MICHIGAN BIOLOGIC PRODUCTS INSTITUTE CFN*1873886
MARTIN LUTHER
SUMMARY OF FINDINGS:
Inspection of this licensed blood derivative and vaccine manufacturer was scheduled as a follow/up to the violative inspection on 11/18-27/96. A Notice of Intent to Revoke letter was sent to the firm on 3/11/97. The blood derivatives portion of the inspection was conducted in accordance with CP. 7342.004.
The previous inspection was conducted on 11/18-27/96. That inspection revealed numerous CGMP deficiencies, in the firms manufacturing of blood derivatives products and rabies vaccine, which led to CBER issuing a Notice of Intent to Revoke letter (NOIR). The NOIR listed numerous CGMP deficiencies, including the lack of QA/QC unit, failure to approve/reject materials, failure to approve/reject SOPs, failure to establish/follow SOPs, failure to establish SOPs to prevent microbial contamination, and lack of validation. The firm did not establish or follow: lab controls, stability programs, cleaning schedules, and they did not separate/define areas to prevent contamination or mix-up. The firm did not: clean building/facilities, clean/maintain equipment, calibrate equipment/instruments and assure that equipment and design was appropriate for intended use. Also, the firm did not store components to prevent contamination; there was no identification and control of rejected components, and there was no assurance that equipment in contact with components/drug products were not reactive.
The current joint inspection (CBER, Team Biologics and DET-DO) covered the firm’s manufacturing and testing of all lots of Anthrax Vaccine Adsorbed and all lots of -________________
- —-—-- - —---— ~—------ — Other products covered during this inspection included: Rabies Vaccine, Diphtheria-Tetanus Toxoid Adsorbed, lot DT4176, and Blood Derivative Products. Also during this inspection we assessed the firm’s commitment to address issues noted in the March 11, 1997 NOIR letter and their approach to upgrading their operating systems and organization infrastructure to ensure full compliance with GMPs. The inspection revealed that the firm has corrected many of the deficiencies noted on the last inspection and that they are working to meet the time frames and commitments identified in the April 9, 1997 Strategic Plan for Compliance and in their January 8, 1998 Strategic Plan Update.
The firm has implemented the renovation of the anthrax building, and currently have suspended the manufacturing of Anthrax vaccine.
At the conclusion of the inspection an FDA 483 was issued for CGMPs deficiencies observed for vaccine and blood derivative products including: testing failures, environmental monitoring deficiencies, facility deficiencies, stability failures, lack of process and analytical method validation and other GMP deficiencies
The firm’s management promised a written response to the FDA 483,
Inspect ional Observations.
Information Requested by CBER’s Office of Compliance:
AHF
(AntihemoDhilic Factor):(DL) As part of the follow-up on the current manufacturing status and capabilities of the company I was asked to determine the current manufacturing status of Antiheinophilic Factor (AHF). As allowed by the company’s PLA 91-0117, MBPI is licensed to manufacture AHF. I was informed that Michigan Biologics Products Institute has no current plans nor currently has the manufacturing capability to produce Factor VIII. Many of the SOPs, and equipment are no longer available or functioning. See Exhibit DL-l, Interoffice Memo from Anthony M. Luttrell.
-
~-~-‘X Manufacturina For Compliance:(DL) The manufacturing process for the ---—--~—--—---material is partially performed in licensed manufacturing areas of the fractionation facility. FDA has been notified about this manufacturing procedure.
V
This is the area of the fractionation facility where-chemical adjustments for Immune Globulin and Albumin _are also performed. A sample is taken for
- - ——-- and sent to the Q.C. lab for testinQ. -
(Exhibit DL-2). I was informed that all the floors leading to the
elevator and into room V- are cleaned before transport begins. The walls are not cleaned before transporting the sealed tank. Cleaning and preparation of the areas for transport of the tank was not reviewed.
The tank of incubated,
- --passed through a pass-through to room
—---- and into the -manufacturing area. At this point, all —~ processing is done in what use to be the AHF manufacturing areas, rooms
Since this manufacturing process is currently under an the AHF manufacturing area where
-- ~— is manufactured, was not inspected. AHF is no longer being produced and MBPI does not intend to or have the capability to produce AHF in this area.
HISTORY OF BUSINESS:
Michigan Biologic Products Institute (MBPI) previously was known as the Michigan Department of Public Health (MDPH), under CBER License 099. MTBPI is owned by the State of Michigan and has administrative and facility support from various state agencies. The firm is currently negotiating with potential buyers to become privatized.
According to Mr. Robert vanRavenswaay, Deputy Director of Strategic Planning, the firm’s estimated volume of sale is
-
approximately,—. of the raw materials purchased are from outside of the state of Michigan; and approximately of the finished product is shipped out of the state of Michigan. Three of their top suppliers of raw materials are:- ~--- ‘~- -, ma~jor customers for finished products
are
-- - ---- ------> -- -. ..
With all incomingshipments of compon~Ets/raw materials the firm receives a Certificate of Analysis and upon request MBPI will supply their finished product customers with a Certificate of Analysis.
The firm currently has approximately.
--- full time employees, and has a formalized training program. The office hours are- -—-—----
----- ~—- —-- on some occasions manufacturing
operations are
—— -- The firm practices daily clean-up. They utilize the exterminator, --and
- —- —- —- of the month) . For garbage coiwection ~ne firm utilizes the following companies:(solid and medical waste),
The
firm does not own any delivery trucks; they ship their products via common carrier such as-
Also, the firm does not engage in any formal advertising for any of their products.
PERSONS INTERVIEWED.INDIVIDUAL RESPONSIBILITIES AND ADMINISTRATIVE DUTIES:
Upon arrival at the firm on 2/4/98, credentials were displayed and Notice of Inspection was issued to Dr. Robert C. Myers,
DVM, Director (Attachment 1). Dr. Myers identified himself as being the most responsible individual at the firm.Upon arrival at the firm on 2/16/98, credentials were displayed and Notice of Inspection was issued to Nancy A. Suinmertorx, Director of Technical Operations. At the time of issuance Ms. Summerton informed us that she was the most responsible individual at the firm (Attachment 2). On this day Diane S. Alexander, Consumer Safety Officer, CBER, was added to the inspectional team.
Upon arrival at the firm on 2/17/98, credentials were displayed and Notice of Inspection was issued to Nancy A. Suminerton, Director of Technical Operations. At the time of issuance Ms. Summerton informed us that she was- the most responsible individual at the firm (Attachment 3). On this day Jeffrey A. Sommers, Investigator was added to the inspectional team.
Upon arrival at the firm on 2/18/98, credentials were displayed and Notice of Inspection was issued to Nancy A. Summertori, Director of Technical Operations. At the time of issuance Ms. Sunimerton informed us that she was the most responsible individual at the
firm (Attachment 4). On this day Donald A. Lebel/Product Specialist was added to the inspectional team.On 2/20/98, FDA 483 Inspectional Observation Form, wao io~ucd to
Dr. Robert C. Myers (Attachment 5).
Per Dr. Robert C. Myers DVM/Director
own admission he is the most responsible individual for the overall operation of the firm. Also, per Mr. Anthony Luttrell/Quality Assurance Manager, own admission he is responsible for the quality of the product that they manufacture/distribute. The firm provided a current table of organization chart (Exhibit PMS-3).The joint inspectional team was comprised of the following people:
Marsha A Major
Peggy M. Speight
Julia Lukas Gorman
Cynthia L. Kelley
Robin Levis, Ph.D.
Diane S. Alexander
Jeffrey Sommers
Donald A. Lebel
National Expert
Team Biologics Investigator
Supervisory Biologist
Microbiologist
Product Specialist
Consumer Safety Officer
Investigator (DET-DO)
Product Specialist
Investigators Major and Speight were present for the entire inspection. Dr. Levis was present 2/4-5/98; Ms. Gorman was present from 2/4-13/98; Ms. Kelley was present from 2/4-17/98; Ms. Alexander was present 2/16-20/98; Investigator Sommers was present from 2/17-20/98 and Mr. Lebel was present from 2/18-20/98.
The inspection was divided into two focus sections: vaccine
product were covered form February 4, through February 17 1998;
Plasma Derivative products were covered February 16 through
February 20, 1998.
Following is a list of the firms officials who accompanied us through the entire inspection and -provided informatien/copies of documents that are contained in this report:
Robert C. Myers
William R. Nuiruny
Anthony Luttrell
Diane Dreyer
Nancy Summerton
Robert vanRavenswaay William White
MBPI Director
-Assistant to the Director
Director of Quality Assurance
Director of Manufacturing
Director of Facilities
Management/Technical Operations Director of Strategic Planning and
Business Development
Manager, Vaccine Production
E. Ann Stephens
Robert S. Merriman
Jan M. Luttrell
Marie D. Pierre
Kathryn J. Ghastin
William A Bursaw
David N. Slabbekoorn
Kevin M. Rodeman
Heidi L. Heuser
Troy M. Wilsey
Carl Yost
Roger Brown
Harry Funtulsis
Frank Masters
Louise Simon
Richard Smith
Gary L. Warren
Keith Bailey
Joseph Marston
Cynthia M. Kramer
LduLd K. Alberts
Christine Sirnmonds
Barbara Kinter
Michael E. Izdebski
Elizabeth Maes
Rajean Potter
Marie Bengal
Coordinator, Stability Program Manager, Analytical Laboratory Section Clinical Health Scientist Manager,
Bioassay Unit
Lab Scientist, Anthrax Vaccine Animal Potency Testing
Lab Scientist, Anthrax Vaccine Animal Potency Testing
Supervisor, Filling and Packaging Section
Anthrax Vaccine Unit
Diphtheria Component Unit
Product Formulation Unit
Lab Scientist, Anthrax Vaccine Unit
Equipment Technician, Qualification/ Validation Services Section
Procurement Supervisor
Formulation Supervisor
Analytical Chemistry & Raw Materials Unit
Chief, Plasma Derivatives
Warehouse Manager
Production Supervisor
WFI Operator
Supervisor, Calibration/Metrology
Lab Scientist
Analytical Laboratory Scientist
Strategic Plan Execution/Document Control
DVM
Animal Care SectionTraining Coordinator
Health and Safety Officer
Supervisor of Product Inspection
Rabies Production Unit
At the end of the day (if necessary) we held briefing sessions with management to discuss concerns and potential deficiencies.
