4 Consideration of the evidence
The appraisal committee reviewed the data available on the clinical and cost effectiveness of evolocumab, having considered evidence on the nature of primary hypercholesterolaemia (heterozygous-familial and non-familial) or mixed dyslipidaemia and the value placed on the benefits of evolocumab by people with the condition, those who represent them, and clinical experts. It also took into account the effective use of NHS resources.
4.1 The committee considered the aim of treating primary hypercholesterolaemia or mixed dyslipidaemia. It heard from the patient experts that having primary hypercholesterolaemia or mixed dyslipidaemia affects day-to-day life, impinging also on family and friends. The committee was aware that the recommendations in NICE's guideline on lipid modification place greater emphasis on managing cardiovascular risk than meeting target cholesterol concentrations, although cholesterol targets are routinely used in clinical practice. The committee understood from the clinical experts that treating people with familial hypercholesterolaemia is a priority because lifelong exposure to high concentrations of low-density lipoprotein cholesterol (LDL‑C) increases the risk of cardiovascular disease (CVD), even if these concentrations are not very high. The committee heard from patient experts that in their experience diet and lifestyle changes, in addition to medication, were important to lose weight and further reduce the risk of CVD. The committee concluded that in clinical practice, lowering LDL-C concentrations in people with primary hypercholesterolaemia or mixed dyslipidaemia aims primarily to prevent CVD, as recommended in the NICE guideline on lipid modification.
4.2 The committee noted that the scope for this appraisal included people with primary hypercholesterolaemia (heterozygous-familial and non-familial) or mixed dyslipidaemia for whom lipid-modifying therapies, in line with current NICE guidance, would be considered. This was consistent with the marketing authorisation for evolocumab, which recommends treatment, as an adjunct to diet, for primary hypercholesterolaemia (heterozygous-familial and non-familial) or mixed dyslipidaemia. The committee discussed whether evolocumab would be used for mixed dyslipidaemia, which is also associated with elevated LDL‑C concentrations. It understood from the clinical experts that although evolocumab was likely to mainly be used for primary hypercholesterolaemia in clinical practice, it may also be used for mixed dyslipidaemia when LDL-C concentrations remain very high despite maximum statin and ezetimibe management.
4.3 The committee considered the current treatment pathway and comparators for people with primary hypercholesterolaemia. It was aware that statins (particularly atorvastatin) are the mainstay of treatment for familial and non-familial hypercholesterolaemia (as described in NICE's guideline on familial hypercholesterolaemia and on lipid modification), but that some people may not tolerate them. The committee noted that NICE technology appraisal guidance on ezetimibe for treating primary heterozygous-familial and non-familial hypercholesterolaemia recommends ezetimibe monotherapy for primary hypercholesterolaemia when statin therapy is contraindicated or not tolerated. It also recommends ezetimibe, co-administered with initial statin therapy, when cholesterol levels are not low enough, even when the dose is increased, or if a person is unable to tolerate higher doses of the statin. The committee concluded that ezetimibe is used to treat primary hypercholesterolaemia (heterozygous-familial and non-familial) in adults who are unable to have a statin, or need to supplement statin therapy, and therefore was an appropriate comparator for evolocumab in this appraisal.
4.4 The committee discussed the prevalence of statin intolerance among people with primary hypercholesterolaemia, noting that in NICE's guideline on familial hypercholesterolaemia intolerance to initial statin therapy was defined as 'the presence of clinically significant adverse effects from statin therapy that are considered to represent an unacceptable risk to the patient or that may compromise adherence to therapy'. The committee understood from the clinical experts that absolute statin intolerance is rare, accounting for less than 5% of people. However, the clinical experts pointed out that up to 25% of people may have some degree of muscle pain that they perceive as being caused by statin intolerance. The committee concluded that absolute statin intolerance was in line with the definition in NICE's guideline on familial hypercholesterolaemia, and that only a few people would have it.
4.5 The committee discussed the clinical situations in which treatment with evolocumab would be started. It heard from the clinical experts that evolocumab would be used in people with a high clinical unmet need such as people with severe forms of heterozygous-familial hypercholesterolaemia, and those who cannot tolerate statins and who are benefitting only marginally from ezetimibe, which will leave them with a high residual risk of CVD. For these people the only option was lipoprotein apheresis, which has its disadvantages (see section 4.6), and so evolocumab would be a welcome alternative. The committee heard that clinicians were likely to use evolocumab as an add-on, rather than an alternative, to statins and ezetimibe because statins and ezetimibe have well-established safety data, are not expensive, and have been robustly shown to improve cardiovascular (CV) outcomes. The committee concluded that, in clinical practice, evolocumab was likely to be reserved for people who are at a high risk of CVD as an add-on to statins and ezetimibe.
