OUP user menu

Clinical and cost-effectiveness of left ventricular assist devices as a bridge to heart transplantation for people with end-stage heart failure: a systematic review and economic evaluation

Andrew J. Clegg, David A. Scott, Emma Loveman, Jill L. Colquitt, Pam Royle, Jackie Bryant
DOI: http://dx.doi.org/10.1093/eurheartj/ehi857 2929-2938 First published online: 7 April 2006

Abstract

Aims To evaluate the clinical and cost-effectiveness of left ventricular (LV) assist devices (LVADs) as a bridge to transplant (BTT) for people with end-stage heart failure (ESHF) through a systematic review and economic evaluation.

Methods and results The systematic review and economic evaluation was conducted according to internationally recognized methods. The search strategy identified systematic reviews, randomized controlled trials, quasi-experimental studies, and observational studies evaluating the effects of LVADs on survival, functional capacity, and quality of life. Cost-effectiveness was assessed through a 5-year decision analytic model to estimate the incremental cost-effectiveness ratio of LVADs compared with usual care. Despite the poor methodological quality of the 18 studies included, LVADs appear beneficial improving survival, functional status, and quality of life. Adverse events are a serious concern. The economic evaluation showed that LVADs had a cost per quality adjusted life year of £65 242 (95% confidence interval £34 194—364 564). Sensitivity analysis showed that post-heart transplant survival gains, pre-heart transplant patient utility, and one-off costs associated with implantation determine cost-effectiveness.

Conclusion Although LVADs appear clinically effective as a BTT for people with ESHF, it is unlikely that they will be cost-effective unless costs decrease or the benefits of their use increase.

  • Heart failure
  • Left ventricular assist devices
  • Systematic review
  • Economic evaluation
  • Health technology assessment

Introduction

Heart transplantation is the accepted form of treatment for people with end-stage heart failure (ESHF), increasing the length and quality of life.1,2 Recent decreases in organ donation have restricted the availability of this treatment option to a few.35 As a consequence, attention has focused on mechanical circulatory support as an alternative. Left ventricular (LV) assist devices (LVADs), which assist the function of the damaged left ventricle and help to restore normal haemodynamics and end organ flow, have been developed to provide support for patients as either a bridge to heart transplantation (BTT), a bridge to myocardial recovery, or as destination therapy. The continued development of LVADs during the last 10 years (miniaturization, low power use, and ease of implantation), declining costs, and improvements in associated care have made their wider use a reality.6 Prior to the adoption of such new technologies, it is essential to evaluate the evidence of their associated benefits, risks, and costs.7 This paper summarizes the results of a systematic review and economic evaluation commissioned by the Department of Health's Health Technology Assessment Programme in England and Wales to examine the clinical and cost-effectiveness of LVADs as a BTT for people with ESHF.8

Methods

We searched 18 electronic databases, including Medline, Cochrane, and Embase, from their inception to December 2004 for published and unpublished studies irrespective of language.8 Additional references were identified through searching bibliographies of related publications and through contact with experts and manufacturers of devices.

We assessed systematic reviews, randomized controlled trials (RCTs), controlled clinical trials, cohort studies, case series, and case studies that evaluated currently available LVADs used as a BTT for people aged 16 years and over with ESHF for inclusion. We placed an emphasis on including studies that used the most rigorous study designs and that included an appropriate comparator, whether another LVAD, heart transplantation, or usual care while on the heart transplant waiting list. We excluded studies if they used LVADs as an emergency rescue strategy during surgery or during the peri-operative period only. Included studies had to assess survival, functional capacity, or quality of life as primary outcome measures. We included studies reported as abstracts or conference presentations if sufficient information was presented on their methodology and results. We assessed the quality of studies that met the stated inclusion criteria using standard criteria.9,10 Two reviewers independently applied the inclusion criteria, extracted data, and assessed methodological quality. A third reviewer independently resolved any differences. Standard information was extracted wherever possible. We assessed clinical effectiveness through a narrative comparison of different outcomes. Differences in, or insufficient details on, study designs, outcomes used, patient characteristics, or intervention used precluded meta-analysis.

