Antiestrogen Therapy for Breast Cancer Modifies the Risk of Subsequent Cutaneous Melanoma

Incidence of cutaneous melanoma is increasing faster than any other cancer in white-skinned populations from industrialized countries (1). Approximately 200,000 new cases of cutaneous malignant melanoma are diagnosed each year worldwide (2). Outstandingly, the highest incidence rates are observed in Switzerland for both males and females after New Zealand and Australia (http://www.nicer.org/Default.aspx).

Intensive sun exposure is a major risk factor for melanoma, together with cutaneous white phenotype, multiple dysplastic nevi, and history of melanoma in a first-degree relative (1, 3).

Striking differences in age-specific patterns of cutaneous melanoma between women and men have suggested modulation of risk by sex. Melanomas in women rarely occur before puberty then rise through the reproductive years until about age 50 and decrease after menopause. With regards to prognosis, women have better survival than males and this survival advantage is not fully explained by differences between sexes in site, histology, or stage (4–6). Sex hormones are known to be associated with increase of melanocytes and melanin production in human skin (7) and may thus be implicated in melanoma genesis. Supporting this hypothesis, benign nevi and melanoma have been shown to express estrogen-binding receptors (7–10).

If estrogens play a role in a woman's risk of developing melanoma, then antiestrogens use may also modify this risk. Women with breast cancer offer a unique possibility to examine the role of antiestrogens on melanoma cancer incidence because of the widespread use of these drugs, such as tamoxifen or aromatase inhibitors, as adjuvant breast cancer therapy. Therefore, we conducted a population-based study to assess the risk of secondary melanoma in a large cohort of breast cancer patients according to the use of antiestrogen therapy.

The population-based Geneva Cancer Registry has been described in detail elsewhere (11). Briefly, the database contains information on all cancer diagnosed in the resident population of the canton of Geneva since 1970. The Cancer Registry extracts information from various sources and is considered accurate (12). Recorded data include sociodemographic information, tumor characteristics (13), hormone receptor status, stage of disease at diagnosis, treatment during the first 6 months after diagnosis, occurrence of other primary cancers, and survival status.

The Cancer Registry regularly assesses survival. The index date refers to the date of confirmation of diagnosis or the date of hospitalization when it preceded the diagnosis and was related to the disease. In addition to passive follow-up (routine examination of death certificates and hospital records), active follow-up is carried out yearly with the files of the Cantonal Population Office in charge of the registration of the resident population.

This study population consisted of all women diagnosed with invasive breast cancer between 1980 and 2005. We excluded breast cancer cases diagnosed at death (n = 64). The final cohort consisted of 7,360 breast cancer patients.

Socioeconomic status was regrouped in 3 levels: low (manual employees, skilled, and unskilled workers), middle (nonmanual employees and administrative staff), and high (professionals, executives, administrators, and entrepreneurs) on the basis of the patient's last occupation or, if unemployed, that of the spouse. Familial risk of breast and ovary cancers was categorized as high (at least 1 first-degree relative with breast or ovarian cancer diagnosed before the age of 50), low (no affected first- or second-degree relatives with breast or ovarian cancer), or moderate (all other known family histories). Information of familial risk of skin cancer was available routinely but without specification of the histologic type. We therefore retrospectively extracted data from medical files of all women reporting family history of skin cancer to specify whether it was a melanoma or not. Breast cancer patients were then classified into 2 categories on the basis of the presence or the absence of at least 1 malignant melanoma in first- or second-degree relatives.

Breast cancer staging was based on the pathologic tumor-node-metastasis (TNM) classification or, when absent, the clinical TNM classification (14). Hormone receptor status was classified as positive (≥10% of tumor cells expressing receptors) or negative (<10% of tumor cells expressing receptors). Treatment was classified as surgery (breast-conservative surgery, mastectomy), radiotherapy (yes/no), chemotherapy (yes/no), and antiestrogens therapy (yes/no). During the study period, antiestrogen therapy consisted mainly of tamoxifen because aromatase inhibitors were prescribed in Switzerland only from 2004.