4.6 The committee discussed the place of lipoprotein apheresis in managing primary hypercholesterolaemia. The committee was aware that, although apheresis is recommended in the NICE guideline on familial hypercholesterolaemia as an option for severe heterozygous-familial hypercholesterolaemia, it is not only costly and onerous for the patient, but also difficult to access because only a few centres offer it. The committee noted that current guidelines recommend lipoprotein apheresis for patients with heterozygous-familial hypercholesterolaemia or other forms of severe hypercholesterolaemia and with progressive coronary heart disease whose low-density lipoprotein cholesterol (LDL‑C) remains above 5.0 mmol/litre or decreases by less than 40% on maximally tolerable doses of combined drug therapy. The committee concluded that treatments that avoid the need for lipoprotein apheresis would be welcomed.
Clinical effectiveness
4.7 The committee considered the randomised controlled trials (RCTs) for evolocumab, noting that the company's submission focussed on the results for LDL-C. It was aware that 2 of the 4 RCTs (LAPLACE‑2 and GAUSS‑2) compared evolocumab directly with ezetimibe, the sole comparator treatment in the scope. However, this was only for the non-familial hypercholesterolaemia population, and none of the trials compared evolocumab with ezetimibe for heterozygous-familial hypercholesterolaemia. RUTHERFORD‑2 and GAUSS‑2 studied evolocumab in 2 subgroups defined in the scope: people with heterozygous-familial hypercholesterolaemia, and those who cannot tolerate statins. The committee was aware that none of the trials studied evolocumab plus ezetimibe in any population. The committee agreed that the RCTs were otherwise relevant, and of good quality. It concluded that the trials were suitable for assessing the clinical effectiveness of evolocumab for primary hypercholesterolaemia (non-familial and heterozygous-familial).
4.8 The committee discussed whether the RCTs for evolocumab represented people who present with primary hypercholesterolaemia in clinical practice in England. It noted that the trials did not include some people with diabetes (see section 3.5), who may also have primary hypercholesterolaemia or mixed dyslipidaemia. The clinical experts did not consider this to have affected the generalisability of the trial results because, in clinical practice, people with diabetes would have their blood glucose levels controlled before being treated for primary hypercholesterolaemia. In general, the clinical experts agreed that the trials included people who reflected those with primary hypercholesterolaemia seen in clinical practice in England. The committee concluded that the trial results could be generalised to clinical practice.
4.9 The committee noted that at both dosages (140 mg every 2 weeks and 420 mg monthly), evolocumab effectively reduced LDL‑C by 60–70% compared with placebo, and around 40% compared with ezetimibe. Consistent results were seen across high-risk subgroups including people with heterozygous-familial hypercholesterolaemia, and those who cannot tolerate statins. The committee also noted that evolocumab was well tolerated. The committee concluded that, compared with placebo or ezetimibe, evolocumab was clinically effective in reducing LDL‑C in people with primary hypercholesterolaemia.
4.10 The committee discussed the effect of evolocumab on CVD in people with primary hypercholesterolaemia. It noted that the RCTs primarily measured surrogate end points (such as LDL‑C), and were not powered to measure CV outcomes, which the committee considered to be an important limitation of the evidence base. The committee was aware that the reduction in CV events with statins has been confirmed by many large RCTs. By contrast, adding other lipid-modifying drugs to statins has not consistently been shown to further decrease CV events. The clinical experts generally considered LDL‑C to be a reasonable surrogate for future CV events, although they advised that this relationship was uncertain when the LDL‑C concentration at baseline is low (below 2.0 mmol/litre). The committee understood that the currently accepted relationship between LDL-C and CV events is based on the Cholesterol Treatment Trialists' (CTT) meta-analysis of statin trials. Further data on the benefit of non-statins on CVD came from RCTs of ileal bypass surgery (POSCH), and recently of ezetimibe (IMPROVE-IT), which showed that when ezetimibe was added to a statin, this further reduced CV events compared with statins alone. The committee noted the consultation comments suggesting that the association between reduced LDL‑C concentrations and improved CV outcomes was well established and shown in many clinical trials. The committee understood that evolocumab should have a beneficial effect on CV outcomes because it has the same ultimate mechanism for LDL‑C reduction as statins. The committee noted the data from OSLER and OSLER‑2 on CV events (see section 3.9). However, it considered that these data were based on an exploratory analysis with few events, and were yet to be validated in larger trials. The committee noted that an ongoing RCT, FOURIER, would test whether or not LDL‑C is a valid surrogate for CV outcomes for evolocumab. It agreed that this trial would give useful data on the direct effect of evolocumab on CVD, and recommended that the consideration of the review of the guidance is scheduled so that the results of FOURIER could be taken into account. The committee concluded that, although it was reasonable to infer that evolocumab would reduce CVD, the extent of this reduction was still uncertain, particularly with low concentrations of LDL‑C at baseline.
