TEXT

Top Abstract Text References

Human immunodeficiency virus type 1 (HIV-1) RNA level in plasma (viral load) is a marker of major importance for the follow-up and management of HIV-1-infected patients. Indeed, an increase in the HIV-1 RNA level in plasma predicts disease progression, whereas in patients treated with antiretroviral drugs, a decrease in the HIV-1 RNA level in plasma indicates treatment efficacy (1, 3, 6, 11). Therefore, the HIV-1 RNA level in plasma is regularly determined for patients undergoing antiretroviral therapy. The commercially available methods for quantitation of HIV-1 RNA in plasma are based on quantitative reverse transcription-PCR assay (7), nucleic acid sequence-based amplification (NASBA) (10), or branched DNA (4); the three methods demonstrated equal reliability (2, 8, 9). The current lower limits of quantitation by these assays are 200, 400, and 500 HIV-1 RNA copies/ml for reverse transcription-PCR, NASBA, and branched DNA, respectively. However, in patients with multidrug combination therapies, HIV-1 RNA levels in plasma frequently drop below these limits. Therefore, there is a need for the development of new assays with increased sensitivity for quantitation of low HIV-1 RNA levels in plasma. Nuclisens HIV-1 QT (Organon Teknika, Boxtel, The Netherlands) is an assay based on the NASBA methodology which presents a lower limit of quantitation than the first-generation NASBA HIV-1 QT assay (80 versus 400 HIV-1 RNA copies/ml). In this study, we evaluated the specificity, reproducibility, precision, accuracy, and linearity of this assay. We also compared Nuclisens HIV-1 QT and NASBA HIV-1 QT for the quantitation of HIV-1 RNA levels in clinical specimens and the quantitation of HIV-1 RNA in plasma infected with non-B clade strains of HIV-1.

This study was conducted in two laboratories, one located in Montpellier (laboratory 1) and the other located in Paris (laboratory 2). All of the plasma specimens were independently analyzed in both laboratories, except the specimens of non-B clade HIV-1 strains, which were only analyzed in laboratory 1. For evaluation of specificity and clinical sensitivity, blood was collected on EDTA and immediately centrifuged. One-milliter aliquots of plasma were immediately placed in the lysis buffer provided by the manufacturer, briefly vortexed, and immediately stored at 70°C until testing. For evaluation of reproducibility, precision, accuracy, and linearity, as well as for quantitation of RNA from non-B clades of HIV-1, HIV-1-infected cell culture supernatants were diluted in a pool of plasma samples obtained from HIV-1-seronegative blood donors. Immediately after preparation, these spiked plasma specimens were aliquoted, placed in lysis buffer, and stored as described above. HIV-1 RNA levels were quantitated with Nuclisens HIV-1 QT or NASBA HIV-1 QT in accordance with the manufacturer's instructions.

Specificity. Specificity was evaluated on 100 plasma specimens obtained from HIV-1-seronegative blood donors. All of the results were lower than the detection limit in laboratory 1. In laboratory 2, a level of 150 HIV-1 RNA copies/ml (2.17 log HIV-1 RNA copies/ml) was obtained for one specimen. This specimen was tested again, and HIV-1 RNA content was found to be lower than the detection limit. The specificity of Nuclisens HIV-1 QT was therefore 100% in laboratory 1 and 99% in laboratory 2 without repeatability of the initial false-positive result.

Reproducibility. Reproducibility was tested on specimens obtained by dilution of a HIV-1 clade B strain (HIV-1_GER) in a HIV-1-seronegative plasma pool. Three levels of HIV-1 RNA levels were tested. In the high HIV-1 RNA level specimen (level A), the HIV-1_GER culture supernatant was diluted to approximately 500,000 HIV-1 RNA copies/ml (5.70 log HIV-1 RNA copies/ml). A low HIV-1 RNA level specimen (level B), which was obtained by diluting the high HIV-1 RNA level specimen 1:200, contained approximately 2,500 HIV-1 RNA copies/ml (3.40 log HIV-1 RNA copies/ml), and a specimen with a very low HIV-1 RNA level (level C) was obtained by 1:10 dilution of the low HIV-1 RNA level specimen (approximately 250 HIV-1 RNA copies/ml; 2.40 log HIV-1 RNA copies/ml). In both laboratories, 10 aliquots of each specimen were tested in one run and the runs were repeated three times on different days. As shown in Table 1, standard deviations obtained with levels A and B were <0.15 log HIV-RNA copy/ml in the within-run, interrun and interlaboratory reproducibility tests; the coefficients of variation (CV) ranged from 11 to 31% when the results were expressed as HIV-1 RNA copies per milliliter and from 1.0 to 3.9% when the results were expressed as log HIV-1 RNA copies per milliliter. Higher standard deviations (0.21 to 0.31 log copies/ml) and CVs (HIV-1 RNA copies per milliliter, 43 to 103%; log HIV-1 RNA copies per milliliter, 8.2 to 15.4%) were obtained with the level C specimen, indicating a decrease in reproducibility when the HIV-1 RNA levels decreased to values near the lower limit of quantitation.

