Article first published online: 28 NOV 2006
Copyright © 2006 American Association for the Study of Liver Diseases
Volume 44, Issue 6, pages 1598–1606, December 2006
How to Cite
Del Rio, R. A., Post, A. B. and Singer, M. E. (2006), Cost-effectiveness of hematologic growth factors for anemia occurring during hepatitis C combination therapy. Hepatology, 44: 1598–1606. doi: 10.1002/hep.21409
See Editorial on Page 1400
Potential conflict of interest: Dr. Post owns stock in Amgen. He is on the speakers' bureau for Schering-Plough, Roche, and Valeant.
- Issue published online: 28 NOV 2006
- Article first published online: 28 NOV 2006
- Manuscript Accepted: 29 AUG 2006
- Manuscript Received: 21 OCT 2005
In hepatitis C virus (HCV)-infected patients who develop anemia during combination therapy, erythropoietic growth factors maintain higher drug treatment levels compared to ribavirin dose reduction, which may lead to an increase in treatment response rates. This study estimated the cost-effectiveness of growth factor therapy in maintaining anemic HCV-infected patients on target drug levels during combination therapy. A decision analysis using a Markov model was developed with 7 health states: Sustained viral response, chronic HCV, compensated cirrhosis, decompensated cirrhosis, hepatocellular carcinoma, liver transplantation, and death. Data sources included population-based studies of growth factor therapy, previously published estimates of costs and natural history of hepatitis C, and recent prospective studies. Our reference case was a 45-year-old Caucasian man with HCV infection (genotype 1, 2, or 3) who developed anemia while undergoing combination therapy with ribavirin and pegylated interferon. We compared growth factor injections (darbepoetin alpha or epoetin alpha) during combination therapy with standard ribavirin dose reduction. Compared to a ribavirin dose reduction strategy, the cost of darbepoetin per additional quality-adjusted life-year was $34,793 for genotype 1 and $33,832 for genotypes 2 or 3 versus $60,600 and $64,311 for epoetin. For all genotypes, the results were sensitive to changes in the cure rates of HCV therapy, the utility of chronic HCV, the costs of growth factors, and the age at which therapy is begun. In conclusion, use of erythropoietic growth factors, specifically darbepoetin, for patients with anemia occurring during HCV combination therapy appears to be cost-effective for genotypes 1, 2, or 3. (HEPATOLOGY 2006;44:1598–1606.)
Over 4 million people in the United States are infected with hepatitis C, making it the most common chronic blood-borne infection in the United States.1 Chronic hepatitis C infection causes progressive liver fibrosis, leading to the development of cirrhosis and hepatocellular carcinoma. Hepatitis C is also the most common indication for liver transplantation, accounting for over one third of all liver transplantations.2 The current recommendation for the treatment of hepatitis C infection is combination therapy with pegylated interferon (alpha 2b or 2a) plus ribavirin (RBV).1, 3, 4
Anemia is a frequent side effect of hepatitis C combination therapy, with over 40% of patients experiencing a drop in hemoglobin greater than 3 g/dL, and it is the most common reason for dose reduction and discontinuation of therapy.5 The etiology of anemia in combination therapy is multifactorial and includes: (1) a hemolytic anemia from ribavirin6; (2) a myelosuppressive effect from interferon7; and (3) a less than expected response in hematopoietic growth factor production for the level of anemia observed.7, 8
The standard of care for the treatment of anemia occurring during combination therapy has been RBV dose reduction.6 However, studies have demonstrated that sustained viral response (SVR) rates are significantly lower for genotype 1–infected patients who undergo RBV dose reduction compared to those who are able to maintain higher doses throughout combination therapy.3, 9, 10
Given the negative impact of lower RBV doses on SVR, RBV dose reduction as the primary strategy of anemia management is being reconsidered. The use of recombinant erythropoietin for anemia during combination therapy has been shown to increase hemoglobin levels to near pretreatment levels in the majority of patients11–15 and allow for the maintenance of higher RBV doses.13–15 Based on these findings, many authors are now considering the off-label use of hematologic growth factors a valid option for the treatment of anemia during hepatitis C therapy, and their use has become common.16–18 However, despite the benefits of growth factors, they are an expensive intervention and may increase the weekly costs of combination therapy by as much as 70%.19
Therefore, the purpose of this study was to perform a decision analysis to determine, among erythropoietin administration and RBV dose reduction, the more cost-effective approach for the treatment of anemia during hepatitis C combination therapy.
