INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Little information is available about the accuracy of different nucleic acid or signal amplification techniques in quantifying human immunodeficiency virus type 1 (HIV-1) RNA in the female genital tract (1, 3, 6, 11). The ability to assess the viral burden in genital secretions is essential in understanding the pathophysiology of sexual and mother-to-child transmissions of HIV infection (8) and in predicting the effect of antiretroviral therapy at the level of the genital compartment (6, 12). This will require the availability of standardized assays for the quantification of HIV-1 RNA which should be altogether sensitive, specific, and reproducible. Recently, two commercial hypersensitive assays conceived to detect and quantify low levels of HIV-1 RNA in plasma have become available. These are the AMPLICOR HIV-1 Monitor Test 1.5 (Roche Diagnostics, Branchburg, N.J.), which uses reverse transcription (RT)-PCR and whose threshold of positivity is 50 copies of HIV-1 RNA/ml (13, 14), and the Quantiplex HIV RNA 3.0 assay (Chiron-Bayer, Emeryville, Calif.), which uses the branched-DNA signal amplification technique and whose lower limit of quantification is 50 copies of HIV-1 RNA/ml (4, 5). The aim of the present study was to compare the ability of both assays to quantify HIV-1 RNA in normal female genital secretions spiked with HIV-1 RNA standards and in clinical cervicovaginal lavage samples from HIV-infected women.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Clinical specimens. Cervicovaginal secretions were obtained from 30 HIV-1-seropositive women and from 30 HIV-seronegative healthy women not at risk for HIV. Women with vaginal discharge, genital bleeding, or suffering from a sexually transmitted infection were excluded from the study. Participants were asked to avoid sexual intercourse and intravaginal medications for 3 days before enrollment. Samples were processed within 1 h of collection. Genital secretions were collected outside the menstrual period. Following the introduction of a speculum, a standardized 60-s lavage was performed with 3 ml of 1 M phosphate-buffered saline (pH 7.2) as previously described (2). This collection procedure introduces a dilution of the genital secretions of about 1:10 (2). After centrifugation of the cervicovaginal lavage at 1,000 × g for 10 min, the cell supernatant was stored at 80°C. Lavage samples were confirmed to be devoid of a significant amount of contaminating blood by measuring traces of hemoglobin using second derivative spectrophotometry, as described earlier (10, 12).

HIV-1 RNA panels. Three HIV-1 RNA panels of 1-ml aliquots were prepared with viral stock sources diluted in pooled cell-free cervicovaginal lavages from the HIV-negative control women. The PELICHECK HIV-RNA-97 genotype B kit standards (prepared by the Central Laboratory of the Blood Transfusion Service, Amsterdam, The Netherlands, for the Viral Diagnostic Quality Control [VQC] Programme) was used as viral stock for the first or VQC panel, with the following predicted HIV-1 RNA quantities (copies/ml): 8,333, 2,500, 833, 250, 83, 25, 8, 2, 1, 0.2, 0.1, and 0.03. The second panel was obtained by spiking pooled HIV-negative cervicovaginal lavages with supernatant from a culture of HIV-1 subtype A strain (a gift from Françoise Barré-Sinoussi, Institut Pasteur, Paris, France), quantified by NASBA-QT assay (Organon Teknika Corporation, Durham, N.C.), with the following HIV-1 RNA predicted values (copies/ml): 60,000, 30,000, 12,000, 2,400, 480, 96, 19, and 3. The third panel corresponded to pooled plasma collected from an antiretroviral drug-naive patient infected with HIV-1 subtype B, quantified by NASBA-QT assay, with the following HIV-1 RNA predicted values (copies/ml): 200,000, 100,000, 40,000, 8,000, 1,600, 320, 64, and 12.

Quantitative assays. Evaluation of the HIV-1 RNA load in cervicovaginal lavage samples was assessed by initial concentration of HIV-1 virions, followed by quantitation of HIV-1 RNA in the resulting pellet, using different initial volumes of vaginal lavage sample for each assay. HIV-1 RNA load in cervicovaginal secretions was finally expressed as RNA copies per milliliter of cervicovaginal lavage sample. Quantitative assays for HIV-1 RNA were used according to the manufacturer's recommendations, with slight modifications. For the AMPLICOR HIV-1 Monitor Test 1.5 assay, the virus concentration was determined by centrifuging 0.5 ml of cervicovaginal lavage samples at 23,600 × g for 60 min at 4°C. The pelleted virus particles were lysed by treatment with 600 µl of working HIV-1 MONITOR lysis buffer containing the quantitation standard, and the released RNA was precipitated with 600 µl of 100% isopropanol. The precipitated RNA was recovered by centrifugation, washed with 1 ml of 70% ethanol, and resuspended in 100 µl of AMPLICOR HIV-1 Monitor specimen diluent. Then, 50 µl of the processed specimen was added to 50 µl of the working AMPLICOR HIV-1 Monitor master mix for the RT-PCR amplification reactions. Further amplification, hybridization, and detection steps were performed according to the manufacturer's instructions. For the Quantiplex HIV RNA 3.0 assay, the virus was concentrated from 1 ml of cervicovaginal lavage samples by centrifugation at 23,500 × g for 60 min at 4°C, and the supernatant was further discarded without disturbing the virus pellet. Pellets were then frozen at 80°C for more than 1 day until they could be tested, as recommended. Further hybridization and detection steps were performed according to the manufacturer's instructions.

