Immunophenotyping. Four‐color flow cytometric measurements were made, using fresh EDTA‐anticoagulated whole blood within 6 h of draw. Cells were treated with FACS lysing solution (Becton Dickinson Immunocytometry Systems), were washed with PBS, were fixed in PBS–1% formaldehyde, and were acquired immediately on a FACSCalibur cytometer. CaliBRITE3, CaliBRITE APC, and QuantiBRITE (Becton Dickinson Immunocytometry Systems), which are bead populations, were used to validate instrument performance and to standardize fluorescence intensity. Quantitative immunofluorescence values in units of antibody‐binding capacity (ABC) were generated by relating phycoerythrin median fluorescence intensity values, which were obtained from cells of interest, to the quantitative standard QuantiBrite. To minimize assay variability, identical reagent lots were used for sample preparation and instrument setup. Lymphocyte subsets were defined as naive (CD45RA+CD62L+) and activated (CD38+HLA‐DR+).
TREC analysis. TREC analysis was done by using genomic DNA from peripheral blood mononuclear cells. The T cell receptor Δ signal joint (sj) TREC standard construct for quantitative competitive polymerase chain reaction was generously provided by Richard Koup and Daniel Douek (University of Texas at Southwestern School of Medicine). The number of TREC molecules per 1 μg of DNA in each sample was calculated by comparing the sample “volume” or band intensity with the volume of the known standard in a reaction where the standard exists in about a 1:1 molar ratio with the sample TREC DNA. The lower limit of the TREC assay was 100 molecules.
Statistics. Descriptive statistics and analytical methods, both parametric and nonparametric, were used to analyze the data. Continuous variables were described in terms of medians and ranges. Categorical variables were expressed as the number and percentage in each group. Group differences were analyzed by using exact Wilcoxon rank sum tests and Fisher’s exact tests for continuous and categorical data, respectively. Regression analyses were done to assess the relationship between plasma HIV RNA level (converted to a log10 scale) and flow cytometry outcome variables, while adjusting for CD4+ lymphocyte counts. Finally, we used a general linear mixed model to assess the longitudinal association between CD4+ lymphocyte count and group status.
Study patients. In total, 28 patients were enrolled in the study: 10 in the stable suppressed group and 18 in the viral relapse group. Although the viral relapse group had fewer men, 11 (61%) of 18, compared with 9 (90%) of 10, the groups were statistically similar with regard to sex, age, and race. The median age was 43.1 years (range, 22.3–56.7 years); 15 (54%) of 28 patients were black.
Antiretroviral treatment and virologic histories.The groups had similar antiretroviral treatment histories. Overall, 18 (64%) of 28 were receiving a protease inhibitor–containing regimen, and 21 (75%) of 28 were receiving >3 antiretroviral medications. For the patients in the stable suppressed group, the median duration of virus suppression was 96 weeks (range, 10–155 weeks), and all patients had plasma HIV RNA levels below the limit of detection (<400 copies/mL) at study entry. For the patients in the viral relapse group, the median duration of relapse was 77 weeks (range, 20–160 weeks), and the median plasma HIV RNA level was 3692 copies/mL (range, 614–292,783 copies/mL) at study entry. CD4+ lymphocytes.The results of longitudinal analyses derived from all available CD4+ lymphocyte counts performed before study entry show that both groups had stable or increasing CD4+ cell counts. For those in the stable suppressed group, CD4+ lymphocytes increased over the duration of viral remission (slope, +1.39 cells/week). For those in the viral relapse group, CD4+ lymphocytes remained stable (slope, +0.19 cells/week). The rate of change in CD4+ cells/week was significantly lower in the viral relapse group, compared with that in the stable suppressed group. At study entry, the viral relapse group and stable suppressed group had similar CD4+ lymphocyte counts (407 vs. 562 cells/mm3; ). However, the viral relapse group had significantly fewer CD4+ lymphocytes (20% vs. 32%; ).