Economic evaluation of early monotherapy versus delayed monotherapy or combination therapy in patients with acute hepatitis C in Germany

(1)

Aus dem Institut für Epidemiologie, Sozialmedizin und Gesundheitssystemforschung

der Medizinischen Hochschule Hannover

Titel der Abhandlung:

Economic evaluation of early monotherapy versus delayed monotherapy or combination therapy in patients with acute hepatitis C in Germany

Dissertation

zur Erlangung des Doktorgrades der Humanbiologie der Medizinischen Hochschule Hannover

vorgelegt von

Charalabos-Markos Dintsios aus Nürnberg

Hannover 2009

(2)

Angenommen vom Senat der Medizinischen Hochschule Hannover am 16.11.2010 Gedruckt mit Genehmigung der Medizinischen Hochschule Hannover

Präsident: Prof. Dr. Dieter Bitter-Suermann

Betreuer der Arbeit: PD Dr. rer. pol. Christian Krauth Referent: Prof. Dr. Matthias Schönermark

Korreferent: Prof. Dr. Hans-Heinrich Wedemeyer Korreferent: Prof. Dr. Uwe Siebert

Tag der mündlichen Prüfung: 16.11.2010

(3)

Inhaltsverzeichnis

Seite

1. Publikation... 1-11 2. Zusammenfassung……… 12-28

2.1 Einleitung………... 12-13 2.2 Zielsetzung………. 13-14 2.3 Methoden……… 14-18 2.4 Ergebnisse……….…. 18-20 2.5 Diskussion……….. 21-23 2.6 Literatur……….. 24-28 Lebenslauf……….. 29-30 Wissenschaftliche Veröffentlichungen………….. 31-36 Danksagung……… 37 Erklärung nach § 2 Abs. 2 Nrn. 7 und 8…………. 38

(4)

Economic evaluation of early monotherapy versus delayed monotherapy or combination therapy in patients with acute hepatitis C in Germany

Charalabos-Markos Dintsios

^a,e

, Alexander Haverkamp

^a

, Johannes Wiegand

^c

, Tilman Gerlach

^d

, Heiner Wedemeyer

^b

, Gerd Pape

^d

, Michael Peter Manns

^b

and Christian Krauth

^a

Background Antiviral treatment of acute hepatitis C virus (HCV) almost doubles the chance of sustained virological response (SVR) compared with that achievable by treating chronic HCV.

Aim To conduct a health economic evaluation comparing early and delayed therapies for acute HCV in Germany.

Methods One hundred and thirty-three patients with acute HCV were evaluated in two early monotherapy (EMT) studies and 60 in a delayed therapy study. Efficacy was determined by SVR. In the EMT studies, patients were treated with either standard or pegylated interferon for 24 weeks. In the delayed therapy study, patients with persisting infection were treated with interferon

monotherapy or combination therapy with ribavirin for a median of 36 weeks. We conducted a cost-effectiveness analysis based on the study results and a linear simulation model based on current treatment recommendations.

Results The SVR rate for the sex-adjusted on-treatment analysis between early and delayed therapies was not significantly different (92.7 vs. 90.9%;P= 0.7). Medication costs accounted for more than 90% in both treatment options. Direct medical costs of early therapy (h7064/patient) wereh321 lower than those of delayed therapy (P= 0.8).

The incremental cost-effectiveness ratio was – 178h/SVR(%)

(confidence interval: – 224 to 360h/SVR(%)). Average

modeled direct medical costs of delayed combination therapy were fromh6745 toh8299 per patient (from approximately 7% less up to 15% higher than EMT).

Spontaneous viral clearance and therapy duration were the most sensitive variables.

Conclusion There was no significant efficacy and cost difference between therapy alternatives in base cases. However, in the majority of scenarios in the sensitivity analyses, EMT was a more cost-effective option in acute HCV therapy. Eur J Gastroenterol Hepatol 00:000–000^c 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins.

European Journal of Gastroenterology & Hepatology2009,00:000–000 Keywords: cost-effectiveness, hepatitis C, interferon, ribavirin, simulation model Departments of^aEpidemiology, Social Medicine and Health System Research,

bGastroenterology, Hepatology and Endocrinology, Hanover Medical School, Hanover,^cMedical Clinic and Polyclinic II, University of Leipzig,^dMedical Department II, Klinikum GroXhadern, Institute for Immunology, Munich and^eDepartment of Heath Economics, Institute for Quality and Efficiency in Health Care, Cologne, Germany

Correspondence to Charalabos-Markos Dintsios MPharm, MA, MPH, Institute for Quality and Efficiency in Health Care, Department of Health Economics, Dillenburger Str. 27,

