Do Breast Implants Cause Disease?

A Review of the Studies Included in the Meta-Analysis
August 2000
By Patricia Lieberman, Ph.D. and Diana Zuckerman, Ph.D.

There are 20 epidemiological studies included in the meta-analysis published in the New England Journal of Medicine (March 16, 2000), to determine whether breast implants cause connective-tissue diseases. This is essentially the same meta-analysis that was conducted by Judge Pointer’s scientific panel. The authors concluded that the studies show no association between silicone breast implants and connective-tissue disease. A careful review of the studies included in the meta-analysis reveals that those studies have a number of flaws.

• Five of 20 studies cited were not published in peer-reviewed journals. Instead they were papers presented at scientific meetings or unpublished doctoral dissertations. There was therefore limited information on methodology available to evaluate the validity of the study designs.

• The studies do not provide a comprehensive evaluation of diseases among breast implant patients. Most evaluate a few connective-tissue diseases, including such rare diseases as scleroderma. Most do not evaluate the “atypical” connective-tissue disease symptoms or fibromyalgia-type symptoms that many patients report.

• Even for the illnesses that they evaluate, the studies have limitations. In order to conduct an accurate study of implant patients’ health, patients should undergo a comprehensive medical exam. In contrast, most of these studies relied on medical records, which might omit vague symptoms that would be reported in the early stages of disease. Several studies relied on self-report, but only the one that found a significant risk due to implants was criticized because patients might exaggerate their health problems. In contrast, studies that determined whether women had implants based on self-report were included and not criticized as biased, even though it would be expected that some women would fail to mention that they have implants. This failure to report implants is especially likely when information was gathered in person or on the telephone, rather than in a questionnaire.

• Several of the studies relied on hospital records. Very few implant patients would have been hospitalized for their symptoms, since most health problems that implant patients have reported do not require hospitalization.

• The studies included women who had implants for a short period of time, such as a few months or years. If implants cause connective-tissue diseases, it would be expected that the disease would develop over a period of years. Diseases might also be more likely after a silicone gel implant breaks, which usually occurs after 7-10 years. Therefore, a well-designed study would include women who had implants for at least 7-10 years, not an average of 7-10 years.

•Many of the studies do not evaluate the safety of implants for mastectomy patients, and therefore the results may not be relevant to them.

• Many of the samples are too small to study rare diseases, and thus, have limited power to detect increases in the rates of disease, even increases as large as 50-100 percent.

• Older implants (from 1964-75) were made of a thicker silicone shell than newer implants. Those implants were less likely to “bleed” silicone through the shell or to break. Therefore, studies with women who had implants for a wide range of years would not be expected to show a “dose” response, and studies with women having implants for an average of 7-10 years often include many women with implants for very short periods and women with these thicker, potentially less damaging, implants. That minimizes the likelihood of results showing significant risks from implants.

• In at least one of the studies, women were included in study even if they had their breast implants removed shortly after they got them. It is impossible to tell from that study how long the women had breast implants. The other studies do not mention whether women who were identified by medical records as having implants still had them years later. Those omissions potentially bias the findings because women who had implants removed do not have the same amount of exposure as women who have implants continuously.

Cohort Studies
Cohort studies compare women with breast implants to a group of women who are similar in terms of age, race, and health who did not have breast implants.

Edworthy, S.M., Martin, L., Barr, S.G., et al. A Clinical Study of the Relationship Between Silicone Breast Implants and Connective Tissue Disease. Journal of Rheumatology 1998; 25: 254-260.

Number of implant recipients: 1576 Number of controls: 727

Does the study include mastectomy patients receiving implants? NO

Diseases studied: Any classic connective-tissue disease including rheumatoid arthritis, lupus, scleroderma, and Sjogren’s syndrome.

