Performance Evaluation of the Aptima Assays in Comparison with the cobas 6800 Assays for the Detection of HIV-1, HBV, and HCV in Clinical Samples
2022; 42(4): 447-456
Ann Lab Med 2024; 44(3): 253-261
Published online December 15, 2023 https://doi.org/10.3343/alm.2023.0306
Copyright © Korean Society for Laboratory Medicine.
Seri Jeong , M.D., Ph.D.1, Su Kyung Lee , M.S.2, Eun-Jung Cho , M.D., Ph.D.2, Han-Sung Kim , M.D., Ph.D.3, Young Kyung Lee , M.D., Ph.D.3, Jae-Seok Kim , M.D., Ph.D.4, Wonkeun Song , M.D., Ph.D.1, and Hyun Soo Kim, M.D., Ph.D.2
1Department of Laboratory Medicine, Hallym University Kangnam Sacred Heart Hospital, Hallym University College of Medicine, Seoul, Korea; 2Department of Laboratory Medicine, Hallym University Dongtan Sacred Heart Hospital, Hallym University College of Medicine, Hwaseong, Korea; 3Department of Laboratory Medicine, Hallym University Sacred Heart Hospital, Hallym University College of Medicine, Anyang, Korea; 4Department of Laboratory Medicine, Kangdong Sacred Heart Hospital, Hallym University College of Medicine, Seoul, Korea
Correspondence to: Hyun Soo Kim, M.D., Ph.D.
Department of Laboratory Medicine, Hallym University Dongtan Sacred Heart Hospital, Hallym University College of Medicine, 7 Keunjaebong-gil, Hwaseong 18450, Korea
E-mail: hskim0901@empas.com
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: Clinical management of patients infected with hepatitis B virus (HBV) or hepatitis C virus (HCV) relies on the viral load (VL). The Cobas 5800 system (Roche Diagnostics) can determine VLs in 200 and 500 μL samples, but the performance of each protocol has not been compared. We evaluated the performance of both protocols for the HBV and HCV tests.
Methods: Precision and linearity were verified using commercial panels. Probit analyses were used to determine limits of detection (LoDs). The results obtained with 336 samples were compared using the 200 and 500 μL protocols. Data from 6,737 retrospective HBV and 768 HCV samples were compared to estimate the effects of the different LoDs on the diagnostic results of the protocols. Correlations between protocols were tested with Spearman’s rank correlation coefficients (rho).
Results: The precision and linearity of both protocols were verified. The LoDs for the 200 and 500 μL protocols were 6.5 and 2.7 IU/mL for HBV and 29.7 and 8.2 IU/mL for HCV, respectively. The agreement between the protocols ranged from 0.8 to 1.0. The results obtained with the HBV and HCV tests showed a strong correlation (rho=0.994). Only 0.4% of HBV and 0.4% of HCV test results were affected by the LoDs of the 200 μL protocol.
Conclusions: The Cobas 5800 200 and 500 μL protocols for the HBV DNA and HCV RNA tests demonstrated excellent performance. These findings establish the 200 μL protocol as a new option for low-volume samples, especially for pediatric and difficult-to-bleed patients.
Keywords: Cobas 5800, Comparison, Hepatitis B virus, Hepatitis C virus, PCR, Protocol, Viral load
Determining the viral load (VL) of hepatitis B virus (HBV) or hepatitis C virus (HCV) is critical for managing patients. VL results support the diagnosis of the disease stage, decisions on treatment initiation or termination, assessments of viral responses to antiviral therapies, and confirmation of compliance [1-6]. An HBV VL threshold of 2,000 IU/mL typically necessitates treatment initiation, and patients with detectable HBV DNA, especially those with cirrhosis, should be treated [3]. Regarding HCV, viral fluctuations of >1 log10 IU/mL may indicate acute HCV infection [2]. The use of ledipasvir–sofosbuvir is indicated for patients with <6 million IU/mL of HCV VL [2]. The goal of therapies is to maintain an undetectable VL measured using molecular methods with a sensitivity of <15.0 IU/mL [4].
