Abstract
BACKGROUND:
In pediatric heart transplantation, surgeons historically avoided donors requiring cardiopulmonary resuscitation (CPR), despite evidence that donor CPR does not change post-transplant survival (PTS). The authors sought to determine whether CPR duration impacts PTS.
METHODS:
All potential brain dead donors age <40 years from 2001–2021 consented for heart procurement were identified in the United Network for Organ Sharing database (n=54,671). Organ acceptance was compared by CPR administration and duration. All recipients age <18 years with donor CPR data were then identified (n=5,680). Survival analyses were conducted using increasing CPR duration as a cutpoint to identify the shortest duration beyond which PTS worsened. Additional analyses were performed with multivariable and cubic spline regression.
RESULTS:
Fifty-one percent of donors (28,012/54,671) received CPR. Donor acceptance was lower after CPR (54 vs. 66%; p<0.001), and across successive quartiles of CPR duration (p<0.001). Of the recipients, 48% (2,753/5,680) belonged to the no-CPR group and 52% (2,927/5,680) belonged to the CPR group. Kaplan-Meier analyses of CPR duration attained significance at 55 minutes, after which PTS worsened (11.1 vs. 9.2 years, p=0.025). There was no survival difference between the CPR≤55 minutes and no-CPR groups (11.1 vs 11.2 years, p=0.571). A cubic spline regression model confirmed PTS increased above 55 minutes of CPR. A Cox regression demonstrated that CPR>55 minutes predicted worsened PTS relative to no CPR (HR=1.51, p=0.007) but CPR≤55 minutes did not (HR=1.01, p=0.864).
CONCLUSIONS:
Donor CPR decreases organ acceptance for transplantation; however, shorter durations (≤55 minutes) had equivalent PTS when controlling for other risk factors.
Graphical Abstract
Pediatric heart transplantation is the standard of care for children with heart failure. Advances in medical technology, surgical technique, and clinical medicine have significantly improved post-heart transplant outcomes;[1] however, waitlist mortality remains high and exceeds all other solid organs regardless of age group.[2, 3] One major factor contributing to waitlist mortality is the scarcity of viable donor organs.[2] Although transplant centers have widened criteria for organ acceptance, the ratio of hearts accepted for transplant to available donors has decreased over time.[4] This gap suggests a potential to identify viable organs already in the donor pool that were rejected for inappropriate reasons and mitigate donor organ scarcity. One subset of donors that have historically been rejected by many centers are brain dead donors who required cardiopulmonary resuscitation (CPR).[5] This practice continues despite evidence from the adult literature indicating no difference in post-transplant outcomes for organs requiring donor CPR.[6, 7] A study of 66 transplant surgeons representing 47 centers at an American Society of Transplantation conference found wide variability in cutoffs used for donor CPR duration—20% cited >20 minutes as an unacceptable duration, 38% said >30 minutes, 23% said >40 minutes, and 20% said >60 minutes.[8] Nevertheless, it seems likely that there exists a duration of CPR beyond which irreversible ischemic injury would negatively affect graft viability and recipient outcomes. The present study sought to identify whether there existed an inflection point in post-transplant survival (PTS) relative to donor CPR duration. The authors hypothesized that while longer CPR duration would portend worse PTS, what constitutes a long CPR duration is likely higher than cutoffs used by contemporary transplant centers.
PATIENTS AND METHODS
The aims of this retrospective, multi-institutional study were to determine 1) whether hearts from brain dead donors requiring CPR were less likely to be accepted than hearts not requiring CPR and 2) whether increasing donor CPR duration had a differential impact on post-transplant outcomes. The primary outcome of interest was long-term PTS. The United Network for Organ Sharing (UNOS) database was utilized, which collates donor and recipient data on all transplants in the United States. The study received institutional review board approval with an informed consent waiver given that all data in the database is publicly available and de-identified (CCHMC IRB# 2018–6837; date of approval: 10/29/2018).
