Journal of Medical Economics
ISSN: 1369-6998 (Print) 1941-837X (Online) Journal homepage: http://www.tandfonline.com/loi/ijme20
Cost-effectiveness of sacubitril/valsartan versus enalapril in patients with heart failure and reduced ejection fraction Liang Lin, David Bin-Chia Wu, Mohamed Ismail Abdul Aziz, Raymond Wong, David Sim, Kui Toh Gerard Leong, Yong Quek Wei, Doreen Tan & Kwong Ng To cite this article: Liang Lin, David Bin-Chia Wu, Mohamed Ismail Abdul Aziz, Raymond Wong, David Sim, Kui Toh Gerard Leong, Yong Quek Wei, Doreen Tan & Kwong Ng (2017): Costeffectiveness of sacubitril/valsartan versus enalapril in patients with heart failure and reduced ejection fraction, Journal of Medical Economics, DOI: 10.1080/13696998.2017.1387119 To link to this article: http://dx.doi.org/10.1080/13696998.2017.1387119
Accepted author version posted online: 29 Sep 2017.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ijme20 Download by: [Australian Catholic University]
Date: 30 September 2017, At: 15:45
Cost-effectiveness of sacubitril/valsartan versus enalapril in patients with heart failure and reduced ejection fraction Lin Lianga, David Bin-Chia Wua, Mohamed Ismail Abdul Aziza, Raymond Wongb, David Simc, Kui Toh Gerard Leongd, Yong Quek Weie, Doreen Tanf, Kwong Nga a
Agency for Care Effectiveness, Ministry of Health, Singapore Department of Cardiology, National University Heart Centre, Singapore c Department of Cardiology, National Heart Centre, Singapore d Department of Cardiology, Changi General Hospital, Singapore e Department of Cardiology, Tan Tock Seng Hospital, Singapore f Department of Pharmacy, Khoo Teck Puat Hospital, Singapore
ip
t
Correspondence: Kwong Ng, College of Medicine Building, 16 College Road, Singapore 169854, email: [email protected]
us
cr
Transparency Statement Declaration of funding No sponsorship/funding requires declaration.
an
Declaration of financial/other relationships The authors have no financial relationships to declare. Peer reviewers on this manuscript have received an honorarium from JME for their review work, but have no other relevant financial relationships to disclose.
Acknowledgments None reported.
pt ed
M
Author contributions LL and DBW contributed equally to this manuscript. LL, DBW, and KN were involved in the conception and design of the study. All authors were involved in the analysis and interpretation of the data. All authors were involved in the drafting of the paper or revising it critically for intellectual content. KN gave the final approval of the version to be published. All authors agreed to be accountable for all aspects of the work.
ce
Previous presentations An abstract of the work was accepted for oral presentation at HTAi 2017 Meeting in Rome, Italy.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
b
Abstract Background: Sacubitril/valsartan reduces cardiovascular death and hospitalisations for heart failure (HF). However, decision-makers need to determine whether its benefits are worth the additional costs, given the low-cost generic status of traditional standard of care.
cr
ip
t
Methods: A Markov model was developed to project clinical and economic outcomes of sacubitril/valsartan versus enalapril for 66-year-old patients with HF over 10 years. Key health states included New York Heart Association classes I to IV and deaths; patients in each state incurred a monthly risk of hospitalisation for HF and cardiovascular death. Sacubitril/valsartan benefits were modelled by applying the hazard ratios (HRs) in PARADIGM-HF trial to baseline probabilities. Primary model outcomes were total and incremental costs and quality-adjusted life years (QALYs) and the incremental costeffectiveness ratio (ICER) for sacubitril/valsartan relative to enalapril
M
an
us
Results: Compared to enalapril, sacubitril/valsartan was associated with an ICER of SGD 74,592 (USD 55,198) per QALY gained. A major driver of cost-effectiveness was the cardiovascular mortality benefit of sacubitril/valsartan. The uncertainty of this treatment benefit in Asian subgroup was tested in sensitivity analyses using a HR of 1 as upper limit, where the ICERs ranged from SGD 41,019 (USD 30,354) to SGD 1,447,103 (USD 1,070,856) per QALY gained. Probabilistic sensitivity analyses showed the probability of sacubitril/valsartan being cost-effective was below 1%, 12%, and 71% at SGD 20,000, SGD 50,000, and SGD 100,000 per QALY gained, respectively.
pt ed
Conclusions: At current daily price of SGD 9.00, sacubitril/valsartan may not represent good value for limited healthcare dollars compared to enalapril in reducing cardiovascular morbidity and mortality in HF in Singapore healthcare setting. This study highlights the cost-benefit trade-off that healthcare professionals and patients face when considering therapy.
ce
Keywords: heart failure; sacubitril/valsartan; enalapril; cost-effectiveness; Asian; Singapore
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Aims: To evaluate the cost-effectiveness of sacubitril/valsartan compared to enalapril in patients with HF and reduced ejection fraction, from the Singapore healthcare payer perspective.
M
an
us
cr
ip
t
Globally, the etiology of HF also varies, but hypertension and coronary artery disease (often associated with obesity and diabetes mellitus) appear to remain important, particularly so in Asia. Increase in these risk factors, related to changing lifestyles and ageing populations, is expected to drive higher burden of HF in Asia [2]. Despite so, Asians are often underrepresented in global landmark trials of HF and therefore the generalizability of trial results to Asians remains uncertain [3]. The latest PARADIGM-HF (Prospective Comparison of ARNI with ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial is one of the largest and most globally representative trial in HF to date, with 18% (n=1,487) of patients recruited in Asia. While the trial showed sacubitril/valsartan, a combination of an angiotensin receptor neprilysin inhibitor (ARNI) and an angiotensin II receptor blocker (ARB), reduced hospitalisations for HF and cardiovascular (CV) mortality by 20% compared with enalapril, an angiotensin-converting enzyme inhibitor (ACEI), subgroup analysis by region showed the risk reduction for both endpoints were non-statistically significant in Asian population [4]. Nonetheless, the study may be under-powered to detect subgroup-level effect; and statistical tests for interaction found no significant association between geographic region and the effect of treatment (p=0.58) [5].
ce
pt ed
Based on evidence from PARADIGM-HF, sacubitril/valsartan has been recommended by international guidelines for HF [6]. While management of Asian patients generally follow international practice, the uncertainty in clinical benefits of sacubitril/valsartan for Asian patients needs to be considered by payers, especially when it comes at a substantially higher upfront treatment cost compared to traditional standard of care (ACEIs) that are mostly generics. Recent cost-effectiveness studies for sacubitril/valsartan are all conducted from the U.S. perspective, and that in the Asian healthcare setting is lacking [7-9]. Therefore we conducted a cost-effectiveness analysis, incorporating the treatment effects observed in the Asian subgroup of the PARADIGM-HF trial, in Singapore to assess the value for money of sacubitril/valsartan.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Introduction Heart failure (HF) is a major public health problem worldwide with extensive burden of disease. Between 1% and 2% of the global population over 40 years of age have HF; this rises to 10% in those over 60 to 70 years old and will continue to increase with an ageing population. It is one of the major causes of hospital admissions [1]. After discharge, around 25% have readmissions, and a further 5 to 10% die within one month. In Asia, data on HF burden is more limited; however, the available estimates are generally similar to that in the West [2]. Asian patients with HF have similarly poor or even worse outcomes than patients from the West, in part due to differences in clinical phenotypes and practice patterns [3].
Methods Model Structure We developed a Markov model, which is a cohort-based state-transition model, to compare ARNI (sacubitril/valsartan) and ACEI (enalapril) for the prevention of hospitalisations for HF and CV mortality - the two primary endpoints of the PARADIGM-HF trial (Figure 1). In line with the mean age of Singaporeans with HF, the starting age of the patient cohort was 66 years [2,10,11]. Health states were New York Heart Association (NYHA) classes I to IV and deaths, with majority of patients starting in class II (72%) and III (24%), based on characteristics from PARADIGM-HF. Patients in each state incurred a monthly risk of hospitalisation for HF and CV death. In addition, hospitalised HF patients were at risk of readmission for HF and in-patient deaths. While patients could experience only one readmission episode during the one-month Markov cycle length, multiple readmissions over the time horizon were allowed to reflect
cr
ip
t
Data and sources Clinical data We estimated baseline probabilities of hospitalisation for HF and CV death for the enalapril group by extrapolating published Kaplan-Meier curves from PARADIGM-HF trial using Weibull and exponential distributions, respectively [14]. The clinical benefits of sacubitril/valsartan were modelled by applying the hazard ratios (HRs) to the baseline probabilities: it reduced hospitalisations for HF by 21% (HR 0.79; 95% confidence interval (CI) 0.71 to 0.89) and CV mortality by 20% (HR 0.80; 95% CI 0.71 to 0.89) relative to enalapril. Its treatment effect in the Asian subgroup was reflected using a HR of 1 as the upper limit in sensitivity analyses, given that the results were non-statistically significant in Asian patients: hospitalisation for HF; 95% CI 0.61 to 1.12 and CV mortality; 95% CI 0.62 to 1.01 [4]. The treatment effect was assumed to continue beyond trial duration of 27 months over 10 years.
