Cost-effectiveness analysis of percutaneous mitral valve repair with the MitraClip delivery system for patients with mitral regurgitation in Japan

Abstract

Aims: The objective of the study is to evaluate the cost-effectiveness of percutaneous mitral valve repair (TMVr) with the MitraClip NT system (MitraClip procedure) for patients with symptomatic severe mitral regurgitation (MR) at high surgical risk in line with the methodological guideline for cost-effectiveness evaluation by the Ministry of Health, Labour and Welfare.

Material and Methods: The cost-effectiveness of MitraClip procedure was evaluated using a Markov model. Patients are classified into four New York Heart Association classes in each cycle. The model considered MitraClip complication (“major vascular complication”, “major bleeding complication”, “non-cerebral thromboembolism”), adverse events, re-implantation with MitraClip device, mitral valve surgery, and congestive heart failure hospitalization. For the evidence on additional benefits, a study compared with propensity score-matched medical therapy group was used in the analysis. The analysis was conducted from the perspective of a public healthcare payer with a discount rate of 2% for both cost and effectiveness.

Results: In the base-case analysis, total cost and quality-adjusted life year (QALY) gained (Life year (LY) gained) were 7,541,151 JPY and 3.23 QALYs (3.85 LYs) for MitraClip group, and 4,699,692 JPY and 1.79 QALYs (2.43 LYs) for medical therapy group, respectively. The incremental cost-effectiveness ratio (ICER) of MitraClip procedure versus medical therapy was 1.97 million JPY/QALY (US$18,570/QALY, US$1 = 106 JPY), which was evaluated to be cost-effective. The probability of ICER of MitraClip procedure versus medical therapy being 5 million JPY/QALY was 96.7%.

Limitations: There are two limitations. Firstly, the parameters for the comparators were based on some assumptions. However, it was a conservative setting against MitraClip group. Secondary, the mortality rate and adverse events of MitraClip group in a lifetime were estimated from data during a year after the procedure.

Conclusions: MitraClip procedure improved life-years and quality of life in patients at high surgical risk and it was also a cost-effective treatment option.

Keywords:

• Cost-effectiveness analysis
• MitraClip NT system
• mitral regurgitation
• percutaneous mitral valve repair
• medical therapy

Introduction

Mitral regurgitation (MR) occurs when the heart’s mitral valve leaflets do not coapt or close properly, allowing the backward flow of blood from the left ventricle (LV) into the left atrium (LA) during systole, decreasing forward flow into the aorta and the systemic circulation. There is inadequate coaptation of the mitral valve leaflets, resulting in regurgitation of blood from the LV into the LA during systole that leads to LV volume overload, reduction in LV afterload, and increase in LV pressure. Severe acute MR presents with cardiogenic shock secondary to pulmonary congestion and lower cardiac output associated with rapid increase in LV volume load. Severe chronic MR is asymptomatic for a while but, over time, LV dilatation occurs followed by remarkable pulmonary congestion and worsened LV ejection fraction (EF). The majority of patients present with chronic MR, and their quality of life (QOL) is worsened as they experience heart failure (HF) symptoms. If worsened, the condition can lead to chronic HF and ultimately, stroke, atrial fibrillation or sudden death.

The management of MR will vary based on the onset and severity of symptoms, the severity of MR, and the degree of LV dysfunction. For most patients with mild to moderate MR, medical management is instituted. Oral medications include angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blocker (ARB), or beta-blockers to inhibit the neuroendocrine system in order to slow progression of LV dilatation and worsening of contraction (LV remodeling), and diuretics or anti-aldosterone to improve symptoms associated with congestion. For an acute-on-chronic state, intravenous injection of vasodilators, diuretics or cardiotonic agents are used. Assisted circulation may be required when the pumping function of the heart is significantly worsened and the patient is poorly responsive to medications. However, medical management is not a radical treatment, and MR progresses over time. Patients with symptomatic severe MR or asymptomatic severe MR with evidence of LV dysfunction or dilation are considered candidates for surgery (valve repair or replacement). Despite its effectiveness in many patients, surgery involves significant trauma and risks, including operational complications, physical impairment, and death, some of which are related to the use of cardiopulmonary bypass. Surgical risks increase with specific co-morbidities such as prior cardiothoracic surgery and advanced age. Additionally, many patients experience an extended post-operative recovery period, which can translate into significant post-discharge healthcare costs, and in some cases, re-hospitalizations. According to Flameng et al., the reoccurrence rate of severe MR (3+/4+) is 3.7%¹. The complications, including both morbidity and mortality associated with re-operation, are substantially greater than for the initial surgical repair or replacement procedure. Surgery is delayed or not an option (i.e. only medically managed) due to its inherent factors explained above, and severe MR in many patients are considered not effectively treated. The number of patients with valvular disease including MR is increasing². In addition, the number of patients at increased risk for surgery is increasing as the overall population ages³. Surgery is the only radical treatment option for MR. As explained above, however, there has been no effective treatment option for patients with contraindication for surgery due to advanced age, comorbid condition, or severe HF.