Patients and Methods
A decision analytic Markov model20, 21 was developed to examine the cost-effectiveness of epoetin alpha or darbepoetin alpha use for anemia occurring during hepatitis C combination therapy with pegylated interferon and ribavirin. Significant anemia requiring intervention was defined as a decrease in hemoglobin >3 g/dL from the patient's baseline hemoglobin level occurring within the first 4 weeks of combination therapy. Two treatment options were compared: (1) erythropoietin administration, with nonresponders also undergoing RBV dose reduction; and (2) RBV dose reduction (Fig. 1).
The model allowed for the simulation of a hypothetical cohort of identical patients experiencing anemia after 4 weeks of hepatitis C combination therapy. Separate analyses were performed for genotype 1 and genotype 2 or 3 infection. The reference case for each analysis was a 45-year-old Caucasian man with hepatitis C virus (HCV) infection not previously treated who developed anemia 4 weeks into combination therapy.
HCV Therapy for Genotype 1.
The model for genotype 1 infection assumed a 48-week course of HCV therapy. HCV therapy in this group consisted of RBV 1,200 mg/d and weekly injections of a pegylated interferon, either alpha 2b (1.5 μg/kg) or alpha 2a (180 μg). Patients who did not attain an early virologic response after 12 weeks of treatment, defined as a >2-log decrease in HCV RNA by polymerase chain reaction, would discontinue therapy. Approximately 29% of genotype 1 patients do not attain an early virologic response.22 A distinction between pegylated interferon alpha 2b or 2a was not made, because both appear to have similar efficacy and costs.3, 4, 19 Our model assumed there was no spontaneous clearance of HCV infection.
HCV Therapy for Genotype 2.
The model for genotype 2 or 3 infection assumed a 24-week course of therapy. HCV therapy consisted of RBV 800 mg/d and weekly injections of a pegylated interferon, either alpha 2b (1.5 μg/kg) or alpha 2a (180 μg). With regard to early virologic response, there were no stopping rules for genotype 2 or 3.
Studies using pegylated interferon plus RBV demonstrate an overall SVR of approximately 45% in patients with genotype 1 infection and 80% in patients with genotype 2 or 3 infection.3, 4 SVR rates appear significantly lower for genotype 1–infected patients who have their RBV doses decreased compared to those who are able to maintain higher doses throughout combination therapy.3, 9, 10 Lower RBV levels, defined as doses <800 mg/d, do not appear to affect SVR rates for those patients infected with genotype 2 or 3.3, 9, 10 However, because hematologic growth factors may also relieve symptoms of anemia and lower the ribavirin discontinuation rate due to anemia, we also evaluated patients with genotype 2 or 3 infection to determine if erythropoietin use in these patients could be cost-effective.
After treatment, the patient cohort entered a Markov process and was followed up until death (Fig. 2). The Markov model was used to represent patient transition over time from one state of health to another.20, 21 The seven health states included in the model were SVR, chronic hepatitis C, compensated cirrhosis, decompensated cirrhosis, hepatocellular carcinoma, liver transplantation, and death. Our model assumed that patients who achieved an SVR were cured from HCV infection. In addition, patients who did not attain an SVR had a prognosis similar to those who had never been treated.
Our analysis was performed from the societal perspective, discounting both costs and health benefits at 3% per year, as per the recommendations of the United States Public Health Service Panel on Cost–Effectiveness in Health and Medicine.20
Treatment of Anemia
Hematologic Growth Factor Therapy.