M
an
us
The probabilities of readmission for HF and in-patient deaths due to CV causes (defined as acute myocardial infarction, arrhythmia, HF, ischaemic heart disease, peripheral artery disease or stroke) within one month were derived from local epidemiological data (2015 data), and assumed to be the same across treatment groups. Background all-cause mortality was derived from 2016 life table data for Singapore population [15]. These rates were adjusted as the cohort aged over the time horizon of the analysis. A two-fold increased risk of death, based on age-adjusted mortality ratios calculated from national case-mix data, was applied to the general all-cause mortality rates from the life tables to reflect the higher mortality burden due to HF. This risk was also assumed to be the same across treatment groups and no treatment effect was modelled for this endpoint.
ce
pt ed
One-month transition probabilities between NYHA health states were derived from an established matrix in cost-effectiveness studies for HF (Table 2) [7]. Patients could transition to a lower NYHA class, for example, from NYHA class II to class I, except post hospitalisations or readmissions where they either remained within the same NYHA class or progressed, for example, from NYHA class II to class III. These probabilities were fixed over time. Conservatively, the probabilities were assumed to be the same for both treatment groups, because it was unclear how sacubitril/valsartan altered the progression between health states relative to enalapril. Utility data Utilities describe the quality of life for different NYHA classes according to individuals’ preferences and range from 1 (perfect health) to 0 (death). They are combined with life expectancy to generate quality-adjusted life years (QALYs). The utilities for NYHA classes were derived from the Care-HF (Cardiac Resynchronization in Heart Failure) trial, in which healthrelated quality of life was captured in patients with chronic HF and reduced LVEF (≤35%) on optimal medical therapy with the European Quality of Life-5 Dimensions instrument [16]. Lower estimates for NYHA class IV (0.51) compared to class II (0.82) indicated poorer quality of life. A one-time disutility of 0.1 was applied, by subtracting patient’s utility in each NYHA class by 0.1, for each hospitalisation or readmission event [7,17]. Utilities were varied over their 95% CIs in sensitivity analyses.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
clinical practice. We used a time horizon of 10 years, based on the average life expectancy of 5 to 10 years for patients with HF [12,13]. The analysis was conducted from Singapore healthcare payer’s perspective. All costs and health outcomes were discounted at 3% annually. Model development and analyses were performed using TreeAge Pro SuiteTM software 2016 (Williamstown, MA. USA). Table 1 shows the list of model inputs.
ip
t
Costs of events were obtained from the national database with a case-mix of 6,663 patients with HF. The costs incurred during hospitalisations for HF of SGD 4,537 comprised doctor consultations, hospitalisations (average 5 days’ LOS), ICU stay (average 2 to 3 days’ LOS), procedures, investigations and laboratory tests. The costs for readmissions for HF of SGD 4,317 were similar to the index hospitalisation, given that clinical management is similar. The costs of in-patient deaths was higher, at SGD 7,170, because these patients required more intensive care and expensive treatment such as dialysis, on top of the costs of hospitalisation.
M
an
us
cr
Main outcome measures The main outcome measures were the number of first hospitalisations for HF and CV deaths prevented, total and incremental costs and QALYs, and the incremental cost-effectiveness ratio (ICER) for sacubitril/valsartan relative to enalapril. The ICER is a ratio of incremental healthcare costs to incremental health outcomes or QALYs gained and expressed in terms of cost per QALY gained. It is typically compared against a cost-effectiveness threshold that reflects willingness-to-pay (WTP). In Singapore, there is no fixed WTP threshold to determine cost-effectiveness. To address this, sensitivity analyses were conducted for various WTP thresholds of between SGD 20,000 and SGD 100,000 per QALY gained.
pt ed
Sensitivity analyses We performed one-way sensitivity analyses and generated Tornado diagram to present the relative impact of varying the model inputs on outcomes (i.e. ICERs). We conducted multivariate probabilistic sensitivity analysis (PSA), using 10,000 Monte Carlo simulations, to evaluate how the simultaneous uncertainties about model inputs across their distributions might influence outcomes.
ce
Results Base-case analysis Among 1,000 66-year-old patients with HF treated with enalapril, our model predicted 306 hospitalisations for HF and 426 CV deaths over 10 years. Relative to enalapril, sacubitril/valsartan prevented an additional 37 hospitalisations for HF and 65 CV deaths (Table 3). The effectiveness of sacubitril/valsartan in reducing hospitalisations for HF and CVdeaths translated into a mean gain of 0.21 QALYs per person. The total projected costs for enalapril and sacubitril/valsartan groups over the 10 years’ time horizon were SGD 2,197 and SGD 17,857, respectively. The additional mean cost of SGD 15,660 for treatment with sacubitril/valsartan was attributable to its substantially higher cost compared to enalapril (additional SGD 3,168 per surviving person annually, for 10 years), which was partially offset by fewer events in the sacubitril/valsartan group. These resulted in a base-case ICER for sacubitril/valsartan versus enalapril of SGD 74,592 (USD 55,198; based on SGD1 = USD0.74 as of September 2017) per QALY gained (Table 4).
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Resource use and cost data Only direct medical costs were used, in accordance with the model’s perspective. Medication costs were based on prices charged to patients at a tertiary national heart centre in Singapore for enalapril 10mg twice daily (SGD 6.00/month) and sacubitril/valsartan 200mg twice daily (SGD 270.00/month). These costs were varied widely (50% in each direction) in one-way sensitivity analyses to account for wide variations across different settings. The costs of additional HF medications and associated doctor visits were not considered; they were assumed to be equivalent between groups.
t
us
cr
ip
Additional analyses on cost of sacubitril/valsartan Given the cost of sacubitril/valsartan - which can be variable depending on country and timing - was among the top three drivers of cost-effectiveness, we performed additional analyses to examine the impact of pricing on model results (Figure 3). The daily cost of sacubitril/valsartan was set to SGD 9.00 in the base-case based on current price. The price would need to drop by 70%, 58% and 32% to SGD 2.70, SGD 3.80 and SGD 6.10, respectively, for sacubitril/valsartan to cost SGD 20,000, SGD 30,000 and SGD 50,000 per QALY gained.
pt ed
M
an
Probabilistic sensitivity analyses The average ICER per QALY gained from multivariate probabilistic sensitivity analyses, which took into account uncertainty of treatment effect in the Asian subgroup, was SGD 74,897 (USD 55,490). Out of 10,000 Monte Carlo simulations, sacubitril/valsartan were cost-effective below 1% of the time, at a WTP of SGD 20,000 or SGD 30,000 per QALY. This increased to 12% and 71% when WTP was increased to SGD 50,000 and SGD 100,000 per QALY gained, respectively (Figure 4).
ce
Model calibration We calibrated the model by comparing the model simulation results to that observed in PARADIGM-HF for both hospitalisations for HF and CV mortality. Using micro-simulation with tracker variables, our model produced estimates of hospitalisations for HF and CV death that compared well with the rates observed in PARADIGM-HF (Figure 5). These results suggest the model structure and baseline probabilities reflect current best estimates of the comparative effects of sacubitril/valsartan and enalapril on hospitalisation for HF and CV death rates.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Sensitivity analysis Deterministic sensitivity analysis showed that the effectiveness of sacubitril/valsartan in reducing the risk of CV death compared with enalapril was the key model driver. When the HR for CV death was varied across the 95% CI from 0.62 to 1.00 – the upper limit reflecting the non-statistically significant treatment effect in Asian subgroup - the ICER ranged from SGD 41,019 (USD 30,354) to SGD 1,447,103 (USD 1,070,856) per QALY gained. The modelled time horizon also had considerable impact; when the time horizon was shortened to the median follow-up duration of PARADIGM-HF trial (27 months), the cost per QALY increased to SGD 271,921 (USD 201,222). Conversely, modelling over a longer time horizon (20 years) resulted in an ICER of SGD 54,777 (USD 40,535). The cost of sacubitril/valsartan resulted in the third biggest impact on ICER, followed by background all-cause mortality rates and HR for hospitalisation for HF. Other variables, including costs of events or quality of life values had little impact on the cost-effectiveness of the treatment (< ±SGD 3,000 in terms of cost per QALY) (Figure 2).
