ABSTRACT
Hepatitis B virus (HBV) DNA quantification is used to establish the prognosis of chronic HBV-related liver disease, to identify those patients who need to be treated, and to monitor the virologic response and resistance to antiviral therapies. Real-time PCR-based assays are gradually replacing other technologies for routine quantification of HBV DNA in clinical practice. The goal of this study was to evaluate the intrinsic characteristics and clinical performance of the real-time PCR Cobas AmpliPrep/Cobas TaqMan (CAP/CTM) platform for HBV DNA quantification. Specificity was satisfactory (95% confidence interval, 98.1 to 100%). Intra-assay coefficients of variation ranged from 0.22% to 2.68%, and interassay coefficients of variation ranged from 1.31% to 4.13%. Quantification was linear over the full dynamic range of quantification of the assay (1.7 to 8.0 log10 IU/ml) and was not affected by dilution. The assay was accurate regardless of the HBV genotype. Samples containing HBV DNA levels above 4.5 log10 IU/ml were slightly underestimated relative to another accurate assay based on branched-DNA technology, but this is unlikely to have noteworthy clinical implications. Thus, the CAP/CTM HBV DNA assay is sensitive, specific, and reproducible, and it accurately quantifies HBV DNA levels in patients chronically infected by HBV genotypes A to F. Samples with HBV DNA concentrations above the upper limit of quantification need to be diluted and then retested. Broad use of fully automated real-time PCR assays should improve the management of patients with chronic HBV infection.
The presence of hepatitis B virus (HBV) DNA in peripheral blood is a reliable marker of active HBV replication. HBV DNA quantification is the best method for monitoring the level of HBV replication in chronically infected patients. Detection and quantification of HBV DNA are useful (i) for establishing the prognosis of liver disease, notably the risk of progression toward cirrhosis and hepatocellular carcinoma, (ii) for identifying patients who need antiviral therapy, and (iii) for monitoring the virologic response and resistance to antiviral therapies (reviewed in reference 9). Commercial assays for detecting and quantifying HBV DNA in clinical practice have been available for several years. Most of them, however, have not been calibrated to the World Health Organization (WHO) standard (14) and still use the nonstandardized measure “copies per milliliter,” whereas international units per milliliter should be preferred (9). These assays are based either on signal amplification following molecular hybridization or on target amplification. Signal amplification assays include the Digene Hybrid-Capture assay and assays based on branched-DNA (bDNA) technology (Versant HBV DNA 3.0 assay; Siemens Medical Solutions Diagnostics, Tarrytown, NY). Target amplification techniques are based on PCR and include the Amplicor HBV Monitor assay (Roche Molecular Systems, Pleasanton, CA) and its semiautomated version, the Cobas Amplicor HBV Monitor (reviewed in reference 10).
HBV DNA quantification assays should ideally be sensitive, specific, precise, reproducible, well calibrated (accurately providing HBV DNA levels in international units per milliliter), automated, and rapid, with minimal hands-on time (16). They should also have clinically relevant dynamic ranges of quantification and should be able both to detect and to quantify HBV DNA, regardless of the HBV genotype or sequence polymorphisms. None of the commercially available HBV DNA assays based on signal amplification or classical PCR meet all these criteria (9, 10). The recent development of “real-time” PCR techniques represents a major advance in the field of viral genome quantification (1, 6, 7, 18). Indeed, these new methods are, at least theoretically, more sensitive than classical PCR techniques, they are not prone to carryover contamination, and they generally have a wide dynamic range of quantification, meaning that they can be used to quantify the full range of viral genome levels observed in treated and untreated patients. Real-time PCR is gradually replacing other technologies for routine quantification of HBV DNA in clinical practice.