D-51105 Cologne, Germany

Tel: + 49 221 35685 261; fax: + 49 221 35685 858;

e-mail: Charalabos-Markos.Dintsios@iqwig.de

Received11 November 2008Accepted30 March 2009

Introduction

An estimated 170 million people worldwide are chroni- cally infected with the hepatitis C virus (HCV): 2.7 million in the USA, 7 million in Europe, and approximately 500 000 in Germany [1–5]. The incidence of newly reported, predominantly chronic HCV cases in Germany was 10.5 per 100 000 inhabitants (8635 cases) in 2001 and 8.0 per 100 000 inhabitants (6600 cases) in 2002. For Germany, HCV-related costs have not yet been fully determined, but are expected to reachh1 billion per year [6]. HCV-associated morbidity already poses a tremendous public health risk factor that will increase in the future, partly because of the low spontaneous viral clearance (SVC) in patients with chronic HCV. Progres-

sion to chronic infection occurs in the majority (55–85%) of infected people [7]. HCV infection is also a substantial risk factor for hepatocellular carcinoma [8] and is currently the main indication for liver transplantation in Germany [3,6].

Early antiviral treatment may lower the burden of HCV- related liver disease. It should take individual prognostic factors such as genotype-dependent treatment schemes into account. Early antiviral treatment almost doubles the chance of sustained virological response (SVR) compared with that achievable by treating chronic hepatitis C.

Therefore, efforts for early identification of newly infected patients should be undertaken [9].

0954-691Xc 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins DOI: 10.1097/MEG.0b013e32832c7b2e

(5)

The German healthcare system, like other healthcare systems, is under great economic pressure. Thus, health economic factors will become important considerations in future drug and technology approvals [10].

Patients who receive treatment for acute HCV are primarily patients who show symptoms. Asymptomatic patients are only diagnosed in rather exceptional cases (e.g. where an infection is suspected and confirmed by blood tests) [11].

This health economic evaluation compared early and delayed treatment options for acute HCV. Early monotherapy (EMT) was given to all affected patients, with either pegylated interferon (peg-IFN) or standard IFN. Delayed therapy involved either type of IFN, as monotherapy or in combination with ribavirin (RBV), only in patients who did not experience SVC. To account for current treatment strategies of combination therapy for acute HCV an adapted healthcare simulation model was additionally performed.

The evaluation was conducted with the support of the German Network of Competence on Viral Hepatitis, which develops nationwide standards for the treatment of viral hepatitis. The German Network of Competence on Viral Hepatitis also coordinates scientific studies and the transfer of knowledge between healthcare institutions [12].

Methods

Evaluated studies

The methods and results of the evaluated trials have been described in detail elsewhere [11–14]. There were two open-label, multicentre, EMT trials and one open- label, bicentre, delayed monotherapy or combination therapy trial. All trials were conducted in Germany.

The treatment approach in the early therapy studies was to initiate treatment without waiting to see whether the course of disease was self-limited. In contrast, in the delayed therapy study, the course of disease was evaluated before a decision to start therapy was made.

Efficacy in all trials was determined by the SVR rate, that is, the absence of detectable levels of serum HCV RNA at 24-week follow-up after the end of therapy.

The EMT I and II studies included 133 adult patients with acute HCV (44 and 89 patients in EMT I and II, respectively). Complete economic data were available for 128 patients. Patients were eligible if they had an acute HCV infection, were HCV RNA positive, and had elevated alanine aminotransferase levels.

The delayed therapy study included 60 adult patients with acute HCV [11]. Diagnosis was based on the

presence of HCV RNA, seroconversion to anti-HCV antibodies, and/or clinical and biochemical criteria.

Therapy schemes

In EMT I study (1998–2001), patients received 5 million units of standard IFN-a2b daily for 4 weeks, and then 3 times per week for another 20 weeks. Treatment was initiated 89 days (mean) after infection [13]. In EMT II study (2001–2004), patients received 1.5mg/kg peg-IFN-a2b weekly for 24 weeks. Treatment was initiated 76 days (median) after infection [14].

In the delayed therapy study (1993–2000), patients with persisting HCV RNA for more than 3 months after the onset of disease were administered antiviral treatment.

The different schemes applied reflect the varying pre- valent treatment approaches during the study period.

These schemes consisted of various dosages and administration frequencies of standard IFN-a or peg-IFN-a2a monotherapy or either type of IFN combined with RBV (Table 1). The doses of peg-IFN and RBV were adjusted to patient weight. Treatment was started 5.7 months (median) after the onset of symptoms. Patients were treated for 35.5 weeks (median).

Table 2 shows the baseline characteristics of patients.

Baseline characteristics between study populations were similar, except for sex distribution; the study population in the EMT II was predominantly male.

Health economic analysis method

We performed a cost-effectiveness analysis calculating the cost of the applied therapy schemes and expressing their effectiveness in SVR rates [15]. We presented the main results of the cost-effectiveness analysis as an incremental cost-effectiveness ratio by dividing the cost differences of direct medical costs by the outcome differences in SVR for the two treatment strategies.

