fb tracking

Testimony on Silicone Gel Breast Implants

Sidney Wolfe, M.D.
Public Citizen’s Health Research Group

Meeting with Acting FDA Commissioner Andrew von Eschenbach, M.D.
Concerning Silicone Gel Breast Implants

More than 17 years ago, in the fall of 1988, I received a plain brown envelope in the mail containing many internal memos written by FDA scientists in CDRH. Among the safety concerns, in addition to an alarmingly frequent occurrence of implant rupture, was the fact that “Silicone gel-filled breast implants all allow slow diffusion of gel through the silicone elastomer shell into the surrounding tissues… Many investigators have reported on the finding of silicone in lymph nodes and an associated lymphadenopathy in women who had undergone mammoplasty with silicone… by placement of a gel-filled prosthesis.” Because of these and other safety concerns, emanating from FDA physicians and scientists and well-documented case reports of rupture and local/regional complications, in November 1988 we petitioned the FDA to ban silicone gel breast implants.

Seventeen years later, earlier this year, in a paper published in the American Journal of Surgical Pathology by pathologists from the Armed Forces Institute of Pathology and Case Western Reserve School of Medicine, lymph nodes from 96 women who had had silicone gel breast implants removed were examined in several ways for the presence of silicone and for abnormal lymph node tissue (silicone lymphadenopathy – pathologic changes in lymph nodes involving silicone). Although there was evidence of implant rupture in only 47 (49%) of these women, lymph nodes from 86 women (90%) contained droplets consistent with silicone. None of the lymph nodes from a control group of 12 women with breast surgery mainly associated with breast cancer but without silicone gel implants contained silicone droplets. Similarly, microscopic examination of the lymph nodes revealed abnormal cells, “foamy macrophages,” occurring much more frequently in women with silicone gel breast implants than in the control group. In the lymph nodes of 91 of the 96 women (95%) with implants, there were these abnormal “foamy macrophages” sometimes replacing normal lymph node tissue, but in the non-implant women, only 4 of the 12 (33%) had these cells, and they were described as “rare.”[1]

The authors commented about the clinical significance of silicone lymphadenopathy: “The possibility that silicone frequently migrates to regional lymph nodes of patients with patients with silicone gel breast implants must be considered in assessing the safety of these medical devices …This phenomenon [silicone lymphadenopathy] does provide a rational basis for suggesting that silicone gel implants may have an impact on the immune system.” Because of these concerns, the investigators who published this study told the FDA that they did not think the silicone gel breast implants should be approved.

2004 FDA Guidance

I would like to spend a few minutes going through the inexplicable agency reversal following FDA’s issue of the 2004 guidance to silicone gel breast implant manufacturers:

This 2004 FDA Guidance set out the standard which implant manufacturers were supposed to follow in order to demonstrate a reasonable assurance of safety. The guidance was very basic and straightforward, stating that manufacturers must produce data that:

  1. Adequately describe the rate of change of implant rupture over the lifetime of the product;
  2. Describe the incidence of gel migration resulting from ruptures; and,
  3. Characterize the health consequences of ruptures and associated migration.

The key point was that the applicant must supply data sufficient to project over a long term.

FDA staff determined that there was a need for a study that includes routine MRI data because 86% of ruptures are silent and can only be detected by an MRI. The standard set forth in the Guidance was clearly not met. Because of limited time, we will mainly focus on rupture rates. Within not too many months after the Guidance, obviously too soon to have collected acceptable long-term data, both Inamed and Mentor submitted their short-term data, such as it was:

Inamed – In the FDA staff summary, it was pointed out that with only 2 data points, it was difficult to make a 10 year projection. The FDA staff, in extrapolations from these two data points, had made the following analyses that were influential in causing the advisory committee to vote against approval of the Inamed devices:

FDA Extrapolations of Rupture Data from the Core Inamed Study

Table 10: Probability of Implant rupture thru 10 years for MRI cohort (silent and symptomatic) for 3 models (from pg 39, FDA Review of Inamed Data)


Constant Hazard

Linearly increasing

Quadratically increasing













Indications combined




Mentor provided even less data, only two years of data, of even less value than Inamed and, because of only one data point, it was not possible for the FDA staff to make the extrapolations they had made with the Inamed data.

