Abstract
Objectives: To predict the role of different clinical and biochemical parameters in identifying nonalcoholic fatty liver disease (NAFLD) among patients with type 2 diabetes mellitus (T2DM) in Abha city, southwestern Saudi Arabia. Methods: A stratified random sample was selected. A detailed clinical and biochemical examinations were performed. Using portable abdominal ultrasound examination, NAFLD was identified. The study used receiver operating characteristic (ROC) analysis. Results: The study covered 237 T2DM patients. NAFLD was detected among 174 patients. Area under the curve (AUC) calculations showed that the ability of age, duration of DM in years, and body mass index to predict NAFLD was poor (AUC < 0.6). Similarly, biochemical factors like HbA1c%, AST, cholesterol, triglycerides, HDL, LDL, and VLDL were poor in discriminating between those with and without NAFLD among T2DM. On the other hand, the ability of ALT to predict NAFLD among T2DM was good (AUC = 0.701, 95% CI: 0.637-0.761). The analysis identified the optimal cutoff point of ALT to be <= 22.1 nmol/L. The corresponding sensitivity was 60.7% (95% CI: 53.0-68.0) and specificity was 62.5% (95% CI: 49.5-74.3). Conclusions: Early identification of NAFLD among T2DM is important. A threshold cutoff value of 22.1 nmol/L of ALT has been identified to predict NAFLD. They should be referred for ultrasound examination for NAFLD.