Background:  Clinical stability has been observed with continued HAART in the setting of partial virologic suppression. A5115 sought to examine the immunologic and virologic consequences of an immediate vs. delayed HAART switch.

Methods: A5115 was a randomized 2-arm pilot study that compared 2 switch strategies in subjects on HAART with stable HIV RNA between 200 and 10,000 copies/mL:  immediate switch at study entry vs delayed switch when HIV RNA ≥10,000 copies/mL or CD4 count 20% below baseline. The primary objectives were to evaluate the level of immune activation measured by the proportion of CD8⁺ T cells expressing CD38⁺ at week 48, and to compare the evolution of genotypic drug resistance at week 48 using a metric that quantifies the number of susceptible drugs at any given time-point (Future Drug Options). The primary immune activation non-inferiority comparison was an intent-to-treat analysis using the Hodges-Lehmann confidence interval (HLCI). The Wilcoxon rank-sum test was used for the FDO comparison.

Results:  We enrolled 47 subjects and followed the for a median of 82 weeks. At baseline, the median CD4 count was 421 cells/mm³ and the median HIV RNA was 3.4 log₁₀ copies/mL; 27(57%) had failed ≥2 HAART regimens. The median (IQR) baseline FDO score was 2.3 (2.3, 3.0) for subjects with 1 prior HAART regimen and 2.0 (1.0, 2.3) for those with ≥2 HAART failures (p = 0.005). In the delayed switch arm, 10 of 23 (43%) subjects met switch criteria (CD4 decline, 7; HIV RNA ≥10,000, 3). The median (IQR) FDO at week 48 (n = 37) was 2.3 (1.0, 2.3) for the immediate switch group and 2.3 (2.0, 3.0) for the delayed switch group (p = 0.21). Only 1 subject (delayed switch arm) had a reduction in FDO by ≥1 prior to switching. No change from baseline FDO was observed for the remaining subjects. New mutations occurred in 4 of 11 sequences in the immediate switch group and 12 of 22 sequences in the delayed switch group. Most of the mutations were minor with the exception of M184V, L90M (1 each), and T215X (2). The majority of the mutations (3 of 4 immediate switch, 8 of 12 delayed switch) were for nucleoside reverse transcriptase inhibitor (NRTI) resistance. At week 48, the median proportion of CD8⁺-expressing CD38⁺ was 73%. HLCI for difference in median proportion of CD8⁺-expressing CD38⁺ was (–7% delayed, 10% immediate switch), thus not sufficient to claim either non-inferiority or inferiority.

Conclusions:  This study of HIV-infected individuals with stable, low-level viremia was under-powered, but results show that delay of switching HAART may not be associated with changes in immune activation or affect treatment options as measured by the FDO score.