Sacubitril valsartan for treating symptomatic chronic heart failure with reduced ejection fraction

Clinical effectiveness

4.2 The committee considered the current treatment pathway for people with chronic heart failure, and the position in the pathway for sacubitril valsartan. It noted that sacubitril valsartan has a marketing authorisation for 'the treatment of symptomatic chronic heart failure with reduced ejection fraction', and therefore includes both patients who have and have not previously had ACE inhibitors or angiotensin II receptor‑blockers (ARBs). The committee heard from the clinical experts that clinical practice is broadly in line with the NICE guideline on chronic heart failure in adults: diagnosis and management in that ACE inhibitors are the gold standard initial treatment and are taken concomitantly with a beta blocker and an aldosterone antagonist. The clinical experts stated that ARBs were also used in clinical practice for people who cannot take ACE inhibitors with concomitant beta blockers and an aldosterone antagonist. The committee discussed whether sacubitril valsartan would be given to patients who were newly diagnosed with chronic heart failure or only to those who were already taking an ACE inhibitor or an ARB. The committee heard from the clinical experts that many of patients with newly diagnosed heart failure may already be taking an ACE inhibitor for other conditions (for example hypertension). The committee heard from one of the clinical experts that people who are newly diagnosed with chronic heart failure but are already taking an ACE inhibitor or ARB would likely be offered sacubitril valsartan. However, the clinical expert explained that they would be reluctant to give sacubitril valsartan to people who had not previously had ACE inhibitors or ARBs because of the lack of evidence for clinical effectiveness and safety of sacubitril valsartan in this population. They also noted that in allowing the marketing authorisation for the drug to extend to patients who have not previously had ACE inhibitors or ARBs, the European Medicines Agency had recommended a lower starting dose. The committee agreed there was a lack of evidence in people who have not previously had ACE inhibitors or ARBs, noting that 99% of patients in the PARADIGM‑HF trial were taking ACE inhibitors or ARBs at entry to the study. The committee discussed comments it had received during consultation about the lack of clinical efficacy data from PARADIGM‑HF in people with chronic heart failure who have not previously had ACE inhibitors or ARBs. It noted that the a priori subgroup analyses of the PARADIGM‑HF trial, based on the time since diagnosis (less than 1 year, 1 to 5 years, and more than 5 years), and a post‑subgroup analysis for patients diagnosed less than 3 months previously did not show any statistically significant interaction between the subgroup demonstrating that the treatment effect would not vary depending up on the time since diagnosis (see section 3.64). However, the committee understood that the PARADIGM‑HF trial was designed to evaluate the efficacy of sacubitril valsartan in people who were already taking an optimised dose of ACE inhibitors or ARBs. It was not persuaded that the analysis showing early onset of treatment effect in the PARADIGM‑HF trial could be used as a proxy to demonstrate the effect in patients who have not previously had ACE inhibitors or ARBs (see section 3.65). The committee agreed that all these analyses included people who were already having optimum treatment for heart failure, and that the evidence of effectiveness within the first 30 days or in people who were recently diagnosed could not be used as a substitute for the lack of evidence in people who have not previously had ACE inhibitors or ARBs. The committee also agreed that in the absence of any trial data for this population, sacubitril valsartan's effectiveness remains uncertain. The committee concluded that sacubitril valsartan should only be offered, in place of ACE inhibitors or ARBs, to patients who are symptomatic despite already taking a stable dose of ACE inhibitors or ARBs.

