Research Article

Patient-important benefits of clearing the hepatitis C virus through treatment: A simulation model Hamish Innes1,2,⇑, David Goldberg1,2, Geoffrey Dusheiko3, Peter Hayes4, Peter R. Mills5, John F. Dillon6, Esther Aspinall1,2, Stephen T. Barclay7, Sharon J. Hutchinson1,2 1

School of Health and Life Sciences, Glasgow Caledonian University, Glasgow, UK; 2Health Protection Scotland, Glasgow, UK; 3UCL Institute of Liver and Digestive Disease, Royal Free Hospital, London, UK; 4Royal Infirmary Edinburgh, Edinburgh, UK; 5Gartnavel General Hospital, Glasgow, UK; 6Ninewells Hospital and Medical School, Dundee, UK; 7Glasgow Royal Infirmary, Glasgow, UK

Background & Aims: Given an appreciable risk of adverseeffects, current therapies for chronic hepatitis C virus (HCV) infection pose a dilemma to patients. We explored, via simulation modelling, patient-important benefits of attaining a sustained viral response (SVR). Methods: We created the HCV Individualised Treatment-decision model (the HIT-model) to simulate, on a per patient basis, the lifetime course of HCV-related liver disease according to two distinct scenarios: (i) SVR attained, and (ii) SVR not attained. Then, for each model subject, the course of liver disease under these alternative scenarios was compared. The benefit of SVR was considered in terms of two patient-important outcomes: (1) the percent-probability that SVR confers additional life-years, and (2) the percent-probability that SVR confers additional healthy lifeyears, where ‘‘healthy’’ refers to years spent in compensated disease states (i.e., the avoidance of liver failure). Results: The benefit of SVR varied strikingly. It was lowest for patients aged 60 years with initially mild fibrosis; 1.6% (95% CI: 0.8–2.7) and 2.9% (95% CI: 1.5–4.7) probability of gaining lifeyears and healthy life-years, respectively. Whereas it was highest for patients with initially compensated cirrhosis aged 30 years; 57.9% (95% CI: 46.0–69.0) and 67.1% (95% CI: 54.1–78.2) probability of gaining life-years and healthy life-years, respectively. Conclusions: For older patients with less advanced liver fibrosis, SVR is less likely to confer benefit when measured in terms of averting liver failure and premature death. These data have important implications. Foremost, it may inform the contempo-

Keywords: Hepatitis C; Chronic hepatitis C; Patient-centred; Patient-important outcomes; Markov model; Simulation model; Antiviral treatment; Adverse effects; Risk-benefit ratio. Received 7 August 2013; received in revised form 20 January 2014; accepted 27 January 2014; available online 5 February 2014 ⇑ Corresponding author. Address: Health Protection Scotland, Meridian Court, 5 Cadogan Street, Glasgow G2 6QE, UK. Tel.: +44 141 3001106. E-mail addresses: [email protected], [email protected] (H. Innes). Abbreviations: HCV, Hepatitis C virus; SVR, Sustained viral response; HIT model, Hepatitis-C Individual-based Treatment-decision model; HCC, Hepatocellular Carcinoma; NSS, Number need to attain SVR; NNT, Number needed to treat; SA, Sensitivity analysis.

rary patient dilemma of immediate treatment with existing therapies (that have poor adverse effect profiles) vs. awaiting future regimens that promise better tolerability. Ó 2014 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Introduction Chronic hepatitis C virus (HCV) infection leads to progressively worsening degrees of liver fibrosis (i.e., scarring of the liver). Eventually, over a period of decades, fibrosis can become so extensive, that liver function is compromised [1]. Often, it is only at this point of hepatic decompensation that HCV infection becomes evident to the patient. Hereafter, in the absence of a liver transplant, long-term prognosis is bleak [2]. It is this typical sequence of events that has led to HCV being dubbed the ‘‘silent killer’’ [3,4]. Chronic HCV infection is treatable, presently through a 16–48 week course of pegylated interferon, ribavirin ± a protease inhibitor [5]. Therapies are not however, a panacea; adverseeffects can be appreciable [6,7] and sometimes severe [8–10]. Moreover, patients failing first generation protease inhibitors can develop resistance to the class [11]. Accordingly, the decision to embark on a course of therapy must involve a judicious assessment of the risks and the benefits – particularly given that de facto, the majority with infection are unlikely to develop overt liver disease within the course of their lifetime [12,13]. The optimal outcome of treatment is permanent eradication of infection, as per the sustained viral response (SVR) surrogate [5]. With current treatment regimens, 67–75% of patients attain this outcome [14–16]; these ‘‘cure’’ rates are emphasised to treatment candidates [17]. However, the question of what value an SVR, in itself, imparts on the patient is often left unstated. Given that many treatment candidates are without symptoms of their infection, patients must be well informed regarding what, by way of prognosis improvement, SVR can offer. Yet, to-date, this issue has received little attention [18], despite being critical to informed consent. In this paper we simulate the lifetime course of liver disease according to SVR status and individualised patient factors. We

Journal of Hepatology 2014 vol. 60 j 1118–1126

JOURNAL OF HEPATOLOGY envisage that these data, in conjunction with information regarding: (i) the probability of SVR, (ii) the adverse-effect profile, and (iii) the prospect that more tolerable and efficacious therapies will be available in the future [19], will arm patients and clinicians alike, with a more complete picture when considering therapy with current and future treatment regimens.

