Abstract Body

Patients with an active opportunist disease or with -200 CD4+/µL and +5 log10 HIV-RNA copies/mL at antiretroviral therapy (ART) start are largely underrepresented in clinical trials; data from large observational studies may help bridging this knowledge gap. We aimed at studying the durability of different initial ART regimens in these patients.

All subjects enrolled in the ICONA Study, who started ART with 1 anchor drug (ritonavir or cobicistat-boosted protease inhibitor [bPI], or non-nucleoside reverse transcriptase inhibitor [NNRTI] or integrase strand transfer inhibitor [InSTI]) plus tenofovir(TDF)/emtricitabine(FTC) or abacavir(ABC)/lamivudine(3TC), CD4+ -200 cells/µL and HIV-RNA >5 log10 copies/mL, and at least 1 HIV-RNA assessed both before and after the start of ART, were included in this analysis. Primary endpoint: treatment failure (TF, defined as virological failure [VF, first of 2 consecutive HIV-RNA >50 copies/mL after >6 months of treatment] or discontinuation of class of the anchor drug) as assessed by Kaplan­Meier method and log-rank test. Independent associations were investigated by Poisson regression analysis, in a model including variables associated with TF at a p-value <.2 at univariate analysis: anchor drug, baseline (BL) HIV-RNA, CDC C stage, HCV co-infection, CD4+, FIB-4, eGRF, ongoing opportunistic disease, nucleos(t)ide backbone.

1127 patients fulfilled inclusion criteria: 729 started ART with a bPI, 305 with an InSTI and 193 with a NNRTI (95% EFV). Their median (IQR) BL CD4+, CD4+/CD8+, HIV-RNA were 63 (27-125) cells/µL, .11 (.05-.2), 5.55 (5.3-5.87) log10 copies/mL. PYFU were 519, 1533, 264 in the NNRTI, bPI and InSTI group; incidence rates (IRs,95%CI) of TF were 18.1 (14.8-22.2), 29.1 (26.5-31.9), 20.8 (16-27.1) per 100 PYFU and IRs of VF were 5.9 (4.4-8.1), 8.0 (6.9-9.3), 5.2 (3.1-8.7) per 100 PYFU in the NNRTI, bPI and InSTI group. Cumulative probabilities of TF and VF are illustrated in figure. At multivariable analysis, compared to those based on bPIs, regimens based on NNRTIs (IRR .65 [.52-.82]; p less than.001) or InSTIs (.7 [.52-.92];p=.012) were associated with a lower risk of TF; BL HIV-RNA >500K (1.38 [1.17-1.63] compared to less than 500K; p<.001) was associated with a higher risk of TF. The type of regimens was not independently associates with VF.

In patients starting ART with -200 CD4+ and +5 log10 HIV-RNA, the durability of regimens based on EFV or InSTIs was longer than that of boosted PI-based regimens.