4.11 The committee discussed the long-term effects of evolocumab. It heard from the clinical experts that the treatment effect could gradually weaken over time, and more likely so when people start treatment with relatively low LDL‑C concentrations, but the committee agreed that evolocumab is only likely to be targeted to people with LDL‑C concentrations at the high end of the spectrum. However, the committee also noted the statement from clinical experts suggesting that with evolocumab, there is a theoretical potential for neutralising antibodies to develop and for treatment to lose its effectiveness, although there was no positive evidence that this would be the case. The committee was aware that long-term data were limited, but what data there were did not show that evolocumab may lose its effect over long treatment durations because of neutralising antibodies. The committee heard from the company that in an integrated safety analysis of more than 6000 patients (representing 7,235 patient years of exposure), anti-evolocumab antibodies were infrequent, non-neutralising, and not associated with clinically relevant adverse events. However, the committee did not consider this analysis to have followed up people for long enough to draw firm conclusions about the long-term effect of evolocumab. Without robust, long-term data the committee could not ascertain whether the effect of evolocumab would be maintained over time at the same level as when therapy was started, although the limited evidence available suggested that any loss of effect would be likely to be infrequent.
Cost effectiveness
4.12 The committee considered the company's model, noting that this reflected the effect of evolocumab on LDL-C, but not on other lipid parameters, including those implicated in mixed dyslipidaemia. The evidence review group (ERG) was concerned about the overall structure of the model, and in particular the 13 composite states, which it considered to be based on many arbitrary assumptions and little evidence (see section 3.24). The committee agreed that the composite states reflected specific combinations of CV events, which were unlikely to be robustly modelled given the existing evidence. The committee acknowledged the company's response to consultation suggesting that the composite states prevented clinically implausible scenarios from occurring in the model, and used assumptions that were endorsed by expert opinion. Although the committee appreciated the logic of using the composite states, it concluded that, without evidence to support the modelling of these states, the internal validity of the model was unclear.
4.13 The committee discussed the modelled populations (that is, non-familial hypercholesterolaemia and heterozygous-familial hypercholesterolaemia). It understood that the company initially assumed that people who can tolerate statins have the same characteristics as those who cannot, although the risk of CVD is likely to be affected by whether or not the person can tolerate statins. In response to consultation, the company used GAUSS‑2, which only included people who could not tolerate statins, to model people with non-familial hypercholesterolaemia who cannot tolerate statins. The committee concluded that it was more appropriate to model people with non-familial hypercholesterolaemia who can or cannot tolerate statins separately.
4.14 The committee discussed the heterozygous-familial hypercholesterolaemia population for which cost-effectiveness results were originally presented (see section 3.39). It noted that the company had modelled patients with or without CVD together. The committee heard from the clinical experts that, in clinical practice, people with CVD are treated more intensively than those without, and so it would be useful to separate the results for each of these groups. In response to consultation, the company split the heterozygous-familial hypercholesterolaemia population by whether or not people had CVD (see section 3.45). The committee concluded that the company's revised analyses more appropriately reflected clinical practice.
4.15 The committee discussed how the company estimated the risk of CVD in the model. It noted the ERG's comment that several assumptions and adjustments were needed to predict the risk of CVD before treatment, even though the company could have estimated the risks directly from its analysis of real‑world data (Clinical Practice Research Datalink [CPRD] and Hospital Episode Statistics) without using risk equations that needed secondary modification. The committee heard from the company that it used risk equations to be able to model the profiles of specific high-risk populations, such as those with CVD who have additional risk factors. The committee concluded that using published risk equations, although introducing additional uncertainty, could be accepted if these reliably estimated the risks of CVD.
4.16 The committee discussed whether the risk equations used by the company to predict the risks of CVD at baseline were appropriate. It noted that the company initially used the Framingham Heart Study risk equations for patients without CVD, and the REACH registry risk equations for patients with CVD. The committee was aware that extensive validation studies had shown that the Framingham risk equations systematically overestimated the risk of CVD in UK patients. In response to consultation, the company used the QRISK2 assessment tool for the non-familial and heterozygous-familial hypercholesterolaemia populations without CVD. The committee was aware that QRISK2 was more widely used in the UK, being recommended in NICE's guideline on lipid modification, and targeted to UK patients. It concluded that QRISK2 was more appropriate than the Framingham risk equations for patients without CVD, acknowledging that neither was derived from people with heterozygous-familial hypercholesterolaemia.
