Prospective Study of Acute HIV-1 Infection in Adults in East Africa and Thailand

Abstract

Background: Acute human immunodeficiency virus type 1 (HIV-1) infection is a major contributor to transmission of HIV-1. An understanding of acute HIV-1 infection may be important in the development of treatment strategies to eradicate HIV-1 or achieve a functional cure.

Methods: We performed twice-weekly qualitative plasma HIV-1 RNA nucleic acid testing in 2276 volunteers who were at high risk for HIV-1 infection. For participants in whom acute HIV-1 infection was detected, clinical observations, quantitative measurements of plasma HIV-1 RNA levels (to assess viremia) and HIV antibodies, and results of immunophenotyping of lymphocytes were obtained twice weekly.

Results: Fifty of 112 volunteers with acute HIV-1 infection had two or more blood samples collected before HIV-1 antibodies were detected. The median peak viremia (6.7 log10 copies per milliliter) occurred 13 days after the first sample showed reactivity on nucleic acid testing. Reactivity on an enzyme immunoassay occurred at a median of 14 days. The nadir of viremia (4.3 log10 copies per milliliter) occurred at a median of 31 days and was nearly equivalent to the viral-load set point, the steady-state viremia that persists durably after resolution of acute viremia (median plasma HIV-1 RNA level, 4.4 log10 copies per milliliter). The peak viremia and downslope were correlated with the viral-load set point. Clinical manifestations of acute HIV-1 infection were most common just before and at the time of peak viremia. A median of one symptom of acute HIV-1 infection was recorded at a median of two study visits, and a median of one sign of acute HIV-1 infection was recorded at a median of three visits.

Conclusions: The viral-load set point occurred at a median of 31 days after the first detection of plasma viremia and correlated with peak viremia. Few symptoms and signs were observed during acute HIV-1 infection, and they were most common before peak viremia. (Funded by the Department of Defense and the National Institute of Allergy and Infectious Diseases.).

Figure 1. Enrollment and Outcomes

Single false reactive results on qualitative nucleic acid testing were common, but acute HIV infection was confirmed in all participants who had two consecutive plasma samples that were reactive for HIV-1 RNA on qualitative nucleic acid testing. Among 112 participants with acute infection, peak viremia could be accurately defined in 54 participants who had at least two samples in which testing for HIV-1 RNA was reactive and enzyme immunoassay was nonreactive. Four of these participants were excluded from virologic and immunologic analyses because they received early antiretroviral therapy (ART). Most infections were of the subtype HIV-1 CRF01_AE in Thailand and of the HIV-1 subtype A and recombinant form of the A, C, and D strains in East Africa. Fiebig stages range from I through VI, with higher stages indicating a more mature stage of antibody response to HIV. CI denotes confidence interval, EIA enzyme immunoassay, and NAT nucleic acid test.

Figure 2. Viral Loads over the First 100 Days of HIV-1 Infection in 50 Participants

Longitudinal viral-load values are plotted against the number of days since the first blood sample was reactive for HIV-1 RNA in 33 participants from East Africa and 17 from Thailand who had two or more blood samples that were nonreactive on enzyme immunoassay (EIA) and were reactive on nucleic acid testing. Day 0 is the day of the first positive nucleic acid test. The box-and-whisker plots show the median, interquartile range, and range for each variable. The vertical box plots show peak and nadir viral loads, and the horizontal box plots show the number of days from the first reactive result on nucleic acid testing to the peak viral load, to reactivity on the EIA, and to the nadir viral load. Median values are shown for each variable.

Figure 3. Viral-Load Associations

Spearman's correlations of peak (Panel A) and nadir (Panel B) viremia with the viral-load set point in participants from Thailand (black dots) and East Africa (red dots) are shown. Five participants were excluded because of missing viral-load data or the initiation of anti-retroviral therapy. The early viremic nadir and viral-load set point were highly correlated (Panel B) and did not differ significantly (Panel C). A significant positive Spearman's correlation of viral-load downslope with the viral-load set point was observed in 44 participants with acute HIV-1 infection (Panel D). Viral-load upslope did not show significant Spearman's correlation with the viral-load set point in the overall sample (Panel E). Seven of 45 participants were excluded from the upslope analysis because the interval between the last test that was negative for HIV-1 RNA and the first test that was reactive for HIV-1 RNA was longer than 10 days. Comparisons with the use of a Wilcoxon rank-sum test did not show regional differences between East Africa and Thailand with respect to peak viral load (Panel F); however, the viral-load set point was higher in participants in Thailand, with a narrower range of values (Panel G). Sex, race or ethnic group, and HIV-1 subtypes also differed between the two groups.

Figure 4. Immune-Cell Counts over 510 Days of Follow-up in 50 Participants

The absolute counts of CD4+ T cells (Panel A), CD8+ T cells (Panel B), natural killer cells (Panel C), and B cells (Panel D) are plotted against the mean visit day. The red dots indicate 33 participants from East Africa, and the black dots 17 participants from Thailand. The blue dotted line indicates the mean viral load. Wilcoxon signed-rank tests were performed on available paired data before the initiation of antiretroviral treatment. A significant decrease in absolute CD4+ counts and a significant increase in absolute CD8+ counts at day 17 roughly coincided with the timing of the peak viral load in 44 participants with data that could be evaluated. By day 510, absolute CD4+ and CD8+ counts had not returned to initial levels in 26 participants. Absolute natural-killer-cell counts decreased at day 17 in 42 participants; however, these counts returned to original levels by day 510 in 24 participants. There was a significant decrease in absolute B-cell counts (P<0.001) at day 17 in 42 participants; however, these counts recovered by day 510 in 24 participants.

Figure 5. Medical Symptoms and Signs before and after Diagnosis of HIV Infection in 50 Participants

The symptoms reported from the medical history (Panel A) and the number of abnormal physical findings (signs) on examination (Panel B) are shown at study visits before, after, and at the time of peak viremia. The dashed lines indicate individual participants.