Our economic evaluation was a cost-utility analysis based on a 5-year decision analytic model examining the benefits and costs of the use of LVADs as a BTT compared with usual medical care on the transplant waiting list. The evaluation focused on devices that were thought to be clinically effective in the systematic review. We adopted the perspective of the UK National Health Service for costs and benefits for 2003 (UK prices). Efficacy was analysed in terms of a change in patient survival, derived from the systematic review of clinical effectiveness. Survival data with an indication of time to an event (e.g. transplant) was limited, with only two studies presenting Kaplan–Meier survival curves for the ‘bridged’ period.11,12 The base case analysis used the study by Aaronson et al.12 as it assessed survival to transplant, post-transplant, and overall survival and included a larger sample. Utility values originated from a study assessing patient utilities before and after LVAD implantation and post-heart transplantation using the standard gamble technique.13 Costs originate from published sources,1419 UK hospitals, and from device manufacturers, with resource-use scenarios developed from the systematic review of clinical effectiveness and expert advice specific to the UK setting (Table 1). We discounted costs at 6% and benefits at 1.5%. The 5-year time horizon adopted following LVAD implantation reflects the duration of the survival curve analysis.12 We carried out a full range of one-way sensitivity analyses and threshold analysis on the LVAD-related costs to yield a cost per quality-adjusted life year (QALY) of £30 000 (the generally accepted cost-effectiveness threshold for the UK).20 We generated confidence intervals around the mean using probabilistic sensitivity analysis (PSA), which randomly samples values from distributions imposed upon model parameters (costs, survival, and utilities) from published data. The confidence intervals represent the influence of parameter uncertainty upon the final cost-effectiveness outcome. These data are used to generate the cost-effectiveness acceptability curve (CEAC), which demonstrates the likelihood that the intervention is cost-effective across a range of thresholds.21,22

View this table:
Table 1

Component costs used in the economic evaluation

Cost (unit)Event rate per patientTotal cost per patient (£)
LVAD costs
Assessments (n=21)a£22021.31252891
Implant operation (n=16)a£36 986136986
Total charge excl. device56645
HeartMate/Novacor devicesb,c48000
MicroMed DeBakey deviceb35000
Jarvik deviced40948e
Heart transplant costs
Preparatory assessmentf3500
Operation1816769
Management following transplantation15,50Months 1–32105g
Months 4–12630g
Months 13–60484g
Ongoing pre-transplant management costs
On LVADMonthly1397
Medical managementMonthly251
  • aCaine N, Papworth Hospital NHS Trust, 2004.

  • bMicroMed Technology Inc., unpublished data, 2003.

  • cThoratec Europe Ltd, unpublished data, 2003.

  • dJarvik Heart, Inc., unpublished data, 2003.

  • eConversion of costs at $1.8316 (US)=£1 (UK).51

  • fLarge S, Papworth Hospital NHS Trust, 2003.

  • gInflated to 2003 prices using the hospital and community health services (HCHS) indices in Netten and Curtis.17

Results

Systematic review of clinical effectiveness

Quantity and quality of evidence

Eighteen studies assessing the clinical effectiveness of LVADs as a BTT met the inclusion criteria (Figure 1), with studies of the HeartMate implanted pneumatic and/or vented electric devices,11,12,2328 Novacor N100,29 Toyobo,30 Thoratec,31 Jarvik 2000,3238 and Micromed DeBakey.3942 Only seven studies, all of them HeartMate, included a comparator, namely the Novacor,24 inotropic agents,12,25,28 or undefined medical care.11,26,27 The methodological quality and the quality of reporting of studies, assessed using available reports, were judged to be weak providing the opportunity for bias (Table 2). The generalizability of the studies was assessed through a comparison of key study characteristics, although limited information was provided (see Supplementary material online, Tables S1S3). Studies tend to focus on male patients aged between 40 and 60 years who suffered from either dilated or ischaemic cardiomyopathy and were in New York Heart Association (NYHA) class IV.

