Methods: HIV-1 was subtyped in env by V3 sequencing or heteroduplex mobility assay. Coreceptor use was determined by the ability of the isolates to replicate in U87 CD4 cells expressing different coreceptors. The Fisher exact test was used to examine the relation between coreceptor use and subtype, clinical stage, and V3 charge. The Kruskall-Wallis nonparametric test was used to examine the association between median CD4 cell counts, coreceptor use, and subtype. Logistic regression was used to examine predicted coreceptor use at different CD4 groupings.

Results: Isolates from 66 participants were analyzed. Thirty-one were infected with subtype A, and 35 were infected with subtype D. Although this work was based on a small sample size, we found statistically significant differences. The probability of having an X4 virus was higher in subtype D infections than in subtype A infections among those with a non-AIDS clinical status (Fisher exact test, P = 0.040). Logistic regression analysis, in which we predicted X4 use by subtype and stratified by CD4 group, confirmed these findings among those with a CD4 count > 200 cells/mu L (likelihood ratio test, P = 0.003). R5 viruses were associated with higher median CD4 cell counts than X4 or X4/R5 (Kruskall-Wallis test, P = 0.0045). A V3 charge of +5 and greater was highly associated with X4 virus (Fisher exact test, P = 0.006).

Conclusions: These subtype differences in coreceptor use may partially explain the faster progression rates we have previously reported in individuals infected with subtype D compared with subtype A. Our observations may have implications for the future use of coreceptor inhibitors in this population.