Discussion As the final common pathway of a myriad of heart diseases, the burden of HF increases with the growing prevalence of CV disease. The World Health Organization has projected that the largest increases in CV disease worldwide are occurring in Asia. Sacubitril/valsartan offers an effective means to reduce the burden of HF. However, given the low-cost generic status of ACE inhibitors, decision-makers and payers need to determine whether the extra benefit with sacubitril/valsartan observed in PARADIGM-HF is worth the additional costs. This study is the first to examine the cost-effectiveness of sacubitril/valsartan versus enalapril, based on costs and resources used in Singapore, a developed country in Asia, to inform decision-making in healthcare. Our study found that relative to enalapril, sacubitril/valsartan prevented an additional 37
M
an
us
cr
ip
t
The key model driver was the effectiveness of sacubitril/valsartan in reducing the risk of CV death that was assumed to last over the time horizon of 10 years. Of note, the uncertainty in its treatment effect in the Asian subgroup translated into an ICER of SGD 1,447,103 103 (USD 1,070,856) per QALY gained in sensitivity analyses. However, subgroup analyses are inherently underpowered. On the other hand, the effectiveness of sacubitril/valsartan in reducing the risk of hospitalisations for HF did not have a significant impact on the model. HF is a condition with significant mortality so the modelled population declined over time due to increasing mortality. As a result, sacubitril/valsartan’s benefit for averting hospitalisations for HF diminished in the latter years. Correspondingly, the model simulated fewer hospitalisations for HF averted compared to CV deaths, and was less sensitive to impact of hospitalisation for HF. Accordingly, that the model did not incorporate endpoints such as reduction in emergency department readmissions for HF or ICU stays was unlikely to affect the cost-effectiveness results, given the relatively small impact of other benefits compared to CV death reduction [19].
ce
pt ed
These findings were consistent with recent cost-effectiveness analyses of sacubitril/valsartan conducted in the U.S. setting. King et al. found the cost-effectiveness of sacubtril/valsartan highly dependent on the duration of treatment effect, ranging from USD 249,411 per QALY at 3 years to USD 50,959 per QALY gained over a lifetime [7]. Gaziano et al. reported similar findings, with ICER of USD 135,964 per QALY gained when the benefit only lasting 27 months (trial duration) was modelled [8]. Sandhu et al. also reported ICER of USD 120,623 per QALY gained if the treatment was effective only for 27 months, with no subsequent reduction in event risks or improvement in quality of life but continued costs [9]. These analyses adopted a longer time horizon up to 30 years that inherently had greater uncertainty given that the effectiveness results were extrapolated from a trial with a duration of 2 years. None of these analyses applied treatment effect observed in the Asian population of PARADIGM-HF trial.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
hospitalisations for HF and 65 CV deaths over 10 years. Sacubitril/valsartan costs payers SGD 74,592 (USD 55,198) per QALY gained compared with enalapril in the Singapore healthcare setting. Locally, the healthcare system is a combination of government healthcare subsidies, mandatory healthcare savings accounts and health insurance. While there is no fixed WTP threshold, sacubitril/valsartan may not be considered cost-effective use of the finite healthcare budget, when benchmarking against other CV drugs in the Singapore healthcare setting with ICER of SGD 10,000 per QALY gained [18]. A price reduction of 32% to 70%, from current price of SGD 9.00/day, will be needed for sacubitril/valsartan to cost between SGD 20,000 and SGD 50,000 per QALY gained and be considered cost-effective compared with other drugs in similar therapeutic area. This will also improve the affordability for patients with HF who currently have no access to the drug due to cost barrier. Given the large and growing number of patients with HF, this will also ensure the sustainability of the healthcare system from a budgetary perspective.
What set our work apart from these studies was contextualising the cost-effectiveness analysis of sacubitril/valsartan in the local healthcare setting. Our modelled population reflected the average age of patients with HF in Singapore and incorporated the treatment effects using results from the Asian subgroup of PARADIGM-HF trial in sensitivity analyses. We used Singapore-specific all-cause mortality rates, with age-adjusted ratios derived from national epidemiological data to reflect higher mortality burden due to HF. Based on national epidemiological data, we applied 16% readmission rates for HF among hospitalised cases and 4% in-patient deaths due to CV causes within one month that were consistent with international figures [20]. We applied quality-of-life values for various NYHA classes; our utility inputs correlated well with the reported scores of 0.78 (SD 0.18) to 0.51 (SD 0.21) for mild to severe disease in patients with HF in a systematic review of utilities for CV disease [21]. Local
us
cr
ip
t
Our findings should be interpreted within the limitations of our analysis. Firstly, we extrapolated the treatment benefits of sacubitril/valsartan for reducing hospitalisations for HF and CV mortality from PARADIGM-HF over 10 years. It remains to be known whether the assumption of sustained treatment effect of sacubitril/valsartan over enalapril beyond trial and throughout the time horizon holds true. Second, the trial included an active run-in period so our results assumed that patients follow the recommended dose of each medication. Dose reduction is associated with higher risk for major CV events [22]. In fact, local clinical experts have provided feedback about hypotensive episodes in real-life patients who are unable to tolerate sacubitril/valsartan’s hemodynamic effects at target dose, especially the older and frailer ones. This early clinical experience mirrored PARADIGM-HF trial data where sacubitril/valsartan was associated with significantly higher rates of hypotension than enalapril (17.6% vs 12.0%) but they rarely required discontinuation of treatment. Other common AEs such as hyperkalemia and renal impairment were not modelled. Given that the discontinuation of study medications due to these events was numerically lower among patients taking sacubitril/valsartan (10.7%) than among patients taking enalapril (12.3%) in PARADIGM-HF, the results from our analysis may be conservative. Furthermore, the analysis did not factor in potential impact of chronic neprilysin inhibition with sacubitril/valsartan on the risk of developing dementia. Studies are currently ongoing to assess its effects on cognitive function.
ce
pt ed
M
an
Conclusions In conclusion, our study reports the long-term cost-effectiveness of sacubitril/valsartan versus enalapril for patients with HFrEF in Singapore healthcare setting. Benchmarking against the ICERs of other drugs in the same setting, sacubitril/valsartan may not represent good value for limited health care dollars at its current price of SGD 9.00/day. This highlights the cost-benefit trade-off that health care professionals and patients face when considering HF therapy. To date, it remains to be proven in the real-world setting whether the mortality reduction benefits in the trial translates into true clinical benefits and actual cost savings from avoiding deaths. Ultimately, the clinical decision on whether to use this novel agent should be individualised and based on shared decision-making that reflects treatment-related benefits, risks, costs and patient preferences.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
costs of hospitalisations and readmissions, however, were lower, at SGD 4,537 (USD 3,357) and SGD 4,317 (USD 3,195), respectively, compared to reported figures of USD 10,698 and USD 11,361 in the U.S [7]. Such differences in costs of care in each country highlight the importance of conducting local cost-effectiveness analyses.
