Sex disparities in the current era of pediatric heart transplantation in the United States

Abstract

BACKGROUND:

While sex-related differences in transplant outcomes have been well characterized amongst adults, there are no sex-specific pediatric heart transplant studies over the last decade and none evaluating waitlist outcomes. In a contemporary cohort of children undergoing heart transplantation in the United States, this analysis was performed to determine if there were sex disparities in waitlist and/or post-transplant outcomes.

METHODS:

Retrospective review of Scientific Registry of Transplant Recipients database from December 16, 2011 to February 28, 2019 to compare male and female children after listing and after transplant. Demographic, clinical characteristics and outcomes were compared unadjusted and after 1:1 propensity matching for selected covariates.

RESULTS:

Of 4089 patients, 2299 (56%) were males. At listing, males were more likely to be older, have congenital heart disease (58% vs 48%), renal dysfunction (49% vs 44%) and implantable cardioverter defibrillator (9% vs 7%). At transplant, males were more likely to have renal (42 % vs 35%) and liver dysfunction (13% vs 10%), PRA >10% (29% vs 22%) and ischemic time >3.5 hours (p <0.05 for all). There were no significant sex differences found in unadjusted rates of transplant or mortality. After propensity matching, females had increased waitlist mortality (HR 1.3, 95%CI 1.04–1.5; p =0.019) compared to males. There were no significant differences in post-transplant morbidity or mortality (HR 1.2, 95% CI 0.93–1.5; p = 0.18) between groups.

CONCLUSION:

In a contemporary pediatric cohort, females have inferior heart transplant waitlist survival compared to propensity-matched males despite lower acuity of illness at listing and similar rates of transplantation. There were no sex-disparities noted in post-transplant outcomes.

Numerous sex-differences have been noted amongst adults presenting with heart failure.^1–3 Such differences have been documented in their heart failure phenotype, their ability to metabolize medications, utilization of heart failure therapy, and outcomes.^2–5

Sex-disparities in adults with advanced heart failure are also very well documented. Adult females constitute a minority of ventricular assist device (VAD) implants^6,7 and are at increased risk for morbidity and mortality post-VAD implantation.^8,9 Also adult females, especially those listed at the highest urgency, are more likely to be on mechanical ventilation and extracorporeal membrane oxygenation (ECMO) and less likely to have a VAD or intra-aortic balloon pump placement compared to males. Adult females have increased waitlist mortality compared to males even after adjusting for various demographic and clinical variables,^10,11 and they are less likely than men to undergo transplantation despite shorter waitlist times.¹² Post-transplantation, adult females have similar survival outcomes to male recipients.^13,14

There is surprisingly limited evidence to date highlighting sex-disparities in heart transplant outcomes for children with advanced heart failure.^15,16 A pediatric study from over a decade ago evaluated the effects of both recipient sex and sex mismatch on post-heart transplant outcomes.¹⁵ They found that after adjusting for possible clinical confounders, females had increased post-transplant mortality (HR [hazard ratio] 1.27). However, donor-recipient sex mismatching did not affect outcomes.¹⁵ A subsequent study mainly focusing on the effects of donor-recipient sex mismatch on post-heart transplant outcomes found that female recipients receiving male donor organs had long-term survival disadvantage.¹⁶ Only one of these older studies focused on the effects of recipient sex on post-heart transplant outcome,¹⁵ and to date there have been no studies evaluating the effects of recipient sex on waitlist outcomes in children listed for heart transplantation. We hypothesize that in the current era of increased VAD utilization in children,^17,18 there are no sex disparities in waitlist or post-transplant outcomes for children undergoing heart transplantation.

Methods

Study cohort

Pediatric (age <18 years) patients listed for heart-only transplantation from December 16, 2011 (date of Food and Drug Administration approval of Berlin EXCOR) to February 28, 2019 were included. Subjects listed for lung transplant (n = 394), heart-lung transplant (n = 43) and re-transplantation (n = 243) were excluded. This study used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donor, wait-listed candidates, and transplant recipients in the United States, submitted by the members of the Organ Procurement and Transplantation Network (OPTN). The Health Resources and Services Administration, United States. Department of Health and Human Services provides oversight to the activities of the OPTN contractor. The study was approved by the Cleveland Clinic IRB and informed consent was waived because data obtained from routine care were completely de–identified by SRTR prior to their transmission to the investigators.

