Date: Fri, 6 Aug 1999 22:03:10 -0700

This very interesting letter to the FDA from Dow has been sent to us by Linda Dintino. It is a Discovery CD document image that has been converted to OCR text. The more I see these "discovery documents" the more I shake my head in disbelief that so much evil has been perpetrated upon implant victims. Thanks Linda for sending our way.

Dow Corning Corporation Medical Products Plant

1635 N. Gleaner Road

Hemlock MI 48626

Telephone: (517)642-5224

Fax: (517)642-8249

April 9, 1992

Mr. Carl C. Reynolds, Director

Detroit District

U. S. Food and Drug Administration

1560 East Jefferson Avenue

Detroit MI 48207

Dear Mr. Reynolds:

This letter responds to the FDA Form 483 issued on March 4, 1992 following the inspection by investigators Nancy Mundo, James E. Szelc, Diane Prince and others. The form noted the following observations:

1. During review of more than 1500 Device Customer Complaint files, approximately 850 complaints from 1985 through 1990 on silicone gel mammary implants that involved reports of serious injuries, potential serious injuries or product malfunctions have not been reported under the Medical Device Reporting Act.

Dow Corning disagrees with the investigator's observation that approximately 850 complaints received between 1985 and 1990, which were not reported under the Medical Device Reporting ("MDR") involved serious injuries or product malfunctions that were required to have been reported. FDA's current disagreement with our previous judgments appears to reast on application of an entirely new interpretation for evaluating the reportability of breast implant complaints. This new "standard" has been incorporated into a document entitled "MDR Reporting Guidance for Breast Implants" which was provied to Dow Corning by FDA the week of February 10. 1992, virtually at the conclusion of the inspection on which this 483 Obeservation was based.

Throughout this period and to the present day, Dow Corning believes that it has substantially complied with FDA's MDR regulation. The results of previous FDA inspections confirmed our beliefs.

As part of the GMP inspection of our Hemlock facility in 1988, FDA specifically made an assessment of our compliance with the medical device reporting regulation. Enclosed is a copy of a six page questionaire which was attached to the Establishment Inspection Report at the time of that inspection. This questionaire reported our procedures for investigating and evaluating complaints, and our decisions about shether or not complaints were reportable under the MDR regulation to be fully adequate and in compliance.

Further, the CDRH had additional access to our complaint information. A log listing of all breast implant complaints from 1985 through 1988 was provided to the Office of Device Evaluation in January, 1989 in response to a specific request from ODE. We, therefore, felt confident that the FDA understood that our interpretations were that a number of the complaints did not meet the criteria for reporting under the MDR. It was not until the 1992 inspection that a contrary opinion was provided to us.

MDR Guidance purports merely to "interpret" the existing MDR regulations to assist manufacturers in identifying product complaints specific to breast implants that are reportable under the MDR regulation. Generally, we believe that there are significant discrepancies between the MDR regulation and our interpretation of policies that appear to be mandated by the guidance document. Our questions and concerns regarding the MDR Guidance document are being addressed in a separate communication to the CDRH. We will copy you on that letter. We will be requesting clarification of FDA's understanding of, and intent relating to, the Guidance document.

2. Complaint investigations are inadequate in that there is no documentation of attempt to obtain information on incomplete complaint reports including: product identification and Lot numbers, patient history and reason for removal, follow up on post operative patients or other information that would assist in complaint evaluation.

For example, the following complaints with "rupture" claims: MW3784, MW3793, MW3805, MW3820, MW3974, MW3976, MW4024, MW4391, MW4392, MW4405.

All complaint details are not normally provided by the complainant with a complaint file is opened. Since the patient (or current physician) may not possess a file listing the manufacturer, model, size, lot numbers, and or catalog numbers, these details may be unknown. During subsequent follow up, more data may become known. A procedure will be developed to document pertinent complaint data supplied and or document attempts to obtain additiona complaint data.Complaint information is not always forwarded immediately to the Dow Corning Complaint Coordinator so that ivestigations can be conducted and MDR reports filed in a timely manner.

a.) Complaint MW2580 Silastic II High Profile Round 175cc, Lot HH054029, where an implant involved in severe capsular contracture was returned to the firm on 9/22/87. The court claim was settled by Dow Corning in August 1998. The complaint information was received by the Dow Corning Hemlock Complaint Coordinator on 12/19/88 and the MDR report phoned in on 12/22/88. The Work Record for the Complaint Analysis is dated 3/16/89 and the return letter to the physician was sent 4/4/89.

In 1989, we adopted a policy that requires complaints such as these to be forwarded by Fax directly from DCW Customer Service group to the Dow Corning Complaint Coordinatior.

b.) Complaint MW3795 was recieved in Hemlock on 3/11/91. The implant was a Silastic II Mammary Implant 150cc, Lot HH034810 and was reportedly implanted from 6/24/86 to 12/22/87. There is no "Nature of Complaint" recorded other than the physician questioned a bubble in the implant on removal. The reason for removal is not reported. There is no copy of the Complaint Analysis Report or Work Record attached to the file. However, the

response to the physician is dated 2/8/91 and the Dow Corning Wright Designated Investigator apologizes for the delay in responding to him. The actual date of the complaint receipt is not reported in the file.

