Assessment of Mortality Among Durable Left Ventricular Assist Device Recipients Ineligible for Clinical Trials | Cardiology | JAMA Network Open | JAMA Network
[Skip to Navigation]
Figure 1.  Kaplan-Meier Estimated Mortality by Trial Eligibility Status
Kaplan-Meier Estimated Mortality by Trial Eligibility Status

Shaded areas indicate 95% CIs.

Figure 2.  Kaplan-Meier Estimated Mortality by Number of Ineligibility Criteria
Kaplan-Meier Estimated Mortality by Number of Ineligibility Criteria

Shaded areas indicate 95% CIs.

Figure 3.  Risk of Mortality by Ineligibility Criteria
Risk of Mortality by Ineligibility Criteria

Forest plot of specific trial criteria associated with increased risk of mortality according to multivariable Cox proportional hazards models for patients who were trial-eligible and met only 1 ineligibility criterion. HR indicates hazard ratio; INR, international normalized ratio; LVEF, left ventricular ejection fraction; MCS, mechanical circulatory support. To convert bilirubin to micromoles per liter, multiply by 17.104 and creatinine to micromoles per liter, multiply by 88.4.

Table 1.  Patient Characteristics by Trial Eligibility Status
Patient Characteristics by Trial Eligibility Status
Table 2.  Cox Proportional Hazards Model for Overall Mortality by Trial Eligibility Status and Number of Exclusions Met
Cox Proportional Hazards Model for Overall Mortality by Trial Eligibility Status and Number of Exclusions Met
1.
Teuteberg  JJ, Cleveland  JC  Jr, Cowger  J,  et al.  The Society of Thoracic Surgeons INTERMACS 2019 annual report: the changing landscape of devices and indications.   Ann Thorac Surg. 2020;109(3):649-660. doi:10.1016/j.athoracsur.2019.12.005 PubMedGoogle ScholarCrossref
2.
Goldstein  DJ, Meyns  B, Xie  R,  et al.  Third annual report from the ISHLT Mechanically Assisted Circulatory Support Registry: a comparison of centrifugal and axial continuous-flow left ventricular assist devices.   J Heart Lung Transplant. 2019;38(4):352-363. doi:10.1016/j.healun.2019.02.004 PubMedGoogle ScholarCrossref
3.
Slaughter  MS, Rogers  JG, Milano  CA,  et al; HeartMate II Investigators.  Advanced heart failure treated with continuous-flow left ventricular assist device.   N Engl J Med. 2009;361(23):2241-2251. doi:10.1056/NEJMoa0909938 PubMedGoogle ScholarCrossref
4.
Rogers  JG, Pagani  FD, Tatooles  AJ,  et al.  Intrapericardial left ventricular assist device for advanced heart failure.   N Engl J Med. 2017;376(5):451-460. doi:10.1056/NEJMoa1602954 PubMedGoogle ScholarCrossref
5.
Milano  CA, Rogers  JG, Tatooles  AJ,  et al; ENDURANCE Investigators.  HVAD: the ENDURANCE Supplemental Trial.   JACC Heart Fail. 2018;6(9):792-802. doi:10.1016/j.jchf.2018.05.012 PubMedGoogle ScholarCrossref
6.
Mehra  MR, Naka  Y, Uriel  N,  et al; MOMENTUM 3 Investigators.  A fully magnetically levitated circulatory pump for advanced heart failure.   N Engl J Med. 2017;376(5):440-450. doi:10.1056/NEJMoa1610426 PubMedGoogle ScholarCrossref
7.
Mehra  MR, Goldstein  DJ, Uriel  N,  et al; MOMENTUM 3 Investigators.  Two-year outcomes with a magnetically levitated cardiac pump in heart failure.   N Engl J Med. 2018;378(15):1386-1395. doi:10.1056/NEJMoa1800866 PubMedGoogle ScholarCrossref
8.
Mehra  MR, Uriel  N, Naka  Y,  et al; MOMENTUM 3 Investigators.  A fully magnetically levitated left ventricular assist device—final report.   N Engl J Med. 2019;380(17):1618-1627. doi:10.1056/NEJMoa1900486 PubMedGoogle ScholarCrossref
9.
Aaronson  KD, Slaughter  MS, Miller  LW,  et al; HeartWare Ventricular Assist Device (HVAD) Bridge to Transplant ADVANCE Trial Investigators.  Use of an intrapericardial, continuous-flow, centrifugal pump in patients awaiting heart transplantation.   Circulation. 2012;125(25):3191-3200. doi:10.1161/CIRCULATIONAHA.111.058412 PubMedGoogle ScholarCrossref
10.
Rose  EA, Gelijns  AC, Moskowitz  AJ,  et al; Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) Study Group.  Long-term use of a left ventricular assist device for end-stage heart failure.   N Engl J Med. 2001;345(20):1435-1443. doi:10.1056/NEJMoa012175 PubMedGoogle ScholarCrossref
11.
Stevenson  LW, Pagani  FD, Young  JB,  et al.  INTERMACS profiles of advanced heart failure: the current picture.   J Heart Lung Transplant. 2009;28(6):535-541. doi:10.1016/j.healun.2009.02.015 PubMedGoogle ScholarCrossref
12.
Birati  EY, Hanff  TC, Maldonado  D,  et al.  Predicting long term outcome in patients treated with continuous flow left ventricular assist device: the Penn-Columbia risk score.   J Am Heart Assoc. 2018;7(6):e006408. doi:10.1161/JAHA.117.006408 PubMedGoogle Scholar
13.
Cowger  J, Sundareswaran  K, Rogers  JG,  et al.  Predicting survival in patients receiving continuous flow left ventricular assist devices: the HeartMate II risk score.   J Am Coll Cardiol. 2013;61(3):313-321. doi:10.1016/j.jacc.2012.09.055 PubMedGoogle ScholarCrossref
14.
Boyle  AJ, Ascheim  DD, Russo  MJ,  et al.  Clinical outcomes for continuous-flow left ventricular assist device patients stratified by pre-operative INTERMACS classification.   J Heart Lung Transplant. 2011;30(4):402-407. doi:10.1016/j.healun.2010.10.016 PubMedGoogle ScholarCrossref
15.
U.S. Food and Drug Administration.  Real-world evidence. Accessed December 31, 2019. https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence
16.
21st Century Cures Act, Pub L No. 114-255 (2016). Accessed November 18, 2020. https://www.govinfo.gov/app/details/PLAW-114publ255
17.
Centers for Medicare and Medicaid Services.  Decision memo for ventricular assist devices for bridge-to-transplant and destination therapy (CAG-00432R). Medicare Coverage Database. Accessed December 31, 2019. https://www.cms.gov/medicare-coverage-database/details/nca-decision-memo.aspx?NCAId=268
18.
Miller  LW, Pagani  FD, Russell  SD,  et al; HeartMate II Clinical Investigators.  Use of a continuous-flow device in patients awaiting heart transplantation.   N Engl J Med. 2007;357(9):885-896. doi:10.1056/NEJMoa067758 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Health Policy
    January 8, 2021