Minimum length of time with implants included in study: Unclear

Average length of time with implants: 13.5 years

Additional notes: Women with breast implants were 44% more likely to have a diagnosis of rheumatoid arthritis (relative risk: 1.44). That difference was not statistically significant. When interviewed about their health, women with implants were significantly more likely to have difficulty solving thought problems, have numbness in their extremities, muscle pain, headache, and hand pain. However, those symptoms were not included in the meta-analysis. This study relied on medical records. The authors did not question or examine patients directly.

Friis, S., Mellemkjaer, L., McLaughlin, J.K., et al. Connective Tissue Disease and other Rheumatic Conditions Following Breast Implants in Denmark. Annals of Plastic Surgery 1997; 39: 1-8.

Number of implant recipients: 2,570 Number of controls: 11,023

Does the study include mastectomy patients receiving implants? YES

If so, how many? 1,435 of 2,570

Were mastectomy patients analyzed separately from augmentation patients? YES

Diseases studied: Any classic connective-tissue disease, including lupus, Sjogren’s syndrome, rheumatoid arthritis, and scleroderma. Also looked at “other and ill-defined” rheumatic conditions.

Minimum length of time with implants included in study: To be in this study a woman could have had implants for less than one year.

Average length of time with implants: 7.2 years for reconstruction group, 8.4 years for augmentation group.

Additional notes: Only women who were hospitalized for connective-tissue disease were categorized as ill, not outpatients. According to the authors, the study had only limited power to detect an increased risk of any specific connective-tissue disease. The control group consisted of women who had breast reduction surgery, or mastectomy without receiving implants. Although the difference was not significant, the rate of scleroderma, lupus, and Sjogren’s syndrome in mastectomy patients receiving implants was 30% higher than expected. The authors found an increase in rheumatic complaint in all of the groups and therefore concluded that breast surgery increases the risk of connective-tissue disease, and that the implants themselves do not cause connective-tissue disease. The authors did not question or examine patients directly.

Gabriel, S.E., O’Fallon, W.M., Kurland, L.T., et al. Risks of Connective-Tissue Diseases and Other Disorders after Breast Implantation. New England Journal of Medicine 1994; 330: 1697-1702.

Number of implant recipients: 749 Number of controls: 1498

Does the study include mastectomy patients receiving implants? YES

If so, how many? 125 of 749

Were mastectomy patients analyzed separately from augmentation patients? YES

Diseases studied: Any classic connective-tissue disease, including lupus, Sjogren’s syndrome, rheumatoid arthritis, and scleroderma. Also looked at other disorders such as Hashimoto’s thyroiditis, cirrhosis, sarcoidosis, and cancer.

Minimum length of time with implants included in study: Women in this study could have had implants for less than one year.

Average length of time with implants: 7.8 + 5.5 years

Additional notes: Women with breast implants had a 35% higher rate of arthritis, which was not statistically significant (relative risk: 1.35). Morning stiffness was 81% higher for implant patients, which was significantly higher than for women without implants (relative risk: 1.81). The authors estimated that they would need to have studied 62,000 women with implants for an average of 10 years to detect a 100% increase (or less) in rare diseases such as scleroderma. This study relied on medical records. The authors did not question or examine patients directly.

Giltay, E.J., Bernelot Moens, H.J., Riley, A.H., et al. Silicone Breast Prostheses and Rheumatic Symptoms: a Retrospective Follow Up Study. Annals of the Rheumatic Diseases 1994; 53: 194-196.

Number of implant recipients: 235 Number of controls: 210

Does the study include mastectomy patients receiving implants? YES

If so, how many? Approximately 56 of 235

Were mastectomy patients analyzed separately from augmentation patients? NO

Diseases studied: Rheumatic complaints, use of anti-rheumatic drugs, and medical consultations regarding rheumatic symptoms. For those reporting rheumatic symptoms, a rheumatologist made an assessment of the likelihood of a rheumatic disease.