t
ip
cr
us
an
M
pt ed
ce
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
References 1. Ambrosy AP, Fonarow GC, Butler J, et al. The global health and economic burden of hospitalizations for heart failure: lessons learned from hospitalized heart failure registries. J Am Coll Cardiol. 2014;63(12):1123-33. 2. Reyes EB, Ha JW, Firdaus I, et al. Heart failure across Asia: Same healthcare burden but differences in organization of care. International journal of cardiology. 2016;223:163-7. 3. Mentz RJ, Roessig L, Greenberg BH, et al. Heart Failure Clinical Trials in East and Southeast Asia: Understanding the Importance and Defining the Next Steps. JACC Heart failure. 2016;4(6):419-27. 4. Kristensen SL, Martinez F, Jhund PS, et al. Geographic variations in the PARADIGM-HF heart failure trial. European heart journal. 2016. 5. McMurray JJV, Packer M, Desai AS, et al. Angiotensin–Neprilysin Inhibition versus Enalapril in Heart Failure. New England Journal of Medicine. 2014;371(11):993-1004. 6. Yancy CW, Jessup M, Bozkurt B, et al. 2016 ACC/AHA/HFSA Focused Update on New Pharmacological Therapy for Heart Failure: An Update of the 2013 ACCF/AHA Guideline for the Management of Heart FailureA Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America. Journal of the American College of Cardiology. 2016. 7. King JB, Shah RU, Bress AP, et al. Cost-Effectiveness of Sacubitril-Valsartan Combination Therapy Compared With Enalapril for the Treatment of Heart Failure With Reduced Ejection Fraction. JACC Heart failure. 2016;4(5):392-402. 8. Gaziano TA, Fonarow GC, Claggett B, et al. COst-effectiveness analysis of sacubitril/valsartan vs enalapril in patients with heart failure and reduced ejection fraction. JAMA Cardiology. 2016. 9. Sandhu AT, Ollendorf DA, Chapman RH, et al. Cost-Effectiveness of Sacubitril-Valsartan in Patients With Heart Failure With Reduced Ejection Fraction. Annals of internal medicine. 2016. 10. Lee R, Chan SP, Chan YH, et al. Impact of race on morbidity and mortality in patients with congestive heart failure: a study of the multiracial population in Singapore. International journal of cardiology. 2009;134(3):422-5. 11. Leong KT, Goh PP, Chang BC, et al. Heart failure cohort in Singapore with defined criteria: clinical characteristics and prognosis in a multi-ethnic hospital-based cohort in Singapore. Singapore medical journal. 2007;48(5):408-14. 12. Allen LA, Yager JE, Funk MJ, et al. Discordance between patient-predicted and modelpredicted life expectancy among ambulatory patients with heart failure. Jama. 2008;299(21):2533-42. 13. Alter DA, Ko DT, Tu JV, et al. The average lifespan of patients discharged from hospital with heart failure. Journal of general internal medicine. 2012;27(9):1171-9. 14. Hoyle MW, Henley W. Improved curve fits to summary survival data: application to economic evaluation of health technologies. BMC Medical Research Methodology. 2011;11(1):139. 15. Yearbook of Statistics Singapore 2016: Population. Department of Statistics, Singapore. Cited February 2017. Available from URL: http://www.singstat.gov.sg/publications/publications_and_papers/reference/yearbook_of _stats.html. 16. Calvert MJ, Freemantle N, Cleland JG. The impact of chronic heart failure on health-related quality of life data acquired in the baseline phase of the CARE-HF study. European journal of heart failure. 2005;7(2):243-51. 17. Yao G, Freemantle N, Flather M, et al. Long-term cost-effectiveness analysis of nebivolol compared with standard care in elderly patients with heart failure: an individual patientbased simulation model. PharmacoEconomics. 2008;26(10):879-89.
t ip cr us an M pt ed ce Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
18. Chin CT, Mellstrom C, Chua TS, et al. Lifetime cost-effectiveness analysis of ticagrelor in patients with acute coronary syndromes based on the PLATO trial: a Singapore healthcare perspective. Singapore medical journal. 2013;54(3):169-75. 19. Packer M, McMurray JJ, Desai AS, et al. Angiotensin receptor neprilysin inhibition compared with enalapril on the risk of clinical progression in surviving patients with heart failure. Circulation. 2015;131(1):54-61. 20. Desai AS, Stevenson LW. Rehospitalization for Heart Failure. Predict or Prevent? Circulation 2012;126(4):501-6. 21. Dyer MTD, Goldsmith KA, Sharples LS, et al. A review of health utilities using the EQ-5D in studies of cardiovascular disease. Health and Quality of Life Outcomes. 2010;8:13. 22. Vardeny O, Claggett B, Packer M, et al. Efficacy of sacubitril/valsartan vs. enalapril at lower than target doses in heart failure with reduced ejection fraction: the PARADIGM-HF trial. European journal of heart failure. 2016;18(10):1228-34.
Table 1. Model inputs Model parameters
Base-case
Range
Intercept Log_scale 0.0904 0.00 0.0328 0.0197 0.0043 to 0.0080
Exponential
Hazard ratios for sacubitril/valsartan compared with enalapril Hospitalisation for HF 0.79
0.71 to 0.89
Log-normal
CV mortality
0.71 to 0.89
Log-normal
ce
M
an
us
0.781 to 0.850 0.693 to 0.749 0.551 to 0.629 0.412 to 0.605 -0.080 to -0.130
3.00 to 9.00 135.00 to 405.00
793.37 to 16,854,73 824.97 to 15,393.09 918.82 to 14,934.49
Extrapolated from published KaplanMeier curves from PARADIGM-HF trial PARADIGM-HF trial PARADIGM-HF trial Local data Local data
t
cr
pt ed
Monthly probabilities for sacubitril/valsartan and enalapril 30-day readmission for HF 0.16 30-day inpatient death due to 0.04 CV causes Utility inputs NYHA class I 0.815 NYHA class II 0.720 NYHA class III 0.590 NYHA class IV 0.508 Disutility for -0.100 hospitalisation/readmission Cost inputs (SGD) Cost of medication (monthly) Enalapril 6.00 Sacubitril/valsartan 270.00 Cost of events (one-time) Hospitalisation for HF 4,537.01 Readmission for HF 4,317.22 Inpatient deaths due to CV 7,170.00 causes
Weibul
Source
ip
0.80
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Clinical inputs Monthly probabilities for enalapril Hospitalisation for HF Intercept=5.2700 Cholesky matrix Log_scale=0.1792 Intercept Log_scale CV mortality 0.0067
Distributi on
Beta Beta Beta Beta Beta
CARE-HF CARE-HF CARE-HF CARE-HF CARE-HF
Local hospital Local hospital
Gamma Gamma Gamma
Local hospital Local hospital Local hospital
Table 2. NYHA transition probabilities per one-month cycle From NYHA
To NYHA Class II
Class III
Class IV
0.9923
0.0064
0.0013
0
Class II
0.0027
0.9936
0.0034
0.0003
Class III
0
0.0116
0.9864
0.0021
Class IV
0
0
0.0189
0.9813
ip
t
Class I
cr
Table 3. Projected clinical events per 1,000 patients with HF over 10 years’ modelled time horizon Hospitalisation for
Difference
CV deaths
HF
Sacubitril/ Valsartan
269
M
306
pt ed
Enalapril
an
Base-case
Difference
us
Treatments
--
426
--
-37
361
-65
Table 4. Projected costs, benefits and incremental cost-effectiveness ratio
Base-case Enalapril
Sacubitril/
Costs (SGD)
ce
Treatments
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Class I
2,197
17,857
Incremental cost (SGD)
--
15,660
QALY
Incremental QALYs
ICER (SGD/ QALY gained)
3.29
--
--
3.50
0.21
74,592
Valsartan ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life-year.
t
ip
or sacubitril/valsartan. Patients who experience clinical events are transitioned through
cr
the health states of New York Heart Association function class I, II, III, IV or death
ce
pt ed
M
an
us
every month.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Figure 1. Markov model structure. Patients enter the model and are on either enalapril
Figure 2. One-way sensitivity analyses (Tornado diagram). The change in incremental cost-effectiveness ratio of sacubitril/valsartan versus enalapril associated with varying key model inputs are shown. The vertical line represents the incremental costeffectiveness ratio in the base-case analysis. HF = heart failure; QALY = qualityadjusted life-year.
t
ip
cr
cost-effectiveness ratio of sacubitril/valsartan versus enalapril associated with varying
us
price of sacubitril/valsartan is shown. The daily price of sacubitril/valsartan used in the
ce
pt ed
M
an
base-case analysis is SGD 9.00.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Figure 3. Sensitivity analysis of cost of sacubitril/valsartan. The change in incremental
Figure 4. Probabilistic sensitivity analyses (cost-effectiveness acceptability curve). The curves show the proportion of the 10,000 model simulations at which sacubitril/valsartan was cost-effective across a range of thresholds. The analyses were performed by independently sampling each model input parameter from its distribution.
t ip cr us an M pt ed
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
ce Ac
Figure 5. Model calibration for a) hospitalization for HF and b) CV death. The KaplanMeier estimates of hospitalization for HF and CV death from the PARADIGM-HF trial (in blue) are shown, along with observed estimates in the simulation model (in orange).