Among the molecular methods used for VL quantification, real-time nucleic acid amplification tests are considered the gold standard because of their higher sensitivity and wider range of measurements [7-9]. Automated systems with high testing volumes in clinical laboratories have been developed and are widely used. The Cobas 5800 system (Roche Diagnostics, Indianapolis, IN, USA) was developed for low- to mid-throughput environments and offers a fully automated workflow, including primary tube handling, nucleic acid extraction, real-time PCR amplification and detection, and data analysis [10]. The Cobas 5800 system shares key processes and reagents with the Cobas 6800 and 8800 systems [11]. The low-volume (200 µL) and high-volume (500 µL) protocols are available on the Cobas 5800 system. The 200 µL protocol is especially useful for pediatric and difficult-to-bleed patients and repeat testing with residual samples on request. However, published reports of the performance of the Cobas 5800 using both 200 and 500 µL protocols are rare.
We evaluated and compared the performances of the Cobas 5800 HBV and HCV tests using the 200 and 500 μL protocols to determine the usefulness of the low-volume protocol in clinical laboratories. We also determined the impact of both protocols on the reported HBV and HCV VLs through a retrospective analysis of data obtained during 2018–2022.
This study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board of Hallym University Dongtan Sacred Heart Hospital, Hwaseong, Korea (HDT number 2023-03-006). Patient consent was waived because the personal information was anonymized.
Plasma samples were loaded into the sample-supply module of the Cobas 5800 instrument. All assays were performed according to the manufacturer’s instructions. We applied both the 200 and 500 µL protocols when analyzing the samples. The results of the HBV and HCV tests are displayed in three formats: numerical values (IU/mL; positive), ‘<titer min’ (meaning less than the titer minimum), and ‘target not detected’ (negative). According to the manufacturer’s instructions, the titer minimum was 25.0 IU/mL for the 200 µL protocol and 10.0 IU/mL for the 500 µL protocol for the HBV assay. For the HCV assay, the titer minimum was 40.0 IU/mL for the 200 µL protocol and 15.0 IU/mL for the 500 µL protocol. The throughput of Cobas 5800 is 144 tests per 8 hrs.
Samples with two concentrations were tested in quintuplicate. The mean low concentrations ranged from 3.3 to 3.4 log10 IU/mL, and the mean high concentrations ranged from 5.3 to 6.4 log10 IU/mL. These processes were repeated for 5 days with both protocols for the HBV and HCV tests. The total SD and CV, as well as the within-run SD and CV, were calculated for each condition.
Linearity was verified using the AccuSpan HBV DNA and HCV RNA Linearity Panel (SeraCare, Milford, MA, USA). The panel comprises samples representing serial log dilutions of HBV DNA- and HCV RNA-positive plasma and one negative sample prepared using HBV DNA- and HCV RNA-negative diluents. Duplicate values were used to determine the linear range of both protocols for the HBV and HCV tests.
HBV DNA samples diluted to 135.0, 45.0, 15.0, 5.0, and 1.7 IU/mL were used to determine the LoD of the HBV DNA test. To determine the LoD of the HCV RNA test, HCV RNA samples diluted to 225.0, 75.0, 25.0, 8.3, and 2.8 IU/mL were used. For the serial dilutions, pooled human serum samples (SeraCare) were used. Twenty replicates at each concentration were assayed. The proportions of positive results for the various concentrations were used for probit regression analysis to determine the LoDs for the 200 and 500 µL protocols. Two LoDs were calculated, depending on whether the values less than the titer minimum were categorized as positive or negative.
A total of 160 and 176 plasma samples from four university hospitals were used for qualitative and quantitative comparisons of the 200 and 500 µL protocols for the HBV and HCV tests, respectively. As the ‘<titer min’ result can be interpreted to indicate a positive or negative sample, we calculated the agreement rates for both interpretations. The quantitative differences and correlations were analyzed based on the data obtained using different analyte concentrations.
The quantitative data from 6,737 patients who requested the HBV test and 768 cases of HCV reported between January 2018 and December 2022 were analyzed to estimate the effects of the 200 and 500 µL protocols on the LoDs. The samples with <20.0 IU/mL of HBV and <15.0 IU/mL of HCV could not be analyzed because specific values were not reported by the Cobas TaqMan system (Roche Diagnostics). The numbers of HBV and HCV samples with values between the LoDs obtained with the 200 µL protocol and the specific values determined using the Cobas TaqMan system were counted. We divided those totals by the respective total numbers of reported HBV and HCV samples.
