HCV RNA detection by TMA during the hepatitis C antiviral long-term treatment against cirrhosis (Halt-C) trial


  • Potential conflicts of interest: Of the authors who have financial relationships with Hoffmann-La Roche, Inc., Timothy R. Morgan is on the speakers' bureau and receives research support; Herbert L. Bonkovsky is a consultant, on the speakers' bureau, and receives research support; Mitchell L. Shiffman is a consultant, on the speakers' bureau, and receives research support; Gregory T. Everson is a consultant, on the speakers' bureau, and receives research support; Karen L. Lindsay is a consultant and receives research support; Anna S. F. Lok is a consultant and receives research support; Adrian M. Di Bisceglie is a consultant, on the speakers' bureau, and receives research support; and William M. Lee is on the speakers' bureau and receives research support. The authors who have other financial relationships related to this project are Chihiro Morishima, David R. Gretch, Herbert L. Bonkovsky, and William M. Lee, who receive research support from Bayer Diagnostics. The authors with no financial relationships related to this project are James E. Everhart, Jules L. Dienstag, and Marc G. Ghany.

  • This is publication number 16 from the HALT-C Trial Group.


For making treatment decisions related to chronic hepatitis C, the utility of HCV RNA tests with increased sensitivity has not been defined. Prior interferon nonresponders with advanced fibrosis (n = 1,145) were retreated with peginterferon alpha-2a and ribavirin. Patients who were HCV RNA-negative by a polymerase chain reaction (PCR)-based assay (Roche COBAS Amplicor™ HCV Test, v. 2.0; lower limit of detection [LOD] 100 IU/mL) at week 20 (W20) received treatment for 48 weeks. Stored specimens were tested using the Bayer VERSANT® HCV RNA Qualitative (TMA) Assay (LOD 9.6 IU/mL) and compared to PCR results for the ability to predict sustained virological response (SVR; defined as undetectable HCV RNA by PCR at W72). Nearly all PCR-positive samples (1006/1007, 99.9%) were positive as assessed by TMA. Among 1,294 PCR-negative samples, 22% were TMA-positive. Negative TMA results were more predictive of SVR than were negative PCR results at W12 (82% vs. 64%, P < .001) and at W20 (66% vs. 52%, P = 0.001). SVR was more likely the earlier TMA had become negative during treatment (82% at W12, 44% at W20, 20% at W24). Among 45 patients who were TMA-positive but were PCR-negative at W20 and W24, none achieved SVR (95% CI: 0%-8%). Approximately 10% of patients with a single positive TMA result at the end of treatment still achieved SVR. In conclusion, negative TMA results at or after W12 were superior to negative PCR results for predicting SVR. In patients with negative PCR results during treatment, a single positive TMA test did not exclude SVR, although persistently positive tests did. (HEPATOLOGY 2006;44:360–367.)

According to current management guidelines, the success or failure of antiviral therapy for chronic hepatitis C is best assessed by hepatitis C virus (HCV) RNA testing using assays sensitive to a level of at least 100 IU/mL.1, 2 Patients who test HCV RNA-negative by such assays at least 24 weeks after completion of therapy are considered to have a sustained virological response (SVR), a critical endpoint associated with durable eradication of infection and long-term remission, if not cure, of disease.3–6

HCV RNA testing during therapy is also helpful for indicating whether a virological response has occurred and for predicting the likelihood of an eventual SVR with prolongation of therapy. Current management guidelines include the recommendation that patients treated with pegylated interferon alpha (peginterferon) and ribavirin have therapy stopped if HCV RNA is still present after 24 weeks of treatment. However, these guidelines were based on clinical trial results that relied on qualitative HCV RNA assays with lower limits of detection of approximately 50–100 IU/mL.1, 2 Recently, several assays with greater sensitivity, having a limit of 5–10 IU/mL or less, have become available.7–10 One of these tests, the Bayer VERSANT HCV RNA Qualitative Assay, based on transcription-mediated amplification (TMA) technology, has been approved by the U.S. Food and Drug Administration for detection of HCV RNA as evidence of active infection. The usefulness and potential role of this assay in monitoring the course and outcome of therapy for hepatitis C has yet to be fully defined.

