Early initiation of antiretroviral therapy: debate over?

In Kevin De Cock and Wafaa El-Sadr’s Comment, they address the importance of early antiretroviral therapy (ART) in the treatment of HIV. That ART substantially reduces HIV transmission in serodiscordant couples and is the most eff ective strategy in curbing the HIV epidemic is not debatable. However, the statement that the question of when to start treatment has been resolved on the basis of the TEMPRANO and STAR trials surprises us. The TEMPRANO trial only shows that deferred ART plus 6-month isoniazid preventive therapy is as eff ective as early ART in avoiding death or severe HIV-related illness in sub-Saharan Africa. In STAR, 4685 patients with a CD4 cell count higher than 500 cells per μL from 35 countries (many with high prevalence of tuberculosis) were randomised to immediate ART or to defer it until their CD4 cell count decreased to 350 cells per μL. In practice, up to 5% of patients (n=118) in the deferred group started ART with a CD4 cell count of 178 cells per μL or lower, which is well below the proposed threshold and could have biased the incidence of opportunistic events in this group. Furthermore, the design of this trial combined several events, not always related to the HIV infection itself, into a single primary endpoint, thus complicating the interpretation of results that are only reported by intention-to-treat analysis. These potentially unrelated events included non-AIDS cancer (eg, bladder, prostate, and gastric adenocarcinoma) and death due to liver disease, diabetes, substance abuse, or suicide, which were more frequent in the deferred-initiation group. The incidence of tuberculosis was also higher in the deferred group, although tuberculosis should not be considered an AIDS-defi ning illness in areas with high tuberculosis prevalence in the absence of immunosuppression. Information about CD4 cell counts when each event occurred and other complementary efficacy analyses would undoubtedly be of great value for clinicians who have to balance the safety and benefi ts of initiating ART in asymptomatic patients who have high or even normal CD4 cell counts. The conclusion reached based on countries with a high prevalence of tuberculosis might not be directly applicable to developed countries. Moreover, the absolute risk reduction of the composite primary endpoint was only 2·2%, which suggests that even if ART were routinely started early, a large number of patients would need to be treated for a long time to avoid one event not always related to HIV infection itself. Therefore, we believe that early ART has not demonstrated clear clinical benefi ts.