So the best that can be said about the Inamed and Mentor rupture data is that we do not know how high the rupture rate is. But Inamed and Mentor, by the terms of FDA’s, Guidance had the obligation to demonstrate that their products have a sufficiently low risk of rupture over the lifetime of the product to provide a reasonable assurance of safety, as required by the law. They have not done this and yet FDA has sent both companies approvable letters, in an apparent refutation of the sound Guidance issued in 2004.

Continuing Concerns (as in 1988) by FDA Scientists:

As you know, over the years I have had many disagreements with the FDA. But here, and in most cases, my disagreement is not with the FDA scientists. Here some of your own scientists have concluded that:

  • The data presented by the industry was “insufficient”, “inadequate” or “non-existent” to determine a reasonable assurance of safety.
  • Long term data was not provided: Inamed had just barely four years and Mentor only presented with two years of complete data

Sharpe-Collis MentorStudy

One of the studies heavily relied upon by Mentor for asserting that their implants are safe is the so-called Sharpe-Collis study, essentially a review of one British plastic surgery practice. Despite assurances of safety, this study was thought by some FDA reviewers to be seriously flawed in the following ways:

  • This was a retrospective, not a prospective study, reviewing augmentation records from 200 implant patients. 
  • The researchers invited them in for screening but only 101 participated
  • The data relate only to augmentation patients, and generalizability to the reconstruction and revision indications is not possible.
    • The data describe a case series from a single surgeon
    • MRI’s were performed based on voluntary participation on only a subset of the patients from this one surgeon’s practice, resulting in possible selection bias.
    • The data exclude patients with implant removal, surgical intervention, clinical evidence of rupture or capsular contracture (who are at higher risk of having rupture).
    • The data describe the point prevalence of rupture rather than rupture rate over time.

FDA staff determined this study to be of limited value and for good reason: 

Because only half of the eligible women participated in the MRI; because of the differences in patient characteristics, implant type, implant placement compared to the U.S. Core Study; and because only one [rather than serial] MRI was performed, the ability of these data to predict a long-term rupture rate or a rate over time is limited. (FDA Panel Transcript p. 173 -175)

Because this study could not possibly meet the standard of reputable scientific inquiry, the study is inadequate from a regulatory perspective.

NCI / Inamed / MentorStudy Using NCI SEER Data

The following chart is are derived from data in the still unpublished SEER study, funded by Inamed and Mentor, using NCI surveillance data. The most striking finding is that the Kaplan-Meier calculated risk of implant removal, 21% at five years after implantation for silicone gel breast implants, had risen to 50.5% by ten years – more than doubling – for women with breast cancer who had had a silicone gel breast implant. In other words, by 10 years more than half of these women had their implants removed. This compares with 40.7% for those women getting saline implants. This is the only population-based study – with the important advantages of better ascertainment – to examine the problem of implant removal. It is noteworthy that of the 6563 women with breast cancer in the geographic areas covered by the study, only 1159 (17.7%) had breast implants.


In summary, if these silicone gel breast implants were prescription drugs, instead of medical devices, there is no possibility they would be approved, knowing how inherently defective they are. If, in violation of the FDA’s legal standard of reasonable assurance of safety, the agency decides to approve silicone gel breast implants, more than 17 years after internal FDA memos voiced serious concern about these safety issues, these devices will clearly be the most defective devices ever approved based on what was known prior to approval. We hope you reverse this extremely unwise decision and rescind the approvable letters to these two companies.

[1] Katzin WE, Centeno JA, Feng LJ, Kiley M, Mullick FG. Pathology of lymph nodes from patients with breast implants: a histologic and spectroscopic evaluation. Am J Surg Pathol. 2005 Apr;29(4):506-11.