Linearity, accuracy, and precision. Linearity, accuracy, and precision were determined with a panel of spiked plasma samples which contained known quantities of HIV-1 RNA copies (7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.7, and 1.0 log HIV-1 RNA copies/ml). This panel was obtained by successive dilutions of a HIV-1 suspension in which the number of HIV-1 particles had been determined by electron microscopy (5). Each sample was tested three times by each laboratory. HIV-1 RNA was detected in none of the six samples containing 1.0 log HIV-1 RNA copy/ml, in two of the six samples containing 1.7 log HIV-1 RNA copies/ml (1.86 and 2.18 log HIV-1 RNA copies/ml), and in four of the six samples containing 2.0 log HIV-1 RNA copies/ml (1.38, 1.46, 2.45, and 2.67 log HIV-1 RNA copies/ml). On the other hand, invalid results due to competition between HIV-1 RNA and the Nuclisens HIV-1 QT calibrators were obtained for five of the six samples containing 7.0 log HIV-1 RNA copies/ml; the result was 6.93 log HIV-1 RNA copies/ml in the remaining sample. Results obtained for input HIV-1 RNA concentrations ranging from 2.0 to 6.0 log HIV-1 RNA copies/ml are presented in Table 2. A linear relationship between the input and measured HIV-1 RNA concentrations was observed throughout this range (r = 0.990; R² = 0.980). According to the input HIV-1 RNA concentration, accuracy varied from 0.11 to +0.10 log HIV-1 RNA copy/ml and precision varied from 0.66 to 0.14 log HIV-1 RNA copy/ml. As observed in the reproducibility tests, the precision decreased when the HIV-1 RNA level dropped near the lower limit of quantitation.

Comparison of Nuclisens HIV-1 QT with NASBA HIV-1 QT for HIV-1 RNA quantitation in clinical samples. In both laboratories, HIV-1 RNA levels were quantitated by NASBA HIV-1 QT and Nuclisens HIV-1 QT in 78 plasma samples obtained from HIV-1-infected patients. For 53 samples, HIV-1 RNA could be quantitated with NASBA HIV-1 QT with values ranging from 2.61 to 5.76 log HIV-1 RNA copies/ml. A linear relationship was observed between the values obtained with NASBA HIV-1 QT and Nuclisens HIV-1 QT (r = 0.964; R² = 0.930). In 25 samples, HIV-1 RNA was not detected with NASBA HIV-1 QT but could be quantitated with Nuclisens HIV-1 QT in 9 (36%) samples in laboratory 1 (range, 1.96 to 3.20 log HIV-1 RNA copies/ml) and 6 (24%) samples in laboratory 2 (range, 2.04 to 3.00 log HIV-1 RNA copies/ml). The HIV-1 RNA levels measured in both laboratories with Nuclisens HIV-1 correlated well (r = 0.980; R² = 0.959).

Quantitation of RNA from non-B clade strains of HIV-1. As shown in Table 3, HIV-1 strains belonging to clade G could be detected by neither NASBA HIV-1 QT nor Nuclisens HIV-1 QT, whereas RNA from strains belonging to the other clades of HIV-1 group M could be quantitated by both assays. Four strains belonging to HIV-1 group O were also tested; RNA could not be detected in the corresponding samples despite detection of HIV-1 antigen (Murex HIV Antigen Mab; Murex Diagnostics, Chatillon, France).

Conclusion. Our results indicate that Nuclisens HIV-1 QT is a reliable method for the quantitation of HIV-1 RNA in plasma. Results obtained with Nuclisens HIV-1 QT are closely related to those obtained with NASBA HIV-1 QT, but Nuclisens HIV-1 QT is more suitable than NASBA HIV-1 QT for quantitation of low levels of HIV-1 RNA.