Background We conducted a population-based, retrospective studyto examine the risk of a variety of connective-tissue diseasesand other disorders after breast implantation.
Methods All women in Olmsted County, Minnesota, who receiveda breast implant between January 1, 1964, and December 31, 1991(the case subjects), were studied. For each case subject, twowomen of the same age (within three years) from the same populationwho had not received a breast implant and who underwent a medicalevaluation within two years of the date of the implantationin the case subject were selected as control subjects. Eachwoman's inpatient and outpatient medical record was reviewedfor the occurrence of various connective-tissue diseases, certainother disorders thought to have an autoimmune pathogenesis (e.g.,Hashimoto's thyroiditis), and cancer other than breast cancer,as well as related symptoms and abnormal results of laboratorytests. The case subjects were categorized according to whetherthey received implants for cosmetic reasons, for reconstructionafter mastectomy for breast cancer, or for reconstruction aftersubcutaneous mastectomy for cancer prophylaxis. Additional controlsubjects (women treated for breast cancer who did not have breastreconstruction) were studied for comparison with the case subjects.
Results A total of 749 women who had received a breast implantwere followed for a mean of 7.8 years, and 1498 community controlswere followed for a mean of 8.3 years. In 5 case subjects, ascompared with 10 subjects in the control group, one of the specifiedconnective-tissue diseases was diagnosed (relative risk, 1.06;95 percent confidence interval, 0.34 to 2.97). Twenty-five casesubjects had signs or symptoms of arthritis, as compared with39 control subjects (relative risk, 1.35; 95 percent confidenceinterval, 0.81 to 2.23). Among the various signs or symptomsexamined, only morning stiffness was significantly increasedamong the women who had received a breast implant (relativerisk, 1.81; 95 percent confidence interval, 1.11 to 2.95).
Conclusions We found no association between breast implantsand the connective-tissue diseases and other disorders thatwere studied. .
We defined connective-tissue disease as a group of disordersthat includes rheumatoid arthritis, systemic lupus erythematosus,Sjogren's syndrome, dermatomyositis, polymyositis, systemicsclerosis, ankylosing spondylitis, psoriatic arthritis, polymyalgiarheumatica, vasculitis, arthritis associated with inflammatorybowel disease, and polychondritis. Other disorders consideredincluded Hashimoto's thyroiditis, primary biliary cirrhosis,sarcoidosis, and cancer other than breast cancer.
Olmsted County is relatively isolated from other urban areas.Medical care is concentrated among a handful of providers, oneof which, the Mayo Clinic, is also a major tertiary care center.A unit record system allows all medical information on eachperson to be accumulated in a single dossier14. Mayo has maintainedthe original records of every patient seen since 1907 (morethan 4.3 million patients to date). The records can be accessedby indexes to the clinical and pathological diagnoses and surgicalprocedures that are incorporated in what is known as the RochesterEpidemiology Project. Thus, medical records from essentiallyall sources of health care available to residents of the countyare linked.
In any given year, over half the residents of the county receivemedical care at the Mayo Clinic (or its affiliates, Saint Mary'sHospital and Rochester Methodist Hospital), and another 20 to25 percent are treated at the Olmsted Medical Group (or itsaffiliate, Olmsted Community Hospital). The records in the RochesterEpidemiology Project include all medical care provided at thesefacilities under the supervision of a physician, except someprivate psychiatry services. In a recent survey of 2122 Rochesterresidents identified by random-digit dialing, every residenthad a medical record at the Mayo Clinic, the Olmsted MedicalGroup, or the Olmsted Community Hospital15. We have estimatedthat 95 percent or more of the local residents (depending onage) will visit at least one of these facilities in a three-yearperiod.
Identification of Case and Control Subjects
All county residents whose medical records indicated that theyhad received a breast implant between January 1, 1964, and December31, 1991, were included as case subjects. For each of thesewomen, two women of the same age (within three years) who hadundergone a medical evaluation within two years of the dateof breast implantation in the case subject were selected fromamong county residents. Among the eligible control subjects,the ones chosen were those with the registration numbers nearestto those of the case subjects. Since unique registration numbersare assigned at the time of the initial registration for care,this method matches a case subject with two control subjectsaccording to the duration of prior medical care in the county.In prior studies, most control subjects selected in this waywere persons who had received routine or minor medical care.The case subjects were classified according to the indicationfor breast implantation -- i.e., cosmetic reasons, reconstructionafter mastectomy for breast cancer, or reconstruction aftersubcutaneous mastectomy for cancer prophylaxis among women athigh risk for breast cancer.