Definition

Sex is defined in the SRTR database as male and female and is a mandatory categorical reporting variable.

Outcomes

Waitlist survival analysis included time from initial listing to a composite outcome of pre-transplant death or removal from the waitlist due to clinical deterioration, censored at transplantation or last date of follow-up (May 31, 2019). Deaths following removal from the waiting list were included in the analysis.

Post-transplant survival analysis included time from heart transplant to graft loss (death/re-transplantation), censored at last graft follow-up.

Statistical analysis

Data were summarized using medians and interquartile ranges for continuous variables and counts and percentages for categorical variables. For comparisons of characteristics at listing and at transplant by recipient sex, the Wilcoxon rank sum test was used for continuous/ordinal characteristics; the Pearson’s chi-square test or Fisher’s exact test was used for categorical characteristics, as appropriate. The Cochran–Armitage trend test was used to assess whether sex distribution changed over time.

For the waitlist cohort, unadjusted competing risk survival analysis was used to model multiple observed outcomes (transplant, death/delisting due to clinical deterioration and being alive). Cumulative death/delisting rates and cumulative transplant rates were compared between males and females using Gray’s test. For the post-transplant cohort, Kaplan-Meier plots were constructed to depict time to death/re-transplant, and the log-rank test was used to compare unadjusted time-to-event curves between both groups. Adjusted waitlist and post-transplant outcomes were compared between male and female candidates using propensity score matched analyses following imputation of missing data; the waitlist analysis was censored at transplantation or last date of follow-up.

Multiple imputation was utilized to handle variables with <5% missing data. This technique addresses data that is missing at random by generating several datasets which replace missing data with predicted values, then combining results from these datasets. Multiple imputation was performed using the MI procedure, and the results of the survival analyses performed on the imputed datasets were combined using the MIANALYZE procedure (SAS 9.4 software, SAS Institute, Cary, NC).¹⁹ Multiple imputation using fully conditional specification methods which generated five imputed datasets was performed separately on the waitlist and post-transplant analysis cohorts, using the regression method for continuous imputed variables and the discrimination function for categorical imputed variables. Variables included in the multiple imputation models are listed below. The survival event indicator and log-transformation of survival time were included in the multiple imputation procedures. As a sensitivity analysis, the cumulative baseline hazard approximated by the Nelson-Aalen estimator was used as replacement of log survival time in the multiple imputation procedures, and survival estimates were similar.

Variables included in the multiple imputation procedure for the waitlist cohort were diagnosis, age, sex, race/ethnicity, height, weight, body mass index (BMI), initial status, blood type, ventricular assist device (VAD), extracorporeal membrane oxygenation (ECMO), ventilator, intravenous (IV) inotropes, implantable cardioverter defibrillator (ICD) use, symptomatic cerebrovascular disease, dialysis, estimated glomerular filtration rate (eGFR), albumin, insurance type, survival event indicator, and log of survival time. The eGFR was calculated according to the Schwartz formula [eGFR = 0.413 × (height/serum creatinine), whereby height is in cm and creatinine is in mg/dl].²⁰

Variables included in the multiple imputation procedure for the post-transplant cohort were infection requiring IV drug therapy 2 weeks prior to transplant, intensive care unit (ICU), inhaled nitric oxide (iNO), bilirubin, ischemic time, cytomegalovirus (CMV) status, donor age, donor/recipient height ratio and weight ratio, donor sex, sex mismatch, race mismatch, Centers for Disease Control and Prevention (CDC) increased risk donor, donor drug abuse, donor eGFR, donor left ventricular ejection fraction (LVEF), donor inotropic support, donor diabetes, donor hypertension, donor cause of death, donor cardiopulmonary resuscitation (CPR) duration, post-transplant stroke, post-transplant dialysis, post-transplant permanent pacemaker, acute rejection episodes prior to discharge, length of stay, as well as characteristics used in the multiple imputation procedure for the waitlist cohort.

Variables with extensive (>30%) missing data were excluded from our survival analyses, including the imputation and matching procedures. These were as follows: intensive care unit (ICU) at listing, mean pulmonary arterial pressure (PAP) and mean pulmonary capillary wedge pressure at listing and transplant, panel reactive antibodies (PRA) at transplant, chest drain > 2 weeks and cardiac re-operation.

