Daily alcohol consumption and sickness absence in the GAZEL cohort.

Background: : Previous studies that examined the association between daily alcohol consumption and sickness absences (SA) were mostly retrospective and did not take into account the characteristics of SA.  : A total of 9907 daily drinkers (8442 men and 1465 women) of the GAZEL prospective cohort were included. Daily alcohol consumption over the three previous years was self-reported at baseline and categorized as low, moderate, high or very high risk according to the World Health Organization. Duration of SA (short: ≤7 days; moderate: 8-28; long: >28) was collected from administrative records as well as causes for long SA. Negative binomial regression models were used to estimate Risk Ratios of SA according to alcohol consumption with low-risk category as reference.  : Duration of follow-up (in years) for SA was 8.4 ± 3.7 in men and 11.2 ± 5.4 in women. Increasing alcohol consumption predicted increasing risk of SA with a dose-response relationship ( P < 0.01 for men; P = 0.01 for women). In men, strength of this association increased with SA duration [e.g. RRs from 1.41 (95% CI: 1.12-1.79) to 2.12 (95% CI: 1.49-3.00) in the very high-risk category, for short and long SA, respectively]. In men, even a moderate consumption predicted increased risk of SA whatever their duration (RR = 1.15; 95% CI: 1.07-1.23). In women, a moderate consumption predicted only long SA (RR = 1.22; 95% CI: 1.00-1.50). Daily alcohol consumption was associated with almost all causes of long SA in men, and with respiratory diseases, digestive diseases and injury in women.  : We found a dose-response relationship between daily alcohol consumption and the risk of SA. Even moderate consumption could increase this risk, particularly in men.