The purpose of this study was to undertake a critical review of the potential role of magnetic resonance imaging (MRI) in the evaluation of low back pain (LBP) and to determine if there were differences in the MRI appearances between various occupational groups. The study group, 149 working men (78 aged 20-30 years and 71 aged 31-58 years) from five different occupations (car production workers, ambulance men, office staff, hospital porters and brewery draymen), underwent MRI of the lumbar spine. Thirty-four percent of the subjects had never experienced LBP. Twelve months later, the examination was repeated on 89 men. Age-related differences were seen in the MRI appearances of the lumbar spine. Disc degeneration was most common at L5/S1 and was significantly more prevalent (P < 0.01) in the older age group (52%) than in the younger age group (27%). Although LBP was more prevalent in the older subjects there was no relationship between LBP and disc degeneration. No differences in the MRI appearance of the lumbar spine were observed between the five occupational groups. Overall, 45% had 'abnormal' lumbar spines (evidence of disc degeneration, disc bulging or protrusion, facet hypertrophy, or nerve root compression). There was not a clear relationship between the MRI appearance of the lumbar spine and LBP. Thirty-two percent of asymptomatic subjects had 'abnormal' lumbar spines and 47% of all the subjects who had experienced LBP had 'normal' lumbar spines. During the 12-month follow-up period, 13 subjects experienced LBP for the first time. However, there was no change in the MRI appearances of their lumbar spines that could account for the onset of LBP. Although MRI is an excellent technique for evaluating the lumbar spine, this study shows that it does not provide a suitable pre-employment screening technique capable of identifying those at risk of LBP.