In this large population-based study of breast cancer patients treated with mastectomy, risks of breast cancer death and nonbreast cancer mortality were lower in women with implants than in women without implants, after adjustment for potential confounders. Postmastectomy breast implants were used by one-fifth of patients who were slightly younger at diagnosis and were more likely to be of white race/ethnicity and to have in situ disease than women without implants. The silicone gel-filled implant was the most common type of implant received.
Although breast reconstruction has been shown to provide psychosocial benefits to breast cancer survivors [11
], concerns have been raised that breast implants may increase the risk of local complications and systemic diseases, including certain cancers and autoimmune diseases [2
]. Breast implants have been suggested to interfere with mammography, thereby facilitating delayed detection of breast tumors, and, consequently, decreased survival [16
]. Despite recent Institute of Medicine recommendations to continue monitoring women with breast implants and to evaluate the potential long-term health effects [1
], few research studies have addressed long-term health outcomes in this group. Moreover, these studies were often conducted on small, nonrepresentative samples without detailed information on implant type and history of use.
Georgiade and colleagues found that the survival time for 101 women undergoing breast reconstruction with breast implants was nonsignificantly better than that for 377 women without reconstruction, after adjustment for tumor grade, histology, lymph node involvement, and age at diagnosis, and after a median of 3 years of follow-up [6
]. With a median of 13 years of follow-up, Petit and colleagues found that the risk of breast cancer death was marginally lower in 146 women who underwent breast reconstruction with silicone gel-filled implants than in a matched group without implants (relative risk, 0.6; 95% confidence interval, 0.3–1.1) [7
]. Vandeweyer and colleagues compared 49 women who received saline-filled breast implants following mastectomy with a matched group of women who did not. They found no difference in the number of breast cancer deaths between the two groups [8
]. In a matched analysis of 176 women with a mean of almost 6 years of follow-up, Park and colleagues found that women with breast implants after mastectomy had approximately a 70% reduced risk of death compared with women without implants (relative risk, 0.33; 95% confidence interval, 0.11–0.92) [9
Our finding of better survival in women with breast implants is consistent with most of this research [6
]. However, our study has the substantial advantages of being population-based, being large, having a long follow-up (median, 12.4 years), and including information on implant type and the implant removal and replacement. It is thus well suited to address public health concerns regarding the long-term survival and use of breast implants in women with early-stage, mastectomy-treated breast cancer. Such concerns have recently been re-evaluated in conjunction with the Food and Drug Administration hearings regarding the safety of silicone gel breast implants and their availability for the general market [17
One explanation for our finding of reduced mortality in patients with breast implants may relate to self-selection rather than to a causal role of implants. Although most women who receive mastectomy are eligible to receive breast implants as part of breast reconstruction, surgeons may not recommend this surgery to women with health conditions such as obesity or a recent history of smoking that may contribute to postoperative complications, and may thus impact on survival [19
]. In our data, the possibility of self-selection based on smoking is supported by the higher proportions of deaths from respiratory cancers and chronic obstructive pulmonary diseases in women without breast implants (Table ). Further investigation is warranted for lifestyle factors (e.g. smoking, diet) and for comorbidities that may account for the survival advantage seen in women with breast implants.
In the present study, women with breast implants had a significant excess proportion of deaths due to suicide compared with women without implants. This finding, albeit based on small numbers, is consistent with observations from studies conducted in cosmetic breast implant patients [21
] and suggests psychiatric consultation should also be considered for breast cancer patients seeking reconstructive surgery with breast implants. In any case, future studies with larger sample sizes are needed to confirm this finding in the breast reconstruction population.
An important bias of common concern in retrospective cohort studies is loss to follow-up. A total 231 (5.3%) of the 4385 patients included in the survival analysis did not have complete follow-up at the end of the study period. However, because of the relatively small percentage of patients lost to follow-up, we know that bias due to loss to follow-up has little impact on our survival findings since we found no substantial change in hazards ratios when we assumed the worst-case scenario that all patients lost to follow-up had all died or assumed that all patients lost to follow-up all lived until the end of the study period.
Furthermore, although our response rate was relatively high, differences between nonresponders and responders on several patient and tumor characteristics could have biased our findings. Although we were able to adjust for reported patient and tumor characteristics in our multivariate analyses, 356 women (nearly 40% of study-eligible patients) who did not participate in the study were deceased. In the unlikely event that all 356 deceased women had received breast implants, it is possible that the exclusion of these cases from our analysis could bias our results towards and beyond the null, and thereby overestimate the protective effects.
Although our survival analyses were adjusted for various demographic and clinical characteristics, our finding of better survival in women with breast implants could reflect uncontrolled confounding by social class, medical care, and psychological factors related to implant usage and survival. Among breast cancer patients treated with mastectomy, those choosing to have breast reconstruction have been shown to differ from women without breast reconstruction on SES, which may be an important factor affecting survival [23
]. In a convenience sample of more than 200,000 breast cancer patients undergoing mastectomy between 1985 and 1995, Morrow and colleagues found that patients with a family income of $40,000 or more were twice as likely as patients with a family income of less than $40,000 to receive postmastectomy breast reconstruction [25
]. Higher income may be a predictor of better survival after breast cancer, as women with higher incomes may have better access to cancer care and treatment. In the present study, SES did not alter the effect of implants on survival in the subset of women for whom SES measures were available. Differences in these area-level measures of SES are thus not likely to contribute substantially to the survival differences between breast cancer patients with and without implants in this study.
Additional unmeasured confounders related to the increased medical care of women with breast implants could explain the protective association of breast implants with cancer survival. Because women with breast implants may be more closely followed in their medical care, they may have recurrences diagnosed and treated earlier; thus they may experience better survival than women without implants. Although our study lacked information on breast cancer recurrence, we were able to examine the impact of subsequently diagnosed primary breast tumors. We observed that the proportion of women with two or more primary breast tumors was lower in women with breast implants than in women without (15% and 21%, respectively).
To address the possibility that a higher incidence of subsequently diagnosed primary breast tumors impacted survival in women without breast implants, we limited survival analyses to women with only one primary tumor (n
= 3535) and found a consistently reduced risk of breast cancer death associated with breast implant usage (hazard ratio, 0.54; 95% confidence interval, 0.42–0.68), after adjusting for similar prognostic factors. Our findings also are consistent with results from studies showing a reduced risk of death in augmentation mammoplasty patients with at least 10 years of follow-up compared with the general population [26
]. Furthermore, psychological factors underlying a woman's decision to obtain breast implants [30
], including body image concerns and self-esteem, may play a role in lifestyle behaviors relevant to survival, although the extent to which they directly impact survival is unclear.
Several biological mechanisms have been proposed to explain how breast implants may influence survival outcomes [16
]. Breast implants may stimulate a local immune response in which cancer cells are more likely to be destroyed [34
]. Breast implants may compress breast tissue, reducing the flow of blood and thereby slowing the rate of cell or tumor growth. Breast implants may decrease the temperature of the breast by separating the breast tissue from the body, thereby decreasing the metabolic rate and slowing the growth rate of residual breast cancer cells [35
]. These mechanisms may provide important clues in cancer prevention and warrant further investigation.