Social support and subsequent premature mortality, gender disparities and role of potential confounders.

To assess the association between social support and subsequent premature mortality (PM) (<70 years), gender difference, and the confounding role of age, occupation, health-related behaviors, obesity, and diseases. Methods: 4118 subjects (2189 men, 1929 women), aged >15 years, randomly selected in north-eastern France, completed in 1996 a postal questionnaire gathering characteristics, smoking, alcohol abuse, obesity, social support from colleagues/family/friends, and physician-diagnosed diseases. The cohort was followed-up until 2008 (12.5 years). Data were analyzed using Poisson models. Results: There were 165 PM (115 men, 50 women) during the follow-up. For all the cohort, social support was strongly associated with PM: crude relative risk (RR) 2.09 (95%CI 1.53-2.85). It decreased to 1.52 (1.11-2.09) when adjusted for age, 1.46 (1.06-2.00) with further adjustment for occupation, 1.44 (1.04-1.98) with further adjustment for smoking, alcohol abuse and obesity, and 1.28 (0.92-1.78) when diseases (cancer, diabetes, mental, nervous-system, cardiovascular, respiratory, genitourinary and musculoskeletal diseases) were taken into account. Similar results were found for men: crude RR 2.46 decreasing to 1.81, 1.70, 1.60, 1.45 respectively. But not for women: close-to-significant crude RR 1.57 (small number of PM). Conclusions: Social support influenced PM in men but not in women. Occupation, health behaviours, and diseases play a role.