Figure 1

Flowchart of describing study selection. From Clegg et al.,8 with permission.

View this table:
Table 2

Methodological quality and quality of reporting of studiesa

ReferenceStudy design (and comparator)Design and allocationSelectionConfoundingBlindingData collectionWithdrawalsIntention to treat
HeartMate LVAD
 El Banayosy et al.24Controlled clinical trial (Novacor LVAD)ModerateWeakStrongWeakWeakStrongCannot tell
 Aaronson et al.12Cohort analytic (inotropic drugs)ModerateWeakWeakWeakWeakNot applicableNot applicable
 Bank et al.25Cohort analytic (inotropic drugs)ModerateWeakWeakWeakWeakNot applicableNot applicable
 Morgan et al.28Cohort analytic (inotropic drugs)ModerateWeakWeakWeakWeakNot applicableNot applicable
 Massad et al.27Cohort analytic (medical care)ModerateWeakWeakWeakWeakNot applicableNot applicable
 Frazier et al.26Cohort (historical) (medical care)WeakWeakWeakWeakWeakStrongNot applicable
 Frazier et al.11Cohort (historical) (medical care)WeakWeakWeakWeakWeakWeakNot applicable
 Grady et al.23Cohort (before and after) (no comparator)WeakModerateNot applicableWeakStrongStrongNot applicable
Novacor N100 LVAD
 Trachiotis et al.29Cohort analytic (no comparator)ModerateWeakNot applicableWeakWeakNot applicableNot applicable
Toyobo LVAD
 Masai et al.30Case series (no comparator)WeakWeakNot applicableWeakWeakNot applicableNot applicable
Thoratec LVAD
 Holman et al.31Case report (no comparator)WeakWeakNot applicableWeakWeakNot applicableNot applicable
 May and Adams43Case report (no comparator)WeakWeakNot applicableWeakWeakNot applicableNot applicable
Jarvik 2000 LVAD
 Frazier et al.32bCase series (no comparator)WeakWeakNot applicableWeakWeakNot applicableNot applicable
 Frazier et al.38bCase series (no comparator)WeakWeakNot applicableWeakWeakNot applicableNot applicable
Micromed DeBakey LVAD
 Klotz et al.42Cohort (before and after) (no comparator)WeakWeakNot applicableWeakWeakNot applicableNot applicable
 Noon et al.39Cohort (before and after) (no comparator)WeakWeakNot applicableWeakWeakWeakCannot tell
 Popatov et al.41Cohort (before and after) (no comparator)WeakWeakNot applicableWeakWeakStrongYes
 Wieselthaler et al.40Cohort (before and after) (no comparator)WeakModerateNot applicableWeakWeakStrongYes
  • From Clegg et al.,8 with permission.

  • aThe different criteria were rated following the guidelines provided by Thomas et al.10 with some adaptations. Where criteria are rated as weak, moderate, or strong, it relates to the study's methodological control of the criteria. As such, if a study is rated as strong it indicates that there is a low risk of bias for the particular criteria. Where a component is considered not to be relevant to a study design, it is referred to as not applicable.

  • bThere may be some overlap between the two studies, although this was not formally identified.