All potential brain dead donors age <40 years old from January 2000 to December 2021 consented for heart procurement with CPR data recorded were first identified and stratified by CPR administration (Y/N) and CPR duration in minutes. Donors after circulatory death were excluded. These cohorts were compared across a number of demographic and clinical variables. Organ acceptance was then compared across groups. Subsequently, all pediatric heart transplant recipients age <18 years old with complete donor CPR data were identified. A series of survival analyses were performed using iteratively increasing CPR duration as a cutpoint to identify the shortest duration above which PTS worsened (i.e, ≤x minutes vs. >x minutes). To confirm the inflection point in PTS as a function of donor CPR duration, we also used Cox proportional hazard with restricted cubic spline analysis. The three groups (no CPR, short CPR duration, and long CPR duration) were then compared. Recipients requiring ventricular assist devices (VAD) included all forms, including right, left, and biventricular VAD. Hepatic dysfunction was defined as total serum bilirubin >1.2 mg/dL. Renal dysfunction was defined as estimated glomerular filtration rate <60 mL/min/1.73 m2 as calculated using the bedside Schwartz equation. Inotrope use was defined as a binary variable (Y/N); data on the specific type, number, and dosage of inotropes is unavailable in UNOS. Ventilator support was similarly a binary variable (Y/N) for both donors and recipients.
Statistical analyses were performed using IBM SPSS Statistics for Windows, version 27 (IBM Corp., Armonk, NY, USA) and R (R Foundation for Statistical Computing, Vienna, Austria). Donor and recipient characteristics were compared using Pearson’s chi-square test with Yates’ continuity correction or Fisher’s exact test (for cross-tabulation cell counts <5) for discrete variables and Mood’s median test for continuous variables. For the recipient groups, post-hoc pairwise comparisons were performed for significant variables (p<0.05). Multivariable logistic regression was used to model donor organ acceptance. The iterative analyses of PTS utilized the Kaplan-Meier method with log-rank (Mantel-Cox) test for significance. Cox proportional hazards regression with backwards stepwise elimination was performed to determine predictors of overall survival.
RESULTS
During the study period, 51% (28,012/54,671) of donors received CPR before organ procurement. Donor acceptance was lower in CPR patients (54 vs. 66%, p<0.001). There were several differences between the CPR cohort and the no-CPR cohort (Table 1). The CPR cohort was older (26 years vs. 25 years), more commonly female (37% vs. 30%), and more commonly white (66% vs. 60%; p<0.001 for all). They were also more likely to die from anoxia or overdose (46% vs. 4%) and less likely to die from a cerebrovascular accident (5% vs. 21%) or blunt (14% vs. 42%) or penetrating trauma (8% vs. 24%; p<0.001 for all). Finally, although more likely to have renal dysfunction (41% vs. 29%) and lower ejection fraction (EF; 15% vs. 10%) at time of procurement, they were less likely to have hepatic dysfunction (18% vs. 26%) or require inotropes (38% vs. 43%; p<0.001 for all).
Table 1.
Donor characteristics by CPR administration status.
| No CPR (n=26,659) | CPR (n=28,012) | p-value | |
|---|---|---|---|
| Demographics | |||
| Age (years) | 25[19–32] | 26[19–33] | <0.001 |
| Weight (kg) | 75[64–89] | 75[61–90] | 0.28 |
| Sex (male) | 18,740(70%) | 17,579(63%) | <0.001 |
| Race/Ethnicity | |||
| Black | 4,827(18%) | 4,433(16%) | <0.001 |
| Hispanic | 4,882(18%) | 4,054(14%) | <0.001 |
| White | 15,968(60%) | 18,542(66%) | <0.001 |
| Mechanism of Death | |||
| Anoxia/overdose | 1,052(4%) | 12,848(46%) | <0.001 |
| CVA | 5,470(21%) | 1,340(5%) | <0.001 |
| Blunt trauma | 11,290(42%) | 3,939(14%) | <0.001 |
| Penetrating trauma | 6,419(24%) | 2,181(8%) | <0.001 |
| Clinical Status | |||
| Renal dysfunction | 7,570(29%) | 11,544(41%) | <0.001 |
| Hepatic dysfunction | 6,833(26%) | 5,119(18%) | <0.001 |
| LVEF <50% | 2,404(10%) | 3,539(15%) | <0.001 |
| Inotrope use | 11,397(43%) | 10,527(38%) | <0.001 |
Values expressed as median [IQR] or n (%) as appropriate. CPR, cardiopulmonary resuscitation; CVA, cerebrovascular accident; LVEF, left ventricular ejection fraction.
