Background: A new version of the Roche Cobas AmpliPrep/Cobas TaqMan HIV-1 assay (CA/CTM v2.0) has been introduced to overcome the underquantification observed with the first version.
Methods: We compared the Roche Cobas CA/CTM v2.0 and Abbott RealTime HIV-1 assays for HIV-1 group M and non-M viral load measurement.
Results: We found a good correlation (r = 0.96) between the 2 techniques for the 260 HIV-1 group M plasma samples tested. The Roche Cobas assay gave significantly higher values than the Abbott assay, and 51 samples (20%) yielded differences greater than 0.5 log10 copies per milliliter. Conversely, 2 samples were more than 0.5 log10 copies per milliliter higher with the Abbott assay than with the Roche Cobas assay. Among the 84 samples with undetectable viral load in the Abbott assay (detection limit 40 copies/mL), 17 (20%) were detectable with the CA/CTM v2.0 assay (detection limit 20 copies/mL), with values ranging from 41 to 897 copies per milliliter. Extrapolation of the Abbott curves led to 10/17 (59%) of these samples being quantifiable. HIV-1 groups O and P were similarly quantified by the two techniques.
Conclusion: The results of the Roche Cobas CA/CTM v2.0 and Abbott RealTime HIV-1 assays correlate well. The new version of the CA/CTM assay shows improved sensitivity. Nevertheless, the 2 assays differ by more than 0.5 log10 copies per milliliter for some samples.
From the *Laboratoire de Virologie, Hôpital Saint-Louis AP-HP, INSERM U941, Université Paris-Diderot, Paris, France; †Unité d'Epidémiologie des Maladies Emergentes, Institut Pasteur, Paris, France; ‡Laboratoire de Virologie, Centre Hospitalier Universitaire Charles Nicolle, Faculté de Médecine et de Pharmacie, Rouen, France; §Service des Maladies Infectieuses et Tropicales, Hôpital Saint-Louis, Université Paris-Diderot Paris, France; ‖Service des Maladies Sexuellement Transmissibles, Hôpital Saint-Louis, Paris, France; and ¶Service de Médecine Interne, Hôpital Saint-Louis, Université Paris-Diderot, Paris, France.
Received for publication August 6, 2010; accepted November 16, 2010.
Presented in part at the 17th Conference on Retroviruses and Opportunistic Infections, February 16-19, 2010, San Francisco, CA, USA (abstract #940).
The authors have no funding or conflicts of interest to disclose.
Correspondence to: Dr. Constance Delaugerre, PharmD, PhD, Laboratoire de virologie, Hôpital Saint-Louis, 1 avenue Claude Vellefaux, 75010 Paris, France (e-mail: firstname.lastname@example.org).
HIV-1 plasma RNA viral load (pVL) is the principal marker of antiretroviral treatment efficacy and a major end-point in clinical trials. Real-time polymerase chain reaction (PCR) techniques such as the Abbott Diagnostics RealTime HIV-1 (Abbott, Abbott Molecular, Rungis, France) and Cobas AmpliPrep/Cobas TaqMan HIV-1 assays (CA/CTM, Roche Diagnostics, Meylan, France) are gradually replacing signal and target amplification for HIV-1 RNA quantification. Indeed, they are faster, suitable for high-throughput purposes, and have a wider dynamic range. As these techniques are fully automated, from viral RNA extraction to quantification, the risk of contamination by amplification products is vastly reduced. Both assays target highly conserved regions of the HIV-1 genome, namely the integrase-encoding region (located in the pol gene) in the Abbott assay, and the gag region in the CA/CTM assay.
