Survival on the Heart Transplant Waiting List

Syed Shahyan Bakhtiyar; Elizabeth L Godfrey; Shayan Ahmed; Harveen Lamba; Jeffrey Morgan; Gabriel Loor; Andrew Civitello; Faisal H Cheema; Whitson B Etheridge; John Goss; Abbas Rana

doi:10.1001/jamacardio.2020.2795

. 2020 Aug 12;5(11):1227–1235. doi: 10.1001/jamacardio.2020.2795

Survival on the Heart Transplant Waiting List

Syed Shahyan Bakhtiyar ¹, Elizabeth L Godfrey ^1,^✉, Shayan Ahmed ², Harveen Lamba ³, Jeffrey Morgan ³, Gabriel Loor ³, Andrew Civitello ³, Faisal H Cheema ⁴, Whitson B Etheridge ¹, John Goss ¹, Abbas Rana ¹

¹Division of Abdominal Transplantation, Baylor College of Medicine, Michael E. DeBakey Department of Surgery, Houston, Texas

²Aga Khan University, Karachi, Pakistan

³Division of Cardiothoracic Transplantation and Circulatory Support, Baylor College of Medicine, Houston, Texas

⁴University of Houston College of Medicine, HCA Research Institute, Houston, Texas

Accepted for Publication: May 26, 2020.

^✉

Corresponding Author: Elizabeth L. Godfrey, BS, Division of Abdominal Transplantation, Michael E. DeBakey Department of Surgery, Baylor College of Medicine, 6620 Main St, Ste 1450, Houston, TX 77030 (elizabeth.godfrey@bcm.edu).

Published Online: August 12, 2020. doi:10.1001/jamacardio.2020.2795

Author Contributions: Drs Bakhtiyar and Rana 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. Dr Bakhtiyar and Ms Godfrey are co–first authors.

Concept and design: Ahmed, Morgan, Loor, Civitello, Cheema, Etheridge, Goss, Rana.

Acquisition, analysis, or interpretation of data: Bakhtiyar, Godfrey, Ahmed, Lamba, Morgan, Cheema, Etheridge, Rana.

Drafting of the manuscript: Bakhtiyar, Godfrey, Ahmed, Morgan, Etheridge, Rana.

Critical revision of the manuscript for important intellectual content: Bakhtiyar, Godfrey, Ahmed, Lamba, Morgan, Loor, Civitello, Cheema, Goss, Rana.

Statistical analysis: Bakhtiyar, Godfrey, Ahmed, Rana.

Obtained funding: Rana.

Administrative, technical, or material support: Bakhtiyar, Lamba, Morgan, Rana.

Supervision: Bakhtiyar, Ahmed, Morgan, Loor, Civitello, Cheema, Etheridge, Goss, Rana.

Conflict of Interest Disclosures: Dr Loor reported receiving grants from TransMedics and JLH Foundation outside the submitted work and additionally discloses grant support from St Jude and Medtronic to his research institutions (University of Minnesota and Baylor College of Medicine) for participation in clinical trials related to left ventricular assist device therapies outside of the present study. He also discloses grant support to the same research institutions from Maquet and United Therapeutics for participation in clinical registries and translational studies pertaining to lung transplantation and lung failure support therapies, and travel and housing reimbursement from the Abiomed Breethe advisory board. No other disclosures were reported.

^✉

Corresponding author.

PMCID: PMC7675100 PMID: 32785619

Key Points

Question

What is the current status of survival of patients on the heart transplantation waiting list?

Findings

In a cross-sectional analysis of 95 323 candidates wait-listed for heart transplantation between 1987 and 2017, associations were consistent with improvements in survival outcomes.

Meaning

Although not a replacement for heart transplant, continued improvement in heart failure therapy outcomes were associated with a survival benefit while patients awaited heart transplantation, although listing practices remain highly variable among centers; this finding may support a future approach of incorporating this survival benefit into the indications for heart transplantation.

Abstract

Importance

With continuing improvements in medical devices and more than a decade since the 2006 United Network for Organ Sharing (UNOS) allocation policy, it is pertinent to assess survival among patients on the heart transplantation waiting list, especially given the recently approved 2018 UNOS allocation policy.

Objectives

To assess survival outcomes among patients on the heart transplant waiting list during the past 3 decades and to examine the association of ventricular assist devices (VADs) and the 2006 UNOS allocation policy with survival.

Design, Setting, and Participants

A retrospective cross-sectional used the UNOS database to perform an analysis of 95 323 candidates wait-listed for heart transplantation between January 1, 1987, and December 29, 2017. Candidates for all types of combined transplants were excluded (n = 2087). Patients were followed up from the time of listing to death, transplantation, or removal from the list due to clinical improvement. Competing-risk, Kaplan-Meier, and multivariable Cox proportional hazards regression analyses were used.

Main Outcomes and Measures

The analysis involved an unadjusted and adjusted survival analysis in which the primary outcome was death on the waiting list. Because of changing waiting list preferences and policies during the study period, the intrinsic risk of death for wait-listed candidates was assessed by individually analyzing, comparing, and adjusting for several candidate risk factors.

