Study of Options for Second-Line Effective Combination Therapy (SELECT)

As more HIV-infected persons start antiretroviral therapy (ART) worldwide, the number needing second-line therapy is increasing. In many settings, an NNRTI-based regimen is the preferred first-line ART, whereas a protease inhibitor (PI)-based regimen is often reserved for second-line ART. Both of these types of regimens usually include two NRTIs.

As with initial ART, second-line regimens ideally should be composed of three fully active drugs that have potent anti-HIV activity to maximize the chances of durable viral suppression. However, such a goal may not be achieved with second-line PI-based regimens containing NRTIs because of resistance mutations from first-line therapy that reduce the activity of the NRTI class. The most significant of these mutations include M184V, thymidine analogue mutations (TAMs), Q151M complex, and K65R. The presence of K65R would result in resistance to most NRTIs (leaving only zidovudine (ZDV) and possibly abacavir (ABC) as active second-line options); the presence of multiple TAMs and/or Q151M alone or in complex with other mutations would also result in resistance to most NRTIs.

Recent data suggest that patients with isolated M184V-related NRTI resistance who subsequently switch to a boosted PI plus lamivudine (3TC)- or emtricitabine (FTC)-based regimen may achieve HIV-1 RNA suppression without the need to switch to more complex regimens [1]. Detection of an isolated M184V NRTI mutation is possible when resources allow for early diagnosis of virologic failure. However, in many resource-limited settings (RLS) early diagnosis of virologic failure is difficult because of infrequent monitoring of viral load or unavailability of viral load testing. This study intended to provide information applicable to the vast majority of RLS where resistance testing is not used routinely for selection of second-line regimens and PIs not needing refrigeration are preferred.

This was a phase III, dual-arm, open-label, randomized, non-inferiority study for participants who were on a failing NNRTI-containing first-line regimen. The study evaluated the difference in virologic failure rate between two treatment arms:

Arm A: LPV/r plus RAL Arm B: LPV/r plus best available NRTIs

Best available NRTI combinations were selected by the site investigator prior to randomization from a list of combinations approved by the study or in consultation with the A5273 Clinical Management Committee (CMC). The NRTIs provided by the study were FTC/TDF, ABC/3TC/ZDV, ABC/3TC, 3TC/ZDV, ABC, 3TC, and ZDV.

Participants took their assigned study therapy until 96 weeks of follow-up or 52 weeks after the last participant was enrolled (protocol amendment), whichever was earlier. This study originally planned to enroll 600 participants (300 per study arm), but the sample size was re-evaluated to 480 participants (240 per study arm), due to emergent data from other studies. Participants were assigned with equal probability to one of the two treatment regimens using permuted blocks. Randomization was stratified by three factors: screening HIV-1 RNA (≥100,000 versus <100,000 copies/mL); screening CD4+ cell count (≥100 versus <100 cells/mm3), and selection of ZDV in the NRTI regimen (yes/no). Furthermore, randomization was balanced by site.

During the study, participants were asked to return to the clinic at Weeks 4, 12, 24, and then every 12 weeks. Visits lasted about 30 minutes. At most visits, participants had a physical exam, had their arm, waist and hip circumference measured, and answered questions about their medical condition and any medications they were taking. At some visits, participants completed questionnaires to see how they were feeling, if they had been hospitalized recently, and how well they were taking their anti-HIV drugs, had their blood drawn and were asked to give urine samples. If the participant was female and able to become pregnant, a pregnancy test was taken at any visit that pregnancy was suspected.