Real-time PCR primers and probes were originally based on the genome of HIV-1 subtype B, the most prevalent subtype in Western countries, but HIV-1 diversity is increasing worldwide. In France, nearly half of all HIV-1 infections diagnosed between January 2003 and March 2005 were due to subtype non-B viruses.1 The proportion of non-B viruses increased from 10.3% in 1995-1996 to 25.5% in 2005-2006,2 and new complex recombinants have been identified.3 An HIV-1 group M/O recombinant was recently reported in France,4 along with a virus representing a new HIV-1 group (group P) identified in a Cameroonian woman living in France.5
Viral load underquantification can have major clinical repercussions and delay the detection of drug resistance.6 Major differences were observed between the previous CA/CTM and Abbott tests,7,8 the former CA/CTM assay giving significantly lower values for the CRF02_AG subtype,7 the most prevalent non-B subtype in France. Moreover, the CA/CTM test could not be used to quantify HIV-1 group O viruses.7
To overcome this lack of sensitivity for HIV-1 group non-B quantification, the newly modified version of the CA/CTM test (CA/CTM v2.0) simultaneously targets the gag and 5′LTR regions in a multiplex PCR approach. The aim of this study was to compare the sensitivity of the CA/CTM v2.0 and Abbott assays for the detection and quantification of HIV-1 RNA in plasma samples containing different HIV-1 groups and subtypes.
PATIENTS AND METHODS
From June 1 to August 15, 2009, all routine viral load measurements for patients managed at Saint-Louis Hospital, Paris, were performed with both the Abbott and CA/CTM v2.0 assays. No additional blood samples were collected for this study. Stored plasma samples from HIV-1 group O-infected patients and culture supernatants of an HIV-1 M/O recombinant strain4 and the newly described HIV-1 group P strain5 were also tested with both techniques.
The CA/CTM v2.0 test was implemented with the COBAS AmpliPrep automated nucleic acid extractor and the TaqMan 96 analyzer in a docked instrument configuration. The Abbott test was implemented with the Abbott Diagnostics m2000sp/m2000rt automated extraction platform. Routine plasma samples were centrifuged at 3000g for 20 minutes, and at least 2 aliquots of 1 mL were stored at −80°C for batch testing. Both techniques were implemented according to the manufacturers' instructions, using plasma aliquots of 1 mL from the same sample, without prior dilution.
Results were expressed in copies per milliliter or log10 copies per milliliter. The lower limit of quantification is 20 copies per milliliter for the CA/CTM v2.0 assay (detection range, 2.0E + 1 to 1.0E + 7) and 40 copies per milliliter for the Abbott assay (detection range, 4.0E + 1 to 1.0E + 7).
Pol nucleotide sequences were determined with a 16-capillary sequencer (ABI PRISM Genetic Analyzer, Applied Biosystems, Les Ulis, France), and genotypes were determined from the protease and reverse transcriptase nucleotide sequences using the Los Alamos HIV sequence database (HIV BLAST, URL: http://www.hiv.lanl.gov/).
The results of the 2 techniques were expressed as the median and interquartile ranges (Q1 and Q3). Samples with viral loads below the working range of each test were assigned the respective lower quantification limit. None of the samples yielded values above the working range of either technique. The Wilcoxon matched-pairs test was used to compare the distribution of values obtained with each method. The Kruskal-Wallis test was used to compare the data for subtype B, CRF02_AG and others. Bland-Altman plot was used to represent the degree of agreement between the 2 assays. The x axis bore the mean values obtained with the 2 techniques for each sample, and the y axis bore the difference between the 2 values obtained with the 2 techniques.9 The correlation between the values obtained with the 2 assays was also assessed with the Spearman rank correlation coefficient. STATA 10 software was used for all statistical analyses.
Viral Load Values in the Overall Population
Consecutive plasma samples from 260 different HIV-1 group M-infected patients were collected and stored at −80°C in our laboratory as part of routine monitoring procedures. The median pVL values were 3.08 (1.51; 4.49) log10 copies per milliliter with the CA/CTM v2.0 assay and 2.81 (1.60; 4.15) log10 copies per milliliter with the Abbott assay (P < 0.001). As shown in Figure 1, the Spearman rank correlation coefficient was r = 0.96.
The median of the differences between the CA/CTM v2.0 and Abbott techniques was 0.13 log10 copies per milliliter [−0.26; 0.40]. The Bland-Altman plot (Fig. 2) confirmed the higher values of the CA/CTM v2.0 assay.
