Effectiveness of integrase strand transfer inhibitors in HIV-infected treatment-experienced individuals across Europe

Objectives To explore the effectiveness and durability of integrase strand transfer inhibitor (INSTI)-based regimens in pre-treated subjects. Methods Treatment-experienced individuals starting an INSTI-based regimen during 2012-2019 were selected from the INTEGRATE collaborative study. The time to virological failure [VF: one measurement of viral load (VL) >= 1000 copies/mL or two >= 50 copies/ml or one VL measurement >= 50 copies/mL followed by treatment change] and to INSTI discontinuation were evaluated. Results Of 13 560 treatments analysed, 4284 were from INSTI-naive, non-viraemic (IN-NV) individuals, 1465 were from INSTI-naive, viraemic (IN-V) individuals, 6016 were from INSTI-experienced, non-viraemic (IE-NV) individuals and 1795 were from INSTI-experienced, viraemic (IE-V) individuals. Major INSTI drug resistance mutations (DRMs) were previously detected in 4/519 (0.8%) IN-NV, 3/394 (0.8%) IN-V, 7/1510 (0.5%) IE-NV and 25/935 (2.7%) IE-V individuals. The 1-year estimated probabilities of VF were 3.1% [95% confidence interval (CI): 2.5-3.8] in IN-NV, 18.4% (95% CI: 15.8-21.2) in IN-V, 4.2% (95% CI: 3.6-4.9) in IE-NV and 23.9% (95% CI: 20.9-26.9) in IE-V subjects. The 1-year estimated probabilities of INSTI discontinuation were 12.1% (95% CI: 11.1-13.0) in IN-NV, 19.6% (95% CI: 17.5-21.6) in IN-V, 10.8% (95% CI: 10.0-11.6) in IE-NV and 21.7% (95% CI: 19.7-23.5) in IE-V subjects. Conclusions Both VF and INSTI discontinuation occur at substantial rates in viraemic subjects. Detection of DRMs in a proportion of INSTI-experienced individuals makes INSTI resistance testing mandatory after failure.