Anemia from HCV combination therapy usually occurs within the first 4 weeks of treatment and persists until the end of therapy.5 Therefore, patients receiving growth factors would begin treatment at week 4 (epoetin alpha: 40,000 units subcutaneously once a week; darbepoetin alpha 3 μg/kg subcutaneously once every 2 weeks) and continue injections until the end of combination therapy. Of those patients who respond to injections, most will experience a response within 4 weeks (http://www.pdr.com). Therefore, patients whose hemoglobin levels did not rise with growth factor therapy after 4 weeks, or nonresponders, would then undergo stepwise dose reduction of RBV.
Recombinant erythropoietin is generally safe and well tolerated, especially when administered subcutaneously. Therefore, our model assumed that any serious adverse events that may have occurred from growth factor treatment were negligible.
RBV Dose Reduction.
Patients undergoing a RBV dose reduction strategy would have RBV doses initially decreased to 800 mg/d for genotype 1 or 600 mg/d for genotype 2 or 3, and then to 600 mg/d or 400 mg/d, respectively, if hemoglobin levels did not rise after 4 weeks. Nonresponders to stepwise dose reduction would discontinue RBV therapy and continue pegylated interferon monotherapy.
The ability of each management strategy to improve hemoglobin levels and maintain higher RBV dosing levels was based on recently published clinical trials.11, 13, 14 Although the exact protocol for hematologic growth factor administration and RBV dose reduction differed somewhat across these studies, the major aspects of each management strategy, such as initial drug dosing and adjustment intervals, were similar across the trials. These studies demonstrated maintenance of RBV dosing >800 mg/d for over 85% of patients treated with growth factors versus <60% for those on a RBV dose reduction strategy. SVR rates based on RBV levels for genotype 1–infected individuals were based on previous studies and averaged almost 60% for patients receiving higher RBV dosing versus approximately 35% for patients receiving lower RBV dosing.3, 9, 10 For genotype 2 or 3 infection, RBV dosing of <800 mg/d did not appear to affect SVR rates and remains approximately 80%. Our model assumed that the probability of achieving an SVR based on RBV dose was independent of the strategy used (growth factors versus RBV dose reduction) or the number of treatment adjustments necessary to improve hemoglobin levels. Estimates for treatment response rates are listed in Table 1.
|Epoetin alpha therapy|
|Maintain RBV level ≥800 mg/d||0.87 (0.73–0.99)||11–14|
|Discontinue RBV due to anemia||0.018 (0.01–0.05)||3,4,13,14|
|Dose reduction of RBV|
|Maintain RBV level ≥800 mg/d||0.58 (0.30–0.65)||11–14|
|Discontinue RBV due to anemia||0.139 (0.05–0.40)||3,4,13,14|
|HCV genotype 1|
|Achieve SVR with RBV level ≥800 mg/d||0.57 (0.40–0.83)||3,9,10|
|Achieve SVR with RBV level <800 mg/d||0.36 (0.17–0.56)||3,9,10|
|Achieve SVR with pegylated interferon monotherapy||0.21 (0.14–0.28)||3,4|
|Early virologic failure||0.29 (0.24–0.34)||22|
|HCV genotypes 2 or 3|
|Achieve SVR with RBV level ≥800 mg/d||0.84 (0.50–0.99)||3,9,10|
|Achieve SVR with RBV level <800 mg/d||0.84 (0.50–0.99)||3,9,10|
|Achieve SVR with pegylated interferon monotherapy||0.45 (0.33–0.57)||3,4|
Natural History of Hepatitis C
Derivation of the estimates of progression to cirrhosis and hepatocellular carcinoma in patients with chronic hepatitis C has been described previously23 and was obtained from the most widely quoted published data.24–33 These estimates were revised to reflect our current understanding of hepatitis C disease progression based on recent research.1, 34 Estimates of liver transplantation rates and survival outcomes were also obtained from the literature.2, 33, 35–37 Transition probabilities are summarized in Table 2.