A difference of 0.5 log10 copies per milliliter or more during follow-up is considered clinically relevant.10,11 The CA/CTM v2.0 assay gave values at least 0.5 log10 copies per milliliter higher than the Abbott assay in 51 cases (20%) and a difference of at least 1 log10 copies per milliliter in 13 cases (5%) (Table 1). As shown by the Bland-Altman representation (Fig. 2), differences above 0.5 log10 copies per milliliter were observed regardless of the absolute pVL value (Table 1). The median pVL values in these 51 samples were 3.98 (2.93; 4.53) log10 copies per milliliter with the CA/CTM v2.0 assay and 3.07 (2.29; 3.61) log10 copies per milliliter with the Abbott assay.
HIV-1 genotyping in 36 cases revealed 17 cases of subtype B infection (47%), 8 cases of subtype CRF02_AG infection (22%), and 11 cases of infection by other subtypes (31%: A, C, D, G, CRF01_AE, CRF06_cpx and undetermined complex recombinant forms).
In 2 cases (0.8%), the Abbott test gave values more than 0.5 log10 copies per milliliter higher than the CA/CTM v2.0 test (Table 1).
Eighty-four samples (32%) were below the detection limit of 40 copies per milliliter in the Abbott technique. Fifty-five of these samples (65%) were below the detection limit of 20 copies per milliliter in the CA/CTM v2.0 test, whereas 12 samples (14%) gave values between 20 and 40 copies per milliliter. Therefore, 17 samples (20%) gave values above 41 copies per milliliter in the CA/CTM v2.0 test, with a median of 90 copies per milliliter (range 41-897 copies/mL). Re-analysis based on extrapolation of the Abbott calibration curves allowed 10 (59%) of these 17 samples to be quantified, with a median value of 27 copies per milliliter. Two samples with values below 40 copies per milliliter in the Abbott assay gave values above 500 copies per milliliter in the CA/CTM v2.0 assay (892 and 897 copies/mL). Drug resistance testing was used to determine whether selection of resistance mutations during antiretroviral treatment could explain the failure of the Abbott technique to detect viremia in these 2 patients. Raltegravir resistance-associated mutations in the integrase gene (G140S, E157K) were detected in one case.
Analysis According to the Viral Subtype
The HIV-1 subtype was determined for 155 of the 260 plasma samples included in this study. Seventy-eight samples (50%) belonged to subtype B, 38 (25%) to the circulating recombinant form CRF02_AG and 39 (25%) to other subtypes (A, C, D, F, G, CRF01_AE, CRF05_DF, CRF06_cpx, CRF09_cpx, and undetermined complex recombinants). The median differences between the CA/CTM v2.0 and Abbott assays were 0.19 (−0.09; 0.44), 0.15 (−0.02; 0.29) and 0.19 (0.11; 0.54) log10 copies per milliliter, respectively, for subtype B, CRF02_AG, and other subtypes (Table 2). Values obtained with the CA/CTM v2.0 assay were significantly higher than those obtained with the Abbott assay, whatever the subtype (subtype B, P < 0.001; CRF02_AG, P = 0.002; and other subtypes, P = 0.001). These discrepancies were not significantly different from one subtype group to another (Kruskal-Wallis rank test P = 0.76). The correlation curves showed good agreement between the assays, with Spearman rank correlation coefficients of 0.97, 0.98, and 0.95, respectively, for subtype B, CRF02_AG, and other recombinant forms (data not shown).
Viral RNA in all 14 plasma samples containing HIV1-group O viruses were quantified by both the CA/CTM v2.0 and Abbott techniques; the median pVL values were respectively, 2.40 (1.30; 3.51) log10 copies per milliliter and 2.52 (1.60; 3.09) log10 copies per milliliter. The median of the differences between the 2 techniques was −0.30 (−0.30; 0.22) (P = 0.548).