Results

In total, 95 323 candidates (72 915 men [76.5%]; mean [SD] age, 51.9 [12.0] years) were studied. In the setting of changes in listing preferences, 1-year survival on the waiting list increased from 34.1% in 1987-1990 to 67.8% in 2011-2017 (difference in proportions, 0.34%; 95% CI, 0.32%-0.36%; P < .001). The 1-year waiting list survival for candidates with VADs increased from 10.2% in 1996-2000 to 70.0% in 2011-2017 (difference in proportions, 0.60%; 95% CI, 0.58%-0.62%; P < .001). Similarly, in the setting of changing mechanical circulatory support indications, the 1-year waiting list survival for patients without VADs increased from 53.9% in 1996-2000 to 66.5% in 2011-2017 (difference in proportions, 0.13%; 95% CI, 0.12%-0.14%; P < .001). In the decade prior to the 2006 UNOS allocation policy, the 1-year waiting list survival was 51.1%, while in the decade after it was 63.9% (difference in proportions, 0.13%; 95% CI, 0.12%-0.14%; P < .001). In adjusted analysis, each time period after 1987-1990 had a marked decrease in waiting list mortality.

Conclusions and Relevance

This study found temporally associated increases in heart transplant waiting list survival for all patient groups (with or without VADs, UNOS status 1 and status 2 candidates, and candidates with poor functional status).

This cross-sectional study assesses survival outcomes among patients on the heart transplant waiting list during the past 3 decades and examines the association of ventricular assist devices and the 2006 United Network for Organ Sharing allocation policy with survival.

Introduction

The 50th anniversary of heart transplantation was recognized in 2018. A previous analysis of the United Network for Organ Sharing (UNOS) data on solid-organ transplants demonstrated that more than 270 000 life-years had been saved during 25 years of solid-organ transplants.¹ That is nearly 5 years of life for each heart transplant performed. It is widely held that heart transplantation offers a significant survival benefit.^1,2,3 Beyond survival, morbidity and quality of life are understood to be superior after transplantation.⁴ However, heart transplantation is in competition with heart failure therapy.

There is a growing body of evidence suggesting that patients on the waiting list for heart transplantation are surviving much longer with ventricular assist devices (VADs).^5,6,7,8 Furthermore, in 2006 the UNOS implemented a new allocation policy to further reduce the number of deaths among patients on the waiting list. It was our aim to reassess the outcomes among patients on the waiting list in the present day. We hypothesize that improvements in both the medical management of heart failure and the mechanical devices used were associated with improved waiting list survival during the last 3 decades. We sought to individually analyze and then concurrently compare the overall survival outcomes, during a 5-year period from the time of wait-listing, of candidates on the heart transplant waiting list and candidates who received a transplant. We also reviewed the association of VADs with waiting list mortality and analyzed waiting list mortality by UNOS status during the past decade (post-2006 UNOS allocation policy).

Methods

Study Population

In this cross-sectional study, we used the UNOS database to retrospectively identify all candidates listed for heart transplantation (≥18 years of age) between January 1, 1987, and December 29, 2017. Our analysis also used the heart registry, with data collected by the Organ Procurement and Transplantation Network. The characteristics of the recipients were reported at the time of transplant. Follow-up information was collected 6 months after transplantation and yearly thereafter. Candidates on the waiting list for all types of combined transplants were excluded (n = 2087). We analyzed a total of 95 323 candidates listed for heart transplant alone. Baylor College of Medicine determined that no separate institutional review board approval was necessary for this study because the data were already deidentified when obtained.

Eras

We arbitrarily stratified the data to create the following time period cohorts: the 1987-1990, 1991-1995, 1996-2000, 2001-2005, 2006-2010, and 2011-2017 eras.

Candidate Risk Factors and Intrinsic Risk of Death

The candidate risk factors and clinical characteristics used in this analysis, with their percentage of entry completion, are listed in Table 1. The covariates included in our analysis are based on the current literature and consistent data completion. Race/ethnicity was as reported in the UNOS database. To account for various listing criteria across different regions and to account for changing listing preferences and policies during the study period, we attempted to depict the intrinsic risk of death for candidates on the waiting list and study it across the different eras. Although it is difficult to assess all factors associated with a candidate’s intrinsic risk of death, we depicted this intrinsic risk of death by analyzing and adjusting for status of diabetes at listing, dialysis status at listing, need for extracorporeal membrane oxygenation (ECMO), presence of intra-aortic balloon pump (IABP), need for inotropes, presence of VADs, ventilator dependence, UNOS status at listing, and functional status at listing (as reported by UNOS).

Table 1. Demographic and Clinical Characteristics by Era.