MANUFACTURING CODES:
The firm does not use a date manufacturing coding system. They are attaching an alphabetical prefix, which is then followed by consecutive numbers (see Exhibit PMS-4 for master prefix list). For example, Rabies Vaccine,
~ Final Anthrax Vaccine,N ~ and -_____________________________ . -
FACILITIES AND
GENERAL PROCESSING:Exhibit PMS-5 is a facility map, listing operations performed in each building. This inspection focused on operations in the following buildings:
Building y:y-----— -
Building
-- _Building
Exhibit DL-2, is the floor plan for building
Exhibit JLG-2, is the floor plan for building
—--- - Rabies vaccine manufactureOBJECTIONABLE CONDITIONS OR PRACTICES:
Below each of the FDA Inspectional Observations, are comments made by the investigator who observed the deficiency.
Marsha A Major (~)
Peggy M. Speight (PMS)
Julia Lukas Gorman (JLG)
Cynthia L. Kelley (CLK)
Robin Levis (RL)
Diane S. Alexander (DA)
Jeffrey Sommers (JS)
Donald A. Lebel (DL)
ANTHRAX Vaccine:
1. The manufacturing process for
Anthrax Vaccine is not validated. For example,a. The formulation
(JLG) The firm has validated the formulation process with media
simulations however this validation did not include holding the media in the formulation tank for extended periods. The last validation of the formulation operation was released by QA on 10/16/96. The last of the three runs did include holding —~ liters of media for~’days. •‘-‘~-- liters of media was sterile filtered into the formulation tank on Sept. 6, 1997; the remainder of the validation was completed on September 12. There is no instruction in the protocol to hold media (Exhibit l.a.1). The previous validation protocol (Exhibit l.a.2) did include instruction for a seven day hold. Integrity of the formulation tanks was validated though iu~ ninute pressure hold challenges only (Exhibit l.a.3). The firm could not provide procedures or data to validate holding of sterile material for up to four months. Exhibit 4A includes the original release test dates for all lots of Anthrax Vaccine; Exhibit 4B includes the filling dates. Exhibit l.a.4 is environmental monitoring records for the formulation of FAVO34 on 9/20/96. Exhibit l.a.5 is environmental monitoring records for the filling of FAVO34 onl/24/97.
b. The formulation tank has not been qualified for mixing time, demonstrating homogeneity of the suspension. Mixing time is not specified in the batch record prior to filling and during th. filling operations. The product is
and settles quickly in the tank.
(1~Th1) To formulate a final product lot, >~-, -
(JLG) The product is adsorbed to
-- —
and readily falls out o~ suspension. The formulation procedure specifies a ~— mix time after formulation and prior to taking samples for testing (Exhibit 3.1, step 7.30). However the firm could not provide validation data to support that the mixing is adequate to create a homogeneous suspension.W. Bursaw described a minimum mix prior to filling and continuous agitation during. However, the filling records do not specify the need to mix the product and therefore do not identify a particular mixing time (Exhibit l.b.l). It should be noted that the filling records are not a coherent batch record, per Se, but a collection of records of individual activities. We were provided with a draft SOP for setting up the tank f or filling that does reference setting up the tank with the mixer, but it does not specify a mixing time and is not yet in effect ( Exhibit l.b.2)
8
c. The firm did not perform media fill challenges to validate aseptic manufacturing after harvest from the holding tank. These operations include the transfer of the sublots from building
Media fills are performed on fermentation and
harvest trains, however not on a scheduled basis.(JLG) At the conclusion of fermentation, the media containing the active ingredient, anthrax protective antigen, is separated from the cell mass through a pre-filter and a filter.
______ --~---- After
the media is harvested, all operations must be aseptic.
/
is aseptically added to the harvest fluid. The-
adsorbs to the protective antigen and miscellaneous media components and precipitates. The precipitate is aseptically harvested.Production steps up to this point have been qualified through media challenges. According to N. Summertori, W. White, and T. Wilsey, these media fills are performed only when deemed necessary due to mechanical maintenance such as replacing a mechanical seal. The last validations were completed in
~— and include fairly detailed protocols. An example is attached as Exhibit l.c.l. These validations include simulations of all manufacturing manipulations from inoculation through precipitate harvest. There is no SOP governing when and how often media challenges of the system are required.After collection of the precipitate, the harvest is transported
to a Biological Safety Cabinet (BSC) in either room
>—-~-- — -There, it is dispensed into centrifuge bottles and centrifuged.
The bottles are returned to the BSC and supernatant is removed.
The pellet/cake is re—suspended in the formulation
-— — —— —~ __ ____________________ ____ ~)and
- ~
The material is then collected into the - sublot bottles. Test samples are taken, the sublot bottle is stoppered and labeled, and then transferred to cold room - tor storage until formulation in building -The
firm has never validated these intermediate operations to assure they are aseptic. As noted in Item l.a. formulation -operations have been qualified by media challenges. The SOP for media challenges of the filling operations is attached as Exhibitl.c.2.
d. There is no validation of
(CLK) SOP 5372.053 -- —-- -- -~ -- used in the — facility for chemical disinfection and decontamination" is used to prepare the --- -- solution for disinfecting surfaces following the use of virulent anthrax in
the —~ facility - anthrax potency
testing) (Exhibit l.d). The - ----- solution is prepared and
used until exhausted. It is kept in plastic spray bottles with no indication as to the date of preparation nor an expiry date.
is also used in the anthrax production facility to disinfect surfaces, specifically in
the biosafetycabi[nets. Validation demonstrating the effectiveness of - - -s ~s a sporicide in these facilities was requested from both Bob Merriman and Bill White. Although both thought that validation did exist, they were unable to provide the validation.
e. The analytical methods for determination of ~--
and - — ~.n Anthrax Vaccine are not validated with
respect to accuracy, precision, linearity, specificity and
limit of detection.
(MM) Exhibit l.e.l is a copy of the firm’s specifications for Anthrax Vaccine. PhemeroJ. (Benzethonium Chloride) is the preservative used in the product. Specifications for the finished
product indicate Phemerol content to be - -- —------ - — —-- —- -
Formaldehyde is also used in production and finished product specifications for formaldehyde are less than Exhibit l.e.2 is a copy of SOP TD6334.l80.3,
- ----
a ~—--—------ - method involving a color change.Exhibit l.e.3 is a copy of SOP ALO4-009--00,
According to Bill White neither of these-methods are validated. He stated they would validate these methods.
f. There is no validation of the length of time sublots are held until they are used in a lot. Sublots have been held longer than
--
For example, sublot AV370 was produced and placed in
coldroom - of building - -- ~further referred to as -in 2/94 until 7/97 at which time it was used to produce lot
FAVO4O.
--
Sublot AV450 was produced and placed in —------ in 5/95 until 3/97 at which time it was used to produce lotFAVO39.
- -
Sublot AV456 was produced and placed in - - - in 5/95 until 3/97 at which time it was transported to the formulation room of building---’ with -- ‘other sublots to make FAVO39. Here it was discovered that AV456 was contaminated with mold, and it was destroyed.(CLK) Following fermentation and harvest, sublots are sampled
for QC release and stored in
- -. bulk form in Cold room -‘-
(Exhibit l.f.
- first and last 4 pages of log book for --- No stability testing has been performed for storage of the bulk vaccine or to determine the integrity of the vaccine components following these extended storage times. QC release testing is not repeated prior to use of a sublot irrespective of the length of time it has been stored.
g. The reference standard used for potency testing is lot FAVOO9, produced March 1991.
(MM) Anthrax Vaccine lot 19 was reportedly used as a reference lot in potency testing from 4/16/90 until 8/95. Anthrax Vaccine lot FAVOO9 was qualified as a reference lot and was initially used on 9/13/94. It continues to be the reference lot.
Exhibit 4C2 are copies of selected test records for Anthrax Vaccine, lot FAVOO9. This lot was filled on 3/12/91 and passed its initial potency test on 3/6/91- with reference lo~ 19. It was retested on 8/30/94 against lot 19 and passed. No other qualifying testing has been performed on lot FAVOO9.
The firm is reportedly trying to qualify a new reference lot, FAVO29, however this has not yet been completed.
h. There are no expiration dates for the working spore concentrations (virulent or avirulent strains). For example, th. production strain, - - -‘ was used to produce sublot AV216 as early as 3/92 and sublot AV450 as lat. as 4/95.
(CLK) _ -- -~ —- details
the preparation of these various spore suspensions from the
-- -~ -------~—---- -<
(Exhibit l.h.). The master seed spore suspensions are prepared from the "Master Seed", the intermediate spore suspensions are prepared from the master seed spore suspensions, and the working spore suspensions are prepared from the intermediate spore suspensions. ~-x-----------~:ubes are produced at each step and are labeled as "B. anthracis, strain -~- --- ------- -
— Seed Culture Lot #XX and date". Additionally, the tubes are numbered for the intermediate and working spore suspensions. However, I was unable to determine how a lot number was assigned each time one of the spore suspensions were prepared. In examining sublot BPR5, working spore stock V-770-NPI-R lot #7 was used to inoculate AV450 in 4/95, however a tube from lot #1 was used to inoculate AV473 in 6/96. ---- tubes out of -- are removed each time a spore suspension is produced to test for purity, contamination and virulence. But no further testing is done on a spore suspension at any time and there are no expiration dates established for any of the spore suspensions. The original Master Seed, from could be found in the original shipping container in the refrigerator used to store the various spore suspensions, among other items.2.. - — testing for Anthrax
sublots used sublot AV462, with a content of 23ppm.
The specification for in Anthrax vaccine is
There is no BF testing at l5ppm or 3oppm.
(MM) Exhibit l.i.l is a copy of results of
-
) testing of Anthrax Vaccine, dated 6/5/96. The testing was performed on sublot AV462 with a content of 23 ppm. According to Robert Merriman, the sublot was-chosen because of its availability.In addition, the firm has not completed work to identify microorganisms present in their environment so the relationship of challenge organisms used in the test to organisms found in the facility has not been determined.
k. Prior to August 1997, the
(JLG) I reviewed the validation for - - — -- filters used in harvest of the anthrax fermentors, approved in August 1997. I questioned W. White, D. Slabbekoorn, and T. Wilsey regarding the filters used prior to this validation. Each reported that the filters used prior to the introduction of the
— — -
filters had not been validated nor were they routinely integrity tested.1. Validation of microbial retention by the
(JLG) The validation of the — filters included a microbial retention challenge (Exhibit 1.1.1). However this challenge was performed only with ). -. media, used for Tetanus toxoid production. I was provided with a matrix which listed the characteristics of the various media/product types relevant to the validation of the filter (Exhibit 1.1.2). A handwritten note on this chart indicates that - media was chosen by MBPI for the validation challenge. When questioned regarding the selection of the media, A. Luttrell stated that they had selected it on advice from -~ ---- -~-- ~, based on
- -
matrix approach. I and MM explained that the matrix approach for filter validation had not been universally accepted by FDA. I asked if MBPI had a justification for the selection of the media as an adequate model for the Anthrax product.A. Luttrell stated that they did not have a documented justification.