Methods Model purpose To delineate the value of an SVR, we created the Hepatitis-C Individualised Treatment-decision model (HIT-model). The HIT-model simulates lifetime liver disease outcomes for individual model subjects, henceforth from two distinct scenarios: (1) SVR scenario: Treatment-induced viral clearance attained, i.e., the patient accepts a course of antiviral therapy and attains SVR. (2) No SVR scenario: Treatment-induced viral clearance not attained, i.e., the same patient alternatively either declines therapy or fails, hence their chronic infection persists over their remaining life course. For each model subject, we then compared the course of liver disease under the SVR scenario vs. the no SVR scenario.

From these two states, subjects can die a HCV-related liver death, according to a per annum risk of 43% and 13% from HCC and decompensated liver cirrhosis states, respectively [33]. Prognosis can be improved via a liver transplant, however, suitable donors are scarce (i.e., hence only a 2% probability of a liver transplant per year [27]). In the year transplantation occurs, the risk of a HCV-related liver death is 14.6%, but this falls to 4.4% for every year thereafter [27]. In the SVR scenario, we assumed subjects with mild/moderate fibrosis could not progress to more advanced states of liver disease. In contrast, subjects with initial cirrhosis (i.e., Ishak-6) could progress through to downstream states, but at a reduced rate relative to the no SVR scenario [34] (see Table 1). At every stage of the model, for both scenarios equally, subjects were susceptible to death from causes unrelated to either their past (i.e., past infection being applicable to the SVR scenario) or current (i.e., current infection being applicable to the no SVR scenario) HCV infection. Based on empirical data from HCV diagnosed individuals, we assumed these deaths occurred at an increased rate relative to the general population [35–39]. Finally, disease progression was simulated in annual cycles, until a maximum age of 90 years. Primary outcomes Our primary outcomes reflect events of importance to patients, as per the following definition set forth by Crowther et al.: outcomes ‘‘perceptible to the patient and of sufficient value that changing their frequency would be of value to the patient’’ [40]. Secondly our outcomes reflect the inceptive rationale for providing therapy; that being, to prevent HCV-related complications and death [41]. Finally our outcomes incorporate patients’ foremost concerns regarding ongoing chronic infection – premature death and disease progression [42]. On this basis, our outcomes were two-fold: (1) Gain of life-years: The percent-probability that SVR confers additional years of life. (2) Gain of healthy life-years: The percent-probability that SVR confers additional years of life in compensated health states (i.e., the avoidance of liver failure).

Model overview The HIT-model is a markov-chain model that simulates the lifetime course of HCV-related liver disease according to SVR status. The model was run as a firstorder Monte Carlo simulation approach – meaning that ‘‘patients’’ were processed through the model as individuals [20,21]. This is in contrast to prior HCV models, which typically consider ‘‘patients’’ at the aggregate (i.e., population) level [22–31]. The disadvantage of this first-order Monte Carlo approach is that it is computationally demanding. The advantage is that it allows one to model individual-level counterfactuals (i.e., in our case, prognosis in each ‘‘patient’’, with and without an SVR). The structure and parameterisation of our model is typical of existing HCV simulation models [22–31], and draws on a wide array of empirical research data; see Fig. 1 and Table 1. Subjects under the no SVR scenario face progressive hepatic fibrosis, culminating in compensated cirrhosis. Probabilities informing progression through these initial stages, were derived from a well-described cohort of chronic HCV patients, who attended a UK tertiary referral clinic and underwent an Ishak-scored [32] liver biopsy. These same fibrosis progression rates, which amount to 12% of persons developing compensated cirrhosis within 20 years of acquiring infection, have been adopted in previous HCV simulation models [22,24,25]. Upon, if at all, reaching Ishak-6 (i.e., compensated cirrhosis), subjects can develop decompensated liver cirrhosis and hepatocellular carcinoma (HCC).

The rationale for the second outcome is that in compensated disease states, liver function is, by and large, unimpaired. In other words, ipso facto, the liver is largely able to compensate for any damage incurred. Accordingly, patient preferences for mild, moderate, and compensated cirrhosis states do not differ [43,44]. We should make clear here, that in contrast to previous HCV models [22–31], our primary outcomes incorporate neither the efficacy (i.e., the SVR rate) nor the adverse effects associated with any particular treatment regimen. Instead, our outcomes reflect the distinct purpose of this analysis; to explore the probability a patient will benefit from an SVR, assuming it can be attained at all. Individualised patient factors Outcomes were stratified according to individualised patient factors; these being initial fibrosis stage (Ishak 0–2, 3–5 and 6) and age (30, 45 and 60 years). Thus in total, disease course was simulated separately for nine types of patient, each patient type being a distinct permutation of initial fibrosis stage and age categories.

Compensated disease states

Decompensated disease states

Death related to HCV

† Mild fibrosis (Ishak 0-2)

Moderate fibrosis (Ishak 3-5)

Decompensated cirrhosis

Compensated cirrhosis (Ishak 6)

Liver transplant

†

Posttransplant

Hepatocellular carcinoma (HCC)

Death unrelated to HCV

† Under scenario 1 (viral clearance) transition to these states is possible,

Non-absorbing state

Absorbing state

but occurs at a reduced rate, relative to scenario 2 Fig. 1. Chronic HCV natural history schematic assumed in the hepatitis C individual-based treatment-decision model.

Journal of Hepatology 2014 vol. 60 j 1118–1126

1119

Research Article Table 1. Parameters incorporated into the HIT-model.