4.3 The committee considered the generalisability of the PARADIGM‑HF trial results to people diagnosed with chronic heart failure with reduced ejection fraction in England. It noted that people in the trial were younger, included a higher proportion of men, were less likely to be using cardiac devices, and had greater tolerability to valsartan (in sacubitril valsartan). The clinical experts acknowledged these differences between the trial population and patients typically seen in clinical practice in England, and stated that the differences would not affect the way they prescribe sacubitril valsartan because the inclusion criteria used in the trial were common to all randomised trials in this disease area. The committee noted the consultation comment that baseline age in the PARADIGM‑HF trial was similar to that seen in other trials that established the effectiveness of ACE inhibitors, beta blockers, aldesterone antagonists, implantable cardiac devices and ivabradine in chronic heart failure. The clinical experts acknowledged that the dose of sacubitril valsartan was roughly twice the dose that would be normally tolerated in clinical practice, and that this suggested the treatment would be less effective in clinical practice than in the trial because of its dose‑dependent nature. However, the clinical experts commented that the dose of enalapril in the trial was also greater than would be typically observed in clinical practice, and that they would therefore expect these differences to cancel each other out, such that the relative treatment effect between sacubitril valsartan and ACE inhibitors in clinical practice to be similar to that in the trial. The committee noted comments from the ERG that the most appropriate choice of ACE inhibitor comparator was ramipril rather than enalapril. It heard from clinical experts that enalapril has the largest evidence base for its effectiveness, but that ramipril is more commonly used in clinical practice. The committee noted that the company had assumed a class effect for ACE inhibitors, based on the findings of a systematic review and network meta‑analysis (Chatterjee et al. 2013). It agreed a class effect for ACE inhibitors was an appropriate assumption, and that the choice between enalapril and ramipril therefore only affected the costs used in the economic modelling. The committee agreed that the generalisability was similar across all trials in this condition, and it concluded that, despite the differences between the trial and the trial eligible population in England, the results of the PARADIGM‑HF trial were relevant to established clinical practice in England.

4.4 The committee examined the clinical‑effectiveness evidence from PARADIGM‑HF comparing sacubitril valsartan with enalapril. The committee considered the PARADIGM‑HF trial was a good quality trial and that the relevant clinical outcomes of mortality and hospital admission were assessed. The committee noted that in the total trial population, the composite primary end point (death from cardiovascular causes or a first hospitalisation for worsening heart failure) significantly favoured sacubitril valsartan compared with enalapril (hazard ratio [HR] 0.80; 95% confidence interval [CI] 0.73 to 0.87, p<0.001). It heard from the clinical experts that such a benefit was considered to be clinically significant. The committee also noted that sacubitril valsartan was associated with a statistically significant benefit compared with enalapril in each of the separate components of the primary end point, and was also associated with a statistically significantly reduced risk of all-cause mortality (see section 3.7). The committee concluded that, for the population included in the PARADIGM‑HF trial, sacubitril valsartan was statistically significantly more clinically effective than enalapril at reducing hospitalisations and improving both overall mortality and cardiovascular mortality.

4.5 The committee considered the subgroup analyses presented by the company. It noted that the company had submitted a large number of prespecified subgroup analyses, and that across all groups sacubitril valsartan was consistently better than ACE inhibitors with regard to the primary end point. The committee was aware that the treatment effect for several subgroups did not reach statistical significance, including the Western Europe group (HR 0.89, 95% CI 0.74 to 1.07), the group who were aged 75 years or older (HR 0.86, 95% CI 0.72, 1.04), the group with New York Heart Failure (NYHA) class III or IV heart failure (HR 0.92, 95% CI 0.79, 1.08), and the group who had not previously had ACE inhibitors (HR 0.92, 95% CI 0.76 to 1.10). The committee understood that for all the subgroups, including the Western Europe subgroup which represented 24% of the total trial population, the comparisons were not powered to detect statistically significant differences in the primary end point, and that the hazard ratio point estimates all suggested a benefit in the sacubitril valsartan group. The committee heard from clinical experts that the lack of statistically significant outcomes among certain subgroups did not affect their assessment of the drug's effectiveness. The committee considered the tests of interaction carried out by the company which showed little evidence of treatment‑effect modifiers for most subgroups. The committee agreed that because the results of subgroup analyses were consistently positive, any differential interpretation of treatment effect in subgroups should be undertaken with caution. The committee noted that the ERG had considered the Western Europe subgroup to be the most representative of clinical practice in England based on race, age and cardiac device use at baseline for the Western Europe subgroup. Baseline characteristics of the Western Europe subgroup were designated academic in confidence by the company and therefore not reported here. The committee was aware that the treatment effect was not statistically significantly different compared with enalapril in this subgroup. The committee also noted the difference in the aetiology of chronic heart failure (ischaemic and non‑ischaemic), baseline risk for the mortality and the clinical management of heart failure between the different regions from which the population in the trial was recruited, and agreed that patients in the Western Europe subgroup were more comparable to patients in clinical practice in England than the total trial population. It concluded that the Western Europe subgroup was the most representative of clinical practice in England, and therefore the magnitude of effect for sacubitril valsartan generalisable to clinical practice in England will be closer to the results in the Western Europe.