Survival

Studies showed a survival benefit for people supported by an LVAD compared with usual medical care or the natural history of ESHF (Table 3). Comparison of the HeartMate LVAD with inotropic agents in three cohort analytic studies showed benefit for patients receiving the LVAD in survival and actuarial survival to transplantation and post-transplantation, although the difference in survival varied between the studies and the period of support and follow-up (range: 0–33%).12,25,28 When comparing patients who had received the HeartMate LVAD with those on usual care, differences were more equivocal. Although a retrospective cohort analytic study found no difference in survival to transplant and following transplantation,27 two retrospective cohort studies with historic controls found a benefit for patients who received an LVAD in survival to transplant (survival difference: 1426 and 59%,11 respectively) and survival post-transplantation (survival difference >60 days: 35%;26 difference actuarial survival 1 and 2 years: 100%11). Studies of other devices did not include a comparator and as a consequence only provide an indication of the additional survival gained from the LVAD compared with the poor prognosis evident from the natural history of the condition. The retrospective cohort analytic study of the Novacor LVAD compared post-transplantation survival of patients supported for <30 days with those supported for ≥30 days, with no difference at 3 years.29 A case series study of three patients with the Toyobo device showed a patient surviving to 64 days before dying, another patient surviving 119 days to transplantation, and a third still alive on the transplant waiting list at 390 days.30 Two case reports of the Thoratec LVAD showed two patients surviving 21 and 60 days before transplantation.31,43 Four cohort studies of the MicroMed Debakey LVAD showed that ∼80% of patients survived at least 30 days post-implantation, with between 34 and 40% undergoing heart transplantation and 10–31% dying while on LVAD support.3942 In two case series of the Jarvik 2000 LVAD, ∼70% were bridged to transplant with 30% having died while on LVAD support.32,44 Direct comparisons of the benefits of the different devices were limited. A controlled clinical trial compared the HeartMate LVAD with the Novacor LVAD showing limited difference in survival.24

View this table:
Table 3

Summary of evidence of effect of LVADs on patient survival

StudyMeasures of survivalLVADComparatorP
HeartMate LVAD
 El Banayosy et al.24Survival to transplantHeartMate 65% (12/20)Novacor 60% (13/20)Not stated
 Aaronson et al.12Survival to transplantHeartMate 73% (48/66)Inotropes 74% (28/38)Not stated
Actuarial survival to transplant
 1 monthHeartMate 81% (SD 5)Inotropes 78% (SD 8)
 3 monthsHeartMate 81% (SD 5)Inotropes 64% (SD 11)
 MedianHeartMate 2.9 monthsInotropes 2.9 monthsP=0.2
Post-transplant survival
 0 dayHeartMate 48 (78%)Inotropes 28 (74%)Not stated
 30 daysHeartMate 47 (71%)Inotropes 24 (63%)
 1 yearHeartMate 31 (47%)Inotropes 17 (45%)
 3 yearsHeartMate 9 (14%)Inotropes 6 (16%)
Post-transplantation actuarial survival
 1 yearHeartMate 98% (SD 2)Inotropes 74% (SD 9)Overall
 3 yearsHeartMate 95% (SD 4)Inotropes 65% (SD 10)P=0.007
 4 yearsHeartMate 95% (SD 4)Inotropes 65% (SD 10)
Overall actuarial survival
 1 yearHeartMate 80% (SD 5)Inotropes 56% (SD 8)
 3 yearsHeartMate 77% (SD 6)Inotropes 44% (SD 9)
 4 yearsHeartMate 77% (SD 6)Inotropes 44% (SD 9)P=0.03
 Bank et al.25Survival post-transplant (6 months)HeartMate 16 (89%)Inotropes 14 (74%)Not stated
 Morgan et al.28Actuarial survival
1 yearHeartMate 92%Inotropes 91%Overall
3 yearsHeartMate 84%Inotropes 84%Log rank
5 yearsHeartMate 74%Inotropes 73%P=0.649
 Massad et al.27Patients transplantedHeartMate 42 (80%)Medical Care 170 (84%)P=ns
Overall survivalHeartMate 48 (91%)Medical Care 174 (86%)P=ns
Actuarial survival
 1 yearHeartMate 94%Medical care: UNOS 1, 91%; UNOS 2, 86%P=ns
 3 yearsHeartMate 85%Medical care: UNOS 1, 86%; UNOS 2, 83%P=ns
 Frazier et al.26Patients transplantedHeartMate 17 (74%)Medical Care 3 (50%)Not stated
Survival >60 daysHeartMate 15 (65%)Medical Care 1 (17%)P<0.05
 Frazier et al.11TransplantationHeartMate 16 (84%)Medical Care 3 (25%)Not stated
Actuarial survival 1 and 2 yearsHeartMate 16 (100%)Medical Care 0 (0)P<0.05
Novacor N100 LVAD
 Trachiotis et al.29Survival post-transplant or explantSurvival of patients with LVAD for <30 daysSurvival of patients with LVAD for LVAD >30 days
 30 days 100% 100%Not stated
 1 year 100% 80%Not stated
 2 years 100% Not reportedNot stated
 3 years 60% 60%Not stated
Toyobo LVAD
 Masai et al.30Survival on LVADPatient 1: transplanted after 119 daysNo comparator
Patient 2: on support after 390 days
Patient 3: died after 64 days
Thoratec LVADs
 Holman et al.31Survival to transplantPatient 1: 60 daysNo comparator
 May and Adams43Survival post-transplantPatient 1: 21 daysNo comparator
Jarvik 2000 LVADs
 Frazier et al.32aSurvived to successful transplant2 (67%)No comparator
Frazier et al.38aSurvival to transplant13 (59%)No comparator
MicroMed DeBakey LVADs
 Klotz et al.42Total mortality3 (30%)No comparator
Pre-transplant mortality1 (10%)No comparator
1 year post-transplant mortality2 (20%)No comparator
 Noon et al.39Probability of survival at 30 days81%No comparator
Number transplanted11 (34%)No comparator
Deaths on support10 (31%)No comparator
 Wieselthaler et al.40Survived on LVAD 8 (80%)No comparator
Transplanted 4 (40%)No comparator
Awaiting transplant 4 (40%)No comparator
Died on LVAD 2 (20%)No comparator
 Popatov et al.41Number transplanted 2 (34%)No comparator
Deaths on support 1 (17%)No comparator
  • UNOS, united network organ sharing; ns=not significant. From Clegg et al.,8 with permission.