Parameter [scenario(s) applicable to*]

Mean annual % probability

Data source(s)

Pre-decompensated health state transitions Mild-moderate fibrosis [2] Mild-moderate fibrosis [1] Moderate fibrosis-cirrhosis [2] Moderate fibrosis-cirrhosis [1] Compensated to decompensated cirrhosis [2] Compensated to decompensated cirrhosis [1]

2.5% 0.0% 3.7% 0.0% 3.9% 0.43%

Wright et al., [25]

Beta (38.06, 1484.38) 1.8-3.3

†

§

Wright et al., [25]

Beta (26.87, 699.30)

2.5-5.2

†

§

§

Compensated cirrhosis to liver cancer [2] Compensated cirrhosis to liver cancer [1]

1.4% 0.38%

Post-decompensated health state transitions Decompensated cirrhosis to liver cancer [1&2] Decompensated cirrhosis to liver related death [1&2] Decompensated cirrhosis/HCC to liver transplant [1&2] HCC to death [1&2] Liver transplant death in first year [1&2] Liver transplant survival post year 2+ [1&2] Non-HCV related mortality Age specific probability of non-HCV related mortality for UK general population [1&2]

Factor increase (relative to general population) assumed for HCV persons 30-39 yr [1&2] Factor increase (relative to general population) assumed for HCV persons 40-49 yr [1&2] Factor increase (relative to general population) assumed for HCV persons 50+ yr [1&2]

Sampling distribution

Fattovich et al., [33] Beta (14.58, 359.21) Fattovich et al., [33], and Beta (14.58, 359.21) Chou et al., [34]** and uniform (0, 0.21) Fattovich et al., [33] Beta (1.92, 135.12) Fattovich et al., [33], and Beta (1.92, 135.12) and Chou et al., [34]†† uniform (0.18, 0.46)

95% CI

§

2.2-6.1 0.0-1.0

0.5-2.8 0.1-1.0

1.4% 13.0% 2.0% 43.0% 14.6% 4.4%

Fattovich et al., [33] Fattovich et al., [33] Hutchinson et al., [27] Fattovich et al., [33] Hutchinson et al., [27] Hutchinson et al., [27]

Beta (5.35, 377.09) Beta (146.9, 983.1) Beta (10.18, 498.71) Beta (116.67, 154.66) Normal (14.6, 1.81) Normal (4.4, 0.46)

0.5-2.8 11.1-15.0 1.0-3.4 37.2-48.9 11.1-18.1 3.5-5.3

E.g, from 0.06% (aged 30-39 yr) to 7.8% (aged 80+ yr) 7.7§§

UK all-cause mortality life tables: years 2007-2009 [66]

§

§

McDonald et al., [38]

Normal (7.66, 0.60)

6.49-8.85

4.9§§

McDonald et al., [38]

Normal (4.90, 0.58)

3.8-6.0

1.7§§

McDonald et al., [38]

Normal (1.67, 0.18)

1.3-2.0

⁄

1 refers to the ‘‘SVR scenario’’; 2 refers to the ‘‘no SVR scenario’’. Assumption, in absence of robust empirical data. § Parameter not varied. ⁄⁄ Systematic review by Chou et al. [34] describes a 79–100% factor reduction in the annual risk of decompensated cirrhosis post SVR.    Systematic review by Chou et al. [34] describes a 54–82% factor reduction in the annual risk of liver cancer with SVR. §§ Modified to 1 for SA-4. CI, credible interval.  

Expressing outcomes through probability distributions Regarding gain of life-years, we assembled the probability distribution for total life years gained in all non-absorbing health states through attaining SVR, vs. the counterfactual scenario of no SVR. From the resulting distribution, we draw particular attention to the percent-probability of gaining >0 additional lifeyear(s). Similarly with regard to gain of healthy life-years, we compiled the probability distribution for life years gained in compensated health states, and emphasise the percent-probability of gaining >0 additional year(s). Of note, as the model was run in annual cycles, gaining >0 years is equivalent to gaining P1 year. Expressing outcomes through the number-needed-to-SVR We determined the number needed to attain SVR in order to prevent one patient dying a premature HCV-related death (NNS1); calculated as the reciprocal of the probability of gaining >0 additional life-years. Equally, we calculated the number needed to attain SVR in order to avert one patient from prematurely developing overt liver disease (NNS2); calculated as the reciprocal of the probability of gaining >0 healthy life-years. The NNS is a HCV-bespoke version of the number needed to treat (NNT). The NNT, referring to the number of patients that need to be treated in order to avert one additional adverse outcome, is well-regarded among clinicians as a meaningful measure of the effectiveness of a medical intervention [45,46]. Thus, the rationale for additionally framing the benefit of SVR in an NNS format, is to assist interpretation by clinicians and their patients – the target audience of this paper.

1120

Uncertainty due to sampling error To quantify the overall uncertainty in our outcomes attributable to sampling error, we assigned probability distributions to model parameters and sampled randomly from these; these distributions are shown in Table 1. As is recommended for microsimulation models [21], we adopted a stratified sampling approach (known as Latin hypercube sampling) which enables a more efficient coverage of the sampling space than non-stratified sampling alone, and so permits a less onerous computational intensity. For each set of random parameter draws, of which we performed 1000 in total, we ran each type of patient through the model 10,000 times. Thus, in total, 10 million simulations were performed for each patient type. We present a 95% credible interval with each outcome to indicate uncertainty. This interval refers to the 2.5th and 97.5th percentile range in each outcome, across the 1000 parameter draws.