|From||To||Baseline, % (Range)||References|
|Chronic hepatitis C||Compensated cirrhosis||1.0 (0.1–7.3)||23,24,29,34,38|
|Compensated cirrhosis||Decompensated cirrhosis||3.9 (2.0–8.3)||23,30,42|
|Hepatocellular carcinoma||1.4 (0–4.4)||30,32|
|Decompensated cirrhosis||Hepatocellular carcinoma||1.4 (1.0–4.4)||30,32|
|Liver transplantation||3.1 (1.0–6.2)||33|
|Hepatocellular carcinoma||Death||42.7 (33–86)||29,30,33,34|
|Liver transplantation||Death (1st year)||21 (6–42)||2,33,35–37|
|Death (after 1st year)||5.7 (2.4–11)||2,33,35–37|
The rate of progression of chronic hepatitis C infection to cirrhosis remains controversial. Studies that focused on symptomatic patients from tertiary care centers who were likely to be rapid progressors have reported rates of 7.3% per year.24 On the other hand, asymptomatic patients are more likely to progress at slower rates.26, 27, 34 The NIH Consensus Statement on the management of hepatitis C estimates a 20-year risk of cirrhosis of approximately 15%–20%.38 To account for this variability, we used a baseline rate of progression of 1% per year, while examining rates as low as 0.1% and as high as 7.3% in the sensitivity analysis.
Quality of Life
Quality of life (QOL) adjustments were based on health state utility scores. Literature on the quality of life associated with hepatitis C combination therapy is limited, especially with regard to treatment for anemia. In addition, the studies that do compare these strategies do not directly measure patients' values or preferences for health outcomes, which is an important dimension of quality of life and health state utilities.39 Nonetheless, the available data suggest that patients receiving growth factors for anemia during combination therapy experience significant improvements in quality of life across many domains (physical, social, mental) compared to patients on a RBV dose reduction strategy.13, 14 Therefore, these utilities were estimated using scores that have been published in cost-effectiveness studies for two other groups of anemic patients who regularly receive erythropoietin: patients with end-stage renal disease on dialysis40 and patients with cancer receiving chemotherapy.41 Although the baseline health state utilities for dialysis and cancer patients are different from patients undergoing hepatitis C therapy, the effect of erythropoietin use in improving health state utilities (by increasing hemoglobin levels and decreasing symptoms) were extrapolated from these cost-effectiveness studies and were used to derive an appropriate utility estimate for anemic patients on hepatitis C combination therapy. The utilities of anemic cancer and dialysis patients were compared to their utilities after erythropoietin treatment and a multiplicative effect of treatment was assumed.
Utilities for the sequelae of chronic hepatitis C infection have been previously reported.23, 42, 43 These scores were updated using data recently published on utility scores obtained from patients with chronic hepatitis C who were in various stages of disease progression.39, 44, 45 Health state utility values for our model are listed in Table 3.
|Chronic hepatitis C||0.82 (0.60–0.90)||23|
|Combination therapy||0.90 (0.70–0.95)||42|
|Interferon monotherapy||0.93 (0.70–1.00)||33|
|Anemia during combination therapy||0.79 (0.50–0.99)||40,41|
|Compensated cirrhosis||0.78 (0.50–0.90)||23|
|Decompensated cirrhosis||0.65 (0.30–0.88)||42|
|Hepatocellular carcinoma||0.25 (0.10–0.80)||23,42|
|Liver transplantation (1st year)||0.50 (0.11–0.70)||23,39|
|Liver transplantation (after 1st year)||0.70 (0.24–0.87)||23|
Baseline drug costs were estimated using the 2005 average wholesale price.19 Because darbepoetin alpha appears to have similar efficacy and safety at far less cost than epoetin alpha, we used the cost of darbepoetin in our base case analysis. Nondrug costs associated with treatment, including laboratory studies (complete blood count, liver profile, HCV genotyping, and viral load) and office visits were estimated using the Medicare fee schedule. Because patients in our model had experienced a complication of combination therapy, namely anemia, these patients were expected to have undergone more intensive follow-up compared to patients who did not experience a complication of therapy. Therefore, based on expert opinion, the nondrug treatment costs for anemic patients were estimated to include a complete blood count, liver profile, and level III established office visit every 4 weeks until the end of therapy (48 weeks for genotype 1, 24 weeks for genotype 2 or 3). There were no additional costs associated with growth factor administration, because, like pegylated interferon, it is a self-administered subcutaneous injection. For all genotypes (1, 2, or 3), costs after therapy included a complete blood count and liver profile at 4, 12, and 24 weeks posttreatment, as well as an HCV viral load and an office visit at 24 weeks posttreatment.
