Accuracy of Noninvasive Indices for Diagnosing Hepatic Steatosis Compared to Controlled Attenuated Parameter (CAP) in a Lebanese Cohort.

Abstract

Background Noninvasive liver indices are valuable tools to identify patients with NAFLD in view of their simplicity and relatively low cost. Aims This study examined the accuracy of several published indices, including hepatic steatosis index (HSI), Framingham steatosis index (FSI), and Dallas steatosis index (DSI) in predicting steatosis compared to Transient elastography controlled attenuation parameter (CAP). Methods This is a retrospective study on 720 patients who underwent Fibro scan at the American University of Beirut hepatobiliary unit between 2015 and 2020. Data on demographics, comorbidities, anthropometrics, laboratory studies were obtained. The following indices: HSI, FSI, DSI were calculated. Sensitivity, specificity and C -statistics of each index was determined and compared to CAP. Results Out of 550 patients who met inclusion criteria ,67% were found to have hepatic steatosis with average CAP 289.5(± 70.9). Mean age 50 (± 15.6) and 70% were male. Those patients were found to have higher BMI, and higher prevalence of diabetes and HTN as well as dyslipidemia than the control group with no steatosis. The FSI out-performed the other 2 indices with fair accuracy in detecting hepatic steatosis based on the C-statistics. HSI and DSI on the other hand showed poor discriminatory ability. AUC: FSI 0.77 95%CI (0.637–0.805), HSI 0.674 95%CI (0.577–0.771), DSI 0.651 95% CI (0.543–0.759). At the cut-off where specificity was 90%, the FSI Had highest sensitivity 49.6% compared to DSI 28.6% and HSI 21.8%. Conclusion Our study demonstrated the accuracy of the non-invasive index FSI in diagnosing steatosis when compared to Transient elastography controlled attenuation parameter (CAP), and highlighted its superiority to other indices the HSI and DSI. Further studies with larger and more heterogeneous population are needed to confirm our results.