Monitoring of viral levels during therapy of hepatitis C: early response and discontinuing HCV therapy
Hepatology; Nov 2002, part 2, volume 36, number 5
Gary Davis, Division of Hepatology, Baylor University Medical Center, Dallas, TX.
ABSTRACT: Alpha interferon therapy of chronic hepatitis C is typically accompanied by a biphasic decrease in hepatitis C virus (HCV) RNA levels: an initial rapid decline during the first 24 to 48 hours, and a second more gradual decline during the following weeks. The rate of second-phase decline correlates with ultimate response to interferon treatment. Thus, assessment of early virological response (EVR) may predict outcome. Data from 2 large clinical trials of peginterferon and ribavirin were combined and analyzed to determine the optimal definition of an EVR which, if not achieved, was associated with a low likelihood of a sustained virological response (SVR). A fall in HCV RNA level to undetectable or by at least 2 log10 units after 12 weeks was found to be the optimal definition of an EVR. Among 965 patients, 778 (80%) achieved an EVR by week 12, including all except 1 patient with genotypes 2 or 3. Among 187 patients without an EVR, only 3 (1.6%) had an SVR. These findings suggest that patients with genotype 1 who do not achieve an EVR should stop treatment after 12 weeks. Use of an early stopping rule reduces treatment costs by at least 16% and avoids the inconvenience and side effects of treatment in the 19% of patients without an EVR who have little chance of a lasting virological response.
Author comments: The early stopping rules based on changes in HCV RNA levels recommended here are based on results from 2 large multicenter trials of peginterferon and ribavirin. The reliability of the definition of an EVR from these studies needs to be independently and prospectively verified in separate studies and in different populations of patients, such as children, the elderly, patients with cirrhosis and decompensation, with renal disease, human immunodeficiency virus coinfection, and after transplantation. Furthermore, with any improvement in efficacy of therapeutic regimens, the reliability of early stopping rules will need to be reassessed.
Research is needed on the implications of viral kinetics during antiviral therapy. The causes of a rapid versus slow or flat phase II response to combination therapy require careful assessment. Thus, viral kinetics can be used as a means of assessing factors that contribute to resistance to antiviral therapy including viral factors (variations in HCV sequence, inhibition of interferon actions by HCV proteins, or resistance caused by quasispecies diversity), as well as immunologic and host genetic causes (blunted T-cell responses or defective interferon signaling) and environmental causes (alcohol use, iron overload, or other medications). Viral kinetics can also be used to assess optimal dosing of antiviral agents and to assess synergy and interactions between interferon and other antiviral agents or immune modulating factors. Thus, viral kinetics, besides being useful clinically, can also provide insights into approaches to improving efficacy of therapies of chronic hepatitis C.
--All patients with chronic hepatitis C should have viral genotype determination before treatment to provide prognostic information as well as to determine the
duration of treatment and dose of ribavirin.
--All patients with genotype 1 should have a quantitative HCV RNA level determined immediately before initiation of treatment and again after 12 weeks. HCV RNA
levels at baseline and week 12 should be measured by the same assay method.
--Early virological response is best defined as a fall in the HCV RNA level by at least 2 log10 units or to an undetectable level at week 12 of treatment.
--Patients with genotype 1 who achieve EVR at week 12 should complete the full 48 weeks of treatment. Those who do not reach an EVR should discontinue therapy.
--Patients with genotype 1 who achieve an EVR, but are still HCV RNA positive at week 12, should be retested for HCV RNA at week 24 using a sensitive qualitative
PCR. Patients who are still HCV RNA positive at 24 weeks should stop therapy.
--Patients with genotype 2 or 3 should be treated for 24 weeks and need not have a 12-week assessment of EVR.
--Assessment of HCV RNA at the end of treatment may be reassuring to the patient and physician, but does not usually alter management.
--Sustained virological response should be determined by measuring HCV RNA by a sensitive method (qualitative PCR or transcription-mediated amplification) 6 months after completing treatment.
--Further HCV RNA testing in follow-up is not necessary unless serum alanine aminotransferase levels are elevated or there is a history of re-exposure to hepatitis C.