Table 1 Delayed therapy schemes for acute hepatitis C N

(total = 26) Medication Dosage Frequency

Cumulative interferon dosages per patient (MU)

5 IFN 3 MU 3x/w 311/261/454/330/276

4 IFN 5 MU 3x/w 656/512/793/255

1 IFN 6 MU 3x/w 470

2 IFN 5/3 MU 3x/w 469/574

1 Peg-IFN 80mg qw NA

1 Peg-IFN 100mg qw NA

3 IFN/RBV 3 MU 3x/w 784/454/211

5 IFN/RBV 5 MU 3x/w 745/793/383/388/386

1 IFN, IFN/RBV 3 MU 3x/w 467

1 IFN, IFN/RBV 5 MU 3x/w 911

1 IFN/RBV,

Peg-IFN/RBV 5 MU, 100mg

qd, qw NA

1 Peg-IFN/RBV 80mg qw NA

IFN, interferon; MU, million units; NA, not available; Peg-IFN, pegylated interferon;

qd, every day; qw, once a week; RBV, ribavirin; 3x/w, three times a week.

2 European Journal of Gastroenterology & Hepatology 2009, Vol 00 No 00

(6)

Discounting, where the value of future costs and benefits are converted into their present value, was not applied because of the short treatment periods (up to 14 months).

Our main focus was on direct costs, which include physician fees as well as laboratory and medication costs. We aggregated the data from the EMT studies to facilitate comparison with the delayed therapy study, as IFN and peg-IFN were used in both the early and delayed treatment strategies.

The recommendations for HCV therapy have been modified several times since 1993 [11]. Various treatment schemes are therefore considered in the calculation of costs.

We estimated the current direct costs of delayed therapy by developing an adapted treatment linear simulation model (Box 1) considering the current therapy strategies [16] and depending on hepatitis C genotype, which is a predictor for SVC. Genotypes 1, 2, and 3 are frequently found in the western world. Genotype 4 is the most predominant genotype in North Africa [17,18]. The recommended treatment period for acute HCV infections is usually 24 weeks for all genotypes. Lower sustained response rates are achieved in patients with genotypes 1 and 4 [16]. Compared with genotype 1 patients, genotype 3 acute hepatitis patients show a high rate of SVC [19].

We present two basic scenarios. The first one assumes the rather high SVC rate of the delayed combination therapy study. The second scenario uses the weighted mean SVC rate of identified relevant and published studies.

The applied model allows a cost comparison between the therapeutic options investigated.

We conducted sensitivity analyses allowing for uncer- tainty by testing whether plausible changes in the values of the main variables affect the results of the analysis [15]. Furthermore, costs and clinical parameters such as efficacy, disease severity, or treatment regimens devia- ting from the study protocol may also have a substantial effect on study results. The variables used in a sensitivity analysis depend on the study design. We performed one or more sensitivity analyses for therapy duration, SVC, and genotype distribution. If the results of the study were not greatly affected over the range of the above- mentioned variables, then the study results were considered to be robust. As assigned patents will expire in the next few years and lower-priced generic RBV will enter the German market, we calculated additionally a ‘break-even’ price for RBV. If introduced, this price would result in equal medication costs for EMT and delayed combination therapy schemes.

Data sources

Prospectively documented data for the health economic analysis for resource use and efficacy results were collected from the case records of study participants.

Direct medical costs were calculated at 2002 market prices to account for the past year when standard IFN and peg-IFN were available in Germany at the same time. Laboratory costs and physician fees were based on the German physician fee scale for outpatient treatment. Medication costs were based on adjusted pharmacy prices, taking discounts for statutory health insurance funds into account [20]. The direct medical costs were thereafter valued at 2007 prices by adapting them for inflation of the respective time span. For the modeled

Table 2 Baseline characteristics of acute hepatitis C patients Variable Early I Early II

P value

Aggregate

early Delayed P value Number

of patients

44 84 128 60

Age 34.9 ± 10.5 37.9 ± 13.6 0.172 36.9 ± 12.7 37.1 ± 14.6 0.892

Sex (F/M) 25/19 28/56 0.024* 53/75 34/26 0.065

Genotype (1/4 vs.

2/3 vs. NA)

27/12/5 53/17/14 0.503 80/29/19 38/14/8 0.873

F, female; M, male; NA, not available.

*P< 0.05.

