------------------------------------- Appendix V:1

Marcia G.  Crosse, Assistant Director
Jacqueline D'Alessio, Assignment Manager
Kurt Kroemer, Project Manager
John E.  Oppenheim, Adviser
Penny Pickett, Communications Analyst
Venkareddy Chennareddy, Referencer
Cynthia S.  Taylor, Writer-Editor


*** End of document. ***