Among the CPR cohort, median CPR duration was 20 minutes (IQR 11–32 minutes). Donor acceptance decreased across successive quartiles of CPR duration (≤10 minutes=58% [3882/6658]; 11–20 minutes=55% [3710/6764]; 21–32 minutes=52% [2643/5044]; ≥33 minutes=52% [2936/5682]; p<0.001). Multivariable logistic regression of organ acceptance (C-statistic=0.791, 95% CI=0.786–0.796) found that longer CPR duration predicted lower organ acceptance versus no CPR (≤10 minutes: HR=0.81 (0.75–0.88); 11–20 minutes: HR=0.77 (0.71–0.83); 21–32 minutes: HR=0.68 (0.63–0.75); ≥33 minutes: HR=0.68 (0.63–0.75); p<0.001 for all), as were renal dysfunction, LVEF <50%, and increasing age, whereas male sex and history of blunt trauma predicted higher organ acceptance (Table 2). A full list of variables included in the regression is included in Table 2.
Table 2.
Univariable and multivariable logistic regression models of organ acceptance.
| Variable | Univariable Model | Multivariable Model | ||
|---|---|---|---|---|
| HR[95% CI] | p-value | HR[95% CI] | p-value | |
| No CPR vs. CPR for | ||||
| ≤10 minutes | 0.72[0.68–0.76] | <0.001 | 0.81[0.75–0.88] | <0.001 |
| 11–20 minutes | 0.63[0.59–0.66] | <0.001 | 0.77[0.71–0.83] | <0.001 |
| 21–32 minutes | 0.57[0.54–0.60] | <0.001 | 0.68[0.63–0.75] | <0.001 |
| >32 minutes | 0.55[0.52–0.59] | <0.001 | 0.68[0.63–0.75] | <0.001 |
| History of blunt trauma | 1.79[1.72–1.86] | <0.001 | 1.54[1.45–1.64] | <0.001 |
| Hepatic dysfunction | 1.01[0.97–1.05] | 0.66 | ||
| Renal dysfunction | 0.71[0.68–0.73] | <0.001 | 0.77[0.73–0.81] | <0.001 |
| LVEF <50% | 0.03[0.03–0.03] | <0.001 | 0.03[0.03–0.03] | <0.001 |
| Inotrope use | 1.07[1.03–1.11] | <0.001 | ||
| Age (years) | 0.97[0.96–0.97] | <0.001 | 0.96[0.95–0.96] | <0.001 |
| White race | 0.83[0.71–0.76] | <0.001 | ||
| Male sex | 1.70[1.64–1.76] | <0.001 | 1.88[1.79–1.98] | <0.001 |
CPR, cardiopulmonary resuscitation; LVEF, left ventricular ejection fraction.
Of the recipient cohort, 52% (2,927/5,680) received donor organs requiring CPR. The iterative Kaplan-Meier analyses achieved significance for donor CPR duration of ≤55 minutes versus >55 minutes (mean survival 11.3 vs. 10.2 years, p=0.03) (Figure 1). The restricted cubic spline analysis demonstrated a nonlinear association between the hazard ratio for PTS and donor CPR duration. Supervised tuning of the cubic spline analysis demonstrated 55 minutes as the optimal cut point of CPR duration (Figure 2). Kaplan-Meier analysis of primary graft dysfunction was conducted as well, which found no difference in graft failure between the no-CPR group and CPR≤55 minutes (median graft survival=10.8 vs. 10.8 years, p=0.251) and a non-significant decreased graft survival for the CPR>55 minutes relative to CPR≤55 minutes (median graft survival=9.9 vs. 10.8 years, p=0.052).
Figure 1.
Representative Kaplan-Meier curves from iteratively increasing survival analyses. a) donor CPR duration ≤ 20 minutes vs. donor CPR duration > 20 minutes. b) donor CPR duration ≤ 40 minutes vs. donor CPR duration > 40 minutes. c) donor CPR duration ≤ 50 minutes vs. donor CPR duration > 50 minutes. d) donor CPR duration ≤ 55 minutes vs. donor CPR duration > 55 minutes. CPR, cardiopulmonary resuscitation.
Figure 2.
Restricted cubic spline graph of hazard ratio of post-transplant survival versus donor CPR duration showing optimal cut point in data at 55 minutes.