The annual costs of care and lost productivity associated with different stages of disease have been reported previously.23, 42 All costs were adjusted to 2005 U.S. dollars (Table 4) using the Medical Care Services component of the consumer price index (http://www.bls.gov/cpi/home.htm).
|Nondrug costs during HCV therapy|
|Complete blood count||$16.46||Medicare fee schedule|
|Liver profile||$21.83||Medicare fee schedule|
|Level 3 outpatient visit||$47.31||Medicare fee schedule|
|HCV RNA test (by PCR)||$105.69||Medicare fee schedule|
|Weekly cost of therapy for:|
|Ribavirin, per 100 mg||$38.63||19|
|Epoetin alpha, per week||$596.50||19|
|Darbepoetin alpha, per week||$278.44||19|
|Annual cost of care for:|
|Chronic hepatitis C||$231.00||23|
|Transplantation (1st year)||$350,362.00||23|
|Transplantation (after 1st year)||$33,738.00||23|
All variables in the model were assigned a broad range of clinically plausible values. When available, 95% confidence intervals were used. Costs were doubled and halved. The model was reanalyzed for each parameter using the end points of the range, while holding all other variables at baseline (univariate sensitivity analysis). A variable was considered potentially influential if it could move the incremental cost-effectiveness ratio (ICER) across $50,000 per quality-adjusted life-year (QALY), implying a possible change in preferred strategy. Other cost-effectiveness analyses evaluating common interventions such as hemodialysis, coronary artery bypass surgery, pneumococcal vaccination, colorectal cancer screening and breast cancer screening, have also used this value as a cutoff, as have previous analyses of hepatitis C combination therapy.23, 33 Thus, $50,000/QALY is a useful reference point for framing the discussion.
To further test the robustness of our conclusions, we conducted a probabilistic sensitivity analysis. To do this, we created probability distributions for all of the parameters in the model. For cost estimates based on specific cost proxies, such as Medicare reimbursements or the average wholesale price for drugs, we used a normal distribution with 95% confidence interval equal to ±25% of the mean cost.46 For all other parameters we used the baseline value for the mean and estimated the standard error based on the approximation that the range used for the one-way sensitivity analysis represented a 95% confidence interval, with range approximately equal to 4 times the standard error.46 For probabilities and utilities we used a beta distribution, and for estimates of the annual cost of care for different health states we used a gamma distribution.46 We used Monte Carlo simulation to create 1,000 samples for which expected values were calculated. We then calculated the proportion of the time each strategy was cost-effective for varying limits of cost-effectiveness. Data analysis was performed using TreeAge Pro (TreeAge Software, Inc, Watertown, MA).
