Using data from a British national cohort of women born in 1946, this study aims to identify menstrual patterns during the first year of perimenopause (based on the frequency of periods, the numbers of days bled each month, and menstrual flow) to see if they are related to health and behaviors earlier in adult life and if they predict entry into menopause and hormone replacement therapy (HRT) use. Three groups of women were identified using cluster analysis: those who experienced more of these characteristics, those who experienced less, and those who experienced few changes. In polychotomous logistic regression models, the likelihood ratio tests indicated that parity and body mass index (BMI) were significant at the 5% level. The odds ratios from the parity models showed a gradient, with women from the Less cluster being most likely to have no children and those from the More cluster most likely to have at least one child. A similar gradient was detected for BMI, with the Less cluster tending to be underweight. The Less cluster came into menopause significantly faster than the Same and the More groups, where the estimated hazard ratios (HR) (95% confidence interval [CI]) were, respectively, 0.61 (0.37-0.99) and 0.24 (0.11-0.52). There was no association between the clusters and later HRT use. The findings suggest that menstrual characteristics should be more carefully studied in population studies of the climacteric.