Survival analysis utilizing propensity score matching:

Propensity Score Matching is a statistical technique used to select matched pairs of patients with similar distributions of baseline covariates in order to reduce confounding and bias in the analysis of observational studies. Propensity score matching of male and female patient pairs was based upon imputed data sets. The scores were estimated for each patient averaging across the imputations; these scores were then used to identify a matched set of patients for each imputed data set.²¹ Variables utilized for propensity score matching at waitlist and post-transplant were those identified to be risk factors in a multivariable model among children listed/transplanted in the same time frame as the current study (December 16, 2011 to February 28, 2019).²² In addition, we added other variables that were of clinical interest for waitlist (blood type, insurance, ICD use, cerebrovascular disease, IV inotrope use and presence of VAD at listing) and post-transplant matching (recipient variables: age, race/ethnicity, BMI, initial status, blood type, cerebrovascular disease, ICD use, presence of VAD, IV inotrope use, ventilator use, eGFR, dialysis, iNO use, in ICU; donor variables: age, sex, donor to recipient height ratio, eGFR, LVEF, and donor cause of death). Patients were matched under a one-to-one nearest neighbor matching scheme without replacement, with the caliper set at 0.1 and the diagnosis forced to be the same within matching pair. Characteristics at listing and transplant were compared per imputed data set after matching using standardized differences (Figures S1–S2 and Tables S1–S2), with absolute differences less than 10% considered acceptable.²³ For survival analysis on the matched samples, robust variance estimators were used in Cox proportional hazards model to account for clustering within matched pairs, and double adjustment was used to remove residual confounding bias. For the analysis of post-transplant morbidity, matched samples were compared on binary outcomes using conditional logistic regression and on log-transformed length of stay using a paired t-test.

Additional sensitivity analyses:

A multivariable Cox regression analysis using the Fine and Gray method for competing risks was performed on imputed data sets using all waitlisted patients to assess the associations between candidate sex and waitlist death/de-listing due to deterioration after adjusting for covariates utilized for propensity score matching at waitlist, which included diagnosis, age, race, BMI, initial status, blood type, VAD, ECMO, ventilator, IV inotropes, dialysis, ICD use, symptomatic cerebrovascular disease, eGFR, albumin, and insurance type.

All tests were two-tailed and performed at an overall significance level of 0.05. SAS 9.4 software (SAS Institute, Cary, NC) was used for all analyses and plots.

Results

Waitlist cohort

During our study period, there were 4089 children listed for heart transplant, of which 2299 (56.2%) were male. Throughout our study period, there were no significant changes in the proportion of children listed by sex with males being more frequently listed for heart transplant (p = 0.79) (Figure 1).

Figure 1.

Sex differences amongst children listed for heart transplantation during the study period.

Differences in patient characteristics at listing

At listing, males were more likely to be older, have congenital heart disease [CHD] (57.8 vs 48.3%), have an implantable cardioverter defibrillator [ICD] (9.0 vs 6.7%) and have renal dysfunction (eGFR <90 ml/min/1.73m²) (49.2 vs 44.3%) (p < 0.05 for all). There were no differences noted in racial/ethnic distribution, listing status, and cardiac support (IV inotropes, mechanical ventilation, ECMO and VAD) between children of either sex (Table 1).

Table 1.

Differences in Demographic and Clinical Characteristics at Listing Between Male and Female Children