  • aThere may be some overlap between the two studies, although this was not formally identified.

Functional status

Two cohort studies of the HeartMate LVAD compared with historical controls showed that patients receiving an LVAD improved their functional status on the NYHA functional classification compared with those on usual care (LVAD: baseline 100% NYHA IV, 60 days post-transplant 94% NYHA I; usual medical care: baseline 83% NYHA IV, 60 days post-transplant 100% NYHA IV) or that expected given the natural history of the condition (LVAD: baseline 100% NYHA IV, 60 days post-transplant 94% NYHA I) (see Supplementary material online, Tables S1S3).11,26 Similar improvements in functional status were experienced by patients receiving a Jarvik 2000 LVAD (LVAD: baseline 100% NYHA IV, post-implantation 70% NYHA I).33

Quality of life

A cohort study of the HeartMate LVAD23 and a case series of the Jarvik 200045 provided the only evidence on changes in quality of life following device implantation (see Supplementary material online, Tables S1S3). Although patients with the HeartMate LVAD reported statistically significant improvements on the Heart Failure Symptom Checklist (change in total score from baseline −0.07, P=0.002), changes on the Quality of Life Index (change in total score from baseline +0.07, P=not significant), Rating Question Form (change in total score from baseline 0, P=not significant), and functional disability from the Sickness Impact Profile (change in total score from baseline +0.02, P=not significant) were equivocal.23 Implantation of the Jarvik 2000 LVAD significantly improved quality of life for patients on all scales of the Minnesota Living with Heart Failure Questionnaire (changes from baseline to 1-month post-implantation: quality of life −23.7; physical subscale −15; emotional subscale −6.5). Improvements continued up to heart transplantation, with a slight worsening subsequently (changes from baseline to 1-month post-heart transplantation: quality of life +12; physical subscale +10; emotional subscale −1.5).45