Sensitivity analyses We performed five one-way sensitivity analyses (SA) to assess whether our findings withstand contentious issues (Supplementary Table 1 outlines the supplementary parameters used for these sensitivity analyses). These five SAs were as follows:

Journal of Hepatology 2014 vol. 60 j 1118–1126

JOURNAL OF HEPATOLOGY (1) SA-1: We increased our base case fibrosis progression parameters, pro rata, to reflect 16% progression to cirrhosis at 20 years, as per alternative progression data on liver clinic attendees [47]. (2) SA-2: We decreased our base case fibrosis parameters, pro rata, to reflect 7% progression to cirrhosis at 20 years. The lower 7% progression rate [48] is appropriate when modelling the general infected population [49]. Models evaluating birth cohort screening interventions [50] are a good example of when these lower progression rates should be considered. (3) SA-3: We assumed that the risk of fibrosis progression increases with age; by a factor of 1.25 when patients are aged 40–59 yrs (vs. 0 years (95% CI) 14.5 (13.5-15.6) 10.0 (9.1-10.8) 6.2 (5.4-7.1) 19.3 (17.7-20.9) 12.9 (11.9-14.0) 7.8 (7.1-8.4) 23.7 (21.8-25.8) 16.4 (15.2-17.8) 9.8 (9.0-10.7)

% Probability of gaining Mean years gained, >0 years (95% CI) among persons gaining >0 years (95% CI) 17.9 (11.5-25.1) 16.0 (14.9-17.2) 9.4 (5.4-14.1) 11.1 (10.2-12.0) 2.9 (1.5-4.7) 7.0 (6.3-7.7) 48.1 (34.4-60.7) 21.6 (19.8-23.4) 34.5 (22.9-46.6) 14.7 (13.5-15.9) 17.1 (10.4-24.9) 8.9 (8.2-9.6) 67.1 (54.1-78.2) 25.7 (23.2-28.7) 60.6 (46.1-73.2) 18.5 (16.8-20.3) 45.4 (32.0-58.1) 11.3 (10.4-12.4)

CI, credible interval.

Table 3. Number needed to attain sustained viral response in order, on average: (i) to avert one patient from a dying a HCV-related death (NNS1), and (ii) for one patient, to gain >0 healthy life-year(s) (NNS2), according to initial fibrosis stage and age.

Initial fibrosis stage Mild (Ishak 0-2)

Moderate (Ishak 3-5)

Compensated cirrhosis (Ishak 6)

Initial age (yr)

NNS1 (95% CI)

NNS2 (95% CI)

30

7.4 (5.1-11.7)

5.6 (4.0-8.7)

45

16.1 (10.4-28.4)

10.8 (7.1-18.5)

60

64.9 (36.8-125.0)

35.0 (21.1-64.5)

30

2.5 (1.9-3.6)

2.1 (1.6-2.9)

45

3.9 (2.8-6.0)

2.9 (2.1-4.3)

60

9.5 (6.1-16.2)

5.9 (4.0-9.6)

30

1.7 (1.4-2.2)

1.5 (1.3-1.8)

45

2.0 (1.6-2.7)

1.6 (1.4-2.2)

60

3.2 (2.4-4.6)

2.2 (1.7-3.1)

CI, credible interval.

will be aged 45 years+ with mild fibrosis, or aged 60 years+ with moderate fibrosis [55]; in other words, patient-groups that gain less from an SVR (assuming SVR is attained at all). Thus, on one hand, at the population level, birth-cohort screening will indeed avert a considerable number of end-stage liver disease cases over the years to come. All the same, from the perspective of any one individual patient, the probability of benefit may not always be sufficient to offset the adverse-effects of therapy. Ultimately, it is for the patient and their clinician, not the authors of this paper, to decide what is, and is not, an attractive risk-benefit ratio. However, it is for the medical-researcher to arm them with an objective assimilation of the data. Accordingly, the data presented herein, hold real value – particularly, in today’s landmark era of rapidly improving treatment regimens. Over the next five years, new therapies will emerge that are far more tolerable and efficacious than the treatment regimens of today. Or in other words, therapies in the near-future will offer the patient a far more attractive risk-benefit ratio. In this context, we envisage that the data herein, will help shape the individual patient view, vis-a-vis how urgently SVR attainment should be pursued, and at what cost to the patient’s immediate quality-of-life. Hitherto HCV simulation models are pitched at policy makers; insofar as they focus on broad population-level outcomes,

1122

far-removed from the individual-level decision to treat [22–31]. Our focus on addressing the questions that matter most to patients, in an intuitive way (i.e., in effect advocating a new patient-centred modelling approach), is the greatest strength of this paper. Various impediments have hindered primary research studies, in their own right, from addressing these issues. In particular, (i) initial infection is asymptomatic, thus inception cohorts are rare, (ii) liver damage thereafter occurs gradually, over a course of decades, rendering follow-up difficult, (iii) antiviral therapy post-diagnosis complicates the picture going forward, and (iv) up until recently, reliable fibrosis staging necessitated a liver biopsy, which is an invasive procedure and thus performed sparingly [13]. The simulation approach adopted here, draws on primary research data to generate the most plausible inference to the question-at-hand. However, results will rest on certain assumptions. In particular, as per hitherto HCV models [22,24,25,27], and as is consistent with observational data [56], the HIT-model assumes a linear course of fibrosis progression over the lifetime of the patient. There is uncertainty in this assumption. Fibrotic accretion may accelerate with time, particularly after 30 years of infection (a follow-up duration few observational studies exceed). In this eventuality our findings may be affected. These uncertainties should be made clear. Nevertheless,

Journal of Hepatology 2014 vol. 60 j 1118–1126

Table 4. Percent-probability that SVR confers additional life-year(s), according to base-case scenario and sensitivity analyses (SA) 1–5.