MitraClip NT system (Abbott Vascular Co., Ltd., Tokyo, Japan) (MitraClip procedure) is a system providing a less invasive treatment to mitigate MR with long-term double orifices (two valve orifices) formed by applying the surgically established suture of double-orifice formation (edge-to-edge repair) and joining the anterior leaflet and posterior leaflet of the mitral valve with percutaneously inserted clip⁴. In Japan, the MitraClip system were reviewed for marketing approval based on the clinical evaluation of the data from the following clinical studies: the Endovascular Valve Edge-to-Edge Repair Study (EVEREST) II randomized controlled trial (RCT), an overseas study which compared MitraClip and surgery in surgical candidates; the EVEREST II High Risk Registry (HRS) and the EVEREST II Real World Expanded Multicenter Study of the MitraClip System (REALISM) Continued Access Study, overseas studies in patients with severe MR (3+/4+) at high surgical risk; the Integrated High Risk Cohort, the pooled analysis of these two studies; the AVJ-514 Japan clinical trial. As result, the review concluded that the percutaneous treatment with the MitraClip system is lower invasive than surgery and beneficial as a novel treatment option for patient with severe symptomatic MR at high surgical risk. Overseas, the MitraClip system has been approved in over 90 countries including the European Union (March 2008) and the United States (October 2013).

As medical technologies advance, the attention to cost-effectiveness of medical technologies has now increased. Actually, the pilot introduction of economic evaluation targeted for new/existing technologies (medical drugs, medical devices) was started in April 2016⁵. Active discussions on price adjustment using the results of the cost-effectiveness analysis is being conducted in the Ministry of Health, Labour and Welfare (MHLW)⁶.

The study is a cost-effectiveness analysis of percutaneous mitral valve repair (TMVr) with MitraClip procedure for patients with symptomatic severe MR at high surgical risk in line with the methodological guideline for cost-effectiveness evaluation by the MHLW.

Methods

Overview and model structure

The target population for this analysis was patients with symptomatic severe MR at high surgical risk. The comparator was patients with medical therapy. The analysis was conducted from the perspective of public healthcare payers, and quality-adjusted life year (QALY) was used as an economic outcome. According to the guideline for cost-effectiveness evaluation in Japan (second edition)⁷, a discount rate of 2% was set in both cost and effectiveness, and lifetime simulation was conducted.

The cost-effectiveness of TMVr with MitraClip procedure was evaluated using a Markov model constructed based on a previous study by Cameron et al.⁸ The analytical model consists of two states: “alive” and “death”. The analysis starts with patients in the “alive” state. They conduct a TMVr with MitraClip procedure (MitraClip group) or continue medical therapy without any additional intervention (medical therapy group). After the start of analysis, patients in the “alive” state are classified into four New York Heart Association (NYHA) classes and gain a utility value by NYHA class in each cycle. The proportions of the patients with each NYHA class were assumed not to change over times and remain to be constant during lifetime. The cycle length of 1 month was used for the Markov model.

The model considered MitraClip complications (“major vascular complication”, “major bleeding complication”, and “non-cerebral thromboembolism”), adverse events, re-implantation with MitraClip device for failed MitraClip procedure, mitral valve (MV) surgery, and, CHF hospitalization. Costs for those events were calculated. Utility decrement for each event was also considered for re-implantation with MitraClip device, MV surgery, and congestive heart failure (CHF) hospitalization.

Adverse events and CHF hospitalization were considered over a lifetime. MitraClip complications and re-implantation with MitraClip device were considered for 1-month post-procedure. Adverse events included the following 8 events; myocardial infarction (MI), stroke, renal failure, non-elective cardiovascular surgery, mechanical ventilation (≥48 h), gastrointestinal (GI) complication requiring surgery, septicemia, and blood transfusion (≥2 units). The adverse events were considered only in MitraClip group as there was no report on those events for medical therapy group. CHF hospitalization was considered in both groups. MV surgery for 1 year from the start of analysis was considered in both groups. All the incidences of MV surgery occurring in the year were calculated in the first cycle of the analytical model. The analytical model is shown in Figure 1.