ISSN: 1369-6998 (Print) 1941-837X (Online) Journal homepage: http://www.tandfonline.com/loi/ijme20
Cost-effectiveness of sacubitril/valsartan versus enalapril in patients with heart failure and reduced ejection fraction Liang Lin, David Bin-Chia Wu, Mohamed Ismail Abdul Aziz, Raymond Wong, David Sim, Kui Toh Gerard Leong, Yong Quek Wei, Doreen Tan & Kwong Ng To cite this article: Liang Lin, David Bin-Chia Wu, Mohamed Ismail Abdul Aziz, Raymond Wong, David Sim, Kui Toh Gerard Leong, Yong Quek Wei, Doreen Tan & Kwong Ng (2017): Costeffectiveness of sacubitril/valsartan versus enalapril in patients with heart failure and reduced ejection fraction, Journal of Medical Economics, DOI: 10.1080/13696998.2017.1387119 To link to this article: http://dx.doi.org/10.1080/13696998.2017.1387119
Accepted author version posted online: 29 Sep 2017.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ijme20 Download by: [Australian Catholic University]
Date: 30 September 2017, At: 15:45
Cost-effectiveness of sacubitril/valsartan versus enalapril in patients with heart failure and reduced ejection fraction Lin Lianga, David Bin-Chia Wua, Mohamed Ismail Abdul Aziza, Raymond Wongb, David Simc, Kui Toh Gerard Leongd, Yong Quek Weie, Doreen Tanf, Kwong Nga a
Agency for Care Effectiveness, Ministry of Health, Singapore Department of Cardiology, National University Heart Centre, Singapore c Department of Cardiology, National Heart Centre, Singapore d Department of Cardiology, Changi General Hospital, Singapore e Department of Cardiology, Tan Tock Seng Hospital, Singapore f Department of Pharmacy, Khoo Teck Puat Hospital, Singapore
ip
t
Correspondence: Kwong Ng, College of Medicine Building, 16 College Road, Singapore 169854, email: [email protected]
us
cr
Transparency Statement Declaration of funding No sponsorship/funding requires declaration.
an
Declaration of financial/other relationships The authors have no financial relationships to declare. Peer reviewers on this manuscript have received an honorarium from JME for their review work, but have no other relevant financial relationships to disclose.
Acknowledgments None reported.
pt ed
M
Author contributions LL and DBW contributed equally to this manuscript. LL, DBW, and KN were involved in the conception and design of the study. All authors were involved in the analysis and interpretation of the data. All authors were involved in the drafting of the paper or revising it critically for intellectual content. KN gave the final approval of the version to be published. All authors agreed to be accountable for all aspects of the work.
ce
Previous presentations An abstract of the work was accepted for oral presentation at HTAi 2017 Meeting in Rome, Italy.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
b
Abstract Background: Sacubitril/valsartan reduces cardiovascular death and hospitalisations for heart failure (HF). However, decision-makers need to determine whether its benefits are worth the additional costs, given the low-cost generic status of traditional standard of care.
cr
ip
t
Methods: A Markov model was developed to project clinical and economic outcomes of sacubitril/valsartan versus enalapril for 66-year-old patients with HF over 10 years. Key health states included New York Heart Association classes I to IV and deaths; patients in each state incurred a monthly risk of hospitalisation for HF and cardiovascular death. Sacubitril/valsartan benefits were modelled by applying the hazard ratios (HRs) in PARADIGM-HF trial to baseline probabilities. Primary model outcomes were total and incremental costs and quality-adjusted life years (QALYs) and the incremental costeffectiveness ratio (ICER) for sacubitril/valsartan relative to enalapril
M
an
us
Results: Compared to enalapril, sacubitril/valsartan was associated with an ICER of SGD 74,592 (USD 55,198) per QALY gained. A major driver of cost-effectiveness was the cardiovascular mortality benefit of sacubitril/valsartan. The uncertainty of this treatment benefit in Asian subgroup was tested in sensitivity analyses using a HR of 1 as upper limit, where the ICERs ranged from SGD 41,019 (USD 30,354) to SGD 1,447,103 (USD 1,070,856) per QALY gained. Probabilistic sensitivity analyses showed the probability of sacubitril/valsartan being cost-effective was below 1%, 12%, and 71% at SGD 20,000, SGD 50,000, and SGD 100,000 per QALY gained, respectively.
pt ed
Conclusions: At current daily price of SGD 9.00, sacubitril/valsartan may not represent good value for limited healthcare dollars compared to enalapril in reducing cardiovascular morbidity and mortality in HF in Singapore healthcare setting. This study highlights the cost-benefit trade-off that healthcare professionals and patients face when considering therapy.
ce
Keywords: heart failure; sacubitril/valsartan; enalapril; cost-effectiveness; Asian; Singapore
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Aims: To evaluate the cost-effectiveness of sacubitril/valsartan compared to enalapril in patients with HF and reduced ejection fraction, from the Singapore healthcare payer perspective.
M
an
us
cr
ip
t
Globally, the etiology of HF also varies, but hypertension and coronary artery disease (often associated with obesity and diabetes mellitus) appear to remain important, particularly so in Asia. Increase in these risk factors, related to changing lifestyles and ageing populations, is expected to drive higher burden of HF in Asia [2]. Despite so, Asians are often underrepresented in global landmark trials of HF and therefore the generalizability of trial results to Asians remains uncertain [3]. The latest PARADIGM-HF (Prospective Comparison of ARNI with ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial is one of the largest and most globally representative trial in HF to date, with 18% (n=1,487) of patients recruited in Asia. While the trial showed sacubitril/valsartan, a combination of an angiotensin receptor neprilysin inhibitor (ARNI) and an angiotensin II receptor blocker (ARB), reduced hospitalisations for HF and cardiovascular (CV) mortality by 20% compared with enalapril, an angiotensin-converting enzyme inhibitor (ACEI), subgroup analysis by region showed the risk reduction for both endpoints were non-statistically significant in Asian population [4]. Nonetheless, the study may be under-powered to detect subgroup-level effect; and statistical tests for interaction found no significant association between geographic region and the effect of treatment (p=0.58) [5].
ce
pt ed
Based on evidence from PARADIGM-HF, sacubitril/valsartan has been recommended by international guidelines for HF [6]. While management of Asian patients generally follow international practice, the uncertainty in clinical benefits of sacubitril/valsartan for Asian patients needs to be considered by payers, especially when it comes at a substantially higher upfront treatment cost compared to traditional standard of care (ACEIs) that are mostly generics. Recent cost-effectiveness studies for sacubitril/valsartan are all conducted from the U.S. perspective, and that in the Asian healthcare setting is lacking [7-9]. Therefore we conducted a cost-effectiveness analysis, incorporating the treatment effects observed in the Asian subgroup of the PARADIGM-HF trial, in Singapore to assess the value for money of sacubitril/valsartan.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Introduction Heart failure (HF) is a major public health problem worldwide with extensive burden of disease. Between 1% and 2% of the global population over 40 years of age have HF; this rises to 10% in those over 60 to 70 years old and will continue to increase with an ageing population. It is one of the major causes of hospital admissions [1]. After discharge, around 25% have readmissions, and a further 5 to 10% die within one month. In Asia, data on HF burden is more limited; however, the available estimates are generally similar to that in the West [2]. Asian patients with HF have similarly poor or even worse outcomes than patients from the West, in part due to differences in clinical phenotypes and practice patterns [3].
Methods Model Structure We developed a Markov model, which is a cohort-based state-transition model, to compare ARNI (sacubitril/valsartan) and ACEI (enalapril) for the prevention of hospitalisations for HF and CV mortality - the two primary endpoints of the PARADIGM-HF trial (Figure 1). In line with the mean age of Singaporeans with HF, the starting age of the patient cohort was 66 years [2,10,11]. Health states were New York Heart Association (NYHA) classes I to IV and deaths, with majority of patients starting in class II (72%) and III (24%), based on characteristics from PARADIGM-HF. Patients in each state incurred a monthly risk of hospitalisation for HF and CV death. In addition, hospitalised HF patients were at risk of readmission for HF and in-patient deaths. While patients could experience only one readmission episode during the one-month Markov cycle length, multiple readmissions over the time horizon were allowed to reflect
cr
ip
t
Data and sources Clinical data We estimated baseline probabilities of hospitalisation for HF and CV death for the enalapril group by extrapolating published Kaplan-Meier curves from PARADIGM-HF trial using Weibull and exponential distributions, respectively [14]. The clinical benefits of sacubitril/valsartan were modelled by applying the hazard ratios (HRs) to the baseline probabilities: it reduced hospitalisations for HF by 21% (HR 0.79; 95% confidence interval (CI) 0.71 to 0.89) and CV mortality by 20% (HR 0.80; 95% CI 0.71 to 0.89) relative to enalapril. Its treatment effect in the Asian subgroup was reflected using a HR of 1 as the upper limit in sensitivity analyses, given that the results were non-statistically significant in Asian patients: hospitalisation for HF; 95% CI 0.61 to 1.12 and CV mortality; 95% CI 0.62 to 1.01 [4]. The treatment effect was assumed to continue beyond trial duration of 27 months over 10 years.
