Drinking pattern as a predictor of cardiovascular harm: A longitudinal study using alternative drinking pattern measures

Background: This study compared measures of drinking pattern at baseline, and subsequent cardiovascular harm in a longitudinal study.

Method: In Winnipeg, Manitoba, Canada, a community sample of 1154 adult men and women was interviewed at baseline in 1990 and 1991, then followed with all‐cause surveillance. Cox proportional hazards regressions were used to assess the “time to event” for morbidity or mortality from coronary heart disease (CHD), hypertension, or other cardiovascular disease. Surveillance was through a 10‐year series of documented physician visits, hospital discharges and deaths, classified by diagnosis. Drinking pattern was defined as either ?8 drinks (80–120 g of alcohol or more) at a sitting in the past 12 months, a report of feeling the effects, or ?5 usual drinks/day.

Results: There were 104 individuals with CHD events in the data. When ?8 drinks at a sitting was the predictor, there were significant hazards for CHD among both men and women [hazard ratio (HR) = 2.32 and 1.07; p = 0.004 and 0.04], and marginally significant hazards for hypertension among men (HR = 1.40; p = 0.08). When feeling the effects or ?5 usual drinks/day were the predictors, there were no significant hazards of drinking pattern.

Conclusion: Eight or more drinks was a stronger predictor of cardiovascular harm in these data than were feeling the effects or ?5 usual drinks.