Characteristic	Candidates still on waiting list, %						Candidates who received a transplant, %
Characteristic	1987-1990 (n = 3835)	1991-1995 (n = 7044)	1996-2000 (n = 6440)	2001-2005 (n = 4323)	2006-2010 (n = 4190)	2011-2017 (n = 9383)	1987-1990 (n = 6155)	1991-1995 (n = 10 573)	1996-2000 (n = 10 074)	2001-2005 (n = 8684)	2006-2010 (n = 9796)	2011-2017 (n = 14 826)
African American race/ethnicity	10.4	13.5^a	13.9^a	17.9^a	21.1^a	25.2^a	8.0	10.5^a	12.2^a	15.3^a	19.5^a	21.9^a
Age, mean (SD), y	47.9 (11.0)	49.4 (10.9)^a	50.8 (11.5)^a	50.4 (12.4)^a	51.7 (12.7)^a	52.7 (12.6)^a	48.7 (10.9)	50.5 (10.8)^a	51.7 (11.3)^a	50.8 (12.3)^a	51.6 (12.5)^a	53.0 (12.7)^a
Age, mean (SD), y	47.9 (11.0)	49.4 (10.9)^a	50.8 (11.5)^a	50.4 (12.4)^a	51.7 (12.7)^a	52.7 (12.6)^a	48.7 (10.9)	50.5 (10.8)^a	51.7 (11.3)^a	50.8 (12.3)^a	51.6 (12.5)^a	53.0 (12.7)^a	BMI, mean (SD)	26.1 (4.9)^b	26.7 (5.0)^a	27.3 (5.2)^a	27.7 (5.4)^a	28.4 (5.5)^a	28.4 (5.1)^a	25.1 (4.7)	25.5 (4.7)^a	26.2 (4.7)^a	26.6 (4.8)^a	27.1 (4.9)^a	27.2 (4.9)^a
Male sex	80.9	80.6	77.3^a	75.1^a	73.3^a	75.1^a	80.7	78.6^a	76.3^a	75.5^a	75.7^a	73.3^a	BMI, mean (SD)	26.1 (4.9)^b	26.7 (5.0)^a	27.3 (5.2)^a	27.7 (5.4)^a	28.4 (5.5)^a	28.4 (5.1)^a	25.1 (4.7)	25.5 (4.7)^a	26.2 (4.7)^a	26.6 (4.8)^a	27.1 (4.9)^a	27.2 (4.9)^a
Diagnosis
Idiopathic dilated myopathy	2.0^c	21.9^a^,^b	32.2^a	30.7^a	30.0^a	34.8^a	30.5	37.0^a	34.0^a	33.0^a	33.7^a	35.6^a
Idiopathic dilated myopathy	2.0^c	21.9^a^,^b	32.2^a	30.7^a	30.0^a	34.8^a	30.5	37.0^a	34.0^a	33.0^a	33.7^a	35.6^a	Ischemic dilated myopathy	1.9^c	25.9^a^,^b	47.4^a	44.3^a	40.6^a	34.0^a	52.0	45.8^a	49.5^a	44.9^a	38.6^a	33.5^a
Diabetes	0^c	16.1^c	21.0	25.9^a^,^d	30.7^a^,^d	31.6^a^,^d	17.9^c	13.8^c	18.8	21.4^a^,^d	26.7^a^,^d	28.9^a^,^d	Ischemic dilated myopathy	1.9^c	25.9^a^,^b	47.4^a	44.3^a	40.6^a	34.0^a	52.0	45.8^a	49.5^a	44.9^a	38.6^a	33.5^a
Dialysis	0^c	0^c	2.3	3.5^a^,^d	4.0^a^,^d	4.4^a^,^d	0^c	0.3^c	1.2	1.9^a^,^d	2.1^a^,^d	2.4^a^,^d
ECMO	NA	NA	0.8	1.2	1.7^a	2.8^a	NA	NA	0.2	0.3	0.5^a	1.0^a
IABP	0	2.8^a	6.8^a	7.9^a	7.4^a	4.6^a	0.1	2.0^a	4.7^a	5.2^a	5.2^a	4.8^a
Inotropes	0	9.3^a	29.5^a	31.3^a	27.7^a	27.4^a	0.2	13.1^a	39.8^a	39.9^a	35.5^a	33.3^a
VAD	NA	NA	6.4^b	13.1^a^,^d^,^e	14.9^a^,^d	29.3^a^,^d	NA	NA	6.4^b	12.6^a^,^d^,^e	16.9^a^,^d	30.3^a^,^d
Ventilator	0	2.7^a	8.3^a	8.6^a	6.5^a	3.3^a	0	1.4^a	4.1^a	4.3^a	2.1^a	1.3^a
UNOS Status
1A	NA	NA	7.1	19.9^a^,^d	17.3^a^,^d	18.1^a^,^d	NA	NA	7.2	20.2^a^,^d	20.1^a^,^d	27.7^a^,^d
1B	NA	NA	5.4	19.3^a^,^d	28.5^a^,^d	41.3^a^,^d	NA	NA	9.4	29.9^a^,^d	39.1^a^,^d	46.9^a^,^d
2	30.2^b	70.1^a	65.4^a	54.7^a	47.5^a	36.7^a	34.7^b	67.5^a	56.8^a	47.3^a	37.2^a	23.4^a
Old status 1	14.8^b	25.1	18.2	NA	NA	NA	15.5^b	29.5^a	24.8^a	NA	NA	NA
Time on waiting list, mean (SD), d	349 (704)	643 (1056)^a	779 (1084)^a	643 (862)^a	580 (730)^a	438 (455)^a	164 (242)	224 (339)^a	254 (402)^a	193 (332)^a	200 (348)^a	196 (267)^a
Functional status
10%	0	0	0	0.8^a	4.2^a	2.6^a	0	0	0	0.5^a	1.9^a	1.4^a
20%	0	0	0	3.2^a	20.7^a	18.4^a	0	0	0	3.9^a	20.3^a	20.3^a
30%	0	0	0	0.8^a	7.0^a	6.7^a	0	0	0	1.6^a	8.9^a	8.9^a
40%	0	0	0	0.7^a	6.1^a	12.7^a	0	0	0	1.1^a	8.3^a	14.0^a
50%	0	0	0	1.1^a	10.4^a	11.6^a	0	0	0	2.0^a	11.9^a	11.4^a
60%	0	0	0	4.3^a	33.4^a	30.3^a	0	0	0.2	7.5^a	32.0^a	28.4^a
>60% and ≤100%	0	0	0	2.5^a	16.2^a	14.4^a	0	0	0.1	3.4^a	15.1^a	12.7^a
1	1.2^b	11.7^a	30.2^a	15.9^a	0	0	1.3	9.4^a	22.8^a	10.3^a	0	0
2	3.5^b	16.7^a	28.4^a	30.0^a	0	0	5.2	16.1^a	26.3^a	26.9^a	0	0
3	0^b	0.9^a	1.1^a	1.8^a	0	0	0	0.8^a	1.3^a	2.1^a	0	0

