World Health Organization guidelines recommend switching nucleoside reverse transcriptase inhibitors (NRTIs) in second-line antiretroviral therapy (ART) after first-line failure. Recent clinical trial data suggest that recycling NRTIs in protease inhibitor (PI)-based second-line ART may have similar efficacy to switched NRTIs. We evaluate this question using programmatic data from East Africa.
We analyzed data from the East Africa International Epidemiology Databases to Evaluate AIDS, which includes programmatic data from public sector clinics in Kenya, Tanzania, and Uganda. We included adults (age ?18 years) with HIV who were switched to atazanavir or lopinavir-based second-line ART after virologic failure on non-nucleoside reverse transcriptase inhibitor (NNRTI)-based first-line ART containing zidovudine (AZT) or tenofovir (TDF). Individuals were included if a switch to second-line ART occurred after introduction of routine viral load testing at the relevant clinic and if ?1 year of post-switch observation time was available. Our outcome of interest was 1-year crude cumulative incidence of viral suppression (HIV-1 RNA <1,000 copies/mL) after switch. We compared individuals with switched versus recycled NRTIs in their second-line regimen, accounting for competing risks of death, lost to follow-up, or transfer. Among those with recycled NRTIs, we compared cumulative incidence of viral suppression for those with recycled TDF versus AZT via Gray's test.
Of 3,240 participants analyzed, median age was 40 (IQR 33 – 47) at the time of switch, and 66% were female. Only 7% (n = 212/3,240) had recycled NRTIs in their second-line regimen, of which 79% (n = 167/212) used recycled TDF. Crude cumulative incidence of viral suppression one year after switch to second-line was 60% (95% CI 53-67%) among those with recycled NRTIs and 69% (95% CI 67-70%) in those with switched NRTIs (p-value = 0.019). Among those with recycled NRTIs, there was no difference in viral suppression rates for those with recycled TDF versus AZT (p-value = 0.901). Limitations include possible confounding by indication since clinical reasons for use of recycled NRTI regimens for participants in this analysis are not known.
In programmatic care in East Africa, we found improved rates of virologic suppression among individuals switching NRTIs in PI-based second-line ART. Recommendations for second-line PI-based regimens with recycled NRTIs should be made with close observation of clinical outcomes.