This was more a product inquiry than a complaint. Follow up with DCW indicates the explant occured in 1987 due to size change on one side. At the time of surgery, no complaint was resgistered and the patient took the intact implant home. In November 1990, the patient noticed bubbles in the implant and became concerned. The unit was sterilized on 11/27/90 and returned to the sales representative in early December 1990. The device was returned toDCW 2/7/91. DCW investigated and responded on 2/8/91. The apology was conveyed because the three month time lag exceeded the time the sales representative promised the patient.

The labeling which accompanies the product includes Surgeons's Note stating:

"Voids may occasionally be observed within the gel of the Silastic II Mammary Implants. These voids may form during the manufacturing process and consist of FREON(R) (dichlorodifluoromethane), the carrier gas used for sterilization. Data is on file that demostrates that the gas in not cytotoxic. These voids are harmless to the patient and generally dissipate in time. The use of the product that contains voids within the gel does not endanger the patient or compromise the safety of the implant."

4. The failure investigation was inadequate for lot HH106486 where gel bleed was found spotting outer surface of both unsold units pulled for the shelf by Dow Corning Wright Designated Investigators during follow up of Complaint MW2387. The investigation was not completer to determine if these units were representative of the entire lot and whether the gel bleed was excessive or of an "acceptable" amount. There is no documentation of the name of the individual responsible for the decision that further investigation was unnecessary.

On November 7, 1988, the Dow Corning Wright Designated Investigator reported the analysis of two pre-operative units returned for investigation. Quoting the TESTING and FINDING sections, the following is recorded.

"TESTING: The units were subjected to bleed testing by being placed anterior surface down on brown Kraft paper in a 300 degress oven for two hours. This testing found no deposition of silicone bleed on the Kraft paper.

FINDINGS: the units were found to be clear, not cloudy, and to have a sticky film on their surfaces. The substance found on the implants was determined to not be a conswquence of gel bleed. The substance had an appearance similar to that from normal body fluids and powder from surgical gloves."

On March 16, 1989, the Dow Corning Wright Designated Investigator reported on two shelf samples from the lot which were examined. The examination disclosed:

"Visible gel residue was found to be deposited grossly on the posterior around the patches as well as potting the entire surface of the envelopes."

In light of this observation, the Designated Investigator prepared a report of the findings, including a diagram depicting the locations of the external gel residues, and submitted the report (as an "internal complaint") to appropriate Manufacturing and Quality Assurance personnel for additional investigation.

The Designated Investigator also offered a recommendation that the barrier coating procedure be reviewed by all parties involved.

Upon reciept of the report and recommendation, Hemlock Quality Assurance personnel evaluated the shelf samples and the relevant records and specifications. Based on this review, the QA person found the samples to meet specifications. The QA report of the investigation noted:

"Our barrier coat procedure is performed according to the mutually agreed upon SOPs. Product is approved based upon Lot Acceptance Requirments."

In view of the thorough multi-level investigations and analyses performed in follow up to this complaint, we are uncertain of the basis for the Observation that the failure investigation was "inadequate." In our view, the complaint investigation was very thorough in that the returned units as well as shelf samples were investigated and assessed against the current specification. The final assessment asserts the product meets thos specifications.

5. Production operators and Quality Assurance inspectors have signed for supervisory/release approval personnel over their own work. The operator signed the supervisor's name, then initialled the signature.

For example: Lot #'s HH071348 (7/30/91), HH031155 (5/7/91), HH057686 (6/11/87). On Lot #HH027686, the inspector appears to have signed her name and the authorized signature without initialing the other person's name.

Our written policy prohibits inspectors from signing the supervisory/release approval for their own work. Personnel are being directed immediately (and retraining will occur within two weeks) to comply with this written policy. Further, confirmation that personnel do not release their own work will become one of the requirements for the final release signature.

6. Final review of Lot History Records and release of the lot is not documented, by signature, in the lot history record.

Lot History Records were reviewed to assure the lot should be released. This review was documented by rubber stamping the date received, but a signature was not added. We have begun adding the release signature with the stamp.

7. Reasons for rejects are not consistently recorded in the Lot History Records when in process units are rejected at specific production steps. For example: Lot HH111571. The Lot History record review indicates that 40% of the total lot was rejected at the "wash & pack" stage prior to the final inspection of the product. There was no documentation of the reason for the rejects. The total rejection rate for this lot was 78%. There is no documentation of a Quality Assurance Investigation of this lot. The remaining implants were released for distribution.

Reasons for rejects were recorded for in process steps that are most likely to produce useful information. Reasons for rejects were not recorded at the "wash & pack" stage, but we were going through a transition to record in process rejects through the use of a PAR (Production Assurance Record) form. This was scheduled to be applied in first quarter 1992. Finally, it should be noted that none of these processes are currently in use nor will they be resumed following our announced departure from the gel filled implant manufacture and distribution.

8. Label Usage and Control forms do not always contain accurate and complete information. Mathematical errors were noted both in Packaging and Q. A. verification of totals.