("in"
was skipped on the 483)n. WFI used in the production of Anthrax sublots in building
(JLG) CLK and I, during our inspection of the Anthrax vaccine facility building -- were shown the stainless steel tank in which WFI is transported. We were told by W. White, N. Suminerton, and T. Wilsey that the water was collected from building in the tank and brought over to building - --‘ T. Wilsey stated that any water not used within the working day is dumped. SOP VPO2-002-OO (Exhibit l.n.l) details these
operations. We (JLG, CLK) asked if the collection and holding of the WFI had been validated. We were told "No."
o. There is no completed cleaning validation of product contact equipment.
(CLK) Production of anthrax vaccine results in the product contacting various kinds of equipment including centrifuge bottles and ~ glass bottles. This equipment is dedicated for anthrax vaccine production and cleaned with detergent.
(SOP 90054.000 - - — - - - -~
-~ --Exhibit l.o.) Cleaning
validation for this equipment showing the removal of residual detergent was requested. Bill White informed us that the cleaning validation of product contact equipment had not been completed.
2. There are no written procedures, including specifications, for the examination, rejection, and disposition of sublots of Anthrax and Rabies.
a. Su.blots are tested at the time of productiOn, and are not retested prior to formulation. For example, sublot AV450 was produced in 5/95 and then used in formulation of lot FAVO39 in 3/97.
(CLK) SOP 5376.010 —~ —---~--—-____ —--- —----- -~ is an
extensive BPR form with sign-off indicated for each step of the production process (Exhibit 2.a.). This SOP was effective on 12/13/96 (revision 4) and is evident in use as early as 6/96 (sublot AV473). It provides for sampling from the final bulk sublot for the various QC release tests, although, regarding mixing prior to sampling, it is quite vague in it’s description. There is no indication as to how to follow-up on the test results. The sublots are placed i-n •-. upon comp~etion of the harvesting and sampling. They are not "quarantined" in ‘. - -until QC testing is completed and found to be adequate. There are no indications on a sublot bottle that it is undergoing testing, or has passed QC testing, etc. There are no limits on the length of time that a sublot can be stored prior to use in formulation of a lot. There are no provisions for examination of a sublot following storage prior to being transported to the formulation room (in building — . There are no guidelines concerning the selection of sublots from - for formulation into a given lot.
b. Quarantined materials are held for extended periods.
- -
For example, sublot AV216 was placed in - on quarantine in 3/92 and was not destroyed, for low antigen content, until 5/97.--
Sublot AV222 was placed in - in 4/92 and was removed and destroyed in 5/97 due to mold.--
Sublot AV493 was manufactured in 8/96 and is still in quarantine in ~— - for low antigen content.(CLK) [See Exhibit 1.f., log of CR 206] D. Slabbekoorn stated that there are no specifications as to why a sublot should or should not be quarantined. Also, there is no procedure for disposing of sublots that fail QC testing (contamination or low antigen content) . There is no follow-up on quarantined sublots.
(JLG) On 2/6/98 I observed the bottle of AV493, produced 8/20/96 quarantined in - — —along with additional quarantined sublots produced in 1997 and 1998.
3. Potency testing of Anthrax Vaccine requires either testing 1 finished product vial, an aliquot from the formulated bulk tank, or a pilot bulk sample. There is no data demonstrating that these samples are representative of the lot.
In addition, expiration dates are assigned based on the latest valid potency test. There is no correlation between this date and formulation of bulk or filling of the finished product.
(MM) Exhibit 3.3 is a copy of SOP 5376.033, -
— —~- --- — — It states that the test may be performed on filled containers, bulk samples or pilot samples. It does not indicate a timefraxne in which the test is to be conducted after the product is formulated.
(CLK) SOP 5376.000 - - -
-- ----—--- - ---- indicates that a —---~-- sample is taken
from the formulation tank after at least ,-~2xours stirring (in the formulation tank at following siphoning of the ---sublots into the tank (Exhibit 3.1). SOP 6166
-~ —-
explains how to obtain a sample from individual sublots prior to formulation that is "a representative sample of the final lot" (Exhibit 3.2). Through interviewing Kathryn Ghastin, who performs the anthrax vaccine potency
testing, I learned that one vial of a filled lot is used for potency testing (need to perform a potency test and there is 5.2 ml per vial) There have been no validation studies to indicate that any of the above three samples would be representative of the lot for potency testing, or any other QC release testing. A lot is potency tested using only one of the three samDles types and released based on that result (SOP
(MM) Expiration dates are assigned based on the
r - - (See discussion of observation 6) There is no
correlation between the date of potency testing and formulation of the bulk, or filling of the finished product. For example, Anthrax Vaccine lot FAVO17 was initially tested for potency from a sample of the bulk on 4/28/92 and filled on 6/18/92 (Exhibit 4ClO); records for lot FAVO18 indicate it was initially tested for potency from a sample of the bulk on 3/9/93, however it was filled on 7/28/92 (Exhibit 4Cll)
Anthrax Vaccine lot FAVO29 was filled on 8/8/95; its initial potency test, conducted on filled units, is dated 7/9/96 (Exhibit 4C22)
. Anthrax Vaccine lot FAVO3O was filled on 8/24/95; its initial potency test, conducted on filled units, is dated 7/30/96 (Exhibit 4C23) . FAVO31 was filled on 9/29/95, with the initial potency on filled units conducted 7/30/96 (Exhibit 4C24); and FAV032 w~ filled on 10/26/95, with the initial potency on filled units conducted 7/30/96 (Exhibit 4C25)Anthrax Vaccine lot FAVOO8 was filled on 3/14/91 with the initial potency test conducted from pilot samples on 3/6/91 (Exhibit 4Cl)
. Lot FAVO11 was filled on 10/17/91 with the initial potency test conducted from pilot samples ~on 7/9/91 (Exhibit-4C4)4. There is no written justification for redating lots of Anthrax vaccine that hav, expired. "Redating" testing consists only of a potency test. There is no documentation of testing for container/closur, integrity or container/closure compatibility for periods up to 7 years. In addition there is no analytical testing identifying and demonstrating the absence of degradants.
There is no written SOP for redating, including when redating will be performed in order to extend the expiration period.
--
Lot FAVO23 was tested for redating 2 times in 1997 and failed. It also failed twice on stability in 1997. It is scheduled for r.dating on 4/21/98.Anthrax lots that are submitted for redating are released by CEER with alternate lot numbers to indicate the redate. However product is not labeled with the alternate lot number.
--
Lot FAVO2O (initial date of potency 4/13/93) was submitted for redating as FAVO2O-2. and was labeled on 2/6/98 as FAVQ2O.For Anthrax Vaccine lots FAVOO8 through FAVO16, the firm unpacked the vials from the cartons and removed the labels (the labels were removed by soaking in alcohol). The firm does not have a written procedure for performing unpackaging of vials and removal of labels. Also, the firm does not hay, documentation of performing reconciliation of the vials before and after this operation.
(MM) The
firm routinely redates Anthrax Vaccine lots that have reached their labeled expiration date. There are no written procedures or justification for redating, including how lots are chosen to be redated and a review of batch records and testing records for the lot chosen to be redated. The firm does not perform container/closure integrity testing nor is there information about container/closure compatibility for redated lots. A potency test is the only test conducted on the lot to be redated.William Bursaw stated he keeps a list of product that has been redated and product that is due for redating. Exhibits 4A and 4B are copies of lists provided by the
firm of Anthrax Vaccine lots and their status.Exhibits 4C1-4C36 are copies of stability records, potency testing results and other selected test information for Anthrax Vaccine lots FAVOO8 through FAVO43. These exhibits will be referred to throughout this report.
As explained by several people at the firm, including A. Luttrell, the product’s initial expiration date is calculated
-~
~see Observation 3).When the product has passed this date it is listed as "outdated" (Exhibit 4B). W. Bursaw stated he will then schedule another potency test for the product. The length of time between the date
of expiry and the next potency test is undefined and is reportedly primarily based the availability of the testing facility.
Once the product has passed potency testing, the
firm notifies CBER of the potency test results and receives from CBER approval for the extension of the expiration date. Product that does not pass the potency test is retested. Exhibit 4Cl6 is a copy of selected test records for Anthrax Vaccine, lot FAVO23, which was filled on 12/13/93. It was initially tested for potency on 3/29/94 and passed and was given an expiration date of 4116/97 It was placed in the stability program (zero time) on 4/2/97. FAVO23 was tested for potency on 4/2/97 as part of the stability program, and also for redating. It did not pass this potency test and is listed as "failed" in the Stability program. It was tested again for potency on 5/14/97 for redating. It did not pass this test and is recorded "Invalid". A third potency test was performed on 8/12/97 and is recorded as failing potency. A fourth test performed on 10/6/97 and passes by 0.01. The deviation report dated 2/10/98 indicates this to be a valid test result and no other testing is required, however, according to W. Bursaw’s records another potency test is scheduled for 4/21/98 for this lot.Product is reportedly labeled when ready to be shipped and may not be labeled while in inventory. However, there are occasions when the product is initially labeled with an expiration date and requires redating. The firm then "relabels" the vials. Again, there is no written procedure for this process so W. Bursaw described the activities involved. Anthrax Vaccine, lots FAVOO8 through FAVO16 were initially manufactured between 3/6/91 through 3/5/92. All of these lots have been submitted to CBER for at least one redating and were "relabeled" according to the firm’s records. "Relabeling" is performed by soaking the original labels off the vials by reportedly standing the vials in a c-ontainer of alcohol. The level of alcohol reportedly is not above the neck of the vial and employees are careful not to immerse the vials in the alcohol. Once the labels are removed1 the vials may or may not be labeled immediately. If they are not to be shipped, the unlabeled vials are reportedly placed back in storage.
The firm’s only reconciliation records maintained for "relabeling" consist of counting the cartons of vials before and after "relabeling". There is no accounting for the individual vials.