M
an
us
The probabilities of readmission for HF and in-patient deaths due to CV causes (defined as acute myocardial infarction, arrhythmia, HF, ischaemic heart disease, peripheral artery disease or stroke) within one month were derived from local epidemiological data (2015 data), and assumed to be the same across treatment groups. Background all-cause mortality was derived from 2016 life table data for Singapore population [15]. These rates were adjusted as the cohort aged over the time horizon of the analysis. A two-fold increased risk of death, based on age-adjusted mortality ratios calculated from national case-mix data, was applied to the general all-cause mortality rates from the life tables to reflect the higher mortality burden due to HF. This risk was also assumed to be the same across treatment groups and no treatment effect was modelled for this endpoint.
ce
pt ed
One-month transition probabilities between NYHA health states were derived from an established matrix in cost-effectiveness studies for HF (Table 2) [7]. Patients could transition to a lower NYHA class, for example, from NYHA class II to class I, except post hospitalisations or readmissions where they either remained within the same NYHA class or progressed, for example, from NYHA class II to class III. These probabilities were fixed over time. Conservatively, the probabilities were assumed to be the same for both treatment groups, because it was unclear how sacubitril/valsartan altered the progression between health states relative to enalapril. Utility data Utilities describe the quality of life for different NYHA classes according to individuals’ preferences and range from 1 (perfect health) to 0 (death). They are combined with life expectancy to generate quality-adjusted life years (QALYs). The utilities for NYHA classes were derived from the Care-HF (Cardiac Resynchronization in Heart Failure) trial, in which healthrelated quality of life was captured in patients with chronic HF and reduced LVEF (≤35%) on optimal medical therapy with the European Quality of Life-5 Dimensions instrument [16]. Lower estimates for NYHA class IV (0.51) compared to class II (0.82) indicated poorer quality of life. A one-time disutility of 0.1 was applied, by subtracting patient’s utility in each NYHA class by 0.1, for each hospitalisation or readmission event [7,17]. Utilities were varied over their 95% CIs in sensitivity analyses.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
clinical practice. We used a time horizon of 10 years, based on the average life expectancy of 5 to 10 years for patients with HF [12,13]. The analysis was conducted from Singapore healthcare payer’s perspective. All costs and health outcomes were discounted at 3% annually. Model development and analyses were performed using TreeAge Pro SuiteTM software 2016 (Williamstown, MA. USA). Table 1 shows the list of model inputs.
ip
t
Costs of events were obtained from the national database with a case-mix of 6,663 patients with HF. The costs incurred during hospitalisations for HF of SGD 4,537 comprised doctor consultations, hospitalisations (average 5 days’ LOS), ICU stay (average 2 to 3 days’ LOS), procedures, investigations and laboratory tests. The costs for readmissions for HF of SGD 4,317 were similar to the index hospitalisation, given that clinical management is similar. The costs of in-patient deaths was higher, at SGD 7,170, because these patients required more intensive care and expensive treatment such as dialysis, on top of the costs of hospitalisation.
M
an
us
cr
Main outcome measures The main outcome measures were the number of first hospitalisations for HF and CV deaths prevented, total and incremental costs and QALYs, and the incremental cost-effectiveness ratio (ICER) for sacubitril/valsartan relative to enalapril. The ICER is a ratio of incremental healthcare costs to incremental health outcomes or QALYs gained and expressed in terms of cost per QALY gained. It is typically compared against a cost-effectiveness threshold that reflects willingness-to-pay (WTP). In Singapore, there is no fixed WTP threshold to determine cost-effectiveness. To address this, sensitivity analyses were conducted for various WTP thresholds of between SGD 20,000 and SGD 100,000 per QALY gained.
pt ed
Sensitivity analyses We performed one-way sensitivity analyses and generated Tornado diagram to present the relative impact of varying the model inputs on outcomes (i.e. ICERs). We conducted multivariate probabilistic sensitivity analysis (PSA), using 10,000 Monte Carlo simulations, to evaluate how the simultaneous uncertainties about model inputs across their distributions might influence outcomes.
ce
Results Base-case analysis Among 1,000 66-year-old patients with HF treated with enalapril, our model predicted 306 hospitalisations for HF and 426 CV deaths over 10 years. Relative to enalapril, sacubitril/valsartan prevented an additional 37 hospitalisations for HF and 65 CV deaths (Table 3). The effectiveness of sacubitril/valsartan in reducing hospitalisations for HF and CVdeaths translated into a mean gain of 0.21 QALYs per person. The total projected costs for enalapril and sacubitril/valsartan groups over the 10 years’ time horizon were SGD 2,197 and SGD 17,857, respectively. The additional mean cost of SGD 15,660 for treatment with sacubitril/valsartan was attributable to its substantially higher cost compared to enalapril (additional SGD 3,168 per surviving person annually, for 10 years), which was partially offset by fewer events in the sacubitril/valsartan group. These resulted in a base-case ICER for sacubitril/valsartan versus enalapril of SGD 74,592 (USD 55,198; based on SGD1 = USD0.74 as of September 2017) per QALY gained (Table 4).
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Resource use and cost data Only direct medical costs were used, in accordance with the model’s perspective. Medication costs were based on prices charged to patients at a tertiary national heart centre in Singapore for enalapril 10mg twice daily (SGD 6.00/month) and sacubitril/valsartan 200mg twice daily (SGD 270.00/month). These costs were varied widely (50% in each direction) in one-way sensitivity analyses to account for wide variations across different settings. The costs of additional HF medications and associated doctor visits were not considered; they were assumed to be equivalent between groups.
t
us
cr
ip
Additional analyses on cost of sacubitril/valsartan Given the cost of sacubitril/valsartan - which can be variable depending on country and timing - was among the top three drivers of cost-effectiveness, we performed additional analyses to examine the impact of pricing on model results (Figure 3). The daily cost of sacubitril/valsartan was set to SGD 9.00 in the base-case based on current price. The price would need to drop by 70%, 58% and 32% to SGD 2.70, SGD 3.80 and SGD 6.10, respectively, for sacubitril/valsartan to cost SGD 20,000, SGD 30,000 and SGD 50,000 per QALY gained.
pt ed
M
an
Probabilistic sensitivity analyses The average ICER per QALY gained from multivariate probabilistic sensitivity analyses, which took into account uncertainty of treatment effect in the Asian subgroup, was SGD 74,897 (USD 55,490). Out of 10,000 Monte Carlo simulations, sacubitril/valsartan were cost-effective below 1% of the time, at a WTP of SGD 20,000 or SGD 30,000 per QALY. This increased to 12% and 71% when WTP was increased to SGD 50,000 and SGD 100,000 per QALY gained, respectively (Figure 4).
ce
Model calibration We calibrated the model by comparing the model simulation results to that observed in PARADIGM-HF for both hospitalisations for HF and CV mortality. Using micro-simulation with tracker variables, our model produced estimates of hospitalisations for HF and CV death that compared well with the rates observed in PARADIGM-HF (Figure 5). These results suggest the model structure and baseline probabilities reflect current best estimates of the comparative effects of sacubitril/valsartan and enalapril on hospitalisation for HF and CV death rates.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Sensitivity analysis Deterministic sensitivity analysis showed that the effectiveness of sacubitril/valsartan in reducing the risk of CV death compared with enalapril was the key model driver. When the HR for CV death was varied across the 95% CI from 0.62 to 1.00 – the upper limit reflecting the non-statistically significant treatment effect in Asian subgroup - the ICER ranged from SGD 41,019 (USD 30,354) to SGD 1,447,103 (USD 1,070,856) per QALY gained. The modelled time horizon also had considerable impact; when the time horizon was shortened to the median follow-up duration of PARADIGM-HF trial (27 months), the cost per QALY increased to SGD 271,921 (USD 201,222). Conversely, modelling over a longer time horizon (20 years) resulted in an ICER of SGD 54,777 (USD 40,535). The cost of sacubitril/valsartan resulted in the third biggest impact on ICER, followed by background all-cause mortality rates and HR for hospitalisation for HF. Other variables, including costs of events or quality of life values had little impact on the cost-effectiveness of the treatment (< ±SGD 3,000 in terms of cost per QALY) (Figure 2).