4.6 The committee noted that there were no head‑to‑head trials comparing sacubitril valsartan with ARBs, and therefore considered the network meta‑analysis carried out by the company to estimate the relative treatment effect for sacubitril valsartan compared with ARBs. The committee noted that the results from the network meta‑analysis suggested that ARBs and ACE inhibitors were broadly equivalent, and that sacubitril valsartan was superior to ARBs with regard to all‑cause and cardiovascular mortality and broadly equivalent with regard to all‑cause hospitalisation. The committee considered the network meta‑analysis to be methodologically sound, noting that it used methods that were in line with the NICE decision support unit's technical support document 2. However, it was aware of the issues raised by the ERG with regard to heterogeneity in the trials underpinning the network, and with regard to the wide confidence intervals associated with the results of the network meta‑analysis. It understood that the company's network meta‑analysis reflected the approach taken by the Cochrane meta-analysis by Heran et al. (2012), and that both analyses had provided similar results. Overall, the committee concluded that although the results of the network meta‑analysis should be treated with caution, the consistency of findings between the network meta‑analyses by Heran et al. and the company provided sufficient reassurance that these results were valid, and were appropriate for the purposes of decision-making regarding the clinical effectiveness for sacubitril valsartan compared with ARBs.

4.7 The committee considered the adverse event profile associated with sacubitril valsartan compared with enalapril. It considered that the overall safety profiles during the double‑blind trial period of PARADIGM‑HF were comparable between the 2 treatment groups, and noted that there were no statistically significant differences with regard to discontinuations because of adverse events. The committee noted that the sacubitril valsartan group had statistically significantly higher rates of hypotension than the enalapril group, with a particularly large hazard ratio of 1.48. The committee considered the potential consequences of the increased rate of hypotension, for example injuries from falls, particularly because people in clinical practice are generally older than the trial population. However, the committee understood that hypotension was related to the greater vasodilator effect of sacubitril valsartan, and noted that there was no increase in the rate of discontinuation in the trial because of possible hypotension‑related adverse events. The committee noted the consultation comment from one of the clinical experts advocating a judicious use of sacubitril valsartan in line with the inclusion criteria of PARADIGM‑HF trial. The main concern raised by the clinical expert was the lack of long‑term data on its effect on cognitive function and a higher incidence of angioedema in people of African family origin. The committee heard from the company that these safety aspects are covered in the risk management plan agreed by the European Medicines Agency and the company. The committee concluded that sacubitril valsartan had a manageable adverse event profile in the population specified in the marketing authorisation.

4.8 The committee explored at what left ventricular ejection fraction level sacubitril valsartan could be considered an appropriate treatment for chronic heart failure. It was aware that the 'reduced ejection fraction' was not specified in the marketing authorisation and that the left ventricular ejection fraction entry criterion for the PARADIGM‑HF trial was changed from 40% or less to 35% or less. The committee heard from the company that improvement in clinical care, attributed to increased use of aldosterone antagonists, had reduced the baseline risk for cardiovascular mortality and hospitalisation. Therefore, the cut‑off was lowered from 40% to 35% in the trial to offset this anticipated decrease in the event rates for the outcomes. The committee noted that subgroup analyses based on baseline left ventricular ejection fraction showed no statistically significant interaction in the relative treatment effect of sacubitril valsartan over enalapril among these subgroups. However, given that most of the patients in the trial had left ventricular ejection fraction of 35% or less, the committee agreed that there was weaker evidence of clinical effectiveness in people with left ventricular ejection fraction of more than 35% to 40%. Furthermore, the lower event rate for people with the higher left ventricular ejection fraction of more than 35% to 40%, acknowledged by the company, meant less potential to benefit from sacubitril valsartan and implied a smaller incremental QALY gain and therefore poorer cost effectiveness (see section 4.20). The committee discussed how the left ventricular ejection fraction level will be determined in clinical practice and whether the necessary tests will be readily available to people who may benefit from sacubitril valsartan. It understood that left ventricular ejection fraction level is usually established with an echocardiogram and additional tests will not necessarily be needed before starting treatment with sacubitril valsartan. The committee noted the consultation comments that there are variations with respect to the reporting of echocardiogram assessments of left ventricular function, but noted that a cut-off of 35% or less has been used in other NICE guidance in this disease area and agreed that this was accepted as an indicator of severe left ventricular dysfunction in clinical practice. The committee concluded that sacubitril valsartan should only be given to people with a left ventricular ejection fraction of 35% or less, normally shown on an echocardiogram.

