Evaluation Of The Fatty Liver Index As A Diagnostic Predictor For Non-Alcoholic Fatty Liver Disease In Type 2 Diabetics

Abstract

The Fatty Liver Index (FLI) has demonstrated utility in diagnosing hepatic steatosis. However, in some middle-income settings such as Peru, its predictive performance has not been validated, nor have optimal cut-off points been established. This study evaluates the FLI as a diagnostic predictor for non-alcoholic fatty liver disease in patients with type 2 diabetes at a tertiary hospital in northern Peru. Materials and Methods: A prospective external validation study of the FLI model was conducted with 175 outpatients at Chimbote Regional Hospital, Peru, during November–December 2024. Using the data collected and the published model equation, the FLI score was calculated for each patient. The model’s predictive performance was assessed through measures of calibration, discrimination, and classification. Results: Among the selected patients, 74.2% were diagnosed with metabolic dysfunction-associated fatty liver disease (MAFLD). The FLI model demonstrated good discriminative ability, with an area under the curve (AUC) of 0.89 (95% CI: 0.83-0.94) and a calibration slope of 1. In our population, an FLI score <30 had a negative likelihood ratio of 0.08, effectively ruling out MAFLD, whereas an FLI score ≥70 produced a positive likelihood ratio of 10.71, confirming the diagnosis. Conclusions: In our study, the FLI model showed excellent predictive performance for MAFLD. Its simplicity and low cost may make it a practical tool in low- and middle-income countries for identifying diabetic patients who require hepatic ultrasound.