Open in a new tab

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pump; NA, not applicable; VAD, ventricular assist device.

^{^a}

Statistically significant compared with 1987-1990 era.

^{^b}

Data entry completion of less than 80%.

^{^c}

Data entry completion of less than 50%.

^{^d}

Statistically significant compared with 1996-2000 era.

^{^e}

Data entry completion of less than 90%.

In the UNOS database, for the 1987-1990, 1991-1995, 1996-2000, and 2001-2005 eras, functional status 1 was reported for candidates who performed activities of daily living with no assistance, functional status 2 for candidates requiring some assistance, and functional status 3 for candidates requiring total assistance. After 2001, UNOS reported the functional status as a percentage on a scale as follows: moribund candidates (10% functional status), very sick candidates for whom hospitalization and active treatment are necessary (20%), severely disabled candidates for whom hospitalization is indicated (30%), disabled candidates requiring special care and assistance (40%), candidates requiring considerable assistance and frequent medical care (50%), candidates requiring occasional assistance (60%), candidates unable to perform normal activity or active work (70%), candidates with some symptoms of disease (80%), candidates who can perform normal activity with minor symptoms of disease (90%), and candidates with no evidence of disease (100%).

Although Table 1 lists a breakdown of the functional status for each era, in our analyses, we grouped candidates with functional status of 50% or less as candidates with poor functional status and grouped those with a functional status of more than 50% as candidates with good functional status. We also grouped status 1 candidates as defined by UNOS before 2000 and status 1A and status 1B candidates as defined currently by UNOS as status 1 for the multivariable regression analysis.

Despite variabilities in data entry from patient registries, most of the risk factors in our analysis had a very high percentage of entry completion. All variables listed in Table 1 that were either associated with the candidates’ intrinsic risk of death or associated with the outcome variable (death on the waiting list) and that were statistically significant in a univariable analysis (P < .05) were included in a multivariable Cox proportional hazards regression model. In our adjusted model analyzing the different eras, the 1987-1990 era was the reference era. We also conducted a second adjusted analysis in which we did not include the 1991-1995 era, owing to a lack of data about use of VADs and ECMO prior to 1993.

Outcome Variables

The primary outcome variable was death on the waiting list. In our waiting list analysis, this outcome was established by the death date in the UNOS database (accounting for publicly reported external sources). For candidates who were delisted for transplantation and lost to the database, we established the date of death by the Social Security Death Master File. However, Social Security Death Master File data were not available after February 2014; to account for candidates who were missing a UNOS database death date and were delisted after this date, we performed a parallel analysis in which waiting list removal for deteriorating clinical condition was counted as death on the waiting list.

Competing Risk Analysis

As the rate of transplantation was significantly higher than the rate of dying while on the waiting list, we performed 2 competing risk regression analyses, one based on the method of Fine and Gray⁹ and the other using the Cox cause-specific proportional hazards regression analysis.¹⁰ The primary outcome was death on the waiting list, while transplantation and removal from the waiting list due to clinical improvement were the competing outcomes (Figure 1A). Competing risk refers to the probability that 1 event—a competing event—can prevent the event of interest from occurring if it occurs first; in our case, removal from the waiting list by either receipt of a transplant or recovery will mean that individual must be mathematically considered no longer at risk of death, either by censoring (Cox proportional hazards regression) or by assigning a zero risk of death (Fine and Gray analysis). Given that the change in the intrinsic risk of death of wait-listed candidates could have been the primary factor associated with the improving outcomes, we stratified our study population into groups based on the risk factors previously identified and performed a competing risk regression analysis for each group. We sought to individually analyze and then concurrently compare the subhazard ratio (SHR) and cause-specific hazard (CSH) for the era of listing among each stratum; all SHRs and CSHs for each subgroup of patients are presented in eTable 4 in the Supplement.

