Anomalous Quantitation Standard Growth Curves in a Laboratory-Developed Hepatitis C Virus (HCV) RNA Quantification Assay Using the TaqMan HCV Analyte-Specific Reagent

The development of ultrasensitive, real-time molecular assays combining the performance characteristics of both quantitative and qualitative reverse transcription-PCR assays has the potential to simplify hepatitis C virus (HCV) diagnostics for both health care providers and clinical laboratories by eliminating the need for complex testing algorithms and reducing inappropriate test ordering (3). While standard quality control practices, including periodic calibration and longitudinal analysis of external controls, may be essential for monitoring and maintaining consistent performance of molecular assays (2, 4, 8), individual assay reaction performance can be assessed only through the introduction of a known quantity of an internal control or quantitation standard into these individual reactions, followed by an accurate measurement of the unique signal of the internal control or quantitation standard (6, 11).

The TaqMan HCV Master Mix Analyte Specific Reagent (TaqMan HCV ASR; Roche Molecular Systems, Inc., Branchburg, NJ) and TaqMan HCV Quantitation Standard (QS; Roche Molecular Systems, Inc.) are commercially available in the United States for use in laboratory-developed HCV assays using the COBAS TaqMan 48 Analyzer (CTM 48; Roche Molecular Systems, Inc.). Laboratory-developed assays using these commercially available reagents can have analytical sensitivities of <10 IU/ml, with dynamic ranges extending up to or exceeding 5.0 × 10⁷ IU/ml. With AMPLILINK software, version 3.1.1, and CTM 48 RNA Test File Template software (Roche Molecular Systems, Inc.), the CTM 48 used in conjunction with these laboratory-developed TaqMan HCV ASR-based assays is uniquely designed to generate a series of result flags alerting operators to a variety of instrument and/or assay problems. Among them are a series of result flags specifically related to the quality of HCV target and QS data obtained from individual reactions, with a 10-character remark preceded by the result flag, either “S” (HCV target) or “Q” (QS), indicating the origin of the problem. Specific parameters used to trigger these result flags and remarks, including “Q, QS_ INVALID,” triggered by a QS critical threshold (C_T) falling outside of the acceptance range, extending from 25 to 40 cycles (S. Rose III, Roche Molecular Systems, Inc., personal communication), and “Q, RFITOOLOW,” triggered by a QS relative fluorescence increase (RFI) below a predetermined minimum threshold, have been established by the manufacturer and incorporated into the test file parameters contained in the CTM 48 RNA Test File Template software. These parameters cannot be accessed, viewed, or modified by the laboratory user.

With these laboratory-developed TaqMan HCV ASR-based assays, the C_T is defined as the fractional cycle number at which reporter dye fluorescence first exceeds a predetermined threshold and begins an exponential growth phase. Thus, the HCV target C_T is inversely related to the quantity of HCV target RNA present in a given sample, while unexpected increases in the QS C_T (obtained from a fixed amount of QS introduced into each sample during processing) may be indicative of failed or suboptimal sample recovery or amplification associated with a given sample. Specifically, when fluorescence from the reporter dye of the QS probe in an individual reaction is adversely affected by PCR inhibitors, procedural failures, or extremely high HCV RNA viral loads, the QS C_T may be delayed significantly or completely inhibited, thereby allowing the calculation of the HCV RNA viral load to be adjusted accordingly or invalidated (i.e., “Q, QS_INVALID”), if deemed appropriate. The establishment of a minimum QS RFI threshold and routine monitoring of the QS RFI among individual reactions further increase the software algorithm's capability of identifying severely inhibited reactions with the potential for producing erroneous viral load results (i.e., “Q, RFITOOLOW”) that may not be readily identified by monitoring the QS C_T alone.

