Research ArticleComparison of liver fibrosis blood tests developed for HCV with new specific tests in HIV/HCV co-infection

Background & AimsWe compared 5 non-specific and 2 specific blood tests for liver fibrosis in HCV/HIV co-infection.MethodsFour hundred and sixty-seven patients were included into derivation (n = 183) or validation (n = 284) populations. Within these populations, the diagnostic target, significant fibrosis (Metavir F ⩾2), was found in 66% and 72% of the patients, respectively. Two new fibrosis tests, FibroMeter HICV and HICV test, were constructed in the derivation population.ResultsUnadjusted AUROCs in the derivation population were: APRI: 0.716, Fib-4: 0.722, Fibrotest: 0.778, Hepascore: 0.779, FibroMeter: 0.783, HICV test: 0.822, FibroMeter HICV: 0.828. AUROCs adjusted on classification and distribution of fibrosis stages in a reference population showed similar values in both populations. FibroMeter, FibroMeter HICV and HICV test had the highest correct classification rates in F0/1 and F3/4 (which account for high predictive values): 77-79% vs. 70-72% in the other tests (p = 0.002). Reliable individual diagnosis based on predictive values ⩾90% distinguished three test categories: poorly reliable: Fib-4 (2.4% of patients), APRI (8.9%); moderately reliable: Fibrotest (25.4%), FibroMeter (26.6%), Hepascore (30.2%); acceptably reliable: HICV test (40.2%), FibroMeter HICV (45.6%) (p < 10−3 between tests). FibroMeter HICV classified all patients into four reliable diagnosis intervals (⩽F1, F1 ± 1, ⩾F1, ⩾F2) with an overall accuracy of 93% vs. 79% (p < 10−3) for a binary diagnosis of significant fibrosis.ConclusionsTests designed for HCV infections are less effective in HIV/HCV infections. A specific test, like FibroMeter HICV, was the most interesting test for diagnostic accuracy, correct classification profile, and a reliable diagnosis. With reliable diagnosis intervals, liver biopsy can therefore be avoided in all patients.