Figure 1. — A, Fine and Gray analysis. Competing outcomes were removed from the list for transplantation or clinical improvement. B, Cox proportional hazards regression analysis. Competing outcomes were removed from the list for transplantation or clinical improvement. C, Waiting list survival by eras for candidates still on the waiting list. D, Waiting list survival by eras for candidates who received a transplant. P < .001 for each group compared with the 1987-1990 era by log-rank test.

Kaplan-Meier Survival Analysis

In our waiting list survival analysis, the primary outcome measure was death on the waiting list. Patients who underwent transplantation or were removed from the waiting list owing to clinical improvement were censored at the time of removal from the waiting list. All patients were followed up from the time of listing to death on the waiting list, transplantation, or removal from the list due to clinical improvement. We further performed Kaplan-Meier survival analyses for wait-listed candidates based on their VAD, UNOS, and functional status at listing.

In our posttransplant survival analysis, candidates were followed up from the time of listing to death after transplantation or to the last known follow-up. Patients who were lost to follow-up or alive on December 29, 2017, were censored at the date of last known follow-up, and only patients who received a transplant were included in the posttransplant survival analysis. Because 5-year outcomes for the 2011-2017 era were truncated, possibly underestimating the 5-year mortality for this era, we also reported 1-year and 3-year survival outcomes to better allow for a comparison between eras.

Statistical Analysis

Data were analyzed using the Stata, version 15 statistical software package (Stata Corp). Continuous variables were summarized using mean (SD) values and compared using the t test. The contingency table and the χ² test were used to compare categorical variables. Percentages between various eras were compared using a 2-sample test for proportions, and the differences in proportions between eras have been reported. Results were considered significant at a 2-tailed P < .05.

Results

Study Population

The study population consisted of 95 323 patients (72 915 men and 22 408 women; mean [SD] age, 51.9 [12.0] years). The study population for the posttransplant survival analysis had 60 108 patients. Demographic and clinical characteristics are summarized in Table 1. The UNOS reports sex as male or female; we arbitrarily reported the percentage of men by era in Table 1 for brevity.

Candidate Risk Factors and Intrinsic Risk of Death

Candidate age and body mass index increased across all eras compared with the 1987-1990 era. Similarly, the percentages of patients with diabetes, undergoing dialysis, receiving inotropes, with an IABP, and needing ventilators all increased over time compared with the reference era (Table 1). Candidates on the waiting list who were receiving ECMO significantly increased from 1.2% in the 2001-2005 era to 2.8% in the 2011-2017 era (difference in proportions, 0.02%; 95% CI, 0.01%-0.02%; P < .001). Similarly, candidates on the waiting list with VADs significantly increased from 13.1% in the 2001-2005 era to 29.3% in the 2011-2017 era (difference in proportions, 0.16%; 95% CI, 0.15%-0.18%; P < .001).

There was an increase in the percentage of patients in each functional status group across the study period. Patients requiring no assistance (functional status 1) increased from 1.2% in the 1987-1990 era to 15.9% in the 2001-2005 era (difference in proportions, 0.15%; 95% CI, 0.14%-0.16%; P < .001), and from 2001 onward, the percentage of patients with a good functional status (>50%) also increased in the 2011-2017 era compared with the 2001-2005 era (Table 1). Similarly, patients requiring total assistance (functional status 3) increased from 0% in the 1987-1990 era to 1.8% in the 2001-2005 era (difference in proportions, 0.02%; 95% CI, 0.01%-0.02%; P < .001), and from 2001 onward, the percentage of patients with a poor functional status (≤50%) also increased in the 2011-2017 era compared with the 2001-2005 era (Table 1).

Competing Risk Analysis

The 1987-1990 era was the reference era. eTable 4 in the Supplement reports the unadjusted SHRs and CSHs by era of wait-listing both overall and after stratifying the study population by risk factors. A higher era of wait-listing is associated with a lower incidence of death on the waiting list. This is seen in the study population overall (SHR, 0.85; 95% CI, 0.84-0.85; P < .001; CSH, 0.86; 95% CI, 0.85-0.86; P < .001) and in each risk factor subgroup. Although SHRs and CSHs can often differ, as statistically speaking SHRs can suggest that a factor is associated with a reduced risk of death on the waiting list when in reality it is associated with an increased likelihood of transplantation and vice versa, there were only small differences between the 2 methods, which cross-corroborates the findings from each.

As seen in Figure 1A, each era after the 1987-1990 era was associated with a lower incidence of death on the waiting list. For comparison, the comparable figure for CSHs, which represents the data as cumulative incidence functions year by year for the eras analyzed, is in Figure 1B. Similar graphs for the estimated cumulative incidence functions of subdistribution hazards for each subgroup of patients by characteristic (eg, need for IABP and geographic region) are shown in eFigures 1, 2, 3, 4, 5, 6, 7, 8, and 9 in the Supplement.

