Second generation INSTIs currently represent the most highly recommended option for first-line ART but superiority to boosted-PI regimens in people with advanced HIV disease (CD4 count <200 cells/mm3 or AIDS), generally underrepresented or excluded from RCTs, has not been demonstrated.
We included ART-naïve patients with CD4 count <200 cells/mm3 or AIDS diagnosis in the Icona Foundation Cohort between 2014-2018, who started a dolutegravir [DTG] or boosted-darunavir [DRV/b (ritonavir or cobicistat)] based ART. We estimated the effect of the difference in risk of a composite endpoint (death, AIDS, serious non-AIDS events – SNAE – viral failure >200 copies/mL, anchor drug discontinuation not followed by a restart of a drug in the same class) between the two strategies using a marginal structural model. We accounted for differences in prognostic factors measured at time of ART initiation. We also accounted for differences in censoring by these same prognostic factors, and time-varying CD4, HIV-RNA and ALT.
Characteristics of the 685 ART-naïve patients were (DTG=416; DRV/b=269; 224 DRV/r and 45 DRV/cobi): male 87%; heterosexual contacts 50%, MSM 37%; born outside Italy 48%; AIDS presenting 36%; median CD4 count 78 cell/mm3 (IQR 30-140); median HIV-RNA 5.25 log10 copies/mL (IQR 4.64, 5.73). All these variables were comparable between the two groups, except for higher proportion of migrant in DTG (51% vs 43%; p <0.001) and higher HIV-RNA values in DRV/b (5.35 vs 5.18 log10/mL; p=0.019). NRTI backbone was TDF/FTC in 58% (80% in DRV/b and 44% in DTG), TAF/FTC in 11% (10% in DRV/b and 12% in DTG), and ABC/3TC in 30% (10% in DRV/b and 44% in DTG) (p <0.001). 116 patients receiving DTG and 145 receiving DRV/b experienced the composite endpoint. The 1-year weighted probability of the composite endpoint was 37% for DRV/b and 21% for DTG (Figure 1a). Patients who initiated DTG were at lower risk of experiencing the composite endpoint compared to those who started DRV/b [aHR 0.50 (95%CI 0.32, 0.79)] (Figure 1b). Calendar year of starting was a key factor but results were consistent across periods of ART initiation.
Under the assumptions of no unmeasured confounding and correct model specification, our analysis suggests that a RCT conducted in the target population of ART-naïve patients with CD4 count<200 or AIDS is likely to show a notable reduction in risk of treatment failure in people initiated with dolutegravir vs. boosted-darunavir based therapies.