Predicting age at menopause.

This article reviews methodologic and clinical aspects of predicting age at menopause. Lifetable methods or logistic models applied to a perimenopausal population represent the most feasible and the least biased methods for estimating the probability of menopause by age. Information is emerging about risk factors besides age which influence risk for an earlier menopause and include a variety of medical, demographic, environmental, and genetic factors. The concept of menopause as a consequence of depleted oocytes suggests that the estimated number of ovulatory cycles might also be a useful predictor. Using these variables in a logistic model yields estimated probabilities of menopause for various risk profiles. Smokers who have accumulated more than 10 pack-years, women estimated to have had more than 300 ovulatory cycles, women with a history of depression, women who have lost one ovary at an early age, and women who have a family history of early menopause have earlier menopause and the greatest shift in the cumulative probability of menopause occurs in women with multiple risk factors.