Correspondence
Breast Augmentation and the Risk of Subsequent Breast Cancer
N Engl J Med 1993; 328:661-663March 4, 1993
- Article
To the Editor:
Berkel and colleagues (June 18 issue)1 performed a population-based cohort-linkage study. A cohort of 11,676 women who underwent cosmetic breast augmentation was compared with the cohort of all 13,552 women in Alberta, Canada, in whom a first primary breast cancer was diagnosed. Breast cancer developed in 41 patients with implants, whereas the number expected on the basis of age-specific and calendar-year-specific incidence rates was 86.2. The authors conclude that women who undergo breast augmentation with silicone implants have a lower risk of breast cancer than the general population.
The authors have not addressed the distribution of risk factors in the cohort of women who underwent augmentation. It is conceivable and likely that women with risk factors for breast cancer were turned down for augmentation at the time of the initial evaluation or were at least discouraged from undergoing the procedure. Thus, women with risk factors for breast cancer (as well as those found to have breast masses at initial evaluation) would be disproportionately represented in the cohort that did not undergo breast augmentation. It is known that more women seek augmentation than actually undergo the procedure. This fundamentally flaws the results and may make the conclusions untenable.
1 ReferencesMark Levin, M.D.
Sanford R. Nalitt Institute for Cancer and Blood Related Diseases, Staten Island, NY 103051
Berkel H ,Birdsell DC ,Jenkins H . Breast augmentation: a risk factor for breast cancer? N Engl J Med 1992;326:1649-1653
Full Text | Web of Science | Medline
To the Editor:
There may be a number of reasons that the population Berkel et al. studied had an apparently lower than expected incidence of breast cancer. Some of these were discussed by the authors, but not ruled out. First, women who elect to undergo breast augmentation for cosmetic reasons are usually not those who are at high risk for familial breast cancer. Over 95 percent of the women in this study were premenopausal, and over 85 percent were under 40 years old. This is the age when familial breast cancer is prevalent. Given the suspected detrimental effect of a foreign-body intrusion, such women would probably avoid it. Second, women with a previous diagnosis of breast cancer were excluded. Third, practically all the women in the study would have had smaller breasts, less breast tissue, and therefore a lower risk of breast cancer.1 Fourth, although the maiden names of the women were used in the linkage, this information would not be available on each record. Forty-five percent of the women studied were under 30 years of age, and therefore, the proportion of unmarried persons was probably high. Subsequent name changes after marriage would impede the identification of patients. Finally, some of the women in whom cancer developed were given a diagnosis elsewhere.
1 ReferencesBaruch Modan, M.D., Dr.P.H.
National Center for Health Statistics, Hyattsville, MD 207821
Hsieh C-C ,Trichopoulos D . Breast size, handedness and breast cancer risk. Eur J Cancer 1991;27:131-135
CrossRef | Web of Science | Medline
To the Editor:
An increased risk of cancer associated with silicone implants has recently been suggested, but several studies1-3 did not confirm such a detrimental effect. We have studied 292 patients with in situ or infiltrating breast carcinoma who were treated by mastectomy at the Gustave Roussy Cancer Institute between 1965 and 1983. One hundred forty-six patients underwent immediate or delayed breast reconstruction between 1976 and 1984 with gel-filled silicone implants and were free of disease at the time of reconstructive surgery. No reconstruction was performed in the control group of 146 patients who were selected from the hospital registry after mastectomy. One control patient was matched to each patient with breast reconstruction with respect to age at diagnosis (within 10 years), year of diagnosis (within 5 years), tumor stage (based on the classification of the International Union against Cancer), histologic type of tumor, histologic prognostic grading, and nodal status (no positive axillary nodes, one to three, or four or more). Each control patient had to be free of disease after a follow-up that was at least as long as the interval between diagnosis and reconstruction for the corresponding case patient. The percentages of patients with in situ carcinoma and negative nodes were identical in both groups (29 percent and 79 percent, respectively). The mean (±SD) age was 43 ±9 years in the reconstruction group and 47 ±9 years in the control group.
The relative risk of death, recurrence, and second primary cancer was estimated by the Cox proportional-hazards model and stratified according to age at diagnosis. The date of origin was the date of reconstructive surgery in the reconstruction group and the corresponding “pseudoreconstruction date” for each control.
