Antidepressant Medications and Change in Mammographic Density in Postmenopausal Women

Some studies suggest prolactin is associated with increased mammographic density (1-3). High mammographic density is a risk factor for breast cancer (4, 5). Antidepressants have been associated with increased prolactin levels (6-9). To our knowledge, no other published report has examined the relation between antidepressant use and mammographic breast density. We examined this association and, furthermore, were able to look longitudinally at changes in density after initiation of antidepressant use.

Study methods have been described previously (10). We identified women within Group Health Cooperative, an integrated health care delivery system based in Seattle, WA, who had two routine screening mammograms within 9 to 28 mo of one another between 1996 and 2006. We excluded women with a history of breast cancer, tamoxifen or raloxifene use, or breast augmentation or reduction at either exam. The analytic cohort consisted of 29,284 women. The Group Health institutional review board approved study procedures.

We classified exposure to each class of antidepressants based on Group Health electronic pharmacy dispensing data, including dispensings from claims data. We ascertained use of selective serotonin reuptake inhibitors, tricyclic antidepressants, and miscellaneous antidepressants before the first mammogram and in between the two mammograms. For each dispensing, we estimated the date when the prescription should have run out based on the date dispensed, quantity dispensed, and instructions for use (10). Successive dispensings with <60 d between the run-out date of one dispensing and fill date of the subsequent dispensing were considered continuous use.

Exposure categories were based on use before the first mammogram and between the first and second mammograms. To be considered a user of a particular class of antidepressants at the time of a mammogram, a woman must have used it for ≥60 d within the previous 6 mo. “Initiators” did not use antidepressants before the first mammogram but started using before their second mammogram. Women who began using antidepressants before their first mammogram and continued between the two mammograms were classified as “continuers.” Women who used antidepressants before the first mammogram but had no dispensings within 6 mo of the second mammogram were classified as “discontinuers.” “Nonusers” had no antidepressant use for ≥60 d in the 6 mo before either mammogram. Women who had antidepressant dispensings from >1 class were considered “mixed” users and treated as a separate exposure category.

Mammographic breast density was collected from radiologists' screening mammogram interpretations using Breast Imaging Reporting and Data System (BI-RADS) classifications: 1, almost entirely fat; 2, scattered fibroglandular; 3, heterogeneously dense; 4, extremely dense (11). We classified change in breast density between the first and second mammograms as increased by ≥1 categories, decreased by ≥1 categories, or stayed the same. For each category of density at the first mammogram, we compared the change in density across exposure categories.

We estimated odds ratios and 95% confidence intervals using polytomous logistic regression with nonusers of any antidepressants as the reference group. Analyses were adjusted for potential confounders selected a priori: age at first mammogram, use of any hormone therapy between mammograms, body mass index, and time between exams. We obtained information on these and other subject characteristics from self-reported information completed as part of the Group Health Breast Cancer Screening Program (12).

Nonusers of antidepressants were less likely to be overweight/obese and to have used hormone therapy compared with the various categories of antidepressant users (Table 1). Initiators, discontinuers, continuers, and mixed users were similar with respect to most characteristics examined; there were no major differences in density at either mammogram across groups. Changes in or continued use of antidepressants were not associated with changes in breast density between mammograms regardless of breast density at the first mammogram or the type of antidepressant (Fig. 1). The odds ratio for the association between initiating any antidepressant use and increased mammographic density was 1.08 (95% confidence interval, 0.84-1.38) for women with BI-RADS of 1 at the first mammogram, 0.98 (95% confidence interval, 0.86-1.12) for BI-RADS of 2, and 0.96 (95% confidence interval, 0.75-1.21) for BI-RADS of 3.

We did not observe changes in mammographic density associated with initiating, continuing, or discontinuing antidepressants use overall or by class. These findings are consistent with recent reviews that do not suggest an association between antidepressant use and breast cancer risk (13-15).

It is possible that only extended use of antidepressants affects breast density; women who initiated antidepressant use between the two mammograms separated by 2 years may not have used antidepressants for long enough to affect breast density in this study. Previously, we were able to detect an association between hormone therapy initiation and change in breast density over a 2-year period (10), which suggests that this interval is not too short to observe density changes associated with initiation of certain medications.

Our study had several important strengths, as previously described (10). With data from 29,284 women, we had 80% power to detect an odds ratio of 1.40 for the association between initiating any antidepressant use and increasing breast density among women who had almost entirely fatty breasts. Additionally, we assessed exposure and outcome independently of each other. Mammographic breast density was interpreted clinically, independent of exposure. Exposure was ascertained from electronic records of prescriptions dispensed and thus not subject to recall bias, a potential limitation of self-report (16). Previous studies report that 97% of patients fill most or all of their prescriptions at Group Health pharmacies (17), underscoring the completeness of these records for research purposes.

To our knowledge, no other published report has examined the relation between antidepressant use and mammographic density. Findings from our study do not suggest that antidepressant initiation, continuation, or discontinuation over a 2-year period is associated with changes in BI-RADS breast density categories.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Jane Grafton for assistance with programming, Hazel Nichols for project support, and Melissa Rabelhofer for technical assistance.