Statistical analysis. Linear regression analysis on log-transformed values was performed to compare observed to expected values obtained with spiked samples and observed values obtained by both assays in clinical samples, by using the software StatView SE⁺ (Abacus Concepts, Inc., Berkeley, Calif.) and to compare both quantitative assays by using the orthogonal regression Deming method with the EVAL kit software, as described elsewhere (7).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Detection of HIV-1 RNA in spiked samples. Table 1 and Fig. 1 depict the raw data of HIV-1 RNA estimations tested in duplicate in the three panels of samples of pooled cervicovaginal lavages from HIV-negative women spiked with serial dilutions of HIV-1 subtype A or HIV-1 subtype B standards. Although both assays were able to detect HIV-1 subtype A as well as HIV-1 subtype B standards, differences between the AMPLICOR HIV-1 Monitor Test 1.5 and the Quantiplex HIV-1 3.0 assays were observed.

View larger version (11K): [in this window] [in a new window]   FIG. 1.   Scatter plots of observed versus expected HIV-1 RNA levels, expressed in log copies/milliliter, estimated in parallel by the AMPLICOR HIV-1 Monitor Test 1.5 assay (left) and the Quantiplex HIV RNA 3.0 assay (right) in pooled cervicovaginal lavages obtained from HIV-negative not-at-risk women spiked with known amount of HIV-1 subtype A or subtype B (n = 28). The vertical and horizontal dashed lines indicate the positivity threshold of both assays (1.69 log copies/ml).

Assay reproducibility. To determine intra-assay reproducibility, two pooled cervicovaginal lavages from HIV-seronegative control women spiked with different levels of HIV-1 subtype B RNA representing high and low copy numbers (cervicovaginal secretions samples 1 and 2, respectively) were analyzed simultaneously 10 times by both assays. The results presented in the Table 2 show that the lower-copy-number HIV-1 RNA cervicovaginal secretion sample 2 yielded a slightly larger expected coefficient of variation (CV), compared to the high-copy-number cervicovaginal secretion sample 1 by both assays. However, these values corresponded to a rather narrow within-run standard deviation of <0.33 log, which is <0.5 log. The interassay reproducibilities were estimated for the two spiked cervicovaginal secretions samples 1 and 2, by 10 successive determinations of HIV-1 RNA load by both assays. Although the inter-assay reproducibilities of the Quantiplex HIV-1 3.0 assay were slightly higher than those of the AMPLICOR HIV-1 Monitor Test 1.5 assay, the mean CV values for between-run precision showed a rather low variability and were found to be in the range of 0.09 to 0.16%, which still corresponds to a narrow between-run standard deviation (<0.40 log).

Quantitative analysis of clinical samples. HIV-1 RNA levels were determined in parallel by both ultrasensitive assays in cervicovaginal lavage clinical samples from 30 HIV-1-seropositive women and 30 HIV-seronegative, not-at-risk women (Fig. 2 and Table 3).

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Correlation of HIV-1 RNA levels, expressed in log copies/milliliter, estimated in parallel by the AMPLICOR HIV-1 Monitor Test 1.5 assay and the Quantiplex HIV RNA 3.0 assay in cervicovaginal lavages obtained from 30 HIV-positive women and 30 HIV-negative not-at-risk controls. , Cervicovaginal HIV-1 RNA load from HIV-infected women; , cervicovaginal HIV-1 RNA load from HIV-negative women. The vertical and horizontal dashed lines indicate the positivity thresholds of both assays (1.69 log copies/ml).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We compared two third-generation hypersensitive assays, the AMPLICOR HIV-1 Monitor Test 1.5 and Quantiplex HIV RNA 3.0, to quantitate HIV-1 RNA levels in cervicovaginal secretions spiked with known amounts of HIV-1 RNA standards and in cervicovaginal lavages clinical samples from HIV-infected and HIV-negative women. Both assays were able to detect similarly HIV-1 subtype A and HIV-1 subtype B RNA mixed in pooled cervicovaginal secretions. The linear ranges of both assays corresponded to those claimed by the manufacturers, except for the lower value of the linear range of the Quantiplex HIV RNA 3.0 assay. Compared to the AMPLICOR HIV-1 Monitor Test 1.5 assay, the Quantiplex HIV-1 3.0 assay yielded higher estimates of HIV-1 RNA concentrations in several tested samples spiked with HIV-1 RNA subtype A or subtype B, particularly those containing low amounts of HIV-1 RNA. The intra-assay and interassay reproducibilities appeared to be nearly similar for both assays. The sensitivities of the AMPLICOR HIV-1 Monitor Test 1.5 assay and of the Quantiplex HIV RNA 3.0 assay were 93 and 97%, respectively, indicating that some samples containing a low amount of HIV-1 RNA could not be detected by the AMPLICOR HIV-1 Monitor Test 1.5 assay. The specificity of the AMPLICOR HIV-1 Monitor Test 1.5 assay and that of the Quantiplex HIV RNA 3.0 assay were 100 and 50%, respectively, indicating clearly that the Quantiplex HIV RNA 3.0 assay gives false-positive results for HIV-negative samples and also overestimates the HIV-1 RNA levels in samples containing copy numbers of HIV-1 RNA under the threshold of detection of the assay. The low specificity of the Quantiplex 3.0 assay used to quantify HIV-1 RNA in cervicovaginal secretions samples suggests that nonspecific hybrization may occur, probably because of the interference with the body fluid matrix. Indeed, we observed that the nonspecific signal was lower when female genital secretions were collected using 10 ml of lavage buffer instead of 3 ml as in the present study (data not shown).

Taken together, our observations demonstrate that the Quantiplex HIV RNA 3.0 assay may lack the specificity to quantify HIV-1 RNA in the acellular fraction of cervicovaginal secretions, the risk of false positivity being particularly marked for the low HIV-1 RNA levels. The AMPLICOR HIV-1 Monitor Test 1.5 assay is highly specific for detecting and quantifying HIV-1 RNA in cervicovaginal secretions, although it may lack sensitivity for some samples with low HIV-1 RNA levels (100 copies/ml).