A frequently used commercial real-time PCR assay for HBV DNA quantification is the Cobas AmpliPrep/Cobas TaqMan (CAP/CTM) platform (Roche Molecular Systems) (8). CAP/CTM combines the Cobas AmpliPrep instrument, which performs fully automated extraction of HBV DNA from 850 μl of patient plasma, and the Cobas TaqMan 48 Analyzer, which performs fully automated real-time PCR amplification and detection, followed by interpretation of HBV DNA levels by means of Amplilink software. The manufacturer's lower limit of detection (LOD) for the HBV DNA assay is 12 IU/ml (1.1 log10 IU/ml), and the dynamic range of quantification is from 54 to 1.1 × 108 IU/ml (1.7 to 8.0 log10 IU/ml).
The goal of this study was to evaluate the intrinsic characteristics and clinical performance of the CAP/CTM assay for HBV DNA quantification.
MATERIALS AND METHODS
Materials. (i) Standards.A standard panel of plasma samples (OptiQuant HBV DNA; AcroMetrix, Benicia, CA) containing different concentrations of HBV DNA from a single source (a patient chronically infected by HBV genotype A) was used to study the analytical performance of the assay. The seven panel samples are designated NAP-000, NAP-HBV2E2, NAP-HBV2E3, NAP-HBV2E4, NAP-HBV2E5, NAP-HBV2E6, and NAP-HBV2E7 and contain no HBV DNA or 2 × 102 IU/ml (2.3 log10 IU/ml), 2 × 103 IU/ml (3.3 log10 IU/ml), 2 × 104 IU/ml (4.3 log10 IU/ml), 2 × 105 IU/ml (5.3 log10 IU/ml), 2 × 106 IU/ml (6.3 log10 IU/ml), and 2 × 107 IU/ml HBV DNA (7.3 log10 IU/ml), respectively.
(ii) Clinical specimens.Plasma samples were obtained from patients managed in the Department of Hepatology and Gastroenterology of Henri Mondor Hospital and from blood donors diagnosed with HBV infection at the Institut National de la Transfusion Sanguine. Group A comprised 205 HBV-seronegative individuals, i.e., subjects with neither HBs antigen (HBsAg) nor anti-HBc antibodies by a third-generation enzyme immunoassay (Vitros ECi; Ortho-Clinical Diagnostics, Raritan, NJ). Group B comprised 52 patients with chronic HBV infection, all of whom had detectable HBsAg, anti-HBc antibodies, and HBV DNA. Initial genotype determination was done with the INNO-LiPA HBV genotyping assay (Innogenetics, Gent, Belgium), a reverse hybridization assay that targets a portion of the sequence encoding both HBsAg and subdomains B and C of the reverse transcriptase domain of the HBV DNA polymerase (overlapping open reading frames). The HBV genotype was confirmed in all cases by sequencing of the gene encoding both HBsAg and the HBV polymerase, followed by phylogenetic analysis. Based on these analyses, group B comprised 12 patients with HBV genotype A, 9 with genotype B, 8 with genotype C, 10 with genotype D, 10 with genotype E, and 3 with genotype F. Group C comprised four patients on nucleoside/nucleotide inhibitor therapy who developed HBV resistance and for whom several on-therapy samples were available. One of these patients was infected with genotype A, one with genotype D, and two with genotype E. Plasma was separated from whole blood by centrifugation, placed in sterile tubes, and stored at −70°C until use in this study.
Assessment of CAP/CTM performance. (i) Specificity.The specificity of the CAP/CTM assay was assessed by testing the 205 HBV-seronegative clinical specimens from group A.
(ii) Linearity, accuracy, and influence of the HBV genotype.The linearity of quantification by the CAP/CTM assay was assessed by testing the seven samples of the OptiQuant HBV DNA standard panel, which contain up to 2 × 107 IU/ml (7.3 log10 IU/ml). Each dilution of each standard was tested three times in the same experiment with both the CAP/CTM assay and the third-generation bDNA-based Versant HBV DNA 3.0 assay. This assay has a dynamic range of quantification between 357 and 17,857,000 IU/ml (2.5 to 7.3 log10 IU/ml). The average measured values were then compared with the expected values. In addition, serial 1/5 dilutions, down to signal extinction, of 12 genotype A, 9 genotype B, 8 genotype C, 9 genotype D, 9 genotype E, and 3 genotype F samples from group B were tested. The dilutions were made with the nucleic acid test dilution matrix (AcroMetrix), a defibrinated, delipidized normal human plasma. We compared the CAP/CTM results for the 52 group B samples and the 13 group C samples with those obtained for the same samples by the third-generation bDNA-based assay. The latter assay has been shown to be precise and accurate and to equally quantify HBV genotypes A to F, through the use of a set of 16 capture and 65 extender oligonucleotide probes spanning the full-length HBV genome.
