Investigating hepatic steatosis: the MISHTI study (Multicentric cross-sectional Indian Study of Hepatic and Metabolic Trends in India)

Background and Aim: Hepatic fibrosis is a critical complication of metabolic disorders, particularly in patients with Type 2 Diabetes (T2D). This study aimed to evaluate the performance of the Fibrosis-4 index (Fib 4) score in detecting significant fibrosis (transient elastography [TE] ≥ 8 kPa) and identify key predictors of advanced fibrosis using logistic regression analysis in patients with T2D. Methods: This retrospective study included propensity-matched T2D and non-T2D patients. Sensitivity, specificity, and Cohen's Kappa were used to assess agreement between Fib 4 score ≥ 1.3 and TE ≥ 8 kPa. Logistic regression models were used to identify independent predictors of significant fibrosis. The predictive performance of the models was evaluated using ROC curves. Results: The Fib 4 score demonstrated high sensitivity (85.3%) but low specificity (13.7%) in the T2D cohort, with a Cohen’s Kappa of -0.01, indicating no agreement with TE. In the non-T2D cohort, specificity improved to 47.2% with a Cohen’s Kappa of 0.16. Logistic regression identified BMI, hypertension, and HbA1c as significant predictors of hepatic fibrosis in T2D patients, with odds ratios of 1.076, 1.824, and 1.279, respectively. Male sex, BMI, AST, and HbA1c were retained in the refined multivariate model, achieving an AUC of 0.735, indicating good discriminatory ability. Elevated transaminases were weakly associated with fibrosis, while BMI and HbA1c showed stronger associations. Conclusion: While Fib 4 is sensitive for detecting significant fibrosis, its low specificity limits its utility as a standalone diagnostic test, particularly in T2D patients. Logistic regression highlighted BMI, AST, and HbA1c as key predictors of fibrosis, emphasising the need to combine non-invasive tools with clinical variables for more accurate risk stratification and improved management of MASLD. Future research should focus on refining diagnostic algorithms to better address the burden of advanced fibrosis in at-risk populations.