Box1: Linear equations for the simulation model

Equation Explanation

1) ^m (cm_i + cl_i + cd_i)

i=k i=k

c_i= Σm

Σ

2) ^m,n cm_i + Σcm_j

i , j=k,l i=k j=l

m

cm_{i , j}= Σ Σ

3)

4) cm₂₄= cm_m24+c_Rbv24

j=ltj

∗Σn

cm_j= cm₂₄+w_{Gt1, 4}∗cm_w

c_i: total costs cm_i: medication costs

cm_j: medication costs after 24th week cl_i: laboratory costs

cd_i: doctors costs

cm_w: weekly medication costs for combination therapy cm₂₄: medication costs for the first 24 weeks cm_m24: monotherapy medication costs at week 24 w_{Gt1, 4}: Genotype 1, 4 distribution

c_Rbv24: Ribavirin costs at week 24

t_j: duration of combination therapy after 24th week

n

(7)

cost calculation of the simulation model, 2007 prices were considered. Table 3 shows an overview of costs for physician fees, hepatitis serology, and medication prices for the different drugs prescribed.

Statistical analysis

Results are presented as rates with confidence intervals (CIs) or as means with standard deviation. Differences in efficacy and costs are presented with respective Pvalues and CIs.Pvalues less than 0.05 were considered significant. All P values were two tailed. For the comparison of baseline characteristics between the two compared therapeutic strategies, we used w² tests and Fisher’s exact tests for dichotomous variables (sex).

Contingency tables analysis with Pearson’s w² test and

Fisher’s exact test with Yates corrections were used for frequencies (genotype). We used Student’s t-tests for continuous variables (age). On account of statistically significant difference in sex distribution between the EMTs, we adjusted for sex. For the calculation of SVR differences, we used Fisher’s exact tests and for the cost differences, Student’s t-tests. Statistical analysis was performed with statistical software package SPSS 12.0 (SPSS Inc., 1989–2003; Chicago, Illinois, USA), sensitivity analysis with TreeAge Pro (1988–2003; Williamstown, Massachusetts, USA).

Results

Efficacy

The pooled EMT and the delayed therapy trials showed comparable treatment efficacy (92.7 vs. 90.9%;P= 0.744, Fisher’s exact test) adjusted for sex. The difference of 1.8% (CI: – 5.6 to 11.9%) in favor of EMTs was not statistically significant. The intention-to-treat analysis (82.8 vs. 74.1%; P= 0.221, Fisher’s exact test) also resulted in a nonsignificant difference of 8.7%

(CI: – 3.5 to 22.1%).

Early monotherapy study I

All patients but one were adherent to therapy. At the end of follow-up (24 weeks after end of treatment), 43 (98%; CI: 88–100%) patients had undetectable levels of HCV RNA.

Early monotherapy study II

A total of 65 (73%) patients were adherent to therapy and completed follow-up. In this subpopulation, the SVR rate at the end of follow-up (24 weeks after end of treatment) was 89% (CI: 79–96%). For the total study population it was 71% (CI: 61–80%).

Delayed therapy study

In six patients, antiviral therapy was initiated immediately after diagnosis. In the remaining 54 patients, the course of HCV was self-limited (loss of HCV RNA) in 24 (44%) patients. Of the 30 patients who developed chronic HCV, 10 did not receive treatment (mainly because of study discontinuations and therapy contra- indications). All the other 20 patients received IFN or peg-IFN monotherapy, or combination therapy with RBV (Table 1). The SVR rate was 80% (median follow-up time: 24 months). The overall SVR rate (self-limited and treatment-induced) was 91% (CI: 78–97%).

Direct costs

Although the average direct medical costs of pooled early therapy strategies per patient (h7064) were lower than the average direct medical costs of delayed therapy strategy per patient (h7385), this difference was not significant (Table 4). The per patient incremental cost effectiveness ratio, taking into account direct medical costs and sex-adjusted SVR for pooled early strategies

Table 3 Medication prices, physician fees, and hepatitis serology costs

Medication prices (2002)

Drug Package Charge 2002 (h)

Rebetol 200 mg capsule (ribavirin) 84 (N1) 652.85

168 (N2) 1214.54

IntronA 1 MU (IFNa-2b) 12 IB, 12 S (N2) 209.43 IntronA 18 MU (IFNa-2b) 2 IB, 12 S (N1) 624.92 12 IB, 72 S (N2) 2873.20^a 2 pens, 24 NE (N1) 624.92 8 pens, 96 NE (N2) 2065.59 IntronA 25 MU (IFNa-2b) 2 IB, 12 S (N1) 851.68 12 IB, 72 S (N2) 3863.33^a IntronA 30 MU (IFNa-2b) 2 pens, 24 NE (N2) 995.15

8 pens, 96 NE (N2) 3265.98 IntronA 60 MU (IFNa-2b) 2 pens, 24 NE (N1) 1774.59 8 pens, 96 NE (N2) 6248.86

PegIntron 50mg 4 PS, 4 NE 602.76

12 PS, 12 NE 1559.47

12 PS, 12 NE 2430.80

12 PS, 12 NE 2984.38

12 PS, 12 NE 3538.34

12 PS, 12 NE 4369.29

Office-based physicians’ care Amount (h) Basic charge per quarter per patient GP: Ins. 13.25; Pens. 23.75