4.17 The committee discussed how the company adjusted the risks of CVD predicted from the Framingham and REACH registry risk equations for the heterozygous-familial hypercholesterolaemia population. It noted that the company applied a rate ratio of 6.1, derived from a study by Benn et al. (2012), to reflect the increased risk of CVD in this population. The committee was aware that the model was highly sensitive to this parameter (see section 3.43). The committee heard from the ERG that this adjustment was not appropriate mainly because Benn et al. compared the risk of CV events between people with heterozygous-familial hypercholesterolaemia and those with non-familial hypercholesterolaemia. However, the company applied the rate ratio estimated from Benn et al. to the RUTHERFORD‑2 trial population, who were already at high risk of CVD from having heterozygous-familial hypercholesterolaemia. Furthermore, the clinical experts considered that the estimate based on Benn et al. was likely to have significantly overestimated the risk of CVD in patients with heterozygous-familial hypercholesterolaemia, although they acknowledged that there was no robust evidence on the increased risk of CVD in these patients compared with the general population. The committee also heard from the ERG that the risk of CVD predicted by the model for people with heterozygous-familial hypercholesterolaemia, both with or without CVD, was much higher than the risks for the same populations in 'real-world' databases, including the CPRD. The committee concluded that the rate ratio from Benn et al. highly overestimated the risk of CVD among people with heterozygous-familial hypercholesterolaemia, and cast doubt about the validity of the estimated cost effectiveness of evolocumab for this population.
4.18 The committee considered how the company captured the lifetime progression of CVD among people with primary hypercholesterolaemia. It noted that the company used estimates from the Cholesterol Treatment Trialists' (CTT) meta-analysis to convert the surrogate outcomes measured in the trials (LDL‑C concentration) to 'real‑world' outcomes (CV events). However, the company did not use the 21 trials of 'statin versus control'. Instead, it used the 5 trials of 'more versus less intensive statin therapy' for most CV events from the CTT meta-analysis published in 2010. The committee was aware that a more recent CTT meta-analysis (including a total of 27 trials) was published in 2012. The committee considered that, without a justification for not doing so, the most recent analysis should be used because it provides the most mature data. It also heard from the clinical experts that using the 2012 meta-analysis was preferable. The committee concluded that the 2012 meta-analysis would be more appropriate than earlier data.
4.19 The committee considered how the company applied the treatment effect in the model. It noted that patients in the model had treatment continually over their lifetime, and that the treatment effect lasted throughout the time horizon at the same level as that observed in the short-term trials. The committee agreed that there were no long-term data on the extent to which evolocumab could reduce CVD, or whether this effect would be sustained over time (see sections 4.10 and 4.11). It noted the company's response to consultation suggesting that there was no evidence that the treatment effect diminished over time, or that neutralising antibodies developed with evolocumab. The committee would have liked the company to have explored further the uncertainty in the long-term effects of evolocumab. The committee concluded that the company's modelling of the treatment effect was uncertain, although the available evidence suggested that it was unlikely to have a significant impact on the cost effectiveness of treatment.
4.20 The committee considered the utility multipliers used in the model. It agreed that these were generally in line with other values used for people with primary hypercholesterolaemia. However, the committee noted that the relationship assumed between age and utility was based on a study by Dolan et al. (1996), which the ERG considered to be crude and outdated by a more recent equation based on the Health Survey for England. In response to consultation, the company used the equation from the Health Survey for England to inform the relationship between age and background health-related quality of life. The committee concluded that the utility multipliers could be accepted in this appraisal.
4.21 The committee considered the cost effectiveness of the 2 dosages of evolocumab (140 mg every 2 weeks and 420 mg monthly). It recognised that the available cost-effectiveness evidence related to the every 2 weeks dosage and that the company had not presented evidence, apart from a scenario analysis, for the monthly dosage. The committee was aware that evolocumab 420 mg monthly was more expensive than evolocumab 140 mg every 2 weeks. Without evidence for the monthly dosage, the committee was unable to recommend evolocumab 420 mg monthly.
