Correspondence
Breast Implants and Connective-Tissue Diseases
N Engl J Med 1994; 331:1231-1235November 3, 1994
- Article
To the Editor:
As rheumatologists who have examined and evaluated a combined total of more than 3000 symptomatic women with silicone breast implants, we feel compelled to point out the shortcomings of the study by Gabriel et al. (June 16 issue).1 This study addresses only classic connective-tissue diseases, such as systemic lupus erythematosus, rheumatoid arthritis, Sjogren's syndrome, and scleroderma, and appropriately concludes that it is unlikely that common rheumatic diseases, such as rheumatoid arthritis, are associated with implants. The sample size is inadequate to assess the frequency of uncommon rheumatic diseases, such as systemic lupus erythematosus or scleroderma, and the study methods did not allow the assessment of the prevalence of Sjogren's syndrome.
We, like many others, believe that many symptomatic women with silicone-gel implants have a new and unique rheumatic syndrome characterized by chronic fatigue, myalgia, polyarthralgia, cognitive dysfunction, symptoms similar to those of the sicca syndrome, rashes, and neurologic disturbances. This disease was first described 20 years ago by the Japanese as “human adjuvant disease” and more recently has been labeled “atypical connective-tissue disease.” The occurrence of this cluster of symptoms is not addressed in the study by Gabriel et al.
Although their study involved a useful cohort for investigation and included an appropriate control group, the study method, which was limited to a retrospective chart review, allows few conclusions to be reached. None of the patients were examined or questioned by the authors, and the patients were not subjected to a standardized evaluation, such as a questionnaire capable of eliciting information about rheumatic symptoms including symptoms of sicca syndrome; a uniform physical examination; or standardized laboratory studies, including tests for autoantibodies. Few if any of the patients were examined by a rheumatologist. The laboratory data that are presented are incomplete. Table 4 of the article by Gabriel et al. shows that 11 of the 749 patients with implants had positive tests for antinuclear antibody, but the number of patients tested is not stated.
We believe that a controlled prospective study investigating the association between silicone breast implants and atypical connective-tissue disease is needed. Such a study can be performed only after preliminary criteria for diagnosing the disease have been established on the basis of a careful analysis of signs, symptoms, and laboratory features in symptomatic women who have been evaluated and examined by board-certified rheumatologists.
Pending definitive studies, we believe that the Food and Drug Administration (FDA) is justified in continuing the ban on silicone-gel implants. The burden of proof should be on implant manufacturers, to provide evidence of safety, not on physicians who treat symptomatic women, to provide proof of injury.
1 ReferencesGary Solomon, M.D.
Hospital for Joint Diseases Orthopaedic Institute, New York, NY 10003Luis Espinoza, M.D.
Louisiana State University Medical Center, New Orleans, LA 70112Stuart Silverman, M.D.
UCLA School of Medicine, Los Angeles, CA 900481
Gabriel SE ,O'Fallon WM ,Kurland LT ,Beard CM ,Woods JE ,Melton LJ III . Risk of connective-tissue diseases and other disorders after breast implantation. N Engl J Med 1994;330:1697-1702
Full Text | Web of Science | Medline
To the Editor:
Both Gabriel et al. and Angell, in her accompanying editorial (June 16 issue),1 point out the limitations in the study of connective-tissue diseases after breast implantation, but they do not discuss a major methodologic problem: the selection of controls. Women who had undergone a medical evaluation within two years of the date of breast implantation in the case subjects were the controls in this study. Some of these control subjects may have sought medical attention because of symptoms presaging connective-tissue disease. The case subjects sought medical attention primarily for breast implantation. The different reasons for seeking medical attention may falsely elevate the incidence of connective-tissue disease among the control subjects and make it more difficult to evaluate the possibility of an increased risk among the case subjects. Once again we are reminded of the problems encountered in the choice of control subjects in clinical studies.
1 ReferencesPaul C. Royce, M.D., Ph.D.
Monmouth Medical Center, Long Branch, NJ 077401
Angell M . Do breast implants cause systemic disease? Science in the courtroom. N Engl J Med 1994;330:1748-1749
Full Text | Web of Science | Medline
To the Editor:
Gabriel et al. found no association between silicone breast implants and connective-tissue disease or any other disorders. Their study, although one of the most comprehensive to date, failed to take into account the population of women with documented ruptures of their silicone-gel implants. The authors assumed that the type of implant was not important because all the implant materials were contained in a silicone envelope; thus, all 749 case subjects were exposed to silicone elastomer. This assumption is false, because silicone-gel breast implants are composed of three physically distinct types of silicone (elastomer, gel, and fluid),1 which have been shown to have different levels of biocompatibility.2
Our own laboratory has shown that only silicone gel is a very potent humoral adjuvant capable of inducing the production of autoantibodies in rats, whereas silicone fluid, which has a much lower molecular weight, and the solid silicone elastomer sheeting possess minimal adjuvant properties.3,4 Therefore, clinical data from the population of women who are exposed to free silicone gel, from either ruptured implants or free injection, need to be compared with data from a control population before a conclusion can be reached about an association between breast implants and disease.