Spec: Ins. 11.75; Pens. 15.00

Consultation charge per quarter 2.50

Intensive medical consultation

> 10 min in case of serious disease

15.00

Hepatitis serology Amount (h)

HBs-Ag 6.10

HCV-RNA quantitative 89.50

HCV-RNA qualitative 40.90

Anti-HAV 6.60

Anti-HIV 4.60

Anti-HCV 11.20

Genotyping 102.30

GP, general practitioner; HCV, hepatitis C virus; IB, injection bottle; IFN, interferon; Ins, insured; MU, million units; NE, needle; Pens, pensioner; PS, prefilled syringe; S, syringe; Spec, specialist.

a2001 prices in German Marks converted to Euros.

(8)

versus delayed therapy strategy, was – 178 h/SVR(%)

(CI: – 224 to 360 h/SVR(%)). This was also not statistically significant. In both the early and delayed studies, medication costs accounted for over 90% of total direct average medical costs, whereas laboratory costs were less than 9%. Physician fees accounted only for a marginal proportion of total costs (0.7% for early therapies and 1.6% for delayed therapy). Applying the 5% pharmacy discount to pharmacy prices for the year 2002 as an approximation for opportunity costs, medication costs amounted to h6208 ± 1029 and h6334 ± 7406 (h6535 ± 1083 and h6667 ± 7796 in 2007 prices, respectively) for the pooled EMTs and the delayed therapy (P= 0.9).

The direct comparison of the two early therapy studies showed significant but small cost differences (Table 5).

Following the opportunity cost approximation approach [20], medication costs yielded h6249 ± 0.05 versus h6760 ± 1412 for EMT I versus EMT II, respectively (P= 0.01).

The slightly higher average total costs of EMT II (h397) were because of the higher prices of peg-IFN compared with standard IFN and on the weight- dependent dosage of peg-IFN. However, adherence considerations support the once-weekly administration of peg-IFN, instead of standard IFN three times per week [21]. Two calculation examples are presented in Box 2 to show the difference in medication costs between different IFN monotherapies.

Modeling of direct costs for delayed combination therapy

Table 6 shows the modeled direct medical costs of delayed combination therapy considering current therapy recommendations for acute HCV [16]. Assuming a hepatitis C genotype distribution of genotype 1 and 4 of nearly 75% [22,23], and similar rates of symptomatic patients to that in the delayed combination therapy study [11], the costs of 24 weeks of delayed combination therapy currently amounts toh11 630 per patient treated.

Taking into account the rate of SVC, two scenarios were calculated for the average direct medical costs per acute HCV patient. In the first scenario, with the SVC rate of the delayed combination therapy study being 44% [11], the average direct medical costs per acute HCV patient amounted toh6745. In the second scenario using an SVC rate of 30% as a weighted mean derived from eight relevant studies [11,24–30], these costs reached h8299.

Table 5 Direct medical costs: early monotherapy I versus II Cost variable

Treatment

Total P= 0.04*

Medication

P= 0.01* Laboratory Physician Early I Mean ± SD (h) 6842 ± 116 6249 ± 0.05 534 54

Proportion (%) 100 91.4 7.8 0.8

Early II Mean ± SD (h) 7239 ± 1399 6760 ± 1412 428 44

Proportion (%) 100 93.5 5.9 0.6

*P< 0.05.

Table 4 Direct medical costs: early monotherapies versus delayed therapy

Cost variable

Treatment

Total P= 0.8

Medication

P= 0.9 Laboratory Physician Early (I + II) Mean ± SD (h) 7064 ± 1064 6535 ± 1083 475 48

Proportion (%) 100 92.5 6.7 0.7

Delayed Mean ± SD (h) 7385 ± 7941 6667 ± 7796 599 119

Proportion (%) 100 90.3 8.1 1.6

Box 2: Calculation examples for medication costs of interferon therapies

Monotherapy example with IFNα-2b: Dosage:

4 weeks: 5 MU/daily subcutaneous [28×5 10⁶U]

20 weeks: 5 MU/3×/weekly subcutaneous [60×5 10⁶U]

Sum: 88 dosages IFNα-2b Calculation based for the year 2001 (in GM), using the former injection bottles with 25 MU IFNα-2b.

A conservative approach for the approximation of the medical costs, assuming an IFN use of 90 dosages [1×12 injection bottles + 3×2 injection bottles (25/5=5 dosages per injection bottle)→ 90 dosages (60+30 dosages)] yields medication costs of 7556.01+3×1605.88= 12 373.65 GM

Monotherapy example with peg-IFNα-2b:

PegIntron (PEG-INFα-2b) week 1−24: 1.5µg/kg weight once weekly subcutaneously [24×dosages]

80 µg (<65 kg): 2002 ( ): 2×12 injection bottles =2×2430.80=4861.60 100 µg (65−75 kg): 2002 ( ): 2×12 injection bottles =2×2984.38=5968.76 120 µg (76−85 kg): 2002 ( ): 2×12 injection bottles =2×3538.34=7076.68 150 µg (>85 kg): 2002 ( ): 2×12 injection bottles =2×4369.29=8738.58

GM, German marks; MU, million units.

