Dr. Zuckerman review of PMAs

Date: Sun, 19 Sep 1999 05:57:25 -0600

Dr. Diana Zuckerman was the Professional Staff Member of the House of Representatives, Human Resources and Intergovernmental Relations Subcommittee of the Committee on Government Operations.

I have found a Sept. 12, 1991 letter from her to Ted Weiss, outlining her preliminary review of the FDA's decision to file the PMAs for the manufacturers. She was not complimentary to the manufacturers in this letter. The letter copy is difficult to read on all three copies I found. It is impossible to scan them into the computer the scanner just won't read them. I have typed the majority of the letter below. It is four pages long.

Quotes from her letter to Ted Weiss.

"The statisticians that reviewed the clinical studies conducted by McGhan, Mentor, and Bioplasty explicitly recommended that the PMAs for each of these companies not be filed because of major flaws in those studies. And yet, all of these PMAs were accepted by filing by FDA. Despite the subcommittee request for all documents regarding the PMAs, we have not received any written justification that explain why the scientists' recommendations were overruled. In fact, there is no written justification of any kind regarding why the seven PMAs were filed by the FDA. This is unusual, within every agency of the HHS there is usually a written justification for any decision of this importance. Since we have been assured by the FDA that the subcommittee has received all documents on this matter, we have to assume that there are no FDA documents supporting the filing of any PMAs for breast implants."

DOW CORNING. In an August 12 memo to the file, the Task Leader of the FDA Breast Prosthesis PMA Task Force wrote that the Dow Corning clinical studies are 'so weak that they cannot provide reasonable assurance of the safety and effectiveness of these devices' because the provide 'no assurance that the full range of complications are included, no dependable measure of the incidences of complications, no reliable measure of the revision rate and no quantitative measure of patient benefit.' In his detailed criticism, he specified that the physicians who conducted the research were instructed 'to report only complications associated with the implant. As a result, the only complications reported are those at the implant site. This prevents those investigating from detecting systemic adverse effects or complications resulting from implantation of the devices.' He also stated this 'causes and underestimate of both types and incidence of complications.' Furthermore, each patient was examined only once after surgery and the number of patients examined at each time point is very small' making it difficult to determine the rate of complications at any point in time.

McGhan. The statistician who reviewed the McGhan PMA pointed out many major problems in their studies of women. For example, in the McGhan prospective clinical study, 10% of the 118 patients were not evaluated at the time they were discharged, and only one third of the implants were assessed at the second required visit (3-6 months). The statistician pointed out that this lack of follow-up makes it impossible to draw any conclusions about long-term safety or effectiveness. In addition, only three reconstruction patients were in the study, making it impossible to draw any conclusions about their experiences. The statistician reported that the company's 'historical cohort study' suffered from 'strong potential for bias' and was therefore of no use in providing support for safety and effectiveness. An FDA biologist pointed out that the company studied only two of the four models listed on the PMA. This obviously makes it impossible to determine safety or effectiveness for the two 'multi lumen' models (made from saline and silicone) that were to be studied. Moreover, only 39 reconstruction patients and 101 augmentation patients were studied, which isn't enough to determine problems (even fatal ones) that affect a small percentage of patients. In addition, many potential medical problems, such as breast disease or carcinoma, were not evaluated for all patients. My review of that PMA indicates that two-thirds of the women included in that study had prostheses implanted in 1989 or 1990, therefore the study could not under any circumstances assess long-term risks.

Bioplasty. Similarly, a statistician reported that in the study of 260 patients with Bioplasty's MISTI single or Double lumen implants, only 6% of the patients were assessed at the 2 year follow-up, and yet the company calculated their claims of safety and effectiveness as if they had followed large numbers of patients for two years. There were only 21 reconstruction patients, which importantly, the company stated that the physicians conducting the study refused to all the company to contact the patients to ask about immune disorders or cancer, 'fearing that it may cause undue concern or violate patient confidentiality.' The company blamed the media, saying 'it created an environment in which gathering that information was, at best difficult.'

Mentor. There were three PMAs submitted by Mentor, for three different types of implants. This in itself is interesting, since most companies submitted one PMA for several types of implants. The studies for the three types of implants were identical, and the statistician criticized the applications for failing to include important information, such as when patients were assessed subsequent to surgery, or whether appropriate steps were taken to avoid bias in the study. My review of the application reveals that the (number is illegible) patients in one study were apparently evaluated on the basis of the medical records, which did not necessarily provide any long-term information. For a second study, 128 women of those (number is illegible) patients were interviewed on the telephone to evaluate their satisfaction with the implants. The 128 women represent 37% of the patients who were selected for the interview: it is therefore impossible to draw any conclusions about patient satisfaction based on that sample. In a third study by Mentor, 273 augmentation patients were included in a retrospective study of complications, but the information available was for an unspecified time, and based on available medical records of the plastic surgeon. Since such records would not be expected to include information on auto immune disease or cancer, this study is inadequate in the safety information it provides.

Conclusions. When new drugs or devices are introduced onto the market, the number of patients evaluated is necessarily small. However, in the case of breast implants, there is a 30 year history involving more that 2 million American women. It is hard to understand why the companies, which have known since at least 1982 that they would probably be required to provide safety data, and which were warned more than three years ago that data would definitely be required, waited so long before they started conducting major studies. In fact, many of the studies were started in 1990 or 1991. Although prospective studies that followed women for many years would have been considered ideal, a reasonable alternative would be to start a study in 1990 that asked patients from the 1970's or 1980's about any medical problems they have had since their implant surgery. That kind of thorough retrospective study was not conducted by any of the manufacturers.

In summary, there are several major problems with most of the studies.

1. Most do not study women for more than two years at the most; this is not sufficient to evaluate the safety of a medical device that is meant to be permanent, especially when allegations have been made that they are likely to rupture after several years.

2. In many cases, the majority of women are lost to the study after a few months; it is therefore impossible to say whether an implant is safe if there is no information at all on most of the women who had the surgery.

3. In several studies, there are no relevant questions regarding immune disorders, cancer, or other possible risks that have been associated with implants. It is not enough to look at medical records kept by plastic surgeons, since women will only return to their plastic surgeon for complications that they recognize to be associated with the surgery.

4. The number of reconstruction patients in most of the studies is so small, that they could not provide persuasive evidence of safety. Robert Sheridan informed me that for the purposes of filing, FDA assumed that the experiences of augmentation patients would be the same as those for reconstruction patients. That assumption is impossible to defend, since there is no data to back it up.

5. Several manufacturers have no studies of women with certain models of implants that they sell, or they have studied fewer than 10 women with particular types of implants. Again, Robert Sheridan told me that for the purposes of filing, the assumption was made that the safety of one model was the same as for other models. Again, that assumption is impossible to defend, since there is no data to back it up.

It well take years to provide meaningful long-term safety and to determine the average lifetime of breast prostheses that have been implanted in women. In some cases, well designed studies are planned but have not yet been started."