Improved analytical sensitivity of the TMA assay has been reported by several laboratories.11–13 Positive TMA results at the end of treatment (week 48, or W48) have improved the identification of patients who subsequently relapse.9, 14–16 However, the clinical usefulness of the TMA assay over less sensitive PCR-based qualitative assays in monitoring responsiveness to interferon alpha–based treatment regimens at other time points is less clear. We sought to determine whether TMA testing during therapy could more accurately predict SVR to peginterferon and ribavirin therapy compared to a standard, less sensitive PCR-based HCV RNA assay.

In the current study, the utility of the TMA assay was evaluated retrospectively using stored serum samples from patients receiving peginterferon alpha-2a and ribavirin for retreatment of hepatitis C during the lead-in phase of the Hepatitis C Antiviral Long-Term Treatment against Cirrhosis (HALT-C) Trial. All patients were nonresponders to a previous course of interferon alpha and had bridging fibrosis or cirrhosis on liver biopsy. Virological results from serum samples obtained at various times during therapy were evaluated using both the Roche COBAS Amplicor HCV Test, v. 2.0 assay (or PCR assay) and the Bayer VERSANT HCV RNA Qualitative Assay (or TMA assay).


HCV, hepatitis C virus; SVR, sustained virological response.

Patients and Methods


The design of the lead-in phase of the HALT-C Trial has been described in detail.17 The 1,145 enrolled patients were nonresponders to a previous course of interferon alpha with or without ribavirin, had tested HCV RNA-positive by PCR, and had bridging fibrosis or cirrhosis on liver biopsy. All subjects were treated with peginterferon alpha-2a 180 μg weekly and ribavirin 1,000–1,200 mg daily in two divided doses. Clinical and other laboratory data were collected from all subjects according to standard procedures. Patients who tested negative for HCV RNA by the Roche COBAS Amplicor HCV Test, v. 2.0 (PCR) assay at week 20 (W20) were continued on combination therapy for a full 48 weeks and were followed thereafter through week 72 (W72) to assess for SVR. Patients who tested positive for HCV RNA at W20 by the PCR assay discontinued combination therapy at week 24 and were offered randomization to receive either long-term maintenance therapy with peginterferon alone or follow-up but no therapy. The current analysis was limited to the combination therapy phase of the HALT-C trial, and patient results beyond randomization are not included. The Institutional Review Boards of all participating institutions approved the study protocols, and written informed consent was obtained from all study subjects.

HCV RNA Qualitative and Quantitative Testing.

All serum samples were frozen at −70°C within 1 hour of blood draw and shipped to the virology core laboratory at the University of Washington for analysis. Serum HCV RNA levels were measured with the quantitative Roche COBAS Amplicor HCV Monitor v.2.0. (Roche Molecular Systems, Branchburg, NJ, LOD 600 IU/mL) as previously described.18 All negative results were retested using the qualitative Roche COBAS Amplicor HCV Test, v. 2.0 assay (Roche Molecular Systems, LOD 100 IU/mL), according to the manufacturer's instructions.

Serum testing with the Bayer VERSANT HCV RNA Qualitative (TMA) Assay (Bayer Diagnostics, Berkeley, CA) was performed according to the manufacturer's instructions, except that five additional negative controls (HCV-seronegative serum samples) and an additional low-positive (sensitivity) control were used in each run of 88 patient samples. All stages of testing by this assay—sample preparation, target amplification, and amplicon detection—were performed within a single tube. Briefly, the capture probe was hybridized to the 5′ untranslated region (UTR) of the HCV genome and the complex captured onto a magnetic microparticle. Transcription-mediated amplification (TMA) was carried out using Moloney leukemia virus reverse transcriptase and T7 RNA polymerase under isothermal conditions. Hybridization of the amplicons to two differentially modified acridinium ester molecules attached to different probes allowed for the simultaneous detection of internal control (IC) and HCV RNA targets in the same tube. Chemiluminescence was measured after oxidation and hydrolysis in relative light units (RLU). Each test result was considered valid if the IC result was reactive for that sample.