Exhibits 4D, 4E, 4F, 4G, are records of the stripping of lots FAVOOS and FAVOO9
Exhibit 4H is lot release documents for the initial release of lot FAVO2O in 1993. Exhibit 41 are the lot release documents for redating FAVO2O, identifying the lot as FAVO2O-1. Exhibit 4J is the records for labeling the lot on February 6, 1998. Exhibit 4K is a copy of the label applied to the redated lot, identifying it still as FAVO2O
5. Regarding the firm’s stability program for Anthrax:
a. The firm’s stability program did not start until 1997. Stability testing consisted only of performing release tests at various intervals and does not address product degradation. There is no justification for placing lots manufactured as early as 1991 into the stability program.
b. The firm does not have a system in place to investigate and report stability failures. (All results noted below are from samples stored at 2-80C. Accelerated stability test results are not included.) For example,
--
Lot FAVO1O was filled on 7/1/91 and submitted for redating on 10/11/94, having passed all potency testing. It was placed in the stability program (zero time) on 10/7/97, and tested for potency. The "zero time" potency for this lot is recorded as "unsatisfactory". There is no investigation into this result nor is there any additional potency testing.--
Lot FAVO11 was filled on 10/17/91 and submitted for r.dating on 10/11/94, having passed all potency testing. It was placed in the stability program (zero time) on 10/27/97, and tested for potency. The "zero time" potency for this lot is recorded as "No Test" having not met the ,> dilution criteria. Test records indicate it was tested for potency 11/24/97. There is no investigation into this result nor isthere
any additional potency testing. In addition, the testis 11 ppm. (Specification is
--
Lot FAVO13 was filled on 10/25/91 and submitted for redating on 11/22/94, having passed all potency testing. It was placed in the stability program (zero time) on 11/12/97,
and is recorded as "No test", having an invalid test. Test records indicate it was tested for potency on 12/5/97. There is no investigation into this result nor is there any additional potency testing.
--
Lot FAVO18 was filled on 7/28/92 and submitted for redating on 6/11/96, having passed all potency testing. It was placed in the stability program (zero time) on 6/10/97, and is recorded as "Unsatisfactory". Test results dated 7/7/97 indicat, it failed potency specifications. There is no investigation into this result nor is there any additional potency testing.--
Lot FAVO22 was filled on 2/9/93 and submitted for redating on 10/15/96, having passed all potency testing. It was placed in the stability program (zero time) on 10/7/97, and is recorded as having an "Unsatisfactory valid test". Test results dated 10/31/97 indicate it failed potency specifications. There is no investigation into this result nor is there any additional potency testing.--
Lot FAVO23 was filled on 12/13/93 and passed a potency test on 3/29/94. It was submitted for r.dating on 4/2/97 and was placed in the stability program (zero time) at the same time. It is reported as failing potency on 4/2/97, and is reported as having an "invalid test" on5/14/97. It was tested again on 8/12/97 and is reported as failing potency. A fourth potency test conducted on 10/6/97 is listed as passing by 0.01. There is no investigation into the original result and justifying the additional testing.
--
Lot FAVO4O was filled on 11/13/97 and placed in the stability program (zero time) on 11/19/97. It is reported as having an "invalid" potency test on 11/19/97. There is no investigation into this result nor is there any additional potency testing.c. The firm’s SOP(s) for handling-manufacturing deviations/departures does not address when a lot should be monitored on stability.
(MM) Stability protocols for Anthrax Vaccine are attached to each lot. As an example, the protocol for lot FAVO18 (Exhibit 4C11) lists testing times and storage temperatures, including labeled temperature storage at 2-80C and accelerated testing at room temperature storage
( — _ -. - and also higher temperatures
of
- / The protocol lists the testing to be performed which is the same as release testing. In addition, there is no provision to place lots that are involved in deviations or departures in the stability program.Please note the observation includes stability testing at the labeled temperature of 2-80C and does not include testing results at the accelerated temperatures. Data reviewed indicates the product is not stable at these temperatures particularly for potency testing and testing for
- - (see Exhibit 4C11 accelerated testing for lot FAVO18). The firm’s stability SOP does not provide for follow-up to out-of-specification test results for product on accelerated stability testing. In addition there is no limit to the length of time accelerated studies are monitored and what the data derived from these studies means.The
firm initiated their stability program for Anthrax Vaccine in 1997. According to the firm’s records the following lots of Anthrax Vaccine have been placed in the stability program, however, there is no justification for the selection of these lots, particularly those manufactured prior to 1997. The firm does not have a system for reporting and investigating product that fails stability testing and these will be listed below. Mr. Luttrell stated that "failing" or "invalid" stability test results would now trigger an investigation and would be reported through the firm’s deviation system.FAVOO8
-- filled on 3/4/91 and submitted for redating on 8/9/94 and a second redating on 8/4/97. It was placed in the stability program (zero time) on 8/4/97. It has met specifications during testing. (Exhibit 4C1)FAVO1O
-- listed in observation. (Exhibit 4C3)FAVO11
-- listed in observation. During the inspection, I stated the phemerol test result was below specification on 10/28/97 at the "Time Zero" station. Dr. Robert Myers stated this lot would be placed in quarantine. (Exhibit 4C4)FAVO13
-- listed in observation. (Exhibit 4C6)FAVO18
-- listed in observation. (Exhibit 4Cll)FAVO22
-- listed in observation. (Exhibit 4C15)
FAVO23
-- listed in observation. (Exhibit 4C16)FAVO29
-- filled on 8/8/95, with a potency test on 7/9/96, it expires 7/13/99. It was placed in the stability program (zero time) on 4/22/97. It has met specifications during testing. (Exhibit 4C22)FAVO32
-- filled 10/26/95, with a potency test on 7/30/96, it expires 7/20/99. It was placed in the stability program (zero time) on 4/3//97. It failed initial sterility testing on 7/21/97- - ~-
It passed a retest. This lot is mentioned in observation 9. (Exhibit 4C25)FAVO35
-- filled 2/5/97, with a potency test on 1/10/97, it expires 1/11/00. It was placed in the stability program on 1/10/97. It failed initial sterility testing on 8/8/97 -k -~- .
It passed a retest. This lot is mentioned in observation 9. (Exhibit 4C28)FAVO4O
-- listed in observation. (Exhibit 4C33)6. There has been no investigation into numerous "invalid" potency test results for lots. For example:
--
Lot FAVO22. was filled on 11/24/92, having a passing potency test. It was tested again on 10/15/96 and failed potency. It was tested again on 1/28/97 for redating and passed. There is no investigation into the test failure nor justification for ret,sting the lot.--
Lot FAV025 was filled on 4/24/94, having a passing potency test. It was tested again on 4/22/97 and failed potency testing. There is no investigation into the test failure.--
Lot FAVO28 was filled on 6/2/95. It was not tested for potency until 7/9/96 when it failed the test. It was tested again on 8/27/96 and passed. There is no investigation into the test failure nor justification for retesting the lot.--
Lot FAVO41 was filled on 11/18/97. It had an "invalid" potency test on 9/30/97. There is no investigation into this invalid test.--
Lot FAV042 was filled on 11/21/97. It had an
"invalid" potency test on 10/29/97. There is no investigation into this invalid test.
--
Lot FAV043 was filled on 12/25/97. It had an "invalid" potency test on 11/18/97. There is no investigation into this invalid test.--
Lot FAVO44 was filled on 1/7/98. It had an "invalid" potency test on 12/8/97. There is no investigation into this invalid test.(MM) Exhibit 3.3 is a copy of SOP 5376.033, --
- It describes the potency test and how to determine whether the test is satisfactory. It does not describe what to do if the test is not satisfactory, or how to report this type of result. According to Mr. Luttrell, there was no mechanism for reporting test results classified as "invalid" into the firm’s deviation or departure system, which would have triggered an investigation. He stated all "invalid" test results would now b~. r~prured in the firm’s deviation reporting system.
During discussions with the firm about the numerous "invalid" results, I pointed out that it appeared the firm began seeing these results during 1997. The reference lot for all of these tests is FAVOO9 which is discussed in Observation 1G. I stated that during 1997, the firm also had numerous valid passing tests, including lot FAVO23 which passed testing on 10/6/97 with a valid test result while lots FAVO1O and FAV022 fail during the same test. Other lots that are listed as having valid passing test results in 1997 are: FAVOOB (8/4//97) (Exhibit 4Cl) ; FAVO12 (11/4/97) (Exhibit 4C5) ; FAVO24 (4/22/97) (Exhibit 4C17) ; FAVO27 (7/14/97) (Exhibit 4C20) ; FAVO29 (10/13/97) (Exhibit 4C22) ; FAV032 (9/22/97) (Exhibit 4C25) . I stated that any investigation into these "invalid" test results would also have to address the passing tests obtained during this-time.
The lots listed in the observation and their respective Exhibit number are:
FAVO21
-- Exhibit 4C14.
(CLK) The present potency test procedure for anthrax vaccine requires that the test meet four criteria to pass (Exhibit 3.3). They include:
reference vaccine, then the firm "invalidates" the test. If any one of these criteria are not met by the test vaccine, then the test vaccine "fails". There are no specifications at present to indicate when a retest is applicable. There are no investigations initiated concerning "failed" or "invalid" tests. For example, lot FAVO43 (a pilot sample) had an "invalid" test result on 11/17/97 because ~‘~‘~of the reference vaccine test animals did not survive the mid-dilution (Exhibit 6.1). However, the test vaccine itself appeared to render an appropriate result. An "invalid" reference has resulted in numerous potency test failures recently (over the last 6-S months). Although the firm is trying to establish a new reference vaccine and trying to validate new criteria for which to evaluate potency test results, there have been no investigations into the cause for the repeated invalid" test results.
7. The firm’s SOP
---dated 9/3/96, requires that vials discarded as rejects be counted, however, it does not specify limits f or when a lot should be investigated or rejected as a result of this lot reconciliation. For example:
--
Lot FAVO16 had 6579 vials rejected due to particulates during post filling inspection. These particulates were not identified, nor was an investigation conducted. The batch was released.--
Lots FAVO2S, FAVO31, FAV033 and FAVO38 had 3323, 2441, 2509, 1347 vials rejected respectively for low volume during post filling inspootion. There was no investigation conducted.--
Lot FAV035 had 409 vials rejected just for faulty closure during post filling inspection. There was no investigation conducted.(CLK) SOP 4430.001 -
details the visual inspection process for all products made at I’EPI and filled into vials of various sizes (Exhibit 7.1). The "Filling
- Inspection Record" for lots FAVO16, FAVO28, FAVO31, FAV033, FAVO35 and FAVO3S indicate the number of vials rejected for the various reasons (Exhibits 7.2, 7.3, 7.4, 7.5, 7.6 and 7.7) . There have been no investigations into the large number of vials rejected for the various reasons stated above.Lot FAVO16 was filled and visually inspected in 1992. It was released by CBER with an initial expiration date of 3/95. In 7/96 it was submitted to CBER for its first redating based on potency retest in 6/96, and released by CBER. On 7/31/97, during QA review, a deviation report, 97DAV70, was issued with regard to the large number of vials rejected in 1992 (Exhibit 7.8). No further investigation was performed.