For each woman receiving an implant after a mastectomy for breastcancer, two additional county residents who had undergone mastectomyfor breast cancer but did not have a breast implant were selectedas control subjects. Thus, there were four control subjectsfor each case subject who had been treated for breast cancer.The primary analysis was performed with control subjects fromthe community. This study was limited to women living in thecounty on the index date (the date of implantation for the casesubjects, the date of the closest medical visit for the controlsubjects, and the date of surgery for women who underwent mastectomyfor breast cancer but did not receive an implant). Follow-upof all women continued from the index date until death, December31, 1991, or the date of the woman's last health care visitin the county.
For each woman studied, the entire medical record, includingall inpatient and outpatient reports by any local provider ofmedical care, was reviewed by one of four trained nurses whosought evidence of the clinical features, laboratory studies,and radiographic studies necessary to make the following diagnoses:systemic sclerosis, systemic lupus erythematosus, rheumatoidarthritis, polymyositis, dermatomyositis, Sjogren's syndrome,keratoconjunctivitis sicca, lymphoproliferative disorders, Hashimoto'sthyroiditis, primary biliary cirrhosis, and cancer other thanbreast cancer. The specific diagnostic criteria are describedelsewhere16,17,18,19. Pertinent clinical and laboratory datawere recorded on a pretested data-collection form, entered intoa computer, and edited with a variety of on-line checks forconsistency and appropriateness of range. The principal investigatorand the study coordinator met weekly with the nurses reviewingthe records to resolve questions about the interpretation ofentries and to monitor adherence to the protocol.
The case and control groups were matched according to age andyear of implant or visit. Since no differences in results wereobserved between matched and unmatched analyses, only the resultsof unmatched analyses are reported. The primary analysis comparedthe incidence of various end points in the case and controlsubjects over the person-years of observation, using a modifiedF-test developed by Cox20 that assumes that the observed eventsfollow a Poisson distribution. The Cox proportional-hazardsmodel21 of time to the occurrence of an end point and the generalizedlinear model (with log-link function)22 of the incidence ofevents over person-years of observation were used to comparethe rates of end points in the case subjects with those in thecontrol subjects, with adjustment for covariates. Both methodsallow for the inclusion of covariates in the models to adjustfor their influence on rates and on the comparison of case withcontrol subjects. Age and index year were included in all analyses.All statistical tests were two-tailed, and a P value of <0.05was considered to indicate statistical significance.
The 749 county residents who received a first breast implantbetween January 1, 1964, and December 31, 1991 (the case subjects),were followed for 5847 person-years (mean, 7.8 years per casesubject). The control subjects from the community included 1498women who were followed for 12,361 person-years (mean, 8.3 yearsper control subject). The control subjects with cancer included306 women with breast cancer who did not receive a breast implantand who were followed for 2703 person-years (mean, 8.8 yearsper subject). There was little difference in mean age or durationof follow-up between the case and control subjects ,but marital status, ethnic background, and smoking history differedsignificantly.
The case subjects were compared with the control subjects withregard to the occurrence of clinical signs and symptoms of theconnective-tissue diseases and other disorders being studied,abnormal laboratory tests suggestive of these conditions, anddiagnoses of these diseases. Five of the 749 case subjects hadbeen given a diagnosis of connective-tissue disease (polymyalgiarheumatica in 2 and Sjogren's syndrome, polychondritis, andarthritis associated with inflammatory bowel disease in 1 each),as compared with 10 of the 1498 community controls (3 with ankylosingspondylitis, 2 each with rheumatoid arthritis and systemic vasculitis,and 1 each with psoriatic arthritis, polymyalgia rheumatica,and systemic sclerosis with polyarthritis) (relative risk, 1.06;95 percent confidence interval, 0.34 to 2.97).
No significant differences were found between the case and controlsubjects in the relative risk of abnormal results on laboratorytesting for antinuclear antibody, rheumatoid factor, or antimicrosomalantibody (Table 4). Abnormal results of the laboratory testfor thyroid-stimulating hormone were significantly more frequentamong the control subjects than among the case subjects (Table 4).
The comparisons in Table 2, Table 3, and Table 4 did not involveadjustment for age or index year, the two variables used inmatching the case subjects and the control subjects, or forethnic background, marital status, or smoking history, variableson which the two groups differed ( Table 1). In the case of endpoints that occurred often enough to warrant a more complexanalysis, the Cox proportional-hazards model was used to obtainestimates of hazard ratios adjusted for age and index year ( Table 5).The other variables were not included, because they eitherhad no effect on the results and were never significant (maritalstatus and smoking history) or were rare (race other than white).Although the majority of these hazard ratios were very similarto the rate ratios, there were some differences, presumablydue to the adjustment for age. As in the unadjusted analysis,the only significantly elevated risks in the proportional-hazardsanalysis were those for serositis, in the women who underwentbreast implantation after mastectomy for breast cancer, andfor morning stiffness, among all women with implants and thosewho had implants after subcutaneous mastectomy for cancer prophylaxis.