Waiting List Outcomes and Survival by Era

The overall waiting list outcomes are reported in eTable 1 in the Supplement. The Kaplan-Meier curves for waiting list survival among candidates still on the waiting list are depicted in Figure 1C and demonstrate an increase in lifespan over time (P < .001; log-rank test for equality of survivor function). The 1-year survival for candidates still on the waiting list has consistently increased from 34.1% for the 1987-1990 era to 67.8% for the 2011-2017 era (difference in proportions, 0.34%; 95% CI, 0.32%-0.36%; P < .001), 3-year survival has consistently increased from 13.9% for the 1987-1990 era to 48.8% for the 2011-2017 era (difference in proportions, 0.35%; 95% CI, 0.33%-0.36%; P < .001), and 5-year survival has consistently increased from 8.9% for the 1987-1990 era to 35.1% for the 2011-2017 era (difference in proportions, 0.26%; 95% CI, 0.25%-0.28%; P < .001).

To analyze the outcome of the 2006 UNOS allocation policy, we grouped candidates listed in the decade before 2006 (1995-2005) as the pre-2006 UNOS group and candidates listed in the decade after 2006 (2007-2017) as the post-2006 UNOS group. In the pre-2006 UNOS group, 1-year survival was 51.1%, 3-year survival was 31.2%, and 5-year survival was 21.9%; in the post-2006 UNOS group, 1-year survival was 63.9% (difference in proportions, 0.13%; 95% CI, 0.12%-0.14%; P < .001), 3-year survival was 43.8% (difference in proportions, 0.13%; 95% CI, 0.11%-0.14%; P < .001), and 5-year survival was 30.8% (difference in proportions, 0.09%; 95% CI, 0.08%-0.10%; P < .001). Figure 1D demonstrates the survival curves, over a 5-year period from the time of listing, for candidates who received a transplant.

Waiting List Outcomes and Survival by VAD Status

The waiting list outcomes for candidates with or without VADs by era are reported in eTable 2 in the Supplement. For candidates with VADs in the 1996-2000 era, 1-year survival was 10.2%, 3-year survival was 4.4%, and 5-year survival was 3.9%. In comparison, candidates with VADs listed in the 2011-2017 era had 1-year survival of 70.0% (difference in proportions, 0.60%; 95% CI, 0.58%-0.62%; P < .001), 3-year survival of 47.6% (difference in proportions, 0.43%; 95% CI, 0.42%-0.45%; P < .001), and 5-year survival of 33.3% (difference in proportions, 0.29%; 95% CI, 0.28%-0.31%; P < .001).

For candidates without VADs listed in the 1996-2000 era, 1-year survival was 53.9%, 3-year survival was 32.5%, and 5-year survival was 23.0%. In comparison, candidates without VADs listed in the 2011-2017 era had 1-year survival of 66.5% (difference in proportions, 0.13%; 95% CI, 0.12%-0.14%; P < .001), 3-year patient survival of 48.9% (difference in proportions, 0.16%; 95% CI, 0.15%-0.18%; P < .001), and 5-year patient survival of 35.5% (difference in proportions, 0.13%; 95% CI, 0.12%-0.14%; P < .001).

The Kaplan-Meier survival curves for candidates on the waiting list, with or without VADs, by era are depicted in Figure 2A and B. Owing to a lack of UNOS data regarding VAD status prior to 1993, the survival curves for the 1987-1990 era and the 1991-1995 era are not shown.

Waiting List Outcomes and Survival by UNOS Status

The waiting list outcomes for UNOS status 1 and status 2 candidates by era are reported in eTable 3 in the Supplement. Among UNOS status 1 candidates in the 2001-2005 era, 1-year survival was 26.2%, 3-year survival was 14.0%, and 5-year survival was 9.9%. After the 2006 UNOS allocation policy, in the 2011-2017 era, 1-year survival improved to 60.8% (difference in proportions, 0.35%; 95% CI, 0.34%-0.36%; P < .001), 3-year survival improved to 41.2% (difference in proportions, 0.27%; 95% CI, 0.26%-0.28%; P < .001), and 5-year survival improved to 27.2% (difference in proportions, 0.17%; 95% CI, 0.16%-0.18%; P < .001).

For UNOS status 2 candidates listed prior to the 2006 UNOS allocation policy in the 2001-2005 era, 1-year survival was 67.3%, 3-year survival was 43.9%, and 5-year survival was 30.2%. In comparison, those listed in the 2011-2017 era had 1-year survival of 77.8% (difference in proportions, 0.11%; 95% CI, 0.09%-0.12%; P < .001), 3-year survival of 59.3% (difference in proportions, 0.15%; 95% CI, 0.14%-0.17%; P < .001), and 5-year survival of 44.9% (difference in proportions, 0.15%; 95% CI, 0.13%-0.16%; P < .001). The Kaplan-Meier survival curves by UNOS status are shown in Figure 2C and D and demonstrate significant change over time (P < .001; log-rank test for equality of survivor function).

