We read the recent article on experimental evidence of nonalcoholic fatty liver disease (NAFLD) exacerbation by tobacco exposure1 and the accompanying editorial2 with interest. As the editorialist correctly pointed out, the question is whether the findings of Azzalini et al.1 have clinical relevance, that is, whether tobacco is associated with NAFLD severity in humans. We have recently shown that heavy smoking is independently associated with liver steatosis and severe fibrosis in patients with chronic hepatitis C, and we have thus provided the first clinical evidence of a link between tobacco exposure and induction of steatosis.3 In order to test this hypothesis in nonalcoholic steatohepatitis (NASH), we investigated the effect of smoking on liver histological lesions in a cohort of 58 consecutive patients with biopsy-proven NASH. Our cohort and methods have been previously described.3, 4 Each patient completed a smoking questionnaire on the day of liver biopsy, and this included the age at which the patient started to smoke or stopped smoking, the duration of smoking, and the number of cigarettes smoked per day. Tobacco consumption was quantified as pack-years (i.e., the average number of packs per day multiplied by the number of years as a smoker). Heavy smokers were considered to be patients with a lifetime consumption of 20 pack-years or more. A single liver pathologist blindly evaluated all biopsy samples according to the classification system proposed by Brunt et al.5 Baseline patient characteristics are shown in Table 1. In all, 36% of patients were smokers, whereas 24% were heavy smokers. In univariate analysis, severe fibrosis was associated with increasing age (45.1 ± 14.3 versus 60.6 ± 9.2 years, P = 0.001) and body mass index (28 ± 3.6 versus 31.1 ± 6.5 kg/m2, P = 0.033), histological grade (1.4 ± 0.8 versus 2.7 ± 0.5, P < 0.001), and smoking (13/45 versus 8/13, P = 0.049), and there was a tendency for an association with heavy smoking (8/45 versus 6/13, P = 0.062). In multivariate analysis, severe fibrosis was independently associated with a higher histological grade (odds ratio = 24.6, P < 0.001), and there was a trend of an association with smoking (odds ratio = 6.645, P = 0.059). When heavy smoking was entered into the multivariate model instead of smoking, it was not associated with severe fibrosis. We further analyzed the potential association of smoking with steatosis severity, but no significant associations were found (data not shown). Our results for a relatively small cohort of patients with histologically proven NASH support the experimental evidence for the association of tobacco exposure with NAFLD and suggest that smoking might aggravate histological lesions. Moreover, it seems that all smokers, rather than those who smoke in excess, are at an increased risk of more severe fibrosis. We believe that a type II error, due to our small sample size, may account for the trend of an association between smoking and severe fibrosis in the multivariate analysis. Our results do not support an association of smoking with the degree of steatosis. However, it has been shown that the degree of steatosis is not associated with fibrosis severity in NASH and that steatosis might actually diminish in cases with cirrhosis.6 Because our cohort was rather small, these results need to be confirmed in larger series of patients. Emmanuel A. Tsochatzis M.D.*, George V. Papatheodoridis M.D.*, * Second Department of Internal Medicine, Athens University Medical School, Hippokration General Hospital, Athens, Greece.