Results
All week 96 visits were completed by October 31, 2011. Baseline demographics and subject characteristics have been detailed previously and were balanced between the 2 groups. Darunavir was the most widely used PI (58% EVG and 59% RAL subjects), and the second background agent was usually a nucleoside reverse transcriptase inhibitor (NRTI) (tenofovir disoproxil fumarate [TDF]-containing 72.4% EVG and 67.2% RAL). Nearly two thirds of subjects had baseline mutations to 2 or more classes of agents. From weeks 48 to 96, an additional 33 and 29 subjects discontinued study drug in the EVG and RAL group, respectively. In the EVG group, most of the discontinuations were because of nonadherence (n = 10), followed by withdrawal of consent (n = 9). In the RAL group, most of the discontinuations were because of lost to follow-up (n = 6), followed by lack of efficacy, nonadherence, and withdrew consent (n = 5 for each category).
Efficacy
Using the prespecified TLOVR algorithm, the proportions of mITT subjects in the EVG and RAL groups, which achieved and maintained an HIV RNA < 50 copies/mL, was 59.0% (207/351) and 57.8% (203/351) through week 48 and 47.6% and 45.0% through week 96, respectively, with a treatment difference at week 96 of 2.6% [95% confidence interval: 4.6% to 9.9%]. Lower efficacy was seen in subjects without resistance at baseline of 33.3% (21/63) for EVG and 40.4% (23/57) for RAL through week 96 (Fig. 1).
(Enlarge Image)
Figure 1.
Efficacy endpoint: HIV-1 RNA,50 c/mL (ITT, TLOVR).
The increase in the number of patients classified as treatment failures in either group between week 48 and week 96 was primarily driven by combined reasons other than virologic failure. Between week 48 and week 96, virologic failure occurred in 15 (4.3%) and 22 (6.3%), discontinued study drug due to an AE or death in 3 (0.9%) and 5 (1.4%), and discontinuation of study drug for other reasons in 22 (6.3%) and 18 (5.1%) for subjects in the EVG and RAL group, respectively.
Among per protocol subjects, efficacy at week 96 (TLOVR) was higher than in the mITT population, with 69.1% (163/236) in the EVG group, compared with 67.8% (154/227) on RAL.
Analyses of efficacy at week 96 imputing missing as failure (M = F) and missing as excluded (M = E) revealed higher and comparable rates of virologic efficacy of 53.6% and 56.4% and 79.0% and 83.2% for each of the analysis, respectively.
Resistance
Eighty-seven EVG group subjects and 93 RAL group subjects were evaluated for genotypic drug resistance. Through week 96, emergent integrase-resistance mutations were detected in 6.6% (23/351) of all EVG (none after week 48) and 7.4% (26/351) for all RAL (an additional 7% after week 48). Most subjects who met criteria for virology resistance analyses did not develop new resistance mutations (had the same resistance mutations at baseline).
Through week 96, a similar percentage of subjects in each group developed new mutations in the reverse transcriptase: NRTI, 2.6% (9/351) and 3.4% (12/351); non-NRTI, 3.4% (12/351) and 2.3% (8/351); and protease, 1.4% (5/351) and 1.4% (5/351), for EVG and RAL, respectively. Most treatment emergent integrase strand-transfer inhibitor mutations occurred in isolation and were not accompanied by new primary mutations in the NRTI, non-NRTI, or PI genes.
Safety
Through week 96, adherence to ≥90% of study drug was 78.3% for EVG and 65.3% for RAL.
For EVG and the RAL groups, the incidences of AEs attributed to study drug through week 96 were similar at 23.7% and 20.4%; rates of grades 3–4 AEs attributed to study drugs were low and similar at 2.3% and 3.1%; study drug-related serious AEs occurred in 1.1% and 2% subjects, respectively. Changes in estimated glomerular filtration rate from baseline to week 96 were similar between groups with a median change of -10.8 mL/min for EVG and -11.7 mL/min for RAL.
AEs of grades 2–4 that occurred in at least 5% subjects are presented in Table 1. Although more subjects (P = 0.02) on EVG reported diarrhea through week 96, most of these events occurred in the first 48 weeks of the trial, with only 3 subjects (0.8%) on EVG and 2 subjects (0.6%) on RAL reporting new onset or worsening diarrhea between week 48 and week 96. No subject discontinued study drug because of diarrhea, and most were self-limited, with low and comparable use of antidiarrheal agents through week 96 in 17.8% and 13.7% on each group respectively.
Through week 96, 11 EVG subjects (3.1%) and 15 RAL subjects (4.2%) discontinued study drugs because of AE. Reasons for discontinuation in more than 1 subject were nausea and vomiting in 3 EVG subjects and hepatitis in 2 RAL subjects.
Three subjects randomized to EVG died: 2 deaths before week 48 and 1 death after (end-stage liver disease caused by hepatitis C, intestinal perforation, and cardiac failure). In the RAL group, 9 subjects died: 8 deaths before week 48 and 1 death after (hemolytic anemia, cardiac arrest, possible coronary event, choking, heroin overdose, cardiomegaly with mitral valve prolapse, automobile accident, non-Hodgkin lymphoma, and Hodgkin lymphoma with sepsis).
Laboratory test abnormalities observed through week 96 were also similar between the EVG and RAL treatment groups, except for alanine transaminase (ALT) and aspartate transaminase (AST) elevations. More subjects randomized to RAL (19/352 or 5.4%) had grades 3–4 elevations in ALT than subjects randomized to EVG (6/349 or 1.7%) (P = 0.013). Eight RAL subjects reported grade 4 ALT elevations compared with none in the EVG group (P = 0.008). Except for 1 subject, all the ALT abnormalities occurred in the first 48 weeks. Similarly, more subjects randomized to RAL (21/352 or 6.0%) had grades 3–4 elevations in AST than subjects randomized to EVG (8/349 or 2.3%) (P = 0.021). Seven of those RAL subjects experienced a grade 4 AST elevations compared with none in the EVG group (P = 0.015). Among subjects with grades 3–4 treatment emergent ALT and/or AST elevations, coinfection with hepatitis B or C virus was reported at baseline in 41.6% EVG- and 45.8% RAL-treated subjects. A low number of hepatitis B or C virus acute infections occurred on study (1.7% for RAL and 0% for EVG).