OBJECTIVES: The purpose of this study was to investigate whether a previous fracture is a risk factor for fracture later in life and its association with tobacco smoking, menopausal age and the use of oral contraceptives. This is an epidemiological population study based on a questionnaire. A sample of 10000 women from seven birth cohorts between 1900-1940 was obtained at random from the population register. The overall response rate was 74.6%. The responders (n = 7459) represented 53% of the total population. METHODS: The women from all these birth cohorts were questioned regarding a possible history of fractures and the year in which they were sustained. The questionnaire also included questions about menopausal age, use of oral contraceptives, previous gynaecological operations and possible oestrogen medication. The women from the 1900-1920 birth cohorts were questioned in detail about urogenital disorders, while tobacco smoking data were recorded only for the 1930 and 1940 birth cohorts. RESULTS: The relative risk of sustaining a further fracture was significantly related to fracture premenopausally before 40 years of age and later fracture postmenopausally after 60 years of age (1.29; 0.97-1.70) compared to controls (0.78; 0.59-1.03)(P = 0.03). Both an early menopause and tobacco smoking were associated significantly with repeated fractures, while use of oral contraceptives had a protective effect against repeated fractures in the 1940 birth cohort. Logistic multiple regression showed that a fracture was a significant independent risk factor for further fracture in both cohorts 1900-1920 (P