Minimum length of time with implants included in study: Two years

Average length of time with implants: 6.5 years with a range of two to 14 years
Additional notes: Women with silicone breast implants reported significantly more rheumatic complaints than controls, but there was no evidence of increased prevalence of common rheumatic diseases, such as fibromyalgia, rheumatoid arthritis, or Sjogren’s disease. If mastectomy patients are more vulnerable to diseases than augmentation patients, the results may not accurately describe the health risks for mastectomy patients, since they were a small minority of the women in the study. The control group consisted of women who had an unspecified cosmetic procedure that did not include silicone products. The study relied on questionnaires completed by the patients. The authors did not question or examine patients directly.

Hennekens, C.H., Lee, I.M., Cook, N.R., et al. Self-Reported Breast Implants and Connective-Tissue Diseases in Female Health Professionals. Journal of the American Medical Association 1996; 275: 616-621.

Number of implant recipients: 10,830 Number of controls: 384,713

Does the study include mastectomy patients receiving implants? YES

If so, how many? 18% of 10,830

Were mastectomy patients analyzed separately from augmentation patients? YES

Diseases studied: Any classic connective-tissue disease including lupus, Sjogren’s syndrome, rheumatoid arthritis, and scleroderma. Also included mixed connective-tissue disease.

Minimum length of time with implants included in study: To be in this study, a woman could have had implants for one year.

Average length of time with implants: Not stated, but the authors analyzed the women in three groups: up to four years, five to nine years, and 10 or more years after receiving implants and showed no increased risk with increased duration of exposure.

Additional notes: Implant patients had a 25% higher rate of connective-tissue disease, whether they were reconstruction or augmentation patients (relative risk: 1.25). This was statistically significant and the researchers concluded that there is a small increased risk of connective-tissue disease among women with implants. Although it is a cohort study, this study was analyzed with case-control and cross-sectional studies in the meta analysis because information about the disease and the patient’s exposure to silicone breast implants was gathered at the same time. The study relied on questionnaires completed by the subjects, who were health professionals. The authors did not question or examine the women directly.

Nyren, O., Yin, L., Josefsson, S., et al. Risk of Connective Tissue Disease and Related Disorders Among Women with Breast Implants: A Nation-Wide Retrospective Cohort Study in Sweden. British Medical Journal 1998; 316: 417-422.

Number of implant recipients: 7,442 Number of controls: 3,353

Does the study include mastectomy patients receiving implants? YES

If so, how many? 3,942 of 7,442

Were mastectomy patients analyzed separately from augmentation patients? YES

Diseases studied: Hospitalizations for classic connective-tissue disease including lupus, Sjogren’s syndrome, rheumatoid arthritis, and scleroderma. Also studied hospitalizations for related diseases.

Minimum length of time with implants included in study: One month

Average length of time with implants: Six years for reconstruction patients, 10.3 years for augmentation patients.

Additional notes: Only women who were hospitalized for connective-tissue disease were categorized as ill, not outpatients. The authors acknowledge that the sample size was too small to draw conclusions about links between breast implants and rare diseases they studied, such as scleroderma. The control group consisted of women who had breast reduction surgery. Both groups who had breast surgery had slightly higher than expected rates of connective-tissue disease. This study relied on hospital records. The authors did not question or examine patients directly.

Park A.J., Black, R.J., Sarhadi, N.S., et al. Silicone Gel-Filled Breast Implants and Connective Tissue Diseases. Plastic and Reconstructive Surgery 1998; 101: 261-268.

Number of implant recipients: 317 Number of controls: 419

Does the study include mastectomy patients receiving implants? YES

If so, how many? 207 of 317 implanted women

Were mastectomy patients analyzed separately from augmentation patients? YES

Diseases studied: Signs and symptoms of connective-tissue disease, such as a antinuclear antibodies, rheumatoid factor, joint pain, fatigue, Raynaud’s syndrome, etc.