The HBV DNA and HCV RNA concentrations were subjected to log10 transformation for our analyses. Analyze-it Method Evaluation Edition software (version 2.26; Analyse-it Software Ltd., Leeds, UK) was used to determine the precision and linearity values. Each LoD was determined as the 95% probability of obtaining a positive result using probit regression analysis, according to MedCalc software (version 19.8; MedCalc Software Ltd., Ostend, Belgium). This software was also used to calculate agreement and correlation values and to construct Bland–Altman plots for comparison-of-methods analysis. The extent of agreement between the 200 and 500 µL protocols was determined using Cohen’s kappa coefficient values. The kappa values were 0.8–1.0 [12]. Spearman’s rank correlation coefficients (rho) were used to test correlations between the results of both protocols. Rho values >0.9 were interpreted as reflecting very strong correlations [13].
The CV values for the 200 and 500 µL protocols on the Cobas 5800 system were higher at low concentrations (0.7% and 0.9%, respectively, for HBV; 2.8% and 3.4%, respectively, for HCV) than at high concentrations (0.3% and 0.4%, respectively, for HBV; 1.4% and 1.6%, respectively, for HCV; Table 1). The CVs for HCV (1.4%–3.4%) were higher than those for HBV (0.3%–0.9%). However, the CVs of all assays were satisfactory, having values within 4.0%. The differences in CVs obtained with either protocol were not statistically significant.
Table 1 . Repeatability and reproducibility of Cobas 5800 HBV and HCV tests using the 200 and 500 µL protocols
Protocol | Mean (log IU/mL) | SD (log IU/mL) | CV (%) | |||
---|---|---|---|---|---|---|
Within-run | Total | Within-run | Total | |||
HBV 500 | 3.3 | 0.02 | 0.02 | 0.6 | 0.7 | |
6.3 | 0.02 | 0.02 | 0.3 | 0.3 | ||
HBV 200 | 3.4 | 0.03 | 0.03 | 0.8 | 0.9 | |
6.4 | 0.02 | 0.03 | 0.3 | 0.4 | ||
HCV 500 | 3.3 | 0.05 | 0.11 | 1.5 | 3.4 | |
5.3 | 0.05 | 0.08 | 1.0 | 1.6 | ||
HCV 200 | 3.3 | 0.07 | 0.09 | 2.0 | 2.8 | |
5.3 | 0.06 | 0.07 | 1.1 | 1.4 |
Abbreviations: HBV, hepatitis B virus; HCV, hepatitis C virus.
We verified the linearity of the assigned values provided by the AccuSpan HBV DNA and HCV RNA Linearity Panel (Fig. 1). The linear range was 2.4–8.2 log10 IU/mL for the HBV 200 and 500 μL protocols. The assigned concentrations for the HCV protocols were 2.2–6.7 log10 IU/mL. The allowable nonlinearity was set to 0.2 log10 IU/mL for all protocols, according to the manufacturer’s instructions.
Probit analysis showed that the LoD of the HBV 500 µL protocol was 2.7 IU/mL at a 95% probability when the ‘<titer min’ results were designated as positive (Fig. 2A). In contrast, the LoD increased to 12.3 IU/mL when the ‘<titer min’ results were designated as negative (Fig. 2B). The LoDs obtained with the HBV 200 µL protocol were 6.5 and 22.5 IU/mL when the ‘<titer min’ results were designated as positive or negative, respectively (Fig. 2C, 2D).
Regarding the HCV assays, the LoD of the 500 µL protocol was 8.2 IU/mL when the ‘<titer min’ results were designated as positive and 16.8 IU/mL when they were designated as negative (Fig. 2E, 2F). For the 200 µL protocol, the LoD increased to 29.7 and 34.4 IU/mL when the ‘<titer min’ results were designated as positive or negative, respectively (Fig. 2G, 2H).
The qualitative agreement results are summarized in Table 2. When the values less than the titer minimum were designated as positive, the total agreement rates were 90.6% and 99.4% for the HBV and HCV tests for the 200 and 500 µL protocols, respectively. Similarly, the total agreement rates when the values less than the titer minimum were designated as negative were >90.0% (96.3% for HBV and 99.4% for HCV). The kappa values for both tests were >0.8 for both protocols, showing almost perfect agreement.