For HCV genotype 1–infected patients, the use of hematologic growth factors was associated with an SVR of 55.6%, compared to 46.1% for patients undergoing RBV dose reduction (Table 5). This increase in SVR led to 0.407 additional QALYs for the growth factor group compared to patients receiving dose reduction. Compared to a dose reduction strategy, the additional costs associated with growth factor use were approximately $14,100 for darbepoetin alpha and $24,600 for epoetin alpha. Therefore, the marginal cost-effectiveness, or ICER, of growth factor use in our model was $34,793 per additional QALY for darbepoetin and $60,600/QALY for epoetin.
|Strategy||SVR (%)||Cost||Effectiveness (QALY)||ICER (ΔC/ΔE)|
|HCV genotype 1|
|ΔC = $14,100||ΔE = 0.407||$34,793|
|HCV genotypes 2 and 3|
|ΔC = $7,000||ΔE = 0.206||$33,832|
For patients with genotype 2 or 3 infection, those undergoing growth factor injections obtained an SVR of 83.3% versus 78.6% for patients on a dose reduction strategy (Table 5). Growth factor use was associated with an additional 0.206 QALYs compared to patients undergoing RBV dose reduction. Compared to a dose reduction strategy, the additional costs associated with growth factor were approximately $7,000 for darbepoetin and $13,200 for epoetin. Therefore, the marginal cost-effectiveness of epoetin use for genotype 2 or 3 infection was $33,832 per additional QALY for darbepoetin and $64,311 for epoetin.
For all genotypes, patients undergoing growth factor administration remained on it for the duration of therapy, regardless of whether dose reduction was also necessary to improve their anemia. Given the significant costs associated with growth factor therapy, we modified our model so that patients who did not respond to growth factors after 4 weeks of administration would then undergo dose reduction, but growth factors would be discontinued for the duration of therapy. Therefore, nonresponders would only receive 4 weeks of drug administration, then begin dose reduction. This dosing modification did not significantly change the ICER for either genotype (data not shown), primarily because most patients responded to growth factor injections and, therefore, continued to receive them throughout their HCV therapy.
A conservative threshold of $50,000/QALY was used as an indication that a model variable could potentially be influential in determining the cost-effectiveness of epoetin alpha use.23, 33 For genotype 1 infection, only a few variables increased the ICER to >$50,000/QALY. The probability of attaining an SVR with RBV dosing >800 mg/d during therapy was estimated to be 57%. Decreasing this probability to <50% led to an ICER for growth factor use of >$50,000/QALY (Table 6). The probability of attaining an SVR with RBV dosing <800 mg/d was 36% in our model. Growth factors were no longer cost-effective if this rate increased to more than 47%. The baseline health state utility of chronic HCV was 0.82 in our model. A utility of 0.90 or higher for HCV led to an ICER of greater than $50,000 for growth factor use. Darbepoetin use remained cost-effective even if its price doubled. In contrast, epoetin use was not as cost-effective (ICER = $60,600/QALY). Treating genotype 1–infected patients older than 58 years of age led to ICERs above the threshold. Sensitivity analysis of the other model variables for genotype 1 infection had little effect on the ICER.
|Variable||Baseline Value||Threshold Value (@$50,000/QALY)|
|HCV genotype 1|
|SVR given RBV ≥800 mg/d||0.57||0.50|
|SVR given RBV <800 mg/d||0.36||0.47|
|Utility of chronic HCV||0.82||0.90|
|Cost of growth factors||$278.44||$400.00|
|Starting age for HCV therapy, years||45||58|
|HCV genotype 2|
|SVR given RBV ≥800 mg/d||0.84||0.72|
|Utility of chronic HCV||0.82||0.90|
|Cost of growth factors||$278.44||$450.00|
|Starting age for HCV therapy, years||45||59|
Sensitivity analysis for genotype 2 or 3 infection demonstrated results similar to that of genotype 1 infection (Table 6). The baseline estimate for attaining an SVR with RBV dosing >800 mg/d during therapy was 84% for genotypes 2 or 3. When this rate decreased to <72%, the use of growth factors had an ICER above the threshold. An increase in the health state utility of chronic HCV from 0.82 to 0.90 also led to an ICER >$50,000/QALY. As with genotype 1, darbepoetin use remained cost-effective even if its price doubled. Epoetin use was not as cost-effective for genotype 2 or 3 infection (ICER = $64,311/QALY). Treating patients with genotype 2 or 3 infection who were older than 59 years of age was not cost-effective using the $50,000/QALY standard. Sensitivity analysis of the other model variables for genotype 2 or 3 infection had little effect on the ICER.
