BACKGROUND & AIMS Physical activity (PA) is inversely associated with nonalcoholic fatty liver disease (NAFLD) prevalence. However, few studies evaluated the effect of PA on NAFLD incidence in regard to visceral adipose tissue (VAT) and insulin resistance (IR). We investigated whether PA at baseline and change in PA during follow-up have any effect on incident NAFLD.

METHODS We enrolled subjects who underwent health screenings between 2007 and 2008 and participated in voluntary follow-up between 2011 and 2013 (median 4.42 years). Incident NAFLD was defined as NAFLD absence at baseline and presence at follow-up by ultrasonography. PA was measured using a detailed questionnaire-based metabolic equivalent at baseline and follow-up; the difference during follow-up was calculated.

RESULTS Of the 1373 subjects enrolled, 288 (21.0%) developed NAFLD. Both total and leisure-time PA at baseline were inversely associated with incident NAFLD (P for trend=.005 and .003 respectively). Decreased PA at follow-up was associated with increased incident NAFLD risk after adjusting for age, gender, body mass index, smoking, hypertension, diabetes and diet [hazard ratio (HR) 1.45, 95% confidence interval (CI) 1.04-2.02, 4th (most decreased PA) vs 1st quartile (increased PA), P=.028]. This relationship was attenuated but remained statistically significant after adjustment for VAT(HR 1.48, 95% CI 1.06-2.06, 4th vs 1st quartile) and IR(HR 1.59, 95% CI 1.11-2.27, 4th vs 1st quartile).

CONCLUSIONS This study shows an independent protective effect of PA at baseline on incident NAFLD after 4-year follow-up. Furthermore, sustained or increased PA had a preventive effect on incident NAFLD independent of VAT and IR.