4.9 The committee explored in which NYHA classes sacubitril valsartan could be considered an appropriate treatment option for chronic heart failure. It was aware that NYHA class was not specified in the marketing authorisation for sacubitril valsartan. The committee noted that the inclusion criteria for the PARADIGM‑HF trial specified patients with NYHA class II to IV, but that most patients in the trial had NYHA class II to III (0.7% [60 patients] had NYHA class IV heart failure). The committee agreed that the representation of patients with NYHA class IV was limited and that the effectiveness of sacubitril valsartan was imprecise (associated with a wide confidence interval) as expected because of the small number of patients in this subgroup. The committee noted the consultation comments that it would be clinically counterintuitive not to recommend sacubitril valsartan in this patient group when it was recommended in patients with milder symptoms of chronic heart failure (NYHA classes II and III). The committee noted the limited adverse events data presented by the company comparing patients with NYHA class II, III and IV symptoms at baseline (see section 3.63) and was reassured by the comparable safety profile of sacubitril valsartan in these groups. The committee understood that it is difficult to recruit severely symptomatic patients, as evident by the low number of NYHA class IV patients in PARADIGM‑HF and other trials in chronic heart failure. Overall, the committee was persuaded that sacubitril valsartan was an appropriate treatment option in people with NYHA class II, III or IV chronic heart failure.

4.10 The committee explored whether a high plasma B‑type natriuretic peptide (BNP) level was necessary for sacubitril valsartan to be considered an appropriate treatment option for chronic heart failure. The committee considered the inclusion criteria for the PARADIGM‑HF trial, noting that they specified a high BNP level of at least 150 pg per ml (or an N‑terminal pro‑B‑type natriuretic peptide [NTproBNP] level of at least 600 pg per ml). It also noted that the inclusion criteria specified that if the patient had been hospitalised within the last 12 months because of heart failure, a slightly lower cut-off BNP (100 pg per ml, or NTproBNP 400 pg per ml) was accepted. The committee understood that BNP is a marker for heart failure, used in the diagnosis of chronic heart failure as well as monitoring the response to treatment. However, it is not specific to heart failure and in the acute setting is mostly used to rule out the presence of acute heart failure in the presence of symptoms like breathlessness. The committee noted that because of its mechanism of action as a neprilysin inhibitor, sacubitril valsartan will affect BNP level, and also that the summary of product characteristics does not recommend BNP as a suitable biomarker of heart failure in patients having sacubitril valsartan. The committee considered a consultation comment that compared the mean baseline BNP level in the trial population (225 pg per ml) with the mean value in stable patients as reported in the literature (93 pg per ml), along with commentary suggesting that sacubitril valsartan may not be as effective in patients with lower BNP. The committee heard from the company that the mean BNP level for people in secondary care reported in Hull Heart Failure Registry was comparable to the baseline value seen in the PARADIGM‑HF trial. The committee understood that high BNP level in patients with chronic failure is an indicator of ventricular dysfunction and although there is no direct correlation between reduced left ventricular ejection fraction and high BNP level, they tend to coexist. The committee noted that the NICE guideline on chronic heart failure in adults: diagnosis and management makes specific recommendations about using BNP or NTproBNP in diagnosing and managing chronic heart failure. It concluded that BNP need not be further specified when considering starting a patient on sacubitril valsartan in addition to reduced left ventricular ejection fraction (see section 4.8).