With an inflection point in survival identified, the recipient cohort was divided into three groups—no CPR, CPR≤55 minutes, and CPR>55 minutes. Several differences were identified between groups. The CPR>55 minutes group was younger (2 years [0–6]) and weighed less (11 kg [7–19]) than the CPR≤55 minutes group (3 years [0–11], p=0.03; 13 kg [7–35], p=0.02) or the no-CPR group (10 years [114], p<0.001; 30 kg [11–54], p<0.001) (Table 3). They were also more commonly status 1A/1 (72%) than the no-CPR group (64%, p=0.01) but not the CPR≤55 minutes group (69%, p=0.32). All three groups had similar rates of renal dysfunction, hepatic dysfunction, and dialysis and were equally likely to require VAD or ECMO support pre-transplant. Waitlist days were similar between groups. Both CPR groups had more congenital heart disease (≤55 minutes = 49%, p<0.001; >55 minutes = 52%, p<0.001) than the no-CPR group (40%).
Table 3.
Pediatric heart transplant recipient characteristics by donor CPR administration status.
| (a) No CPR (n=2,753) | (b) CPR ≤55 min (n=2,687) | (c) CPR >55 min (n=240) | p-value | |
|---|---|---|---|---|
| Recipient Demographics | ||||
| Age (years) | 10[1–14] | 3[0–11] | 2[0–6] | a-b, <0.001 a-c, <0.001 b-c, 0.03 |
| Weight (kg) | 30[11–54] | 13[7–35] | 11[7–19] | a-b, <0.001 a-c, <0.001 b-c, 0.02 |
| Sex (male) | 1,567(57%) | 1,437(53%) | 129(54%) | a-b, 0.01 a-c, 0.38 b-c, 0.99 |
| Race/Ethnicity | ||||
| Black | 593(22%) | 513(19%) | 36(15%) | a-b, 0.027 a-c, 0.021 b-c, 0.14 |
| Hispanic | 501(18%) | 564(21%) | 46(19%) | a-b, 0.01 a-c, 0.776 b-c, 0.56 |
| White | 1,469(53%) | 1,391(52%) | 133(55%) | 0.35 |
| Recipient Clinical Status | ||||
| Status 1A/1 | 1,738(64%) | 1,826(69%) | 171(72%) | a-b, <0.001 a-c, 0.01 b-c, 0.32 |
| Waitlist days | 58[19–140] | 58[23–126] | 71[22–133] | 0.30 |
| Renal dysfunction | 346(13%) | 318(12%) | 25(11%) | 0.53 |
| Hepatic dysfunction | 550(20%) | 480(18%) | 44(19%) | 0.17 |
| Dialysis | 98(4%) | 87(3%) | 10(4%) | 0.64 |
| Inotrope use | 1,283(47%) | 1,345(50%) | 120(50%) | a-b, 0.01 a-c, 0.35 b-c, >0.99 |
| Mech. ventilation | 348(13%) | 438(16%) | 44(18%) | a-b, <0.001 a-c, 0.02 b-c, 0.47 |
| VAD | 703(26%) | 659(25%) | 70(30%) | 0.23 |
| ECMO | 116(4%) | 108(4%) | 16(7%) | 0.15 |
| Diagnosis | ||||
| CHD | 1,105(40%) | 1,310(49%) | 125(52%) | a-b, <0.001 a-c, <0.001 b-c, 0.33 |
| Cardiomyopathy | 1,373(50%) | 1,178(44%) | 100(42%) | a-b, <0.001 a-c, 0.02 b-c, 0.59 |
Values expressed as median [IQR] or n (%) as appropriate. CPR, cardiopulmonary resuscitation; mech. ventilation, mechanical ventilation; VAD, ventricular assist device; ECMO, extracorporeal membrane oxygenation; CHD, congenital heart disease.
The recipient groups differed significantly in donor characteristics as well (Table 4). Donors for the CPR>55 minutes group were younger (2 years [1–6]) and smaller (14 kg [10–27]) than donors for the CPR≤55 minutes group (3 years [0–12], p<0.001; 17 kg [9–48], p<0.001) or the no-CPR group (12 years [2–17], p<0.001; 45 kg [14–66], p<0.001). There were similar rates of low donor EF (<50%) across all groups. There was a stepwise increase in donors dying by anoxia or overdose (no CPR = 3%; CPR ≤55 minutes = 34%; CPR >55 minutes = 64%; p<0.001 for all). There were also no statistically signifnicant differences between 30-day mortality rates between groups (No-CPR=2.4%, CPR≤55 minutes=3.0%, CPR>55 minutes=4.8%; p=0.069). Although there were no differences in acute rejection prior to discharge between groups (No-CPR=14.6%, CPR≤55 minutes=13.3%, CPR>55 minutes=12.7%; p=0.297), there was significantly less rejection in the CPR≤55 minutes group than either of the other two cohorts (No-CPR=20.9%, CPR≤55 minutes=17.3%, CPR>55 minutes=20.3%; p=0.009).