Waiting List Survival for Candidates With Poor Functional Status

In the 2001-2005 era, 1-year survival for candidates with poor functional status (≤50%) was 29.8%, 3-year survival was 10.6%, and 5-year survival was 6.9%. In the 2006-2010 era, 1-year survival was 37.3% (difference in proportions, 0.08%; 95% CI, 0.02%-0.13%; P = .01), 3-year survival was 20.2% (difference in proportions, 0.10%; 95% CI, 0.06%-0.14%; P < .001), and 5-year survival was 12.6% (difference in proportions, 0.06%, 95% CI, 0.02%-0.09%; P = .01). Compared with the 2001-2005 era, in the 2011-2017 era, waiting list survival further increased to 1-year survival of 59.3% (difference in proportions, 0.29%; 95% CI, 0.24%-0.35%; P < .001), 3-year survival of 40.3% (difference in proportions, 0.30%; 95% CI, 0.26%-0.34%; P < .001), and 5-year survival of 26.7% (difference in proportions, 0.20%, 95% CI, 0.17%-0.23%; P < .001). The Kaplan-Meier survival curves for candidates with a poor functional status are shown in eFigure 10 in the Supplement.

Multivariable Analysis

Table 2 reports the multivariable analysis for waiting list survival. Each time period after the 1987-1990 era had a marked decrease in waiting list mortality. After adjusting for candidate risk factors, including factors associated with the candidates’ intrinsic risk of death, the CSH ratio for candidates was 0.84 (95% CI, 0.80-0.89; P < .001) in the 1996-2000 era, 0.80 (95% CI, 0.75-0.85; P < .001) in the 2001-2005 era, 0.64 (95% CI, 0.57-0.71; P < .001) in the 2006-2010 era, and 0.35 (95% CI, 0.32-0.39; P < .001) in the 2011-2017 era.

Table 2. Multivariable Analysis: Waiting List Mortality.

Risk factor	Cause-specific HR (95% CI)	P value
Listed between
1996-2000	0.84 (0.80-0.89)	<.001
2001-2005	0.80 (0.75-0.85)	<.001
2006-2010	0.64 (0.57-0.71)	<.001
2011-2017	0.35 (0.32-0.39)	<.001
Age at listing	1.02 (1.01-1.02)	<.001
BMI at listing	0.98 (0.98-0.99)	<.001
Male sex	1.14 (1.09-1.18)	<.001
Diagnosis
Idiopathic dilated cardiomyopathy	0.77 (0.74-0.81)	<.001
Ischemic dilated cardiomyopathy	0.86 (0.83-0.90)	<.001
Diabetes	1.12 (1.08-1.17)	<.001
Dialysis	1.76 (1.63-1.91)	<.001
ECMO at listing	2.47 (2.19-2.79)	<.001
IABP at listing	1.35 (1.26-1.44)	<.001
Inotropes at listing	1.40 (1.33-1.46)	<.001
Ventilator at listing	1.45 (1.35-1.56)	<.001
VAD at listing	1.13 (1.07-1.19)	<.001
Poor functional status at listing^a	1.23 (1.12-1.35)	<.001
Good functional status at listing^b	0.84 (0.76-0.92)	<.002
UNOS status 1 (1A, 1B, or old status 1)	1.81 (1.72-1.90)	<.001

Open in a new tab

Abbreviations: BMI, body mass index; ECMO, extracorporeal membrane oxygenation; HR, hazard ratio; IABP, intra-aortic balloon pump; UNOS, United Network for Organ Sharing; VAD, ventricular assist device.

^{^a}

Functional status of 50% or less as reported by UNOS.

^{^b}

Functional status of more than 50% as reported by UNOS.

Discussion

Survival on the heart transplant waiting list has increased during the past 3 decades. Our analysis of candidate risk factors, especially factors associated with the intrinsic risk of death (status of diabetes at listing, dialysis status at listing, need for ECMO, presence of IABP, need for inotropes, presence of VADs, ventilator dependence, UNOS status at listing, and functional status at listing), shows a significant increase in all of these covariate risks across the study period, indicating sicker listed individuals. Nevertheless, our overall unadjusted and adjusted analyses, as well as our competing risk analysis for each individual risk factor, were associated with increased waiting list survival time. We cannot fully account for selective listing practices among sites, but we note the associations with improved survival time despite increases in known risk factors associated with a higher intrinsic of death.

Our adjusted analysis showed a 36% reduction in the risk of death on the waiting list in the 2006-2010 era and a 65% reduction in the 2011-2017 era (compared with the reference period). We recognize the limitations of our analysis given the multiple decades, centers, and variations in practice; thus, we cannot definitively assert that improvements in heart failure therapy and use of device-based therapies are solely responsible for the observed improvements in waiting list survival. We suggest that the survival benefit of heart transplantation continue to be reassessed with forthcoming data. Outcomes in transplantation are relative to the success of organ replacement therapy.

Overall waiting list survival time has seen considerable increases during the 2006-2010 and 2011-2017 eras, probably owing at least in part to continued advancements in VADs during those periods, as well as in association with the 2006 UNOS allocation policy changes. The survival curve of waiting list candidates has never been closer to that of candidates who received transplants, and this greater waiting list survival time is seen in patients with VADs, patients without VADs, UNOS status 1 candidates, UNOS status 2 candidates, and candidates with poor functional status.

