Performance Characteristics of the MultiCode-RTx Hepatitis C Virus Load Test Targeting the 3′ Untranslated Region

Standards and clinical samples.Clinical specimens were submitted to Emory University Hospital Molecular Diagnostics Laboratory for routine quantitative HCV load testing by using the TaqMan test. Remaining plasma was stored at −70°C until testing with the MultiCode assay. This study was approved by the institutional review board. A total of 170 positive specimens were selected to represent a range of HCV loads and genotypes as follows: 73 HCV-positive specimens, with viral loads of >1 million IU/ml; 67 HCV-positive specimens, with viral load ranges of >10,000 IU/ml and <1 million IU/ml; 30 HCV-positive specimens, with viral loads of <10,000 IU/ml; and 32 HCV-negative specimens. Genotype information was available for 140 specimens; the remaining 30 specimens had a viral load of <10,000 and could not be genotyped. The genotype composition was as follows: genotype 1, 76 samples; genotype 2, 37 samples; genotype 3, 22 samples; and genotype 4, five samples. HCV genotyping was performed using the TruGene HCV 5′NC genotyping system (TruGene 5′NC; Siemens Healthcare Diagnostics, Inc., Tarrytown, NY). An HCV genotype panel was purchased from SeraCare Life Sciences (Milford, MA) and contained nine members (1a, 1b, 2b, 3a, 4a, 4a, 5a, 6a) and one negative panel member with expected concentrations determined by COBAS Amplicor HCV Monitor (HCV Monitor; Roche Diagnostics Corporation, Indianapolis, IN).

Viral quality assurance panels were produced by the Division of AIDS Viral Quality Assurance Laboratory (DAIDS VQA Laboratory) at Rush University Medical Center, Chicago, IL. A plasma sample was obtained from an HCV-infected individual (genotype 1b) following an institutional review board-approved protocol. The expected concentration of the HCV RNA in the plasma sample was determined using the COBAS Amplicor HCV Monitor v2.0 test. The stock material was diluted with normal human plasma to a concentration of 7.0 to 1.0 log₁₀ IU/ml. The panel included seven replicates of each concentration. Aliquots were stored at −70°C prior to testing.

Viral load assays.The MultiCode and the TaqMan tests were performed following the manufacturers' recommendations with the exception of the nucleic acid extraction procedure. For both assays, the MagNA Pure extraction instrument was used with the MagNa Pure LC total nucleic acid isolation kit. A total of 200 μl of specimen or standard was extracted and eluted into 65 μl with 50 μl to be used in the final reaction. For the limit-of-detection studies, 500 μl of specimen was used and eluted into 65 μl with 50 μl in the final reaction. The MultiCode test includes an RNA sample processing reference (SPR) that is extracted with the specimens and standards (0.325 μl SPR per specimen). The assay was run on an ABI 7500 with a standard block, and the MultiCode test analysis software (provided by the manufacturer) was used to determine HCV load.

A subset of samples were extracted using the NucliSens easyMAG following the manufacturers' recommendations; 200 μl of specimen was extracted and eluted into 65 μl with 50 μl to be used in the final reaction.

Study design.Linearity and reproducibility were determined with HCV Optiquant RNA panels (AcroMetrix, Benicia, CA); 10 replicates each of seven panel members (0; 50; 500; 5,000; 50,000; 500,000; 5,000,000 IU/ml) were run in duplicate across five runs. Each run contained a high and low positive control and a negative control. Limits of detection were determined with 44 replicates of a five-member panel (10; 25; 50; 100; 200 IU/ml) which were diluted from Optiquant RNA and run in quadruplicate across 11 runs. Agreement was determined by testing the 170 clinical specimens as described above.

Statistical analysis.HCV load values were log₁₀ transformed for analysis. Descriptive statistics and regression line equations were calculated with the analysis tool pack of Microsoft Excel 2007 (Microsoft Corp., Redmond, WA). Agreement between viral loads was assessed by the method of Bland and Altman (2).