Table 4.
Pediatric heart transplant recipient’s donor characteristics by donor CPR administration status.
| (a) No CPR (n=2,753) | (b) CPR ≤55 min (n=2,687) | (c) CPR >55 min (n=240) | p-value | |
|---|---|---|---|---|
| Donor Demographics | ||||
| Age (years) | 12[2–17] | 3[0–12] | 2[1–6] | a-b, <0.001 a-c, <0.001 b-c, <0.001 |
| Weight (kg) | 45[14–66] | 17[9–48] | 14[10–27] | a-b, <0.001 a-c, <0.001 b-c, <0.001 |
| Sex (male) | 1,644(60%) | 1,611(60%) | 158(66%) | 0.18 |
| Race/Ethnicity | ||||
| Black | 635(23%) | 653(24%) | 48(24%) | 0.55 |
| Hispanic | 564(20%) | 546(20%) | 32(13%) | a-b, 0.91 a-c, 0.01 b-c, 0.01 |
| White | 1,466(53%) | 1,356(50%) | 137(57%) | a-b, 0.04 a-c, 0.28 b-c, 0.06 |
| Donor Clinical Status | ||||
| Renal dysfunction | 472(17%) | 649(24%) | 89(37%) | a-b, <0.001 a-c, <0.001 b-c, <0.001 |
| Hepatic dysfunction | 571(21%) | 341(13%) | 17(7%) | a-b, <0.001 a-c, <0.001 b-c, 0.02 |
| LVEF <50% | 57(2%) | 55(2%) | 3(1%) | 0.80 |
| Inotrope use | 1,214(44%) | 1,053(39%) | 92(39%) | a-b, <0.001 a-c, 0.12 b-c, 0.92 |
| Cause of death | ||||
| Anoxia/overdose | 91(3%) | 924(34%) | 154(64%) | a-b, <0.001 a-c, <0.001 b-c, <0.001 |
| CVA | 379(14%) | 131(5%) | 7(3%) | a-b, <0.001 a-c, <0.001 b-c, 0.22 |
| Blunt trauma | 1,389(50%) | 627(23%) | 23(10%) | a-b, <0.001 a-c, <0.001 b-c, <0.001 |
| Penetrating trauma | 510(19%) | 158(6%) | 2(1%) | a-b, 0.27 a-c, 0.04 b-c, 0.13 |
Values expressed as median [IQR] or n (%) as appropriate. CPR, cardiopulmonary resuscitation; CVA, cerebrovascular accident; LVEF, left ventricular ejection fraction.
Given the recipient group differences, a Cox regression (C-statistic=0.631, 95% CI=0.613–0.649) was performed to control for confounding variables in PTS. In the final model, CPR>55 minutes predicted worse PTS (HR=1.4 [1.03–1.90]) relative to no-CPR, but CPR≤55 minutes did not (HR=1.02 [0.90–1.17]) (Table 5). Other significant variables included donor age (HR=1.02 [1.01–1.03]), recipient white race (HR=0.73 [0.64–0.83]), recipient renal dysfunction (HR=1.53 [1.30–1.81]) and dialysis (HR=1.71 [1.31–2.23]), recipient ECMO (HR=1.64 [1.26–2.13]), and recipient diagnosis of congenital heart disease (HR=1.67 [1.46–1.90]). A full list of variables included in the regression is included in Table 5.
Table 5.