    Assessment of Mortality Among Durable Left Ventricular Assist Device Recipients Ineligible for Clinical Trials

    Author Affiliations
    • 1Department of Cardiac Surgery, Michigan Medicine, University of Michigan, Ann Arbor
    • 2Center for Healthcare Outcomes and Policy, University of Michigan, Ann Arbor
    • 3Division of Cardiovascular Medicine, Department of Internal Medicine, Michigan Medicine, University of Michigan, Ann Arbor
    • 4Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor
    • 5Department of Health Management and Policy, School of Public Health, University of Michigan, Ann Arbor
    JAMA Netw Open. 2021;4(1):e2032865. doi:10.1001/jamanetworkopen.2020.32865
    Key Points

    Question  What proportion of real-world recipients of left ventricular assist devices (LVADs) meet clinical trial eligibility, and what prognoses are associated with ineligibility for randomized clinical trials?

    Findings  In this cohort study of 14 679 real-world recipients of durable LVADs, a total of 43.8% were ineligible for trial inclusion. Each additional criterion of ineligibility met by participants was associated with an increase in risk of mortality.

    Meaning  These findings suggest that while treatment for individuals ineligible for clinical trials should be weighed against medical management, more consideration could be given to designing trials with eligibility criteria that reflect real-world experience.

    Abstract

    Importance  While wide-scale adoption of durable left ventricular assist devices (LVADs) can be attributed to favorable randomized clinical trial outcomes, restrictive selection criteria may be associated with a lack of generalizability to real-world experience.

    Objective  To estimate the proportion of LVAD recipients who are eligible for clinical trials and to assess whether an association exists between trial eligibility and mortality.

    Design, Setting, and Participants  This cohort study examined 14 679 patients undergoing primary, intracorporeal continuous-flow LVAD implantation (with or without a right ventricular assist device) in 181 North American centers from January 1, 2012, to June 30, 2017, identified in the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS). To simulate a trial population, trial criteria from the Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Therapy With HeartMate 3 (MOMENTUM 3) were mapped to INTERMACS variables. Patients were categorized as eligible for trial inclusion or ineligible for trial inclusion and by number of ineligibility criteria met. Follow-up in INTERMACS was complete for all patients through October 31, 2017. Data were analyzed from July 2019 through November 2020.

    Exposures  Undergoing durable LVAD implantation.

    Main Outcomes and Measures  Trial eligibility and postimplant mortality were analyzed using Kaplan-Meier estimates and Cox proportional hazards models.

    Results  Among 14 679 recipients, mean (SD) age was 57 (13) years, 11 503 individuals (78.4%) were men, and 11 406 individuals (77.7%) presented with New York Heart Association class IV heart failure. A total of 6429 recipients (43.8%) were ineligible for trial inclusion, of whom 4226 individuals (65.7%) met 1 ineligibility criterion, 1442 individuals (22.4%) met 2 criteria, and 761 individuals (11.8%) met 3 or more criteria. Estimated mortality for recipients who were trial-ineligible was higher than for recipients who were trial-eligible (1-year mortality: 25.3% [95% CI, 24.2%-26.5%] vs 16.2% [95% CI, 15.4%-17.1%]; 3-year mortality: 42.8% [95% CI, 41.3%-44.4%] vs 36.4% [95% CI, 35.0%-37.8%]; log-rank P < .001 for both). Patients who were trial-ineligible had increased risk of mortality compared with patients who were trial-eligible if they met 1 trial ineligibility criterion (hazard ratio [HR], 1.16 [95% CI, 1.08-1.24]; P < .001), 2 trial ineligibility criteria (HR, 1.51 [95% CI, 1.36-1.67]; P < .001), or 3 or more trial ineligibility criteria (HR, 2.09 [95% CI, 1.84-2.39]; P < .001). Among patients meeting only 1 ineligibility criterion, 4 criteria were independently associated with mortality: prior or ongoing mechanical circulatory support (HR, 1.63 [95% CI, 1.23-2.16]; P = .001), elevated creatinine level (HR, 1.42 [95% CI, 1.17-1.72]; P < .001), elevated bilirubin level (HR, 1.39 [95% CI, 1.17-1.66]; P < .001), and low albumin or prealbumin level (HR, 1.18 [95% CI, 1.05-1.33]; P = .007).

    Conclusions and Relevance  These findings suggest that while treatment for patients who are ineligible for LVAD trial inclusion should be weighed against medical management, more consideration could be given to designing trials with eligibility criteria that reflect real-world experience.