4 ReferencesJohn O. Naim, Ph.D.
Raymond J. Lanzafame, M.D.
Rochester General Hospital, Rochester, NY 14621Carel J. van Oss, Ph.D.
State University of New York, Buffalo, NY 142141
LeVier RR ,Harrison MC ,Cook RR ,Lane TH . What is silicone? Plast Reconstr Surg 1993;92:163-167
Web of Science | Medline2
Picha GJ ,Goldstein JA . Analysis of the soft-tissue response to components used in the manufacture of breast implants: rat animal model. Plast Reconstr Surg 1991;87:490-500
CrossRef | Web of Science | Medline3
Naim JO ,Lanzafame RJ ,van Oss CJ . The adjuvant effect of silicone-gel on antibody formation in rats. Immunol Invest 1993;22:151-161
CrossRef | Web of Science | Medline4
Naim JO ,van Oss CJ ,Lanzafame RJ . The induction of autoantibodies to thyroglobulin in rats with silicone gel as adjuvant. Surg Forum 1993;44:676-678
To the Editor:
Gabriel et al. were unable to detect an association between breast implants and a number of connective-tissue diseases. As they appropriately point out, however, the study had inadequate statistical power to assess the association between implants and many of the disorders in question -- in particular, systemic sclerosis. This is important, because scleroderma is the connective-tissue disease most strongly suspected to be associated with prior exposure to silicone.1,2
We were also surprised by the small number of women in their study who were identified as having any rheumatic disease. On the basis of previously published data on the incidence of rheumatoid arthritis among women in Olmsted County, Minnesota (the county where the women studied by Gabriel et al. resided),3 one would have predicted 12 new cases of rheumatoid arthritis in the total cohort of 2247 women followed for an average of approximately eight years. Alternatively, on the basis of prevalence data from Olmsted County,3 one would have predicted 30 cases of rheumatoid arthritis in that cohort of women. (It is unclear whether prevalence or incidence figures are appropriate; the authors do not specifically state that women with recognized connective-tissue disease at the index date were excluded.) In fact, only two cases of rheumatoid arthritis were detected by the authors. One wonders whether the sensitivity of the study design was even lower for the diseases that are more difficult to diagnose.
3 ReferencesRobert F. Spiera, M.D.
Allan Gibofsky, M.D., J.D.
Hospital for Special Surgery, New York, NY 10021-4898Harry Spiera, M.D.
Mount Sinai Hospital, New York, NY 100291
Sanchez-Guerrero J ,Chur PH ,Sergent JS ,Liang MH . Silicone breast implants and rheumatic disease: clinical, immunologic, and epidemiologic studies. Arthritis Rheum 1994;37:158-168
CrossRef | Web of Science | Medline2
Spiera RF ,Gibofsky A ,Spiera H . Silicone gel filled breast implants and connective tissue disease: an overview. J Rheumatol 1994;21:239-245
Web of Science | Medline3
Linos A ,Worthington JW ,O'Fallon WM ,Kurland LT . The epidemiology of rheumatoid arthritis in Rochester, Minnesota: a study of incidence, prevalence, and mortality. Am J Epidemiol 1980;111:87-98
Web of Science | Medline
To the Editor:
Our group has seen 20 patients with scleroderma-related disorders after augmentation mammaplasty, whereas Gabriel et al. were not able to find any such disorders among 749 women. This discrepancy can be explained by several methodologic flaws and misunderstandings.
First, the literature that Gabriel et al. cite on autoimmune disease in women with implants (references 3 through 12) suggests an interval of 8 to 10 years between implantation and the onset of disease. The women studied by Gabriel et al. were followed for a mean of only 7.8 years.
Second, none of the patients were interviewed or examined; chart reviews were performed not by physicians but by nurses. Medical records often do not include musculoskeletal or constitutional symptoms if the patients are not directly asked about them. Many of the records we have reviewed do not mention breast implants. Many of our patients admitted that they had never mentioned the implantation to their doctors; some had not even told their husbands. Most of our patients with scleroderma had only outpatient, office-based surgical records, and since they were young and healthy, they did not seek medical attention for 5 to 10 years, until rheumatic symptoms developed. Such patients with “proto-rheumatic disease” would have been absent from the Mayo Clinic data base for a long time.