Univariable and multivariable Cox proportional hazards regression of post-transplant survival.
| Variable | Univariable Model | Multivariable Model | ||
|---|---|---|---|---|
| HR[95% CI] | p-value | HR[95% CI] | p-value | |
| No CPR vs. CPR for | ||||
| ≤55 minutes | 0.96[0.85–1.09] | 0.52 | 1.02[0.90–1.17] | 0.72 |
| >55 minutes | 1.35[1.02–1.80] | 0.04 | 1.40[1.03–1.90] | 0.03 |
| Age (years) | 1.01[1.00–1.01] | 0.10 | ||
| Donor age (years) | 1.01[1.00–1.02] | 0.007 | 1.02[1.01–1.03] | <0.001 |
| Sex (male) | 0.94[0.84–1.06] | 0.35 | ||
| Donor Sex (male) | 1.11[0.98–1.26] | 0.10 | ||
| Recipient race (white) | 0.75[0.66–0.84] | <0.001 | 0.73[0.64–0.83] | <0.001 |
| Donor race (white) | 1.00[0.89–1.13] | 0.95 | ||
| Waitlist status 1A/1 | 1.09[0.96–1.24] | 0.19 | ||
| Days on waitlist | 1.00[1.00–1.00] | 0.99 | ||
| Renal dysfunction | 1.76[1.52–2.05] | <0.001 | 1.53[1.30–1.81] | <0.001 |
| Donor renal dysfunction | 0.97[0.83–1.12] | 0.65 | ||
| Hepatic dysfunction | 1.34[1.16–1.54] | <0.001 | ||
| Donor hepatic dysfunction | 0.98[0.84–1.15] | 0.81 | ||
| Dialysis | 2.24[1.76–2.86] | <0.001 | 1.71[1.31–2.23] | <0.001 |
| Inotrope use | 1.06[0.94–1.19] | 0.37 | ||
| Donor inotrope use | 1.03[0.92–1.17] | 0.60 | ||
| Mechanical ventilation | 1.42[1.22–1.65] | <0.001 | ||
| VAD | 0.99[0.85–1.14] | 0.85 | ||
| ECMO | 2.06[1.64–2.58] | <0.001 | 1.64[1.26–2.13] | <0.001 |
| CHD | 1.50[1.33–1.69] | <0.001 | 1.67[1.46–1.90] | <0.001 |
| Donor LVEF <50% | 0.97[0.65–1.44] | 0.86 | ||
| Donor mechanism of death (Anoxia/overdose) | 1.00[0.86–1.17] | 0.99 | ||
Multivariable model built using backward stepwise elimination. HR, hazard ratio; CI, confidence interval; CPR, cardiopulmonary resuscitation; VAD, ventricular assist device; ECMO, extracorporeal membrane oxygenation; CHD, congenital heart disease; LVEF, left ventricular ejection fraction.
COMMENT
In the present analysis of over five thousand pediatric heart transplants, donor CPR≤55 minutes did not predict PTS whereas donor CPR>55 minutes predicted worsened PTS. Although prior studies have evaluated the impact of donor CPR duration, they typically divide CPR duration by arbitrary values based on either convention or statistical measures of center, neither of which produce clinically useful tools for transplant clinicians.[6, 7, 9–14] As such, this study fills a major gap in the literature by identifying the point at which donor ischemia has an adverse impact on PTS.
There were significant differences noted between the CPR and no-CPR donor cohorts. Donors who required CPR were sicker, with more renal dysfunction and reduced EF than their no-CPR counterparts. Notably, however, they had less hepatic dysfunction and inotrope requirement than no-CPR donors. Additionally, they were far more likely to have presented secondary to anoxia or overdose than trauma, which is likely reflective of different life support algorithms in trauma; CPR is not typically a first-line treatment for traumatic cardiac arrest. However, even when controlling for these differences in the logistic regression, longer CPR duration predicted organ non-acceptance, suggesting that transplant surgeons discriminated against these hearts on the basis of CPR duration alone, potentially using an arbitrary cutoff time.
There were significant differences among the recipients as well. Both the recipients and donors of hearts requiring CPR were smaller and younger than their no-CPR counterparts. This trend may reflect a relative scarcity in organs suitable for these patient populations (i.e., infants), and therefore a greater likelihood of accepting organs requiring CPR. In addition, there exists a belief, whether substantiated by literature or not, that infant donors would tolerate ischemic periods better than other donors. Recipients of donor organs requiring CPR were also overall sicker; they were more likely to be status 1A/1, have congenital heart disease, or require mechanical ventilation and/or inotropic support. Again, given their overall sicker nature, this trend may be explained by pressure to find a suitable organ acutely and an inability to reject an organ based solely on CPR duration. Interestingly, although many of the potential donor trends persisted in the actual donor cohort, rates of low EF were equivalent regardless of CPR duration and in the range of 1–2%. As seen in our potential donor analysis, the single greatest predictor of organ non-acceptance was EF<50% (HR 0.03 [0.03–0.03]), suggesting that transplant centers evaluate any organ with low EF with extreme skepticism.
