Authors’ reply to: Anti‐Müllerian hormone and breast cancer risk: is the correlation possibly associated with PCOS

Abstract

Dear Editor, Dr. Blumenfeld suggests that our observation of an increased risk of breast cancer with increasing anti-Müllerian hormone (AMH) concentration 1 could be due to a higher prevalence of polycystic ovary syndrome (PCOS), among women with high AMH concentration. This hypothesis is based on the observations that AMH concentration is elevated among women with PCOS 2,3 and that some studies have reported an association between PCOS and risk of breast cancer. 4,5 To explore the hypothesis proposed by Dr. Blumenfeld, we examined the association of PCOS with AMH and with risk of breast cancer in three cohorts (Breakthrough Generations Study, ORDET, and The Sister Study) that collected information on PCOS. We did observe higher AMH concentrations in women who reported PCOS than in women who did not, though differences were not statistically significant. In controls, the ageadjusted and cohort-adjusted geometric mean was 2.20 pmol/L (95% CI: 1.20, 4.04) for women with PCOS (n = 43) vs. 1.48 pmol/L (95% CI: 1.34, 1.64, p = .16) for women without. Similarly, in cases, the age-adjusted and cohort-adjusted geometric mean was 3.23 pmol/L (95% CI: 1.33, 7.86; n = 27) vs. 1.99 pmol/L (95% CI: 1.79, 2.21; p = .14). These observations are consistent with studies of AMH concentration in women diagnosed with PCOS. However, the proportion of women who reported a history of PCOS was slightly lower for breast cancer cases (2.7%) than controls (3.4%, p = .36). This lack of association, which is in agreement with the results of the majority of studies that have examined the association of PCOS with breast cancer risk, 6 does not support the proposed hypothesis. It should be noted, though, that we cannot exclude that underdiagnosis contributed to the lack of association we observed. The diagnosis of PCOS is difficult due to the heterogeneous nature of disease presentation and some women are asymptomatic or have nonspecific symptoms. 3,7 As a result, PCOS is often underdiagnosed, by up to 50% in some estimates, 8 with the overall prevalence estimated to be 6-10%. 9 The low prevalence of PCOS we observed suggests that underdiagnosis is indeed a possibility in our study.Another result of our study does not support the hypothesis of Dr. Blumenfeld. We observed a positive trend between AMH and breast cancer risk, with multivariate-adjusted odds ratios of 1.18, 1.34, and 1.62 for AMH quartiles 2, 3, and 4, respectively. Such a trend is not consistent with the association being driven by PCOS, even assuming a prevalence of 10%. Further, PCOS is often associated with hyperandrogenism, 10 yet the odds ratios for the AMH-risk association did not change appreciably when testosterone concentration was added to the multivariateadjusted model (AMH quartile 4 OR: 1.58).Given the lack of association between PCOS and risk of breast cancer and the positive trend of increasing risk with increasing AMH concentration we observed, our study does not support the hypothesis that the AMH-breast cancer risk association is explained by PCOS.Sincerely,