Definitions of Response.

All patients had HCV RNA testing in weeks 12, 20, 24, and 48 of combination therapy. Follow-up testing was performed at weeks 60 and 72. Early virological response, on-treatment virological response, sustained virological response, and relapse were defined on the basis of Roche COBAS Amplicor PCR results in accordance with standard definitions.2 Briefly, early virological response (EVR) was defined as at least a 2 log10 decline in HCV RNA levels or the absence of HCV RNA by the PCR test in treatment week 12. On-treatment response was defined as undetectable serum HCV RNA determined by PCR during antiviral therapy. Sustained virological response (SVR) was defined as undetectable serum HCV RNA determined by PCR after 48 weeks of combination therapy and 24 weeks of follow-up. Subjects with undetectable serum HCV RNA at W20 but with reappearance of serum HCV RNA at two or more consecutive timepoints after W20 and before or at W48 were considered breakthrough patients. Finally, any subject with detectable serum HCV RNA after W48 and by W72 was considered a relapse patient. Patients with W20 PCR-positive results and those defined as having breakthrough or relapse were considered non-SVR patients in outcome analyses.

Sample Selection for TMA Testing.

All samples from the W20 and W48 visits were tested with the TMA assay, as shown in Fig. 1, prior to testing samples from other visits. W12 samples were tested from all patients with unquantifiable HCV RNA levels by the Monitor assay at W12, and those with W20 PCR-negative results. W24 samples were tested from all patients who had PCR-negative results at W20 or at W24. Of the 2,301 samples tested by TMA, all but 20 were previously unthawed aliquots of frozen serum. The other 20 had previously been thawed once, and all were TMA-negative.

Figure 1.

HALT-C lead-in phase design and number of samples tested. The left side of the figure illustrates the design of the lead-in phase of the HALT-C Trial. Patients with positive week 20 PCR results while on combination therapy were randomized to the main HALT-C Trial, as described in the Patients and Methods section. Only those with negative PCR results at week 20 continued combination therapy for 48 weeks and had follow-up monitoring through week 72. The number of samples from each visit tested by PCR and TMA assays is shown to the right.

HCV Genotype.

HCV genotype was determined with the INNO-LiPA HCV II kit (Bayer Diagnostics, Emeryville, CA) according to the manufacturer's instructions. This method failed to determine genotype in three samples, and these were genotyped by CLIP™ sequencing (Bayer Reference Testing Laboratory, Berkeley, CA).

HCV RNA Standard.

The Nucleic Acid Panel (NAP) HCV RNA panel member (Acrometrix, Berkeley, CA) containing 500 IU/mL of HCV RNA was diluted serially in HCV RNA-negative serum to obtain samples for TMA sensitivity testing at concentrations of 20, 10, 5, and 2 IU/mL. The sensitivity of the Amplicor assay was tested on 100 IU/mL samples diluted from the 500 IU/mL panel member and the panel member containing 50 IU/mL. The source for NAP-positive members was genotype 1b HCV RNA-infected patient serum samples, calibrated by the manufacturer directly against the WHO First International Standard for HCV RNA using the Quantiplex HCV RNA version 2.0 bDNA assay (Bayer Diagnostics).

Statistical Analysis.

Standard statistical tests (chi-square test, odds ratios, McNemar test, and exact confidence intervals for proportions) were performed with SAS release 9.1 (SAS Institute, Cary, NC). Generalized estimating equations (SAS Proc Genmod) were used as needed to adjust for multiple time points per patient.