(iii) Precision and reproducibility.In order to assess precision (intra-assay reproducibility), each sample in the OptiQuant HBV DNA standard panel was tested in triplicate. To assess interassay reproducibility, the low-positive control (LPC) and the high-positive control (HPC) provided with the kits were tested 35 times in the corresponding runs on different days.
HBV DNA quantification. (i) CAP/CTM.HBV DNA was extracted from 850 μl of plasma by the Cobas AmpliPrep instrument according to the manufacturer's instructions. The Cobas TaqMan 48 Analyzer was used for automated real-time PCR amplification and detection of PCR products according to the manufacturer's instructions. The data thus generated were analyzed with Amplilink software. HBV DNA levels were expressed in international units per milliliter.
(ii) bDNA.In the Versant HBV DNA 3.0 assay, HBV DNA was recovered from 50 μl of plasma and quantified by the semiautomated System 340 bDNA analyzer (Siemens Medical Solutions Diagnostics) according to the manufacturer's instructions. HBV DNA levels were expressed in international units per milliliter.
HBV genotype determination.The HBV genotype was determined by directly sequencing a portion of the overlapping genes encoding HBsAg and the B and C subdomains of the reverse transcriptase domain of HBV polymerase, followed by phylogenetic analysis. A heminested PCR procedure was used to amplify a 492-bp fragment. The first-round PCR used external sense and antisense primers POL-1 and POL-2 (12), and the second round used the same sense primer, POL-1, and the internal antisense primer HBPr-94 (15). PCR products were directly sequenced in both directions; the sequences were aligned together with prototype sequences of HBV genotypes A to H obtained from GenBank; and phylogenetic relationships were inferred.
Statistical analysis.Descriptive statistics are shown as means ± standard deviations (SD) or as medians and interquartile ranges as appropriate. Comparisons between groups were made using the Kruskal-Wallis test or the Mann-Whitney test. The relationship between quantitative variables was studied by means of regression analysis. P values of <0.05 were considered significant.
RESULTS
Intrinsic performance of the CAP/CTM assay. (i) Specificity.The specificity of the CAP/CTM assay was assessed by testing 205 samples from anti-HBV-seronegative patients (group A). None of these sample tested positive (above the LOD of 12 IU/ml); the results were expressed as “target not detected” in every case (specificity, 100%; 95% confidence interval, 98.1 to 100%).
(ii) Precision and reproducibility.Precision (intra-assay reproducibility) was assessed by testing the seven samples of the OptiQuant HBV DNA standard range, which contain 0, 2.3, 3.3, 4.3, 5.3, 6.3, and 7.3 log10 IU/ml, respectively, three times in the same experiment. As shown in Table 1, the coefficients of variation ranged from 0.22% to 2.68%. Interassay variability was assessed by testing both the HPC and the LPC 35 times in different experiments. The coefficients of variation were 1.31% and 4.13%, respectively (Table 1).