Adverse events

Comparison of the HeartMate and Novacor LVADs showed similar rates of mortality from adverse events.24 The HeartMate LVAD appeared to have a higher rate of device-related complications than the Novacor LVAD, including technical problems with controllers (70 vs. 10%, P<0.001), device-related infections (55 vs. 20%, P=0.02), and driveline infections (45 vs. 20%, P=0.09). Other studies of the HeartMate device showed ∼20% of patients having mechanical device failures.23,25 Patients with the Novacor LVAD suffered slightly higher rates of bleeding (40 vs. 35%, P=0.7), re-operation (30 vs. 20%, P=0.5) and thromboembolic events (20 vs. 0%, P=0.1) than those with a HeartMate LVAD.24 Other studies showed similar rates of thromboembolic events (026 and 5%11) and rates of bleeding (39%26) for the HeartMate device. Several studies showed that implantation of the HeartMate device resulted in right heart failure among ∼20% of patients24,26 and a study found that 40% of patients suffered psychiatric complications.23 Studies of the HeartMate LVAD showed lower rates of adverse events following heart transplantation than for those receiving usual care, with fewer patients suffering end organ failure, re-operation, rejection, disability, or infection.11,25,27 Although the adverse events experienced by recipients of other LVADs were less clearly reported, it was evident that patients of the Novacor and Thoratec devices suffered from infection, thromboembolism, and required re-operations as a consequence.29,31,43 Studies of the Jarvik 2000 LVAD reported technical problems with power cable connections breaking and incurring infections, but thromboembolic events, serious device infections, or haemolysis were not significant problems.32,33 Two cohort studies of the Micromed DeBakey LVAD reported a death, one from device-related sepsis41 and another from unspecified adverse events.39

Cost-effectiveness of LVADs as a BTT for ESHF

With no relevant cost-effectiveness studies available,8 an economic evaluation was produced to consider cost-effectiveness in the UK. Comparison of the use of the HeartMate LVAD with usual medical care (base case) showed that the LVAD offered an additional 18.3 (discounted) months of survival per patient over the 5-year duration of the model at an additional cost of £99 475 (discounted), producing a base case cost per QALY of £65 242 (Table 4).

View this table:
Table 4

Cost per quality adjusted life year gained

InterventionCost (£)Incremental cost (£)Effectiveness (quality adjusted life years)Incremental effectiveness (quality adjusted life years)Incremental cost effectiveness (£/QALY)
LVAD135587994753.61.52565242
Usual medical care361222.075

A range of one-way sensitivity analyses showed that the model is sensitive to survival gains, utility benefit in the pre-transplant period, and to the one-off costs associated with the LVAD (assessment, operation, and device costs) (Table 5). Threshold analysis showed that LVADs only became cost-effective (assumed to be £30 000/QALY) when total LVAD costs were around £37 500 per patient compared with a base case cost £87 877. PSA using stochastic simulation was undertaken over 1000 iterations to produce 95% confidence limits of £34 194–364 564 per QALY. Cost and effectiveness pairs from the PSA were used to generate the CEAC (Figure 2). The probability that the intervention is cost-effective at a cost/QALY of £30 000 is negligible.

Figure 2

Cost-effectiveness acceptability curve. From Clegg et al.,8 with permission.

View this table:
Table 5

Sensitivity analysis

Model parameter variedBase case valueRange of values in sensitivity analysisCost/QALY (£)
Base case65242
LVAD operation cost (£)3698629589–4438360391–70094
LVAD management cost per month (£) 13971118–167764399–66086
Heart transplant operation cost (£)141146511–2297064932–65603
Medical management cost per month (£)  251201–139759828–65478
Post-transplant management costsSee Table 1Base case minus 20% to base case plus 20%64442–66042
HeartMate/Novacor device cost (£)a4800038400–5760058946–71539
MicroMed DeBakey device (£)a3500028000–4200052125–61307
Jarvik device cost (£)a4098432758–4913755246–65988
Costs discount rate (%)63–964999–65517
Benefits discount rate (%)1.50–663343–70925
Assuming equal pre-transplant duration64667
Assuming no overall survival gain29.7246.90497708
Assuming equal proportions reaching transplant64667
Assuming no utility benefit in pre-transplant period72710
Post-transplantation utility0.9640.75–162866–84152
  • From Clegg et al.,8 with permission.