(9)

Thus, average direct medical costs per acute HCV patient of delayed combination therapy scheme are, depending on the SVC rate, approximately 7% less up to 15% higher than average direct medical costs per acute HCV patient of peg-IFN EMT (h7239) assuming an identical bodyweight distribution of patients in both therapy schemes.

In the delayed therapy study, patients with asymptomatic disease did not experience SVC. Immediate antiviral treatment for such patients has been proposed in the past [11,31]. Following this recommendation, the modeled medication costs would decrease from h6224 to h4868 according to current therapy recommendations of 24-week treatment, because the nine patients with asymptomatic hepatitis C would immediately receive IFN monotherapy instead of delayed combination therapy with IFN and RBV.

Sensitivity analysis

The most sensitive variables within the direct medical costs are the duration of the therapies and the SVC as shown in the tornado diagram (Fig. 1).

SVC rates of 16–50% have been documented in the literature [11,24–30,32]. The respective range of average direct costs per patient for a 24-week therapy duration lies between h6079 (SVC 50%) and h9854 (SVC 16%).

Therefore, even when patients show a relatively high viral clearance rate of up to 37%, delayed combination therapy still remains more expensive than EMT in most cases. Figure 2 shows the results of the two-way sensitivity analysis of medication costs for the variables SVC and therapy duration for the two treatment schemes.

The medication costs for delayed combination therapy are modeled following current therapy recommendations of 24-week therapy duration. Given that medication costs account for more than 90% in both therapy schemes, these seem to be the most important cost parameters.

Furthermore, medication usage is not as strongly dependent on the therapy protocol as laboratory and physician costs. Overall, it seems to be meaningful to restrict comparison to medication costs. Different duration of EMT with varying but at least comparable short- time efficacy results has been found among the spec- trum of acute HCV studies (Table 7 offers an overview).

Therefore, the duration was varied in the sensitivity

Fig. 1

1100 2500 3900 5300 6700 8100 9500 10 900 12 300 17 200

Duration (weeks): 4−24 Spontaneous viral clearance (%): 15−50 Combination therapy duration (weeks): 24−48

Tornado diagram of the most sensitive variables.

Table 6 Modeled direct medical costs according to current therapy recommendations Cost variable (h)

Treatment

Total Medication Laboratory Physician

SVC (%) 44 30 44 30 44 30 44 30

Delayed combination Mean 11 630 11 115 446 69

Proportion (%) 100 95.6 3.8 0.6

Self-limited course Mean 528 — 425 103

Proportion (%) 100 — 80.5 19.5

All patients Mean 6745 8299 6224 7781 437 440 84 79

Proportion (%) 100 92.3 93.8 6.5 5.3 1.2 0.9

SVC, spontaneous viral clearance.

(10)

analyses at a range of 4–24 weeks (Fig. 2). Depending on the SVC rate, combination therapy only starts to be favorable compared with the immediate monotherapy in case of a therapy duration over 19 weeks.

In some cases, combination therapy schemes of 48-week duration for the treatment of acute HCV are still applied [39], for example, for HIV co-infected patients. To bear this in mind, we calculated the direct medical costs per patient within a 16–50% SVC range for a 48-week delayed combination therapy scheme depending on genotype distribution. For this scenario, genotype 1 and 4 patients showing no self-limited course were treated for 48 weeks, genotype 2 and 3 patients for 24 weeks. The costs for a genotype 1 and 4 distribution of 75% amounted from h10 439 (SVC 50%) up toh17 179 (SVC 16%) per patient, respectively. In case of a therapy continuation beyond 24 weeks, the applied therapy duration range for the combination therapy scheme within the sensitivity analysis also contained 24–48 weeks. Figure 3 shows the results of a three-way sensitivity analysis for the prolonged combination therapy over 24 weeks in comparison with the immediate monotherapy for 24 weeks. The SVC and combination therapy duration over 24 weeks was varied in accordance with two genotype 1 and 4 extreme distributions of 100 and 50%. These extreme distribution values include the assumed distribution of 75% as being representative for Central Europe. Only genotype 1 and 4 patients are treated again in this sensitivity analysis beyond 24 weeks. After week 36, combination therapy is dominated by immediate monotherapy assuming a very high SVC rate of 50%.