Discussion As the final common pathway of a myriad of heart diseases, the burden of HF increases with the growing prevalence of CV disease. The World Health Organization has projected that the largest increases in CV disease worldwide are occurring in Asia. Sacubitril/valsartan offers an effective means to reduce the burden of HF. However, given the low-cost generic status of ACE inhibitors, decision-makers and payers need to determine whether the extra benefit with sacubitril/valsartan observed in PARADIGM-HF is worth the additional costs. This study is the first to examine the cost-effectiveness of sacubitril/valsartan versus enalapril, based on costs and resources used in Singapore, a developed country in Asia, to inform decision-making in healthcare. Our study found that relative to enalapril, sacubitril/valsartan prevented an additional 37
M
an
us
cr
ip
t
The key model driver was the effectiveness of sacubitril/valsartan in reducing the risk of CV death that was assumed to last over the time horizon of 10 years. Of note, the uncertainty in its treatment effect in the Asian subgroup translated into an ICER of SGD 1,447,103 103 (USD 1,070,856) per QALY gained in sensitivity analyses. However, subgroup analyses are inherently underpowered. On the other hand, the effectiveness of sacubitril/valsartan in reducing the risk of hospitalisations for HF did not have a significant impact on the model. HF is a condition with significant mortality so the modelled population declined over time due to increasing mortality. As a result, sacubitril/valsartan’s benefit for averting hospitalisations for HF diminished in the latter years. Correspondingly, the model simulated fewer hospitalisations for HF averted compared to CV deaths, and was less sensitive to impact of hospitalisation for HF. Accordingly, that the model did not incorporate endpoints such as reduction in emergency department readmissions for HF or ICU stays was unlikely to affect the cost-effectiveness results, given the relatively small impact of other benefits compared to CV death reduction [19].
ce
pt ed
These findings were consistent with recent cost-effectiveness analyses of sacubitril/valsartan conducted in the U.S. setting. King et al. found the cost-effectiveness of sacubtril/valsartan highly dependent on the duration of treatment effect, ranging from USD 249,411 per QALY at 3 years to USD 50,959 per QALY gained over a lifetime [7]. Gaziano et al. reported similar findings, with ICER of USD 135,964 per QALY gained when the benefit only lasting 27 months (trial duration) was modelled [8]. Sandhu et al. also reported ICER of USD 120,623 per QALY gained if the treatment was effective only for 27 months, with no subsequent reduction in event risks or improvement in quality of life but continued costs [9]. These analyses adopted a longer time horizon up to 30 years that inherently had greater uncertainty given that the effectiveness results were extrapolated from a trial with a duration of 2 years. None of these analyses applied treatment effect observed in the Asian population of PARADIGM-HF trial.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
hospitalisations for HF and 65 CV deaths over 10 years. Sacubitril/valsartan costs payers SGD 74,592 (USD 55,198) per QALY gained compared with enalapril in the Singapore healthcare setting. Locally, the healthcare system is a combination of government healthcare subsidies, mandatory healthcare savings accounts and health insurance. While there is no fixed WTP threshold, sacubitril/valsartan may not be considered cost-effective use of the finite healthcare budget, when benchmarking against other CV drugs in the Singapore healthcare setting with ICER of SGD 10,000 per QALY gained [18]. A price reduction of 32% to 70%, from current price of SGD 9.00/day, will be needed for sacubitril/valsartan to cost between SGD 20,000 and SGD 50,000 per QALY gained and be considered cost-effective compared with other drugs in similar therapeutic area. This will also improve the affordability for patients with HF who currently have no access to the drug due to cost barrier. Given the large and growing number of patients with HF, this will also ensure the sustainability of the healthcare system from a budgetary perspective.
What set our work apart from these studies was contextualising the cost-effectiveness analysis of sacubitril/valsartan in the local healthcare setting. Our modelled population reflected the average age of patients with HF in Singapore and incorporated the treatment effects using results from the Asian subgroup of PARADIGM-HF trial in sensitivity analyses. We used Singapore-specific all-cause mortality rates, with age-adjusted ratios derived from national epidemiological data to reflect higher mortality burden due to HF. Based on national epidemiological data, we applied 16% readmission rates for HF among hospitalised cases and 4% in-patient deaths due to CV causes within one month that were consistent with international figures [20]. We applied quality-of-life values for various NYHA classes; our utility inputs correlated well with the reported scores of 0.78 (SD 0.18) to 0.51 (SD 0.21) for mild to severe disease in patients with HF in a systematic review of utilities for CV disease [21]. Local
us
cr
ip
t
Our findings should be interpreted within the limitations of our analysis. Firstly, we extrapolated the treatment benefits of sacubitril/valsartan for reducing hospitalisations for HF and CV mortality from PARADIGM-HF over 10 years. It remains to be known whether the assumption of sustained treatment effect of sacubitril/valsartan over enalapril beyond trial and throughout the time horizon holds true. Second, the trial included an active run-in period so our results assumed that patients follow the recommended dose of each medication. Dose reduction is associated with higher risk for major CV events [22]. In fact, local clinical experts have provided feedback about hypotensive episodes in real-life patients who are unable to tolerate sacubitril/valsartan’s hemodynamic effects at target dose, especially the older and frailer ones. This early clinical experience mirrored PARADIGM-HF trial data where sacubitril/valsartan was associated with significantly higher rates of hypotension than enalapril (17.6% vs 12.0%) but they rarely required discontinuation of treatment. Other common AEs such as hyperkalemia and renal impairment were not modelled. Given that the discontinuation of study medications due to these events was numerically lower among patients taking sacubitril/valsartan (10.7%) than among patients taking enalapril (12.3%) in PARADIGM-HF, the results from our analysis may be conservative. Furthermore, the analysis did not factor in potential impact of chronic neprilysin inhibition with sacubitril/valsartan on the risk of developing dementia. Studies are currently ongoing to assess its effects on cognitive function.
ce
pt ed
M
an
Conclusions In conclusion, our study reports the long-term cost-effectiveness of sacubitril/valsartan versus enalapril for patients with HFrEF in Singapore healthcare setting. Benchmarking against the ICERs of other drugs in the same setting, sacubitril/valsartan may not represent good value for limited health care dollars at its current price of SGD 9.00/day. This highlights the cost-benefit trade-off that health care professionals and patients face when considering HF therapy. To date, it remains to be proven in the real-world setting whether the mortality reduction benefits in the trial translates into true clinical benefits and actual cost savings from avoiding deaths. Ultimately, the clinical decision on whether to use this novel agent should be individualised and based on shared decision-making that reflects treatment-related benefits, risks, costs and patient preferences.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
costs of hospitalisations and readmissions, however, were lower, at SGD 4,537 (USD 3,357) and SGD 4,317 (USD 3,195), respectively, compared to reported figures of USD 10,698 and USD 11,361 in the U.S [7]. Such differences in costs of care in each country highlight the importance of conducting local cost-effectiveness analyses.