Probabilistic Sensitivity Analysis.
We performed Monte Carlo simulation to create 1,000 samples for each genotype model to conduct probabilistic sensitivity analysis (Fig. 3). For all genotypes, the strategies were equally cost-effective when willingness to pay was approximately $40,000/QALY. Using a willingness-to-pay threshold of $50,000/QALY resulted in darbepoetin being cost-effective in approximately 60%–64% of the samples.
Although hematologic growth factors have been shown in prospective studies to benefit patients who experience anemia during HCV combination therapy, they are expensive; therefore, it is necessary to evaluate their costs in relation to their expected benefit. This study evaluated the cost-effectiveness of growth factors for anemia during HCV therapy compared to RBV dose reduction. In our analysis, darbepoetin was cost-effective, with an incremental cost-effectiveness ratio of $34,793/QALY for genotype 1 infection and $33,832/QALY for genotype 2 or 3 infection. Epoetin was not as cost-effective with an ICER of $60,600/QALY for genotype 1 infection and $64,311 for genotype 2 or 3 infection.
For both genotypes, sensitivity analysis demonstrated that relatively few variables significantly influenced the marginal cost-effectiveness of growth factor use. These variables include: (1) the SVR rate for patients on higher RBV levels; (2) the health state utility associated with chronic HCV infection; (3) the patient age at which HCV therapy is begun; and (4) the costs of hematologic growth factors.
We estimated an SVR of 57% for patients on standard RBV doses of >800 mg/d, with growth factors no longer being cost-effective if SVR rates were <50% for genotype 1 infection or <72% for genotype 2 or 3 infection. In a retrospective cohort study of 511 patients, McHutchison et al.9 reported an SVR of 63% for patients with higher RBV dosing (>10.6 mg/kg) who were adherent to therapy. In a randomized, prospective study of 1,311 patients, Hadziyannis et al.10 achieved SVR rates of 57% for genotype 1–infected patients without cirrhosis receiving combination therapy with pegylated interferon and standard RBV doses. In addition, patients with a low viral load experienced SVR rates of 65%. McHutchison et al. reported SVR rates of >90% for genotype 2 or 3 infection, and Hadziyannis et al. reported SVR rates of 83% in patients without cirrhosis. Although sensitivity analysis revealed that low SVR rates would affect our results, these studies demonstrate that it is very unlikely that the true SVR in these groups would be low enough to render the use of growth factors too expensive.
Most cost-effectiveness studies evaluating HCV therapy have estimated the rate of progression to cirrhosis to be approximately 5.4%-7.3% per year.23, 29, 33, 47–49 However, the rate of progression to cirrhosis for chronic HCV infection in patients at average risk appears to be approximately 15%–20% over 20 years.1, 34, 38 Therefore, the rate of cirrhosis in our study was estimated to be 1% per year, as assumed in a previous cost-effectiveness analysis of screening for hepatitis C.42 However, applying previous annual rates of progression would make our current ICER estimates for growth factor therapy much lower than the current values of $34,793/QALY for genotype 1 and $33,832/QALY for genotypes 2 or 3.
Regarding quality of life estimates, our model was sensitive only to changes in the utility for chronic hepatitis C, which had a baseline value of 0.82. Growth factor therapy was no longer cost-effective if the health states' utility for chronic HCV became greater than 0.90 for genotypes 1, 2, or 3. Most studies report a utility value of chronic HCV below 0.90, including studies with patient-elicited utilities; therefore, it is unlikely that the true utility value of chronic HCV is higher than 0.90.
