To the Editor In their article in JAMA Psychiatry, Georgakis et al1 report an odds ratio of 0.98 (95% CI, 0.96-0.99) for a 2-year increment in age at menopause in association with depression. I was surprised that the authors attribute this minor effect to estrogen exposure and that they recommend estrogen-based therapies to treat depression in menopausal women. I wonder how the authors came to such conclusions. In outcomes with rather low probability, the odds ratio corresponds to the risk ratio; thus, we may infer that in women who enter menopause 2 years after the population average, their relative risk is 2% smaller (hence, 0.98). Assuming that women with average age at menopause have a risk of 10% for major depression, then women with a 2-year prolonged age at menopause would have a risk of 9.8% (0.098/0.10 = 0.98). Thus, the absolute risk reduction is only 0.2%. Even when extrapolated to a 10-year increment, which is approximately twice the SD of age at menopause, the absolute risk reduction is only about 1%.
Hengartner MP. Estrogen-Based Therapies and Depression in Women Who Naturally Enter Menopause Before Population Average. JAMA Psychiatry. 2016;73(8):874. doi:10.1001/jamapsychiatry.2016.0709