VanWagner et al1 performed a post hoc analysis from the CARDIA cohort study (Coronary Artery Risk Development in Young Adults Study) and found that 58% of the patients with nonalcoholic fatty liver disease self-reported “moderate alcohol” use. This finding deserves robust comment. First, self-reported alcohol use is poorly reliable and the relevance of the binary classification (drinker/nondrinker) for a 25-year period of time when questioning people only 3 times (at 15, 20, and 25 years) only for the amount drunk over the last month, contrasted with the sophistication of the radiologic investigations.