    Introduction

    Durable left ventricular assist device (LVAD) implantation is an accepted surgical treatment for advanced heart failure refractory to guideline-directed medical therapy, with more than 25 000 implantations in the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) to date.1,2 The rise in use of durable LVAD has followed clinical trials3-9 of newer continuous-flow rotary pumps demonstrating improved outcomes with fewer adverse events. These clinical trial outcomes4-10 were followed by US Food and Drug Administration (FDA) approval for commercial use for both bridge-to-transplant and destination therapy indications, and this approval has broadened the application of LVAD therapy. To date, use of LVAD therapy in practice remains guided by clinical trial findings, which are derived from a highly selected pool of patients.

    However, the extent to which recipients of durable LVADs in practice meet clinical trial eligibility criteria is unknown. If real-world patients receiving LVADs have different outcomes than patients in clinical trials, clinicians should understand these differences when informing patients about the expected benefits and prognosis of LVAD therapy. Applying trial eligibility criteria to real-world recipients of durable LVADs may help determine the impact of trial inclusion and exclusion criteria on mortality. This may help clinicians better understand whether clinical trial outcomes can be generalized to patients who would have been ineligible for the trials that led to the commercial availability of these devices.

    The objectives of this study were to determine the generalizability of LVAD clinical trial findings by (1) using INTERMACS to characterize trial eligibility among real-world recipients of LVADs, (2) comparing mortality between recipients who were eligible for trial inclusion and those who were ineligible, and (3) examining associations of specific trial-ineligibility criteria with long-term mortality to inform patient prognosis. We hypothesized that LVAD recipients who did not meet trial criteria would have increased risk and worse risk-adjusted outcomes compared with recipients who met trial criteria.

    Methods
    Data Source

    The University of Michigan Institutional Review Board approved this cohort study and waived the requirement for informed consent because it met the board’s 4 criteria for a waiver: (1) the research involved no more than minimal risk to participants; (2) research could not practicably (ie, feasibly) be carried out without the waiver or alteration; (3) the waiver or alteration did not adversely affect the rights or welfare of the study participants; and (4) whenever appropriate, the participants or their legally authorized representative were provided with additional pertinent information after participation. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. The primary data source was INTERMACS, previously funded in part by the National Heart, Lung, and Blood Institute (NHLBI) under contract No. HHSN268201100025C and a joint effort among the NHLBI, FDA, and Centers for Medicare & Medicaid Services (CMS). The registry was established in 2005 at the University of Alabama at Birmingham and is currently administered by The Society of Thoracic Surgeons. This study was undertaken prior to the acquisition of INTERMACS by The Society of Thoracic Surgeons. The registry is prospectively maintained and includes a real-world population of patients undergoing FDA-approved implantation of durable mechanical circulatory support (MCS) devices, including LVADs, right ventricular assist devices, and total artificial hearts. The study period coincided with 3 trials: ENDURANCE,4 ENDURANCE supplement,5 and the Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Therapy with HeartMate 3 (HM3) (MOMENTUM 3).6-9 Some HeartMate II (HMII; Abbott Laboratories) devices in INTERMACS may have been in MOMENTUM 3 as control devices, but MOMENTUM 3 study devices (ie, HM3s; Abbott Laboratories) were not included in INTERMACS. Although INTERMACS has a variable for study inclusion, the variable is frequently missing and unverifiable, so we were unable to determine which patients were in a clinical trial.

    The data set for this study was provided by the INTERMACS Data and Clinical Coordinating Center to the University of Michigan through permission from the NHLBI. The study was performed under a data use agreement with the Research Data Assistance Center.

    Patient Population

    Patients undergoing primary, intracorporeal continuous-flow LVAD implantation (with or without right ventricular assist device) between January 1, 2012, and June 30, 2017, were identified in INTERMACS. Patients receiving an isolated right ventricular assist device or total artificial heart and those with a history of durable MCS device implantation when entered into INTERMACS were excluded. Each patient was categorized as eligible for trial inclusion or ineligible for trial inclusion through 2 analyses, using inclusion and exclusion criteria from MOMENTUM 3 applied to INTERMACS variables (eTable 1 in the Supplement). Eligibility criteria in MOMENTUM 3 included 7 inclusion and 25 exclusion criteria.6

    The primary (ie, limited) analysis included defined trial inclusion and exclusion criteria that could be directly mapped to INTERMACS (eg, bilirubin >2.5 mg/dL [to convert to micromoles per liter, multiply by 17.104] and creatinine >2.5 mg/dL [to convert to micromoles per liter, multiply by 88.4]) (eTable 2 in the Supplement). We excluded 2 inclusion criteria that were not feasible to match to INTERMACS variables: signing an informed consent form and female patients of childbearing age agreeing to use adequate contraception. The primary analysis captured all 5 remaining MOMENTUM 3 inclusion criteria and 11 of 25 exclusion criteria.

    To use INTERMACS variables to identify more patients who would have been eligible for the MOMENTUM 3 trial, an additional analysis (ie, comprehensive) was performed matching MOMENTUM and INTERMACS variables that were similar but not identical (eg, MOMENTUM exclusion variable: presence of an active, uncontrolled infection; INTERMACS exclusion variable: events this hospitalization — major infection [yes]) (eTable 1 in the Supplement). The comprehensive analysis captured the same 5 inclusion criteria as the primary analysis and 21 of 25 exclusion criteria.

    Outcomes

    The primary outcome was trial eligibility. Patients failing to meet 1 or more inclusion criteria or meeting 1 or more exclusion criteria were considered trial-ineligible, while patients meeting all inclusion criteria and no exclusion criteria were considered trial-eligible. Number of reasons for trial ineligibility was determined for each patient, and patients were categorized by number of ineligibility criteria they met (ie, 0, 1, 2, or ≥3).