Analytical Sensitivity of the PCR and TMA Assays.

The analytical sensitivities of the PCR and TMA assays were determined using the genotype 1b Acrometrix HCV RNA NAP panel (Table 1). The PCR assay detected HCV RNA in 100% of samples (10/10) at 100 IU/mL and in 88% of samples (63/72) at 50 IU/mL. By contrast, the TMA assay detected HCV RNA in 100% of samples (10/10) at 5 IU/mL and in 90% (9/10) at 2 IU/mL. Using the standard threshold of at least 95% positivity to define the limit of detection for each assay, the lower limit of HCV RNA detection was 100 IU/mL by the PCR assay and 5 IU/mL by the TMA assay in our laboratory. Below these concentrations, detection of HCV RNA was variable.

Table 1. Analytical Sensitivity of PCR (Amplicor) and TMA Assays
# Positive/Total% Positive# Positive/Total% Positive
  • *

    NAP = Nucleic Acid Panel, genotype 1b HCV RNA standard (Acrometrix, Berkeley, CA).

5002/2100%Not tested 
10010/10100%Not tested 
5Not tested 10/10100%
2Not tested 9/1090%

Comparison of PCR and TMA Test Results.

Of the 1,145 patients enrolled in the HALT-C Trial who were treated with peginterferon and ribavirin, 373 (33%) tested HCV RNA-negative by PCR at W20 and were continued on combination therapy beyond W24 (Fig. 1). Virological response was defined by negative HCV RNA results obtained using a PCR-based test. Using this criterion, 348 patients continued combination therapy to W48 and were tested, of whom 319 were considered end-of-treatment responders. At W72, 24 weeks after discontinuation of treatments, 180 patients were considered to have achieved SVR. Figure 1 shows the number of patients tested by the PCR and TMA assays at each visit.

A comparison of the PCR and TMA test results is shown in Table 2. Of the 1,007 PCR-positive samples tested, TMA was positive in all but one (99.9% agreement). The single discrepant sample was obtained at the W20 visit of a patient infected with HCV genotype 1. A total of four aliquots from this blood draw were available for testing. Two of these four aliquots, tested in two separate runs, were HCV RNA-positive by PCR. Three of the four aliquots (two previously unthawed) were tested with the TMA test and found to be negative for HCV RNA. Additional PCR tests using oligonucleotide primers specific for regions of core and E1 in three of the four aliquots were also negative. At Week 24, this patient tested HCV RNA-negative by both PCR and TMA tests. Because of the greater than 99% concordance between the TMA and PCR positivity results, further TMA testing focused on samples that were HCV RNA-negative by PCR.

Table 2. Comparison of HCV RNA Test Results According to PCR (Amplicor) and to TMA by Visit
VisitHCV RNA-Positive by PCRHCV RNA-Negative by PCR
NPositive by TMANPositive by TMA
Week 12198198 (100%)24189 (37.0%)
Week 20697696 (99.9%)37399 (26.5%)
Week 248383 (100%)36172 (19.9%)
Week 482929 (100%)31919 (6%)
All Visits1,0071,006 (99.9%)1,294279 (21.6%)

Of the 1,294 samples that were HCV RNA-negative by the PCR assay, 279 (22%) were positive by the TMA assay (Table 2). When examined according to the week the samples were obtained, the proportion of samples that were PCR-negative but TMA-positive for HCV RNA decreased from 37% at W12 to 6% at W48 (P < .001 by test for trend). Percent agreement was lower for patients with genotype 1 than for other genotypes. For all PCR-negative samples at all time points, TMA results were positive in 24% of patients with genotype 1 infection and in 13% of those with other genotypes (data not shown). The least agreement between the two tests occurred at the W12 visit; among the 241 patients who tested HCV RNA-negative by PCR, 46% of 162 patients with genotype 1 and 19% of 79 patients with other genotypes tested positive by TMA (genotype 1 vs. other genotypes, P = .0001).