In our survival analyses, we sought to identify an evidence-based inflection point in PTS contingent on donor CPR duration. The series of Kaplan-Meier curves identified an unexpectedly high value for that inflection point—nearly a full hour. However, that exact point was corroborated using restricted cubic spline analysis. Similarly, the Cox regression confirmed that even when controlling for other donor and recipient variables, patients with CPR≤55 minutes had similar outcomes to no CPR. Although this duration appears quite long, it is consistent with literature studying ischemia/reperfusion in both myocardial infarction and warm ischemia time in transplant.[15–17] The exact safe ischemic period has not been defined in humans; however, tissue salvage in both animals and humans has been possible hours following ischemic events.[15] Donors who survive cardiac arrest may even benefit from a shortterm period of ischemic preconditioning secondary to the biochemical response to the initial ischemia/reperfusion event, which could be protective during warm ischemic time.[15] Nevertheless, we also found in our analysis that CPR>55 minutes predicted worse outcomes even when other factors were controlled. In fact, the vast majority of recipients (99%) had a normal EF. Although transplant surgeons may use recovered EF as evidence of graft quality after cardiac arrest, our analysis suggests that prolonged CPR>55 minutes predicts adverse outcomes independently of EF.
The present study has a number of limitations. By nature, using retrospective data such as the UNOS database limits the ability to make causal inferences and is subject to reporting and selection biases as well as confounding. Analysis is also similarly limited by the specific variables that UNOS chooses to collect for each patient. Mandatory reporting to UNOS of all transplants in the United States does limit the risk of reporting/selection bias, however. Furthermore, it is valuable for transplant surgeons to have an understanding of how CPR duration predicts PTS, regardless of causation. Another potential limitation is the quality of the CPR data available from UNOS; granular data on number of arrest events or time between arrest and procurement is lacking. Indeed, there are many donor and recipient variables that would likely impact ischemic injury tolerance and post-transplant performance, many of which are not available in UNOS. However, the values in this analysis are the exact same values seen by transplant clinicians in DonorNet as they evaluate potential donors, so understanding the impact of these values remains highly clinically applicable. Further research should investigate factors that may protect against myocardial injury after CPR, which could inform more nuanced organ selection. Finally, the potential donors in the database were identified as those consented for procurement; we have no way of looking at those potential donors that local organ procurement organizations choose not to pursue for donation. Further research should survey organ procurement organizations to see which organs they choose not to pursue and whether those practices are standardized between regions.
The rate of pediatric heart transplants in the United States is limited by availability of donor organs. Although there is a trend toward using less traditionally ‘desirable’ organs, the fraction of transplants performed to available donors has decreased over time.[4] There are likely donors within that pool that have been rejected who would have been suitable for transplant. The present analysis finds that only durations of cardiopulmonary resuscitation lasting over nearly an hour predict poor post-transplant survival. Less than 4% of potential donors undergo CPR lasting >55 minutes. If rates of acceptance were similar between donors without CPR and donors with CPR under 55 minutes and normal EF, an additional 1,100 potential transplants could have occurred during the study period. Conversely, for those rare donors who have CPR durations >55 minutes, transplant surgeons should be skeptical of organ quality even with recovery of EF. These trends may be used to expand acceptance and utilization of quality donor organs in pediatric heart transplantation.
Funding Statement:
National Institutes of Health, R01HL147957: “Novel Methods to Grow the Impact of Pediatric Thoracic Transplantation;” Principal Investigators: David L.S. Morales, MD, Farhan Zafar, MD, MS
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest Statement: Dr. Morales is a consultant for Abbott, Inc., Azyio, Inc., Berlin Heart, Inc., CorMatrix, Inc., Peca, Inc., Syncardia, Inc., and Xeltis, Inc., and serves as a principal investigator for FDA trials sponsored by Peca, Inc. and Xeltis, Inc. Dr. Zafar has financial relationship (employment) with TransMedics, Inc. The remaining authors report no disclosures.
Waiver of Informed Consent: The present study was given institutional review board approval via waiver of informed consent (CCHMC IRB# 2018–6837; date of approval: 10/29/2018).