BACKGROUND: The level of hepatitis C virus (HCV) RNA in the blood of patients with chronic hepatitis C reflects a balance between virus production and clearance. Kinetic studies and mathematical modeling estimate that there is a constant high rate of production of virions in chronic hepatitis C, ranging from 1010 to 1012 per day. In addition, the half-life of HCV is short, averaging only a few hours.2 This balance of production and clearance is altered with initiation of antiviral therapy. Treatment with alpha interferon results in a decline in HCV RNA levels that can be resolved mathematically into 2 phases. After a latent period of 8 to 10 hours following the injection of interferon, there is an initial rapid decline in HCV RNA levels, the magnitude of which correlates with dose of interferon and viral genotype. The first phase decline is usually measured at 24 or 48 hours and averages 0.9 log10 units (range, 0.3 to 3.0 log10 units).1-6 This first phase decline likely reflects direct inhibition of intracellular HCV production and release.4 The models estimate that interferon's efficacy in inhibiting HCV production in this first phase ranges from
about 70% (approximately 0.7 log10 units) for standard interferon given thrice weekly to more than 90% (1 log10 unit) for high daily doses of standard interferon or once weekly peginterferon. The addition of ribavirin does not appear to alter first phase kinetics. The rate of the first phase decline is approximately 50% greater for genotypes 2 and 3 than for genotype
The second phase decline in HCV RNA levels during interferon therapy begins after 24 to 48 hours, and is slower and more variable than the first phase. The second phase decline is felt to reflect continued inhibition of replication and the gradual elimination of virus-infected cells. The second phase decay correlates less with interferon dose than the first phase, but is more rapid with peginterferon as compared with standard interferon preparations, and is considerably more rapid in patients with genotypes 2 and 3 than in those with genotype 1 infection. As observed with the first phase, ribavirin does not appear to influence second phase kinetics. There is marked patient-to-patient variation in second phase decay rates.
As might be expected, viral elimination kinetics correlate with responsiveness to interferon and, at least in part, to ultimate outcome of therapy. Most studies have reported little correlation between first phase kinetics and treatment outcome, although the failure to achieve at least a 1 log10 unit drop in HCV RNA levels within 24 hours after the first dose has been found in some studies to correlate with subsequent treatment failure. In contrast, the rate of second phase decay of HCV RNA levels appears to be a reliable predictor of durable treatment outcome. Rates of second phase decline in HCV RNA can be measured in log10 units per day and can be categorized into 3 patterns: flat, slow, or rapid. In a kinetic study comparing standard with peginterferon, the loss of detectable HCV RNA during therapy (end-of-treatment response) occurred in 4 of 13 (31%) patients with flat response, 9 of 11 (82%) with a slow response, and 7 of 9 (78%) with a rapid second phase response, rates being higher with pegylated than standard interferon. The sustained virological response (SVR) rates in the combined interferon treatment groups were 0% (none of 13) in flat responders, 27% (3 of 11) in slow responders, and 67% (6 of 9) in rapid responders. These results suggested that patients with a flat response are unlikely to have a sustained response and could discontinue treatment early, perhaps within 4 to 12 weeks of starting therapy (once the pattern of viral decline could be reliably defined).
Use of detailed kinetic studies in clinical practice, however, is not practical because of the cost and inconvenience of frequent sampling, the variability in viral quantitation assays, the complexity of mathematical model calculations, and the known patient-to-patient variability. However, data derived from small kinetic studies suggest that testing for HCV RNA at critical time points might allow clinicians to identify nonresponders early in the treatment course. These patients could stop therapy and avoid the further cost, side effects, and inconvenience of treatment.
Practical approaches for stopping therapy early were initially developed using qualitative polymerase chain reaction (PCR) determinations of HCV RNA at 12 or 24 weeks of treatment. Studies with standard interferon monotherapy showed that patients who remained HCV RNA positive at 12 weeks were unlikely to have an SVR, which led to recommendations at the 1997 Consensus Development Conference that patients receiving alpha interferon should stop therapy at 12 weeks if they remain HCV RNA positive by sensitive PCR assays. However, this early stopping rule did not perform well with more effective regimens. Studies of patients receiving alpha interferon and ribavirin indicated that a sizeable proportion of patients who achieved an SVR were still HCV RNA positive at week 12, but that persistence of this viral marker at week 24 was associated with a very low likelihood of a sustained response, findings that led to a 24-week stopping rule for combination therapy. With development of peginterferon and trials of combination therapy using peginterferon, this stopping rule needed to be reassessed. Preliminary data from 1 large clinical trial of peginterferon and ribavirin suggested that the most appropriate time point for assessing response with that regimen was also 24 weeks. Nevertheless, data from kinetic studies suggested that use of quantitative assays for HCV RNA could refine the predictive models and allow for an accurate prediction of a nonresponse at earlier time points. Most of these studies, however, were too small to provide reliable information on the accuracy of these models, and the quantitative assays used were neither standardized nor commercially available.
