Xpert HIV-1 Viral Load Assay and VERSANT HIV-1 RNA 1.5 Assay: A Performance Comparison

To the Editors:

The study by Gueudin et al¹ provides data that show a very good performance, both on B and non-B subtypes, of the Xpert HIV-1 Viral Load assay (Xpert, Cepheid, Sunnyvale, CA) in HIV-1 RNA detection and quantification in comparison with the Abbott RealTime HIV-1 assay. Evaluating the performance of different HIV-1 RNA quantification assays is a crucial task in laboratories involved in the management of anti-HIV-1 patients; indeed, an accurate HIV-1 RNA quantification is recommended by national and international guidelines to monitor the effectiveness of antiretroviral therapy and also represents an important tool in various phases of the diagnosis of HIV-1 infection. Although primers used in commercial tests have been designed to target the most conserved regions located in the pol gene of the viral genome, such as the integrase region, discrepancies in HIV-1 RNA quantification keep being detected. In fact, considering the extremely high genetic HIV-1 variabiality [ie, 4 distinct groups: M, N, O e P; 9 pure subtypes within group M: A–D, F–H, J, and K; 6 sub-subtypes (A1–A4 and F1–F2), and at least 79 circulating recombinant forms (CRFs)] it cannot be expected that primers recognize all HIV-1 field variants with the same efficiency. Therefore, any new available realtime assay must be tested to evaluate its capability to identify and correctly quantify all HIV-1 subtypes. The Xpert is a real-time assay, designed for rapid quantification of HIV-1 RNA, in which the entire process, from the extraction to the final result, is given within 90 minutes by using a single cartridge containing all reagents and internal controls. One milliliter of plasma sample in the cartridge is enough to obtain HIV-1 RNA quantification. Its recent availability has prompted studies to evaluate its performance. In addition to the study by Guedin et al, other studies were recently published. They provide data obtained by comparison of Xpert with Aptima HIV-1 Quant Dx assay, Abbott RealTime HIV-1 assay, NucliSens EasyQ HIV-1 v. 2.0, and Roche Taqman v. 2.0 assay.^2–4 On the whole, these studies highlight a very good performance of the Xpert in HIV-1 RNA quantification, both of B and non-B subtypes, in comparison with known and routinely used real-time assays.

With the present letter we give our contribution, reporting on the results obtained comparing the analytical performance of the Xpert with the VERSANT HIV-1 RNA 1.5 Assay (Siemens HealthCare Diagnostics, Tarrytown, NY) (kPCR).

Forty-five HIV-1 positive samples, quantified by kPCR, routinely used in the Hygiene Laboratory, IRCCS AOU San Martino-IST, Genoa, Italy, were selected for evaluating analytical sensitivity of Xpert. To assess its specificity, 5 negative samples were also tested.

The claimed range of detection by Xpert from 40 to 10,000,000 copies per milliliter (1.6–7 Log) was considered for sample selection. On the whole, the selected samples ranged from 40 copies per milliliter (1.6 Log) to 821,500 copies per milliliter (5.91 Log). A variety of B and non-B subtypes, as well as CRFs were included in the selected series (25 subtypes B, 1 A, 3 C, 3 F, 1F1, 3 G, 3 CRF02_AG, 2 CRF01_AE, 4 CRF12_BF).

HIV-1 subtype determination was obtained by using the Trugene HIV-1 Genotyping Kit (Siemens HealthCare Diagnostics, Tarrytown, NY), according to the manufacturer's recommendations after automated extraction through the NucliSens EasyMAG instrument (BioMérieux, Boxtel, NL). Phylogenetic analysis was performed as previously described.⁵ The Xpert was used on the GeneXpert Instrument Systems according to manufacturer's instructions.

Statistical analyses were performed to compare the performance of the 2 assays by linear regression and graphically represented by the Bland–Altman plot.

It was not possible to test one subtype B sample for a damaged cartridge. Both assays correctly did not detect HIV-1 RNA in the 5 samples from HIV-1 negative controls. Three of 44 (6.8%) results were correctly detected as positive by both systems but were not used in the comparison. Pointedly, 3 samples quantified as 40 (subtype B), 41 (subtype B), and 65 copies per milliliter (CRF01_AE) by kPCR were positively detected and quantified as <40 copies per milliliter by Xpert. Qualitative HIV-1 RNA results were consistent between the 2 methods in all the 41 comparable samples. Indeed, quantitative HIV-1 RNA results obtained by the 2 methods were significantly correlated when analyzed by linear regression (R² = 0.93; slope of the regression line of 0.9745; intercept of +0.3298) and also by Bland–Altman plot since the mean difference was −0.13 (CI: 0.24 to −0.44). All the Xpert quantitative results gave a slightly higher viral load than kPCR except for 8 samples (5 subtype B, 1 subtype F and 2 CRF12_BF–mean difference = 0.08). However, differences between individual viral load values exceeding 0.5 Log were found in only 5/41 (12.2%) samples (1 subtype B, 1 subtype C, 2 CRF12_BF and 1 CRF02_AG–mean differences = −0.94); in all those cases, Xpert gave the higher viral load. As differences in HIV-1 RNA quantification have been equally observed in B and non-B subtypes, they do not seem to be related to any specific HIV-1 variants. Therefore, the 2 tests seem to use primers with similar efficiency in binding different HIV-1 variants (Table 1).

TABLE 1.: Results of K-PCR Siemens and Xpert HIV-1 Viral Load Cepheid

The samples used in this study are representative of almost all HIV-1 clades circulating in our region. Overall, despite the limitations of this study represented by the small sample size and by an intrinsic bias in sample selection (ie, based on the viral load levels obtained with kPCR), a highly significant correlation among viral load results was achieved. Differences exceeding 0.5 Log were found only in 5/41 (12.2%) samples. On the whole, both methods recognize all HIV-1 field variants with the same efficiency; when differences were observed, it seems safe to assume that the highest value was the more accurate. The intrinsic characteristics of Xpert, designed to obtain the results in a run time of 90 minutes with all PCR processes performed in a single cartridge, without the need for any particular requirements for personal skills or laboratory setups, added to its capability to correctly detect and quantify HIV-1 subtypes and CRFs, make the assay very useful not only for routine but especially for urgent diagnosis, and in developing countries, where it could facilitate and improve patients follow-up.