Cigarette smoking and the age at menopause.

In a sample of 15,464 apparently healthy Australian women, the mean age at natural menopause of women who smoke ten or more cigarettes a day was 1.3 years lower than that of other women. These distributions were well described by logistic distributions with the same variances. A parallel non-parametric Kaplan-Meier analysis confirmed that heavy smokers reached menopause earlier than other women, and also showed that the women accurately recalled their age at menopause. Logistic regression analyses were used to show that obesity, alcohol intake and regular use of aspirin and other analgesics, sedatives, tranquillizers or anti-depressants did not affect the distribution of the age at natural menopause. The age-specific probability of natural compared with surgically induced menopause was shown to be independent of smoking history and obesity. Logistic regression analysis is a powerful statistical technique for analysing data such as age-specific proportions of still menstrual women. It is easy to apply and interpret, and its aptness for any particular data set is readily assessed.