    The secondary outcome was postimplant mortality. Time-to-event analyses specified death as a failure event, while patients who underwent heart transplantation, explant without exchange, or LVAD decommissioning with no new device or had other reason for LVAD removal were censored. Patients undergoing LVAD exchange remained in the analysis and were not censored. Death status and event dates were collected from INTERMACS, with complete follow-up performed for all patients through October 31, 2017.

    Statistical Analysis

    Bivariate comparisons used 2-sample t tests or χ2 tests, as appropriate. Missing preoperative variables were treated as a separate category. Missing values for trial exclusion criteria were considered not present, which provided the most conservative estimates of trial ineligibility.

    Time-to-event analyses were performed using log-rank test and Kaplan-Meier estimates to compare cumulative mortality by trial eligibility status (ineligible vs eligible), number of trial ineligibility criteria per patient (ie, 0, 1, 2, or ≥3), INTERMACS patient profile of advanced heart failure number (ie, 1, 2, 3, or 4-7),11 and trial eligibility status stratified by INTERMACS profile. Sensitivity analyses included comparing mortality by eligibility status after stratifying by device type and treating transplantation as a competing risk to mortality rather than a censored event.

    Multivariable Cox models incorporating clinical characteristics and device type were created to report risk-adjusted mortality with hazard ratios (HRs) and 95% CIs; specific hospital for procedure was included as a random effect through a shared frailty model. Cox models were created to compare patients ineligible for trial vs those eligible for trial (eTable 3 in the Supplement), then to compare patients meeting 1, 2, or 3 or more trial ineligibility criteria. The number of trial ineligibility criteria met (ie, 0, 1, 2, or ≥3) was considered as both a categorical and continuous variable in separate analyses. Lastly, the most frequent reasons for exclusion were evaluated for association with mortality through Cox models in 3 ways: with any number of exclusions present, when each exclusion was the only criterion present, and when the specific criterion plus 1 or more criteria were present. All models were adjusted for age, sex, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), blood type O, White race, implantable cardioverter defibrillator presence, INTERMACS patient profile, bridge-to-transplant status, prior coronary artery bypass or valve surgery, concomitant cardiac surgery, device type, and year of implant.

    Statistical significance was defined as P < .05, and P values were 2-sided. All analyses were performed in Stata statistical software version 16.0 (StataCorp) from July 2019 to November 2020.

    Results
    Patient Population and Trial Eligibility

    Among 14 679 recipients of LVAD, mean (SD) age was 57 (13) years, 11 503 (78.4%) individuals were men, 9704 individuals (66.1%) were White, mean (SD) BMI was 28.6 (7.1), and 11 406 individuals (77.7%) presented with New York Heart Association class IV heart failure (Table 1). In total, 2489 patients (17.0%) were INTERMACS patient profile 1, 5041 patients (34.3%) were profile 2, 4879 patients (33.2%) were profile 3, and 2186 patients (14.9%) were profiles 4 through 7. The HMII was implanted in 10 652 patients (72.6%), and HeartWare LVAD (Medtronic) was implanted in 4027 patients (27.4%).

    In total, 6429 patients (43.8%) were trial-ineligible in the limited analysis and 7888 patients (53.7%) were trial-ineligible in the comprehensive analysis (eFigure 1 in the Supplement). The percentage of patients who were ineligible for trial inclusion was consistent over the study period in these analyses (eFigure 2 in the Supplement). In the limited analysis, 8250 patients (56.2%) met no criteria conferring ineligibility (eg, were trial-eligible), 4226 patients (28.8%) met 1 ineligibility criterion, 1442 patients (9.8%) met 2 ineligibility criteria, and 761 patients (5.2%) met 3 or more ineligibility criteria (eFigure 3 in the Supplement). In the comprehensive analysis, 6791 patients (46.3%) met no ineligibility criteria, 4563 patients (31.1%) met 1 ineligibility criterion, 2015 patients (13.7%) met 2 ineligibility criteria, and 1310 patients (8.9%) met 3 or more ineligibility criteria (eFigure 3 in the Supplement). The most common reasons for trial ineligibility in both analyses were low albumin or prealbumin levels (2281 patients [15.5%]), thrombocytopenia (1369 patients [9.3%]), and elevated bilirubin levels (1293 patients [8.8%]) (eFigure 4 in the Supplement).

    A greater proportion of patients who were trial-ineligible, compared with patients who were trial-eligible, had INTERMACS profile 1 (1813 patients [28.2%] vs 676 patients [8.2%]; P < .001) and underwent other concomitant cardiac surgery (3367 patients [52.4%] vs 2908 patients [35.2%]; P < .001). A greater proportion of patients who were trial-ineligible , compared with patients who were trial-eligible, had a left ventricular ejection fraction of less than 20% (4078 patients [63.4%] vs 5688 patients [68.9%]; P < .001), and a smaller proportion of patients had an implantable cardioverter defibrillator (4674 patients [72.7%] vs 6873 patients [83.3%]; P < .001) (Table 1).

