Objectives The aim of the study was to estimate associations between aerobic fitness among men at age 18–19 years and work absence due to musculoskeletal sickness 5–15 years later.

Methods All 321 975 men born between 1967 and 1976 in Norway were identified and followed up in several national registers. Men who completed an aerobic fitness test at military conscription during 1985–1995 (N=227 201) were followed from 2000 through 2003 with respect to a first musculoskeletal absence. Cox regression was conducted to estimate HRs between aerobic fitness (high, medium, poor) and musculoskeletal absence.

Results A total of 26 061 men had a musculoskeletal absence (absolute risk 0.115). Absence was associated with fitness level. Associations were confounded by other conscript characteristics (intellectual capacity, body mass index, musculoskeletal condition) and parental education level and were restricted to non-injury absence. With high fitness as reference, the adjusted non-injury HR estimates were 1.18 (95% CI 1.12 to 1.24) and 1.39 (1.31 to 1.47) for medium and poor fitness, respectively. Poor fitness men were more likely to achieve low educational attainment and employment in high-absence industries and enterprises. The impact of intellectual capacity and parental education level on absence was considerably larger than the effect from fitness. A subset analysis with fitness data of better quality yielded moderately stronger associations.

Conclusions Aerobic fitness among conscripts was moderately associated with non-injury musculoskeletal absence 5–15 years later. However, the overall impact of intellectual capacity and parental education appears to be greater than that of aerobic fitness.