BACKGROUND: Cross-sectional studies have reported a lower prevalence of abdominal obese persons among frequent drinkers than among nonfrequent drinkers. OBJECTIVE: We tested the hypothesis that drinking frequency is associated with subsequent changes in waist circumference. DESIGN: Data come from a prospective cohort study conducted in 1993-1997 (baseline) and 1999-2002 (follow-up) and included 43 543 men and women. Baseline information on alcohol drinking frequency was related to 1) change in waist circumference by linear regression and 2) major gain and major loss in waist circumference (defined as waist change in the lowest or highest quintile of waist changes) by polytomous logistic regression, also taking into account amount of alcohol intake. RESULTS: Drinking frequency was inversely associated with changes in waist circumference in women and was unassociated with changes in waist circumference in men. Drinking frequency was unassociated with major waist loss but was inversely associated with major waist gain: odds ratios among men were 0.97 (95% CI: 0.73, 1.28), 0.95 (95% CI: 0.81, 1.12), 0.88 (95% CI: 0.77, 0.99), 0.82 (95% CI: 0.71, -0.95), and 0.79 (95% CI: 0.69, 0.9) for never drinking, drinking on 1, 2-4, 5-6, and 7 d/wk, respectively, compared with men who drank alcohol on <1 d/wk (P for trend < 0.0001). Results for women were similar. Adjustment for the amount of alcohol intake or total energy intake did not affect results considerably. CONCLUSIONS: Drinking pattern may be associated with development of abdominal obesity; in this prospective study, drinking frequency was inversely associated with major waist gain and was unassociated with major waist loss
Udgivelsesdato: 2008/4