4.11 The committee considered how sacubitril valsartan will be prescribed in clinical practice. It heard from clinical experts that a heart failure specialist in secondary care with access to a multidisciplinary team should initiate sacubitril valsartan. The committee noted that the NICE guideline on chronic heart failure in adults: diagnosis and management defined a specialist as a physician with a subspecialty interest in the management of heart failure and who leads a specialist multidisciplinary heart failure team of professionals with appropriate competencies from primary and secondary care. The committee also noted that the inclusion criteria of the trial specified that patients must have been taking a stable dose of an ACE inhibitor or an ARB for at least 4 weeks before entering the study. It recalled its previous discussions (see section 4.2) that that there was a lack of evidence for sacubitril valsartan in people who had not previously had ACE inhibitors or ARBs, and it heard from clinical experts that sacubitril valsartan would only be considered for people who are already receiving a stable, optimised dose of an ACE inhibitor or an ARB. The committee concluded that sacubitril valsartan should be started by a heart failure specialist (in line with the NICE guideline) with access to a multidisciplinary heart failure team, in people who are receiving a stable, optimised dose of an ACE inhibitor or an ARB. The committee acknowledged the consultation comments highlighting the lack of available GPs with a special interest in heart failure and heart failure specialist nurses in the community. It further noted comments suggesting that the dose monitoring and titration could be done by GPs who may not have a special interest in heart failure. The committee noted that NICE's guideline on chronic heart failure in adults: diagnosis and management recommended that the multidisciplinary heart failure team should decide regarding who is the most appropriate team member to address a particular clinical problem. Therefore, the committee concluded that dose titration and monitoring should then be done by a heart failure specialist or in primary care by the most appropriate team member, as specified in NICE's guideline on chronic heart failure in adults: diagnosis and management.

Cost effectiveness

4.12 The committee discussed the company's economic model and the ERG's critique of this model. It heard from the clinical experts that the model captured the outcomes that were clinically relevant to chronic heart failure. The committee considered the company's model to be generally well structured, a good predictor of the PARADIGM‑HF trial outcomes and noted it was of a similar structure to those previously published, including the model submitted during the development of NICE technology appraisal guidance on ivabradine for treating chronic heart failure. The committee noted that the company's model made use of patient‑level data in the base‑case analysis, and that the ERG had considered this was not completely justifiable (see section 3.43). It understood that the company had developed the model allowing the user to run it using average patient characteristics in each cohort, and that the ERG's exploratory analyses had been carried out using this alternative approach. The committee considered there were advantages and disadvantages for both the patient‑level and the cohort‑model approaches, and it was aware that similar model outcomes were observed for both modelling approaches. The committee therefore considered that both the cohort model using average patient characteristics and the patient‑level approach were acceptable. The committee concluded that the company's model was sufficiently robust for assessing the cost effectiveness of sacubitril valsartan.

4.13 The committee considered the age of patients entering the economic model. It noted that the mean baseline age in the company's base case (64 years) reflected the PARADIGM‑HF trial. It was aware that in exploratory analyses the ERG had adjusted the model to a mean baseline age of 75 years, and agreed that this more closely reflected the age of patients generally seen in clinical practice. It discussed the ERG's concerns that the modelling approach taken by the company resulted in an inflexible economic model, and that despite its adjustment to the baseline age, the model could not be changed to portray an older population at baseline in order to generalise the model results. It understood that the model was accurate in replicating the trial data, but that there were issues of face validity (such as 21‑year olds having the same life expectancy as 87‑year olds; see section 3.48). The committee heard from clinical experts that these findings could not be explained from a clinical perspective, and the committee agreed that there was some uncertainty as to whether the ERG's additional analysis in 75‑year olds was fully reflective of the true cost effectiveness of sacubitril valsartan in an older population. The committee was aware that this issue was a result of the economic model being structured to closely reflect the population and outcomes from the PARADIGM‑HF trial, and further recalled its concern about the sacubitril valsartan dose used in the trial. Nevertheless, the committee concluded that, despite the inflexibility of the company's economic model and the resulting constraints in generalising the model results to portray an older population, the ERG's use of a baseline age of 75 years was a reasonable attempt to generalise the model results to the heart failure population in England, and was appropriate for the purposes of decision‑making.

4.14 The committee considered the population used in the economic model. It noted that the company had used the results of the full analysis set population to inform its model, and that in exploratory analyses the ERG had used only the company's Western Europe subgroup analysis results. The committee recalled its earlier conclusions regarding the generalisability of clinical effectiveness that the Western Europe subgroup was the most representative of clinical practice in England (see section 4.5), and it therefore considered the use of the results closer to this subgroup more appropriate for the cost‑effectiveness analyses comparing sacubitril valsartan with ACE inhibitors and ARBs.

