3 Committee discussion
The appraisal committee (section 5) considered evidence submitted by Bayer, a review of this submission by the evidence review group (ERG), and the technical report developed through engagement with stakeholders. See the committee papers for full details of the evidence.
The appraisal committee was aware that several issues were resolved during the technical engagement stage, and agreed that:
It is more clinically plausible to substitute transition probability values used in the economic model that were numerically zero with non-zero values taken from the REACH registry.
The company's economic model underestimates the effect of varying the stratified mortality outcome 'cardiovascular death' in its sensitivity analyses. The company used the hazard ratio (HR) value for 'all cardiovascular death' from the COMPASS trial and assumed the same HR for all stratified death events in its model. Each mortality HR was then varied separately in the deterministic and probabilistic sensitivity analyses. Because each death rate only included a small proportion of all cardiovascular deaths, the effect of varying the HRs individually in the sensitivity analyses was underestimated. This issue had no effect on the results of the base-case analysis and only related to the uncertainty in the sensitivity analyses.
The committee recognised that there were remaining areas of uncertainty associated with the analyses presented (see technical report, table 2, page 7), and took these into account in its decision making. It discussed the following issues (issues 1, 2 and 3), which were outstanding after the technical engagement stage.
3.1 The patient expert explained that people with coronary or peripheral artery disease have serious concerns about the risk of having events such as heart attack and stroke. They noted that coronary or peripheral artery disease is a challenging condition requiring significant lifestyle adjustments, including diet and exercise, which can affect the whole family. It is important that people understand why they are being offered a dual therapy and how the drugs work, because this can help with adherence to treatment. The patient expert explained that patients rely on clinicians to discuss with them the best treatment options. All options should be discussed, weighing up the potential benefits and risks of potentially long-term medication. The committee concluded that patients rely on clinicians to present the best treatment options to them, and all potential benefits and risks should be fully discussed.
Patients would welcome an additional treatment option with an acceptable benefit-risk ratio for the prevention of atherothrombotic events
3.2 NICE's guideline on myocardial infarction: cardiac rehabilitation and prevention of further cardiovascular disease recommends dual antiplatelet therapy for people who have had an acute myocardial infarction. NICE's guideline on management of stable angina recommends considering aspirin 75 mg daily for secondary prevention of cardiovascular disease. NICE's technology appraisal guidance recommends ticagrelor in combination with aspirin for preventing atherothrombotic events in people who have had a myocardial infarction and are at high risk of a further event. The clinical experts noted that despite widespread use of aspirin in both acute and secondary care, the risk of stroke, myocardial infarction and dying from cardiovascular disease remains high for these patients. An additional treatment option that could reduce this risk would be welcomed by patients. The potential clinical benefit from treatment would need to be assessed, taking into consideration the potential risk of major bleeding events. The committee concluded that people with coronary artery disease or peripheral artery disease would welcome an additional treatment option with an acceptable benefit-risk ratio for the prevention of atherothrombotic events.
3.3 COMPASS is a double-blind randomised clinical trial comparing rivaroxaban plus aspirin against aspirin alone in people with stable coronary artery disease or peripheral artery disease who are at high risk of ischaemic events. The company presented data for the overall trial population and also for 3 subpopulations that it identified as being at especially high baseline risk of ischaemic events: people with coronary artery disease and peripheral artery disease, people with coronary artery disease and heart failure, and people with coronary artery disease and poor kidney function. The risk of ischaemic events is related to the person's medical history and the extent of atheroma. People with heart failure, diabetes, poor kidney function or diffuse atherosclerosis affecting several areas (such as the coronary and peripheral arteries) are at higher risk of events. However other factors also increase risk including diabetes, high body mass index and smoking. At technical engagement, it was queried whether the company's subgroups are clinically relevant and represent people at highest risk of atherothrombotic events who are likely to benefit most from treatment with rivaroxaban plus aspirin. The clinical experts agreed that although these subgroups are clinically identifiable and relevant, there are other groups of people who are also at high risk. These include people who have had myocardial infarction or stroke, and those with multi-vessel coronary disease and diabetes. These people could derive similar benefit from rivaroxaban plus aspirin as seen in the 3 subgroups, because there was no between-group heterogeneity in relative treatment effects reported in COMPASS. The committee acknowledged the company's attempt to predict people with higher baseline risk using the 3 subgroups but noted that although the subgroups are illustrative of people at high risk, they do not identify all people at highest absolute risk of ischaemic events. The clinical experts also highlighted the difficulty in identifying people who are at greatest risk in clinical practice and agreed that, because there is no inter-group difference in treatment effects in COMPASS, the results from the overall population should be considered for decision making. However, the committee was aware that, within the overall eligible population, clinicians are likely to identify people at highest absolute risk of ischaemic events and offer rivaroxaban to those who are likely to have the highest potential absolute benefit. People at high risk of ischaemic events and no known increased bleeding risk will have the most favourable benefit−risk profile. The committee concluded that the efficacy results from the whole COMPASS population should be considered rather than the 3 subpopulations identified by the company, and it did not consider the subgroups further.
3.4 The primary efficacy outcome in COMPASS was a composite of 3 major cardiovascular events: myocardial infarction, ischaemic stroke and 'cardiovascular death'. The committee noted that rivaroxaban plus aspirin showed a statistically significant relative risk reduction of 24% in major cardiovascular events compared with aspirin (HR 0.76, 95% confidence interval [CI] 0.66 to 0.86; p<0.001). Two of the individual components of the primary composite outcome also showed statistically significant relative risk reductions in the treatment arm: 42% for ischaemic stroke (HR 0.58, 95% CI 0.44 to 0.76; p<0.001) and 22% for cardiovascular death (HR 0.78, 95% CI 0.64 to 0.96; p=0.02). The committee concluded that rivaroxaban plus aspirin reduces the risk of cardiovascular events compared with aspirin alone, and that the greatest effect is for ischaemic stroke.