Minimum length of time with implants included in study: Not specified

Average length of time with implants: Six years for reconstruction patients, five years for augmentation patients.

Additional notes: Because the sample size was so small, the authors acknowledge that a health risk would have to exceed 320% for reconstruction patients and 1600% for augmentation patients in order to be statistically significant. In addition, approximately half of the women had implants for less than six years. Because of these shortcomings, this study does not provide useful information. The study included two controls for each implantation patient. Half of the controls were maternity patients and half were outpatients from the plastic surgery department. The authors did not specify what types of procedures the plastic surgery controls received. The study subjects were interviewed and received a medical examination.

Sanchez-Guerrero, J., Colditz, G.A., Karlson E.W., et al. Silicone Breast Implants and the Risk of Connective-Tissue Diseases and Symptoms. New England Journal of Medicine 1995; 332: 1666-1670.

Number of implant recipients: 1,183 Number of controls: 86,318

Does the study include mastectomy patients receiving implants? YES

If so, how many? 525 of 1183 for cancer or prophylaxis

Were mastectomy patients analyzed separately from augmentation patients? NO

Diseases studied: Any classic connective-tissue disease, including lupus, Sjogren’s syndrome, rheumatoid arthritis, and scleroderma. Excluded women with milder or atypical cases of connective-tissue disease.

Minimum length of time with implants included in study: One month

Average length of time with implants: 9.9 + 6.4 years

Additional notes: According to the authors, the study does not exclude small health risks of implants that would be of public health importance. The study was designed to minimize “reporting bias” of health problems by implant patients by excluding any health problems diagnosed after May 1990, which was six months before the major media coverage of implant problems. They did not minimize bias in the opposite direction; for example, they included women who only had implants for one month. Also, they should have excluded women who reported receiving breast implants from 1952 to 1961, since breast implants had not yet been invented. Including these women and their inaccurate statements increased the average years of implantation. The study relied on questionnaires completed by the subjects, who were health professionals. The authors did not question or examine the women directly, although, for a random sample of 100 women, they verified whether the women had breast implants by looking at her medical records.

Schusterman, M.A., Kroll, S.S., Reece, G.P., et al. Incidence of Autoimmune Disease in Patients after Breast Reconstruction with Silicone Gel Implants Versus Autogenous Tissue: A Preliminary Report. Annals of Plastic Surgery 1993; 31: 1-6.

Number of implant recipients: 250 Number of controls: 353

Does the study include mastectomy patients receiving implants? YES, all were mastectomy patients.

Diseases studied: Patients were considered to have rheumatic disease if they had been seen by a physician who made the diagnosis on clinical grounds with corroborating laboratory evidence and had prescribed therapy.

Minimum length of time with implants included in study: 10 months

Average length of time with implants: Less than 2.5 years

Additional notes: Length of follow-up was too short to be meaningful. The authors state that the report must be considered preliminary because the onset of autoimmune disorders could occur two to 21 years after implantation. Also, if Friis and Nyren are correct, any breast surgery patient would be at increased risk for an autoimmune disease.

Wells, K.E., Cruse, C.W., Baker, J.L. Jr., et al. The Health Status of Women Following Cosmetic Surgery. Plastic and Reconstructive Surgery 1994; 93: 907-912.

Number of implant recipients: 222 Number of controls: 80

Does the study include mastectomy patients receiving implants? NO

Diseases studied: Study looked at the incidence of 23 symptoms and the diagnosis of connective-tissue disease such as rheumatoid arthritis, lupus, scleroderma, and Raynaud’s disease.