The median follow-up after the diagnosis of breast cancer was 12 years, and the median follow-up after the reconstructive procedure was 9 years. The risk of death or distant metastasis was significantly lower in the reconstruction group than in the control group (Table 1Table 1
Relative Risk of Death, Relapse, and Second Primary Cancer in 292 Patients with Breast Cancer, According to Whether the Patients Underwent Reconstruction with Silicone Implants.). The risk of a local recurrence, a second breast cancer, or a second primary cancer in a site other than the breast was not significantly different between the two groups of patients. The slightly increased risk of a contralateral breast cancer in the reconstruction group might be explained by the frequent association of a contralateral reduction mammaplasty, allowing the histologic diagnosis of occult breast carcinomas. The early detection of contralateral occult carcinomas could also partially explain the better survival rate in the reconstruction group.In conclusion, our results do not support the hypothesis of a carcinogenic effect of exposure to gel-filled implants in patients with breast cancer; however, our study was not powerful enough to detect small carcinogenic effects of silicone implants.
3 ReferencesJ.Y. Petit, M.D.
M. Le, M.D.
H. Mouriesse, M.Sc.
Institut Gustave-Roussy, 94805 Villejuif, France1
Deapen DM ,Pike MC ,Casagrande JT ,Brody GS . The relationship between breast cancer and augmentation mammaplasty: an epidemiologic study. Plast Reconstr Surg 1986;77:361-368
CrossRef | Web of Science | Medline2
Berkel H ,Birdsell DC ,Jenkins H . Breast augmentation: a risk factor for breast cancer? N Engl J Med 1992;326:1649-1653
Full Text | Web of Science | Medline3
Malone KE ,Stanford JL ,Daling JR ,Voigt LF . Implants and breast cancer. Lancet 1992;339:1365-1365
CrossRef | Web of Science | Medline
To the Editor:
The discrepancy found by Berkel et al. might be partly explained by the fact that there were fewer in situ carcinomas diagnosed in the group that underwent augmentation. The widespread use of screening mammography has resulted in ductal carcinoma in situ accounting for 10 to 20 percent of breast carcinomas diagnosed. It is usually identified by the presence of subtle clustered microcalcifications. An indeterminate number of these nonpalpable lesions, perhaps as many as 75 percent, may never become clinically apparent invasive cancers1. Relatively few women with breast implants in the Berkel study would be expected to have such lesions identified by mammography, because until recently, these women would not have undergone the now standard “implant push-back” views2. These views allow more breast tissue to be visualized, as well as permitting more optimal compression of the tissues that are imaged.
The findings of Berkel et al. could also be partly explained if the Alberta Cancer Registry categorized lobular carcinoma in situ as a carcinoma. This diagnosis would be expected to be made less frequently in the cohort that received breast augmentation, because those women would have undergone fewer mammographically directed biopsies.
2 ReferencesFerris M. Hall, M.D.
Beth Israel Hospital, Boston, MA 022151
Schnitt SJ ,Silen W ,Sadowsky NL ,Connolly JL ,Harris JR . Ductal carcinoma in situ (intraductal carcinoma of the breast). N Engl J Med 1988;318:898-903
Full Text | Web of Science | Medline2
Eklund GW ,Busby RC ,Miller SH ,Job JS . Improved imaging of the augmented breast. AJR Am J Roentgenol 1988;151:469-473
Web of Science | Medline
To the Editor:
The study of breast augmentation by Berkel et al. provides good evidence that smooth-walled silicone-filled and saline-filled breast implants are not associated with an increased risk of breast cancer. However, the chief concern expressed in the introduction to the paper is that polyurethane-sponge-coated breast implants are potentially carcinogenic; the risk of cancer associated with smooth-walled implants is not mentioned. Since the study only included women with smooth-walled implants, the risk of cancer associated with polyurethane-coated devices cannot be assessed. It seems that a well-founded answer has been provided, but not to the question posed.
Susan Catnach, M.R.C.P.
Peter Fairclough, F.R.C.P.
St. Bartholomew's Hospital, London EC1A 7BE, United KingdomTo the Editor:
We are pleased to see Berkel et al. replicate our finding of a low risk of breast cancer among patients who had undergone augmentation mammaplasty in Los Angeles1. In their discussion of possible explanations for this finding in both studies, the authors refer to patients lost to follow-up. It is important to note that, unlike Berkel et al., we confirmed the residence of each of our patients throughout our study. Thus, when a patient moved from Los Angeles County and our ability to report subsequent cancer diagnoses was therefore lost, we also removed her from our estimate of the number of cancers expected to occur.
In their summary of our results, Berkel et al. report that the average length of follow-up in their study (10.2 years) was almost twice as long as ours. They had not seen our most recent report2: after a median of 10.6 years of follow-up, the risk remains well below that expected, at 66 percent of the expected rate.