Performance characteristics of the MultiCode HCV test.The limit of detection of the MultiCode assay was determined by testing multiple replicates of samples ranging in concentration from 1.0 to 2.3 log₁₀ IU/ml (Table 1). Based on detection of 95% of replicates, the limit of detection was 2.3 log₁₀ IU/ml (200 IU/ml). The linear range was assessed by testing aliquots of HCV RNA ranging in concentration from 1.7 to 6.7 log₁₀ IU/ml. Data shown in Fig. 1 are a mean (± standard deviation [SD]) for samples tested in duplicate in five separate runs. The assay was linear from 1.7 to 6.7 log₁₀ IU/ml. The reproducibility varied through the linear range of the assay (0.12 to 0.43 log₁₀ IU/ml), with the greatest precision at higher viral loads and lowest precision at low viral loads of 1.7 log₁₀ IU/ml (Table 2).

• Open in new tab
• Download powerpoint

FIG. 1.

Linear range of the MultiCode test using plasma samples. Data are means (±SD) for samples tested in duplicate in five separate runs (n = 10).

Genotype bias.To determine whether genotype-specific quantification bias occurs in the MultiCode test, a total of 140 clinical specimens were tested, and the results were compared to concentrations determined by the TaqMan test. Mean viral loads (±SD) for each genotype are shown in Table 3. There did not appear to be any genotype bias (genotypes 1, 2, and 3) for the MultiCode test compared to the TaqMan test. There were too few genotype 4 specimens (total of five) to draw any conclusions; however, it is interesting that four of five specimens had higher values with the MultiCode test than with the TaqMan test. The results of the HCV RNA genotype panel are shown in Table 4. The viral load values obtained with the MultiCode test were consistently lower than the expected concentrations which were determined by the Monitor HCV test, with differences ranging from 0.16 to 0.70 log₁₀ IU/ml.

Agreement between the MultiCode test and the TaqMan assay.A total of 170 clinical specimens that had detectable viral loads by the TaqMan test were run with the MultiCode test (Fig. 2). For these samples, the population mean (±SD) was 5.4 log₁₀ IU/ml (1.4 log₁₀ IU/ml) for the MultiCode test and 5.5 log₁₀ IU/ml (1.3 log₁₀ IU/ml) for the TaqMan assay. Based on the agreement analysis, the mean difference between the two tests was −0.10 log₁₀ IU/ml (95% limits of agreement, −0.76 to 0.56 log₁₀ IU/ml) (Fig. 2).

• Open in new tab
• Download powerpoint

FIG. 2.

Agreement plot for 170 plasma samples with a detectable viral load in the MultiCode and TaqMan tests. The dashed line indicates the mean difference of −0.10 log₁₀ IU/ml. The solid line represents zero bias for the samples tested, and the dot-dashed line represents 95% limits of agreement of −0.76 to 0.56 log₁₀ IU/ml (n = 170).

Viral quality assurance panel.The 77-member panel obtained from the VQA laboratory had specimens ranging in concentration from 0 to 7.0 log₁₀ IU/ml. One of seven negative specimens tested positive (Table 5). The mean values obtained with the MultiCode test were in good agreement with the expected concentrations, with mean differences ranging from 0.1 to 0.3 log₁₀ IU/ml. Interestingly, all seven replicates of the 2.0 log₁₀ IU/ml sample were detected.

Comparison of extraction methods.To evaluate the effect of different extraction methods on viral loads, plasma specimens were tested by the MagNA Pure and easyMAG extraction methods. Twenty-one positive specimens were extracted. For all 21 specimens, the viral load attained with the easyMAG was higher than that observed for the MagNA Pure. The differences ranged from 0.11 to 1.14 log₁₀ IU/ml, with a mean difference of 0.45 log₁₀ IU/ml (data not shown).