Recently, several standardized commercial assays for quantitation of HCV RNA levels in serum have become available. The role of these quantitative tests in
assessing early virological response (EVR) or nonresponse to the combination of peginterferon and ribavirin has now been assessed in large numbers of patients
participating in multicenter registration trials. The goal of these analyses was to determine whether reduction of the level of HCV RNA during the first weeks of combination treatment could reliably predict nonresponse, and whether this information could be used to formulate early stopping rules before 24 weeks of
treatment. Thus, early stopping rules would be developed based on a predefined virological response or endpoint at a specified time during the first weeks of treatment. The viral response and timing of this response would be chosen to optimize treatment responses. Treatment could be discontinued in those who do not achieve this EVR.
Development of early stopping rules relied on 2 important assumptions. First, rules must minimize discontinuation of treatment in patients who might ultimately respond after completion of the full course of therapy. This goal is quite different from trying to maximize the proportion of patients achieving sustained response. While this distinction may seem subtle, it is not. Minimizing loss of potential responders is the most important clinical goal in defining an early stopping rule because it provides the most sustained virological responders (though fewer patients may stop treatment early). This definition of the rule relies on maximizing the negative predictive value of the early stopping rule (i.e., ensuring that the highest possible proportion of patients not achieving EVR would in fact not respond if treatment were continued). In contrast, rules that select groups based on optimizing the positive predictive value produce the highest proportion of sustained responders (higher rate of response among those continuing treatment) and may discontinue treatment in more patients (thus reducing immediate costs), but will necessarily exclude some patients who would achieve response with continued therapy and reduce the total number of sustained responders. Second, early stopping rules assume that SVR (i.e., viral eradication) is the only beneficial endpoint of treatment. This assumption is controversial because some investigators have proposed that prolonged interferon therapy may be beneficial in decreasing necroinflammatory activity and slowing the progression of fibrosis, even in the
absence of a viral response.19 However, this hypothesis remains unproven and is currently being tested prospectively in several controlled clinical trials.
To examine whether EVR could predict treatment outcome, data from 2 recent large international clinical studies of peginterferon and ribavirin were made
available by the study sponsors, the Schering Plough Research Institute (Kenilworth, NJ) and Hoffmann-La Roche Inc. (Nutley, NJ), after agreement of the study investigators. Only those treatment groups receiving the optimum regimen in the studies were included. These regimens were peginterferon alfa-2a, 180 µg weekly combined with ribavirin 1,000 to 1,200 mg daily (n = 453); and peginterferon alfa-2b, 1.5 µg/kg weekly and ribavirin 800 mg daily (n = 512). All patients were treated for 48 weeks; early stopping rules were not used. Overall SVR rates were similar with the 2 peginterferon regimens, allowing for pooling of results. Quantitative HCV RNA was measured at baseline, 4 weeks, 12 weeks, and 24 weeks by competitive PCR, using Superquant (National Genetics Institute; Los Angeles, CA) in the peginterferon alfa-2b study and the Cobas Amplicor HCV Monitor Test (version 2.0; Roche Diagnostics, Branchburg, NJ) using appropriate dilutions of high titer samples in the peginterferon alfa-2a trial.20,21 A total of 965 patients were analyzed, 446 (67%) with genotype 1 and 277 (29%) with genotype 2 or 3. SVRs were achieved in 529 patients (55%).
These data were analyzed to answer the following questions:
1.What is the best definition of EVR?
2.What is the optimal time to determine EVR?
3.Do these stopping rules accurately predict nonresponse to peginterferon and ribavirin?
4.Should factors such as treatment compliance, genotype, and fibrosis affect the use of stopping rules?
5.What are the cost savings of early stopping rules?