Figure 1. Model structure. Abbreviations. NYHA, New York Heart Association; MV, mitral valve; CHF, congestive heart failure. MitraClip complications, adverse events, re-implantation with MitraClip, MV surgery and CHF hospitalization are considered in the analytical model.

Data source

Base-case analysis

Evidence for additional benefits of MitraClip procedure compared with medical therapy was collected by a systematic literature review (SLR). Among the collected evidence, direct comparison studies or propensity score-matched comparison studies were selected. As a result of the SLR, two studies were selected as candidate^9,10. Both of them were propensity score-matched study, the study by Velazquez et al.¹⁰ was selected as evidence for base-case analysis because its larger sample size and it included both functional and degenerative MR.

Velazquez et al. pooled the data of two studies on clinical efficacy of MitraClip procedure for patients with symptomatic severe MR (MR severity 3 or 4+) at high surgical risk; EVEREST II HRS¹¹ and REALISM¹². With respect to the comparator, patients without procedure were extracted from two databases; Duke Echocardiography Laboratory Database (DELD) and Duke Databank for Cardiovascular Disease (DDCD). Among those, patients with matched background characteristics were selected for the comparator by propensity score-matched using age, sex, a history of MI, stroke, chronic obstructive pulmonary disease, renal failure and diabetes, previous cardiac surgery, NYHA class III/IV and LVEF of both groups¹⁰. There was no report on outcomes other than 30-day and 1-year mortality. Therefore, although the report by Velazquez et al. was used for mortality, other clinical outcomes of MitraClip group were obtained from re-analysis of the same group by Abbott Vascular Japan Co., Ltd.

Although the length of stay (LOS) was 14.4 days in AVJ-514 trial¹³, as this is the data in clinical trial, it is expected to be shorter in actual clinical practice. By reference to a medical expert’s experience that a clinical-path of hospitalization for MitraClip procedure was the same as for percutaneous coronary angioplasty (PCI) surgery in the United States, 5.8 days, which was the average LOS of Diagnosis Procedure Combination (DPC) hospitalization for “Percutaneous coronary angioplasty” (K 546) in Japan, was used for the LOS of hospitalization for MitraClip procedure^14,15. The average age of target population in AVJ-514 trial (80 years old) tended to be older because one of the eligibility criteria for the trial was the Society of Thoracic Surgeons (STS) score ≥8%.¹³ In the actual indication, there is no criterion for STS score ≥8% and the average age is considered to be lower. Therefore, in the base-case analysis, the initial age was set to 74 years old as the average age reported by Velazquez et al.¹⁰

Scenario analysis

An analysis using clinical evidence used in a paper of cost-effectiveness analysis of MitraClip procedure published by Cameron et al.⁸ was conducted as scenario analysis.

Cameron et al. used the results of EVERST II HRS, the study on the clinical efficacy of MitraClip procedure for patients with symptomatic severe MR (MR severity 3 or 4+) at high surgical risk. Regarding the comparator, patients without procedure were retrospectively reviewed after recruiting the patients with severe MR at high surgical risk for the trial described above. Mortality was estimated by applying Weibull distribution to each survival curve. The initial age was 77 years old, which is the average age in EVEREST II HRS used in the analysis by Cameron et al.

Model inputs

Transition probabilities

Transition probabilities used for the analyses are summarized in Table 1.

Table 1. Transition probability.