    Postimplant Mortality

    In the limited analysis, 1-year mortality was higher for patients who were trial-ineligible compared with those who were trial-eligible (25.3% [95% CI, 24.2%-26.5%] vs. 16.2% [95% CI, 15.4%-17.1%]; P < .001) (Figure 1A). Similarly for the comprehensive analysis, 1-year mortality was higher for patients who were trial-ineligible compared with those who were trial-eligible (25.1% [95% CI, 24.1%-26.1%] vs. 14.5% [95% CI, 13.6%-15.5%]; P < .001) (Figure 1B). Additionally, 3-year mortality was higher for patients who were trial-ineligible compared with those who were trial-eligible in the limited analysis (42.8% [95% CI, 41.3%-44.4%] vs 36.4% [95% CI, 35.0%-37.8%]; log-rank P < .001) and the comprehensive analysis (43.8% [95% CI, 42.3%-45.2%] vs 33.9% [95% CI, 32.3%-35.4%]; log-rank P < .001). With cardiac transplantation as a competing risk, the cumulative incidence of mortality was higher for patients who were trial-ineligible compared with those who were trial-eligible in both analyses (eFigure 5 in the Supplement). These mortality differences were comparable between HMII and HeartWare VAD devices (eFigure 6 in the Supplement). Median (interquartile interval) follow-up was 1.2 (0.5-2.4) years, and this was longer for patients who were trial-eligible compared with those who were trial-ineligible (1.3 [0.6-2.5] years vs 1.1 [0.5-2.3] years; P < .001). Trial ineligibility was independently associated with increased risk of mortality in the limited analysis (HR, 1.29 [95% CI, 1.21-1.37]; P < .001) and comprehensive analysis (HR, 1.42 [95% CI, 1.33-1.52]; P < .001) (Table 2; eTable 3 in the Supplement).

    Number of ineligibility criteria per patient was associated with mortality in both analyses, with higher estimated mortality for each additional criterion for trial ineligibility (Figure 2). Estimated mortality at 1 year for patients meeting 3 or more ineligibility criteria was 39.9% (95% CI, 36.3%-43.7%) in the limited analysis and 37.6% (95% CI, 34.9%-40.4%) in the comprehensive analysis; estimated 1-year mortality for patients who were trial-eligible was 16.2% [95% CI, 15.4%-17.1%] in the limited analysis and 14.5% [95% CI, 13.6%-15.5%] in the comprehensive analysis. The full distribution of criteria determining trial ineligibility is detailed in eTable 4 in the Supplement.

    An increased number of trial ineligibility criteria was associated with increased mortality risk for the limited analysis (1 criterion: HR, 1.16 [95% CI, 1.08-1.24]; P < .001; 2 criteria: HR, 1.51 [95% CI, 1.36-1.67]; P < .001; ≥3 criteria: HR, 2.09 [95% CI, 1.84-2.39]; P < .001) and comprehensive analysis (1 criterion: HR, 1.29 [95% CI, 1.19-1.38]; P < .001; 2 criteria: HR, 1.54 [95% CI, 1.40-1.69]; P < .001; ≥3 criteria: HR, 2.10; [95% CI, 1.88-2.34]; P < .001). Mean increased mortality risk per additional ineligibility criterion was similar in the 2 analyses (Table 2). Among patients meeting only 1 ineligibility criterion in the limited analysis, 4 ineligibility criteria were independently associated with mortality: prior or ongoing MCS other than intra-aortic balloon pump (HR, 1.63 [95% CI, 1.23-2.16]; P = .001), serum creatinine greater than 2.5 mg/dL (HR, 1.42 [95% CI, 1.17-1.72]; P < .001), bilirubin greater than 2.5 mg/dL (HR, 1.39 [95% CI, 1.17-1.66]; P < .001), and low albumin or prealbumin (HR, 1.18 [95% CI, 1.05-1.33]; P = .007); no independent association was found for meeting any other single criterion (Figure 3). Meeting any 2 or more ineligibility criteria was associated with an increased risk of mortality (eTable 5 in the Supplement). In addition, meeting 3 or more ineligibility criteria was associated with an even higher risk of mortality, with the specific combination of low albumin or prealbumin, elevated bilirubin, and elevated creatinine levels associated with the worst prognosis (eTable 6 in the Supplement).

    Among INTERMACS profiles, 1-year mortality was highest for profile 1 (28.8% [95% CI, 26.9%-30.8%]; log-rank P < .001) compared with profile 2 (20.6% [95% CI, 19.4%-21.8%]), profile 3 (16.8% [95% CI, 15.7%-17.9%]), and profiles 4 through 7 (17.4% [95% CI, 15.8%-19.2%]), with no significant difference between profile 3 and profiles 4 through 7 (eFigure 7 in the Supplement). Stratified by profile, 1-year mortality in the limited analysis differed between patients who were trial-ineligible vs those who were trial-eligible among patients with profile 1 (33.0% [95% CI, 30.8%-35.4%] vs 17.2% [95% CI, 14.3%-20.5%]; P < .001), profile 2 (25.3% [95% CI, 23.4%-27.3%] vs 17.1% [95% CI, 15.7%-18.6%]; P < .001), and profile 3 (20.5% [95% CI, 18.4%-22.8%] vs 15.2% [95% CI, 14.0%-16.5%]; P < .001) (eFigure 8 in the Supplement). In contrast, mortality did not differ by trial eligibility status among patients with profiles 4 through 7 (18.5% [95% CI, 16.1%-21.3%] vs 16.6% [95% CI, 14.5%-18.9%]; P = .44) (eFigure 8 in the Supplement). In the comprehensive analysis, patients who were trial-ineligible in all 4 groups of profiles had higher mortality compared with patients who were trial-eligible (eFigure 9 in the Supplement).

    Discussion

    This cohort study of nearly 15 000 patients receiving primary, durable LVAD implants in INTERMACS found that approximately one-half of patients were ineligible for clinical trial enrollment based on MOMENTUM 3 trial criteria. Ineligible status was associated with higher mortality compared with being eligible for trial inclusion, with mortality estimates of patients who were trial-eligible comparable to postimplant mortality rates published in recent clinical trials.5-8 Additionally, prior or ongoing MCS (other than intra-aortic balloon pump), elevated creatinine levels, elevated bilirubin levels, and low albumin or prealbumin levels were isolated ineligibility criteria independently associated with mortality. Meeting any other single ineligibility criterion was not associated with increased risk of mortality. In contrast, meeting any 2 or more ineligibility criteria was associated with increased risk of mortality compared with being eligible for trial inclusion, with each additional ineligibility criterion associated with increased mortality risk, independent of device type. Collectively, these data suggest that not meeting clinical trial eligibility criteria was associated with different mortality rates compared with being eligible for trial inclusion and that meeting increased numbers of ineligibility criteria was associated with progressively increased mortality.