Intra-assay and interassay reproducibility of the CAP/CTM HBV real-time PCR assay
Accuracy, linear quantification, and influence of the HBV genotype. (i) Linear quantification of standard panel dilutions.The OptiQuant HBV DNA genotype A standard panel, composed of samples containing 2 × 102 IU/ml (2.3 log10 UI/ml) to 2 × 107 IU/ml (7.3 log10 IU/ml), was used to assess the linearity of HBV DNA quantification by the CAP/CTM assay. The panel was tested three times in the same experiment by both the CAP/CTM and bDNA assays. As shown in Fig. 1A, a significant relationship was found between the average HBV DNA levels measured by the CAP/CTM assay and the expected levels (r = 0.9988; P < 0.0001). The difference between the average measured and expected HBV DNA levels ranged from 0.13 to 0.52 log10 IU/ml and was larger for the highest HBV DNA levels. A significant relationship was also found between the average HBV DNA levels measured by the bDNA assay and the expected levels (r = 0.9995; P < 0.0001) (Fig. 1B). The difference between the average measured and expected HBV DNA levels ranged from 0.00 to 0.13 log10 IU/ml. The HBV DNA levels of two standards, NAP-000 and NAP-HBV2E2, were below the LOD of the bDNA assay (357 IU/ml [2.55 log10 IU/ml]).
CAP/CTM (A) and bDNA (B) assay quantification of a commercial standard panel containing 2 × 102 (2.3 log10) to 2 × 107 (7.3 log10) IU of HBV DNA/ml (OptiQuant HBV DNA; AcroMetrix, Benicia, CA). The average measured values are shown as a function of the expected values (actual HBV DNA content of the panel sample). The dashed line is the equality line.
(ii) Quantification of HBV DNA in clinical samples containing HBV genotypes A to F.Fifty-two samples from patients with chronic hepatitis B due to genotypes A to F (group B; see Materials and Methods) were tested by both the CAP/CTM and the third-generation bDNA assay. All these samples were within the dynamic range of quantification of both assays. As shown in Fig. 2, there was a significant relationship between the HBV DNA levels obtained for each sample by the CAP/CTM and bDNA methods, whatever the HBV genotype. The regression lines were always slightly below the expected equality line, because lower values were obtained by the CAP/CTM assay than by the bDNA method for most samples, usually in the higher range of HBV DNA levels (Fig. 2).
Correlation between measurements by the CAP/CTM and bDNA assays of HBV DNA levels in 52 clinical samples (group B) containing HBV genotypes A, B, C, D, E, and F.
Figure 3A shows a Bland-Altman plot of HBV DNA levels measured for the 52 group B samples by the CAP/CTM and bDNA methods. The figure plots the difference between the two measured values (the CAP/CTM minus the bDNA value) as a function of the mean of the two measurements. A moderate underestimation of HBV DNA levels by the CAP/CTM assay compared to the bDNA method was observed with 39 (75.0%) of the 52 samples, containing all HBV genotypes (median difference, −0.36 log10 IU/ml). HBV DNA levels were underestimated by the CAP/CTM assay compared to the bDNA assay for almost all samples with levels over 4.5 log10 IU/ml (on average, −0.42 ± 0.19 log10 IU/ml for samples with HBV DNA levels over 4.5 log10 IU/ml). Below 4.5 log10 IU/ml, HBV DNA levels were often moderately overestimated by the CAP/CTM assay compared to the bDNA method (average difference, +0.11 ± 0.35 log10 IU/ml, significantly different from the average difference above 4.5 log10 IU/ml [P < 0.0001]) (Fig. 3A). None of the samples had a difference (result by the CAP/CTM assay minus result by the bDNA assay) less than −1.96 times the mean difference, whereas four samples (two genotype A, one genotype D, and one genotype E sample) had a difference more than +1.96 times the mean difference. However, the individual differences between CAP/CTM and bDNA values were always below 1.0 log10 IU/ml for these samples. Box plots of individual differences between the two methods are shown for each genotype in Fig. 3B. They confirm the global, moderate underestimation of HBV DNA levels by the CAP/CTM assay compared to the bDNA method, independently of the HBV genotype. The median differences were −0.50 log10 IU/ml for genotype A, −0.40 log10 IU/ml for genotype B, −0.38 log10 IU/ml for genotype C, −0.03 log10 IU/ml for genotype D, and −0.25 log10 IU/ml for genotype E (differences were not significant). HBV genotype F is not shown in Fig. 3B, because only three samples were tested.