Meeting Presentation: Oral presentation at the 59th Annual Meeting of the Society of Thoracic Surgeons, January 21–23, 2023, San Diego, California, USA.
REFERENCES
- 1.Morales DLS, MD, Dreyer WJ, MD, Denfield SW, MD, et al. Over two decades of pediatric heart transplantation: How has survival changed?. The Journal of Thoracic and Cardiovascular Surgery 2007;133:632–9. [DOI] [PubMed] [Google Scholar]
- 2.Dipchand AI. Current state of pediatric cardiac transplantation. Annals of cardiothoracic surgery 2018;7:31–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Almond CSD, Thiagarajan RR, Piercey GE, et al. Waiting list mortality among children listed for heart transplantation in the united states. Circulation 2009;119:717–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dharmavaram N, Hess T, Jaeger H, et al. National trends in heart donor usage rates: Are we efficiently transplanting more hearts?. Journal of the American Heart Association 2021;10:e019655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Godown J, Kirk R, Joong A, et al. Variability in donor selection among pediatric heart transplant providers: Results from an international survey. Pediatric transplantation 2019;23:e13417, n/a. [DOI] [PubMed] [Google Scholar]
- 6.Quader MA, MD, Wolfe LG, MS, Kasirajan V, MD. Heart transplantation outcomes from cardiac arrest–resuscitated donors. The Journal of heart and lung transplantation 2013;32:1090–5. [DOI] [PubMed] [Google Scholar]
- 7.Cheng A, MD, Schumer Erin M., MD, MPH, Trivedi Jaimin R., MD, MPH, Van Berkel Victor H., MD, PhD, Massey HT, MD, Slaughter MS, MD. Does donor cardiopulmonary resuscitation Time Affect heart transplantation outcomes and Survival?. The Annals of thoracic surgery 2016;102:751–8. [DOI] [PubMed] [Google Scholar]
- 8.Kobashigawa J, Khush K, Colvin M, et al. Report from the american society of transplantation conference on donor heart selection in adult cardiac transplantation in the united states. American journal of transplantation 2017;17:2559–66. [DOI] [PubMed] [Google Scholar]
- 9.Ali AA, Lim E, Thanikachalam M, et al. Cardiac arrest in the organ donor does not negatively influence recipient survival after heart transplantation. Eur J Cardiothorac Surg 2007;31:929–33. [DOI] [PubMed] [Google Scholar]
- 10.L’Ecuyer T, Sloan K, Tang L. Impact of donor cardiopulmonary resuscitation on pediatric heart transplant outcome. Pediatric transplantation 2011;15:742–5. [DOI] [PubMed] [Google Scholar]
- 11.Mehdiani A, Immohr MB, Sipahi NF, et al. Successful heart transplantation after cardiopulmonary resuscitation of donors. The Thoracic and cardiovascular surgeon 2021;69:504–10. [DOI] [PubMed] [Google Scholar]
- 12.Quader M, Wolfe L, Katlaps G, Kasirajan V. Donor heart utilization following cardiopulmonary arrest and resuscitation: Influence of donor characteristics and wait times in transplant regions. Journal of Transplantation 2014;2014:519401–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Roth S, M’pembele R, Nucaro A, et al. Impact of cardiopulmonary resuscitation of donors on days alive and out of hospital after orthotopic heart transplantation. Journal of clinical medicine 2022;11:3853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Southerland KW, MD, Castleberry Anthony W., MD, MMCi, Williams JB, MD, Daneshmand MA, MD, Ali Ayyaz A., MD, PhD, Milano CA, MD. Impact of donor cardiac arrest on heart transplantation. Surgery 2013;154:312–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ostadal B. The past, the present and the future of experimental research on myocardial ischemia and protection. Pharmacol Rep 2009;61:3–12. [DOI] [PubMed] [Google Scholar]
- 16.Soares ROS, Losada DM, Jordani MC, Évora P, Castro-E-Silva O. Ischemia/reperfusion injury revisited: An overview of the latest pharmacological strategies. International Journal of Molecular Sciences 2019;20:5034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kalogeris T, Baines CP, Krenz M, Korthuis RJ. Ischemia/reperfusion. Comprehensive Physiology 2011;7:113–70. [DOI] [PMC free article] [PubMed] [Google Scholar]