Factor	Female (N = 1,790)		Male (N = 2,299)		p-value
	N	Median [P25, P75] or N (%)	N	Median [P25, P75] or N (%)
Age at Listing (years)	1,790		2,299		<0.001b
< 1		707 (39.5)		802 (34.9)
1 – 10		650 (36.3)		763 (33.2)
11 – 17		433 (24.2)		734 (31.9)
Diagnosis	1,790		2,299		<0.001c
CMP		898 (50.2)		942 (41.0)
CHD		865 (48.3)		1,329 (57.8)
Other		27 (1.5)		28 (1.2)
Weight (kg)	1,787	11.0 [5.4, 32.5]	2,298	14.3 [5.7, 42.2]	<0.001b
Height (cm)	1,774	85.0 [59.0, 139.0]	2,285	96.5 [61.0, 151.8]	<0.001b
Race/Ethnicity	1,790		2,299		0.12c
Caucasian		894 (49.9)		1,240 (53.9)
African American		390 (21.8)		450 (19.6)
Hispanic		362 (20.2)		446 (19.4)
Asian		78 (4.4)		83 (3.6)
Other		66 (3.7)		80 (3.5)
UNOS Listing Status	1,790		2,299		0.76c
Status 1A		1,193 (66.6)		1,503 (65.4)
Status 1B		289 (16.1)		374 (16.3)
Status 2		271 (15.1)		367 (16.0)
Temporarily inactive		37 (2.1)		55 (2.4)
ICU	721	454 (63.0)	932	562 (60.3)	0.27c
ICD	1,778	119 (6.7)	2,286	205 (9.0)	0.008 c
Blood Type	1,790		2,299		0.40 c
A		630 (35.2)		797 (34.7)
AB		55 (3.1)		84 (3.7)
B		259 (14.5)		300 (13.0)
O		846 (47.3)		1,118 (48.6)
ECMO	1,790	112 (6.3)	2,299	168 (7.3)	0.19c
Ventilator	1,790	369 (20.6)	2,299	479 (20.8)	0.86c
IV Inotropes	1,790	854 (47.7)	2,299	1,066 (46.4)	0.39c
VAD	1,790	220 (12.3)	2,299	286 (12.4)	0.89c
eGFR < 90 mL/min/1.73m2	1,768	783 (44.3)	2,279	1,121 (49.2)	0.002 c
Albumin ≤3.5 g/dL	1,739	867 (49.9)	2,232	1,129 (50.6)	0.65c
Mean PAP (mm/Hg)	968	22.0 [16.0, 30.0]	1,314	20.5 [16.0, 30.0]	0.046 b
Mean PCWP (mm/Hg)	917	15.0 [11.0, 21.0]	1,213	15.0 [10.0, 20.0]	0.081b
Insurance	1,790		2,298		0.38c
Private		748 (41.8)		1,010 (44.0)
Medicaid		855 (47.8)		1,059 (46.1)
Other		187 (10.4)		229 (10.0)

Abbreviations: CMP indicates cardiomyopathy; CHD, congenital heart disease; ICD, implantable cardioverter defibrillator; ICU, intensive care unit; IV, intravenous; ECMO, extracorporeal membrane oxygenation; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; UNOS, United Network for Organ Sharing VAD, ventricular assist device; eGFR, estimated glomerular filtration rate.

p-value: b=Wilcoxon Rank Sum test, c=Pearson’s chi-square test.

Bolded p-values: statistically significant (p < 0.05).

Waitlist outcomes in males and females listed for heart transplantation

There were no significant differences noted in the transplant rates between male and female children in the unadjusted cohort after listing (p = 0.93) (Figure S3). Unadjusted waitlist survival between male and female children listed for heart transplantation in the current era were not significantly different (p=0.19). (Figure S4) However, after propensity score matching, females were noted to have increased waitlist mortality compared to males (HR_ps [hazard ratio after propensity score matching] 1.3, 95%CI 1.04–1.5; p=0.019) (Figure 2). The transplant rates after listing between propensity matched male and female children were not significantly different (p = 0.78) (Figure S5). Multivariable analysis using Cox regression analysis confirmed that females were at increased risk for waitlist mortality compared to males (aHR 1.22; 95% CI 1.03–1.44, p = 0.020) (Table S3).

Figure 2.

Propensity score matched analysis comparing waitlist survival for male and female children listed for heart transplantation.

Differences in recipient and donor characteristics at transplant

Of 2865 children transplanted, 1608 (56.1%) were male. At transplant, males continued to be older, more likely to have CHD (51.9 vs 44.6%), ICD (10.5 vs 7.5%) and renal dysfunction (41.6 vs 34.7%). Males were also more likely to have liver dysfunction [bilirubin ≥ 2mg/dl] (13.4 vs 9.5%), panel reactive antibody >10% (28.5 vs 21.7%) and longer ischemic times (p < 0.05 for all). No significant differences were noted between male and female children in terms of racial/ethnic distribution, blood type, cardiac support (IV inotropes, mechanical ventilation, ECMO and VAD) or waitlist times (Table 2).