Cohort Initial fibrosis stage

Mild (Ishak 0-2)

Moderate (Ishak 3-5)

Base-case scenario (95% CI)

SA 1-16% cirrhosis progression at 20 yr (95% CI)

SA 2-7% cirrhosis progression at 20 yr (95% CI)

SA 3-accelerating fibrosis with age (95% CI)

SA 4-no excess mortality assumed (95% CI)

SA 5-increased excess mortality assumed (95% CI)

30

13.6 (8.5-19.4)

17.4 (11.5-24.1)

8.4 (4.3-13.9)

17.5 (11.1-24.4)

18.8 (12.0-26.2)

12.6 (7.8-17.9)

45

6.3 (3.5-9.6)

8.4 (5.0-12.4)

3.7 (1.7-6.6)

9.7 (5.7-14.5)

8.8 (5.1-13.0)

n.a.

60

1.6 (0.8-2.7)

2.2 (1.1-3.5)

0.9 (0.4-1.8)

3.1 (1.6-5.1)

2.5 (1.4-4.0)

n.a.

30

40.0 (27.9-51.7)

44.1 (32.4-55.2)

32.6 (20.1-45.7)

43.2 (31.2-55.3)

50.4 (36.2-63.7)

37.0 (26.1-48.1)

45

26.0 (16.7-36.1)

29.3 (19.6-39.4)

20.5 (11.6-30.8)

30.7 (20.3-41.7)

32.5 (21.5-44.0)

n.a.

60

10.6 (6.2-16.2)

12.3 (7.1-18.2)

8.1 (4.1-13.3)

14.4 (8.5-21.3)

14.8 (9.0-21.7)

n.a.

30

57.9 (46.0-69.0)

n.a.

n.a.

n.a.

66.7 (54.2-78.2)

54.5 (42.9-65.4)

45

49.3 (36.7-60.9)

n.a.

n.a.

n.a.

55.6 (42.5-67.8)

n.a.

60

31.6 (21.5-42.1)

n.a.

n.a.

n.a.

38.1 (27.1-49.3)

n.a.

CI, credible interval; n.a., not applicable.

Table 5. Percent-probability SVR confers additional healthy life-years, according to base-case scenario and sensitivity analyses (SA) 1–5.

Cohort Initial fibrosis stage Mild (Ishak 0-2)

Moderate (Ishak 3-5)

Compensated cirrhosis (Ishak 6)

Initial age (yr) 30

Base-case scenario (95% CI)

SA 1-16% cirrhosis progression at 20 yr (95% CI)

SA 2-7% cirrhosis progression at 20 yr (95% CI)

SA 3-accelerating fibrosis with age (95% CI)

SA 4-no excess mortality assumed (95% CI)

SA 5-increased excess mortality assumed (95% CI)

17.9 (11.5-25.1)

22.7 (15.4-30.5)

11.2 (6.0-18.0)

23.2 (15.1-31.6)

24.1 (15.9-33.1)

16.6 (10.5-23.0)

45

9.4 (5.4-14.1)

12.3 (7.6-17.7)

5.6 (2.7-9.6)

14.4 (8.7-21.0)

12.5 (7.6-18.1)

n.a.

60

2.9 (1.5-4.7)

4.0 (2.2-6.2)

1.7 (0.7-3.2)

5.6 (3.0-8.7)

4.4 (2.5-6.8)

n.a.

30

48.1 (34.4-60.7)

52.6 (39.8-64.3)

39.6 (24.9-54.3)

52.1 (38.5-64.7)

58.9 (43.5-72.3)

44.7 (32.1-56.5)

45

34.5 (22.9-46.6)

38.6 (27.0-50.4)

27.4 (16.0-40.3)

40.6 (28.0-53.5)

41.5 (28.6-54.4)

n.a.

60

17.1 (10.4-24.9)

19.6 (12.4-27.6)

13.2 (6.9-21.3)

22.8 (14.2-32.5)

22.3 (13.9-31.4)

n.a.

30

67.1 (54.1-78.2)

n.a.

n.a.

n.a.

74.5 (61.3-85.5)

64.0 (50.9-75.0)

45

60.6 (46.1-73.2)

n.a.

n.a.

n.a.

66.2 (51.8-78.5)

n.a.

60

45.4 (32.0-58.1)

n.a.

n.a.

n.a.

51.5 (37.4-64.9)

n.a.

CI, credible interval; n.a., not applicable.

1123

JOURNAL OF HEPATOLOGY

Journal of Hepatology 2014 vol. 60 j 1118–1126

Compensated cirrhosis (Ishak 6)

Initial age (yr)

Research Article it is reassuring that our conclusions withstand plausible acceleration assumptions made in SA-3. Secondly despite stratifying by age and disease-stage, patients and clinicians may argue that the HIT-model still remains too broad-brush. In particular, alcohol accelerates fibrosis progression [57] but we do not model its effect explicitly. So, because the same transition probabilities are applied to all patient types equally, then tacitly, the same level of alcohol use is also assumed. This perhaps is not a realistic assumption. One might argue that ongoing alcohol use will be more prevalent for patients who have developed cirrhosis by the age of 30, than for the 60-year-old patient who has mild fibrosis. Clearly this is a limitation, but one that underscores an important point: that whilst some population-level models do adjust for alcohol [27,29,30], there is a general lack of primary data characterising liver disease progression according to lifetime alcohol history. Research redressing this shortfall, would improve modelling estimates. Thirdly, the HIT-model frames the benefit of SVR in terms of averting liver failure and premature death; sequelae that concern HCV patients most [42]. However, these outcomes may not capture the totality of HCV-induced adversities, and as such may understate the benefits of therapy. In particular: (a) Compensated cirrhosis is frequently asymptomatic, but not always [58]; some patients present with non-specific symptoms (dyspepsia, asthenia, upper abdominal discomfort, etc.). The precise symptomatic-asymptomatic ratio is undetermined. Fattovich et al. report that among patients referred to tertiary clinics following diagnosis of compensated cirrhosis, 43% demonstrated such symptoms, and 57% did not – although N.B., a referral bias in favour of symptomatic patients is clearly likely to operate in this type of cohort [33]. (b) HCV infection can predispose one to various non-liverrelated conditions (cryoglobulinaemia, lichen planus, vitiligo, porphyria cutanea tarda and B-cell lymphomas [59,60]) which are not considered in our model. However, barring cryoglobulinemia, these conditions are relatively infrequent. (c) Our simulation model does not include the possibility that SVR benefits the patient vis-à-vis non-liver-related mortality. An association between chronic HCV infection per se and increased non-liver mortality is equivocal and not supported by many key studies [61–64]. Ongoing scrutiny however, particularly with regard to vascular disease [65], is required. There are important implications to these data. In the shortterm, this work will inform the contemporary patient dilemma of immediate treatment with existing interferon-centred therapies, vs. holding out for future regimens that promise higher SVR rates and improved tolerability. Longer-term, it urges that broadened access to therapy be twinned with efforts to more objectively communicate, to the individual patient, the benefits of therapy.

