Development and validation of the HABIT score: a practical and robust tool to predict hepatic steatosis using HbA1c, BMI, and triglycerides in patients with unexplained elevated liver enzymes

INTRODUCTION AND OBJECTIVES: Unexplained elevated liver enzymes (ELE) are common and often linked to metabolic dysfunction-associated steatotic liver disease (MASLD). Diagnosis in primary care is difficult due to limited access to advanced imaging, highlighting the need for simple non-invasive tools. This study aimed to develop and validate a pragmatic, laboratory-based score to predict hepatic steatosis in patients with unexplained ELE. PATIENTS AND METHODS: A derivation cohort of 206 patients with 40 classified as steatosis-positive with a NAFLD Activity Score (NAS) ≥4 was analysed. Candidate predictors were tested by univariate analysis and entered into multivariate logistic regression. Regression coefficients were scaled to construct an integer-based score (HABIT). Validation was performed in a retrospective cohort of patients with unexplained ELE attending our hepatology clinic between 2019 and 2021 (n = 648), all undergoing standardized diagnostic work-up including ultrasound, elastography, and laboratory testing. RESULTS: Multivariate analysis identified HbA1c, body mass index (BMI), and triglycerides as independent predictors of steatosis. The HABIT score was derived as: (61 × HbA1c [%]) + (3 × BMI [kg/m²]) + Triglycerides [mg/dL]. Diagnostic accuracy was high in the derivation cohort (AUROC: 0.83) and robust in validation (AUROC: 0.81), outperforming the Hepatic Steatosis Index (HSI, AUROC: 0.77). Cut-offs <490 and >630 reliably ruled out (sensitivity 93.6 %) or confirmed (specificity 92.5 %) steatosis. Score values correlated significantly with ultrasound grades and CAP. CONCLUSIONS: The HABIT score is a simple, robust tool for identifying hepatic steatosis in unexplained ELE. It outperforms HSI and may facilitate early detection and risk stratification in primary care with limited diagnostic resources.