    Multiple landmark clinical trials3-9 have demonstrated improved outcomes and fewer adverse events with the use of newer axial and centrifugal, continuous-flow, durable LVAD devices for both bridge-to-transplant and destination therapy. Most recently, in the MOMENTUM 3 trial,6-8 a totally magnetically levitated centrifugal-flow LVAD (HM3) was demonstrated to be superior to an axial-flow LVAD (HMII), with 77% of patients free from death, disabling stroke, or reoperation at 2 years. While clinical trial outcomes have been promising, the generalizability of these findings to the current real-world LVAD population has remained unclear. By evaluating generalizability within a contemporaneous real-world cohort, this study’s data may not only inform patient prognosis but may also be used for future clinical trial design, treatment decision-making, and informing regulatory policy for LVAD use and reimbursement.

    Importantly, in addition to a higher risk of mortality for patients who were trial-ineligible, the number of ineligibility criteria (ie, meeting 1, 2, or 3 or more criteria) and the specific types of ineligibility criteria varied in their association with mortality. The expected findings that temporary MCS other than intra-aortic balloon pump, elevated bilirubin or creatinine levels,12 and low albumin or prealbumin levels in isolation each conferred worse mortality in patients who were otherwise-eligible for trial inclusion should be used by clinicians implanting commercial LVADs to improve prognostication and facilitate shared decision-making. Despite the knowledge that ineligibility factors, such as low albumin or prealbumin and elevated bilirubin levels, are associated with mortality and morbidity, the prevalence of trial ineligibility among recipients of LVADs has not changed over time.13 The FDA and clinical trialists may use findings from this study to refine future trial eligibility criteria to be more representative of real-world LVAD recipients and may consider eliminating criteria that may have no association with mortality (eg, thrombocytopenia). Understanding the association of each eligibility criterion with mortality may allow clinicians to better prognosticate, individualize treatment plans, and counsel candidates for LVAD on risks and benefits of therapy relative to trial outcomes.

    A 2011 clinical trial14 demonstrated increasing mortality risk with increasing severity of illness based on INTERMACS profiles. Specifically, that study found that patients with less acute illness (ie, profiles 4-7) had superior survival and shorter length of stay compared with patients with more acute illness (ie, profiles 1 and 2). As expected, mortality in our study was highest for patients with INTERMACS profile 1, followed by profile 2, but there was no significant difference between profile 3 and profiles 4 through 7. Interestingly, 1-year mortality rates among patients who were trial-eligible with INTERMACS Profiles 1, 2, 3, and 4 through 7 in this study were all comparable with 1-year survival found in ENDURANCE II (16.9% in the HeartWare LVAD group and 17.8% in the HMII control group)5 and MOMENTUM 3 (15.9% for the HMII control group).8 These data for mortality by trial eligibility and profile number in the limited and comprehensive analyses suggest that ineligibility criteria are associated with prognosis, independent of INTERMACS profile.

    Our findings may also have important implications for regulators and payers. To provide estimates of benefit closer to real-world experience, regulators, such as the FDA, could limit the number and types of exclusion criteria used in clinical trials. Clinical registries could be designed to more closely follow patients receiving commercial durable LVAD devices who were ineligible for clinical trials to understand outcomes in these patients, potentially by mandating registry arms for patients who are trial-ineligible. Regulators may wish to incorporate these data to support coverage decisions, develop guidelines and decision support tools, monitor postmarket safety through more robust postmarket approval studies, and develop innovative treatment approaches, as required by the 21st Century Cures Act of 2016.15,16 Payers could also consider our findings in future payment coverage determinations. For instance, CMS originally based reimbursement for LVAD therapy under the National Coverage Determination on inclusion criteria from the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial.10,17 Incorporating trial criteria in coverage determinations may be associated with reimbursement of a smaller population, which may be associated with reduced incentives among clinicians of using durable LVAD therapy in patients at higher risk of mortality.

    However, with a 1-year mortality of approximately 25%, patients who were trial-ineligible in this study may still benefit from LVAD implantation over medical therapy. Connecting exclusion criteria to reimbursement could also be associated with increased incentives among trialists to limit exclusion criteria in trials to expand the patient population eligible for reimbursement, which may provide more real-world estimates of benefit and may ultimately benefit a larger group of patients undergoing durable LVAD therapy instead of medical therapy, which is associated with an extremely poor prognosis. More informed regulatory and payment decisions are critical for optimizing both patient outcomes and the societal burden of durable LVAD therapy.

    Limitations

    This study has several limitations. First, participants in the MOMENTUM 3 trial receiving the HM3 device and patients receiving devices not approved by the FDA were not included in the analysis. While this study found worse outcomes for recipients of HMII or HeartWare LVAD who were trial-ineligible, the most recent INTERMACS report found that HM3 outcomes were comparable between patients in MOMENTUM 3 and those in INTERMACS.1 Although recipients of HM3 were not included in this study, patients receiving HMII as part of the control device group in MOMENTUM 3 were included in INTERMACS and are represented here. Additionally, exclusion criteria among major LVAD trials3-9,18 have remained relatively consistent, and our intention in choosing MOMENTUM 3 criteria was to most accurately inform present-day device selection. Furthermore, when these trial criteria were applied temporally within this study population, real-world recipients of LVAD were just as likely to be ineligible for trial in 2012 as in 2017. Second, not every trial criterion from MOMENTUM 3 could be implemented through INTERMACS. However, 2 sets of criteria (limited and comprehensive) were used to together provide a full assessment of trial eligibility in this population. Third, this analysis focuses exclusively on mortality, while other outcomes, such as disabling stroke, permanent dialysis, and patient-reported outcomes, including quality of life, must be considered to fully characterize the utility and drawbacks of durable LVAD therapy.