Table 2 . Agreement between the 200 and 500 µL protocols of the Cobas 5800 system*
Variable | Roche 500 µL protocol | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
HBV† | HBV‡ | HCV† | HCV‡ | |||||||||
Positive | Negative | Positive | Negative | Positive | Negative | Positive | Negative | |||||
Roche 200 µL protocol | Positive | 88 | 2 | 78 | 0 | 100 | 1 | 99 | 0 | |||
Negative | 13 | 57 | 6 | 76 | 0 | 75 | 1 | 76 | ||||
Positive agreement (%) | 87.1 (79.0–93.0) | 92.9 (85.1–97.3) | 100.0 (96.4–100.0) | 99.0 (94.6–100.0) | ||||||||
Negative agreement (%) | 96.6 (88.3–99.6) | 100.0 (95.3–100.0) | 98.7 (92.9–100.0) | 100.0 (95.3–100.0) | ||||||||
Total agreement (%) | 90.6 (85.0–94.7) | 96.3 (92.0–98.6) | 99.4 (96.88–99.99) | 99.4 (96.9–100.0) | ||||||||
Kappa value | 0.8 (0.7–0.9) | 0.9 (0.9–1.0) | 1.0 (1.0–1.0) | 1.0 (1.0–1.0) |
*Agreement rates are expressed as the 95% confidence interval.
†The ‘<titer min’ results of both the 200 and 500 µL protocols were designated as positive.
‡The ‘<titer min’ results of both the 200 and 500 µL protocols were designated as negative. The results of HBV and HCV tests are displayed in three formats: numerical value (IU/mL) (positive), ‘<titer min’ (less than titer minimum), and ‘target not detected’ (negative). According to the manufacturer’s instructions, the titer minimum was 25 IU/mL for the 200 µL protocol and 10 IU/mL for the 500 µL protocol for the HBV assay. For the HCV assay, the titer minimum was 40 IU/mL for the 200 µL protocol and 15 IU/mL for the 500 µL protocol.
Abbreviations: HBV, hepatitis B virus; HCV, hepatitis C virus.
For the positive samples, the correlations between the HBV and HCV test results for the 200 and 500 µL protocols were very strong (rho=0.994 for both HBV and HCV; Fig. 3A, 3B). The Bland–Altman plot depicts the agreement between the protocols (Fig. 3C, 3D). The mean difference HBV quantification for the 200 µL protocol was 0.1 log10 IU/mL, which was lower than that for the 500 µL protocol. In contrast, the mean difference of HCV quantification for the 200 µL protocol was 0.04 log10 IU/mL, which was higher than that for the 500 µL protocol.
The distributions of the HBV and HCV levels reported during 2018–2022 (Fig. 4) were used to determine the effects of the LoDs on the quantitative results of the protocols used. For HBV, only 27 of 6,737 patients’ results (0.40%) were affected when the LoD was set at the highest concentration (22.47 IU/mL). None of the 27 patients were of pediatric age; the ages of these outpatients diagnosed as having chronic HBV infection or cirrhosis ranged from 34 to 66 yrs. For HCV, 3 of 768 patients (0.39%) had concentrations ranging from 15 to 34 IU/mL, which may have been influenced by the assigned LoDs. The ages of these outpatients with chronic HCV infection or cirrhosis ranged from 46 to 70 yrs. None of the quantitative results were affected by the LoDs with the 500 µL protocols.
We showed the precision, linearity, detection limit, and potential impact of the LoD on the Cobas 5800 system HBV and HCV test results using the 200 and 500 µL protocols. According to a recent study, the clinical performance of the Cobas 5800 system was comparable to that of the Cobas 6800 system [10]. The Cobas 5800 system may represent a new option for optimizing the sample volume in low- to mid-throughout laboratories. To the best of our knowledge, this study represents the first comparative evaluation of 200 and 500 µL protocols of the Cobas 5800 system. We investigated the potential impact of the LoD on the HBV and HCV test results. We also determined the qualitative agreement and quantitative correlation between the low-volume 200 µL protocol and the standard 500 µL protocol to encourage the use of the low-volume protocol in clinical laboratories.
Precision, indicated by the within-run and total SD and CV, was satisfactory for the low and high concentrations evaluated in the HBV and HCV tests. All CVs for the 200 and 500 µL protocols were <4.0%. Relatively high values at low concentrations were expected owing to the inherent variability of PCR with low VL. The variability was predominantly related to within-run factors. Intrinsic variations, such as random errors in sample preparation and pre-analytical processing, should be considered. The results of a previous study designed to evaluate the Cobas 5800 system at multiple centers showed that the within-run SDs for the HBV test were 0.03–0.19, and the total SDs were 0.04–0.19 [10]. The within-run SDs for the HCV test ranged from 0.03 to 0.27, and the total SDs ranged from 0.06 to 0.33. Our data showed high precision and were consistent with the manufacturer’s performance specifications. We also verified that our data obtained with all studied protocols fell within the allowable degree of nonlinearity (0.2 log10 IU/mL) specified by the manufacturer.