Items			MitraClip	Medical therapy	Source
Proportion of patients with each NYHA class^*1	Base-case	Class I	41.8%	3.8%	^10
		Class II	43.0%	22.2%
		Class III	13.3%	59.5%
		Class IV	1.9%	14.6%
	Scenario	Class I	33.3%	3.2%	^11
		Class II	40.7%	12.9%
		Class III	24.1%	64.5%
		Class IV	1.9%	19.4%
Mortality (/month)	Base-case		HR: 0.66*² (1.96% (/month))	3.16% (/month)	^10
	Scenario		HR: 0.492^*2	S(t)=exp(−0.1004t^0.6875)	^8
Re-implantation with MitraClip device for failed MitraClip procedure	Base-case		2.1%	-	^10
	Scenario		1.28%	-	^11
Probability of MV surgery (30 days)	Base-case		2.5%	2.5%^*3	^10
	Scenario		0.0%	14%	^11
Probability of CHF hospitalization (/month)	Base-case		4%	13%	^10
	Scenario		3%	8.7%	^11
Probability of MitraClip complication (30 days)^*4	Base-case	Major vascular complication	6.4%	-	^10
		Major bleeding complication	8.5%	-
		Non-cerebral thromboembolism	0.0%	-
	Scenario	Major vascular complication	6.4%	-	^11
		Major bleeding complication	17.9%	-
		Non-cerebral thromboembolism	0%	-
Adverse events (30 days)	Base-case	MI	1.3%	-	^10
		Stroke	2.5%	-
		Renal failure	2.1%	-
		Non-elective cardiovascular surgery	0.4%	-
		Mechanical ventilation (≥48 h)	2.5%	-
		GI complication requiring surgery	0.0%	-
		Septicemia	1.3%	-
		Blood transfusion (≥2 units of blood)	14.8%	-
Adverse events (after the 2nd month) (/month)	Base-case	MI	0.16%	-
		Stroke	0.12%	-
		Renal failure	0.32%	-
		Non-elective cardiovascular surgery	0.0%	-
		Mechanical ventilation (≥48 h)	0.36%	-
		GI complication requiring surgery	0.12%	-
		Septicemia	0.28%	-
		Blood transfusion (≥2 units of blood)	1.51%	-
Adverse events (30 days)^*5	Scenario	MI	2.6%	-	^11
		Stroke	2.6%	-
		Renal failure	3.8%	-
		Mechanical ventilation (≥48 h)	2.6%	-
		GI complication requiring surgery	1.3%	-
		Septicemia	0.0%	-
		Blood transfusion (≥2 units of blood)	28.2%	-
Adverse events (after the 2nd month)(/month)^*5	Scenario	MI	0.3%	-
		Stroke	0.0%	-
		Renal failure	0.2%	-
		Mechanical ventilation (≥48 h)	0.0%	-
		GI complication requiring surgery	0.2%	-
		Septicemia	0.3%	-
		Blood transfusion (≥2 units of blood)	1.0%	-

Abbreviations. NYHA, New York Heart Association; HR, hazard ratio; MV, mitral valve; CHF, congestive heart failure; MI, myocardial infarction; GI, gastrointestinal; OS, overall survival.

^*1For MitraClip group, 12-month postoperative values in the clinical trials (6-month postoperative values only in AVJ-514) were used. The values at baseline of MitraClip group were used for medical therapy group.

^*2HR of OS for MitraClip against medical therapy.

^*3It was assumed to be the same as the value for MitraClip group in Velazquez¹⁰.

^*4Events including MitraClip complications; Major vascular complication: hematomas exceeding 6 cm of puncture, retroperitoneal hematoma, arteriovenous fistula, symptomatic peripheral ischemia/clinical signs or nerve injury with symptoms persist for more than 24 h, surgical repair of puncture vessel, Ipsilateral deep venous thrombosis, Major bleeding complication: bleeding requiring surgery, bleeding requiring blood transfusion (more than 2 units), Non-cerebral thromboembolism: paroxysmal atrial fibrillation, persistent atrial fibrillation, arrhythmia requiring cardiac block/pacemaker. MitraClip complications were not considered in the scenario analysis 3 since there is no report in AVJ-514 on complications targeted in the analysis.

^*5Non-elective cardiovascular surgery was not considered in the scenario analysis 2 since it was not evaluated in the previous study by Cameron et al.⁸

The proportions of patients with NYHA class from the clinical trial used in each analysis were applied. For MitraClip group, the proportions were at 1 year post-operative. As there was no report for the medical therapy group, the proportions of patients before MitraClip procedure (baseline) were used. The mortality rates (/month) for the setting assuming a hazard ratio (HR) to be constant (the base-case analysis, Scenario 1, 3) were estimated by the following formula: (1) $$\text{Mortality} \left(/\text{month}\right) =1-\hspace{0.17em}\text{exp}\hspace{0.17em}(-(-\hspace{0.17em}\text{ln}(1-x))/\left(T\right))$$(1)

(Cumulative mortality rate at T months in each clinical study (x))

In the case that the mortality estimated by the above formula in the analytical model was below the average mortality in general population, the mortality in general population was selected and increase in mortality by age was considered.

