Recent cross-sectional studies have been reported the possibility that light to moderate alcohol consumption might be negatively associated with fatty liver. However, there has been no large-scale longitudinal study addressing an impact of alcohol consumption on a development of fatty liver diagnosed by ultrasonography. Thus, we investigated the impact of alcohol consumption on a natural history of fatty liver.We analyzed 5437 apparently healthy Japanese who received the health checkup programs repeatedly over 10 years. In this study, we used a standardized questionnaire for addressing the medical history and lifestyle and used a standardized ultrasonographic diagnosis for fatty liver. The total amount of alcohol consumed per week was calculated and classified into four grades; none or minimal, light, moderate, or heavy alcohol consumption (< 40, 40-140, 140-280 or > 280 g/week, respectively). The hazard risks of alcohol consumption for the development of fatty liver were calculated by Cox hazard model after adjusting age, BMI, and parameters for lifestyle.During 10 years of follow-up, fatty liver was continuously diagnosed just in 10

作者：Yoshitaka, Hashimoto；Masahide, Hamaguchi；Takao, Kojima；Yasuhiro, Ohshima；Akihiro, Ohbora；Takahiro, Kato；Naoto, Nakamura；Michiaki, Fukui

来源：Journal of gastroenterology and hepatology 2015 年 30卷 3期