On 2/10/98, when inspecting the cold rooms in building’,
--where final container products are stored, I visually inspected random vials from lots FAVOO8, FAVO16, FAVO23 and FAVO35 for the presence of particles, any signs of stopper degradation and container closure. In the randomly selected samples from each lot, I observed no deficiencies or particles.8. The firm does not have specifications for time limits at which the product can be exposed to room temperature conditions during filling, labeling and packaging operations (repeat observation).
--
Lot FAVO36 was at room tmperature for 20 hours and then the filling operation was aborted, it was placed back in the refrigerator (deviation report 97DAV34).In addition, there is no stability information regarding product exposure to room temperature. Prior to 1996 the firm did not monitor the length of time at which the product was exposed to room temperature conditions during the filling operations
(FAVOO9-FAVO15).
(JLG) Exhibit 8A is SOP 6830- - - — - -
- I~ approved by Quality control on 6/28/94. The 1996 date comes from Exhibit 8B dated June 28, 1996 stating, that temperature exposure was not tracked during the filling of lots FAVOO9, FAVO1O, FAVO11, FAVO12, FAVO13, FAVO14 and FAVO15. Page 2 is the room temperature tracking sheet for the second packaging of FAVO12. Exhibit SC is a similar memo dated October
1, 1996 for lot FAVO16. Although the time at room temperature is recorded, there is no set limit.
(CLK) The firm was cited in both 1993 (Anthrax Vaccine prelicense inspection and bacterial products inspection) and in 1994 (annual inspection, observation specifically in reference to the inspection of filled vials of albumin observed by the inspectors) for not monitoring or recording the time final containers are exposed to room temperature during the filling, inspection and packaging processes. In addition, SOPs 4430-001 —- — —
- ~—-~-~-- - and 5376 —--—~----—-~
- ~—--- —---——-~--do not provide for recording the time ~ne product is exposed to room temperature nor set limits on room temperature exposure (Exhibits 7.1 and 3.1).
Except for information gathered through the stability program, started in 1997, no testing has been performed to determine the effect of room temperature exposure on the product(s) at any stage of processing. For example, deviation report 97DAV34, which addresses the exposure of formulated, bulk lot FAVO36, indicates that the product is "not affected" by the exposure to room temperature (Exhibit SD) . The report appears to indicate that there was no real investigation into the effect of room temperature exposure on the product because there are "no set time limit" for exposure.
9. The firm’s procedures for Environmental Monitoring of critical production areas do not require that additional cleaning and increased sampling be performed when environmental action limits are exceeded. When environmental monitoring action limits are exceeded during filling, investigations do not consider environmental monitoring results during production of subJ.ots, sterility test results of su.blots and sterility test results of final product. In addition, when a sterility retest is performed during stability testing, no investigation is performed. For example:
--
Lot FAVO29 was filled on 8/11/95 and passed sterility testing. On 5/23/57, during stability testingf
it required a sterility retest. The contaminant was identified as P.nicilliwn species. The product passed a retest. Production records for sublotu used to produce PAVO29 indicate -~ ~~-were bulked to produce the lot. Sublot AV383 had an initial sterility failure on 5/17/94 (Rhodococcus species); Sublot AV390 had an initial sterility failure on 7/19/94 (ProDionibacteriuzn acnes). Both sublots
passed sterility retest. There is no environmental monitoring data from preparation of the sublots. On 8/11/95, during filling of lot FAVO29, environmental monitoring testing found the following on critical surfaces:
Cladosiorium species, Alternaria species, Micrococcus species, Bacillus subtilis, Stavhvlococcus saDrophvticus, StaDhvlococcus eDidermidis,
and StaDhylococcus caDitis.--
Lot FAVO32 was filled on 10/26/95. On 7/28/97, during stability testing x’~ V"~- it required a sterility retest. The contaminant was identified as Penicillium species. The product passed a retest. The lot was formulated on 9/21/95. Two -operators performing the formulation exceeded action limits on viable monitoring. Four CFEJ were sampled from one of th. operator’s gloves and identified as Penicilliuzn species.--
Lot 7AV035 was filled on 2/5/97. On 8/11/97, during stability testing ~ it required a sterility retest. The contaminant was identified as Bacillus cereus. The product passed a retest. The lot was formulated on 1/9/97. Environmental monitoring exceeded action limits in the gowning area prior to formulation identifying the following:StaDhvlococcus caDitis, Micrococcus species, Bacillus coaoulans, and Corvnebacteriuzn species. In addition, photohelic gauges were out of range during this time indicating insufficient air pressure in critical areas.
--
The firm does not trend multiple contamination’s with microorganisms in sublots. For example, between 4/94 and 2/95, ----- -~ were produced of which 23 were discarded due to some kind of microbial contamination. Lots FAV029, FAVO3O and FhV031 were whole or partially formulated from those sublots not discarded in this time period. In January and February 1997 of-12 consecutive sublots produced, 5 were discarded for microbial contamination. The others were included in lot FKVO39. And in september and October 1997. of - ~-~-~-produc.d, 6 had contamination and two of those were retested and released for formulation. The remaining sublots were formulated into FAV045 orFAVO46.
(MM) Exhibit 12A is a copy of the Environmental Monitoring SOP.
Selected records for Anthrax Vaccine, lot FAVO29 are attached as Exhibit 4C22.
Selected records for Anthrax Vaccine, lot FAVO32 are attached as Exhibit 4C25.
Selected records for Anthrax Vaccine, lot FAVO35 are attached as Exhibit 4C28.
(CLK) Attached is a sublot list produced by the firm at my request (Exhibit 9A). Due to the large number of contaminated sublots, most of which were destroyed, I attempted to identify the bacterial contaminants. If "unidentified" is written, then there was no identification of the contaminant by the firm. Otherwise, the contaminant is written on the list. Of those contaminants identified between 4/94 - 2/95 and 1-2/97, the majority are "environmental" organisms (gram negative rods) Since many of the contaminations are described on the sublot list as "Contaminated, p~m7oducLiur1 f~rm. sainp. #2", I asked W. White to clarify this description. W. White informed me it meant a contamination was discovered in the sample taken from the ~‘. -fermentation step. I asked W. White if there were any investigations into these multiple contaminations at that step in the process. W. White said there were no investigations.
Of the 6 sublots contaminated in September and October 1997, 4 were contaminated with B. anthracis, one with Bacillus cereus and one identified only as Bacillus species. The firm initiated investigations into these contaminations (97DAV42, 97DAV53, 97DAV49, 97DAV52 and 97DAV64; I did not investigate one of the six lots). It was determined by thc firm that, other than technician error (in reference to the contamination in sublots AV636 and AV637), a change in filters due to the previous filter being discontinued, was the cause of the contamination with B. anthracis in several sublots. All sublots which were made with these filters in place were quarantined until all release testing was finished and QA released the sublots. As an exa~p1e, the deviation report for sublot AV646 is included (Exhibit 9B)
10. Recording of data in building-- from room to log books in room -~ ts accomplished by viewing the results through the UV pass box and is not checked for accuracy prior to discarding the original data.
(CLK) Anthrax vaccine potency testing is performed in building (BL3) . Data collected to determine the actual
strength of the challenge inoculumn (colony counts from the dilutions performed) and relating to the guinea pig deaths on
each day is written on small pieces of paper and taped to the CV pass box so they can be read in room ~‘> and transcribed onto the official records for each potency test. The following day, upon entering the — room, I viewed these papers being discarded. There is no procedure by which a second technician checks the transcribed information for accuracy, nor is it rechecked the next day prior to being discarded.
11. Specifications for the release of subJ.ots were not formally established until 1995.
(JLG) Exhibit 2.a is the current sublot batch record and includes Quality Assurance specifications for sublots on page 39 of 40. I ask~c1 A~ Luttrell when these specifications were first set. He stated November 1995. Sublot batch records prior to that do include testing for -
- ~~j.~but specifications for these were not listed (example, Exhibit llB, batch record for sublot AV85 produced in
1991)
It should be noted that W. White repeatedly informed the team that the - -------~ test listed in the current specifications was an indicator/in-process test and not a release criteria for sublots.
12. The firm does not have a current SOP for environmental monitoring in the Anthrax production facility. The firm has replaced it’s previous environmental monitoring SOP with a centralized procedur. that references area specific monitoring plans. However, the Anthrax specific plan has not been finalized.
(JLG) Exhibit 12A is the current -
‘ SOP 6776.00, Rev. 1, effective 3/27/97. This SOP sets out the general guides for the environmental monitoring program, including action and alert limits for areas, classifications and corrective actions. Throughout, the SOP refers to "area specific plans" for monitoring guidelines and frequencies (i.e. Section 9, page 10 of 23). R. Merriman informed me that the Anthrax building - plan was in draft and not yet in effect.
Exhibit l2B is the firm’s previous environmental monitoring program, SOP 6776, effective date 1/18/96. It includes the monitoring guidelines for the Anthrax vaccine facility.
Anthrax Building Facilities Conditions:
13. In room - q of
- the Anthrax production
facility, we observed peeling paint, exposed duct and pipe work, insulation peeling off the pipes, and rusty steam and gas lines.
(JLG) On 2/4/98 and 2/6/98, CLK and I (JLG) inspected the Anthrax vaccine manufacturing facility on the ~—. floor of building ~y We were accompanied by W. White, N. Suxnrnerton, and
T. Wilsey. A floor plan can be found in Exhibit 12B Attachment
7. The firm had begun decommissioning the facility for planned renovations. Operators were performing pre-shutdown qualifications of equipment.
Throughout the facility, the floors are linoleum tile (apparently unsealed), and the walls are painted brick. All piping, ductwork and lighting fixtures are exposed, either suspended overhead or along the walls. Room -
- The wall paint was chipping and flaking, particularly behind the exposed piping along the walls. We observed chipped paint on the insulation covering a steam line within three feet of the inoculation port, just outside the short vinyl curtain surrounding the laminar flow area over the inoculation port. Insulation on another stretch of steam line was shredding. The gas line delivering compressed air for positive pressure transfers was rusted.
Room --- —- — -- -- — - - -~ System 1 is used as a back-up train. The walls in this room seemed to be in slightly better condition. Some equipment, i.e. the centrifuge, had been removed as part of the decommissioning. Exposed steam drain lines were rusted. As in room . the duct work and piping was exposed.