As for MitraClip complications, the incidence rates at 30 days evaluated in the clinical study used for each analysis were set. The incidences of those complications were not considered for the medical therapy group as they were complications related to MitraClip procedure. The incidence rates at 30 days and 12 months evaluated in the clinical study used for each analysis were set for the adverse events rates. The 30-day results were used for at 1 month from the start of the analysis. From the second month onwards, the incidence rate per month of each event was estimated using the incidence rate through 2 to 12 months in the clinical trials. The incidence rates of the adverse events from the second month onwards were estimated by the following formula. (2) $$\text{Incidence} \text{rate} \text{of} \text{event} \left(/\text{month}\right)=\left(X_{T2}-X_{T1}\right)/\left(T2-T1\right)$$(2)

X_T2: Incidence rate at T2 months in each clinical study.

X_T1: Incidence rate at T1 months in each clinical study.

Although the adverse events are expected to occur in both of MitraClip group and medical therapy group, the incidence rate of the adverse events was assumed to be 0% as there was no report on the adverse events for a medical therapy group. The probability of MV surgery for MitraClip group was set using the rate of MV surgery until 12 months evaluated in the clinical study used for each analysis. The probability of MV surgery for medical therapy group was conservatively set to be the same as MitraClip group. The probability of CHF hospitalization for MitraClip group was set using the CHF hospitalization rate until 12 months evaluated in the clinical study used for each analysis. For medical therapy group, the rate before MitraClip procedure (baseline) in the same clinical study was used.

Costs

Cost parameters used in the analysis are summarized in Table 2.

Table 2. Cost.

Items				MitraClip	Medical therapy	Source
Device cost for MitraClip (JPY)				3,057,535	-	^16
Technical fee for MitraClip procedure (JPY)				349,300	-	^17
Cost other than device cost and technical fee for MitraClip (Interatrial septum paracentesis) (JPY)				124,400	-	^17
Cost for MitraClip procedure hospitalization (JPY/event)				262,654	-	^14^,^17
Cost for MV surgery (JPY/event)				7,365,553	-	Claims data calculation
Cost for CHF hospitalization (JPY/event)				997,544	-	Claims data calculation
Treatment cost for MitraClip complication (JPY/event)			Major vascular complication	216,321	-	^17
			Major bleeding complication	5,500	-	^17
			Non-cerebral thromboembolism^*1		-
Drug cost (JPY/month)	0–1 month			14,865	15,678	^13^,^18
	≥2 months			14,807	15,678
Follow-up cost (JPY/month)^*2	0–1 month			15,760	4,668	Expert opinions,^17
	2–12 months			5,860	4,668
	≥1 year			4,668	4,688
Adverse events (JPY/event)	0–1 month	MI^*3		1,927,462	-	Claims data calculation
		Stroke^*3		1,382,356	-	Claims data calculation
		Renal failure^*3		1,465,396	-	Claims data calculation
		Non-elective cardiovascular surgery		5,500	-	^17
		Mechanical ventilation (≥48 h)^*4		45,864	-	Claims data calculation,^17
		GI complication requiring surgery^*5		46,000	-	^14^,^17
		Septicemia^*6		438,350	-	^14
		Blood transfusion (≥2 units)^*7		36,638	-	Claims data calculation,^17
	2 month	MI		1,927,462	-	Claims data calculation
		Stroke		1,382,356	-	Claims data calculation
		Renal failure		1,465,396	-	Claims data calculation
		Non-elective cardiovascular surgery		5,500	-	^17
		Mechanical ventilation (≥48 h)^*4		184,652	-	Claims data calculation,^17
		GI complication requiring surgery		273,910	-	^14^,^17
		Septicemia		438,350	-	^14
		Blood transfusion (≥2 units)^*7		36,638	-	Claims data calculation,^17

Abbreviations. MV, mitral valve; CHF, congestive heart failure; MI, myocardial infarction; GI, gastrointestinal; DPC, Diagnosis Procedure Combination; LOS, length of stay.

^*1It is not considered in the analysis since there is no incidence of “non-cerebral thromboembolism” in the group pooled with EVEREST II HRS and REALISM.

^*2Follow-up cost for MitraClip group after the 2nd year onwards is used for follow-up cost for medical therapy group.

^*3As the hospitalization costs for “MI”, “stroke” and “renal failure” are higher than the cost for “MitraClip procedure hospitalization”, the cost for “MitraClip procedure hospitalization” are included in the hospitalization cost for each event.

^*4As the hospitalization cost for “mechanical ventilation >48h” in 30 days is included in the cost for “MitraClip procedure hospitalization”, only the medical fee for mechanical ventilation (45,864 JPY) is accounted. From the 2nd month onwards, the hospitalization cost for 6.8 days (138,788 JPY) calculated by claims data calculation and the medical fee for mechanical ventilation (45,864 JPY) are accounted.