Prediction of SVR Based on PCR or TMA Results at Week 12 of Therapy.

Of 1,145 enrolled patients, 75 withdrew before W20, and 24 of the 373 who continued therapy beyond W24 did not return for the W72 visit, so that prediction of SVR from the results of W12 testing could be assessed in only 1,046 patients (Table 3). Patients with a positive PCR result at W20 were not eligible for a full course of therapy and were considered virological nonresponders. W12 results could then be assessed for the predictive value of both positive and negative PCR or TMA tests. For these analyses, all positive PCR results were assumed to be TMA-positive (as mentioned earlier in the Results section and in Table 2).

Table 3. Comparison of PCR (Amplicor) and TMA in Predicting SVR
 PCRTMAP value*
  • *

    P values are based on a McNemar test comparing the correct prediction of SVR by a negative TMA test to the correct prediction of SVR by a negative PCR test.

Negative at W12140/21963.9%113/13881.9%< .0001
Negative at W20180/34951.6%170/26766.2%.0016
Negative at W12 and W20140/21066.7%112/12688.9%< .0001
Negative at W48177/31057.1%175/29160.1%NS

At W12, the likelihood of SVR was low if HCV RNA was detectable by either PCR or TMA; SVR occurred in 7.3% of patients who were HCV RNA-positive by TMA, compared to 4.8% of those positive by PCR (data not shown). Thus, a positive result for HCV RNA at W12 did not exclude the possibility of SVR, although it was uncommon. In contrast, the absence of detectable HCV RNA at W12 was highly predictive of SVR and was more likely with a negative TMA test (82%) than with the standard PCR-based test (64%, P < .0001, Table 3). The odds ratio for SVR of a negative versus positive TMA at W12 was 57.3 (95% CI = 34.7–94.4), compared with an odds ratio of 35.5 (95% CI = 23.3–54.3) for SVR of a negative versus positive PCR result at W12.

Use of TMA to Predict SVR Among Patients With PCR-Negative Results at Week 20.

Among 349 patients who were HCV RNA-negative at week 20 according to PCR, 180 (52%) ultimately achieved an SVR with 48 weeks of combination therapy (Table 3). In comparison, among 267 patients who were HCV RNA-negative at week 20 according to TMA, 170 (66%) achieved an SVR (P = .0016 compared to PCR). SVR could not be evaluated in patients who were HCV RNA-positive at week 20 according to PCR because these patients were considered nonresponders, and combination therapy was stopped at week 24. Patients with negative results at both W12 and W20 by TMA were found to achieve SVR more frequently than those with negative results by PCR (67% vs. 89%, respectively, P < .0001).

The added value of TMA for predicting SVR when PCR was negative was evaluated (Fig. 2). At each visit (weeks 12, 20, 24, and 48), a significantly greater percentage of patients with PCR-negative results achieved SVR when their TMA test was negative rather than positive (P < .0001 for all four comparisons). Most patients (89%–91%) with a negative PCR but positive TMA result at weeks 20, 24, or 48 experienced a virological breakthrough or relapse. However, 9% to 11% of patients with a single TMA-positive result at either W20, W24, or W48 still achieved SVR.

Figure 2.

Patients with both PCR- and TMA-negative results were more likely to achieve SVR than those with PCR-negative but TMA-positive results. The percentages of patients who achieved SVR are shown according to whether they had PCR-negative, TMA-positive results (open columns) or PCR-negative, TMA-negative results (black columns) at each indicated visit. The total number of patients with PCR-negative, TMA-positive or PCR-negative, TMA-negative results is shown below the corresponding column.

Use of Sequential TMA Testing for Predicting SVR.