4.22 The committee discussed the company's revised incremental cost-effectiveness ratios (ICERs), including the patient access scheme, presented as part of the company's new evidence in response to consultation (see section 3.45). It discussed the ICERs for evolocumab 140 mg every 2 weeks in people without CVD and, separately, in those with CVD, noting that they all needed considerable caution in their interpretation:
-
Non-familial hypercholesterolaemia population: the committee noted that the company's base-case ICERs with the patient access scheme for the non-familial hypercholesterolaemia population with or without CVD (£31,000–78,500 per quality-adjusted life year [QALY] gained) were all above the maximum acceptable ICERs normally considered to represent a cost-effective use of NHS resources (£20,000–30,000 per QALY gained).
-
Heterozygous-familial hypercholesterolaemia population: the committee noted that the ICERs for people without CVD (£21,900–25,600 per QALY gained) were lower than those for people with CVD (£25,300–32,700 per QALY gained). This was inconsistent with the results for the non-familial hypercholesterolaemia population, and counter-intuitive because people with CVD have a higher risk of CV events, and so would be expected to gain more QALYs from treatment than those without CVD. The committee heard from the company that people without CVD may be benefitting from the prevention of a first CV event. However, it considered that this did not explain why the non-familial hypercholesterolaemia population without CVD would not benefit in the same manner, and have lower ICERs than the population with CVD. The committee heard from the ERG that different CV events were assumed for the non-familial and heterozygous-familial hypercholesterolaemia populations and, further, the calibration of CV events for the non-familial hypercholesterolaemia population was event-specific, whereas a single rate ratio (6.1) from Benn et al. (2012) was applied to all CV events for the heterozygous-familial hypercholesterolaemia population. Therefore, the committee had doubts about the resulting ICERs and their face validity, especially those for people without CVD.
In addition to the magnitude of the company's ICERs, and the concerns about their validity for certain subpopulations, the committee recalled its misgivings about using the composite states in the model (see section 4.12), and modelling treatment continually over the person's lifetime, while assuming that its effect would be maintained throughout the time horizon (see section 4.19). Furthermore, the committee had concerns about using Benn et al. (2012) to reflect the increased risk of CVD for heterozygous-familial hypercholesterolaemia (see section 4.17). Because of this, it concluded not to recommend evolocumab 140 mg every 2 weeks for the primary hypercholesterolaemia (heterozygous-familial and non-familial) or mixed dyslipidaemia population as a whole.
Subgroups
4.23 The committee considered the cost effectiveness of evolocumab in clinically relevant subgroups. It agreed that the company's analyses had limitations, which made the committee question the validity of the results, particularly those for heterozygous-familial hypercholesterolaemia. Nevertheless, the committee agreed that there was merit in exploring potential subgroups of patients with the highest need. It also noted that most responses to consultation advocated using evolocumab in selected subgroups. The committee considered how it could reconcile the uncertainty in the evidence base with the clinical unmet need in the primary hypercholesterolaemia or mixed dyslipidaemia population. It noted the consistent trend in the company's results suggesting that the cost effectiveness of evolocumab would improve within a given population as the risk of CVD increases (see section 3.50). The committee acknowledged that evolocumab is a new therapy with a novel mechanism of action, which consistently reduced LDL‑C concentrations compared with placebo and ezetimibe, while also being well-tolerated by patients. Taken together, the committee concluded that, although the ICERs were not as robust as it would have liked, they could be used to check a proposed set of recommendations guided by the clinical unmet need in the primary hypercholesterolaemia or mixed dyslipidaemia population.
4.24 The committee discussed which subgroups would have a high clinical unmet need. It noted that the Royal College of Pathologists and a clinical expert considered that evolocumab would be particularly valued for:
-
non-familial hypercholesterolaemia with progressive, symptomatic CVD, and persistently high LDL‑C concentrations above 4.0 mmol/litre despite maximal tolerated lipid-lowering therapy
-
heterozygous-familial hypercholesterolaemia with progressive, symptomatic CVD, and persistently high LDL‑C concentrations above 4.0 mmol/litre despite maximal tolerated lipid-lowering therapy
-
severe heterozygous-familial hypercholesterolaemia with pre-treatment LDL‑C concentrations above 8.0 mmol/litre and persistently high LDL‑C concentrations above 4.0 mmol/litre despite maximal tolerated lipid-lowering therapy.
The committee understood that 'persistent' LDL-C means on-treatment LDL-C concentrations confirmed by repeated measures. Other consultation comments received also suggested specific subgroups. However, these suggestions would be difficult to implement in the NHS because the subgroups did not reflect clinical practice. The committee agreed that the suggestions from the Royal College of Pathologists and the clinical expert highlighted the areas where the clinical unmet need was highest, concluding that it would be appropriate to use them as a starting point in its decision-making.