Financial support This work was supported through funding from the Scottish Government. The Scottish Government had no influence or involvement in any aspect of this study. 1124

Conflict of interest H Innes: None; D Goldberg: None; G Dusheiko: has received research and grant support from, has acted as a clinical investigator or chief investigator and has served on advisory boards for Gilead Sciences, Vertex, Abbott, Boehringer Ingelheim, Bristol Myers, Glaxo Smith Kline, Pharmasett, Merck, Janssen; P Hayes: None; P Mills: None; J Dillon: has received research and grant support from, has acted as a clinical investigator, has received speaker honoraria or has served on advisory boards for Gilead Sciences, Abbott, Boehringer Ingelheim, Bristol Myers, Glaxo Smith Kline, Merck, Janssen; E Aspinall: None; S Barclay: Has received travel grant support from, has acted as a clinical investigator, received speaker honoraria or has served on advisory boards for Gilead Sciences, MSD, Janssen and Roche,; S Hutchinson: has received speaker honoraria from AbbVie, Gilead Sciences, Janssen, MSD and Roche. Acknowledgement This work was supported through funding from the Scottish Government. The Scottish Government had no influence or involvement in any aspect of this study.

Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2014. 01.020.

References [1] Marcellin P, Asselah T, Boyer N. Fibrosis and disease progression in hepatitis C. Hepatology 2002;36:S47–S56. [2] D’Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognostic indicators of survival in cirrhosis: a systematic review of 118 studies. J Hepatol 2006;44:217–231. [3] Edlin BR. Perspective: test and treat this silent killer. Nature 2011;474: S18–S19. [4] Detonating a viral time bomb-the hepatitis C pandemic. Lancet 2013;381: 178. [5] EASL Clinical Practice Guidelines. Management of hepatitis C virus infection. J Hepatol 2011;55:245–264. [6] Manos MM, Ho CK, Murphy RC, Shvachko VA. Physical, social, and psychological consequences of treatment for hepatitis C: a communitybased evaluation of patient-reported outcomes. Patient 2013;6:23–34. [7] Fried MW. Side effects of hepatitis c and their management. Hepatology 2002;36:S237–S244. [8] US Food and Drug Administration. Drug safety communication. http:// www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm332860.htm. Accessed January 2013. [9] Hezode C, Fontaine H, Dorival C, Larrey D, Zoulim F, Canva V, et al. Triple therapy in treatment-experienced patients with HCV-cirrhosis in a multicentre cohort of the french early access programme (anrs c020-cupic)nct01514890. J Hepatol 2013;59:434–441. [10] Maasoumy B, Port K, Markova AA, Serrano BC, Rogalska-Taranta M, Sollik L, et al. Eligibility and safety of triple therapy for hepatitis C: lessons learned from the first experience in a real world setting. PLoS One 2013;8:e55285. [11] Wyles DL. Antiviral resistance and the future landscape of hepatitis C virus infection therapy. J Infect Dis 2013;207:S33–S39. [12] Alter HJ, Seeff LB. Recovery, persistence, and sequelae in hepatitis C virus infection: a perspective on long-term outcome. Semin Liver Dis 2000;20: 17–35. [13] Seeff LB. Natural history of chronic hepatitis C. Hepatology 2002;36: S35–S46.