t
ip
cr
us
an
M
pt ed
ce
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
References 1. Ambrosy AP, Fonarow GC, Butler J, et al. The global health and economic burden of hospitalizations for heart failure: lessons learned from hospitalized heart failure registries. J Am Coll Cardiol. 2014;63(12):1123-33. 2. Reyes EB, Ha JW, Firdaus I, et al. Heart failure across Asia: Same healthcare burden but differences in organization of care. International journal of cardiology. 2016;223:163-7. 3. Mentz RJ, Roessig L, Greenberg BH, et al. Heart Failure Clinical Trials in East and Southeast Asia: Understanding the Importance and Defining the Next Steps. JACC Heart failure. 2016;4(6):419-27. 4. Kristensen SL, Martinez F, Jhund PS, et al. Geographic variations in the PARADIGM-HF heart failure trial. European heart journal. 2016. 5. McMurray JJV, Packer M, Desai AS, et al. Angiotensin–Neprilysin Inhibition versus Enalapril in Heart Failure. New England Journal of Medicine. 2014;371(11):993-1004. 6. Yancy CW, Jessup M, Bozkurt B, et al. 2016 ACC/AHA/HFSA Focused Update on New Pharmacological Therapy for Heart Failure: An Update of the 2013 ACCF/AHA Guideline for the Management of Heart FailureA Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America. Journal of the American College of Cardiology. 2016. 7. King JB, Shah RU, Bress AP, et al. Cost-Effectiveness of Sacubitril-Valsartan Combination Therapy Compared With Enalapril for the Treatment of Heart Failure With Reduced Ejection Fraction. JACC Heart failure. 2016;4(5):392-402. 8. Gaziano TA, Fonarow GC, Claggett B, et al. COst-effectiveness analysis of sacubitril/valsartan vs enalapril in patients with heart failure and reduced ejection fraction. JAMA Cardiology. 2016. 9. Sandhu AT, Ollendorf DA, Chapman RH, et al. Cost-Effectiveness of Sacubitril-Valsartan in Patients With Heart Failure With Reduced Ejection Fraction. Annals of internal medicine. 2016. 10. Lee R, Chan SP, Chan YH, et al. Impact of race on morbidity and mortality in patients with congestive heart failure: a study of the multiracial population in Singapore. International journal of cardiology. 2009;134(3):422-5. 11. Leong KT, Goh PP, Chang BC, et al. Heart failure cohort in Singapore with defined criteria: clinical characteristics and prognosis in a multi-ethnic hospital-based cohort in Singapore. Singapore medical journal. 2007;48(5):408-14. 12. Allen LA, Yager JE, Funk MJ, et al. Discordance between patient-predicted and modelpredicted life expectancy among ambulatory patients with heart failure. Jama. 2008;299(21):2533-42. 13. Alter DA, Ko DT, Tu JV, et al. The average lifespan of patients discharged from hospital with heart failure. Journal of general internal medicine. 2012;27(9):1171-9. 14. Hoyle MW, Henley W. Improved curve fits to summary survival data: application to economic evaluation of health technologies. BMC Medical Research Methodology. 2011;11(1):139. 15. Yearbook of Statistics Singapore 2016: Population. Department of Statistics, Singapore. Cited February 2017. Available from URL: http://www.singstat.gov.sg/publications/publications_and_papers/reference/yearbook_of _stats.html. 16. Calvert MJ, Freemantle N, Cleland JG. The impact of chronic heart failure on health-related quality of life data acquired in the baseline phase of the CARE-HF study. European journal of heart failure. 2005;7(2):243-51. 17. Yao G, Freemantle N, Flather M, et al. Long-term cost-effectiveness analysis of nebivolol compared with standard care in elderly patients with heart failure: an individual patientbased simulation model. PharmacoEconomics. 2008;26(10):879-89.
t ip cr us an M pt ed ce Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
18. Chin CT, Mellstrom C, Chua TS, et al. Lifetime cost-effectiveness analysis of ticagrelor in patients with acute coronary syndromes based on the PLATO trial: a Singapore healthcare perspective. Singapore medical journal. 2013;54(3):169-75. 19. Packer M, McMurray JJ, Desai AS, et al. Angiotensin receptor neprilysin inhibition compared with enalapril on the risk of clinical progression in surviving patients with heart failure. Circulation. 2015;131(1):54-61. 20. Desai AS, Stevenson LW. Rehospitalization for Heart Failure. Predict or Prevent? Circulation 2012;126(4):501-6. 21. Dyer MTD, Goldsmith KA, Sharples LS, et al. A review of health utilities using the EQ-5D in studies of cardiovascular disease. Health and Quality of Life Outcomes. 2010;8:13. 22. Vardeny O, Claggett B, Packer M, et al. Efficacy of sacubitril/valsartan vs. enalapril at lower than target doses in heart failure with reduced ejection fraction: the PARADIGM-HF trial. European journal of heart failure. 2016;18(10):1228-34.
Table 1. Model inputs Model parameters
Base-case
Range
Intercept Log_scale 0.0904 0.00 0.0328 0.0197 0.0043 to 0.0080
Exponential
Hazard ratios for sacubitril/valsartan compared with enalapril Hospitalisation for HF 0.79
0.71 to 0.89
Log-normal
CV mortality
0.71 to 0.89
Log-normal
ce
M
an
us
0.781 to 0.850 0.693 to 0.749 0.551 to 0.629 0.412 to 0.605 -0.080 to -0.130
3.00 to 9.00 135.00 to 405.00
793.37 to 16,854,73 824.97 to 15,393.09 918.82 to 14,934.49
Extrapolated from published KaplanMeier curves from PARADIGM-HF trial PARADIGM-HF trial PARADIGM-HF trial Local data Local data
t
cr
pt ed
Monthly probabilities for sacubitril/valsartan and enalapril 30-day readmission for HF 0.16 30-day inpatient death due to 0.04 CV causes Utility inputs NYHA class I 0.815 NYHA class II 0.720 NYHA class III 0.590 NYHA class IV 0.508 Disutility for -0.100 hospitalisation/readmission Cost inputs (SGD) Cost of medication (monthly) Enalapril 6.00 Sacubitril/valsartan 270.00 Cost of events (one-time) Hospitalisation for HF 4,537.01 Readmission for HF 4,317.22 Inpatient deaths due to CV 7,170.00 causes
Weibul
Source
ip
0.80
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Clinical inputs Monthly probabilities for enalapril Hospitalisation for HF Intercept=5.2700 Cholesky matrix Log_scale=0.1792 Intercept Log_scale CV mortality 0.0067
Distributi on
Beta Beta Beta Beta Beta
CARE-HF CARE-HF CARE-HF CARE-HF CARE-HF
Local hospital Local hospital
Gamma Gamma Gamma
Local hospital Local hospital Local hospital
Table 2. NYHA transition probabilities per one-month cycle From NYHA
To NYHA Class II
Class III
Class IV
0.9923
0.0064
0.0013
0
Class II
0.0027
0.9936
0.0034
0.0003
Class III
0
0.0116
0.9864
0.0021
Class IV
0
0
0.0189
0.9813
ip
t
Class I
cr
Table 3. Projected clinical events per 1,000 patients with HF over 10 years’ modelled time horizon Hospitalisation for
Difference
CV deaths
HF
Sacubitril/ Valsartan
269
M
306
pt ed
Enalapril
an
Base-case
Difference
us
Treatments
--
426
--
-37
361
-65
Table 4. Projected costs, benefits and incremental cost-effectiveness ratio
Base-case Enalapril
Sacubitril/
Costs (SGD)
ce
Treatments
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Class I
2,197
17,857
Incremental cost (SGD)
--
15,660
QALY
Incremental QALYs
ICER (SGD/ QALY gained)
3.29
--
--
3.50
0.21
74,592
Valsartan ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life-year.
t
ip
or sacubitril/valsartan. Patients who experience clinical events are transitioned through
cr
the health states of New York Heart Association function class I, II, III, IV or death
ce
pt ed
M
an
us
every month.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Figure 1. Markov model structure. Patients enter the model and are on either enalapril
Figure 2. One-way sensitivity analyses (Tornado diagram). The change in incremental cost-effectiveness ratio of sacubitril/valsartan versus enalapril associated with varying key model inputs are shown. The vertical line represents the incremental costeffectiveness ratio in the base-case analysis. HF = heart failure; QALY = qualityadjusted life-year.
t
ip
cr
cost-effectiveness ratio of sacubitril/valsartan versus enalapril associated with varying
us
price of sacubitril/valsartan is shown. The daily price of sacubitril/valsartan used in the
ce
pt ed
M
an
base-case analysis is SGD 9.00.
Ac
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
Figure 3. Sensitivity analysis of cost of sacubitril/valsartan. The change in incremental
Figure 4. Probabilistic sensitivity analyses (cost-effectiveness acceptability curve). The curves show the proportion of the 10,000 model simulations at which sacubitril/valsartan was cost-effective across a range of thresholds. The analyses were performed by independently sampling each model input parameter from its distribution.
t ip cr us an M pt ed
Downloaded by [Australian Catholic University] at 15:45 30 September 2017
ce Ac
Figure 5. Model calibration for a) hospitalization for HF and b) CV death. The KaplanMeier estimates of hospitalization for HF and CV death from the PARADIGM-HF trial (in blue) are shown, along with observed estimates in the simulation model (in orange).