The LoDs of the Cobas 5800 HBV test for the 200 and 500 µL protocols (6.5 and 2.7 IU/mL for titer minimum-positive samples and 22.5 and 12.3 IU/mL for titer minimum-negative samples) were lower than or equal to those determined in a previous study using the Cobas 6800 system (20 IU/mL for the 200 µL protocol and 10 IU/mL for the 500 µL protocol) [11]. For the HCV test, the obtained LoDs for both protocols (8.2–34.4 IU/mL) were also lower than or equal to those of the Cobas 6800 system (40 IU/mL for the 200 µL protocol and 15 IU/mL for the 500 µL protocol).
We reviewed reported VLs for the HBV and HCV tests performed on the Cobas TaqMan system within a 5-yr period (2018–2022) to estimate the effects of the LoDs on the results obtained with different protocols. Previous findings demonstrated excellent agreement between the Cobas 6800 and Cobas TaqMan assays (kappa=1.0 for both the HBV and HCV tests; R=1.0 for the HBV test and 1.00 for the HCV test) [11]. In another study, the multisite performance of the Cobas 5800 was compared with that of the Cobas 6800 and 8800 systems [10]. The correlation between the Cobas 5800 and 6800 systems was extremely high (correlation coefficient, r2=0.990–0.999), demonstrating their equivalent performance. Based on these results, we estimated the influence of the LoDs from the data obtained with each protocol using data reported for the Cobas TaqMan assays because a direct comparison was unavailable. Only 0.4% and 0.4% of the HBV and HCV test results, respectively, were considered affected by the LoDs for the used protocols. Most patients were follow-up outpatients diagnosed as having chronic infection and were older than 18 yrs. The 200 µL protocol used for low-volume samples (predominantly those from pediatric patients) may not affect the HBV and HCV test results. The clinically important HBV VL threshold is at least 2,000 IU/mL [3, 14, 15], which is markedly higher than the LoDs presented here.
We observed almost perfect agreement and a very strong correlation between the 200 and 500 µL protocols for the HBV and HCV tests. The cycle threshold (Ct) values of the discordant HBV samples were 35.7–39.6, reflecting a low VL. The Ct of one discordant HCV sample was 38.5 for the positive 200 µL (less than the titer minimum) and negative 500 µL cases. The discordant HCV samples had a Ct of 35.9 for the 500 µL protocol when the titer minimum was designated as negative (i.e., less than that for the titer minimum; Ct=37.7). These discrepancies mostly resulted from the inherent vulnerability of real-time PCR data to a low VL. Clinically, the low VL did not seem to be an issue because it could not be used to predict the absence of a sustained viral response [16, 17]. Sample preparation, reagent-handling techniques, and assay conditions should be considered to prevent the formation of bubbles and foam and potential under-aspiration of samples, resulting in variability of the molecular test results [11, 18-20]. In a previous study of the Cobas 6800 system, a very strong correlation was found between the 200 and 500 µL protocols [11]. In addition, the correlations between the VL values generated using the Cobas 5800 and 6800 systems were extremely high, with rhos >0.9 [10].
The limitation of this study is that we did not determine the HBV and HCV genotypes from the patient samples. These tests are not routinely requested for patient management. Under- or over-quantification associated with the genotypes has been reported [21-24]; however, recent data demonstrated reliable quantification with the Roche assays [11, 25]. Retrospective analysis was performed to estimate the impact of LoDs on the results obtained with different protocols. However, the analysis of samples with <20 IU/mL of HBV and 15 IU/mL of HCV was restricted because specified values were not reported for the Cobas TaqMan system.
In conclusion, we performed the low-volume 200 µL and standard-volume 500 µL protocols for the HBV DNA and HCV RNA tests with the Cobas 5800 system. To our knowledge, our results demonstrated practically useful features for the first time. Comparing the precision, linearity, LoD, and outcomes obtained using both protocols revealed satisfactory results. The 200 µL protocol should be valuable in clinical laboratories for samples obtained from pediatric and difficult-to-bleed patients.
The authors appreciate the staff at Roche Diagnostics for providing reagents and technical support. We also appreciate the Biobank at Chung-Ang University.
Kim HS designed the study, analyzed the data, and edited the manuscript. Lee SK performed the experiments and collected the data. Jeong S analyzed the data and wrote the manuscript. Cho E-J, Kim H-S, Lee YK, Kim J-S, and Song W collected the data and reviewed the manuscript. All authors approved the final version to be submitted.
None declared.
None declared.