Introduction Several observations have led to the hypothesis of a hormonal dependence of cutaneous melanoma. However, previous research yielded conflicting results as to the influence of reproductive and hormonal factors on melanoma risk, and most findings relied on case-control designs.
Methods To assess the role of these factors on the risk of melanoma, we conducted a prospective analysis of 91,972 French women from the National Education System, aged 40–65 years at inclusion into the E3N cohort study. Data on reproductive and hormonal factors were collected at baseline and were updated every two/three years. Between 1990 and 2005, 460 incident first primary melanoma cases were reported and confirmed through pathology reports. Relative risks (RR) were computed using Cox proportional hazards regression models, adjusted for host factors (number of naevi, freckling, skin colour, hair colour, skin sensitivity to sun exposure), mean daily ultraviolet radiation dose in regions of birth and inclusion, history of endometriosis, history of uterine fibroma, age at menarche, menstrual cycle length, parity, age at menopause, and educational level.
Results We observed significantly inverse relationships between melanoma risk and late age at menarche (RR=0.68, 95% confidence interval (CI)=0.47–0.99; for menarche ≥15 years compared to 13–15 years), and irregularity of menstrual cycles (RR=0.54, 95% CI=0.31–0.91; compared to regular cycles of 25–31 days periodicity). Early age at natural menopause was marginally significantly associated with risk (RR=0.72, 95% CI=0.51–1.02, for menopause <48 years compared to 48–52 years). Other reproductive and hormonal factors (parity, number of pregnancies, age at first live birth or pregnancy, or history of miscarriages) were not significantly associated with melanoma risk.
Conclusion Our findings suggest an inverse association between melanoma and some menstrual factors in women. These results from a large prospective cohort provide some evidence for a hormonal-dependent etiological component of this cancer. Biological mechanisms that could explain these associations remain unclear and should be further explored in experimental and epidemiological studies.