Development and validation of a clinical and laboratory-based nomogram to predict nonalcoholic fatty liver disease

Background and aim Nonalcoholic fatty liver disease (NAFLD) is becoming the leading cause of chronic liver disease in China. The early identification and management of patients at risk are essential. We aimed to develop a novel clinical and laboratory-based nomogram (CLN) model to predict NAFLD with high accuracy. Methods We designed a retrospective cross-sectional study and enrolled 21,468 participants (16,468 testing and 5000 validation). Clinical information and laboratory/imaging results were retrieved. Significant variables independently associated with NAFLD were identified by a logistic regression model, and a NAFLD prediction CLN was constructed. The CLN was then compared with four existing NAFLD-related prediction models: the fatty liver index (FLI), the hepatic steatosis index (HSI), the visceral adiposity index (VAI) and the triglycerides and glucose (TyG) index. Area under the receiver operator characteristic curve (AUROC) and decision curve analysis (DCA) were performed. Results A total of 6261/16,468 (38.02%) and 1759/5000 (35.18%) participants in the testing and validation datasets, respectively, had ultrasound-proven NAFLD. Six variables were selected to build the CLN: body mass index (BMI), diastolic blood pressure (DBP), uric acid (UA), fasting blood glucose (FBG), triglyceride (TG), and alanine aminotransferase (ALT). The diagnostic accuracy of the CLN for NAFLD (AUROC 0.857, 95% CI 0.852–0.863) was significantly superior to that of the FLI (AUROC 0.849, 95% CI 0.843–0.855), the VAI (AUROC 0.752, 95% CI 0.745–0.760), the HSI (AUROC 0.828, 95% CI 0.822–0.834), and the TyG index (AUROC 0.774, 95% CI 0.767–0.781) (all p