Impact of alcohol consumption and body mass index on mortality from nonneoplastic liver diseases, upper aerodigestive tract cancers, and alcohol use disorders in Korean older middle-aged men

Abstract

Supplemental Digital Content is available in the text

Body mass index and gastrointestinal cancer mortality in Korean adults: A prospective cohort study

“…Therefore, reverse causality and residual confounding by smoking are unlikely to be mechanisms behind our findings. Higher alcohol consumption relative to bodyweight in persons with a lower BMI may partially explain the findings …”

“…Several potential mechanisms might be suggested to explain these inverse associations. First, the effects of risk factors for GI cancers such as alcohol and smoking may be stronger in lean persons due to the higher concentrations of hazardous substances such as alcohol relative to bodyweight than in obese persons . Second, reverse causality and residual confounding by smoking may partially explain these associations.…”

“…Body mass index values were grouped into seven categories (kg/m 2 : <18.5, 18.5–20.9, 21.0–22.9, 23.0–24.9 [reference], 25.0–27.4, 27.5–29.9, and ≥30) using cut points established by the World Health Organization . A cubic spline transformation of BMI with three knots (5th, 50th, 95th percentile; 19.4, 23.9, and 29.0 kg/m 2 ) was used to examine nonlinear associations.…”

“…Body mass index values were grouped into seven categories (kg/m 2 : <18.5, 18.5-20.9, 21.0-22.9, 23.0-24.9 [reference], 25.0-27.4, 27.5-29.9, and ≥30) using cut points established by the World Health Organization. [12][13][14] A cubic spline transformation of BMI with three knots (5th, 50th, 95th percentile; 19.4, 23.9, and 29.0 kg/m 2 ) was used to examine nonlinear associations. BMI was analyzed as a continuous variable to estimate the effect of each 5 kg/m 2 increase or decrease on GI cancer mortality within the ranges of 10-24.9 kg/m 2 (termed the "lower range"), 25-50 kg/m 2 (termed the "upper range"), and 10-50 kg/m 2 (termed the "full range"), assuming linear associations within each range.…”

Interaction Between Alcohol Use and Metabolic Risk Factors for Liver Disease: A Critical Review of Epidemiological Studies

“…Only one study considered waist circumference in addition to BMI (Åberg et al, ). BMI < 25 kg/m 2 was used as the reference group in interaction analyses in 8 of 15 studies; 3 studies used BMI < 30 kg/m 2 as reference, one study used BMI 25 to 27.4 kg/m 2 (Yi et al, ), and 2 studies did not provide sufficient data for relative risks compared to a reference group. Four studies explicitly stated to exclude underweight subjects from the reference group.…”