Variability of CD4+ Cell Counts in HIV-1-Uninfected Volunteers Who Are Eligible for a Phase I HIV Vaccine Study.

Objective: Vaccines and biologics containing CD4 molecules or HIV-1 gp120 might induce antibodies targeting CD4. We evaluated temporal variability of CD4 levels in healthy volunteers to quantify declines that could indicate true adverse events.
Design: Prospective observational cohort study of 100 healthy adults without HIV-1 infection from the Baltimore region.
Methods: Participants enrolled and consented to blood draws for immunologic laboratory panels performed once every 8 weeks for 48 weeks. The primary CD4 measurements were CD4 absolute count (cells/mm) and CD4 percentage (CD4%, total CD4 cells/total lymphocyte cells). CD4 changes over time were modeled using fold changes for CD4 absolute counts and differences for CD4 percentages.
Results: Variation of average CD4 cell counts and percentages were highly participant-specific (P < 0.001 for both). However, changes in both CD4 measurements over time were stable in the population. We proposed thresholds to flag unusual drops using 1.5 SD estimates, calculated as 1.5-fold declines for CD4 count and 6.4% declines for CD4 percentage. In this healthy cohort, flagging simultaneous declines in both measurements corresponded to a low false-positive rate (5.26%).
Conclusions: Normal biological variation in large lymphocytes should be taken into account to establish thresholds for adverse changes in clinical trials. The inherent subject-specific variability in CD4 levels makes establishing absolute cutoffs difficult. However, this study proposes that thresholds for declines using 1.5 SDs from these data (50% in absolute count and 6.4% for CD4 percentage) allow a small false-positive rate (∼5%) that could maintain sensitivity for true adverse events in a clinical trial.