While there have been a number of published evaluations of various laboratory-developed TaqMan HCV ASR-based assays (1, 3, 7), none have assessed the overall performance of the QS and associated software algorithms among large groups of clinical specimens. As a result, the impact of PCR inhibitors or poor RNA recovery on accurate HCV RNA detection and quantification by these laboratory-developed assays performed using the CTM 48 remains unknown. Furthermore, Roche Diagnostics Corp. has issued software bulletin 07-234 (9), which identifies the potential for QS growth curve anomalies characterized by QS RFI values of ≤3.0 that may not be detected by current research-use-only (RUO) assay software and may result in erroneous viral load results. Although this bulletin applies specifically to RUO assays, it may also have implications for the performance of similar laboratory-developed TaqMan HCV ASR-based assays. We conducted a retrospective study examining QS growth curves (as outlined in software bulletin 07-234 [9]) among 1,000 clinical serum specimens tested by a laboratory-developed TaqMan HCV ASR-based assay to determine whether these anomalous QS growth curves can occur with this assay.

HCV RNA was extracted from 500-μl sample aliquots by using a MagNA Pure LC (MP) instrument (Roche Diagnostics Corp., Indianapolis, IN) and the MP “Total Nucleic Acid Large Volume Serum_Plasma” protocol in conjunction with an MP Total Nucleic Acid Isolation Kit—Large Volume (Roche Diagnostics Corp.) (5). The QS was incorporated into the assay during MP sample processing by adding it directly to MP lysis/binding buffer just prior to the start of automated processing. To process an assay run consisting of 24 samples, 115 μl of QS was added to 38.0 ml of MP lysis/binding buffer and gently mixed prior to being dispensed into the appropriate MP reagent reservoirs to achieve the manufacturer's suggested final QS concentration of ∼1,000 copies/reaction. The final MP elution volume for each sample was set at 75 μl. MP sample processing was performed on four different instruments, all using MP software, version 3.0.11.

Following the completion of MP sample processing, a vial containing TaqMan HCV ASR working master mix and a K-carrier (containing previously opened K-tubes) were added to a K-carrier postelution handling block (noncooling prototype) designed specifically for use with the MP. The addition of working master mix (50 μl) and sample eluate (50 μl) to individual K-tubes was automatically performed using the MP's postelution handling capability. Immediately following the addition of sample eluate to each K-tube, the contents were automatically mixed up and down three times. Once additions and mixing were complete, K-carriers were removed from the MP and individual K-tubes were manually sealed. To reduce the potential for invalid assay results, K-carriers containing sealed K-tubes were gently tapped on the countertop to remove any air bubbles (introduced during the mixing process) from the contents of the individual K-tubes before the K-carrier was loaded into the CTM 48.

Viral load results were generated on four different CTM 48 instruments by using AMPLILINK software, version 3.1.1, CTM 48 RNA Test File Template software, and the manufacturer's suggested amplification and detection profile for use with the TaqMan HCV ASR and QS on the CTM 48 (3, 10). During the course of this study, two lots of TaqMan HCV ASR were used following independent calibration of each lot, which was performed by testing eight replicates of a commercially available panel of HCV standards (OptiQuant HCV RNA Quantification Panel; AcroMetrix Corp., Benicia, CA).

Aliquots of customized negative, low-positive, and high-positive HCV controls (AcroMetrix Corp.) were included in each assay run of up to 24 samples. Individual assay runs were considered acceptable if the negative control yielded a “Target Not Detected” result and neither the low- nor high-positive HCV control results violated the established 1_3s quality control rules, which were used to determine routine run acceptability (2, 8). All data from failed runs were excluded from further study.

