LAUSR

Direct‐acting antiviral (DAA) failure, which is mainly associated with the selection of resistance‐associated substitutions (RASs), is not rare in HCV treatment. RAS data collected from published literature and RAS prevalence were integrated using meta‐analysis. DAA‐failure‐associated RASs were identified by comparing their prevalence between DAA‐failure and DAA‐naïve patients. Prevalences of emerging RASs that occurred during treatment were also estimated. A total of 2932 DAA‐naïve patients and 1466 DAA‐failure patients were included. Significant differences in the prevalence of RASs were found in 76 scenarios that involved 34 RASs (11 in NS3, 18 in NS5A and 5 in NS5B), 4 genotypes (GTs) (GT1a, GT1b and GT3‐4) and 14 DAAs (6 NS3 protease inhibitors [PIs], 6 NS5A inhibitors and 2 NS5B inhibitors). For NS3, the DAA‐failure‐associated RASs included V36L, Y56H, Q80K/R, R155K, A156T and D168A/E/L/T/V/Y. Substitutions at R155 and D168 were dominant for most NS3 PIs. For NS5A, DAA‐failure‐associated RASs included K24R, Q30R, L31M, and P32L in GT1a; R30Q/H, L31F/I/M/V, P58S, and Y93H in GT1b; A30K, L31M and Y93H in GT3; and M31V and Y93H in GT4. Y93H was the most prevalent RAS for NS5A inhibitors. DAA‐failure‐associated RASs were found at only five positions in NS5B. The majority of DAA‐failure patients relapsed. A significant difference was detected for only four RAS sites between relapse patients and nonresponse/breakthrough patients. The RAS prevalence in DAA‐failure patients varied among the HCV GTs and DAA regimens. The identified treatment‐selected resistance patterns for broadly used DAA regimens will enable the selection of optimized retreatment options.