^*5As the hospitalization costs for “GI complication requiring surgery” is lower than the cost for “MitraClip procedure hospitalization”, the cost for the events occurred in 30 days is included in the cost for “MitraClip procedure hospitalization”. Technical fees and material costs are counted.

^*6As the hospitalization cost for “septicemia” is set based on DPC point. “Septicemia” is assumed to occur in the same hospitalization for “MitraClip procedure”, if it occurs in 30 days, the higher hospitalization cost whichever “septicemia” or “MitraClip procedure” is counted. Therefore, the hospitalization cost for “septicemia” is not calculated in the scenario analysis 1 because “MitraClip procedure hospitalization” exceeds the hospitalization cost of “septicemia”.

^*7As for hospitalization for blood transfusion, because the units of transfusion are calculated for the patients with MV surgery, LOS for the event cannot be calculated for those patients. Thus, only the cost for blood transfusion is included.

In order to set costs related to MitraClip procedure for the analysis, a standard treatment flow was discussed with medical expert and costs were estimated for hospitalization, MitraClip procedure, drug, and outpatient follow-up. Costs for MitraClip procedure include material cost (3,057,535JPY)¹⁶, (average 1.77 clips used per case), technical fee (K559-3) (349,300JPY)¹⁷ and other cost for procedure. Other cost for procedure was assumed to be a cost of interatrial septum paracentesis. Drug cost (concomitant cardiac drugs administering during hospitalization) was set based on prescription data of concomitant drugs in AVJ-514 trial. Outpatient follow-up cost was calculated based on the expert’s opinion. Patients were assumed to make follow-up visits monthly and take a blood test, electrocardiogram, transthoracic echocardiography, chest X-ray examination at 1-, 6-, and 12-month visit, and then once a year. Medical fee points in April 2017 were used as unit costs for each resource usage.

Hospitalization cost for MV surgery, cost for CHF hospitalization, hospitalization cost for major adverse event (MI, stroke, renal failure) treatment and treatment cost for adverse events (mechanical ventilation (≥48 h), blood transfusion (≥2 units)) were calculated using claims database provided by Medical Data Vision (MDV) (April 2008 through December 2016).

With reference to the expert’s opinion, medical treatment for adverse events (non-elective cardiovascular surgery due to adverse events related to a device or procedure) was assumed to be “percutaneous pericardial drainage”, and medical fee for the corresponding procedure was calculated. With reference to the expert’s opinion, medical treatment for adverse events (GI complication requiring surgery) was assumed to be “endoscopic gastrointestinal hemostasis (K654)”. As for hospitalization cost for septicemia, points for corresponding DPC (A419) was used¹⁴. The details of the type of MitraClip complications were cited from the report of Velazquez et al.¹⁰. Costs associated to those complications were set to be costs of representative treatments based on the expert’s opinion.

Health utility

Health Utility in the analysis was set as the same as the study evaluating the cost-effectiveness of MitraClip procedure by Cameron et al. In the study by Cameron et al., utility by NYHA class in an “alive” state, utility decrement for CHF hospitalization, MitraClip procedure and MV surgery were considered and cited from other literature⁸. The utility parameters used in the analysis are summarized in Table 3.

Table 3. QOL value.

Items		MitraClip	Medical therapy	Source
NYHA class	Class I	0.900	0.900	^19
	Class II	0.830	0.830
	Class III	0.740	0.740
	Class IV	0.598	0.598
Disutility for MitraClip procedure		−0.043	-	^20
Disutility for MV surgery		−0.074	−0.074	^21
Disutility for CHF hospitalization		−0.064	−0.064	^22

Abbreviations. QOL, quality of life; NYHA, New York Heart Association; MV, mitral valve; CHF, congestive heart failure.

Analysis

The incremental cost-effectiveness ratio (ICER) was calculated by dividing incremental costs by incremental QALYs between MitraClip group and the medical therapy group. The threshold for ICER, i.e. the threshold to be judged as cost-effective, was set as 5 million JPY/QALY in the pilot cost-effectiveness evaluation in Japan²³.

Both one-way sensitivity analysis (1-way SA) and probabilistic sensitivity analyses (PSA) were conducted to explore the uncertainty around model input parameters. A range in 1-way SA was set as the 95% confidence interval (CI) of each parameter for the incidence rates, costs, and utility values. The range was set as 20% for variables without 95% CI reported or not be able to estimate it. In the PSA, beta distribution was applied to transition probabilities and utilities, and gamma distribution was applied to cost parameters. Log normal distribution was applied to a HR.