Patients were followed for cutaneous melanoma occurrence from 3 months after the date of breast cancer diagnosis until December 31, 2008. For each patient who developed a melanoma, information on the anatomic area where melanoma developed was extracted from medical files.

We compared patient and tumor characteristics among women with versus without antiestrogen treatment by χ² test of heterogeneity.

Person-years at risk were calculated to the date of end of follow-up, date of departure from the canton, date of skin melanoma diagnosis, or date of death, whatever came first.

Expected numbers of melanoma cases among the female Geneva population were calculated on the basis of the cantonal melanoma incidence rate for each 5-year age- and calendar year–group. We then calculated Standardized Incidence Ratios (SIR) by dividing the observed numbers by the expected numbers. Statistical significance and 95% CIs were estimated assuming a Poisson distribution. We calculated SIRs separately for women with and without antiestrogen therapy.

Characteristics of melanoma patients who did and did not use antiestrogens were compared with the Fisher exact tests. Melanoma incidence hazards among women with and without antiestrogen therapy were assessed by the Kaplan–Meier method.

Characteristics of the 7,360 breast cancer patients are presented in Table 1. Antiestrogen therapy was administered to 3,358 (54.4%) patients and among these, 2,941 (87.6%) were treated during the most recent period of the study (1991–2005). Compared with women without antiestrogen therapy, those with therapy were significantly older (mean age at diagnosis 59.8 and 63.7 years, respectively; P < 0.001). Social classes were similarly distributed in the 2 groups.

As expected, tumors of women who received antiestrogen therapy were more likely to express estrogens and progesterone receptors. Women with antiestrogen therapy more often underwent radiotherapy and less often underwent surgery or chemotherapy than women without antiestrogens.

The proportion of breast cancer patients who left Geneva and were lost to follow-up was not different among women with antiestrogen therapy (8.2%) and those without therapy (10.8%). The median follow-up period for the whole cohort was 7.0 years (6.7 years among antiestrogens users and 7.4 among nonusers, P < 0.0001). The cohort yielded 62,610 person-years (24,466 among antiestrogens users and 38,144 in nonusers). From July 1980 until December 2008, 34 breast cancer patients developed subsequent cutaneous melanoma, 11 among women with antiestrogens and 23 among women without antiestrogens (Table 2). The melanoma incidence rates per 100,000 person-years were 45.0 versus 60.3, respectively (P = 0.296).

SIRs of secondary melanoma according to antiestrogen use are described in Table 2, and curves of melanoma occurrence by time elapsed at least 3 months after breast cancer diagnosis are displayed in Fig. 1. All but 1 melanoma was diagnosed within 4 years of breast cancer diagnosis among women with antiestrogen therapy, whereas melanoma diagnoses continued for 14 years among those without therapy (Fig. 1). Compared with the general population, the SIR of melanoma was 1.32 (95% CI: 0.93–1.71, P = 0.07) for all breast cancer patients (Table 2). It was increased among women who did not receive antiestrogens (SIR: 1.60, 95% CI: 1.08–2.12, P = 0.02). On the contrary, it was close to 1 (SIR: 0.98, 95% CI: 0.40–1.56, P = 0.57) for women with antiestrogens. Similar results were obtained in an analysis restricted to the period 1990 to 2008 (SIR: 2.81, 95% IC: 1.56–4.68, P = 0.001 for women without antiestrogens and SIR: 1.44, 95% CI: 0.73–2.57, P = 0.165 for women with antiestrogens).

Table 3 presents characteristics of malignant melanoma occurring among women with versus without antiestrogens. No significant differences were observed for Breslow thickness, Clark level, histologic type, and anatomic site of melanoma. Also, the familial history of melanoma did not differ between the 2 groups.

This population-based study suggests that antiestrogen treatment for breast cancer can modify the risk of subsequent melanoma. We observed a 60% increased risk of melanoma among breast cancer patients who did not receive antiestrogens, whereas no excess of risk was found among women treated with antiestrogens.