The earlier TMA first became negative during therapy, the better a predictor it was of SVR: 85% at W12, 44% at W20, and 21% at W24 (Table 4). A first negative TMA result at either W12 or W20 better predicted SVR than a first negative PCR result at the same time point. Importantly, if TMA was repeatedly positive through 24 weeks of therapy, no patient achieved SVR, even if HCV RNA was undetectable by PCR. Among those who achieved EVR, all patients who were HCV RNA-positive at both weeks 20 and 24 according to TMA did not achieve an SVR (Fig. 3). Whereas a single positive result by TMA could not rule out the possibility of an SVR, serial sample testing could. Thus, achieving an SVR appeared futile for patients receiving a full course of therapy who had positive TMA results through W24.

Table 4. Sustained Virological Response (SVR) Among Virological Responders at Week 20 (N = 346) According to Week of Therapy When PCR (Amplicor) or TMA First Became Negative*
First NegativeTMAPCR
NSVR (%)95% CINSVR (%)95% CI
  • *

    Excludes one patient with missing samples at W12 and two patients who were TMA-positive at week 20 and were not tested at week 24.

Week 1213385.0%78%–91%21066.7%60%–73%
Week 2012944.2%35%–53%13628.7%21%–37%
Week 244320.9%10%–36%   
Positive on All410%0%–9%   
Figure 3.

Sequential TMA-positive test results at weeks 20 and 24 predicted treatment failure. The percentages of patients who achieved SVR are shown according to their TMA results at weeks 20 and 24 (as indicated below the x-axis). All patients had early virological response (EVR) and undetectable serum HCV RNA by PCR at week 20. The total number of patients in each category is shown below the corresponding column, and 95% confidence intervals are included.

Prediction of Relapse by TMA Testing at End of Treatment.

Among the 310 patients who were HCV RNA-negative by PCR at W48, 177 (57%) had an SVR, compared to 175 of the 291 (60%) who were negative by TMA (Table 3, P = NS). The 19 patients with discrepant results (PCR-negative but TMA-positive) included two (11%) who subsequently achieved SVR. Both these patients had had at least two negative TMA results at previous visits. Thus, although highly predictive of relapse (89%), end-of-treatment testing for HCV RNA by TMA was not significantly better than PCR in predicting SVR, and detection of HCV RNA by TMA in patients who were negative by PCR did not exclude the possibility of an SVR. For the 291 patients who were TMA-negative at the end of therapy, the duration of TMA-negativity on treatment largely determined the likelihood of SVR: 86% of the 125 patients who were noted to be TMA-negative for at least 36 weeks (TMA first negative at week 12), 46% of the 146 patients who were TMA-negative for 24–36 weeks (TMA first negative at week 20 or 24), and none of the 20 patients who had been negative for less than 24 weeks (TMA first negative at week 48) achieved SVR.


The current report represents the largest and most comprehensive analysis of HCV RNA detection by the TMA assay in patients undergoing treatment for chronic hepatitis C with peginterferon and ribavirin. Our results demonstrate nearly perfect (99.9%) agreement between positive PCR and TMA results, with extensive additional testing showing the one discordant specimen to be a false-positive PCR result. The results also confirm the increased analytical sensitivity of TMA in comparison to the standard PCR-based assay for HCV RNA used in clinical practice. The approximately 1.5 log10 improvement in sensitivity likely accounted for the finding that 22% of the samples obtained during therapy tested HCV RNA-negative by PCR but HCV RNA-positive by TMA. These discrepancies were most frequent during the first 24 weeks of therapy and with genotype 1 infection, findings compatible with the known kinetics of HCV RNA elimination during interferon-based therapy, in that HCV RNA usually falls to undetectable levels by PCR (<100 IU/mL) in virological responder patients by week 24 of therapy and that viral clearance is more rapid and more frequent with genotypes 2 and 3 than with genotype 1 infection. Thus, as found in our study, HCV RNA levels from some early time points and samples from genotype 1–infected patients would have been expected to be just below the level of detection of current PCR-based assays and potentially detectable by more sensitive tests.