(A) Bland-Altman plot of HBV DNA levels in the 52 group B samples measured by the CAP/CTM and bDNA assays. The difference between the CAP/CTM and bDNA measurements is represented as a function of the mean of the two values. Different genotypes are represented by different colors. The gray area corresponds to the mean difference ± 1.96 SD. (B) Distribution of the differences between HBV DNA levels in the same samples measured by the CAP/CTM and bDNA assays according to the HBV genotype (A to E). The differences were not significant.
Figure 4 shows the distribution of individual differences (result by the CAP/CTM assay minus result by the bDNA assay) for each HBV genotype. This figure confirms the global, modest underestimation by the CAP/CTM assay relative to the bDNA method, independently of the HBV genotype. In the majority of cases, the amount of underestimation was less than −0.5 log10 IU/ml, and it was never more than −1.0 log10 IU/ml (11 samples had a difference between −0.54 and −0.83 log10 IU/ml). Overestimation relative to bDNA values was rare and exceeded +0.5 log10 IU/ml in only two cases (Fig. 4).
Individual differences between HBV DNA levels measured by the CAP/CTM and bDNA assays, ranked in increasing order of bDNA values, according to the HBV genotype. Different HBV genotypes are represented by different colors. Bars above the zero line correspond to samples for which the CAP/CTM result was higher than the bDNA result, whereas bars below that line correspond to samples for which the CAP/CTM result was lower than the bDNA result.
(iii) Quantification of HBV DNA in clinical samples containing nucleoside/nucleotide analogue-resistant HBV.One to six samples from four patients receiving nucleoside/nucleotide analogue therapy for chronic hepatitis B who developed viral resistance were tested by both the CAP/CTM assay and the third-generation bDNA assay. The results are shown in Table 2. The average difference (result by the CAP/CTM assay minus result by the bDNA assay) was −0.2 ± 0.4 log10 IU/ml (range, −0.5 to 1.0 log10 IU/ml).
HBV DNA levels measured by the CAP/CTM and bDNA-based assays in clinical samples from four patients with chronic hepatitis B treated with nucleoside/nucleotide analogues who developed resistance to these drugs
(iv) Linear quantification of serial dilutions of HBV-infected plasma.Serial 1/5 dilutions down to signal extinction were tested for 12 genotype A, 9 genotype B, 8 genotype C, 9 genotype D, 9 genotype E, and 3 genotype F samples from group B. The curves were always linear, whatever the HBV genotype, with significant Pearson's coefficients ranging from 0.9916 to 0.9994 for HBV genotype A, 0.9978 to 0.9988 for HBV genotype B, 0.9982 to 0.9999 for HBV genotype C, 0.9919 to 0.9999 for HBV genotype D, 0.9927 to 0.9992 for HBV genotype E, and 0.9946 to 0.9997 for HBV genotype F. Figure 5 shows individual examples of HBV DNA levels measured by the CAP/CTM assay in serial 1/5 dilutions of samples containing HBV genotypes A, B, C, D, E, and F. The expected difference between two successive 1/5 dilutions was 0.70 log10 IU/ml. The mean difference ± SD between the undiluted sample and the first 1/5 dilution, and between each dilution and the subsequent dilution, were 0.76 ± 0.14, 0.68 ± 0.09, 0.64 ± 0.11, 0.69 ± 0.09, 0.77 ± 0.10, 0.77 ± 0.13, and 0.78 ± 0.16 log10 IU/ml, respectively, for genotypes A, B, C, D, E, and F (no significant difference).
HBV DNA levels in serial 1/5 dilutions of patients' clinical samples containing HBV genotypes A through F (panels A through F, respectively), measured by the CAP/CTM assay. Two examples are shown for each genotype.