4.15 The committee considered the modelling of health‑related quality of life. It noted that the company used a linear mixed regression model based on EQ‑5D trial data from PARADIGM‑HF to predict utility scores. It further noted that the company had assumed a small but statistically significant EQ‑5D treatment effect in favour of sacubitril valsartan even after controlling for the effects of hospitalisations and adverse events. The committee discussed whether using a utility benefit of 0.011 in patients having sacubitril valsartan over and above those generated by the difference in these outcomes was reasonable. The committee considered the ERG's concerns, in particular that the trial (and consequently the model outcomes) could potentially be biased in favour of the sacubitril valsartan group if, for example, patients in this group had a better quality of life at baseline, and this healthier state may be carried through to the trial and result in better outcomes. The committee noted that in exploratory analyses the ERG had explored changing the baseline utility value to reflect the utility value in the publication by Berg et al., and it had adopted a simplified quality of life modelling approach linked to the incidence of adverse events, hospitalisation events and disease progression (see section 3.56). The committee heard from clinical experts that hospitalisation rates were a good surrogate for determining patients' quality of life, and it understood that the ERG's simplified approach adequately captured the effect of reduced hospitalisation. It also noted that the ERG had not included the absolute utility increment (0.011) with sacubitril valsartan because of concerns regarding the difference in the baseline EQ‑5D values between the 2 arms of the PARADIGM‑HF trial (see section 3.55). The committee heard from the company that it had adjusted for any imbalance in the baseline EQ‑5D values using an analysis of covariance (ANCOVA). The committee agreed that there may have been other benefits due to symptomatic improvement of the chronic condition that may not have been captured in the ERG's analyses, and so was persuaded that some additional utility benefit for patients having sacubitril valsartan – as suggested by the company – could be accepted for the purposes of decision‑making.

4.16 The committee noted that the company had chosen to model enalapril as the ACE inhibitor comparator although ramipril is more commonly used in clinical practice (see section 4.3). The committee noted that in its exploratory analyses, the ERG had included the cost of ramipril and assumed drug doses for ramipril that reflected the way it is given in clinical practice in the UK. This had only a modest effect on the cost-effectiveness results, but the committee agreed that the use of ramipril costs rather than enalapril costs more appropriately reflected clinical practice in England.

4.17 The committee noted that the company's assumptions regarding the daily drug doses were not consistent across different treatments (see section 3.52). The ERG carried out exploratory analyses to reflect a consistent approach using target doses for estimating drug costs. The committee noted that this change had almost no effect on the cost‑effectiveness results, but concluded that a consistent approach to the use of drug doses was more appropriate to inform its decision‑making.

4.18 The committee considered the incremental cost‑effectiveness ratios (ICERs) presented by the company for sacubitril valsartan compared with ACE inhibitors, as well as the ERG's exploratory analyses. It noted that the company's base-case deterministic ICER for sacubitril valsartan compared with ACE inhibitors was £18,000 per quality‑adjusted life year (QALY) gained (incremental costs £7,514, incremental QALYs 0.42). The committee noted that the company had done a number of scenario analyses that had shown the ICER was relatively robust to the changes explored. The committee then considered the ERG's exploratory analyses. It noted that the ERG's exploratory analyses, including all of its preferred parameters or assumptions (see sections 4.10 to section 4.14), resulted in a deterministic ICER for sacubitril valsartan compared with ACE inhibitors of £29,500 per QALY gained (incremental costs £6,841, incremental QALYs 0.33). The committee was aware that the ERG had considered its exploratory analyses to be associated with significant uncertainty because of the lack of generalisability of the results from the PARADIGM‑HF trial and the lack of statistical significance associated with the Western Europe subgroup. The committee recalled its conclusion that the outcome expected to be seen in clinical practice in England would be closer to the results for the Western European group than the overall trial population (see section 4.5). The committee agreed that it would accept the central estimate of effect for the Western European subgroup. The committee recalled its conclusion that the company's approach to modelling utility benefit with sacubitril valsartan was plausible (see section 4.15) and noted that during consultation the company, applying all of the ERG's assumptions with the exception of its approach to modelling utility values, estimated an ICER of about £26,000 per QALY gained for sacubitril valsartan compared with ACE inhibitors (see section 3.70).