Table 2.

Differences in Demographic and Clinical Characteristics at Transplant Between Male and Female Children

Factor	Female (N=1,257)		Male (N=1,608)		p-value
	N	Median [P25, P75] or N (%)	N	Median [P25, P75] or N (%)
Age at Transplant (years)	1,257		1,608		<0.001b
<1		386 (30.7)		451 (28.0)
1 – 10		498 (39.6)		539 (33.5)
11 – 17		356 (28.3)		570 (35.4)
≥ 18		17 (1.4)		48 (3.0)
Diagnosis	1,257		1,608		<0.001c
CMP		679 (54.0)		755 (47.0)
CHD		560 (44.6)		834 (51.9)
Other		18 (1.4)		19 (1.2)
Weight (kg)	1,257	14.5 [7.2, 40.5]	1,608	19.0 [7.8, 50.7]	<0.001b
Height (cm)	1,241	95.0 [66.0, 146.0]	1,597	107.4 [68.0, 159.0]	<0.001b
Race/Ethnicity	1,257		1,608		0.17c
Caucasian		633 (50.4)		876 (54.5)
African American		254 (20.2)		309 (19.2)
Hispanic		267 (21.2)		303 (18.8)
Asian		62 (4.9)		63 (3.9)
Other		41 (3.3)		57 (3.5)
UNOS Listing Status	1,257		1,608		0.31c
Status 1A		851 (67.7)		1,087 (67.6)
Status 1B		218 (17.3)		276 (17.2)
Status 2		171 (13.6)		208 (12.9)
Temporarily inactive		17 (1.4)		37 (2.3)
ICU	1,257	745 (59.3)	1,608	941 (58.5)	0.69c
ICD	1,253	94 (7.5)	1,604	168 (10.5)	0.006 c
Blood Type	1,257		1,608		0.53c
A		439 (34.9)		583 (36.3)
AB		43 (3.4)		68 (4.2)
B		187 (14.9)		224 (13.9)
O		588 (46.8)		733 (45.6)
IV Drug Therapy 2 Weeks prior to Transplant	1,249	222 (17.8)	1,596	295 (18.5)	0.63c
ECMO	1,257	76 (6.0)	1,608	110 (6.8)	0.39c
Ventilator	1,257	513 (40.8)	1,608	663 (41.2)	0.82c
IV Inotropes	1,257	792 (63.0)	1,608	1,021 (63.5)	0.79c
VAD	1,257	341 (27.1)	1,608	452 (28.1)	0.56c
eGFR < 90 mL/min/1.73m2	1,241	431 (34.7)	1,595	664 (41.6)	<0.001c
Albumin ≤3.5 g/dL	1,223	588 (48.1)	1,568	779 (49.7)	0.40c
Bilirubin ≥ 2 mg/dl	1,239	118 (9.5)	1,586	212 (13.4)	0.002 c
Mean PAP (mm/Hg)	751	22.0 [17.0, 30.0]	1,046	21.0 [16.0, 30.0]	0.12b
Mean PCWP (mm/Hg)	732	15.0 [10.0, 20.0]	1,006	14.0 [10.0, 20.0]	0.048 b
Insurance	1,257		1,608		0.35c
Private		511 (40.7)		694 (43.2)
Medicaid		617 (49.1)		765 (47.6)
Other		129 (10.3)		149 (9.3)
PRA >10%	451	98 (21.7)	536	153 (28.5)	0.014 c
Ischemic Time > 3.5 hours	1,255	663 (52.8)	1,601	908 (56.7)	0.038 c
Waitlist Time (months)	1,257	2.1 [0.80, 4.6]	1,608	2.1 [0.80, 4.6]	0.83b

Abbreviations: CMP indicates cardiomyopathy; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; ICD, implantable cardioverter defibrillator; ICU, intensive care unit; IV, intravenous; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; UNOS, United Network for Organ Sharing; PRA, panel reactive antibody; VAD, ventricular assist device; eGFR, estimated glomerular filtration rate.

p-value: b=Wilcoxon Rank Sum test, c=Pearson’s chi-square test.

Bolded p-values: statistically significant (p < 0.05).