The optimal definition of EVR in this analysis was a fall in HCV RNA from baseline by at least 2 log10 units or to an undetectable level by qualitative PCR at week 12 of therapy. Only 187 of 965 (19.4%) treated subjects failed to achieve this EVR, and only 3 of these patients (1.6%) had an SVR. If treatment were discontinued based on this definition of EVR, only 3 of the 529 (0.6%) patients who had an SVR would have treatment stopped prematurely (thus, negative predictive value was 98.4%). Using a fall in HCV RNA by at least 1 log10 unit as a less rigorous definition of EVR would not have reduced the number of potential treatment responders who would have stopped therapy early (no improvement in negative predictive value). In contrast, if a more rigorous cutoff of 3 logs were used, more subjects would have stopped treatment early but the number of sustained responders who would have stopped therapy early would have nearly tripled (negative predictive value,
Overall, only 68% of patients who achieved an optimally defined EVR (2 log10 unit decline in HCV RNA) had an SVR. For several reasons, this
definition of an EVR does not guarantee eradication of virus. First, not surprisingly, the positive predictive value decreases as the definition of EVR is made less rigorous. This simply reflects the fact that patients who have a less pronounced decrease in virus (e.g., a 2 log10 unit decline only) have a lower chance of a sustained response. Indeed, SVRs were achieved in 80% of patients who were HCV RNA negative at week 12 of treatment, but in only 40% in those who had a 2 log10 unit decline in HCV RNA level while remaining HCV RNA positive (data not shown). Fortunately, the majority of subjects with an EVR (greater than a 2 log10 unit decrease) at 12 weeks were HCV RNA negative (89%). At issue is whether the 11% of patients who achieve an EVR by 12 weeks but are still HCV RNA positive should be retested at 24 weeks to assess continued positivity and whether this reliably predicts nonresponse. Further analysis was available from the peginterferon alfa-2b trial on patients who had a 2 log10 unit decline in viral levels but were still HCV RNA positive at 12 weeks. Among 23 such patients, 13 were HCV RNA negative at week 24, and 6 of these patients subsequently had an SVR; in contrast, 10 patients were still HCV RNA positive at 24 weeks, and none had an SVR. Although the number of subjects in these subgroups was small, these observations suggest that patients with a 12-week EVR who are still HCV RNA positive should be retested for HCV RNA by a sensitive qualitative PCR at week 24 of treatment and stop treatment if they remain HCV RNA positive. This secondary decision point would allow for early discontinuation of therapy in a further proportion of patients.
The second important explanation for a low positive predictive value for the 12-week EVR is the inability of some patients to complete treatment. In a more detailed analysis of the data from the peginterferon alfa-2a trial, Ferenci et al.22 noted that SVR was achieved in 75% of patients who achieved EVR (2 log10 unit drop or HCV RNA negative) and received at least 80% of the dose and duration of both drugs, but was only 48% in those who could not take the full course. The reduced response was mainly because of discontinuations rather than dose modifications during treatment: sustained responses occurred in 67% of subjects who underwent dose reduction because of side effects compared with only 12% of those who required early discontinuation.22 Most dose reductions were for anemia and were limited to ribavirin alone. Similarly, in the peginterferon alfa-2b study, dose reduction or discontinuation accounted for all nondurable responders among the group of patients who were HCV RNA negative at 12 weeks. It should be encouraging to patients who achieve an EVR by 12 weeks to know that their chance of having a sustained response is between 67% and 75% if they are able to complete their treatment course, even if dose reduction is required.
Viral genotype correlates significantly with both kinetic phases of the antiviral response of HCV to interferon therapy. Genotypes 2 and 3 are far more sensitive to interferon, and the more rapid decay of HCV in these cases probably explains why sustained response rates are higher and shorter courses of therapy more effective than in patients with genotype 1. Nonetheless, the EVR definition described previously for the overall group appears to be the best early endpoint for all patients, regardless of genotype with negative predictive value of 99% for genotype 1 and 91% for genotype 2 or 3. Although 96% of patients with genotype 2 or 3 meet early response criteria, an even earlier assessment of response was not possible because of the poor negative predictive value (loss of potential responders). The high EVR at 12 weeks in patients with genotypes 2 and 3 raises the question of whether it is cost effective to perform virological testing during treatment.