Results

In the base-case analysis, total cost and QALY gained were 7,541,151 JPY and 3.23 QALYs for MitraClip group, and 4,699,692 JPY and 1.79 QALYs for medical therapy group, respectively. The ICER of MitraClip procedure versus medical therapy was 1.97 million JPY/QALY (US$18,570/QALY, US$1 = 106 JPY), which is within the range that MitraClip procedure was evaluated to be cost-effective compared with medical therapy.

The results of the 1-way SA are shown in Figure 2. As a result of 1-way SA, the ICER was above 5 million JPY/QALY if the HR of Overall Survival (OS) for MitraClip procedure against medical therapy exceeds 0.97. There was no other variable exceeding 5 million JPY/QALY within the setting range of the analysis. The results of PSA clarified that the probability of ICER of MitraClip procedure versus medical therapy being 5 million JPY/QALY was 96.7%. The cost-effectiveness plane and cost-effectiveness acceptability curve are shown in Figure 3.

Figure 2. Tornado diagram. For parameters with asterisk (*), 95% CI is used as a range of sensitively analysis. The range of ±20% of baseline value is chosen for the rest of parameters. Abbreviations. ICER, incremental cost-effectiveness ratio; QALY, quality adjusted life year; HR, hazard ratio; OS, overall survival; CHF, congestive heart failure; NYHA, New York Heart Association; MI, myocardial infarction; MV, mitral valve; CI, confidence interval.

Figure 3. Probabilistic sensitivity analysis. (A) Cost-effectiveness plane, and (B) cost-effectiveness acceptability curve. Abbreviations. QALY, quality adjusted life year; WTP, willingness to pay.

In the scenario analyses, ICERs of MitraClip procedure versus medical therapy were also below 5 million JPY/QALY. The results of base-case and scenario analysis are reported in Table 4.

Table 4. Results of base-case and scenario analysis.

	QALY	LY (years)	Cost (JPY)	ICER (JPY/QALY)	ICER (JPY/LY)
(1) Base-case analysis
MitraClip	3.2312	3.8519	7,541,151	1,968,389	1,996,778
Medical therapy	1.7876	2.4289	4,699,692
(2) Scenario analysis^*1
MitraClip	4.6709	5.6642	8,153,892	1,304,314	1,198,355
Medical therapy	1.9266	2.6772	4,574,425

Abbreviations. QALY, quality adjusted life year; LY, life year; ICER, incremental cost-effectiveness ratio.

^*1Analysis using parameters on Canadian CEA (Cameron et al.⁸).

Discussions

In Japan, as the national medical expenditure continues to expand due to the aging population and the soaring prices by the increase of development expenses with the advancement of medical technology, concerns about the continuation of universal health insurance are increased year by year. The total population of Japan as of 2016 is 126.93 million, continuously decreasing since 2011, while the percentage of the population aged 65 or over in the total population is a record high of 27.2%. Japan has the highest percentage of the population aged 65 or over among the major countries (as of 2015). In terms of the national medical expenditure, it reached 42 trillion and 364.4 billion yen in 2015, and it is steadily increasing²⁴. Japan has an urgent issue in improvement in the tightening of social security expenses due to the imbalance between the income decrease and the expense increase, caused by the decrease in the labor force population.

Under these circumstances, discussions towards the introduction of cost-effectiveness evaluation to the medical system have begun in Japan from May 2012, and verification towards full-scale introduction in the 2019 fiscal year is in progress. Therefore, as well as foreign countries, concern about the cost-effectiveness of medical drugs and medical devices is rising in Japan not only in government ministries but also at medical sites. Especially in innovative technologies, cost-effectiveness evaluation is important to evaluate quantitatively their values for money

Our base-case analysis showed that ICER of MitraClip procedure versus medical therapy was 1.97 million JPY/QALY (2.00 million/LY) and MitraClip procedure was evaluated as cost-effective. The results of the 1-way SA showed that the HR of OS for MitraClip procedure against medical therapy had a large influence on the analysis result. As the results of PSA clarified that the probability of ICER of MitraClip procedure versus medical therapy being 5 million JPY/QALY was 96.7%, the robustness of these results was confirmed. In the 1-way SA about the base-case analysis, it was shown that the long-term transition of OS has a large influence on the analysis result. Therefore, a scenario analysis quoting a transition probability using survival curves based on observation results in a longer-term by Cameron et al.⁸ was conducted. The results of the analysis were similar to the results of the base-case analysis.