    Conclusions

    The findings of this cohort study suggest that nearly one-half of real-world LVAD recipients did not meet current trial criteria. This ineligibility for trial inclusion was associated with increased mortality, and prior or ongoing MCS, elevated bilirubin or creatinine levels, and low albumin or prealbumin levels were independently associated with mortality. Patients meeting any 3 or more ineligibility criteria had more than 2-fold the risk of mortality compared with patients who were trial-eligible. Importantly, these data demonstrated an association between trial ineligibility criteria and postimplant prognosis, independent of baseline heart failure severity. However, these patients who were trial-ineligible may still be considered for potential LVAD therapy by comparing prognosis with medical management alone. Clinical trialists, regulators, and payers should consider the generalizability and prognostic value of individual eligibility criteria to further refine LVAD trial populations. Likewise, clinicians should use trial criteria to inform clinical decision-making when evaluating LVAD candidates and prognosticating in real-world practice.

    Back to top
    Article Information

    Accepted for Publication: November 12, 2020.

    Published: January 8, 2021. doi:10.1001/jamanetworkopen.2020.32865

    Correction: This article was corrected on June 21, 2020, to fix errors in the colors in the key for Figure 2.

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Brescia AA et al. JAMA Network Open.

    Corresponding Author: Alexander A. Brescia, MD, MSc, Department of Cardiac Surgery, Michigan Medicine, University of Michigan, Frankel Cardiovascular Center, 1500 E. Medical Center Dr, Ann Arbor, MI, 48109 (abrescia@med.umich.edu).

    Author Contributions: Drs Thompson and Brescia had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Brescia and Watt contributed equally to this manuscript.

    Concept and design: Brescia, Watt, Pagani, Cascino, Shore, Likosky, Aaronson, Thompson.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Brescia, Watt, Pagani, Zhang, Likosky.

    Critical revision of the manuscript for important intellectual content: Brescia, Watt, Pagani, Cascino, McCullough, Shore, Likosky, Aaronson, Thompson.

    Statistical analysis: Brescia, Watt, Zhang, Thompson.

    Obtained funding: Pagani.

    Administrative, technical, or material support: Brescia, Pagani, Likosky, Thompson.

    Supervision: Pagani, McCullough, Likosky, Thompson.

    Conflict of Interest Disclosures: Dr Pagani reported receiving personal fees from FineHeart outside the submitted work and serving as chair of The Society of Thoracic Surgeons Interagency Registry for Mechanically Assisted Circulatory Support National Database Task Force. Dr Likosky reported receiving a grant from the National Institutes of Health and personal fees from the American Society of Extracorporeal Technology outside the submitted work. Dr Aaronson reported receiving grants from Abbott and HeartWare and personal fees from Medtronic outside the submitted work. Dr Thompson reported receiving partial salary support from Blue Cross Blue Shield of Michigan outside the submitted work. No other disclosures were reported.

    Funding/Support: Dr Brescia was supported by the National Research Service Award postdoctoral fellowship No. 5T32HL076123. Drs Likosky, Pagani, Aaronson, and Zhang were supported by grant No. R01HS026003 AHRQ from the Agency for Healthcare Research and Quality. Drs Likosky, Zhang, and Pagani were supported by grant No. HL146619-01A1 from the National Heart, Lung, and Blood Institute.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Michigan Congestive Heart Failure Investigators: Donald S. Likosky, PhD and Francis D. Pagani, MD, PhD (Department of Cardiac Surgery, Michigan Medicine, University of Michigan, Ann Arbor; Center for Healthcare Outcomes and Policy, University of Michigan, Ann Arbor); Keith Aaronson, MD, MS (Division of Cardiovascular Medicine, Department of Internal Medicine, Michigan Medicine, University of Michigan, Ann Arbor, Michigan); Min Zhang, PhD (Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor); Jeffrey S. McCullough, PhD (Department of Health Management and Policy, School of Public Health, University of Michigan, Ann Arbor); Supriya Shore, MBBS (Division of Cardiovascular Medicine, Department of Internal Medicine, Michigan Medicine, University of Michigan, Ann Arbor); Alexander Brescia, MD, MSc; Michael P. Thompson, PhD; and Michelle Hou, MS (Department of Cardiac Surgery, Michigan Medicine, University of Michigan, Ann Arbor; Center for Healthcare Outcomes and Policy, University of Michigan, Ann Arbor); Thomas Cascino, MD, MSc (Center for Healthcare Outcomes and Policy, University of Michigan, Ann Arbor; Division of Cardiovascular Medicine, Department of Internal Medicine, Michigan Medicine, University of Michigan, Ann Arbor); Allison M. Janda, MD and Michael R. Mathis, MD (Department of Anesthesiology, Michigan Medicine, University of Michigan, Ann Arbor); Ashraf Shaaban Abdel Aziz Abou El Ela, MB, BCH and Paul C. Tang, MD, PhD (Department of Cardiac Surgery, Michigan Medicine, University of Michigan, Ann Arbor); Tessa M. F. Watt, MD, MSc and Michael J. Pienta, MD, MS (Department of Cardiac Surgery, Michigan Medicine, University of Michigan, Ann Arbor; Center for Healthcare Outcomes and Policy, University of Michigan, Ann Arbor).

    Disclaimer: Opinions expressed in this manuscript do not represent those of The Society of Thoracic Surgeons Interagency Registry for Mechanically Assisted Circulatory Support; National Heart, Lung, and Blood Institute; Centers for Medicare & Medicaid Services; Food and Drug Administration; Agency for Healthcare Research and Quality; or Department of Health and Human Services.