Donors for male recipients compared to female recipients were more likely to be older, have significantly lower donor to recipient height and weight ratio, and more likely have head trauma as cause of death (50.1 vs 44.1%) or history of drug abuse (13.5 vs 10.9%). Female recipients were more likely to receive a sex mismatched donor heart (57.5 vs 35.8%) (p < 0.05 for all) (Table S4).

Post-heart transplant outcomes in male and female recipients

Unadjusted post-heart transplant survival between male and female children in the current era were not significantly different (p = 0.55) (Figure S6). Even after propensity matching, there were no significant differences noted in post-heart transplant survival between male and female recipients (HR_ps 1.2, 95% CI 0.93–1.5; p = 0.18) (Figure 3).

Figure 3.

Propensity score matched analysis comparing post-transplant survival for male and female children undergoing heart transplantation.

The unadjusted analysis revealed that males were significantly more likely than females to undergo post-transplant dialysis (7.7 vs. 5.3%, p=0.014) and cardiac re-operation (11.0 vs 6.8%, p=0.027) (Table S5). However, after propensity matching there were no differences noted in the morbidity outcomes such as post-transplant stroke, dialysis, permanent pacemaker placement, cardiac re-operation, acute rejection episodes prior to discharge and length of stay between male and female recipients undergoing heart transplantation (Table 3).

Table 3.

Differences in Post-Transplant Morbidity Outcomes in a Propensity Matched Cohort of Male and Female Children Undergoing Heart Transplantation

Factor	Female (N=1,179)		Male (N=1,179)
	N	Median [P25, P75] or N (%)	N	Median [P25, P75] or N (%)	p-value
Stroke	1,179	39 (3.3)	1,179	47 (4.0)	0.35a
Dialysis	1,179	65 (5.5)	1,179	77 (6.5)	0.29a
Permanent Pacemaker	1,179	10 (0.85)	1,179	9 (0.76)	0.82a
Acute Rejection Episodes prior to Discharge	1,179	166 (14.1)	1,179	155 (13.1)	0.51a
Length of Stay (days)	1,179	20.0 [13.0, 36.0]	1,179	19.0 [13.0, 35.0]	0.93b

p-value: a=Conditional logistic regression with multiple imputation, b=Paired t-test on log-scale with multiple imputation.

Discussion

There are four important findings in our study of a contemporary cohort of children undergoing heart transplantation in the United States. First, females form a minority of listings compared to males, and this trend did not change over the study period. Second, male children have a higher acuity of illness at listing and heart transplantation compared to females. Third, in the current era, female children were noted to be at increased risk for waitlist deaths or removal from the waitlist due to clinical deterioration after matching for clinical confounders. Finally, there was no evidence of disparity in post-heart transplant morbidity or mortality currently. (Central Illustration).

As noted in our study, adult studies both from the United States and Europe have revealed that females form a minority of heart transplant listings, accounting for 26% of listings in the United States and 18% in Europe.^3,24 At first glance, it appears that females in our study were less sick than males. They were less likely to have congenital heart disease at listing and were less likely to be sensitized or have end organ dysfunction at transplant. We speculate that the lower prevalence of females with CHD at listing maybe because of the known sex-differences in CHD phenotype. It is recognized that females are more likely to have CHD such as atrial septal defects, patent ductus arteriosus and Ebstein’s anomaly, while males are more likely to have left sided heart lesions (aortic stenosis, aortic coarctation, hypoplastic left heart syndrome) and conotruncal lesions (d-transposition of the great vessels and double outlet right ventricle).^25–28 The CHD lesions seen in males are more likely to present with heart failure which may partly explain the higher number of males being listed with CHD.²⁹ Unfortunately, the SRTR database does not provide granular data on the type of CHD to confirm our beliefs. The SRTR also does not provide data on children that were evaluated and never listed for heart transplant; hence it is unknown if an implicit bias exists amongst providers when listing female compared to male children with advanced heart failure. Literature in adults supports the notion that there is an implicit sex/gender bias when diagnosing and treating females who have cardiovascular diseases,^30,31 and in allocation of advanced heart failure therapies.^3,32 Future studies evaluating implicit sex bias in pediatric listing practices may be able to address this question.