We also conducted a sensitivity analysis for RBV price, taking the genotype distribution in HCV patients into account. RBV was approved in Germany in 1999 (Rebetol) and had a sales price of more thanh1000 for 168 capsules in 2002 (Fig. 4). Assuming an SVC rate of 44% as found in the

Fig. 2

Therapy duration (weeks)

Spontaneous viral clearance

4 9 14 19 24

0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15

Combination therapy Immediate monotherapy

areas indicate less expensive treatment Sensitivity analysis on duration of therapy and probability spontaneous viral clearance.

Table 7 Summary of trials with standard and peg-IFNa-2b or 2a

Study Type Therapy start (median) Treatment duration (weeks) Male (%) Genotype 1 (%) SVR (%)

Standard interferon

Vogelet al. [33] NRCT After establishment of diagnosis Until normalisation of ALT 33 42 74

Jaeckelet al. [13] NRCT 89 days after infection 24 43 61 98

Delwaideet al. [25] NRCT 110 days after infection 8 50 43 75

Nomuraet al. [27] RCT^a Early intervention (109 days) 4( + 20) 67 80 87–100

Late intervention 4( + 20) 60 87 40–53

Pegylated interferon

Kamalet al. [29] RCT^b 12 weeks observation 24 57 37 80–85

Santantonioet al. [34] NRCT 12 weeks observation 24 69 38 94

De Rosaet al. [9] NRCT 31 days after diagnosis 12 68 58 74

Kamalet al. [35] RCT^c 8 weeks observation 12 49 37 95

12 weeks observation 12 42 37 92

20 weeks observation 12 47 37 76

Kamalet al. [36] RCT^d 8–12 weeks observation 8 26 47 68

12 32 44 82

24 26 47 91

Wiegandet al. [14] NRCT 76 days after infection 24 63 66 71

Dominguezet al. [37] NRCT^e 14 weeks after diagnosis 24 100 29 71

Matthewset al. [38] NRCT^e 28 weeks before treatment 24 100 67 > 80

NRCT, Nonrandomised controlled trial; peg-IFN, pegylated interferon; RCT, randomised controlled trial; SVR, sustained virological response.

aExamines the SVR of a Japanese target population in dependence of waiting time before therapy start. For those who do not respond at week 4 the therapy is continued for another 20 weeks.

bOpen-label controlled trial with mono- versus combination therapy, but reduced external validity due to mixed target population.

cExamines the SVR in dependence of waiting time before monotherapy start.

dExamines the SVR in dependence of monotherapy duration.

eTarget population with HIV co-infection.

(11)

delayed therapy study [11] and an average body weight of 75 kg, the calculated ‘break-even’ price of an RBV package containing 168 capsules is h139.60 (referring to 24 weeks immediate monotherapy’s medication costs of h7057).

Prices for both RBV medications available in Germany (Rebetol and Copegus) currently lie substantially above this threshold, at nearlyh1000.

Discussion

The vast majority of patients with chronic hepatitis C are not seen during the acute phase. It may be that treatment of acute hepatitis C will have little impact on the overall cost of managing chronic hepatitis C-related liver diseases.

Study results, however, confirm that early treatment is associated with a better outcome compared with that achievable in chronic hepatitis. As therapy of acute HCV is worthwhile and there are more therapeutic options available to do this, a health economic evaluation to figure out the most efficient alternative seems to be reasonable.

Many health economic analyses have been conducted worldwide for different treatment options in chronic

HCV, among them six concerning patients in Germany [40–45]. To our knowledge, this analysis is the first health economic analysis considering alternative treatment options for acute HCV.

In most cases, IFN therapy in patients with acute HCV leads to SVR and an improvement in clinical and laboratory outcomes [46–48]. The success of IFN therapy seems to be associated with its initiation at an early stage of disease [27,49,50]. Patients with no SVR after immediate monotherapy may receive re-treatment with combination therapy, when diagnosed with chronic disease. This may be an important part of the immediate treatment strategy, but was beyond the scope of this study. This is a limitation of the study.

Only one controlled trial has been published comparing peg-IFN monotherapy versus combination therapy with peg-IFN and RBV for acute HCV. This study included 40 healthcare workers in Egypt and Germany, and, as in the studies we evaluated, also showed similar SVR rates between treatment groups [29]. This study is of limited external validity because of context reasons such as the profession of the patients and the location. Therefore, a cost comparison of immediate monotherapy versus delayed combination therapy including the selected studies and a performed linear treatment model based on current guideline recommendations seemed to be a suitable approach in our health economic evaluation.

Fig. 4

Ribavirin costs ( )

Medication costs

0 200 400 600 800 1000

11 200 10 600 10 000 9400 8800 8200 7600 7000 6400

Combination therapy Immediate monotherapy Threshold values:

Ribavirin package costs = 139.60 Expected value = 7057

Sensitivity analysis on ribavirin package costs. EV, expected value.

