Body mass index modifies the association between frequency of alcohol consumption and incidence of hypertension in men but not in women: a retrospective cohort study.

Alcohol consumption is one of the major modifiable risk factors of hypertension. The aim of the present retrospective study was to assess the clinical impact of obesity on the association between alcohol consumption and the incidence of hypertension. The present study included 5116 male and 6077 female university employees with a median age of 32 (interquartile range 27-39) who underwent annual health checkups between January 2005 and March 2013. Self-reported drinking frequency was recorded at their first checkup and categorized into rarely and 1-3, 4-6, and 7 days/week. During the median observational period of 4.9 years (interquartile range 2.1-8.3), hypertension, defined as systolic/diastolic blood pressure of ≥140/90 mmHg and/or self-reported treatment for hypertension, was observed in 1067 (20.9%) men and 384 (6.3%) women. Poisson regression models adjusted for clinically relevant factors revealed a dose-dependent association between drinking frequency and the incidence of hypertension in men (adjusted incidence rate ratio [95% confidence interval] of men who drank rarely, 1-3, 4-6, and 7 days/week was 1.00 [reference], 1.12 [0.97-1.30], 1.42 [1.19-1.70], and 1.35 [1.14-1.59], respectively; Ptrend < 0.001). However, this association was not observed in women. The dose-dependent association was significant in nonobese men (body mass index (BMI) < 25 kg/m2), but not in obese men (BMI ≥25 kg/m2) (P for interaction between drinking frequency and BMI = 0.072). The present study provides clinically useful evidence to identify the drinkers who may reap the health benefits of abstinence from alcohol consumption.