Surprisingly, while adult studies show that females on the heart transplant waitlist have higher acuity of illness (more likely to be mechanically ventilated, on inotrope, ECMO) and are less likely to receive a VAD compared to their male counterparts, our findings were disparate. Female children had a lower acuity of illness both at listing and at transplantation. Except for ICD use which was higher in males both at listing and transplantation, we did not find any disparity in the listing status, or the amount of cardiac support in children of either sex. The higher ICD use is in males can at least partially be explained by the fact that males were more likely to have CHD, a subgroup inherently prone to arrhythmias.³³ It is also possible that there is an implicit bias when considering ICD implantation in females,^25,34 as is seen in adults.^3,35

In contrast to adult studies that have highlighted the increased waitlist mortality in females,^10,11,36,37 ours is the first pediatric study to evaluate disparity in waitlist outcomes for female children undergoing heart transplantation. In a propensity matched cohort, we found that females had a 30% higher risk for waitlist mortality compared to males in the current era. This difference in mortality persisted after adjusting for baseline differences in demographic and clinical variables captured by SRTR. It is known that while the SRTR captures a variety of variables, both demographic and clinical, it does not capture a variety of important variables.³⁸ In children with heart failure, laboratory parameters such as serum sodium, natriuretic peptide levels, organ perfusion parameters (lactate, mixed venous saturation), echocardiographic parameters such as ventricular ejection fraction and ventricular dimensions (wall thickness and dilation) are all known to affect outcomes.^39–43 None of these variables were available to us through the SRTR database. Moreover, hemodynamic information which is also helpful in prognostication of children with advanced heart failure, was not available for all patients.⁴⁴ Given the lack of availability, differences in these variables could not be adjusted for which may explain the observed differences in waitlist survival. Also, the SRTR data is only available at listing and at transplantation and the trajectory of advanced heart failure may be very different in males compared to females and this needs to be explored in subsequent studies.

Compared to a previous pediatric study utilizing the OPTN/SRTR data from over a decade ago that found that females had a 27% increased risk for post-heart transplant mortality,¹⁵ we found no such differences in post-transplant survival in the current era which is encouraging. Similar to our findings, adult studies have shown that females typically have similar or superior survival as compared to males after heart transplantation,^13,45 The prior pediatric study¹⁵ also did not evaluate post-transplant morbidity as was done in our study. We found that after propensity matching, post-transplant morbidity outcomes are no different for male and female recipients undergoing heart transplantation currently.

Study limitations

While the main strength of our study is the fact that we were able to analyze the largest possible cohort of children undergoing heart transplantation in the current era and that there was 100% capture of recipient sex in this database, there are a few limitations that we would like to highlight. The adjudication of recipient sex is by the individual entering the SRTR data and it is possible although unlikely that there may be errors in entering this data which could affect the interpretation of our results. While propensity matching allowed us to make our cohorts of interest (males and females) similar at baseline, we were only able to match on available clinical risk factors. As we highlighted in our discussion, we adjusted for reported demographic and clinical variables in the SRTR, which still misses key laboratory, hemodynamic and echocardiographic data. It is entirely possible that differences in these unmeasured clinical variables could help explain the reported differences in waitlist survival.²³ In our analysis, we were able to match 94% females, but only 73% males both at listing and at transplant. While propensity matching is a robust technique to compare groups of patients that are similar on risk factors, the matched patients may be not be representative of the whole cohort. In our study, the propensity score analysis of matched males and females indicates that sex alone is a risk factor for waitlist mortality among otherwise similar patients. Our additional sensitivity analysis using all male and female patients in a multivariable regression model confirms that sex is an independent risk factor for waitlist mortality. Finally, we had no information regarding children of either sex who were evaluated but not listed for heart transplantation. It is known in adults that implicit bias exists amongst providers when considering females for advanced heart failure therapies;^3,32 whether these biases exist when listing pediatric patients is unknown and will need to be explored.

Conclusions

In the current era of pediatric heart transplantation, females form a minority of listed recipients. They have a lower acuity of illness at listing and at transplant, however, they have increased waitlist mortality after adjusting for clinical confounders. Post-heart transplant outcomes however, are similar for children of either sex undergoing transplantation. Future studies incorporating key laboratory, echocardiographic and hemodynamic parameters may help explain the noted differences in waitlist survival. Also, studies evaluating implicit physician bias when listing children of different sex need to be explored.