Minimum length of time with implants included in study: Not specified

Average length of time with implants: 4-5 years

Additional notes: The authors compared women who had breast implants to those who had liposuction, nose jobs, or eyelid lifts. The average age of women getting breast implants was almost 10 years younger than those getting the other cosmetic procedures. Tender and swollen glands under the arm were more likely in implanted women. Symptoms that were more frequent in implanted women but did not achieve statistical significance were: easily tired, muscle pain, swollen and tender glands in the neck, change in hand color with cold, weight gain, swollen and painful joints, and general stiffness. The authors acknowledged that the small sample size could explain why the differences did not achieve statistical significance. The authors reported no cases of scleroderma or lupus. Arthritis was present in 5% of implanted women and 3% of controls. One implanted woman reported Raynaud’s disease. The study relied on questionnaires completed by the subjects. The authors did not question or examine the women directly.

Case-control or cross-sectional studies
These studies compare women suffering from a particular disease (cases) to those who are healthy (controls) and determine whether breast implants are more common in the ill women.

Burns, C.J., Laing, T.J., Gillespie, B.W., et al. The Epidemiology of Scleroderma Among Women: Assessment of Risk from Exposure to Silicone and Silica. Journal of Rheumatology 1996; 23: 1904-1911.

Number of cases: 274 Number of controls: 1184

Diseases studied: Scleroderma

Additional notes: This study revealed no increased likelihood that women with scleroderma reported having silicone breast implants. However, women with scleroderma were significantly more likely to report other exposures to silicone. Women with scleroderma were identified by contacting rheumatologists, hospitals, and a scleroderma support group. They were then interviewed on the telephone to determine past exposure to silicone or silica.

Dugowson, C.E., Daling, J., Koepsell, T.D., et al. Silicone Breast Implants and Risk for Rheumatoid Arthritis. Arthritis and Rheumatism 1992; 35: Suppl:S66.

Number of cases: 300 Number of controls: 1,456

Disease studied: Rheumatoid arthritis

Additional notes: This study was a non-peer-reviewed abstract from a scientific meeting. One case and 12 controls had breast implants before diagnosis. There was no increase in the likelihood that rheumatoid arthritis patients reported having breast implants. The study was based on a questionnaire sent to women with rheumatoid arthritis and age-matched controls asking if they had breast implants.

Englert, H.J., Brooks, P., et al. Scleroderma and Augmentation Mammoplasty – A Casual Relationship? Australia and New Zealand Journal of Medicine 1994; 24: 74-80.

Number of cases: 286 Number of controls: 253

Disease studied: Scleroderma.

Additional notes: This study found no increased likelihood that women with scleroderma reported having breast implants, although the authors acknowledged that the study lacked the power to detect an increased risk of lower than 150-200%. The study was based on a telephone questionnaire. The information on whether the women had implants was self-reported to the interviewer on the telephone and unverified.

Goldman, J.A., Greenblatt, J., Joines, R., et al. Breast Implants, Rheumatoid Arthritis, and Connective Tissue Diseases in a Clinical Practice. Journal of Clinical Epidemiology 1995; 48: 571-82.

Number of cases: 721 Number of controls: 3,508

Disease studied: Rheumatoid arthritis and other connective-tissue disease.

Additional notes: Instead of comparing sick women to healthy women, all of the women in this study were patients in a rheumatology practice. The authors found no increased likelihood that women with rheumatoid arthritis and other connective-tissue disease reported having breast implants. The women who had breast implants were significantly younger than those who did not have implants. The authors acknowledged that since the study took place in the practice of a single clinician, there is the potential for referral or selection bias. Also, many patients were seen for only a single assessment (Fewer than half were seen in that practice for more than one year). Additionally, the author acknowledged that losing women to follow up could have resulted in a selection bias. The authors relied on medical records to determine who had breast implants.

Hochberg, M.C., Perlmutter, D.L., Medsger, T.A. Jr., et al. Lack of Association Between Augmentation Mammoplasty and Systemic Sclerosis (Scleroderma). Arthritis and Rheumatism 1996; 39: 1125-1131.

Number of cases: 837 Number of controls: 2,507

Disease studied: Scleroderma.