Although the TMA-based assay for HCV RNA was more sensitive than the PCR-based assay and detected more samples that were positive during antiviral therapy, the more sensitive assay was not always the more clinically useful one. The finding that the absence of HCV RNA at W12 according to TMA was more predictive of SVR than the absence according to PCR (82% vs. 64%) provides encouragement for the patient to complete the planned therapy. However, a high likelihood of virological response does not guarantee it in a specific patient. Furthermore, a positive result at W12 according to TMA did not exclude the possibility of SVR and was less reliable than a positive result according to PCR at predicting nonresponse. The same difficulty applied to samples from weeks 20 and 24: a negative result by TMA was more predictive of SVR than a negative result by PCR (66% vs. 52% at W20), but a positive result did not exclude the possibility of SVR. Thus, a single positive test for HCV RNA at weeks 20 or 24 during combination therapy could not be used as the basis for a decision to stop therapy. On the other hand, the finding that a positive reaction at both weeks 20 and 24 by TMA had a 100% negative predictive value for SVR in this cohort, suggests that repeatedly positive results by TMA could be used as a basis to stop therapy at 24 weeks, even in patients with undetectable HCV RNA by a standard PCR assay.

Our testing with the TMA assay identified a subset of patients with persistent low-level viremia (PCR-negative but TMA-positive) on combination therapy who did not achieve SVR. Desombere et al.19 described a small number of treatment-naive patients who exhibited a similar pattern of treatment nonresponsiveness and who subsequently relapsed. Thus, within the context of the HALT-C study, consecutive TMA testing at weeks 20 and 24 could have been used in conjunction with EVR at W12 to avoid unnecessary therapy in an additional 45 patients, without missing any subjects who could have achieved SVR. These findings warrant further verification in other cohorts of treated patients, particularly patients with milder forms of chronic hepatitis C and patients undergoing initial therapy, neither of which was characteristic of this patient cohort.

Our results agree with those of previous studies reporting that TMA may be effective in identifying patients at the end of treatment who will relapse after peginterferon therapy is discontinued.9, 14–16 However, an important caveat from our cohort was that 2 of 19 patients with undetectable HCV RNA by PCR at W48 had a positive TMA result and still achieved SVR. This finding may have been a result of the specialized characteristics of the subjects in our study, all of whom had advanced liver disease and had not been responsive to prior interferon therapy. However, repeatedly TMA-positive, but PCR-negative, results on therapy clearly identified patients certain to relapse when treatment was terminated, indicating the presence of low-level clinically relevant viremia. Whether relapse could have been prevented by continuing treatment for a longer duration remains to be confirmed in prospective controlled trials.

The detection of HCV RNA by TMA in samples that were nonreactive by PCR suggests the presence of low levels of circulating virus, with HCV RNA levels between 5 and 100 IU/mL. In accordance with this hypothesis, breakthrough, relapse and ultimate nonresponse were predicted for TMA-positive but PCR-negative patients. Between 89% and 91% of patients who tested HCV RNA-positive by TMA, but HCV RNA-negative by PCR, were virological nonresponders through 48 weeks of combination therapy. On the other hand, 9%–11% of patients with such responses achieved SVR and had apparent long-term eradication of virus. One possibility is that these patients may undergo delayed relapse, as has been described for a minority of patients years after achieving SVR.6, 20, 21 Alternatively, these results suggest that some patients clear HCV RNA late, even after discontinuation of therapy (in those patients with an SVR who were HCV RNA-positive by TMA at the end of therapy). Finally, these results may be falsely positive or indicate the detection of low levels of noninfectious virus particles of little clinical significance. These explanations remain hypotheses, however, because of the lack of long-term follow-up, as well as a gold standard against which to measure the specificity of TMA testing for HCV RNA.