Even more recently, two case-control studies reported estimates of the relative risk of breast cancer associated with breast implants of 0.8 and 0.23. Although these studies included small numbers of patients who had undergone augmentation mammaplasty, the consistency of these findings with those of the two cohort studies provides strong evidence that such patients are not at increased risk for breast cancer.
3 ReferencesDennis M. Deapen, Dr.P.H.
Garry S. Brody, M.D.
University of Southern California School of Medicine, Los Angeles, CA 900331
Deapen DM ,Pike MC ,Casagrande JT ,Brody GS . The relationship between breast cancer and augmentation mammaplasty: an epidemiologic study. Plast Reconstr Surg 1986;77:361-368
CrossRef | Web of Science | Medline2
Deapen DM ,Brody GS . Augmentation mammaplasty and breast cancer: a 5-year update of the Los Angeles study. Plast Reconstr Surg 1992;89:660-665
CrossRef | Web of Science | Medline3
Malone KE ,Stanford JL ,Daling JR ,Voigt LF . Implants and breast cancer. Lancet 1992;339:1365-1365
CrossRef | Web of Science | Medline
Author/Editor Response
Dr. Berkel replies:
To the Editor: Dr. Levin points to exactly the same (theoretical) possibility that my colleagues and I addressed in our paper, when we said that “one could hypothesize that women who undergo augmentation mammaplasty have a much lower a priori risk of breast cancer.” At this time we do not have information on the prevalence and distribution of risk indicators. In the next phase of our investigations, however, we are planning a nested case-control study that is expected to provide answers to this question.
Dr. Modan gives a number of reasons why, in his opinion, the results of our study are flawed. We have already addressed the issue of prevalence of risk indicators. His second point is not relevant to the discussion. We set out to evaluate the relation between cosmetic breast augmentation and the subsequent risk of breast cancer. Obviously, women who received implants for reconstructive reasons, who were excluded from the study, are in a totally different category. However, Dr. Petit et al. provide some reassuring evidence in that regard. Dr. Modan's belief in the relation between small breasts and a lower risk of breast cancer is not shared by many investigators, as we discussed in our paper, and we do not think that this factor has a major impact on our findings. Unfortunately, Dr. Modan missed the sentence in our paper describing the linkage variables used. In addition to each patient's full name (including first initial, maiden name, and surname), the date of birth and the Alberta Health Care Insurance Plan number -- a unique number -- were used in the linkage. Moreover, we selected from the cancer registry a random sample of women with breast cancer and reviewed their clinical records. As we described in our paper, in none of the cases in this sample was evidence of the presence of an implant found. We therefore do not believe that underascertainment is a reasonable explanation of our findings.
The comment by Dr. Hall is interesting. However, Alberta had no breast-cancer screening program until late 1990. The relative frequency of in situ carcinomas in the women with implants was 12.2 percent (n = 5), whereas in the cohort with breast cancer 3.5 percent of the tumors were diagnosed as in situ tumors. Both ductal and lobular carcinomas in situ are included in this diagnosis.
We are afraid that Drs. Catnach and Fairclough misread our paper. We stated that “polyurethane-sponge-covered implants were not used in Alberta during the study period.” We therefore could not -- and never intended to -- evaluate this issue. As a matter of fact we mentioned that it would probably take another 5 to 10 years before such a study could be undertaken.
Drs. Deapen and Brody are correct in saying that at the time we wrote our paper, the results of their extended follow-up study were not available. It is indeed reassuring to see that two large cohort studies with extensive follow-up have yielded consistent results. We agree with Deapen and Brody that there is now strong evidence that cosmetic breast augmentation does not increase the risk of breast cancer.
Hans Berkel, M.D., Ph.D.
Alberta Cancer Board, Edmonton, AB T5K 2L9, Canada- Citing Articles (3)
Citing Articles
1
Kenneth A. Kern, John T. Flannery, Paul G. Kuehn. (1997) Carcinogenic Potential of Silicone Breast Implants: A Connecticut Statewide Study. Plastic & Reconstructive Surgery 100:3, 737-747
CrossRef2
Jack W. Snyder. (1997) Silicone breast implants. Journal of Legal Medicine 18:2, 133-220
CrossRef3
Louise A. Brinton, Kathleen E. Malone, Ralph J. Coates, Janet B. Schoenberg, Christine A. Swanson, Janet R. Daling, Janet L. Stanford. (1996) Breast Enlargement and Reduction: Results from a Breast Cancer Case-Control Study. Plastic & Reconstructive Surgery 97:2, 269-275
CrossRef