Additional notes: The study revealed no difference in the likelihood that women with scleroderma reported having breast implants, although the authors noted that 1,000 cases and 3,000 controls would be needed in order to detect a two-fold increase in scleroderma. For women with scleroderma, information about whether she had breast implants was gathered using a self-administered questionnaire. Controls were given the identical questionnaire over the telephone. For both groups, the information was unverified.

Lacey, J.V. Jr., Laing, T.J., Gillespie, B.W., et al. Reply to Letter: Epidemiology of Scleroderma Among Women: Assessment of Risk from Exposure to Silicone and Silica. Journal of Rheumatology 1997; 24: 1854-1855.

Number of cases: 189 Number of controls: 1,043

Disease studied: Scleroderma

Additional notes: This study was briefly described in a letter in the Journal of Rheumatology. It was not peer-reviewed. In a telephone interview, researchers asked who were diagnosed with scleroderma about their exposure to silicone (including silicone gel breast implants) and compared the likelihood with similarly aged controls. One case and 10 controls reported having silicone breast implants. There was no increased likelihood that women with scleroderma reported having breast implants.

Laing, T.J., Gillespie B.W., Lacey, J.V. Jr., et al. The Association Between Silicone Exposure and Undifferentiated Connective Tissue Disease Among Women in Michigan and Ohio. Arthritis and Rheumatism 1996; 39: Suppl:S150.

Number of cases: 206 Number of controls: 2,239

Disease studied: Undifferentiated connective-tissue disease

Additional notes: This study was a non-peer-reviewed abstract from a meeting. In a telephone interview, researchers asked women with undifferentiated connective-tissue disease about their silicone exposure and compared the exposure with similarly aged controls. Although there were no raw data in the abstract, the authors state that women with undifferentiated connective-tissue disease were significantly more likely to report having all types of implanted devices, including breast implants. For silicone breast implants, the adjusted odds ratio was elevated, but did not achieve statistical significance (women with undifferentiated connective-tissue disease were 127% more likely to report having silicone breast implants than controls). Women with undifferentiated connective-tissue disease were significantly more likely to report having other types of devices containing silicone, such as, internal fixation devices, artificial joints, pacemakers, non-CNS shunts or catheters.

Strom, B.L., Reidenberg, M.M., Freundlich, B., et al. Breast Silicone Implants and Risk of Systemic Lupus Erythematosus. Journal of Chemical Epidemiology 1994; 47: 1211-1214.

Number of cases: 133 Number of controls: 100

Disease studied: Lupus

Additional notes: The study did not detect an increased likelihood that women with lupus had breast implants, although the small number of cases and controls severely limited the statistical power of this study. Only one woman in the study reported that she had breast implants. Information was gathered by telephone interview.

Teel, W.B., A Population-Based Case-Control Study of Risk Factors for Connective Tissue Diseases. (Ph.D. dissertation. Seattle: University of Washington, 1997.

Number of cases: 427 Number of controls: 1577

Disease studied: All connective-tissue diseases

Additional notes: Non-peer-reviewed doctoral dissertation. No information available.

Wolfe, F., Silicone Breast Implants and the Risk of Fibromyalgia and Rheumatoid Arthritis. Arthritis and Rheumatism 1995; 38: Suppl:S265.

Number of cases: 1,270 Number of controls: 1,134

Disease studied: Compared women with rheumatoid arthritis and fibromyalgia (though fibromyalgia data not included in meta-analysis) to women with osteoarthritis and healthy women.

Additional notes: This study was a non-peer-reviewed abstract from a meeting. It compared 533 patients with fibromyalgia and 637 with rheumatoid arthritis to 479 with osteoarthritis and 655 women from the general population. Only fourteen women reported having breast implants in the study. Women with fibromyalgia or rheumatoid arthritis were no more likely to report having silicone breast implants than controls. The information on whether the women had implants was self-reported and unverified. Patients were asked to fill out questionnaires and controls (healthy women) were questioned on the telephone.