The cost-effectiveness of MitraClip procedure was evaluated by Haute Autorité de Santé (HAS)²⁵ in France, Health Quality Ontario (OHTAC)²⁶ in Canada, and Medical Services Advisory Committee (MSAC)²⁷ in Australia. Medical therapy was set as a comparator in the evaluations by all three organizations of health technology assessment (HTA).

This is the first analysis to evaluate the cost-effectiveness of MitraClip procedure in patients with symptomatic severe MR at high surgical risk in Japan. Newer data was used for the clinical evidence, and it was a comparison study with propensity score-matched with medical therapy¹⁰. Since Velazquez et al. reported only 30-day and 1-year mortality as outcomes, other clinical outcomes for MitraClip group were newly obtained by re-analysis in the same population and used in the analysis.

The report of the Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients with Functional Mitral Regurgitation (COAPT) trial, a multicenter, randomized, controlled, parallel-group, open-label trial, was recently published²⁸. In evaluation of long-term outcomes within 24 months of follow-up in the trial, it was shown that death from any cause occurred in 29.1% of the patient in the device group and 46.1% in the control group (HR, 0.62; 95% CI, 0.46 to 0.82; p < .001), and the annualized rate of all hospitalizations for HF in each group was 35.8% per patient-year and 67.9% per patient-year (HR, 0.53, 95% CI, 0.40 to 0.7, p < .001), respectively. Since the ICER value is affected by the HRs of both outcomes and the magnitudes of the event rates in the control group, the results cannot be judged unconditionally. Nevertheless, from the result of the 1-way SA shown that both outcomes have a large impact on the analysis results, it is expected that the cost-effectiveness of MitraClip procedure will be shown similarly to the result of this analysis. Cost-effectiveness analysis based on COAPT trial is expected in order to evaluate the robustness of the result of this analysis. Before the publication of the COAPT trial, Percutaneous Repair with the MitraClip Device for Severe Functional/Secondary Mitral Regurgitation (MITRA-FR trial), a multicenter, randomized, open-label, controlled phase 3 trial²⁹ was also reported. In the evaluation of the cost-effectiveness of MitraClip procedure, it might be needed to consider uncertainties of the analysis results caused by the learning curve effect on MitraClip procedure^30,31, the follow-up period (12 months) of this trial and differences in the baseline disease state³².

The cost-effectiveness of MitraClip procedure compared with medical therapy for patients with symptomatic severe MR at high surgical risk was evaluated in the present study. However, there are two limitations regarding result interpretation. Firstly, there was a limit to the setting of parameters for the comparator. In the base-case analysis, the result of medical therapy group with propensity score-matched to the result of the clinical trial of MitraClip procedure was used for the comparator. However, data of direct comparison for parameters other than mortality was not obtained because there was no report on other parameters. Therefore, the incidence rate of adverse events for medical therapy was set to be 0%, which was a conservative setting against MitraClip group. In addition, the proportions of patients with NYHA class, CHF hospitalization were assumed to be the same as before MitraClip procedure, and data of MitraClip group at baseline was used. Although the analysis using parameters of previous study⁸ was conducted in Scenario analysis, its comparability has limitation because EVEREST II HRS used by Cameron et al. was not an RCT nor propensity score-matched study. Secondly, the mortality rate and adverse events of MitraClip group in lifetime were estimated from data during a year after the procedure. As for the proportions of patients with NYHA class, the state at 1 year after procedure was assumed to continue over lifetime. Although the average of time horizon in lifetime simulation was short, which is 4.17 years for MitraClip group and 2.55 years for medical therapy, an analysis considering long-term prognosis after MitraClip procedure is required after accumulation of longer-term clinical data.

Conclusions

Using best available comparative studies of MitraClip procedure and medical therapy, the cost-effectiveness of MitraClip procedure compared with medical therapy for patients with symptomatic severe MR at high surgical risk was evaluated. The ICER in the base-case analysis was 1.97 million JPY/QALY, and the robustness of the result was confirmed in the sensitivity analysis and the scenario analysis. It was shown that MitraClip procedure improved life-years and QOL in patients at high surgical risk and it was also a cost-effective treatment option.

Transparency

Declaration of funding

This study was funded by Abbott Vascular Japan Co., Ltd.

Declaration of financial/other interest

The authors have no conflicts of interest. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

All authors were involved in the conception and design of this study. SI conducted the analyses. KN, TM, and HS were involved in the interpretation of data. All authors agree to be accountable for all aspects of the work.

Acknowledgements

None reported.