4.25 The committee was aware that there was an ongoing NICE appraisal of another PCSK9 inhibitor (alirocumab) and that the evidence submitted for that appraisal differed from this one in a number of ways. Firstly, the risks of CVD at baseline for most populations were estimated directly from observational data without using risk equations, and then modifying the estimated risks to approximate the characteristics of the respective population. Secondly, the 2012 CTT meta-analysis was used to convert the surrogate outcomes measured in the trials to 'real world' outcomes, which the committee agreed was more appropriate than using the earlier meta-analysis published in 2010 (see section 4.18). Thirdly, the ICERs were for triple therapy (that is, alirocumab plus statin plus ezetimibe), which the committee heard was likely to be the combination in which PCSK9 inhibitors would be mainly used. For these reasons, the committee was generally satisfied that the ICERs in the appraisal of alirocumab were suitable for decision-making. The committee recognised that evolocumab and alirocumab had similar efficacy in clinical trials, and that clinicians generally regarded them as being clinically equivalent. Because of this, the committee concluded that it should refer to the appraisal of alirocumab in its discussion about the subgroups of people who would be prioritised for treatment with PCSK9 inhibitors.
4.26 The committee considered the subgroups identified as being at high risk of CVD in the appraisal of alirocumab. It was aware that that appraisal defined 2 levels of CVD risks:
-
High risk of CVD: previous acute coronary syndrome (such as myocardial infarction or unstable angina needing hospitalisation), coronary or other arterial revascularisation procedures, coronary heart disease, ischaemic stroke, or peripheral arterial disease.
-
Very high risk of CVD: recurrent CV events or CV events in more than 1 vascular bed (that is, polyvascular disease).
The committee noted that the CV events defining these risk levels meant that the definition of CVD was more restricted in the appraisal of alirocumab than in this appraisal (see section 3.22). The committee concluded that it would be mindful of the distinction between high and very high risk of CVD in the appraisal of alirocumab when considering subgroups with a high unmet need.
Non-familial hypercholesterolaemia with CVD
4.27 The Royal College of Pathologists and the clinical expert recommended persistently high LDL‑C concentrations above 4.0 mmol/litre for non-familial hypercholesterolaemia with progressive, symptomatic CVD. The committee understood that most people in this subgroup will not have been able to tolerate statins, and would consequently have a high residual risk of CVD. The committee agreed that the company's subgroup with acute coronary syndrome could be considered a reasonable proxy for progressive, symptomatic CVD. The company estimated that people with a mean LDL‑C concentration of 4.0 mmol/litre and acute coronary syndrome who can tolerate statins had ICERs of £37,700 (LAPLACE‑2) and £29,200 (CPRD) per QALY gained for evolocumab plus statin compared with ezetimibe plus statin. These ICERs would be lower for a minimum, as opposed to a mean, LDL‑C concentration of 4.0 mmol/litre. The committee recognised that the company's ICERs for evolocumab were high. However, it noted from the ongoing appraisal of alirocumab that when a more restricted definition of CVD is used, the ICERs were likely to decrease. The committee agreed that the remaining uncertainty around the ICER could be accepted because evolocumab would represent an important treatment option for this group of patients, who have a relatively high risk of CVD and few treatment options. The committee concluded that it could recommend evolocumab 140 mg every 2 weeks for primary non-familial hypercholesterolaemia with CVD in people with LDL‑C concentrations persistently above 4.0 mmol/litre.
4.28 The committee discussed whether there would be people with LDL‑C concentrations below 4.0 mmol/litre at very high risk of CVD who could benefit from evolocumab. It considered the response to the second appraisal consultation document, which suggested that specifying a minimum LDL‑C concentration above 4.0 mmol/litre to start treatment would exclude some people at high risk of CVD because of a poor response to statins, statin intolerance, or a high pre-treatment LDL‑C concentration. The response noted that in one-sixth of people, high-dose statins will reduce LDL‑C concentrations by less than 39%. The committee heard from the clinical experts that LDL‑C concentrations below 4.0 mmol/litre in themselves would not pose as high a risk of CVD as those above 4.0 mmol/litre, but other factors will sometimes increase the risk of CVD. The committee recalled that in the ongoing appraisal of alirocumab, a group considered to be at even higher risk of CVD than the non-familial hypercholesterolaemia population with CVD as a whole was defined (group at 'very high risk of CVD'; see section 4.26). The committee accepted that requiring a minimum LDL‑C concentration of 4.0 mmol/litre may exclude a small group with additional risk factors for CVD who do not have an effective option to reduce their risk of CVD. It revisited the company's subgroup analyses, noting that these suggested that evolocumab would be cost effective for people with more than 1 risk factor, or with disease in multiple vascular beds (see section 3.50). The committee discussed whether a lower LDL‑C concentration of 3.5 mmol/litre for starting treatment would be more appropriate than 4.0 mmol/litre for that group. It heard from the clinical experts that the benefit of treatment on CV outcomes below an LDL‑C concentration of 3.5 mmol/litre was still being researched. Therefore, the committee concluded that for primary non-familial hypercholesterolaemia with very high risk of CVD (as defined in section 4.26), evolocumab 140 mg every 2 weeks could be recommended in people with LDL‑C concentrations persistently above 3.5 mmol/litre.