Room - had already been emptied of all equipment, except the and Chemical hood. This room was used for component
weighing, inoculum preparation, and harvest and formulation of sublots. -
14. Compressed air used to perform positive pressure transfers of sterile products (Anthrax
and Rabies) is central plant air and is not monitored. The ~—~---—‘--~> filters used at the point of. use are not integrity tested.(JLG) Transfers into and out of bottles and fermentors during the production of rabies and anthrax vaccines is accomplished
using positive pressure. These transfers include formulation of Anthrax and Rabies sublots. N. Summerton told us (JLG, CLK) that the compressed air in building -, is "house" air, and when questioned, stated that air was not monitored. She confirmed that the house air system does not use an ~-~---~
The transfer bottles we observed in the rabies facility building included disk air filters. T. Wilsey confirmed that similar
bottles and filters are used in Anthrax production. M. Bengal and T. Wilsey confirmed that the air filters are single use, but that they are not integrity tested.
15. Rooms .- - ,‘v~/Y~, I’y are not environmentally controlled. There was no active environmental monitoring of aseptic manufacturing activities until 1996. During the 1996 production of sublots, more than half of the sublots had environmental monitoring excursions. There was no tracking of these events and no significant corrective action taken until 10/96.
(JLG) Nancy Summerton stated that the general Anthrax vaccine facility was not environmentally controlled. She indicated that it would have been difficult to control the environment considering the limitations of the facility. Exhibit l2B, SOP 6776 attachment 7, indicates that room classifications were suggested but not formalized. As noted in the discussion of Observation 12, the current version of SOP 6776, Exhibit 12A, does not address the specifics of monitoring the Anthrax facility.
I requested environmental monitoring records for the C~ 2~tn
building and was provided with records for 1996 and 1997. I
was informed by R. Merriman that there were no earlier records,
as the firm had not started monitoring until the beginning of
1996. Exhibit l2B - SOP 6776,~
lhas an effective date of January 18, 1996. reviewed data for the L 7~ in room L-. Dfocusing on sublot formulation activities. Out of ( imonitaring days, 16 had excursions over action limits. A total of 6 out of 9 sublot formulation days had C )results. Although deviation reports were written, the corrective actions consisted largely of a stated intention to clean gloves, bottles and equipment more
carefully with _ ~ (Exhibits 15A and 15B).
Monitoring excursions between August 18, 1996 and September 17, 1996 were universally responded to with a memo written on 10/7/96 and 10/10/96 by "DNS" ( Exhibits l5C and l5D). It appears that
the corrective action may have been effective as the rate ofK 3 results is greatly reduced for the rest of 1996 and all of 1997.
16. Plant steam is used for sterilizing production vessels and glassware in buildingsJ 2~and is not monitored or controlled.
(JLG) According to N. Summerton, the autoclaves in building ~ j1are supplied with plant steam from a central boiler. When
asked, she stated that they did not routinely monitor the quality of the steam. She said she believed they had tested it once several years ago from a single port. She stated that they do not use any boiler additives.
17. Poor facility arrangements exist for aseptic processes in building C.. ~room r 3m that media is made, dishes washed, equipment and glassware autoclav.d, a.s well as the production processes of fermentation, inoculation, and harvest all occur in this one room simultaneously.
(JLG) A floor plan can be found in Exhibit 12B, attachment 7. Room r jis a multipurpose room with little segregation of activities. For example, the’~ - used for inoculum preparations and aseptic sublot formulations is within two feet of the sink where glass washing occurs and within six feet of the autoclave, which sits on the floor in the room (it is not built-in) . We (JLG, CLK) asked if there was an SOP controlling the activities in this room. T. Wilsey stated that they did not have one. He went on to explain that each of the fermentor trains had separate teams of operators. When questioned further, he stated that there were no controls that would prevent different teams from performing incompatible operations (i.e. aseptic collection of the precipitate from the harvest vessel and inoculation) at the same time.
18. Regarding cold storage of critical seed stock:
a. In-the Anthrax production suite the logs for the refrigerator/freezers Care incomplete. The
logs do not match the refrigerator/freezer contents. The Anthrax refrigerator/freezer contained unlabeled vials.
(CLK) JLG and I inspected the Anthrax production facility in bui1ding~ ~ -~ floor with W. White and T. Wilsey. We inspected the refrigerator/freezers where either Anthrax or
materials are stored, as well as reviewed the logs for these (Exhibits lS.a.l and l8.a.2). The refrigerator/freezers have been moved recently in preparation for renovating the production facility, although they remain on the second floor of building L 2 Both Anthrax Vaccine and are campaigned in this facility. Both refrigerator/freezers were connected to the alarm system. I was unable to match the log books for each refrigerator/freezer to the contents. The majority of the contents were sparsely labeled, as in just a sublot number and no dates or further description. There were boxes of old sera, that were hand labeled but not dated. There were unlabeled vials (looked like media only) in a rack labeled "anthrax spore suspension" in the anthrax refrigerator. In the anthrax freezer there were unlabeled glass vials with rubber stoppers that was possibly sera. In general, both refrigerator/freezers contained items that were not involved in the production of Anthrax Vaccine or — — - —------— and/or that were unknown as to their content.
b. There is no segregation of the master spore concentrations and the working spore concentrations of both the virulent in building ~ind j. 2 strains in building ‘~ Z~in both Anthrax production and potency testing facilities.
(CLK) In both the Anthrax potency testing facility (building and the Anthrax production facility (building , one refrigerator in each facility contains both the master and working spore concentrations. There is no separation within the refrigerators of the similar spore concentrations (master versus working) . Also, all of the master spore concentration for production ~‘ ~ strain) can be found stored in one refrigeratorin building ~ 2and all of the master spore concentration for challenges in potency testing (virulent strain) can be found stored in one refrigerator in building ci A
c. The keys for all refrigerator/freezers in building Ci D (second floor) and building Lilwere found on top.
(CLK) When JLG and I inspected the refrigerator/freezers on the second floor of buildings Dthe keys to open them were retrieved from the top of the refrigerator/freezers. i questioned as to whether that was the normal storage place for the keys. T. Wilsey said that when the refrigerator/freezers are in their normal location on the second floor, there is a desk close by where the keys are kept. The desk is unlocked and all who work
in the production facility know where the keys are located. To enter to the second floor of building\ > L- is required.
When I (CLK) inspected the ‘— ifacility, building I - ,~/ room ½he keys to the refrigerator which contains both the master
and working virulent spore suspensions was located on top of the refrigerator. I questioned M. Pierre as to whether that was where the keys were always located. M. Pierre said that was the keys normal location. There are two keys to the refrigerator in room -~ -~and both were on top. To enter building ~- -.
~,is required.
19. There is no SOP for change over in building _ Anthrax Biosafety Cabinets (BSC). Both inoculum preparations and aseptic sublot formulation occur in these hoods.
(JLG) T. Wilsey and D. Slabbekoorn informed me that the BSCs in building 2e used for both inoculuxn preparation and for post harvest and formulation operations. They also stated that these operations could occur on the same day. I questioned them regarding the change-over procedures. Both men, separately, explained that they cleared the equipment and washed the BSC down
with r ~, I asked D.
Slabbekoorn for a change-over procedure either as a separate document or as part of another SOP or batch record. The firm was not able to supply a change-over procedure. It is not included in the BPR (Exhibit 2.a).
BABIES:
26. The manufacturing process for Rabies Vaccine is not validated.
- -
Media fills have not been performed to validate the aseptic preparation of sublots and hold times prior to filling. Th. product does not pass through a ? ~micron filter during production.(MM) Product is passed through a~_ Jfilter during production. There is no sterilizing step in the manufacture of Rabies vaccine. There are no records documenting the qualification of the aseptic processing.
(JLG) There was discussion regarding whether the
in the product is a sterilant. A. Luttrell stated that MBPI’s position was that the is not functioning as a sterilant in the product, but as an inactivant. He stated that they consider the Rabies production to be an aseptic production process that would need appropriate validation of sterility assurance including media fills.
- -
There has been no determination or identification of environmental organisms present during the manufacturing process.(MM) According to R. Merriman, the firm has not established a baseline of environmental organisms present in the manufacturing facility.
- -
There is no validation of the length of time sublots are held until they are used in a formulation. Sublots have been held as long as 14 months prior to use. There is no stability data to support this hold time.(JLG) Exhibit 26B is the sublot inventory log for rabies sublots. There is no official expiry date for the sublots. The log includes listed expiry dates. W. White stated that expiry dates were arbitrarily assigned to sublots for a period of time, but as the dating period was not based on data, the practice was discontinued. Examples of sublots that have been held for extended periods: RSL837 was placed in storage on 8/7/96, and was removed on 11/21/97 for use in lot RV16O. Sublots RSL839,
RSL847, and RSL85O were placed in storage on 8/7/96, and was removed on 1/21/98 for use in lot RV161.
- -
The autoclave used to sterilize glassware used in the preparation of sublots and formulated vaccine is supplied with plant steam and has no vent filter.(JLG) There are ~ ~yautoclaves in room ~jj - ~ One was labeled as not in use awaiting validation. M. Bengal stated
that they were using the - 2 autoclave, Sterilizer _ - and
confirmed that this was the autoclave used for sterilizing rabies production glassware. NM noted that the autoclaves were of an old style and asked if they had vent filters. N. Suminertofl stated that she did not believe so. We (JLG, MM) could not identify a filter on the unit. N. Surninerton also stated that the autoclaves were supplied with plant steam.
--
Glassware used in production is not depyrogenated. In addition, glassware/tubing assemblies used for harvesting the product are autoclaved and then placed on a shelf for up to7A. days. This storage time has not been validated.(JLG) MM and I observed the ~1ass carboys with tubing and filter assemblies, used for rabies production, lined up on a table in room~ iy These were labeled with an expiry date. When questioned, M. Bengal stated that they autoclave the glassware and tubing assemblies and then give it a <. ~ day expiry. We asked if there was validation to support the expiry date. M. Bengal responded that they had arbitrarily selected that time period. Condensate was noted in the bottles. Exhibit 26A - SOP 1365
- -
WFI supplied to building ci ~ used to prepare sublots and rinse glassware, is filled from the point of use in building C)into glass bottles and transported to building~ These bottles are used, rinsed with tap water, rinsed with
WYL from another transport bottle, dried in~an incubator at c. 2. covered with blue paper wrap, and placed on a shelf until filled again. There is no assurance this water meets WFI specifications.(JLG) M. Bengal explained the use and transport of WFI in building to MM and I. She stated that the water was collected and used within 24 hours. We asked if they had validated that the water retained its WFI quality throughout this period.
Management answered that they had not.
-
- The analytical method for determination of-4in Rabies vaccine has not been validated.