This study has several limitations. First, our model assumes that patients successfully treated with hematologic growth factors continue on this therapy for the remainder of hepatitis C therapy. However, it is possible growth factors may be able to be withdrawn later in therapy without a significant reduction in SVR, making their use even more cost-effective. Second, because there are currently no published health state utilities for HCV-infected anemic patients on erythropoietin, utility values for the incremental effect of anemia were estimated using data from patients on dialysis and cancer patients receiving chemotherapy. However, anemia has several clinical manifestations similar to all patient groups regardless of etiology, such as shortness of breath and fatigue. Therefore, the application of these utilities to the HCV population has face validity. In addition, sensitivity analysis demonstrated that the utility value for anemia had no significant influence on the ICER for growth factors across a broad range of values, regardless of HCV genotype. A third limitation was that erythropoietin and RBV dose reduction strategies were developed based on algorithms followed in the literature13–15 or recommended by drug manufacturers.6 However, no standard algorithms are currently in place for hematologic growth factor administration or stepwise RBV dose reduction. Shiffman et al.50 recently demonstrated that RBV dose reductions occurring after week 20 of combination therapy had no significant impact on SVR. The patients in that study were previous nonresponders to HCV therapy, whereas our analysis is based on treatment-naïve patients. Regardless, if true, this finding could have important implications. Specifically, it is possible that the use of hematologic growth factors could be limited to only the first 20 weeks of therapy. However, given that our model suggests these growth factors are cost-effective even if used throughout HCV therapy, the ability to decrease or eliminate their use later in HCV therapy would only make them a more cost-effective option. In addition to changes in growth factor administration (such as those incorporating early stop rules for growth factors), there are multiple permutations for RBV dose reduction that could have been used in our model. Small changes in RBV dose reduction in our model would primarily affect the costs associated with therapy. Therefore, minor changes in RBV dose reduction strategies would lead to inconsequential differences in cost-effectiveness, because our model was not sensitive to the cost of RBV.
In conclusion, this cost-effectiveness analysis demonstrates that compared to RBV dose reduction, the use of hematologic growth factors, specifically darbepoetin, is cost-effective by current standards regardless of genotype. The results of our model should be confirmed by a large randomized control trial evaluating the effect of erythropoietic growth factors on HCV treatment response rates. Factors that were shown to influence the cost-effectiveness of therapy and that may also require further study to develop more accurate disease estimates include: (1) the SVR rate for patients on higher RBV levels; (2) the quality of life associated with chronic hepatitis; and (3) the age at which HCV therapy is begun. The costs of hematologic growth factors also significantly affect their cost-effectiveness, and a standard dosing schedule that provides high efficacy while limiting their use is needed, especially for epoetin alpha.
- 6Rebetol (ribavirin, USP) product information. Kenilworth, NJ: Schering Corporation, 2003. [Package insert.]
- 15Darbepoetin alpha (DA) for ribavirin-induced anemia in patients with chronic hepatitis C (CH-C) treated with pegylated interferon and ribavirin (PEG-IFN/RBV): a preliminary analysis [Abstract]. Gastroenterology 2004; 126( Suppl 1): 83A., , , , , , et al.
- 18Current status of the use of growth factors and other adjuvant medications in patients receiving peginterferon and ribavirin. Rev Gastroenterol Disord 2004; 4( Suppl 1): S39–S47., , .
- 19CohenHE, editor. Drug Topics. Red Book. Montvale, NJ: Thomson PDR, 2005.
- 20Cost-effectiveness in Health and Medicine. New York: Oxford University Press, 1996., , , .
- 21Meta-analysis, Decision Analysis and Cost-effectiveness Analysis. New York: Oxford University Press, 1994..
- 36Liver transplantation for chronic viral hepatitis. Viral Hepatitis Rev 1996; 2: 219–228., , .
- 39Health-state utilities and quality of life in hepatitis C patients. Am J Gastroenterol 2003; 98: 630–638., , , , , , et al.Direct Link:
- 43Assessment of utilities and health-related quality of life in patients with chronic liver disease. Am J Gastroenterol 2001; 96: 579–583., , , , .Direct Link: