Efficacy and tolerability of initial antiretroviral therapy: a systematic review

Objectives: Successful antiretroviral therapy (ART) is largely attributed to the type of third drug. Guidelines recommend regimens without systematic review of all factors that might affect treatment Success, such as study design, eligibility criteria and participant characteristics. Why patients cease ART has not been systematically studied. Design and setting: Systematic review of initial ART studies (64 randomized, 15 cohort). Group-based analysis was by weighted, forward, stepwise, linear regression. Main outcome measures: Treatment efficacy (undetectable plasma HIV viral load by intention-to-treat) and cessation for adverse events. Results: Seven variables were independently associated with study groups reporting higher treatment Success (mean 59%, r(2)=0.79): nonwhite race (P=0.002), exclusion for low haemoglobin (P=0.0006), lower CD4 cell count (P=0.014), closing relative to food (P=0.001), dual-nucleoside backbone (favouring didanosine or tenofovir with emtricitabine or lamivudine; P=0.002), nonnucleoside analogue or ritonavir-boosted protease inhibitor as the third drug (P<0.0001), and shorter follow-up (P=0.0004). Although the most common cause of treatment cessation (9.0%), adverse events were reported in only half of studies and were significantly more likely in studies that were phase 2 or 3, academia-sponsored and less than 36 months duration, and in older participants. Nausea was the. only adverse event significantly associated with treatment Success (r=0.277). Conclusion: Multiple reasons influence initial ART success beyond the type of third drug and should be considered when designing and comparing studies. Most studies are too short and report insufficient adverse event data. Didanosine is an effective option for initial ART, with particular relevance to resource-limited settings. ART guideline development might benefit from systematic review. (c) 2009 Wolters Kluwer Health vertical bar Lippincott Williams & Wilkins