For example: Lot #HH061324. The packaging count does not match the label print quantity. The Q. A. tally of units Rejected and Destroyed or Unused, only records unused labels.

Lot HH071348. The Q. A. verification of Label Control in final packaging was not completed. The lot was released for distirbution.

RESPONSE: The intended use of the Label Use and Control form appears to be misunderstood, since packaging operators are not returning damaged labels to the inspector and the inspector is not reconciling the error. This misunderstanding will be corrected by retaining during the next two weeks. Further, proper completion of the form will be confirmed prior to final release of the lot.

9. No documentation is maintained for temperature and viscosity of the envelop(sic) dispersion that is monitored by the Dipping Area personnel. There is also no documentation for time and pressure of vacuum(sic) chamber used during the gel filling/deairing process.

We believe that documentation of the dipping process temperature and viscosity is not required. The temperature operating limits which are regularly checked in the calibration program have alarms set for the upper and lower boundaries so that the operator is notified of an out of control situation. The viscosity is controlled by a set point controller at the desired viscosity. The operating set point is established at the start of each lot of dispersion and rechecked with weekly lab samples. Vacuum chamber time and pressure were inadvertently not maintained because use of the process was optional per the SOP. Finally, it should be noted that none of these processes are currently in use nor will they be resumed following our announced departure from gel filled implant manufacture and distribution.

10. Process validation of equipment used in intermediate and final manufacturing procedures has not been conducted since the inital installation/qualification of the equipment. Initial qualification documentation of the equipment was not available for review.

Review of more than 500 Lot History Records disclosed wide variations in lot to lot rejection rates.

a) Two (2) envelope dipping units, installed in 1973 and 1975, had rejection raes that ranged from 2-100%

b) Patch placement unit, installed in 1976, had rejection rates that range from 0-54%

c) Rotocoater, installed in 1981, had rejection rates that ranged form 0-57%

d) Gel filler, installed in 1971, had rejection rates that ranged form 0-88%

As is clearly evident in the Observation, the processes involved for which the reject rates are quoted are old processed (in existence for 10 or 20 or more years). The state of the art in manufacturing control and engineering was clearly not the same when these processes were first utilized as that which is avialable today. Futher, our standards are such that we frequently reject product for appearance criteria, unrelated to product safety or performance.

That further explains the apparent relatively high rejection rates. Finally, it should be noted that none of these processes are currently in use nor will they be resumed following our announced departure from gel-filled implant manufacture and distribution.

11. No quantitative limits for pressure applied during patch seal replacement operation have been implicated in gel bleed from the mammary envelopes.

We believe that knowing the precise pressure utilized for patch seal placement is not of significant value. The ram pressure is adjustable in order to accommodate the needs of different types of product (e.g., single lumen, double lumen).

The operators adjust the pressure to achieve the desired final result, not a predetermined process level. Further, setting limits around this value range would not improve the product quality.

Finally, it should be noted that none of these processes are currently in use nor will they be resumed following our announced departure from gel filled implant manufacture and distribution.

12. Incoming raw material product specifications have not been formally approved for ETCH (1,ethynl 1,cyclohexanol), utilized in mammary envelope production as a cure inhibitor.

RESPONSE:

Raw material product specifications do exist for ETCH and were in effect. A copy of that specification, dated 03/Sep/86, was provided to Investigator Mundo during the course of the recent GMP inspection. The specification has been approved and is being currently utilized to accept or reject incoming raw material lots. We have not fully incorporated the documentation format into the system because we are continuing to work with the vendor to obtain vendor agreement to test and to certify on a lot to lot basis.

13. Corporate Technology Changes, Good Manufacturing Practices Compliance Manual, Appendix 6, in addition to specific information regarding formal and approvals, also contains the form to be used in the procedure. The current approved revision date is 6/1/91. Technology Change forms, currently in use at this facility, are dated 3/1/88, indicating the revised form was not implemented for use in a timely manner.

Immediately upon being alerted by your investigators to this fact, the change was implimented and we are currently using the 1991 dated form. The technology change form was changed in only minor ways between March 1, 1988 version and the June 1, 1991 version. Because of the extensive involvement of the Hemlock quality assurance staff in the effort to prepare the breast implant PMAAs for submission, the form change was inadvertently not implemented.

14. Bioassurance method EM-001, Environmental Monitoring Program, General Information, dated 6/1/90 references supervisory review, via signature, of data generated by the technician performing and monitoring a specific test. Current review indicates no supervisory signature or review is documented.

The bioassurance supervisory review was conducted consistently. However, the supervisor failed to document that review by signature as required by our SOP.

Immediately upon observation that the required signature to document the review was not being undertaken, the bioassurance supervisor began to implement that required procedure and her review is currently being documented by signature.

If you have any further questions, please do not hesitate to contact me.

Very truly yours,

Kenneth L. Montague

Quality/Supply Manager

cc: Harvey L. Steinberg, Dow Corning

Nancy N. Mundo, FDA

******

This document was OCR'd (converted to text) from Discovery CD #119 DCC. It consisted of 10 images, DCC 010000253 thru DCC 010000263, combined into one document to make it easier to read.