Viral RNA concentrations measured in culture supernatants of the HIV-1 M/O recombinant strain were similar with the CA/CTM v2.0 test (8.78 log10 copies/mL) and the Abbott test (8.67 log10 copies/mL), as were those measured in supernatants of cells infected by HIV-1 group P (8.20 and 8.54 log10 copies/mL, respectively).
We evaluated the performance of the new CA/CTM assay (v2.0) and the Abbott assay on plasma samples from 260 HIV-1 group M-infected patients. Values obtained with the 2 techniques showed a strong correlation, but the CA/CTM v2.0 assay gave significantly higher values than the Abbott assay. This difference was observed throughout the range of pVLs and regardless of the subtype (subtype B, CRF02_AG, and other circulating recombinant forms). Because of the relatively small number of patients studied, we were unable to identify the reasons for this discrepancy. The CA/CTM v2.0 assay has been reported to quantify HIV-1 RNA more accurately than the CA/CTM v1.0 test12,14,15 and the CAT test.13,15 Our results confirm the higher values obtained with the CA/CTM v2.0 test relative to the Abbott test.14,15 The median difference between the techniques was below the clinically significant threshold of 0.5 log10 copies per milliliter, although individual values for 20% of plasma samples exceeded this threshold. Only 2 samples were at least 0.5 log10 copies per milliliter higher with the Abbott assay than with the CA/CTM v2.0 assay. Thus, the difference between the assays can be clinically significant in some cases.
The CA/CTM v2.0 assay successfully quantified viruses belonging to HIV-1 group O, with values similar to those obtained with the Abbott assay, whereas the first version of the CA/CTM assay was unable to quantify this group.7 Finally, viral RNA contained in supernatants of cells infected by the newly described HIV-1 group P strain and an HIV-1 M/O recombinant were successfully quantified by both the CA/CTM v2.0 and Abbott assays.
Seventeen samples (20%) with undetectable viral load in the Abbott assay were quantifiable (≥40 copies/mL) with the CA/CTM v2.0 assay, 2 of these samples containing more than 500 copies per milliliter. Extrapolation of the Abbott calibration curves enabled 10 (59%) of these 17 samples to be quantified. These results may be related to the higher sensitivity of the CA/CTM v2.0 assay (20 copies/mL) compared with the Abbott test. It was reported that plasma samples containing HIV-1 RNA levels at the critical threshold of 50 copies per milliliter were more frequently quantifiable with the CA/CTM v1.0 technique than with the CAT test.16 Moreover, real-time PCR is more susceptible to artefacts due to plasma processing procedures than are standard PCR17 and branched-DNA assays.18
Quantification with the Abbott test might be impaired by mutations in the integrase target region. We identified 2 mutations in the integrase gene of one of the viruses in the 2 samples that were above 500 copies per milliliter in the CA/CTM v2.0 test and undetectable in the Abbott test, but analysis of this sequence showed no mismatches in the primer and probe-binding regions (data provided by Abbott Diagnostics). Moreover, we recently reported that raltegravir resistance mutations detected in a large cohort of patients did not affect pVL quantification with the Abbott test compared with the CA/CTM v1.0 assay.19 Thus, in this particular case, the overquantification observed with the CA/CTM v2.0 assay seems to reflect ongoing replication.
As in most studies comparing HIV-1 RNA quantification techniques, we were unable to assess reproducibility owing to the limited available plasma volume. Moreover, we did not study batch-to-batch variability because we did use several lots of reagents for both assays.
In conclusion, this study shows a good correlation between viral load values obtained with the CA/CTM v2.0 assay and the Abbott real-time PCR assay. Like the Abbott test, the CA/CTM v2.0 assay was able to detect a large diversity of HIV-1 strains. The CA/CTM v2.0 assay seems to be more sensitive than the Abbott assay. Thus, as previously reported, the new CA/CTM v2.0 assay represents a real improvement over the previous version. Nevertheless, the differences in pVL values obtained by the 2 assays were sometimes clinically significant (up to 1.5 log10 copies/mL). As a result, clinicians and patients must be aware that the CA/CTM v2.0 assay may show low-level viral load in patients whose viral load is undetectable with other methods.