  • aDevice costs represent base case costs for the separate devices, however the base case used in the original model was £48 000 for the HeartMate LVAD.

Discussion

Our systematic review and economic evaluation have identified four important findings. First, LVADs appear beneficial as a BTT for patients with ESHF. They seem to benefit patients through improved survival, functional status, and quality of life when compared with inotropic agents, optimum care, or no care. However, their use is associated with serious adverse events, showing the need for improvements in patient care and design of devices. Second, there is insufficient evidence to discriminate between the different LVADs. The more established devices, such as the HeartMate, have a greater evidence base, but there are theoretical benefits to the newer non-pulsatile devices, such as the Jarvik 2000 and MicroMed Debakey. The benefits, including miniaturization, low power use, and ease of implantation, are beginning to be shown in practice. However, further investigations are needed to assess their long-term effectiveness, any adverse effects of non-pulsatility, and to take account of the continual developments in the devices, implantation techniques, and patient care that tend to occur with increased use.6,46 Third, that LVADs are not cost-effective with a cost per QALY of £65 242, which is outside the assumed threshold for acceptability within the UK of £30 000 per QALY. Sensitivity analysis has shown that substantial reductions in the cost of the devices and/or operation or improvements in the length and quality of life after heart transplantation would be required before LVADs would be cost-effective. Finally, a scarcity of evidence and several methodological shortcomings in the available studies limit the findings. LVADs are a relatively new and fast changing technology with limited diffusion. As a consequence, the studies of efficacy are often observational, retrospective, and non-comparative and data for populating economic evaluations are imperfect. With the pace of development likely to continue, considerable efforts will be needed to ensure appropriate assessment of the technology. Initiatives are underway to improve the availability of comparative data but further trials are needed.47 Although additional research may help to clarify the clinical and cost-effectiveness of these devices, it is clear that the lack of donor organs severely limits the opportunity for people to benefit from LVADs as a BTT.35,8 Heart donation has decreased by 14% in the USA48 and 63% in the UK8 between 1993 and 2003. Without an adequate provision of donor organs to meet the need and demand for heart transplantations,49 it is likely that the future of LVADs lies in their use as long-term chronic support (destination therapy) for people with ESHF.8

We applied consistent methods for undertaking systematic reviews and economic evaluations, with support throughout the process from an expert advisory group of clinicians, academics, and health professionals from the US and UK. We contacted manufacturers, the major specialist centres implanting LVADs in the UK, and authors of studies to try and obtain additional information, with varying success. Although the economic evaluation used the most up-to-date information available, it suffered from certain inadequacies including poor-quality evidence of clinical effectiveness, limited patient utility data, and cost data. It is likely that the poor evidence base reflects the fact that LVADs are a relatively new and fast changing technology, with limited diffusion. The pace of development is set to persist through the emergence of new devices and guidance for their use and management. It will be important to ensure that further primary research is appropriately conducted to high methodological standards and that this systematic review and economic evaluation is regularly updated.

Several areas for research have emerged, including the epidemiology of ESHF to assess incidence, prevalence, and characteristics of people; long-term (at least 2 years follow-up) RCTs focusing on head-to-head comparisons of different devices encompassing a broad range of patient groups and assessing patient survival, functional ability, quality of life, and adverse events; economic evaluations as part of clinical trials with prospective collection of data on quality of life, utilities, resources, and costs. Studies should assess the use of LVADs among different subgroups of people with ESHF, where clinical and cost-effectiveness may vary. In addition, given the declining availability of donor organs for heart transplantation and the potential for using LVADs as destination therapy for people with ESHF, it will be important to produce a systematic review of clinical and cost-effectiveness to assess the evidence base.

Conclusions

Although LVADs appear to be a clinically effective treatment as a BTT for people with ESHF, it is unlikely that they will be a cost-effective option unless the costs of the devices and their implantation are reduced substantially and/or the length and quality of life resulting from implantation are improved. With the limited and declining donation of hearts for transplantation, it appears that the future of the technology is in its use as destination therapy.