Fig. 3

Additional combination therapy duration over 24 weeks

Spontaneous viral clearance

24 30 36 42 48

0.50 0.45 0.40 0.35 0.30

0.25 0.20 0.15

Combination therapy (Genotype 1, 4 50 %)

Immediate monotherapy

areas indicate less expensive treatment Genotype 1, 4

Combination therapy (Genotype 1, 4 100 %)

Three-way sensitivity analysis on additional combination therapy duration over 24 weeks (clearance, genotype, duration).

(12)

In the EMT II study, differences in SVR rates between the total study population and the patients adherent to therapy were mainly because of patients who dis- continued therapy as a result of adverse events or who were lost to follow-up. Differences in SVR rates between the EMTs were also shown. This may have been caused by a higher proportion of patients from a low social background in EMT II study, as indicated by significant differences with regard to the source of infection (a higher rate of infection through intravenous drug abuse in EMT II: P= 0.04, w² test). This is an expected shortcoming when pooling data from different studies, even when patient inclusion and exclusion criteria of the studies are similar.

Two recent studies [25,51] on early peg-IFN monotherapy lasting 8 and 12 weeks achieved SVR rates of 75 and 74%, respectively. This indicates that a shorter duration of therapy (and a reduction of costs) may be associated with a reduction in efficacy. Furthermore, the efficacy derived from studies with a short time design is uncertain because there are no long-term follow-ups confirming the achieved SVR. In contrast, the long-term success (maintenance of SVR for up to 4 years after treatment) of EMT with a treatment duration of 24 weeks has been demonstrated earlier [52].

Regarding the cost results, the most important variable of medication costs and thereby of the direct medical costs is the SVC and the therapy duration. An extension of combination therapy for acute hepatitis C over 24 weeks, as is still performed in some cases, is only favorable in combination with very high SVC. As such high SVC rates have been documented rather seldom in the literature, the delayed combination therapy is dominated by the immediate monotherapy in most cases of prolonged duration. For very short therapy durations under 24 weeks, the domi- nance of immediate monotherapy shown in the sensitivity analysis is connected with uncertain outcomes and should not be relied on. To guarantee a comparable effectiveness, the duration should be 24 weeks for both therapeutic schemes, as recommended in the guidelines [16].

Indirect costs should also be taken into account. For example, therapy-related adverse effects may occur, leading to inability to work and lost earnings [14]. To estimate indirect costs, we collected data by means of a self-developed questionnaire including questions about employment, absence from work, and patients’ expenses related to HCV. This questionnaire was distributed only to participants (physicians and patients) of EMT II study, as this was the most recent study and an adequate response rate could be expected. The response rate to the EMT II questionnaire was 71%. Average absence from work of employed persons amounted to 23 days.

This equaled a profession-dependent loss of income of

approximatelyh2998 per employee, which is a noticeable potential cost factor.

Further research is needed to evaluate the incidence of adverse effects occurring with early therapy compared with delayed therapy. Both therapy schemes are claimed to induce fewer adverse effects than the other: early therapy because of the use of IFN alone [14,53], and delayed therapy because of the avoidance of treatment of patients with self-limited disease [11]. Adverse effects may substantially affect the quality of life of the individual patient, and with approximately the same cost structures of the therapy schemes, treatment-related quality of life may be a relevant decision-making factor [15].

Furthermore, we should consider that all three studies were conducted under clinical trial conditions. In the usual healthcare setting, conditions will differ and dropouts may reach higher levels. There is a substantial treatment mix in drugs, doses, and duration in the observed data. Although the authors have tried to partially adjust for these variables in the linear modeling, this heterogeneity is still a severe limitation while interpreting and comparing the empirical results from different studies. Other limitations of this evaluation include incomplete data, small study samples, and nonrandomized designs. In contrast, resource use data were collected prospectively. Only the data on absence from work were obtained retrospectively by survey of the patients and doctors who took part in EMT II study, so a complete estimate of indirect costs is not possible.

To consider the detrimental effects of therapy on quality of life, a cost-utility analysis is necessary. This is a special type of cost-effectiveness analysis in which the costs per unit of utility (units that relate to a person’s well-being) are calculated. Its results could lead to a change in preferred treatment options for acute HCV. For chronic HCV, quality of life studies have already been conducted [54–56].

In a previous very small and therefore not conclusive study, two thirds of 12 icteric patients with acute HCV infection cleared HCV [57]. Another study showed similar results [58]. This unexpectedly high spontaneous recovery rate was observed in East German women to whom HCV-contaminated immunoglobulin was administered in 1978–1979. This rate, however, is not considered to be representative because of specific patient characteristics. In this population, overall SVC (symptomatic and asymptomatic patients) reached 49%, which was within the range of our sensitivity analysis.

Finally, we did not conduct an analysis of the price sensitivity of IFN. IFN is not expected to be available as a generic agent in the near future, because of its