Occupational physical activity and risk for breast cancer in a nationwide cohort study in Sweden

Objective: Our purpose was to investigate effects of physical activity on risk for breast cancer. Methods: From the Swedish nationwide censuses in 1960 and 1970 we defined three partly overlapping cohorts of women whose occupational titles allowed reproducible classification of physical demands at work in 1960 (n=704,904), in 1970 (n=982,270), or with the same demands in both 1960 and 1970 (n=253,336). The incidence of breast cancer during 1971-89 was ascertained through record linkage to the Swedish Cancer Register. We used Poisson regression to estimate relative risks (RR). Results: A total of 20,419, 22,840, and 8261 breast cancers, respectively, were detected in the three cohorts. In all three cohorts the risk for breast cancer increased monotonically with decreasing level of occupational physical activity and with increasing socioeconomic status. Among women with the same estimated physical activity level in 1960 and 1970 the RR was 1.3 for sedentary as compared with high/very high activity level (95% CI 1.2-1.4; p for trend < 0.001). Adjustment for socioeconomic status virtually eliminated this association (RR 1.1; 95% CI 0.9-1.2; p for trend 0.12) leaving a statistically significant 30% gradient only among women aged 50-59 years at follow-up. The association between socioeconomic status and breast cancer risk was largely unchanged after adjustment for occupational physical activity. Conclusion: The protective effect of occupational physical activity on breast cancer risk, if any, appears to be confined to certain age groups.