A comparison of three highly active antiretroviral treatment strategies consisting of non-nucleoside reverse transcriptase inhibitors, protease inhibitors, or both in the presence of nucleoside reverse transcriptase inhibitors as initial therapy (CPCRA 058 FIRST Study): a long-term randomised trial

Background: Long-term data from randomised trials on the consequences of treatment with a protease inhibitor (PI), non-nucleoside reverse transcriptase inhibitor (NNRTI), or both are lacking. Here, we report results from the FIRST trial, which compared initial treatment strategies for clinical, immunological, and virological outcomes. Methods: Between 1999 and 2002, 1397 antiretroviral-treatment-naive patients, presenting at 18 clinical trial units with 80 research sites in the USA, were randomly assigned in a ratio of 1:1:1 to a protease inhibitor (PI) strategy (PI plus nucleoside reverse transcriptase inhibitor [NRTI]; n=470), a non-nucleoside reverse transcriptase inhibitor (NNRTI) strategy (NNRTI plus NRTI; n=463), or a three-class strategy (PI plus NNRTI plus NRTI; n=464). Primary endpoints were a composite of an AIDS-defining event, death, or CD4 cell count decline to less than 200 cells per mm³ for the PI versus NNRTI comparison, and average change in CD4 cell count at or after 32 months for the three-class versus combined two-class comparison. Analyses were by intention-to-treat. This study is registered with ClinicalTrials.gov, number NCT00000922. Findings: 1397 patients were assessed for the composite endpoint. A total of 388 participants developed the composite endpoint, 302 developed AIDS or died, and 188 died. NNRTI versus PI hazard ratios (HRs) for the composite endpoint, for AIDS or death, for death, and for virological failure were 1·02 (95% CI 0·79-1·31), 1·07 (0·80-1·41), 0·95 (0·66-1·37), and 0·66 (0·56-0·78), respectively. 1196 patients were assessed for the three-class versus combined two-class primary endpoint. Mean change in CD4 cell count at or after 32 months was +234 cells per mm³ and +227 cells per mm³ for the three-class and the combined two-class strategies (p=0·62), respectively. HRs (three-class vs combined two-class) for AIDS or death and virological failure were 1·15 (0·91-1·45) and 0·87 (0·75-1·00), respectively. HRs (three-class vs combined two-class) for AIDS or death were similar for participants with baseline CD4 cell counts of 200 cells per mm³ or less and of more than 200 cells per mm³ (p=0·38 for interaction), and for participants with baseline HIV RNA concentrations less than 100 000 copies per mL and 100 000 copies per mL or more (p=0·26 for interaction). Participants assigned the three-class strategy were significantly more likely to discontinue treatment because of toxic effects than were those assigned to the two-class strategies (HR 1·58; p<0·0001). Interpretation: Initial treatment with either an NNRTI-based regimen or a PI-based regimen, but not both together, is a good strategy for long-term antiretroviral management in treatment-naive patients with HIV.