Finally, the onset of systemic lupus erythematosus, scleroderma, and some other autoimmune processes tends to decrease after menopause. We would venture to say that the Mayo Clinic cohort was older than our group of patients in Los Angeles (mean age at the onset of scleroderma, 39.5 years), which included a much smaller number of patients with breast cancer and many people in the entertainment industry.
Daniel J. Wallace, M.D.
Eileen Schwartz, M.D.
UCLA School of Medicine, Los Angeles, CA 90024To the Editor:
Gabriel et al. describe a timely and well-conducted study. The results regarding morning stiffness and serositis, however, need clarification. There were statistically significant increases in the rate ratios for morning stiffness and serositis among the women who received breast implants after mastectomy, as compared with the community controls. The authors state that the increased risks were more likely to be related to the underlying cancer than to the implants, because the rates in patients with cancer who had implants were similar to those in patients with cancer who did not have implants. However, no data or analyses are presented. Given the importance of the results, it is essential to provide all the relevant data so that readers can make their own assessment. In this regard, it is also essential to discuss the potential problems of using the relatively small group of 306 control subjects with breast cancer as the basis for comparison.
In discussing the literature on connective-tissue diseases and breast implantation, Gabriel et al. are overly critical of several previously published case-control studies.1-3 In particular, Gabriel et al. criticize these studies for their case-control design, small size, and low statistical power. Although, in general, cohort studies may be more desirable, they may not be appropriate for rare diseases, as is evident from the authors' own discussion of the sample size required to detect an increased risk of scleroderma. In fact, the retrospective cohort study by Gabriel et al. offers no data on two major connective-tissue diseases: scleroderma and rheumatoid arthritis. Case-control studies may be the only practical alternative when the disease is rare.4,5 The case-control design has long been accepted as a valuable tool by epidemiologists.
I agree with Gabriel et al. that the previous case-control studies were relatively small and that their statistical power may have been low. For these reasons, no definitive conclusion can be drawn from the individual studies. However, the consistency of the findings of these studies is reassuring. For rheumatoid arthritis in women with breast implants, Dugowson et al.1 reported an odds ratio of 0.41, and Goldman et al.2 reported an odds ratio of 0.76. For scleroderma, Goldman et al.2 reported an odds ratio of 0.0 (i.e., none of the women with scleroderma had breast implants), and Wigley et al.3 reported no increase in the risk (no odds ratio was given). In addition, two recent studies on scleroderma and breast implants were not cited by Gabriel et al.; Hochberg et al.6 reported an odds ratio of 0.31, and Burns7 reported an odds ratio of 0.61.
Although each of these studies was relatively small, the consistency and the totality of the data are quite impressive. The appropriate epidemiologic approach to deal with rare diseases and small individual studies is to summarize the data through the use of a meta-analysis.8 On the basis of a meta-analysis of the studies by Dugowson et al.1 and Goldman et al.,2 the odds ratio for rheumatoid arthritis in women with breast implants is 0.70 (95 percent confidence interval, 0.36 to 1.40). Similarly, a meta-analysis of the studies by Goldman et al.,2 Hochberg et al.,6 and Burns7 indicates that the odds ratio for scleroderma in women with breast implants is 0.45 (95 percent confidence interval, 0.15 to 1.32). Judging by the 95 percent confidence intervals, the meta-analyses based on the combined data have sufficient statistical power to rule out an increase in risk as small as 40 percent for rheumatoid arthritis or scleroderma.
Thus, these case-control studies, together with the study by Gabriel et al., provide strong evidence that there is no association between breast implants and connective-tissue diseases.
8 ReferencesOtto Wong, Sc.D.
Applied Health Sciences, San Mateo, CA 944011
Dugowson CE ,Daling J ,Koepsell TD ,Voigt L ,Nelson JL . Silicone breast implants and risk for rheumatoid arthritis. Arthritis Rheum 1992;35:Suppl:S66-S66 abstract.
Web of Science2
Goldman JA ,Lamm SH ,Cooper W ,Cooper L . Breast implants are not associated with an excess of connective tissue disease (CTD). Arthritis Rheum 1992;35:Suppl:S65-S65 abstract.
