The Ability of the Framingham Steatosis Index (FSI) to Predict Non-alcoholic Fatty Liver Disease (NAFLD): A Cohort Study

Background The utilization of indexes for the diagnosis of non-alcoholic fatty liver disease (NAFLD) can be valuable. This study was conducted to determine the ability of the Framingham steatosis index (FSI) to distinguish between people with NAFLD and those without and to predict people at risk of NAFLD to establish the need for lifestyle modifications in such individuals. Methods Our study was conducted in two phases from 2009-2010 (phase I) to 2016-2017 (phase II). A total of 4670 people in northern Iran were included. NAFLD was diagnosed by ultrasound. The FSI was calculated based on age, sex, hypertension, diabetes mellitus status, liver enzyme levels and triglyceride levels. Receiver operating characteristic (ROC) analysis was conducted to determine the discriminatory and predictive abilities of the FSI. To remove the confounding effects of potential mediators, logistic regression was performed in which NAFLD was considered the outcome and the FSI as the predictor. Results The odds ratios of the FSI when the outcome was the prevalence of NAFLD in phase I and when the outcome was new cases of NAFLD from 2009–2010 to 2016-2017 were 4.909 (4.243–5.681) and 2.453 (2.024–2.972), respectively (P<0.001). The areas under the curve (AUCs) for the discriminatory and predictive abilities of the FSI were 0.8421 (95% CI: 0.8314–0.8527) and 0.7093 (95% CI: 0.6863–0.7322), respectively. Conclusion The FSI has a strong ability to diagnose NAFLD while it has an acceptable ability to predict the occurrence of new cases of NAFLD.