1. Semaille C, Barin F, Cazein F, et al. Monitoring the dynamics of the HIV epidemic using assays for recent infection and serotyping among new HIV diagnoses: experience after 2 years in France. J Infect Dis
2. Chaix ML, Descamps D, Wirden M, et al. Stable frequency of HIV-1 transmitted drug resistance in patients at the time of primary infection over 1996-2006 in France. AIDS
3. Frange P, Galimand J, Vidal N, et al. New and old complex recombinant HIV-1 strains among patients with primary infection in 1996-2006 in France: the French ANRS CO06 primo cohort study. Retrovirology
4. Vessiere A, Leoz M, Brodard V, et al. First evidence of a HIV-1 M/O recombinant form circulating outside Cameroon. AIDS
5. Plantier JC, Leoz M, Dickerson JE, et al. A new human immunodeficiency virus derived from gorillas. Nat Med
6. Delaugerre C, Denis B, Peytavin G, et al. Clinical and resistance consequences of misquantification of plasma and cerebrospinal fluid human immunodeficiency virus type 1 (HIV-1) RNA in samples from an HIV-1 subtype G-infected patient. J Clin Microbiol
7. Gueudin M, Plantier JC, Lemee V, et al. Evaluation of the Roche Cobas TaqMan and Abbott RealTime extraction-quantification systems for HIV-1 subtypes. J Acquir Immune Defic Syndr
8. Wirden M, Tubiana R, Marguet F, et al. Impact of discrepancies between the Abbott realtime and cobas TaqMan assays for quantification of human immunodeficiency virus type 1 group M non-B subtypes. J Clin Microbiol
9. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
11. Yeni P. Prise en Charge Médicale des Personnes Infectées Par le VIH
. La Documentation Française. Paris, France; 2008.
12. Damond F, Avettand-Fenoel V, Collin G, et al. Evaluation of an upgraded version of the Roche Cobas AmpliPrep/Cobas TaqMan HIV-1 test for HIV-1 load quantification. J Clin Microbiol
13. De Bel A, Marissens D, Debaisieux L, et al. Correction of underquantification of human immunodeficiency virus type 1 load with the second version of the Roche Cobas AmpliPrep/Cobas TaqMan assay. J Clin Microbiol
14. Scott L, Carmona S, Stevens W. Performance of the new Roche Cobas AmpliPrep-Cobas TaqMan version 2.0 human immunodeficiency virus type 1 assay. J Clin Microbiol
15. Pas S, Rossen JW, Schoener D, et al. Performance evaluation of the new Roche Cobas AmpliPrep/Cobas TaqMan HIV-1 test version 2.0 for quantification of human immunodeficiency virus type 1 RNA. J Clin Microbiol
16. Lima V, Harrigan R, Montaner JS. Increased reporting of detectable plasma HIV-1 RNA levels at the critical threshold of 50 copies per milliliter with the Taqman assay in comparison to the Amplicor assay. J Acquir Immune Defic Syndr
17. Rebeiro PF, Kheshti A, Bebawy SS, et al. Increased detectability of plasma HIV-1 RNA after introduction of a new assay and altered specimen-processing procedures. Clin Infect Dis
18. Garcia-Bujalance S, Ladron de Guevara C, Gonzalez-Garcia J, et al. Elevation of viral load by PCR and use of plasma preparation tubes for quantification of human immunodeficiency virus type 1. J Microbiol Methods
19. Amiel C, Schneider V, Lescure FX, et al. Analysis of the impact of a raltegravir-containing regimen on HIV-1 RNA quantification by the Abbott RealTime HIV-1 assay. Paper presented at: 12th European AIDS Conference/EACS, 2009; Cologne, Germany.
Keywords:Copyright © 2011 Wolters Kluwer Health, Inc. All rights reserved.
group O; group P; HIV-1 diversity; real-time PCR; RNA quantification