Web of Science3
Wigley FM ,Miller R ,Hochberg MC ,Steen V . Augmentation mammoplasty in patients with systemic sclerosis: data from the Baltimore Scleroderma Research Center and Pittsburgh Scleroderma Data Bank. Arthritis Rheum 1992;35:Suppl:S46-S46 abstract.
CrossRef | Web of Science4
Wong O . A practical guide for non-epidemiologists. Occup Health Saf 1981;50:31-39
Medline5
IdemA practical guide for non-epidemiologists. Part 2. Occup Health Saf 1981;50:21-26
Medline6
Hochberg MC ,White B ,Steen V ,Medsger TA ,Weisman M ,Wigley FM . The association of augmentation mammoplasty with systemic sclerosis: preliminary results from a case-control study. Arthritis Rheum 1993;36:Suppl:S71-S71 abstract.
CrossRef | Web of Science7
Burns CG. The epidemiology of systemic sclerosis: a population-based case-control study. (Ph.D. dissertation. Ann Arbor: Department of Epidemiology, University of Michigan, 1994.)
8
Wong O ,Raabe GK . A critical review of cancer epidemiology in petroleum industry employees, with a quantitative meta-analysis by cancer site. Am J Ind Med 1989;15:283-310
CrossRef | Web of Science | Medline
To the Editor:
The article by Gabriel et al. states that the study was supported by grants from the Plastic Surgery Educational Foundation and the National Institutes of Health. Although technically the Plastic Surgery Educational Foundation was the conduit for $147,000 of funding, in reality, half the funds were from manufacturers of silicone breast implants (Dow Corning, McGhan, and Bristol-Myers-Squibb), and the other half came from the American Society of Plastic and Reconstructive Surgeons, both industry groups with a financial interest in the continued sales of silicone breast implants.
The 1994 budget of the Plastic Surgery Educational Foundation has been verified by Dr. Mary McGrath, treasurer of the foundation, in sworn testimony1:
Mr. Levine: I'm sure the papers go into more detail, but it appears that Dow Corning gave at least half a million dollars to the [foundation] for [silicone] studies. Does that jog your memory?
Dr. McGrath: I think it's higher.Clearly, this reflects a pooling of funds and payment under an alias. This commingling of funds constitutes scientific money laundering.
1 ReferencesAaron M. Levine
Plaintiff's Steering Committee, 1320 19th St. NW, Washington, DC 20036Author/Editor Response
The authors reply:
To the Editor: Solomon et al. and Wallace and Schwartz allude to patients in their specialty practices. We performed a population-based, controlled study specifically to avoid the biases (e.g., due to referrals or an absence of population controls) inherent in observations such as theirs.
Solomon et al. criticize us for not including a “new and unique rheumatic syndrome” in our study. This cannot be done until consensus is reached on the diagnostic criteria for such a syndrome. We examined not only the classic connective-tissue diseases but also 14 associated clinical signs, symptoms, and laboratory values, along with several related autoimmune diseases, such as Hashimoto's thyroiditis. Solomon et al., Spiera et al., and Wallace and Schwartz reiterate the limitations of our retrospective cohort design, which we fully acknowledged in our article. Although a multicenter, prospective study with over 10 years of follow-up would be more definitive, not only would such a study be expensive and time-consuming, but it would be virtually impossible to achieve unbiased ascertainment of the outcomes emphasized by Solomon et al. (e.g., chronic fatigue and myalgias).
Royce's concern about our method of selecting controls is misplaced. The control pool included over 95 percent of the population of women, and most of the controls were seen for routine care, primarily for annual pelvic examinations or mammograms. Connective-tissue diseases were no more common in the controls than in the population generally.
We agree with Naim et al. that data on the complications of implantation, including not only rupture but contracture, leakage, and infection, are needed. We have recently completed an analysis of these results (unpublished data). During the preparation of that analysis, we noted that two of the case subjects, who were included in the group of women with cosmetic implants, should have been included in the group with prophylactic implants. Since none of the events of interest occurred in either subject, all that was required was a minor adjustment in the person-years of follow-up. All calculations involving the total case group, the control group, and the cancer subgroup, as well as all conclusions, remain unchanged.
