External validation of serological scores for the detection of liver steatosis among people with HIV

Abstract

Abstract Background Fatty liver index (FLI) and hepatic steatosis index (HSI) are serological scores used to detect liver steatosis. However, their diagnostic performance in people with HIV (PWH) remains unclear. We performed an external validation of FLI and HSI in the Swiss HIV Cohort Study (SHCS). Methods We systematically performed vibration controlled transient elastography (VCTE) among SHCS participants at Bern University Hospital between November 2019 and August 2021. Individuals with viral hepatitis and pregnant women were excluded. We defined liver steatosis as controlled attenuation parameter (CAP) ≥248dB/m using VCTE. Model discrimination was assessed using the c-index, and model calibration using calibration plots. A decision curve analysis was performed to compare the clinical usefulness of both scores. Results Of 321 participants, 91 (28.4%) were female, median age was 51.4 years (interquartile range 42-59), 230 (71.7%) were Caucasian, and 164 (51.1%) had a body mass index >25kg/m2. VCTE-confirmed liver steatosis was present in 158 (49.2%) participants. Overall, 125 individuals (38.9%) had an FLI ≥60, and 128 (39.9%) had an HSI ≥36. At these cut-offs, the c-index to diagnose liver steatosis was 0.85 for FLI (95% confidence interval [CI] 0.80-0.89), and 0.78 for HSI (95% CI 0.73-0.83). Whereas FLI was well calibrated, HSI overestimated the risk for steatosis. Both models showed a positive net benefit, with FLI having a greater net benefit compared with HSI. Conclusions FLI and HSI are valid tools to detect liver steatosis in PWH. FLI should be the preferred score, given its better performance and greater clinical usefulness.