Background:
Bictegravir (BIC) has been shown to be non-inferior to dolutegravir (DTG) as initial antiretroviral therapy (ART) in non-inferiority randomized controlled trials (RCTs). Mexico established in 2019 a national policy to start ART with BIC to every ART naïve person with HIV (PWH). However, BIC has not been compared to efavirenz (EFV) or raltegravir (RAL) (commonly prescribed in Mexico) and no superiority trials have been performed.
Methods:
We used data from the national antiretroviral surveillance system in Mexico to perform three target trial (TT) emulations for initial ART among PWH starting ART from 2019 onwards: BIC vs DTG, BIC vs EFV, and BIC vs RAL. Our approach consists of specifying the protocol of an index TT (the RCT we would ideally conduct) and emulate it with observational data while adjusting for confounding to approach randomization. Baseline was the date of ART start. People with missing viral load (VL), VL <500 copies/mm3, and/or missing CD4 count within the last year before ART start, those started on double-dose DTG and/or reduced-dose emtricitabine, lamivudine or tenofovir disoproxil fumarate were excluded. We performed individual logistic regressions while adjusting for treatment-outcome confounders (baseline VL and CD4, age, gender, imprisonment status, state of residence, year of ART start) to estimate risk of undetectable VL (UVL) (<50 copies/mm3) for each person at 3 and 12 months. We performed intention-to-treat analyses. We calculated the mean risk of UVL for each group, risk ratios (RR) and 95% percentile-based confidence intervals using 500 bootstrap resamples.
Results:
26178 PWH were included: 18894 started BIC, 2530 DTG, 562 RAL, and 4192 EFV. Most PWH were cis men (85%), median age was 31 years. 57% had a baseline viral load <100k copies, and median CD4 cell count was of 213. Only 1% of PWH were in a prison. At three months, the risk of UVL was of 79.2% vs 78.4% (BIC vs DTG, RR 1.01 [0.98-1.03]), 79.7% vs 71% (BIC vs RAL, RR 1.12 [1.03-1.22]), and 79.7% vs 63.6% (BIC vs EFV, RR 1.25 [1.21-1.3]). At twelve months, the risk of UVL was of 85.9% vs 83.6% (BIC vs DTG, RR 1.02 [0.98-1.08]), 85.9% vs 79% (BIC vs RAL, RR 1.08 [1-1.23]), and 86.6% vs 83% (BIC vs EFV, RR 1.04 [1.01-1.08]).
Conclusions:
BIC is superior to RAL and EFV, but not DTG, as initial ART in treatment naïve PWH for reaching UVL at 3 and 12 months. These results support existing RCTs and argue in favor of this type of country-wide strategy.