Supplementary material

Supplementary material is available at European Heart Journal online.

Acknowledgements

A.J.C. developed the research protocol; assisted in development of search strategy; assessed studies for inclusion in systematic review of clinical and cost-effectiveness; extracted data from, and quality assessed, included studies; synthesized evidence; assisted in development of economic model; collected data for economic evaluation; drafted and edited final report; and project managed study. D.A.S. developed research protocol; assessed studies for inclusion in systematic review of cost-effectiveness; extracted data from, and quality assessed, included studies; synthesized evidence; developed the economic model; collected data for economic evaluation; and drafted final report. E.L. developed research protocol; assessed studies for inclusion in systematic review of clinical effectiveness; extracted data from, and quality assessed, included studies; synthesized evidence; and drafted report. J.L.C. developed research protocol; assessed studies for inclusion in systematic review of clinical-effectiveness; extracted data from, and quality assessed, included studies; synthesized evidence; and drafted report. P.R. developed research protocol; developed literature search strategy; and drafted report. J.B. developed research protocol; assessed studies for inclusion in systematic review of clinical and cost-effectiveness; extracted data from, and quality assessed, included studies; and drafted report. We thank the advisory group for this study for advice and peer review of a draft of the original report for the NHS R&D HTA Programme, including: Prof. S. Ball, Leeds General Infirmary, Leeds, UK; Prof. Martin Buxton, University of Brunel, UK; Prof. Noreen Caine, Deputy Director of Research and Development, Department of Health, UK; Prof. J. Cleland, University of Hull, Hull, UK; Dr Mick Davies, University Hospital Birmingham NHS Trust, Birmingham, UK; Mr Stephen Large, Papworth Hospital NHS Trust, Cambridge, UK; Dr S. Ludgate, Medicines and Healthcare Products Regulatory Agency, London, UK; Dr P. McCarthy, Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic, Cleveland, USA; Dr A. Moskowitz, INCHOIR, Columbia University, New York, USA; Prof. John Pepper, Royal Brompton Hospital, London, UK; Mr L. Vale, Health Economics Research Unit, University of Aberdeen, Aberdeen, UK; Mr Stephen Westaby, John Radcliffe Hospital, Oxford, UK; Prof. Sir M.H. Yacoub, Imperial College, London, and The Magdi Yacoub Institute NHLI at Heart Science Centre, Harefield, Middlesex, UK. In addition, several people who provided data, translated papers, or assisted in extracting data for this study, including: J. Fawell, Papworth Hospital NHS Trust, Cambridge, UK; Liz Hodson, University of Southampton, UK; Dr Christine Clar, Germany; Sian Thomas, Orkney Isles, UK; Mariam Brazzelli, University of Aberdeen, Scotland; Sacci Thomas, Winchester, UK; Dr Goro Matsumiya, Osaka University Graduate School of Medicine, Japan; Dr Deereck Wheeldon, WorldHeart, UK; Dr Charles Glanville, Thoratec Europe Ltd, Cambridge, UK; Karen Brennan, Thoratec Europe Ltd, Cambridge, UK; Dr P. Jansen, Jarvik Heart, Inc., New York, USA; Dr Janine Meehan, Jarvik Heart, Inc., New York, USA; Dallas Anderson, MicroMed Technology, Inc., Houston, TX, USA; Yosuke Kobayashi, Toyobo Co. Ltd, Osaka, Japan; Dr Bill Gutteridge, Medical Adviser, NSCAG, Department of Health, London, UK; Dr N. Banner, Royal Brompton and Harefield NHS Trust, London, UK. ' This project was funded by the UK Department of Health through its Health Technology Assessment Programme (HTA project reference 01/12/02). The opinions and conclusions expressed here are those of the authors and do not necessarily reflect those of the UK National Health Service or the Department of Health'.

Conflict of interest: none declared.

References

View Abstract