3.5 The primary safety outcome in COMPASS was major bleeding based on a modification of the International Society on Thrombosis and Haemostasis (ISTH) criteria. Major bleeding was defined as a composite of fatal bleeding, and/or symptomatic bleeding in a critical area or organ (such as intracranial, intraspinal, intraocular, retroperitoneal, intraarticular or pericardial, intramuscular with compartment syndrome, or bleeding into the surgical site requiring re-operation), and/or bleeding leading to hospitalisation (with or without an overnight stay). The risk of major bleeding, as defined by the modified ISTH criteria, increased by 70% in the rivaroxaban plus aspirin group compared with aspirin alone (HR 1.70, 95% CI 1.40 to 2.05; p<0.001). The bleeds in the rivaroxaban plus aspirin group were mostly gastrointestinal, with non-significant differences for the other components of the primary safety outcome. The committee concluded that rivaroxaban plus aspirin increases the risk of major bleeding compared with aspirin alone.
Rivaroxaban should only be used in people at high risk of ischaemic events defined by the inclusion criteria of COMPASS
3.6 The clinical experts highlighted that COMPASS is a highly selected population that includes people at high risk of ischaemic events but excludes people with a known increased bleeding risk. The committee noted that the benefits and risks of rivaroxaban plus aspirin are only known for the specific population of people in COMPASS; that is, people with a high risk of ischaemic events as defined by the inclusion criteria of the trial. The committee recalled that the clinical benefits and risks associated with rivaroxaban plus aspirin are finely balanced (see section 3.4 and section 3.5). The committee noted that rivaroxaban plus aspirin is contraindicated in cases of active clinically significant bleeding and in people with lesions or conditions considered to be a significant risk factor for major bleeding. It agreed that a person's risk of bleeding should be assessed before rivaroxaban is considered as a treatment option. The decision to start treatment should only be taken after an informed discussion about the risks and benefits of rivaroxaban, weighing up the risk of ischaemic events against the bleeding risk. The committee concluded that rivaroxaban should only be considered a treatment option for people at high risk of ischaemic events as defined by the inclusion criteria of COMPASS.
3.7 The company did not present evidence for rivaroxaban plus aspirin compared with clopidogrel. The clinical experts explained that clopidogrel is the preferred antiplatelet treatment for people with peripheral artery disease (based on NICE's technology appraisal guidance for the prevention of occlusive vascular events), but it is not a relevant comparator for the whole COMPASS population. This is because aspirin is the preferred treatment for secondary prevention of cardiovascular disease in people with stable coronary artery disease, and clopidogrel is only recommended when aspirin is unsuitable because of contraindication or hypersensitivity. The committee concluded that clopidogrel is not a relevant comparator for the overall COMPASS population and that it is appropriate to base its recommendation on a comparison of rivaroxaban plus aspirin with aspirin alone.
The indirect comparison of rivaroxaban against ticagrelor does not provide reliable estimates of relative effectiveness or risk of bleeding
3.8 To estimate the relative efficacy of rivaroxaban plus aspirin compared with ticagrelor plus aspirin, the company did an indirect treatment comparison of COMPASS against data from another trial (PEGASUS). The ERG highlighted that the indirect comparison is methodologically sound, but there are important differences between the patients included in the 2 trials. The proportion of people who had had a myocardial infarction was 100% in PEGASUS, but 62% in COMPASS. Also, the time since a myocardial infarction was between 1 and 3 years in PEGASUS, but in COMPASS people could have had a myocardial infarction at any time within the past 20 years. A clinical expert also noted that people in PEGASUS and COMPASS did not necessarily have the same baseline bleeding risk, and the trials used different methods to define significant bleeding (modified ISTH in COMPASS and TIMI in PEGASUS), making it difficult to compare the rates of bleeding in the 2 trials. The committee noted that the company did not use the results of the indirect comparison in its economic model. It concluded that COMPASS and PEGASUS have too many differences to allow a reliable estimate of the relative efficacy or bleeding risk of rivaroxaban plus aspirin compared with ticagrelor plus aspirin.
3.9 The company modelled cost effectiveness using a Markov model with 5 states (event-free, non-fatal myocardial infarction, ischaemic stroke, intracranial haemorrhage and death). These were subdivided into acute event (0 to 3 months after an acute event), post-event (3 or more months after an acute event) and second acute event. The efficacy and clinical parameters in the model were derived from COMPASS for the comparison of rivaroxaban plus aspirin against aspirin alone. For the comparison against ticagrelor plus aspirin, the results from the indirect comparison of COMPASS and PEGASUS were not directly used to inform the economic model. Instead, the model used HRs from the COMPASS and PEGASUS trials, which the committee considered appropriate. The committee concluded that the structure of the company's model is appropriate for decision making.
3.10 The company's base-case incremental cost-effectiveness ratio (ICER) for rivaroxaban plus aspirin compared with aspirin alone is £14,185 per quality-adjusted life year gained. The committee noted that the ICER is in the range normally considered cost effective and therefore concluded that rivaroxaban plus aspirin is a cost-effective use of NHS resources in people who are at high risk of ischaemic events as defined by the inclusion criteria of COMPASS (see section 3.6).