Although the presence of significant hepatic fibrosis, gender, African-American race, and human immunodeficiency virus coinfection may affect the response to interferon-based therapy, there is currently insufficient information from these databases to determine whether these factors influence response to a degree that would necessitate the use of different early response criteria. Ongoing studies with different dose durations and in different patient cohorts should address these questions.
Assessment of EVR requires that quantitative HCV RNA levels be determined at baseline (before treatment) and at some predetermined time point during the course of therapy. The cost that this testing adds to the management of patients is offset by the potential savings from early discontinuation of treatment in those who do not achieve EVR. The analysis presented here is based on treatment costs only and does not consider indirect costs from missed work, decreased quality of life, or the future medical costs derived from missed responders. The percent of patients who would discontinue treatment early, the percent of sustained virological responders who would be missed, and the estimated treatment cost reduction for each definition of EVR are summarized in Table 3. If EVR is defined as a fall of the baseline HCV RNA by more than 2 log10 units or to an undetectable level, 19.4% of patients would stop treatment, a savings of 16% of total costs of treating all patients with
a full course. Only 0.6% of potential responders would be lost by premature discontinuation of therapy. While other definitions of EVR would allow greater cost savings, more patients who would have had a sustained response would have been prematurely discontinued, resulting in the potential for greater long-term disease and societal costs.
The benefit of viral testing during treatment in patients with genotypes 2 and 3 is less certain. Only 3.95% of these patients fail to achieve EVR. If the chance of SVR were similar for 24 and 48 weeks of treatment, as suggested by a recent international study,23 early discontinuation of therapy would save the cost of only an additional 12 weeks of treatment, resulting in overall cost savings that would not offset the costs of the quantitative HCV RNA testing for the cohort of patients with genotypes 2 and 3. Furthermore, 1 of 10 patients with genotype 2 or 3 who did not achieve an EVR in the combined database subsequently had a sustained response. Therefore, HCV RNA testing to assess EVR is not cost effective in patients with genotype 2 or 3; these patients should receive 24 weeks of treatment regardless of the EVR.
Virological responses during treatment with peginterferon and ribavirin, particularly loss of detectable virus by qualitative PCR, usually persist for the duration of treatment. Although end-of-treatment virological responses (no detectable HCV RNA by qualitative PCR at the end of treatment) can be confirmed by virological testing, the result does not alter clinical management. Relapse occurs in 20% to 30% of patients with an end-of-treatment response. Recently, it has been reported that 35% to 64% of patients with end-of-treatment response who are destined for relapse can be identified by using the more sensitive transcription-mediated amplification method to measure HCV RNA.24,25 Currently, knowing this information does alter management.
All patients who complete treatment with interferon-based treatment regimens should have HCV RNA tested by a sensitive method (qualitative PCR or
transcription-mediated amplification) 6 months after completion of therapy. Undetectable HCV RNA at this time connotes SVR and makes it likely that HCV has been eradicated. Relapse or reinfection during subsequent follow-up is extremely unusual. Prospective follow-up for 4 years in 322 patients with an SVR after treatment with interferon and ribavirin found that 97% to 99% were durable.26 Thus, follow-up virological testing is probably not necessary once a 6-month post-treatment SVR has been documented, unless the patient continues to have ongoing risk factors for acquisition of hepatitis C or serum aminotransferase values are found to be elevated.
Treatment guidelines and endpoints that rely on virological testing raise issues that need to be considered by the physician and testing laboratory. Phlebotomy
centers and laboratories must process samples for viral testing under conditions that do not allow degradation of HCV RNA.27 Samples should be refrigerated or
serum separated from the clot within 2 hours. Quantitative testing should be performed by a method with a wide enough dynamic range to allow for accurate assessment of the pretreatment HCV RNA level. The same assay method should be used for assessment of both the pretreatment virus level and the EVR. Despite the adoption of an international standard, different assay methods are not equivalent and wide differences in results can occur. Accurate determination of EVR is not possible unless the same assay is used for both time points.
The sensitivity and precision of testing for HCV RNA continue to improve. The availability of more sensitive qualitative tests for HCV RNA, such as transcription-mediated amplification, will not affect the early stopping rules proposed here because they are based on quantitative changes as well. More sensitive testing does appear able to identify many patients who have nondurable responses to treatment. While earlier identification of these cases does not currently change their management, future treatment protocols may need to be designed to consolidate antiviral effects in these patients.