(MM) The preservative used in Rabies Vaccine is C According to W. White, the analytical method for determining
_in the product is not validated
--
Reagents used in the preparation of the product are prepared in building C ii and assigned an expiration date of 6 months. There is no documentation that the bottles used to Store the reagents are sterile when filled, nor has the expiry period been validated. Potency testing is not performed on the reagents.(JLG) Reagents used in the preparation of media and buffers were observed by MM and myself stored in room ~ When questioned,
M. Bengal explained that the reagents were assigned a 6 month expiry date without any validation to support this dating period. Example of reagents: & —, prepared 10/3/97, expires 4/3/98; prepared 9/30/97, expires 3/30/98; ii prepared 11/25/97, expires 11/25/98.
27. The specification for potency is the geometric mean of _ test results and shall be J —~ This allows out-of-specification results to be averaged with in-Specification results to obtain a passing final average. In addition, potency tests have been conducted more than three times, three results selected and a passing averag, reported with no justification for the additional testing. For example:
-Rabies Vaccine, lot 150, was tested for potency 13 times in October and November, 1995 (2.06, 3.08, 2.62, 0.94, 1.01, 0.92, 1.86, 1.74, 1.94, 0.92, 2.52, 1.93, 5.09 lU/mi). The geometric mean of three results (2.06, 3.08, and 2.62) were determined and a potency result of 2.6 lU/mi reported.
-Rabies Vaccine, lot 152, was tested for potency 6 times in March, 1996 (5.09, 3.18, 1.86, 2.27, 1.40, 3.49 lU/mi) with a geometric mean of 2.6 lU/mi reported.
-Rabies Vaccine, lot 158, was tested for potency 5 times in July and August, 1997. (0.84, 4.45, 3.40, 2.73, 1.68) Two of the results are marked "Invalid test" (0.84, 1.68). There is no documentation justifying the
invalidation of these two results. A geometric mean of 3.46 lu/mi was reported.
Exhibit 27A - Testing records for lot RV15O.
Exhibit 27B - Summary report and testing records used for the
potency calculation for Lot RV15O.
Exhibit 27C - Summary report and testing records for Lot RV152
Exhibit 27D - Summary report and testing records for Lot RV158.
Exhibit 27E - Sop 6144 < - -
(RL) The rabies virus vaccine potency test is a live animal test requiring many animals and numerous variables which contribute to a valid test. It is the industry standard for potency and therefore it is important to justify and document the need for multiple rounds of testing. It is also important to document invalid test results.
28. SOP QAO2-Mol-01, dated 10/11/97, states ~sainple will be placed on stability each year. Two samples were placed on stability in 1996 and no samples for 1997.
Lots 152 manufactured 3/7/96 and 153 manufactured 4/26/96 were placed on stability (zero time) on 8/27/96.
The initial stability test for Lot 152, initiated 9/6/96, is classified "Invalid" with only 2 of 3 results reported (2.11 lU/mi; 1.69 lU/mi), with no investigation. The results were not reviewed until 11/12/97. The 3 month stability test results (11/26/96) were reported 11/12/97 (1.12; 4.99; 3.14 lU/mi).
The initial stability test results for Lot 153, initiated 9/6/96, are 2.11 lU/mi. The 12 month stability test results, dated 8/28/97, are listed as "invalid", and reported 12/4/97 (1.86;
2.03). There was no investigation.
Exhibits 28B and 28C
- Summary stability charts for lots RV152and RV153.
-Exhibit 28D
- Summary report, and testing data for "Time Zero" stability testing for Lot RV152. This includes a memo dated 9/12/96 stating that the mice were mixed up for the final test set, with a recommendation that the test be invalidated. The results were not reviewed until 11/12/97.Exhibit 28E
- Summary report, and testing data for three month
stability testing for Lot RV152.
Exhibit 28F
- Summary report, and testing data for twelve month stability testing for Lot RV153.(RL) of rabies virus vaccine per year is put on stability testing. Stability testing consists of potency and
- content testing at~ ~months
for stability testing is appropriate.
MBPI has two lots of rabies virus vaccine on stability from 1996
-
lot RV152 and RV153. RV152 had an invalid potency test at time zero and has then tested above the standard of 2.5 IU/ml at months 3, 6, 9, and 12. This lot will be tested at 18 months starting in February, 1998. RV153 had a test result of 2.11 at zero time (below the standard of 2.5 IU/ml) and an invalid test at twelve months. This lot will also be tested at 18 months starting in February, 1998. No lots were put on stability during 1997. Lot RV159 has been put on stability for 1998.It is understood throughout industry and at FDA/CBER that the rabies virus vaccine potency test is difficult. It is a live animal test which is dependent on many variables. However, it is still the standard used to measure potency and so is an important part of a stability program. As described above, record reviews for potency tests for lots RV152 and RV153 on stability were done as much as a year later. Invalid or below standard test results during stability testing must be investigated and verified at the time of testing and repeated if necessary.
29. On 9/16/97 samples of Rabies cell lines were sent to
—-
. for evaluation of their abilityto form colonies in soft agarose. There is no documentation of review, particularly by Senior Quality management of the results received from. this company which were inconclusive. There is no written justification for the continued use of this cell line in production. Thor. are no written procedures for the notification of Senior Quality management of product rejections, investigations
and failures.(RL) See discussion of qualification of cell lines. Given the potential for transforming viruses in these cell lines, I informed the firm of my concern that the positive result reported
by~- )was not investigated more
thoroughly. A discussion with MBPI management addressed this
issue and they stated that they would further check the characteristics of these cells. Exhibit 29A - Batch records for lots RV1E5, RV156, and RV157 sublots that indicate altered morphology of the cell lines. Exhibit 29B - Report from&
iDindicating the soft agar growth.
30. Rabies Vaccine sublots RSL9Q2 and RSL9O3 contained errors, yet these batch records were released by QA, without any corrections.
(RL) In reviewing the batch records for sublots RSL9O2 and RSL9O3 errors were detected in the notebook (Exhibits 30A and 30B) . These errors were not noticed when the notebook was reviewed by QA. We had a discussion with MBPI management regarding this and they have tightened the control on QA review.
31. Cells ~batch 66, used for rabies manufacturing (i.e.
lot 155 to current) have not been qualified.
(RL) This is a repeat observation for the past two inspections. In discussing this with W. White, he presented me with data
from -"on the qualification of
a new 2cell bank and of a new working cell bank. As discussed above, MBPI is in the process of validating these new cell banks and will then submit a supplement to FDA/CBER to include these new cell lines in their manufacturing process. MEPI was encouraged to work quickly on this submission.
Diphtheria-Tetanus Vaccine, Lot DT4176:
32. There is no environmental monitoring data for production, formulation and filling of the product in 1994.
(CLK) In general, environmental monitoring was not performed at MBPI at the time lot DT4176 was produced. A. Luttrell said the environmental-monitoring program did not really begin~until after the 1995 inspection. There was documentation in the BPR for lot DT4176 of sterilization runs performed in the autoclave and viable monitoring by the use of settling plates and £ ii plates during the filling operation.
33. There are no finished product specifications for release of Diphtheria-Tetanus vaccine.
(CLK) The firm was unable to find documentation as to the release specifications for DT. A. Luttrell informed me that there were no "formal" release specifications, however I was provided with an unofficial copy of bulk product and filled lot specifications for Diphtheria and Tetanus Toxoids, Adsorbed (Exhibit 33.1).
During the discussion with management, R. Myers stated that he was sure these specifications existed and he would submit them with the firm’s response.
34. There is no procedure for zedating of the vaccine including release testing specifications.
(CLK) I inquired as to whether the firm had an SOP f or extending the dating of the DT product. I was informed by A. Luttrell that there was not an SOP for extension dating of this product. However, I was provided with "Release Protocol, Extension of Dating" for lot DT4176-1 (Exhibit 34.1). The release protocol for lot DT4176 from 1994 is attached to the release protocol for lot DT4176-1 from 1997. The potency assays were repeated in 1997 for both diphtheria and tetanus. The diphtheria component failed the first potency test. Both the diphtheria and tetanus components were retested and both passed. A deviation report was filed, 97DDTO1 (filed 10/22/97), blaming the failing potency test for the diphtheria in the first test on the small size of the serum pool at the day ~bleed (because i~ guinea pigs died as a result of the day C ~bleed) (Exhibit 34.2). CBER released lot DT4176-1 on December 11, 1997 based on the potency test initiated on July 18, 1997.
35. Diphtheria-Tetanus vaccine lot DT4176 is not on atability.
(CLK) There are stability programs for diphtheria and tetanus toxoids separately, however, there is no stability program for the combination vaccine.
36. The analqtical methods for c - )is not validated.
(NM) The preservative used in DT vaccine is < According to W. White, the analytical method for determining
~‘
in the product is not validated.37. c > testing and preservative effectiveness testing have not been performed on DT vaccine.
(MM) According to R. Merriman, j. ~3 testing and preservative effectiveness testing has not been performed.
Blood Derivatives Products:
38. Training of employees performing examination of incoming plasma does not include examples of defects including: excess red blood cells, hemolysis, microbial contamination, and examination of product labels.
(PMS) Review of the firin’~ SOP 3101.3,
• -
(Exhibit 38), revealed that incoming plasma can berejected for such reasons as: excess red blood cells, hernolysis, microbial contamination, positive for leukemia and syphilis. I asked Ms. Simon if they have examples of units that the examiner uses to pass or reject incoming units of plasma. Ms. Simon informed me that they do not have examples of units that the examiner can compare to during the inspection. Ms. Simon informed me that if the examiner notices anything unusual they will bring
it to the attention of the supervisor.39. Media fills do not represent all product manipulations after sterile filtration. Immune Globulin and Albumin (Human) are sterile filtered in building~ ~ and may be stored in bulk for up to— i days prior to being moved to buildingc i~for filling. There has been no media challenge of this storage period.
(MM) Exhibit 39 is a copy of SOP 5458.000, 5
stating that a sample from the bulk tank is taken and tested for sterility. It also states that if the standing time exceeds ~ days, a second sample will be drawn by manufacturing and sent to QC for sterility testing. If the results are acceptable the product may proceed to filling.
Media fills conducted by the firm have been performed in building 1 in the filling suite only. There is no media fill data to support a holding period in buildingC C
40. There is no SOP defining the actions to take, including when the Quality Unit is to be notified, when the following situations occur in production:
a. the sterile filter fails in building
b. the filling line is stopped and product must be stored
and moved to another
tank in building ~c. when the product requires refiltration.