Heterozygous-familial hypercholesterolaemia with CVD
4.29 The Royal College of Pathologists and the clinical expert recommended persistently high LDL‑C concentrations above 4.0 mmol/litre for heterozygous-familial hypercholesterolaemia with progressive, symptomatic CVD. The committee recognised that people in this subgroup with LDL‑C concentrations above 5.0 mmol/litre would be eligible for apheresis according to the current guidelines (see section 4.6), reflecting a high clinical unmet need. Although it recalled its misgivings about the calculations, the committee noted that the company's ICER for people with an LDL‑C concentration above 4.0 mmol/litre would be lower than £30,200 per QALY gained. It also considered responses to the second appraisal consultation document, suggesting that a minimum LDL‑C concentration above 4 mmol/litre to start treatment would exclude some people at high risk of CVD. The committee noted that the company's ICER for heterozygous-familial hypercholesterolaemia with CVD in people with a mean LDL‑C concentration of 3.5 mmol/litre was £33,600 per QALY gained, but would be lower for people with progressive symptomatic CVD whose minimum, rather than mean, LDL‑C concentration was 3.5 mmol/litre. Also, restricting the definition of CVD as in the appraisal of alirocumab would further reduce the ICER. The committee, accepting that a minimum LDL‑C concentration of 4 mmol/litre may exclude a group at high risk of CVD, agreed that a lower LDL‑C concentration of 3.5 mmol/litre for starting treatment would be more appropriate than 4.0 mmol/litre for heterozygous-familial hypercholesterolaemia with CVD. The committee concluded that for this population, it could recommend evolocumab 140 mg every 2 weeks in people at high risk, or very high risk, of CVD (as defined in section 4.26) with LDL‑C concentrations persistently above 3.5 mmol/litre.
Heterozygous-familial hypercholesterolaemia without CVD
4.30 For the subgroup with heterozygous-familial hypercholesterolaemia without CVD, the Royal College of Pathologists and the clinical expert recommended a pre-treatment LDL‑C concentration above 8.0 mmol/litre, and a persistently high LDL‑C concentration above 4.0 mmol/litre. The company's estimated ICER for people with a mean LDL‑C concentration of 4.0 mmol/litre was less than £23,500 per QALY gained. However, the committee had doubts about the validity of the ICERs for the heterozygous-familial hypercholesterolaemia population without CVD (see section 4.22). The committee was aware that in the ongoing appraisal of alirocumab the risks of CVD for most populations were estimated directly from real-world data without using the rate ratio from Benn et al., and the ICERs for heterozygous-familial hypercholesterolaemia were higher for people without CVD than those with CVD. The committee noted the response to the second appraisal consultation document suggesting that the 4.0 mmol/litre threshold already defined a higher risk subgroup within the heterozygous-familial hypercholesterolaemia population, and that the additional criterion for pre-treatment LDL‑C concentration (above 8.0 mmol/litre) would exclude some people at high risk of CVD in whom evolocumab would be cost effective. The committee understood from the clinical experts that the risk of CVD varies widely among people with heterozygous-familial hypercholesterolaemia. A pre-treatment LDL‑C concentration above 8 mmol/litre, corresponding to the 90th percentile, has historically been used to define 'severe heterozygous-familial hypercholesterolaemia', which reflects long exposure to high LDL‑C concentrations. This is associated with an increased risk of CVD of 25% compared with non-severe heterozygous-familial hypercholesterolaemia. The committee agreed that a minimum LDL‑C concentration on repeated measures, rather than a single pre-treatment measure, would capture severe heterozygous-familial hypercholesterolaemia with less potential to exclude people at high risk of CVD (young people, for example, because LDL‑C concentrations increase with age). The committee was aware that the current criteria for apheresis require the LDL‑C concentration to remain above 5.0 mmol/litre. Without further evidence on how to identify people at high risk of CVD within this subgroup, the committee agreed that this threshold would adequately capture such people. The committee concluded that it could recommend evolocumab 140 mg every 2 weeks for heterozygous-familial hypercholesterolaemia without CVD in people with LDL‑C concentrations persistently above 5.0 mmol/litre.