4.19 The committee considered the ICERs for sacubitril valsartan compared with ARBs in people who cannot have an ACE inhibitor. It noted that the company's base-case deterministic ICER for sacubitril valsartan compared with ARBs was around £16,500 per QALY gained (incremental costs £8,513, incremental QALYs 0.52). The committee was mindful of its earlier conclusions that the results of the network meta‑analysis were appropriate for the purposes of its decision‑making (see section 4.6). It noted that for this analysis the ERG had presented equivalent exploratory analyses to the comparison with ACE inhibitors, resulting in an ICER of £30,100 per QALY gained. The committee did not have an exact figure for the ICER underpinning its preferred assumptions, which included all assumptions in the ERG's preferred scenario except the additional utility benefit with sacubitril valsartan. However, from the effect of these assumptions on the ICER for sacubitril valsartan compared with ACE inhibitors (see section 4.18), the committee concluded that the most plausible ICER for sacubitril valsartan compared with ARBs in people who cannot have an ACE inhibitor would be less than £30,000 per QALY gained.

4.20 The committee considered whether sacubitril valsartan was a cost‑effective use of NHS resources. It was aware that the ICERs for the comparisons of sacubitril valsartan with ACE inhibitors and with ARBs were at the upper end of the range that would normally be considered a cost‑effective use of NHS resources (£20,000 to £30,000 per QALY gained). The committee was also aware that NICE's guide to the methods of technology appraisal states that above a plausible ICER of £20,000 per QALY gained, judgements about the acceptability of the technology as an effective use of NHS resources will specifically take account of a number of other factors, including the innovative nature of the technology. The committee recognised the innovative nature of sacubitril valsartan in that the inhibition of neprilysin is a novel development in the pharmacological management of heart failure. The committee also considered comments from the clinical and patient experts that this is a disease area that has been historically underinvested. In addition, the committee was aware that sacubitril valsartan has been granted an Early Access to Medicines Scheme positive opinion by the Medicines and Healthcare Products Regulatory Agency. The committee concluded that sacubitril valsartan was innovative and that it offered a small step-change in the management of this condition. The committee considered that, given its innovative nature, the most plausible ICERs for sacubitril valsartan – between £26,000 and £30,000 per QALY gained – could be considered to represent a cost‑effective use of NHS resources. However, it acknowledged that there were considerable uncertainties in the data including: non‑significant results in the Western European subgroup; a high sacubitril valsartan dose that may not be replicable in clinical practice; limited representation of patients with NYHA class IV symptoms; the weaker evidence (as well as lower event rate and therefore less potential to benefit) for patients with a left ventricular ejection fraction of over 35%; and the absence of direct evidence for patients who have not previously had ACE inhibitors or ARBs. These latter 2 groups (people with a left ventricular ejection fraction of over 35% and people who have not previously had ACE inhibitors or ARBs) could not be assumed to have the same cost‑effectiveness level as that modelled for the core trial population. The committee therefore considered that its recommendations should closely reflect the core population in the PARADIGM‑HF trial and agreed that people with NYHA class IV should not be excluded from the recommendations. The committee concluded that sacubitril valsartan is a cost‑effective use of NHS resources for treating chronic heart failure with reduced ejection fraction, only in people with NYHA class II to IV chronic heart failure, who are already taking a stable dose of ACE inhibitors or ARBs and have a left ventricular ejection fraction of 35% or less.

4.21 The committee discussed whether there were any equality issues it should consider before making its recommendations. It noted the comments received during consultation had stated that there were higher rates of angio‑oedema in those of African family origin having ACE inhibitors, and that extra vigilance would be needed because of the low numbers of these patients included in the trial (5%). Mindful of its recommendations for sacubitril valsartan, the committee concluded that there was no unfairness or unlawful discrimination and no need to alter or add to its recommendations.

4.22 The committee considered whether it should take into account the consequences of the 2014 Pharmaceutical Price Regulation Scheme (PPRS), and in particular the PPRS payment mechanism, when appraising sacubitril valsartan. The committee noted NICE's position statement in this regard, and accepted the conclusion 'that the 2014 PPRS payment mechanism should not, as a matter of course, be regarded as a relevant consideration in its assessment of the cost effectiveness of branded medicines'. The committee heard nothing to suggest that there is any basis for taking a different view on the relevance of the PPRS to this appraisal. It therefore concluded that the PPRS payment mechanism was not applicable when considering the cost effectiveness of sacubitril valsartan.

Summary of appraisal committee's key conclusions