(MM) During the inspection, I reviewed the firm’s deviation reports, three of which are attached as exhibits, and involve Albumin (Human). Deviation report 97DHA02, dated 2/15/97 involves filling line problems. The line was stopped and the remaining bulk was ref iltered (Exhibit 40A). The second report, 97DHA04, dated 7/1/97, involves an incident where the sterilizing filter collapsed and the product was ref iltered (Exhibit 40B). The third report, 97DHA14, dated 11/5/97, involves a failure of the filling pump with the bulk tank being placed in the cold room and filling continuing the next day (Exhibit 40C).
The following SOPs, provided by the firm as coverina the incidents that occurred, are attached: 2404, ~-
- (Exhibit 40D) ; 7137, ii~Exhibit 40E) . SOP 2404 does
not address when a product may or may not be ref iltered and when it is necessary to contact the Quality Unit prior to ref iltering the product. SOP 7137 indicates that a new number must be given to a lot that requires a change in fill. It does not address placing the product back in storage, with limits on the amount of time it may be stored, and when refiltration is required, nor when the quality unit should be notified.
During the inspection of 11/96, the firm was cited for inadequate investigations. As part of their strategic plan, a new system for deviation reporting is to be in place in the next few weeks.
I explained that in reviewing these 3 deviation reports, it appeared production was not notifying the Quality Unit of problems that may impact on product quality. In addition, despite changes in their deviation reporting system, the two SOPs that would cover these incidents do not require the completion of a deviation report, nor notification of the Quality Unit. I pointed out that it appeared the production unit simply ref iltered the product involved in the filter collapse (97DHA04) on 7/1/97 and did not notify the Quality Unit, since the memo covering the incident is dated 7/9/97.
41. There has been no investigation into out of specification pressure differential readings in building ~
»
for the time period of 6/97 to 1/98. Thecurrent SOP 2303.300 for monitoring air pressure differentials
does not require notification of the Quality Unit when differentials are 005.
(JS) The firm has established a log to record pressure differential photohelic gauge readings. These gauges were verified in place during inspection of the area. Review of the log showed that some readings were outside of the established specifications in )of the areas he-ma mcrnThcred (daily
recordings) between rooms ~ — -- -.
~at various times from June 1997 to January 1998 (Exhibit 4lA). The out of specification readings were indicated by notation as to action taken or referral but some areas continued to be out of specifications. The firm’s current SOP 2303.300,
i~indicates reporting and work order requirements but does not include the Quality Unit in this process (Exhibit 41-B). The firm’s new version of SOP 2303.300 draft includes a new step ("F" under "V. Procedure") which specifies the issuance of deviation reports for out of specification readings which would then be investigated with Quality Unit input (Exhibit 41-C).
42. The analytical method for determination of L ~in Immune Globulin has not been validated.
(NM) Exhibit 42 is a copy of the specifications for Immune Globulin, including
~L .7~ which is the preservative in the product. According to the firm the method for analysis of 1. Tim Immune Globulin is not validated.43. - . testing for Immune
Globulin was performed on product with a 100 ppm Thimerosal. The specification for Thimerosal in the product is ~_ ~ppm. There is noQ~ 7)testing at 85 or 115 ppm. There is no information that the antimicrobial effectiveness of Thimerosal is effective through the expiration date of the product.
(MM) ~ Thtesting was performed on
12/23/96 on I~nmune Globulin lots 126, 127~, 128 (Exhibit 43) These lots were analyzed and contained 100 ppm of Thimerosal. The product specifications indicates the range for Thimerosal is to be between 2 7711 There is no data forZ Z~ testing at these levels in the product.
44. SOP 5431, ~
does not include reporting incidents that may occur while
transporting tanks and vials between building -~
(DL) Filling, incubation, inspection and packaging of lots are performed in buildingl_ After the vials of Albumin are filled, they are transported from building~ to building ,. 5After review of SOP 5431,
~iExhibit 44), it was discovered that the SOP made no provision for reporting of any deviations pertaining to possible accidents or unscheduled incidents during the transport of the vials and or filled tank from one building to another.
45. Contract testing labs are used for HBsAg, HCV, HTLV-1, HIV1/2 anti-HBc, HIV-1 Ag testing. Samples of Fraction I supernatant or pooled Recovered Plasma are sent to the lab for analysis. No qualification of the contract lab, nor tests performed has been conducted.
(MM) Exhibit 45 is a copy of testing performed for MBPI by -~ including the tests listed above. These tests are indicated for use in testing units of donor blood, specifically the plasma or serum portion of blood intended for transfusion.
The firm has no documentation demonstrating the validity of test results from testing samples of Fraction I supernatant or of pooled Recovered Plasma. According to A. Luttrell these tests are being performed as an industry standard.
46. A Climet is used to monitor temperature and humidity of production rooms in building Calibration documentation of this instrument indicated it cannot be calibrated because it lacks accuracy of measurement. No other instrumentation is used to monitor these parameters in these areas.
(JS) The firm uses Climet model § -- instruments to monitor particulate counts and temperature/humidity room conditions in those production areas that are both classified and unclassified for buildings 2~ The Climets provide continuous monitoring with print-outs maintained with the monitoring records. Examples of these print-outs are attached to the forms entitled, "In Process Airborne Particulate Counts (Climet, Met One)", Exhibit 46-A. The Climets are calibrated on an annual basis by the outside firm -
i2~zith calibration certificates on file. I reviewed the most recent calibration certificates for the five Climets used by the firm in buildings ~ i4ith serial nos. 967082 and 967068 identified as used in building md serial nos. 967083, 967059, and 967077 for buildingç
- ~See Exhibit 46-
B). These certificates showed calibration of the particulate counts but not the temperature/ humidity measurements. The certificate contained a statement that the temperature and relative humidity readings could not be controlled during calibration due to the wide operating range of the unit
(~.The
firm relies on the Climets formonitoring of room conditions.
General:
47. The procedures (6108 and 6107) for routine sampling of the
WFI loop at a point of use (POU) require a
.. §ininute flush ofthe POU prior to collecting the sample. There is no requirement
f or this flush prior to production (building
(JS) SOPs 6108 and 6107.020 for sampling and testing of WFI systems in buildings 17 require a .ninute flush at each use point during sampling. See page 4 for each of these SOPs, attached as Exhibits 47-A and 47-B. SOP 6108 has an attachment
3", which describes cleaning of outlets containing tubing, but does not specify the flushing of outlets prior to production use (Exhibit 47-C). Management indicated that flushing is normally performed prior to production use but there is no written directive for this flush with time specified (at least the same duration as sampling time) and no documentation that flushing is performed.
48. Pressure differential readings are recorded in each batch record. There is no written procedur, for monitoring pressure differentials, including instructions for follow-up of pressure differentials that are out of limits in buildingL
(JS) During review of batch records for filling and packaging of Immune Globulin and Albumin lots made in 1997, I noted that the records contained readings of the Photohelic gauges that showed air pressure differential readings between these rooms. The batch records contained the established limits for these areas but the firm has not established a procedure for the monitoring and follow-up activities if any out-of-limits are encountered. did not observe any readings that were outside of the indicated air pressure differential specifications shown in the batch records.
49. No annual review of BPR5 (repeat observation).
(CLK) The firm was cited in the 1994 inspection for this same observation. The batch oroduction records for Anthrax Vaccine Adsorbed, ~ Rabies Vaccine Adsorbed, and Diphtheria-Tetanus Toxoids Adsorbed, Lot DT4176 contained a variety of errors including simple transcription errors, missing records and records from the wrong lots. A careful, annual review by QA is needed to correct the majority of these errors. This was discussed with A. Luttrell and he agreed with our assessment and said that they had really been trying to get a document room organized and were working on reviewing BPRs for all their products.
50. Training of employees performing visual inspection of finished product containers does not include examples of types of particulates and discoloration they are to look for. Finished containers are not held up against a black and white background during inspection. In addition, there is no requirement that employees performing the inspection demonstrate their ability to detect defects.
(MM)
As stated in Observation 7, the firm does not have a SOP for disposition of a lot when the number of discarded finished product containers is high. In addition, there is no specification for an unacceptable number of finished product discards in a product lot.Exhibit 7.1 is a.copy of SOP 4430.001,4—
) Exhibit 50B is a copy of SOP 4432,C,
) Exhibit 50C is a copy of SOP
SOP 4431, C
These SOPs indicate reasons for rejecting finished product containers, however, the training of employees inspecting these containers does not include examples of the defects they are to look for. According to Rajean Potter, employees are given a training session during which they read the SOP, however no examples of product defects are demonstrated. Exhibit 50E is a copy of a training document for Visual Inspection of Filled Vaccines, dated 9/2/97. R. Potter stated employees are trained on the job, explaining that during training, all products are reinspected by a second person, who is instructed to notify their supervisor if they find anything unusual. However, there is no assurance they will see all defects during this training period, nor is there documentation of the results of the supervisory
review of their inspection. No trending is performed.
During the inspection we went to the inspection room and R. Potter explained the inspection process. The inspectors inspect vials on two hour shifts. Only one inspection station was in the room and R. Potter explained the containers to be inspected were shaken and then held up against a piece of paper. The paper was gray in color and hung from the lamp at the station. I explained that usually we see a background of both white and black to assure a contrast when looking for particulates in a container.
51. According to the firm’s SOP 6430,
raw materials that do not have an expiration date will be automatically assigned a jz~irear expiration date by the firm. The firm does not have data to support a 4- Dyear expiration date on raw materials including:
a.
1
used in the manufacture of Rabies vaccine.b.
_ 22 used in themanufacture of Anthrax vaccine.
c.
22 used in themanufacture of
d.
~ -- ~—-- —- 4-~ used inthe manufacture of Plasma DerivativeS.
This SOP does not indicate when a supplier will be requalified.
This procedure does not address raw material monographs updates. For example, during certification of the vendor for
tJSP, in 1989, the
firm did not perform microbial testing. The USP XXIII requires microbial testing of this raw material which has not been performed by the firm.(PMS) R~vjew of the firm’s SOP 6430, ~ -
,Exhibit 51A), revealed that incoming raw materials which do not have an assigned expiration date by the manufacturer, are automatically assigned a year expiration date by the firm. F. Masters provided me with a list of raw materials that are used in the manufacturing of Anthrax Vaccine, C
-~
Rabies Vaccine, and Blood Derivatives (Exhibits 5lB-51E respectively). I asked F. Masters if they have data to support the C j)year expiry dating. F.
Masters informed me that they do not have data to support the year expiration dates. I asked F. Masters how did they come to the decision to label products with a~
F. Masters informed me that some of their suppliers of various other raw materials will assign~
- ~year expiration date so they adopted that practice for Jmflcoming raw materials that do not come in with an e