Ultrasensitive monitoring of HIV-1 viral load by a low-cost real-time reverse transcription-PCR assay with internal control for the 5' long terminal repeat domain

Abstract

Background: Current HIV-1 viral-load assays are too expensive for resource-limited settings. In some countries, monitoring of antiretroviral therapy is now more expensive than treatment itself. In addition, some commercial assays have shown shortcomings in quantifying rare genotypes.

Methods: We evaluated real-time reverse transcription-PCR with internal control targeting the conserved long terminal repeat (LTR) domain of HIV-1 on reference panels and patient samples from Brazil (n = 1186), South Africa (n = 130), India (n = 44), and Germany (n = 127).

Results: The detection limit was 31.9 IU of HIV-1 RNA/mL of plasma (> 95% probability of detection, Probit analysis). The internal control showed inhibition in 3.7% of samples (95% confidence interval, 2.32%-5.9%; n = 454; 40 different runs). Comparative qualitative testing yielded the following: Roche Amplicor vs LTR assay (n = 431 samples), 51.7% vs 65% positives; Amplicor Ultrasensitive vs LTR (n = 133), 81.2% vs 82.7%; BioMerieux NucliSens HIV-1 QT (n = 453), 60.5% vs 65.1%; Bayer Versant 3.0 (n = 433), 57.7% vs 55.4%; total (n = 1450), 59.0% vs 63.8% positives. Intra-/interassay variability at medium and near-negative concentrations was 18%-51%. The quantification range was 50-10,000,000 IU/mL. Viral loads for subtypes A-D, F-J, AE, and AG yielded mean differences of 0.31 log(10) compared with Amplicor in the 10(3)-10(4) IU/mL range. HIV-1 N and O were not detected by Amplicor, but yielded up to 180 180.00 IU/mL in the LTR assay. Viral loads in stored samples from all countries, compared with Amplicor, NucliSens, or Versant, yielded regression line slopes (SD) of 0.9 (0.13) (P < 0.001 for all).

Conclusions: This method offers all features of commercial assays and covers all relevant genotypes. It could allow general monitoring of antiretroviral therapy in resource-limited settings.

Figure 1.

Probit regression analysis to determine the limit of detection of the LTR assay. Probability (y axis) is plotted against RNA concentration in 16 parallel test samples per data point (x axis). The plot depicts the observed proportion of positive results in parallel experiments (□), as well as the derived predicted proportion of positive results at a given input concentration of RNA. The solid line is the prediction, and the dashed lines are the 95% confidence limits for the prediction. Note that the tested range of concentrations extends farther than shown here (see the text).

Figure 2.

Testing of 2 different subtype reference panels. (A), Panel 1, German National Reference Centre for Retroviruses; Panel 2, NIBSC panel. (B), plasma samples from patients with non-B HIV-1. ▪, viral load as determined by real-time RT-PCR LTR assay; □, viral load as determined by Roche Cobas Amplicor (Ver. 1.5). The log₁₀ differences in viral loads are plotted at the bottom of both panels. Letters displayed along the x axis of each panel are identifiers of HIV strains present in the tested patients. ∗, not detected.

Figure 3.

Correlation of viral loads as determined by real-time RT-PCR LTR assay (y axis) and commercial (comparison) assays (x axis) in samples from different countries. (A), samples from South Africa, assayed with the BioMerieux NucliSens NASBA assay. (B), samples from India, assayed with the Roche Amplicor (Ver. 1.5) ultrasensitive protocol. (C), samples from Brazil assayed with the BioMerieux NucliSens NASBA assay. (D), samples from Brazil, assayed with the Bayer Versant (Ver. 3.0) bDNA assay. (E), samples from Brazil, assayed with the Roche Amplicor (Ver. 1.5) standard protocol. (F), samples from Germany, assayed with the Roche Amplicor (Ver. 1.5) ultrasensitive protocol. Each panel shows the number of samples included in the analysis; these samples were selected to be within the linear ranges of both the LTR assay and the respective comparison assays (refer to Materials and Methods for details).

Figure 4.

Courses of viral loads in 5 patients with first-time diagnoses of HIV-1 infection, covering the time point at which ART was initiated (arrows). PCP, Pneumocystis carinii pneumonia.