Journal of Hepatology 2014 vol. 60 j 1118–1126

JOURNAL OF HEPATOLOGY [14] Poordad F, McCone Jr J, Bacon BR, Bruno S, Manns M, Sulkowski MS, et al. Boceprevir for untreated chronic HCV genotype 1 infection. N Engl J Med 2011;364:1195–1206. [15] Jacobson IM, McHutchison JG, Dusheiko G, Di Bisceglie AM, Reddy KR, Bzowej NH, et al. Telaprevir for previously untreated chronic hepatitis C virus infection. N Engl J Med 2011;364:2405–2416. [16] Innes HA, Hutchinson SJ, Allen S, Bhattacharyya D, Bramley P, Delahooke TE, et al. Ranking predictors of a sustained viral response for patients with chronic hepatitis C treated with pegylated interferon and ribavirin in Scotland. Eur J Gastroenterol Hepatol 2012;24:646–655. [17] Brau N. Evaluation of the hepatitis C virus-infected patient: the initial encounter. Clin Infect Dis 2013;56:853–860. [18] Trembling P, Tanwar S, Rosenberg W, Dusheiko G. Treatment decisions and contemporary vs. pending treatments for hepatitis C. Nat Rev Gastroenterol Hepatol 2013;10:713–728. [19] Dore GJ. The changing therapeutic landscape for hepatitis C. Med J Aust 2012;196:629–632. [20] Wever EM, Hugosson J, Heijnsdijk EA, Bangma CH, Draisma G, de Koning HJ. To be screened or not to be screened? Modeling the consequences of PSA screening for the individual. Br J Cancer 2012;107:778–784. [21] Sharif B, Kopec J, Wong H, Fines P, Sayre E, Liu R, et al. Uncertainty analysis in population-based disease microsimulation models. Epidemiol Res Int 2012;2012:14. http://dx.doi.org/10.1155/2012/610405. [22] Martin NK, Vickerman P, Miners A, Foster GR, Hutchinson SJ, Goldberg DJ, et al. Cost-effectiveness of hepatitis C virus antiviral treatment for injection drug user populations. Hepatology 2012;55:49–57. [23] Rein DB, Smith BD, Wittenborn JS, Lesesne SB, Wagner LD, Roblin DW, et al. The cost-effectiveness of birth-cohort screening for hepatitis C antibody in U.S. primary care settings. Ann Intern Med 2012;156:263–270. [24] Grieve R, Roberts J, Wright M, Sweeting M, DeAngelis D, Rosenberg W, et al. Cost effectiveness of interferon alpha or peginterferon alpha with ribavirin for histologically mild chronic hepatitis C. Gut 2006;55: 1332–1338. [25] Wright M, Grieve R, Roberts J, Main J, Thomas HC. Health benefits of antiviral therapy for mild chronic hepatitis C: randomised controlled trial and economic evaluation. Health Technol Assess 2006;10:1–113. [26] Davis GL, Alter MJ, El-Serag H, Poynard T, Jennings LW. Aging of hepatitis C virus (HCV)-infected persons in the United States: a multiple cohort model of HCV prevalence and disease progression. Gastroenterology 2010;138: 513–521. [27] Hutchinson SJ, Bird SM, Goldberg DJ. Modeling the current and future disease burden of hepatitis C among injection drug users in Scotland. Hepatology 2005;42:711–723. [28] Rein DB, Wittenborn JS, Weinbaum CM, Sabin M, Smith BD, Lesesne SB. Forecasting the morbidity and mortality associated with prevalent cases of pre-cirrhotic chronic hepatitis C in the United States. Dig Liver Dis 2011;43:66–72. [29] Deuffic-Burban S, Deltenre P, Louvet A, Canva V, Dharancy S, Hollebecque A, et al. Impact of viral eradication on mortality related to hepatitis C: a modeling approach in France. J Hepatol 2008;49:175–183. [30] Deuffic-Burban S, Deltenre P, Buti M, Stroffolini T, Parkes J, Muhlberger N, et al. Predicted effects of treatment for HCV infection vary among European countries. Gastroenterology 2012;143:974–985. [31] Deuffic-Burban S, Babany G, Lonjon-Domanec I, Deltenre P, Canva-Delcambre V, Dharancy S, et al. Impact of pegylated interferon and ribavirin on morbidity and mortality in patients with chronic hepatitis C and normal aminotransferases in France. Hepatology 2009;50:1351–1359. [32] Ishak K, Baptista A, Bianchi L, Callea F, De Groote J, Gudat F, et al. Histological grading and staging of chronic hepatits. J Hepatol 1995;22: 696–699. [33] Fattovich G, Giustina G, Degos F, Tremolada F, Diodati G, Almasio P, et al. Morbidity and mortality in compensated cirrhosis type C: a retrospective follow-up study of 384 patients. Gastroenterology 1997;112: 463–472. [34] Chou R, Hartung D, Rahman B, Wasson N, Cottrell EB, Fu R. Comparative effectiveness of antiviral treatment for hepatitis C virus infection in adults: a systematic review. Ann Intern Med 2013;158:114–123. [35] Amin J, Law MG, Bartlett M, Kaldor JM, Dore GJ. Causes of death after diagnosis of hepatitis B or hepatitis C infection: a large community-based linkage study. Lancet 2006;368:938–945. [36] Duberg AS, Torner A, Davidsdottir L, Aleman S, Blaxhult A, Svensson A, et al. Cause of death in individuals with chronic HBV and/or HCV infection, a

[37]

[38]

[39]

[40] [41]

[42] [43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58]

