FLAMINGO: how much rosier can antiretroviral therapy get?

In the SPRING-2 trial, dolutegravir showed noninferiority to raltegravir, another integrase inhibitor, in antiretroviral-naive adults with HIV-1 infection. In the SINGLE trial, dolutegravir showed superiority over a non-nucleoside reverse transcriptase inhibitor, efavirenz. The next obvious step of the dolutegravir targeted strategy was to compare dolutegravir with a boosted protease inhibitor such as darunavir plus ritonavir. The comparison between dolutegravir and a boosted protease inhibitor is particularly interesting because boosted protease inhibitors are very potent and, by contrast with raltegravir and efavirenz, extremely resilien t to HIV resistance development even when used as monotherapy. In The Lancet, Bonaventura Clotet and colleagues present the FLAMINGO study, in which 484 antiretroviral-naive HIV-infected patients were randomly assigned to receive two nucleos(t)ide reverse transcriptase inhibitors (tenofovir–emtricitabine or abacavir–lamivudine) plus either dolutegravir (50 mg) or the boosted protease inhibitor darunavir (800 mg) plus ritonavir (100 mg). After 48 weeks of follow-up, more than 80% of patients in each group achieved virological suppression (217 [90%] patients receiving dolutegravir and 200 [83%] patients receiving darunavir plus ritonavir had HIV-1 RNA of less than 50 copies per mL, adjusted diff erence 7·1% [95% CI 0·9–13·2]), and no patient developed drug resistance. It is diffi cult to imagine a better outcome for a clinical trial in a disease that just two decades ago did not have an eff ective treatment. In a prespecifi ed secondary analysis, dolutegravir also showed superiority to darunavir plus ritonavir. The data suggest that superiority was driven both by better tolerability (nine [3%] patients in the dolutegravir group and 20 [9%] patients in the darunavir plus ritonavir group discontinued for non-virological reasons) and better effi cacy (virological success in 217 [90%] patients in the dolutegravir group vs 200 [83%] patients in the darunavir plus ritonavir group), especially in the 25% of patients who started with viral loads greater than 100 000 copies per mL. But is dolutegravir really superior to darunavir plus ritonavir? This is a diffi cult question to answer for various reasons. First, FLAMINGO is an open-label clinical trial with two pills a day taken by patients in the dolutegravir group versus four pills a day taken by those in the comparator group (darunavir plus ritonavir). The openlabel design might have led to patients, disappointed with their treatment assignment, choosing not to continue. In fact, six patients withdrew in the darunavir plus ritonavir group very early on compared with the one patient who withdrew in the dolutegravir group. Second, results cannot be automatically extrapolated to all antiretroviral-naive patients because most patients recruited were young white men whose median CD4 cell count was almost 400 cells per μL. This cell count was within the threshold for starting therapy for most guidelines but above the mean CD4 cell count at HIV diagnosis in patients in high-income countries (336 cells per μL). Third, integrase inhibitors tend to reduce viral load very rapidly, often reaching undetectability in less than 8 weeks, whereas protease inhibitors tend to cause a more gradual decrease. This eff ect could lead to patients starting with a very high viral load in the darunavir plus ritonavir group still having a detectable but falling viral load, even after 48 weeks. They might have reached a very low viraemia of 50–200 copies per mL but not virological suppression at that time, even though viral load might have subsequently become undetectable. The detailed viral load results of 48-week outcomes in patients with very high viral loads at initiation (>100 000 copies per mL and >500 000 copies per mL), and the 96-week study data are analyses important in understanding the outcomes. Fourth, since no patient developed resistance after virological failure in either group, the superiority of dolutegravir does not translate into an advantage with