    References
    1.
    Teuteberg  JJ, Cleveland  JC  Jr, Cowger  J,  et al.  The Society of Thoracic Surgeons INTERMACS 2019 annual report: the changing landscape of devices and indications.   Ann Thorac Surg. 2020;109(3):649-660. doi:10.1016/j.athoracsur.2019.12.005 PubMedGoogle ScholarCrossref
    2.
    Goldstein  DJ, Meyns  B, Xie  R,  et al.  Third annual report from the ISHLT Mechanically Assisted Circulatory Support Registry: a comparison of centrifugal and axial continuous-flow left ventricular assist devices.   J Heart Lung Transplant. 2019;38(4):352-363. doi:10.1016/j.healun.2019.02.004 PubMedGoogle ScholarCrossref
    3.
    Slaughter  MS, Rogers  JG, Milano  CA,  et al; HeartMate II Investigators.  Advanced heart failure treated with continuous-flow left ventricular assist device.   N Engl J Med. 2009;361(23):2241-2251. doi:10.1056/NEJMoa0909938 PubMedGoogle ScholarCrossref
    4.
    Rogers  JG, Pagani  FD, Tatooles  AJ,  et al.  Intrapericardial left ventricular assist device for advanced heart failure.   N Engl J Med. 2017;376(5):451-460. doi:10.1056/NEJMoa1602954 PubMedGoogle ScholarCrossref
    5.
    Milano  CA, Rogers  JG, Tatooles  AJ,  et al; ENDURANCE Investigators.  HVAD: the ENDURANCE Supplemental Trial.   JACC Heart Fail. 2018;6(9):792-802. doi:10.1016/j.jchf.2018.05.012 PubMedGoogle ScholarCrossref
    6.
    Mehra  MR, Naka  Y, Uriel  N,  et al; MOMENTUM 3 Investigators.  A fully magnetically levitated circulatory pump for advanced heart failure.   N Engl J Med. 2017;376(5):440-450. doi:10.1056/NEJMoa1610426 PubMedGoogle ScholarCrossref
    7.
    Mehra  MR, Goldstein  DJ, Uriel  N,  et al; MOMENTUM 3 Investigators.  Two-year outcomes with a magnetically levitated cardiac pump in heart failure.   N Engl J Med. 2018;378(15):1386-1395. doi:10.1056/NEJMoa1800866 PubMedGoogle ScholarCrossref
    8.
    Mehra  MR, Uriel  N, Naka  Y,  et al; MOMENTUM 3 Investigators.  A fully magnetically levitated left ventricular assist device—final report.   N Engl J Med. 2019;380(17):1618-1627. doi:10.1056/NEJMoa1900486 PubMedGoogle ScholarCrossref
    9.
    Aaronson  KD, Slaughter  MS, Miller  LW,  et al; HeartWare Ventricular Assist Device (HVAD) Bridge to Transplant ADVANCE Trial Investigators.  Use of an intrapericardial, continuous-flow, centrifugal pump in patients awaiting heart transplantation.   Circulation. 2012;125(25):3191-3200. doi:10.1161/CIRCULATIONAHA.111.058412 PubMedGoogle ScholarCrossref
    10.
    Rose  EA, Gelijns  AC, Moskowitz  AJ,  et al; Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) Study Group.  Long-term use of a left ventricular assist device for end-stage heart failure.   N Engl J Med. 2001;345(20):1435-1443. doi:10.1056/NEJMoa012175 PubMedGoogle ScholarCrossref
    11.
    Stevenson  LW, Pagani  FD, Young  JB,  et al.  INTERMACS profiles of advanced heart failure: the current picture.   J Heart Lung Transplant. 2009;28(6):535-541. doi:10.1016/j.healun.2009.02.015 PubMedGoogle ScholarCrossref
    12.
    Birati  EY, Hanff  TC, Maldonado  D,  et al.  Predicting long term outcome in patients treated with continuous flow left ventricular assist device: the Penn-Columbia risk score.   J Am Heart Assoc. 2018;7(6):e006408. doi:10.1161/JAHA.117.006408 PubMedGoogle Scholar
    13.
    Cowger  J, Sundareswaran  K, Rogers  JG,  et al.  Predicting survival in patients receiving continuous flow left ventricular assist devices: the HeartMate II risk score.   J Am Coll Cardiol. 2013;61(3):313-321. doi:10.1016/j.jacc.2012.09.055 PubMedGoogle ScholarCrossref
    14.
    Boyle  AJ, Ascheim  DD, Russo  MJ,  et al.  Clinical outcomes for continuous-flow left ventricular assist device patients stratified by pre-operative INTERMACS classification.   J Heart Lung Transplant. 2011;30(4):402-407. doi:10.1016/j.healun.2010.10.016 PubMedGoogle ScholarCrossref
    15.
    U.S. Food and Drug Administration.  Real-world evidence. Accessed December 31, 2019. https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence
    16.
    21st Century Cures Act, Pub L No. 114-255 (2016). Accessed November 18, 2020. https://www.govinfo.gov/app/details/PLAW-114publ255
    17.
    Centers for Medicare and Medicaid Services.  Decision memo for ventricular assist devices for bridge-to-transplant and destination therapy (CAG-00432R). Medicare Coverage Database. Accessed December 31, 2019. https://www.cms.gov/medicare-coverage-database/details/nca-decision-memo.aspx?NCAId=268
    18.
    Miller  LW, Pagani  FD, Russell  SD,  et al; HeartMate II Clinical Investigators.  Use of a continuous-flow device in patients awaiting heart transplantation.   N Engl J Med. 2007;357(9):885-896. doi:10.1056/NEJMoa067758 PubMedGoogle ScholarCrossref
    ×