Reports of connective-tissue diseases and other disorders3,4,5,6,7,8,9,10,11,12,23,24,25,26,27,28,29,30among women with breast implants have raised the question ofa cause-and-effect relation. We believe that a definitive answerto this question would require a well-conducted prospectivecohort study using a multicenter controlled design with follow-upof a large population of women for at least 10 years. Such astudy would be very expensive and time-consuming. Instead, weconducted a retrospective cohort study using the extensive dataresources of the Rochester Epidemiology Project.
Other studies have addressed the question of whether silicone-containingbreast implants cause connective-tissue disorders31,32,33,34,35.However, these studies were limited by substantial methodologicflaws, including uncontrolled, retrospective designs,31 lackof objective validation of the survey replies,31 very smallsamples, inadequate statistical power, large losses to follow-up(nearly 75 percent of the initial sample in one study),31 andthe use of referral-based cohorts31,32,33,34,35. In contrast,our controlled study allowed extensive follow-up of a population-basedcohort of women with silicone-containing breast implants andthe corresponding control group.
We found no statistically significant elevation in the relativerisk of any of the specified connective-tissue diseases or otherdisorders among the women with breast implants as compared withthe control subjects. Two conditions, morning stiffness andserositis, were significantly more frequent among the subgroupof the case subjects who received their implant after mastectomyfor breast cancer (Table 5), but our data suggest that theseevents are more likely to be related to the underlying cancerthan to the implant itself, because the incidence rates of theseevents were similar to those among the control women with breastcancer who did not receive a breast implant after mastectomy.
Although our results do not support the hypothesis that womenwith breast implants have an increased risk of connective-tissuediseases or other disorders, our study has several limitations.First, we had limited power to detect an increased risk of rareconnective-tissue diseases, such as systemic sclerosis. Indeed,we calculated that it would require a sample of 62,000 womenwith implants and 124,000 women without implants, followed foran average of 10 years each, for a doubling of the relativerisk of this condition to be detected among women with implants,assuming that the annual incidence of systemic sclerosis is1.6 cases per 100,000 women36. Our results, therefore, cannotbe considered definitive proof of the absence of an associationbetween breast implants and connective-tissue disease. Othercontrolled studies of this question are ongoing. Their results,together with ours, will help to resolve this controversy.
The use of data bases linked to medical records, such as theRochester Epidemiology Project, presents a number of potentialproblems37. For a disorder to be recorded in the data base,the condition must be recognized by a clinician, recorded accurately,and retrieved. Because we screened and reviewed the symptomsand laboratory results directly, events could be included whenthey met our definitions, even if the attending physician failedto make (or record) the diagnosis. For example, we found thatcontrary to previous reports,30 the incidence of abnormal testresults for antinuclear antibodies in women with breast implantsdid not differ significantly from the incidence in women withoutimplants in whom the test was performed. Still, conditions thatare unlikely to come to medical attention may not be detectedin such a data set -- for example, some episodes of "dry eyes"that could have been related to sicca syndrome may have beenmissed. However, we reviewed each subject's complete medicalrecord covering an average of eight years, and it is unlikelythat serious or persistent problems, such as Sjogren's syndromeor cancer, went unrecorded by every attending physician.
Our study had a long average follow-up, lasting at least 10years for 36 percent of the cohort. Nevertheless, very lateoutcomes, occurring more than 10 years after surgery, couldnot be evaluated adequately. It may be desirable to extend thefollow-up of this cohort. Finally, some racial and ethnic groupsare underrepresented in Olmsted County, where the populationwas 96 percent white in 1990. Conditions that occur preferentiallyin other groups could not be assessed. However, most previousreports linking breast implants and connective-tissue diseaserelate to white women, and 97 percent of our case subjects werewhite. Thus, the results of our population-based study are morelikely to apply to whites in general than are the findings frominvestigations of referral-based cohorts. In conclusion, ourresults do not support an association between breast implantsand connective-tissue diseases or other disorders that werestudied.
Supported by research grants from the Plastic Surgery EducationalFoundation and by a grant (AR30582) from the National Institutesof Health.
We are indebted to Susan Greteman, B.S., and Joanne Benson,B.A., for performing the statistical analysis, to Beth Ryanand Lori Norby for assistance in the preparation of the manuscript,and to Joseph Duffy, M.D., for providing the idea from whichthis study developed.
From the Division of Rheumatology and Internal Medicine (S.E.G.), the Department of Health Sciences Research (S.E.G., W.M.O., L.T.K., C.M.B., L.J.M.), and the Department of Surgery (J.E.W.), Mayo Clinic and Mayo Foundation, Rochester, Minn.