The influence of occupational physical activity on early pregnancy failure was examined in a population of 24 cleaners, 36 kitchen staff and 110 clerical workers from 39 Dutch hospitals who were enrolled before becoming pregnant between August 1987 and January 1989. The occupational energy expenditure was defined by an intensity and a fatigue score, which were studied alone and in combination with working hours and working speed. The occupational biomechanic load was defined by a peak and a chronic pressure score. The intensity and the fatigue score of the work, whether or not in combination with working hours and working speed, were not related to the occurrence of spontaneous abortion. Work involving a high biomechanic load, in particular high peak pressure scores, showed an odds ratio (OR) for spontaneous abortion of 3.1 (95% confidence interval [CI]: 1.1-8.9). There were strong indications that work involving bending (OR = 3.2; 95% CI: 1.3-9.8) rather than lifting (OR = 1.1; 95% CI: 0.3-3.4) was the main cause.