Backgound: Current ART is effective in suppressing but not eliminating HIV-1 infection. Previously, we used a real-time polymerase chain reaction (PCR) assay with single-copy sensitivity to investigate HIV-1 viremia in suppressed patients and detected a stable level of viremia in patients on therapy. The level of biological variation in HIV-1 viremia in individual patients over time remains uncertain. We applied a poisson statistical analytic approach to investigate fluctuations in viremia over time in individual patients.

Methods:  Patients (n = 25) with HIV-1 viremia suppressed <50 copies/mL on standard antiretroviral regimens were identified and plasma samples were subjected to single-copy assay (SCA) in triplicate. Probability mass function and Poisson homogeneity testing was used to investigate intra-assay variability of the replicate samples and determine the likelihood that SCA determinations were distributed according to Poisson statistics. In 13 patients with longitudinal sampling over 1 year, Poisson regression statistics and maximum-likelihood ratio testing were used to investigate whether fluctuations in viremia were present.

Results:  Analysis of SCA values from individual patient time points were distributed according to predicted Poisson probabilities in 20 of 25 patients, indicating assay variation was sufficiently low to permit detection of biological variation. Analysis of longitudinal samples revealed 5 of 13 patients had biological variation with statistically significant fluctuations in viremia over 1 year sampling. In contrast, 8 of 13 patients exhibited no detectable inter-assay variability, indicating minimal if any variation in viremia. Throughout the analysis period, 3 patients had <0.3 copies/mL. SCA of as much as 14 mL plasma (performed in replicates of 7 mL plasma) from individual time-points revealed that 2 of 3 patients had <1 detectable copy of HIV-1 RNA, indicating a remarkably low level of viremia; the remaining patient had 1 copy/7 mL plasma.

Conclusions:  SCA is remarkably robust with assay variation lower than random variation in the majority of patients studied. HIV-1 infection can be maintained with <1 copy/7 mL plasma, but biological variation in HIV-1 viremia was detectable in patients using SCA even at low levels of viremia, indicating that fluctuations in viremia still occur despite suppressive therapy.