Several previous studies have shown that breast cancer patients are at increased risk of developing secondary melanoma (15–23). These findings have been attributed to shared lifestyle and genetic risk factors, carcinogenic effect of treatment (in particular radiotherapy), and most commonly to better medical surveillance of women with previous breast cancer. Most of those studies addressing melanoma risk after breast cancer concerned patients treated in the distant past and therefore with treatment regimens not including hormonal therapy (15, 16), or lacked antiestrogen data or when available, did not evaluate if hormonal therapy influenced melanoma risk (17, 19, 20–22, 24). To our knowledge, only 2 prior studies evaluated melanoma cancer risk among breast cancer patients according to hormonal therapy (23, 24). The large study from the Netherlands of a population-based breast cancer cohort (23) showed an increased risk of secondary melanoma (SIR: 1.69, 95% CI: 1.44–1.97). As in our study, the excess was absent in women with hormonal therapy. A case–control study nested in a cohort (24) evaluated the relation between antiestrogen use and subsequent secondary cancers after breast cancer: a significant dose–response relationship between duration of antiestrogen treatment and cancer risk was limited to endometrial cancer. As only 14 cases of melanoma were observed in the cohort the lack of the association could be due to the low power of this study. Our results are also compatible with data from a hospital-based cohort reporting an excess of melanoma risk greater for premenopausal patients (19) who received antiestrogen treatment less frequently than older women.

However, we cannot draw conclusions on a putative modulation of melanoma risk by antiestrogens from those previous studies because of the sparse results so far available, together with differences in study design and studied populations.

Shared risk factors for breast cancer and melanoma could explain the increased risk of melanoma after breast cancer. The issue of whether exogenous and endogenous hormones influence melanoma risk has been investigated since the 1970s in numerous epidemiologic studies (25) with conflicting results. Most of these studies (25, 26), as well as a recent meta-analysis (27), did not support an association of pregnancy, oral contraceptives, and hormonal replacement therapy with melanoma.

In contrast, Koomen and colleagues (28) observed an increased risk of melanoma associated with estrogen use, with a cumulative dose-dependent effect, and with use of hormone replacement therapy in a study linking pathology and pharmacology databases from the Netherlands. Finally, a large prospective cohort study (29) reported significantly decreased risks of melanoma in women with late age at menarche, irregular menstrual cycles, earlier age at natural menopause, and in women with a shorter length of ovulatory cycle, suggesting a reduced melanoma risk associated with decreased exposure to ovarian hormones. These very recent results support our findings, that is, women with antiestrogen therapy would be less exposed to ovarian hormones and accordingly at lower melanoma risk than women without antiestrogen therapy.

The differences in melanoma risks found in our study for women with and without antiestrogen therapy are likely not explained by radiotherapy as this treatment was more often administered among women with antiestrogens. Similarly, they are likely not explained by differences in the length and periods of follow-up between the 2 groups as shown by the similar results obtained with the analysis restricted to 1990–2008. However, the lack of melanoma diagnoses at longer length of follow-up may be due to the small number of persons remaining at risk in the antiestrogen group (Fig. 1).

When interpreting our results, one must consider the strengths and limitations of the study. The cohort included a large number of unselected breast cancer patients obtained from a population-based cancer registry and the follow-up for secondary melanoma is virtually complete. Therefore, this study provides reliable estimates of melanoma risks in general population. However, use of cancer registry data has some limitations. Information on oral contraceptives, hormonal replacement therapy, antiestrogen characteristics (types, doses, and duration) as well as phototype, nevi number, and history of sun exposure were not available. Also, we cannot rule out that the significant association between melanoma and antiestrogen therapy may be due to chance given the small number of secondary melanoma.

Our hypothesis was that if estrogens play a role in melanoma development, then antiestrogens would affect melanoma incidence. Our results suggest that breast cancer patients who did not receive antiestrogens could be at increased risk of subsequent melanoma. These findings are in agreement with the hypothesis that estrogens may play a role in melanoma occurrence. However, they were based on a small number of melanomas and need to be replicated in large-scale studies with data on potential confounders such as previous clinical trials assessing the efficacy of antiestrogens on breast cancer outcomes.

No potential conflicts of interest were disclosed.

This work was supported by grant no. OCS-02070-04-2007 from the Fondation Recherche Suisse contre le cancer.

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.