Assay results and QS growth curve data for 1,000 unique clinical serum specimens tested in 50 consecutive valid assay runs performed during July 2008 were manually reviewed. A total of 517 (51.7%) specimens contained detectable HCV RNA, yielding a median viral load of 672,240 IU/ml (5.83 log₁₀ IU/ml), with a range extending from <10 IU/ml to >50,000,000 IU/ml. The invalid result rate among all 1,000 clinical specimens evaluated was 0.30% (95% confidence interval, 0.06% to 0.87%), with all three of the invalid results automatically flagged as “Q, QS_INVALID.” Of these three clinical specimens flagged as “Q, QS_INVALID,” only one had sufficient quantity remaining for repeat testing, which yielded the result obtained previously. Severely depressed or anomalous QS growth curves yielding RFI values of ≤3.0 (as described in software bulletin 07-234 [9]) were also identified in 0.40% (95% confidence interval, 0.11% to 1.00%) of the specimens analyzed, with “Target Not Detected” (n = 3) and HCV RNA-positive (n = 1, viral load of 10,601 IU/ml) results (Fig. 1). Of these four clinical specimens with anomalous QS growth curves, only two (both with a result of “Target Not Detected”) had sufficient quantity remaining for repeat testing, and both specimens yielded the same “Target Not Detected” results obtained previously but with QS RFI values of 10 and 13. No other types of QS or target growth curve abnormalities were noted among the data obtained for these 1,000 clinical specimens.

While the vast majority of our data appeared to be free of growth curve abnormalities, severely depressed or anomalous QS growth curves with RFI values of ≤3.0 were found to occur among clinical specimens yielding “Target Not Detected” and HCV RNA-positive results by our laboratory-developed TaqMan HCV ASR-based assay (Fig. 2). While these anomalies occurred at very low frequency (at <1%, consistent with the estimation described in software bulletin 07-234 [9]), failure of the assay software to automatically flag and invalidate such results suggests that, like those contained in RUO assay software, the test file parameters contained in the current CTM 48 RNA Test File Template software may be insensitive to partial PCR inhibition and may not accurately identify all specimens with the potential for erroneous viral load results.

Our findings are limited to the specific laboratory procedures described above, and results may vary with other laboratory-developed TaqMan HCV ASR-based assay protocols. Nevertheless, these findings are relevant and important to clinical laboratories performing HCV RNA quantification assays that use the TaqMan HCV ASR and QS because ASR manufacturers are restricted by the U.S. Food and Drug Administration (FDA) in their ability to communicate important issues relating to the use of their ASRs in laboratory-developed assays to their customers. In addition, Roche Molecular Systems, Inc., is unlikely to upgrade the CTM 48 RNA Test File Template software as the ASR will no longer be available commercially after November 2009 due to the recent FDA approval of the COBAS AmpliPrep/COBAS TaqMan HCV Test (Roche Molecular Systems, Inc.). Software upgrades intended to automatically flag and invalidate severely depressed or anomalous QS growth curves with RFI values of ≤4.0 have been developed for use with the COBAS AmpliPrep/COBAS TaqMan HCV Test. However, further evaluation of the effectiveness of such software upgrades will be required. In the interim, routine manual review of QS growth curves generated by laboratory-developed TaqMan HCV ASR-based assays using the CTM 48 with AMPLILINK software, version 3.1, and CTM 48 RNA Test File Template software may be warranted, with subsequent retesting of specimens showing anomalous QS growth curves, as suggested in Roche Diagnostics Corp. software bulletin 07-234 (9).

• Open in new tab
• Download powerpoint

FIG. 1.

Examples of valid but anomalous severely depressed QS growth curves exhibiting QS RFI values of ≤3.0 occurring in an HCV RNA “Target Not Detected” clinical specimen (A) and an HCV RNA-positive clinical specimen (B) are shown. Of note, QS and target growth curves generated with the CTM 48 RNA Test File Template software automatically form a baseline at an RFI value of 1 rather than 0, as described in Roche Diagnostics Corp. software bulletin 07-234 (9).

• Open in new tab
• Download powerpoint

FIG. 2.

Distribution of QS RFI values found among 1,000 clinical serum specimens tested by a laboratory-developed TaqMan HCV ASR-based assay is shown. Data from three invalid (i.e., “Q, QS_INVALID”) specimen results were excluded from this analysis.

• Copyright © 2009 American Society for Microbiology