DISCUSSION
This analysis of the intrinsic performance of the CAP/CTM HBV DNA quantification assay shows its excellent specificity. The CAP/CTM assay was also precise and reproducible, as previously reported (2, 5, 13). Although the LOD of the CAP/CTM assay was not specifically validated here, the fact that all the standards containing more than the stated LOD (i.e., 12 IU/ml) tested positive suggests that the analytical sensitivity of the assay is in keeping with that stated by the manufacturer. We observed a strong, significant relationship between HBV DNA levels obtained for a given sample by the CAP/CTM assay and by the third-generation bDNA-based assay, regardless of the HBV genotype. The use of the third-generation bDNA assay as the comparator was justified by the fact that this assay is accurate, precise, and reproducible, is well calibrated to the WHO HBV DNA standard (as confirmed in this study [Fig. 1B]), quantifies HBV DNA independently of the HBV genotype, and is relatively immune to sequence polymorphisms, owing to the use of a large number of capture and amplification probes located at various positions along the HBV genome (4, 11, 17).
Recently, we published an extensive analysis of the performance of the CAP/CTM assay for hepatitis C virus (HCV) RNA quantification (3) and identified two technical issues with possible clinical implications: (i) a substantial global overestimation of HCV RNA levels, of the order of +0.5 to +0.7 log10 IU/ml, likely due to the presence of substances interfering with the real-time PCR in patients' blood, and (ii) marked underestimation of HCV RNA levels in approximately 15% of patients infected by HCV genotype 2 and 30% of those infected by HCV genotype 4, likely owing to mismatches in the primers and/or probe hybridization regions, related to natural polymorphisms (3). Neither problem was observed with the CAP/CTM HBV DNA assay. In particular, HBV DNA quantification was not affected by plasma dilution.
In contrast with the HCV assay, we observed modest underestimation of HBV DNA levels in the six samples of the standard panel, mostly above an HBV DNA level of 4.5 log10 IU/ml. A similar modest underestimation of HBV DNA levels above 4.5 log10 IU/ml was observed for clinical samples relative to the bDNA assay, the calibration of which we confirmed as excellent (Fig. 1B). This underestimation, which had already been observed (13), was independent of the HBV genotype. It is not surprising, since target amplification methods generally cannot reach the same level of accuracy as hybridization-based methods; this is the price to pay for better analytical sensitivity and broader dynamic ranges. Overall, the underestimation of HBV DNA levels was modest and should have no noteworthy clinical implications. In addition, quantification by the CAP/CTM assay did not appear to be affected by the presence of common amino acid substitutions associated with nucleoside/nucleotide analogue resistance.
In conclusion, this study shows that the CAP/CTM HBV DNA assay is sensitive, specific, and reproducible and that it accurately quantifies HBV DNA in samples from patients with chronic HBV infection. Quantification is linear over the full dynamic range of quantification, which covers levels observed for both treated and untreated patients with chronic hepatitis B. However, the upper limit of quantification (8.0 log10 IU/ml) is still too low to cover the full range of possible values, which may reach very high levels in some cases. Therefore, any sample above the upper limit of quantification will need to be retested after dilution, a step that does not affect quantification. In our hands, the CAP/CTM assay appeared to be suitable for large-scale routine analysis of samples containing HBV genotypes A to F. The impact of occasional nucleotide polymorphisms remains to be tested. Broad use of fully automated real-time PCR assays should improve the management of patients with chronic HBV infection, as well as the monitoring of antiviral therapy and drug resistance.
ACKNOWLEDGMENTS
The CAP/CTM kits used in this study were kindly provided by Roche Diagnostics (Meylan, France). This work is part of the activity of the VIRGIL European Network of Excellence in Antiviral Drug Resistance, which is supported by a grant (LSHM-CT-2004-503359) from the Priority 1 “Life Sciences, Genomics and Biotechnology for Health” program in the 6th Framework Programme of the European Union.
We are grateful to Françoise Darthuy, Jocelyne Rémiré, Guillaume Dameron, and Françoise Bouchardeau for technical assistance.
FOOTNOTES
- Received 20 June 2007.
- Returned for modification 30 October 2007.
- Accepted 8 February 2008.
- Copyright © 2008 American Society for Microbiology