Overall conclusion
4.31 The committee agreed that evolocumab would be a clinically and cost-effective use of NHS resource in certain subgroups. It concluded that it could recommend evolocumab, only if:
-
the dosage is 140 mg every 2 weeks
-
the person has primary non-familial hypercholesterolaemia and
-
a history of any of the following: acute coronary syndrome (such as myocardial infarction or unstable angina needing hospitalisation); coronary or other arterial revascularisation procedures; coronary heart disease; ischaemic stroke; or peripheral arterial disease and
-
an LDL‑C concentration persistently above 4.0 mmol/litre
-
-
the person has primary non-familial hypercholesterolaemia and
-
recurrent CV events or CV events in more than 1 vascular bed (that is, polyvascular disease) and
-
an LDL‑C concentration persistently above 3.5 mmol/litre
-
-
the person has primary heterozygous-familial hypercholesterolaemia and
-
a history of any of the following: acute coronary syndrome (such as myocardial infarction or unstable angina needing hospitalisation); coronary or other arterial revascularisation procedures; coronary heart disease; ischaemic stroke; or peripheral arterial disease and
-
an LDL‑C concentration persistently above 3.5 mmol/litre
-
-
the person has primary heterozygous-familial hypercholesterolaemia and
-
no CVD and
-
an LDL‑C concentration persistently above 5.0 mmol/litre
-
-
the company provides evolocumab with the discount agreed in the patient access scheme.
Without evidence for the monthly dosage, the committee was unable to recommend evolocumab 420 mg monthly for primary hypercholesterolaemia (heterozygous-familial and non-familial) or mixed dyslipidaemia.
4.32 The committee was aware that the company's model did not include people with mixed dyslipidaemia. It discussed whether its recommendations for primary hypercholesterolaemia could be generalised to mixed dyslipidaemia. The committee recalled its conclusion that evolocumab may be used for mixed dyslipidaemia when LDL-C concentrations remain very high despite maximum statin and ezetimibe management (see section 4.2). The committee was aware that people with mixed dyslipidaemia also have elevated LDL‑C concentrations, and that treatment for mixed dyslipidaemia is partly determined by the LDL-C concentration. The committee was aware that people with heterozygous-familial hypercholesterolaemia are considered to be at high risk of CVD not only because of the high LDL-C concentrations, but also because of the lifelong exposure to such concentrations. Because of this, the committee concluded that the recommendations for mixed dyslipidaemia should follow those for non-familial hypercholesterolaemia because mixed dyslipidaemia is not associated with additional risk factors that warrant intervention at lower LDL-C concentrations.
4.33 The committee was aware that the Royal College of Pathologists and the clinical expert recommended that LDL‑C concentrations had to be persistently high despite maximal tolerated lipid-lowering therapy. It recalled that statins with or without ezetimibe are the mainstay of treatment for primary hypercholesterolaemia. However, some people may be misidentified as being unable to tolerate statins (see section 4.4), and this may worsen the cost effectiveness of subsequent treatment. Because of this, the committee emphasised that its recommendations for evolocumab should only apply when maximal tolerated lipid-lowering therapy has failed. It clarified that this meant that either the maximum dose has been reached or further titration is limited by intolerance (as defined in NICE's guideline on familial hypercholesterolaemia). The committee was aware that there may also be people in whom statins are contraindicated. These people have the same unmet clinical need as those who cannot tolerate statins, and so should have the same treatment options. In addition, there is no biologically plausible reason for the effect to differ between these 2 groups. Because of this, the committee concluded that its recommendations should also apply to people in whom statins are contraindicated.
4.34 The committee discussed whether the ICERs presented reflect the cost of evolocumab to the NHS. It understood that the actual discount received by the NHS may be less than the percentage discount offered in the patient access scheme. This is because people may move from secondary to primary care after 2–3 years, and simple discounts do not apply when drugs are prescribed through GP's FP10 prescriptions. The committee considered that the subgroups for which evolocumab is recommended have severe hypercholesterolaemia and a high risk of CVD, so people should continue treatment under secondary care where simple patient access schemes apply. The committee concluded that the discounted patient access scheme price of evolocumab would be consistently applied for all people for whom evolocumab is recommended.
4.35 The committee considered whether it should take into account the consequences of the Pharmaceutical Price Regulation Scheme (PPRS) 2014, and in particular the PPRS payment mechanism, when appraising evolocumab. 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 with regard to the relevance of the PPRS to this appraisal of evolocumab. It therefore concluded that the PPRS payment mechanism was not applicable for considering the cost effectiveness of evolocumab.