This analysis demonstrates that survival time on the heart transplant waiting list is increasing with time. There is a significant and growing body of corroborating evidence in the literature, particularly with respect to documented improvements in survival outcomes after VADs.^5,6,7,8 Other studies have demonstrated improvements in survival for all wait-listed candidates.¹¹ Our analysis reaffirms these studies with a robust adjusted analysis, and contemporaneous data demonstrate a continued trend. The reasons behind these observed improvements are likely multifactorial, including variations in listing preferences, overall better medical management, improved mechanical devices, and the greater use of mechanical devices and implantable cardioverter-defibrillators prior to transplantation.¹¹ Although there are numerous reasons for survival on the waiting list to be genuinely improving, some component of changed and optimized listing behaviors is likely also a factor.

Our data suggest that survival on the waiting list is increasing faster than survival after transplantation. Continued improvement in outcomes for heart failure therapy among patients on the waiting list may support continued attention to balancing the survival benefit of existing supportive therapies against the indication for heart transplantation.

Limitations

Our study has some limitations, including that the database used has changing definitions and data completion rates over time, which makes certain characteristics and outcomes difficult, if not impossible, to assess. This study is also not able to assess the center-level behaviors underlying changing listing and allocation decisions, which may be critical factors leading to longer survival times on the waiting list.

Conclusions

This study found increased heart transplant waiting list survival during the past 30 years, despite listing sicker and older candidates. Waiting list survival time is increasing more quickly than posttransplant survival.

Supplement.

eTable 1. Overall Outcomes of Candidates Wait-listed for Heart Transplantation by Era

eTable 2. Outcomes of Candidates Wait-listed for Heart Transplantation by VAD Status at the Time of Listing

eTable 3. Outcomes of Candidates Wait-listed for Heart Transplantation by UNOS Status at the Time of Listing

eTable 4. Sub-distribution and Cause-Specific Hazards of Wait-list Mortality vs Era per Patient Group

eFigure 1. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 2. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 3. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 4. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 5. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 6. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 7. Estimated cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 8. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 9. Estimated Cumulative Incidence Functions from Competing Risk Analysis Depicting Probability of Death on the Waiting List

eFigure 10. Waiting List Survival by Era for Candidates With a Poor Functional Status

Click here for additional data file.^{(1.4MB, pdf)}

References

1.Rana A, Gruessner A, Agopian VG, et al. Survival benefit of solid-organ transplant in the United States. JAMA Surg. 2015;150(3):252-259. doi: 10.1001/jamasurg.2014.2038 [DOI] [PubMed] [Google Scholar]
2.Hunt SA, Haddad F. The changing face of heart transplantation. J Am Coll Cardiol. 2008;52(8):587-598. doi: 10.1016/j.jacc.2008.05.020 [DOI] [PubMed] [Google Scholar]
3.Van Trigt P, Davis RD, Shaeffer GS, et al. Survival benefits of heart and lung transplantation. Ann Surg. 1996;223(5):576-584. doi: 10.1097/00000658-199605000-00013 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Czyżewski Ł, Torba K, Jasińska M, Religa G. Comparative analysis of the quality of life for patients prior to and after heart transplantation. Ann Transplant. 2014;19:288-294. doi: 10.12659/AOT.890190 [DOI] [PubMed] [Google Scholar]
5.Trivedi JR, Cheng A, Singh R, Williams ML, Slaughter MS. Survival on the heart transplant waiting list: impact of continuous flow left ventricular assist device as bridge to transplant. Ann Thorac Surg. 2014;98(3):830-834. doi: 10.1016/j.athoracsur.2014.05.019 [DOI] [PubMed] [Google Scholar]
6.Park SJ, Tector A, Piccioni W, et al. Left ventricular assist devices as destination therapy: a new look at survival. J Thorac Cardiovasc Surg. 2005;129(1):9-17. doi: 10.1016/j.jtcvs.2004.04.044 [DOI] [PubMed] [Google Scholar]
7.Long JW, Healy AH, Rasmusson BY, et al. Improving outcomes with long-term “destination” therapy using left ventricular assist devices. J Thorac Cardiovasc Surg. 2008;135(6):1353-1360. doi: 10.1016/j.jtcvs.2006.09.124 [DOI] [PubMed] [Google Scholar]
8.Park SJ, Milano CA, Tatooles AJ, et al. ; HeartMate II Clinical Investigators . Outcomes in advanced heart failure patients with left ventricular assist devices for destination therapy. Circ Heart Fail. 2012;5(2):241-248. doi: 10.1161/CIRCHEARTFAILURE.111.963991 [DOI] [PubMed] [Google Scholar]
9.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94(446):496-509. doi: 10.1080/01621459.1999.10474144 [DOI] [Google Scholar]
10.Latouche A, Beyersmann J, Fine JP. Comments on ‘Analysing and interpreting competing risk data’. Stat Med. 2007;26(19):3676-3679. doi: 10.1002/sim.2823 [DOI] [PubMed] [Google Scholar]
11.Lietz K, Miller LW. Improved survival of patients with end-stage heart failure listed for heart transplantation: analysis of organ procurement and transplantation network/U.S. United Network of Organ Sharing data, 1990 to 2005. J Am Coll Cardiol. 2007;50(13):1282-1290. doi: 10.1016/j.jacc.2007.04.099 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials