Aims: This study was designed to assess the potentially confounding influences of social integration and depression on the form of the relationship between alcohol consumption and all-cause mortality. DESIGN, PARTICIPANTS AND MEASUREMENT: Respondents from the 1984 US National Alcohol Survey (N = 5177) were followed by searching the National Death Index (NDI) through 1995; 540 were identified as deceased. Predictor variables in a Cox proportional hazards model included gender, ethnicity, marital status, income, smoking, age and alcohol consumption (volume and patterns). Two social variables and their interactions with alcohol consumption were added, the Center for Epidemiological Studies Depression (CES-D) scale and an eight-item social isolation scale.
Findings: The J-shaped risk curve for all-cause mortality by volume was approximated for men but not significantly for women. In addition heavy drinking occasions independently contributed to mortality in men. Low social integration (bottom 12%) had no significant effects on mortality or on the relationship between alcohol consumption and mortality curve. Inclusion of the interaction between alcohol consumption and depression proved significant for heavy male drinkers (> six drinks on average per day) and for female former drinkers with heavy drinking occasions. In both cases, the respective subgroup, which additionally was depressed, had about four times the risk of a life-time abstainer.
Conclusions: The relationship of alcohol consumption to 11-year all-cause mortality in a general population indicated little confounding effect of social isolation, but revealed important interactions with depression for heavy male drinkers and heavy female ex-drinkers.