Although obtained in a cohort of subjects with advanced liver disease who have been nonresponsive to previous interferon therapy, the data presented here are compelling in that they suggest an alternative and possibly improved method of monitoring patients during standard peginterferon and ribavirin therapy. A large study of previously untreated patients demonstrated that a long duration of viral negativity with a less sensitive test was associated with SVR,22 underscoring the importance of accurate determination of HCV RNA-negative status. Furthermore, serial positive tests could permit the early discontinuation of therapy in persons in whom further treatment would not be beneficial. Thus, the time to HCV RNA-negativity by TMA (or other tests with similar sensitivity for HCV RNA detection) may be the best means of predicting likelihood of SVR, and serial testing for HCV RNA by TMA, the best means of predicting nonresponse. The increased cost of repeated measurements of HCV RNA should be weighed against the increased cost and continued side effects of continuing therapy without benefit. Results from this study deserve further assessment in other cohorts of patients, particularly in those undergoing initial therapy regimens. Finally, the ability to detect extremely low levels of serum HCV RNA may allow clinicians to develop more individualized protocols for treatment duration based upon the degree and rapidity of virological response to therapy.


Funding and kits for the current study were supplied by Bayer Diagnostics through a Clinical Trial Agreement with the National Institutes of Health and a Research and Technology Development Agreement with the University of Washington. In addition to the authors of this article, the following individuals at participating institutions were instrumental in the planning, conduct, and/or care of patients enrolled in this study: Gyongyi Szabo, M.D., Maureen Cormier, R.N., Donna Giansiracusa, R.N., and Savant Mehta, M.D. (University of Massachusetts Medical Center, Worcester, MA); Michelle Kelley, R.N., A.N.P. (University of Connecticut Health Center, Farmington, CT); Bruce Bacon, M.D., Brent Neuschwander-Tetri, M.D., Debra King, R.N. (Saint Louis University School of Medicine, St Louis, MO); Andrea E. Reid, M.D., Raymond T. Chung, M.D., Wallis A. Molchen, Loriana Di Giammarino (Massachusetts General Hospital, Boston, MA); Jennifer DeSanto, R.N., Carol McKinley, R.N., Brenda Easley, R.N. (University of Colorado School of Medicine, Denver, CO); John Hoefs, M.D., Muhammad Sheikh, M.D., M. Mazen Jamal, M.D., M.P.H., Choon Park, R.N. (University of California - Irvine, Irvine, CA); Peter F. Malet, M.D., Nicole Crowder, L.V.N., Rivka Elbein, R.N., B.S.N. (University of Texas Southwestern Medical Center, Dallas, TX); Carol B. Jones, R.N., Susan L Milstein, R.N. (University of Southern California, Los Angeles, CA); Robert J. Fontana, M.D., Pamela A. Richtmyer, L.P.N., C.C.R.C. (University of Michigan Medical Center, Ann Arbor, MI); Richard K. Sterling, M.D., Charlotte Hofmann, R.N., Paula Smith, R.N. (Virginia Commonwealth University Health System, Richmond, VA); T. Jake Liang, M.D., Yoon Park, R.N., Elenita Rivera, R.N., Vanessa Haynes-Williams, R.N. (National Institute of Diabetes and Digestive and Kidney Diseases, Liver Disease Branch, Bethesda, MD); Leonard B. Seeff, M.D., Patricia R. Robuck, Ph.D., Jay H. Hoofnagle, M.D. (National Institute of Diabetes and Digestive and Kidney Diseases, Division of Digestive Diseases and Nutrition, Bethesda, M.D.); Minjun Chung, B.S., A.S.C.P., Rohit Shankar, B.S., A.S.C.P., Natalia Antonov (University of Washington, Seattle, WA); Deepa Naishadham, M.S., Kristin K Snow, M.Sc., Sc.D., Linda Massey (New England Research Institutes, Watertown, MA); and Zachary D. Goodman, M.D (Armed Forces Institute of Pathology, Washington, DC). Finally, the authors thank David Hendricks, Ph.D. (Bayer Diagnostics, Emeryville, CA) for his advice and support of this study.