A 12-year follow-up study was carried out to evaluate the risk factors for unspecified low-back pain and sciatic pain. Middle-aged farmers who did not report any low-back or neck-shoulder pain during the previous year in a postal questionnaire in 1979 were included in the follow-up study in 1992 (n = 537). In total, 366 (68%) of these farmers were interviewed by telephone. In 1992, the one-year prevalence rates of unspecified low-back pain (13.3%) and sciatic pain (9.6%) were low. Full-time farmers had a significantly higher prevalence of sciatic pain than did part-time or retired farmers. In the logistic regression modelling of sciatic pain in men, the odds ratio was 9.6 (95% CI: 2.7-65.2) for current smokers and 13.1 (95% CI: 1.7-53.0) for ex-smokers as compared to never smokers. Mental stress, body height, body mass index and production factors did not predict unspecified low-back pain or sciatic pain. Farmers who are free of back problems in middle age seem to stay well in the long run despite the heavy work. The close association between smoking and the prevalence of sciatic pain found in men is in line with the hypothesis that smoking is a causal risk factor for lumbar disc disease.