The aim of this study was to examine whether the volume and intensity of physical activity are associated with subsequent sickness absence spells of different lengths, and how much of these associations can be explained by socioeconomic position, body mass index (BMI) and physical health functioning. Baseline data were collected by questionnaire surveys in 2000-2002 among 40-60-year-old employees of Helsinki City (n=6465, 79% women). Sickness absence data were derived from the employer's registers (mean follow-up time 3.9 years). Associations of physical activity with shorter (< or =14 days) and longer (>14 days) sickness absence spells were examined, using Poisson's regression analysis. The volume of physical activity was weakly and somewhat inconsistently associated with sickness absence. However, men and women who were vigorously active systematically had reduced risk of sickness absence, whereas the same volume of moderately intensive physical activity did not reduce the risk of sickness absence. Adjusting for BMI and in particular physical health functioning attenuated these associations, after which the associations lost statistical significance. The results suggest that vigorous physical activity is associated with sickness absence and may contribute to better work ability.