nationwide community-based register study. J Viral Hepat 2008;15: 538–550. El-Kamary SS, Jhaveri R, Shardell MD. All-cause, liver-related, and non-liverrelated mortality among HCV-infected individuals in the general US population. Clin Infect Dis 2011;53:150–157. McDonald SA, Hutchinson SJ, Bird SM, Mills PR, Dillon J, Bloor M, et al. A population-based record linkage study of mortality in hepatitis C-diagnosed persons with or without HIV coinfection in Scotland. Stat Methods Med Res 2009;18:271–283. Prasad L, Spicher VM, Negro F, Rickenbach M, Zwahlen M. Little evidence that hepatitis C virus leads to a higher risk of mortality in the absence of cirrhosis and excess alcohol intake: the Swiss Hepatitis C Cohort Study. J Viral Hepat 2009;16:644–649. Crowther MA. Introduction to surrogates and evidence-based mini-reviews. Hematol Am Soc Hematol Educ Program 2009:15–16. Ghany MG, Strader DB, Thomas DL, Seeff LB. Diagnosis, management, and treatment of hepatitis C: an update. Hepatology 2009;49: 1335–1374. Minuk GY, Gutkin A, Wong SG, Kaita KD. Patient concerns regarding chronic hepatitis C infections. J Viral Hepat 2005;12:51–57. Hsu PC, Krajden M, Yoshida EM, Anderson FH, Tomlinson GA, Krahn MD. Does cirrhosis affect quality of life in hepatitis C virus-infected patients? Liver Int 2009;29:449–458. Chong CA, Gulamhussein A, Heathcote EJ, Lilly L, Sherman M, Naglie G, et al. Health-state utilities and quality of life in hepatitis C patients. Am J Gastroenterol 2003;98:630–638. Laupacis A, Sackett DL, Robert RS. An assessment of clinically useful measurements of the consequences of treatment. N Engl J Med 1988;318: 1728–1733. Chatellier G, Zapletal E, Lemaitre D, Menard J, Degoulet P. The number needed to treat: a clinically useful nomogram in its proper context. BMJ 1996;312:426–429. Thein HH, Yi Q, Dore GJ, Krahn MD. Estimation of stage-specific fibrosis progression rates in chronic hepatitis C virus infection: a meta-analysis and meta-regression. Hepatology 2008;48:418–431. Freeman AJ, Dore GJ, Law MG, Thorpe M, Von Overbeck J, Lloyd AR, et al. Estimating progression to cirrhosis in chronic hepatitis C virus infection. Hepatology 2001;34:809–816. Dore GJ, Freeman AJ, Law M, Kaldor JM. Is severe liver disease a common outcome for people with chronic hepatitis C? J Gastroenterol Hepatol 2002;17:423–430. Smith BD, Morgan RL, Beckett GA, Falck-Ytter Y, Holtzman D, Ward JW. Hepatitis C virus testing of persons born during 1945–1965: recommendations from the centers for disease control and prevention. Ann Intern Med 2012;157:817–822. Martin N, Vickerman P, Grebly J, Hellard M, Hutchinson S, Lima VD, et al. Hepatits C virus treatment for prevention among persons who inject drugs: modelling treatment scale-up in the age of direct-acting antivirals. Hepatology 2013;58:1598–1609. Mathers B, Degenhardt L, Bucello C, Lemon J, Wiessing L, Hickman M. Mortality among people who inject drugs: a systematic review and metaanalysis. Bull World Health Organ 2013;91:102–123. Allen E, Taylor A, Palmateer N, Hutchinson S, Rees C, Johnston L, et al. The Needle Exchange Surveillance Initiative (NESI): Prevalence of HCV and injecting risk behaviours among injecting drug users attending needle exchanges in Scotland, 2008/2009. Available online at http://www.hepatitisscotlandc.org.uk/media/50084/nesi-report-08-09.pdf. Yoshida H, Tateishi R, Arakawa Y, Sata M, Fujiyama S, Nishiguchi S, et al. Benefit of interferon therapy in hepatocellular carcinoma prevention for individual patients with chronic hepatitis C. Gut 2004;53:425–430. McGarry LJ, Pawar VS, Panchmatia HR, Rubin JL, Davis GL, Younossi ZM, et al. Economic model of a birth cohort screening program for hepatitis C virus. Hepatology 2012;55:1344–1355. Kim WR, Poterucha JJ, Benson JT, Therneau TM. The impact of competing risks on the observed rate of chronic hepatitis C progression. Gastroenterology 2004;127:749–755. Innes HA, Hutchinson SJ, Barclay S, Cadzow E, Dillon JF, Fraser A, et al. Quantifying the fraction of cirrhosis attributable to alcohol among chronic hepatitis C virus patients: Implications for treatment cost-effectiveness. Hepatology 2013;57:451–460. Schuppan D, Afdhal NH. Liver cirrhosis. Lancet 2008;371:838–851.

Journal of Hepatology 2014 vol. 60 j 1118–1126

1125

Research Article [59] El-Serag HBHampel H, Yeh C, Rabeneck L. Extrahepatic manifestations of hepatitis C among United States male veterans. Hepatology 2002;36:1439–1445. [60] Giordano TP, Henderson L, Landgren O, Chiao EY, Kramer JR, El-Serag H, et al. Risk of non-hodgkin lymphoma and lymphoproliferative precursor diseases in US veterans with hepatitis C virus. JAMA 2007;297:2010–2017. [61] Uto H, Stuver SO, Hayashi K, Kumagai K, Sasaki F, Kanmura S, et al. Increased rate of death related to presence of viremia among hepatitis C virus antibody-positive subjects in a community-based cohort study. Hepatology 2009;50:393–399. [62] Harris HE, Ramsay ME, Andrews NJ. Survival of a national cohort of hepatitis C virus infected patients, 16 years after exposure. Epidemiol Infect 2006;134:472–477.

1126

[63] Seeff LB, Buskell-Bales Z, Wright EC, Durako SJ, Alter HJ, Iber FL, et al. Longterm mortality after transfusion-associated non-A, non-B hepatitis. The National Heart, Lung, and Blood Institute Study Group. N Engl J Med 1992;327:1906–1911. [64] Kielland KB, Skaug K, Amundsen EJ, Dalgard O. All-cause and liver-related mortality in hepatitis C infected drug users followed for 33 years: a controlled study. J Hepatol 2013;58:31–37. [65] Petta S, Torres D, Fazio G, Camma C, Cabibi D, Di Marco V, et al. Carotid atherosclerosis and chronic hepatitis C: a prospective study of risk associations. Hepatology 2012;55:1317–1323. [66] Life Tables for Great Britain, 2008–2010. http://www.ons.gov.uk/ons/ taxonomy/index.html?nscl=Interim+Life+Tables. Accessed January 2013.

Journal of Hepatology 2014 vol. 60 j 1118–1126