|Home | About | Journals | Submit | Contact Us | Français|
The Cobas Taqman assay for hepatitis B virus (HBV) DNA showed linear detection over 7 logs for genotypes A to D. The coefficient of variation was 1.2% at ≥1,000 IU/ml and 22.0% at 10 IU/ml. In 97 clinical samples, the log HBV DNA/ml differed by 0.11 between Cobas Amplicor and Cobas Taqman (r2 = 0.97).
Hepatitis B virus (HBV) infection is a major cause of chronic liver disease and may result in liver cirrhosis or hepatocellular carcinoma. In chronic infection, viremia in general persists lifelong but at highly variable levels, from more than 109 copies/ml to below 100 copies/ml. Quantitation of HBV DNA is important for staging and pretreatment evaluation (5, 6, 11, 15). Moreover, antiviral therapy requires highly sensitive detection of viremia, both for monitoring the initial response and later for identifying increasing HBV DNA levels indicating drug resistance (10, 12, 14).
Methods for analyzing HBV DNA by quantitative PCR have had limitations, mainly concerning the detection range (7, 9, 18). In the last few years, real-time PCR methods with wider detection ranges have been reported (1-4, 8, 10, 12, 13, 16, 17, 19-24). Here we evaluate the detection range, reproducibility, and clinical applicability of the Cobas Taqman HBV assay (Roche Molecular Systems, Branchburg, NJ), a real-time PCR method which includes an internal quantitation standard (23).
The analyses were performed with a Cobas Taqman 48 instrument according to the manufacturer's instructions. First, HBV DNA was manually isolated from 500 μl serum. A known number of quantitation standard (QS) molecules were introduced into each specimen and were carried through the specimen preparation, amplification, and detection steps, serving as both a quantitation standard and an inhibition control. The DNA was eluted in a volume of about 80 μl, of which 50 μl was used for PCR in a reaction mixture of 100 μl, amplifying a 105-bp segment of the precore-core region. The CT values, i.e., the cycles in which the fluorescence becomes detectable for target HBV and QS, are used to calculate the target HBV concentration, which essentially equals to [QS](2ΔCT), where [QS] represents the concentration of added QS and ΔCT is the difference in CT for HBV and QS.
Linearity panels consisting of samples (genotypes A, B, C, and D) with high HBV DNA levels (>109 copies/ml as measured by Cobas Amplicor [Roche Diagnostics, Branchburg, NJ]) were prediluted to approximate levels of 108 copies/ml and serially diluted in 1:10 steps to around 102 copies/ml. At each level, two replicates were analyzed by Cobas Taqman (DNA extraction and real-time PCR). As shown in Fig. Fig.1,1, there was a good linearity over 7 logs for all genotypes (A to D). The r2 values were 0.997, 0.997, 0.997, and 0.991 for genotypes A, B, C, and D.
Reproducibility was evaluated by analyzing a genotype D sample that was diluted to six replicates at four different levels (5 × 107, 5 × 105, 5 × 103, and 50 copies/ml). Each of the 24 replicates was analyzed by Cobas Taqman (DNA extraction and real-time PCR) on 4 subsequent occasions. As shown in Fig. Fig.22 and Table Table1,1, the results were highly reproducible, and all 24 replicates were detected at all the 4 levels. At the 10-IU/ml level, 9 of 24 samples were detectable below the range of quantitation, i.e., <6 IU/ml; at this level the CT values for the 24 samples ranged from 35.6 to 37.9 (mean, 36.7; SD, 0.62; coefficient of variation [CV], 1.7%). As shown in Fig. Fig.2,2, bottom panel, the relation between input HBV DNA and ΔCT was not perfectly linear. The Cobas instrument corrects this nonlinearity by an algorithm, which applies three coefficients that are supplied for each lot of reagents. We achieved a good fit by a third-degree equation, log HBV DNA/ml = a(ΔCT)3 + b(ΔCT)2 + cΔCT + 3.5, where 3.5 represents the log concentration of QS (corresponding to 3,200 IU/ml). This provisional equation was applied on raw data to estimate and plot HBV DNA below the range of detection in three patients on therapy (see below).
Correlation between Cobas Amplicor and Cobas Taqman assays was evaluated with 97 samples previously quantified by Cobas Amplicor in clinical diagnostics. Samples that were above the linear detection range for Cobas Amplicor (200,000 copies/ml) were prediluted up to 1:100,000 prior to retesting by Cobas Amplicor. Samples that showed HBV DNA above the linear detection range for Cobas Taqman (110 million IU/ml) were retested after predilution at 1:1,000. There was a good correlation, with an r2 value of 0.97 (Fig. (Fig.3).3). The slope was 0.99, and the mean difference between Taqman and Amplicor log values was 0.11. For one sample, repeatedly discordant results were obtained, with Cobas Taqman values being around 2 logs higher than Amplicor values. Sequencing of this sample showed a genotype C sequence with mutations causing mismatches in the Amplicor probe region (position 1899 to 1922), probably explaining the lower HBV DNA level obtained by Cobas Amplicor. This finding illustrates that although the use of probes in quantitative PCR is advantageous in adding specificity, it confers a risk of underestimating certain samples due to mismatches in the probe region.
The clinical performance for measuring low viremia levels during therapy was evaluated with three patients. After liver transplantation (Fig. (Fig.4,4, top and middle panels), an initial reduction of around 2 logs was followed by a gradual decay of HBV DNA during 4 weeks, with a t1/2 value of ≈6 days. For patient A, HBV DNA was detected even after 90 days, possibly reflecting reinfection of the new liver. In samples drawn during long-term lamivudine therapy of a third patient (Fig. (Fig.4C),4C), there was an early decline of viremia with a t1/2 value of ≈9 days, followed by a decay rate that seemed to slow down over time (with a t1/2 value increasing from ≈80 days to ≈270 days). HBV DNA was detected at very low levels after 2.2 years of clinically effective treatment and was then negative on two occasions. Treatment was stopped after 3.3 years, whereupon viremia reappeared but at a lower level than before therapy was given. Thus, it is uncertain if even repeatedly negative HBV DNA tests can predict persistent loss of replication during long-term therapy. Eradication of HBV is more likely to be achievable after liver transplantation, when repeatedly negative HBV DNA may indicate that stopping antiviral or immunoglobulin therapy is possible. Conversely, this is probably not possible if viremia persists months after transplantation, as was the case for one of the patients studied here (Fig. (Fig.4A4A).
In comparison with other real-time PCR methods, the Cobas Taqman adds certain qualities. First, the results are expressed in IU/ml, which is important for comparison of results between different laboratories and assays. Second, it includes an internal combined inhibition control/QS, which is carried along with the sample through the extraction and PCR process. The QS signal also was maintained at very high HBV DNA levels and was significantly suppressed only at levels above 107 to 108 IU/ml. Although this stable QS signal is remarkable, its suppression at very high levels contributes to limiting the detection range. Thus, samples with levels above 110 million IU/ml need to be reanalyzed by Cobas Taqman after predilution. Therefore, it may be advisable to predilute all hepatitis B e antigen (HBeAg) samples 1:100, because some of them probably have levels above 110 million IU/ml. Such a dilution may become more important in the future if analysis of early kinetics proves of value for predicting sustained response.
In summary, the Cobas Taqman assay was shown to have a wide linear range and a high reproducibility, also at low viremia levels. This makes the assay useful for clinical detection and quantification of HBV, as well as for studying viral kinetics during treatment and after transplantation.
We thank Giuseppe Colucci for technical support and advice during preparation of the manuscript and Clementina Garcia, Lena Tollén, and Katarina Lindström Johansson for technical assistance.