The calculations of Spiera et al. of the expected incidence of rheumatoid arthritis among our study subjects are incorrect, because their calculations are based on a study that included all cases of probable, definite, and classic rheumatoid arthritis, according to the 1958 diagnostic criteria of the American Rheumatism Association.1 Our annual incidence rate of 11.1 per 100,000 (based on the 1987 diagnostic criteria of the American College of Rheumatology) is similar to a more recently published incidence rate (16.5 per 100,000).2 Since our study excluded outcomes that had already occurred at the index date, published prevalence rates are not applicable. We refer Spiera et al. to Wong's letter regarding studies of scleroderma. To our knowledge, there has not been a single controlled, population-based study showing a statistically significant association between breast implants and connective-tissue disease. Another large case-control study with negative results was recently reported.3
Wallace and Schwartz suggest that the mean follow-up of 7.8 years in our study was inadequate. Even though 36 percent of the women in our cohort were followed for over 10 years, we intend to extend the follow-up period. Wallace and Schwartz incorrectly assume that the women in our cohort, who had an average age of 34.4 years, were older than the women in theirs (average age, 39.5 years).
Wong requests clarification of the results regarding morning stiffness and serositis among the patients with cancer. With adjustments for age and type and severity of cancer, the rates for both morning stiffness and serositis were similar in the group of 125 case subjects who had implants and the group of 306 control subjects with breast cancer who did not have implants (P = 0.69 and 0.19, respectively).
We remind Levine, cochairman of the Science and Causation Committee of the Plaintiff's Steering Committee, of our June 17, 1994, statement to the press:
In fact, the design was complete and the study was already under way, with funds from the Mayo Foundation, before the Plastic Surgery Educational Foundation awarded its grant.
Our interest in this issue is strictly scientific. The same cannot be said of the plaintiffs' lawyers, who “have a huge economic incentive in the outcome.”4
4 ReferencesSherine E. Gabriel, M.D.
W. Michael O'Fallon, Ph.D.
Leonard T. Kurland, M.D.
Mayo Clinic, Rochester, MN 559051
Linos A ,Worthington JW ,O'Fallon WM ,Kurland LT . The epidemiology of rheumatoid arthritis in Rochester, Minnesota: a study of incidence, prevalence, and mortality. Am J Epidemiol 1980;111:87-98
Web of Science | Medline2
Dugowson CE ,Daling J ,Koepsell TD ,Voigt L ,Nelson JL . Silicone breast implants and risk for rheumatoid arthritis. Arthritis Rheum 1992;35:Suppl:S66-S66 abstract.
Web of Science3
Englert HJ ,Brooks P . Scleroderma and augmentation mammoplasty -- a causal relationship? Aust N Z J Med 1994;24:74-80
CrossRef | Medline4
The $4.3 billion mistake. Wall Street Journal. June 17, 1994:A14.
Author/Editor Response
A spokesperson for the Plastic Surgery Educational Foundation replies:
To the Editor: As president of the Plastic Surgery Educational Foundation, I want to take this opportunity to describe the foundation's process for the selection of research projects for funding. For nearly 50 years, the foundation's mission has been to identify and respond to the educational, research, and public-service needs associated with plastic and reconstructive surgery, particularly as these needs affect our patients.
Recently, the foundation has devoted substantial resources to an examination of the issues raised by the FDA relating to breast prostheses. The foundation has raised over $2.7 million from plastic surgeons and breast-implant manufacturers to address these concerns. Currently, 14 independent research projects are specifically addressing the concern about breast implants expressed by the FDA.
Essential to the integrity of the foundation is the process by which research proposals are evaluated for funding. After an initial call for proposals, a committee of foundation members uses rigorous standards and rating systems established by the National Institutes of Health to select the proposals that will be funded. Although periodic updates on the progress and results of the research projects are required, the foundation never seeks to exert any influence over the research.
Under no circumstances are the financial donors to the foundation given a voice in the selection of research projects, nor are these donors permitted to influence the outcome of the research in any fashion. To ensure that this standard is never compromised, the foundation's board of directors adopted the following policy at its meeting on March 13, 1993, in Washington, D.C.:
The foundation is proud to continue its tradition of support for research focused on meeting the needs of our patients.
Edward A. Luce, M.D.
Plastic Surgery Educational Foundation, Arlington Heights, IL 60005- Citing Articles (3)
Citing Articles
1
H. James Williams, Michael H. Weisman, Charles C. Berry. (1997) Breast implants in patients with differentiated and undifferentiated connective tissue disease. Arthritis & Rheumatism 40:3, 437-440
CrossRef2
M. L. Cuellar, O. Gluck, J. F. Molina, S. Gutierrez, C. Garcia, R. Espinoza. (1995) Silicone breast implant — Associated musculoskeletal manifestations. Clinical Rheumatology 14:6, 667-672
CrossRef3
(1995) More on Breast Implants and Connective-Tissue Diseases. New England Journal of Medicine 332:19, 1306-1308
Full Text
- Letters
