This study is based on data gathered by means of a postal questionnaire from a cohort of 640 38-year-old subjects. At the age of 14 years these subjects had been interviewed by their school doctor to ascertain whether any of them had a history of low back pain (LBP), and X-rays of the thoracic and lumbar spine were taken. The questionnaire contained related groups of questions, with LBP as the main topic. We wanted to identify probable risk factors in developing LBP. The results show a cumulative life-time prevalence of LBP of 70%, a 1-year prevalence of 63% and a point prevalence of 19%, independent of gender. Women reported a higher incidence of LBP than men during the month and week before they filled out the questionnaire, they also reported a higher incidence of sciatica and greater use of the health care system and analgestics over the previous year. Heavy manual work was associated with LBP and sciatica, and smoking (more than 16 cigarettes per day) was more common among unemployed and sick-listed subjects, Severe LBP was associated with increased morbidity, reduced work capacity, deterioration in social life, mental and sexual problems, and increased smoking. A stepwise logistic regression analysis of "early" independent variables indicated that severe LBP is positively correlated with low social class for men and with menstruation and pregnancy for women.