OBJECTIVE: To assess the effects of duration of exposure to biomechanical strains on various types of low back pain (LBP). METHODS: The population study was a random sample from the GAZEL cohort. Durations of exposure to selected biomechanical strains during subjects' working lifetime and potential confounders were assessed in 1996 by self-administered questionnaires. Data on LBP in the previous 12 months were collected in 2001. Relations between various dimensions of LBP and durations of exposure to the biomechanical strains were analysed with multivariate regression models. Polytomous models were built to determine whether some biomechanical strains were specifically associated with some types of LBP. RESULTS: Analyses were performed separately for men (n = 2218) and women (n = 383). Significant associations were observed (ORs reported are those for 20 years of exposure) between LBP and durations of driving and bending/twisting for men (OR 1.24 and 1.37 respectively); LBP for more than 30 days and duration of exposure to bending/twisting for men and women (OR 2.20 and 2.00 respectively) and duration of driving for women (OR 3.15); LBP radiating to the leg and duration of driving for men (OR 1.43) and bending/twisting for women (OR 1.95); LBP radiating below the knee and duration of exposure to pulling/pushing/carrying for men (OR 1.88). Bending/twisting in both men and women, and driving for women appeared to be stronger risk factors for LBP for more than 30 days. Pulling/pushing/carrying heavy loads appeared to be a risk factor specific for LBP radiating below the knee for men. CONCLUSION: This study suggests that exposure to biomechanical strains has long-term effects and a dose-response relation with duration of exposure and specific effects for some types of LBP.