ABSTRACT: Purpose: To study the association of baseline (BL) HIV protease and reverse transcriptase (RT) genotype with week 24 plasma RNA concentration in ACTG 320, which was a Phase III clinical trial of zidovudine plus lamivudine with or without indinavir in zidovudine-experienced, lamivudine- and PI-naïve patients with CD4 counts of ¾200 cells/mm3. There was a 50% reduction in the development of AIDS or death in the zidovudine/lamivudine/indinavir arm in ACTG 320, and 50% of patients in the indinavir arm had plasma HIV RNA <500 copies/ml at 24 weeks.
Methods: A total of 392 patients were selected from the indinavir arm for sequence analysis of HIV protease (codons 10-99) and RT (codons 1-480), using the Perkin Elmer/Applied Biosystems HIV Genotyping kit. Group 1 was a case-cohort sample of 113 patients that included a randomly selected cohort of 80 controls (including four who reached a clinical endpoint), and the remaining 33 cases that reached a clinical endpoint. Group 2 included all 279 patients who were not in Group 1, had complete RNA data to week 24, and remained on their initial study treatment through week 24.
Results: To date, baseline samples from 102 subjects have been sequenced and analyzed, including 80 from group 1 (26 cases) and 22 from group 2. Approximately 55% of patients had T215Y/F, 23% had K70R, and 39% had M41L RT mutations. None had M184V. Ten percent had G333E, 62% had L63P, 40% had A77T/V, and 18% had L101/R/V protease polymorphisms. Only one percent had V82A; none had changes at codons 48 or 90 of the protease. To date, we have found no associations of week 24 plasma HIV RNA concentration with either BL zidovudine resistance mutations, number of zidovudine resistance mutations, presence of zidovudine resistance mutations in combination with protease polymorphisms, or presence of G333E.
Conclusions: In this preliminary analysis, we found no correlation between BL HIV protease or RT resistance mutations and week 24 virological response. Further exploration of the relationship between genotype and RNA and clinical responses will make use of statistical methods that accommodate the large number of possible patterns of mutation.
