Abstract
Background:
Pediatric heart transplant (HT) candidates experience high waitlist mortality due to a limited donor pool that is constrained in part by anti-HLA sensitization. We evaluated the impact of CDC and Flow donor-specific crossmatch (XM) results on pediatric HT outcomes.
Methods:
All pediatric HT between 1999 and 2019 in the OPTN database were included. Donor-specific XM results were sub-categorized based on CDC and Flow results. Primary outcomes were treated rejection in the first year and time to death or allograft loss. Propensity scores were utilized to adjust for differences in baseline characteristics.
Results:
A total of 4695 pediatric HT patients with T-cell XM data were included. After propensity score adjustment, a positive T-cell CDC-XM was associated with two times higher odds of treated rejection (OR 2.29 (1.56,3.37)) and shorter time to death/allograft loss (HR 1.50 (1.19,1.88)) compared to a negative Flow-XM. HT recipients who were Flow-XM positive with negative/unknown CDC-XM did not have higher odds of rejection or shorter time to death/allograft loss. An isolated positive B-cell XM was also not associated with worse outcomes. Over the study period XM testing shifted from CDC- to Flow-based assays.
Conclusions:
A positive donor-specific T-cell CDC-XM was associated with rejection and death/allograft loss following pediatric HT. This association was not observed with a positive T-cell Flow-XM or B-cell XM result alone. The shift away from performing the CDC-XM may result in loss of important prognostic information unless the clinical relevance of quantitative Flow-XM results on heart transplant outcomes is systematically studied.
Keywords: Pediatric heart transplant, crossmatch, Histocompatibility
INTRODUCTION
The presence of anti-Human Leukocyte Antigen (HLA) antibodies prior to heart transplant (HT) is associated with higher levels of rejection, cardiac allograft vasculopathy, and allograft failure post-HT.1–4 Historically, centers required a negative prospective donor-specific crossmatch (XM) in patients with high levels of anti-HLA antibodies. Requiring a negative XM, however, has been associated with longer waitlist times and higher mortality.2 With the increasing prevalence of anti-HLA antibodies in the pediatric population, particularly after procedures to palliate congenital heart disease, transplant across a positive XM is more commonly considered as a viable option.5
Several studies have shown higher rates of rejection, cardiac allograft vasculopathy, and mortality in complement dependent cytotoxicity (CDC)-XM positive pediatric HT recipients compared to those with a negative CDC-XM.4,6–11 The CDC-XM detects the presence of complement-fixing anti-HLA antibodies, and while this assay identifies clinically relevant antibodies, it is limited by its sensitivity, availability of reagents, technical variability, time to perform, and cost.12,13 Thus, the methods for performing donor-specific XM testing have evolved over time with fewer histocompatibility labs offering CDC-XM testing and instead offering flow cytometry-based XM. Flow-based assays are more sensitive, reproducible, and faster to perform, however the clinical implications of a positive Flow-XM on pediatric HT outcomes are less well understood.12–14
In this study, we sought to understand the impact of a positive CDC- and Flow-XM on the risk of rejection and allograft loss after pediatric heart transplantation utilizing twenty years of contemporary data from the Organ Procurement and Transplantation Network (OPTN).
METHODS
Study Design & Patient Population
Utilizing OPTN data as of September 2020, all pediatric heart transplants (recipient age <21 years) between January 1, 1999 and August 31, 2019 were identified. Patients were excluded if they had multi-organ transplants (e.g., heart-lung) or if they had no XM data available. Data collected included patient characteristics (age at transplant, weight, height, sex, race, UNOS listing status at transplant, cardiac diagnosis, hemodynamic support, blood type, creatinine, bilirubin, year of transplant), donor/graft-specific data (donor age, sex, graft ischemic time, cause of death), and XM data. Patients were categorized based on donor-specific T-cell XM results into four groups: 1) CDC-XM positive, 2) Flow-XM positive/CDC-XM unknown, 3) Flow-XM positive/CDC-XM negative, 4) Flow-XM negative (Figure 1A). In secondary analysis, we assessed the impact of B-cell XM on outcomes conditional on a negative T-cell XM result. We first identified patients with a negative T-cell Flow-XM and then categorized these patients into the same four groups based on B-cell XM results (Figure 1B). For initial comparison of XM positive to XM negative patients, subgroups 1, 2, and 3 formed the XM positive group while group 4 formed the XM negative control group. Patients without Flow-XM results were excluded from the negative control groups in both T- and B-cell analyses to minimize misclassification bias. Outcomes of interest were time to allograft loss, time to death, and treated rejection within the first year post-transplant.
Figure 1.
Patient Stratification based on CDC and Flow Crossmatch Results
(A) Patients were categorized based on donor-specific T-cell CDC and Flow crossmatch (XM) testing into four groups: 1) CDC-XM positive (Blue), 2) Flow-XM positive/ CDC-XM unknown (Green), 3) Flow-XM positive/CDC-XM negative (Red), 4) Flow-XM negative (Purple). Patients with no XM results and those who were T-Cell CDC-XM Negative/Unknown and Flow-XM Unknown were excluded from analysis.
(B) For the B-Cell XM analysis, only patients with a negative T-Cell Flow XM were included. Patients were then classified in a similar manner according to the results of the B-cell donor-specific CDC- and Flow-XM testing into four groups.
This project involved de-identified external registry data and was exempt from review by the Boston Children’s Hospital IRB. The study was compliant with the ISHLT Ethics statement.
Statistical Analyses
For continuous variables, patient and donor characteristics were summarized using medians with 25th and 75th percentiles, and positive versus negative XM results were compared using the Wilcoxon rank sum. Categorical variables were summarized as frequencies and percentages and compared using the chi-square test. Time from transplantation to death or graft loss was estimated using the Kaplan-Meier method and compared by XM status using the Cox proportional hazards model. Patients who did not experience the outcome event were censored at the time of last known follow-up. Schoenfeld residual plots were used to evaluate the assumption of proportional Hazards. Treated rejection within the first year post-transplant was compared using logistic regression. Hazard ratios (HR) and odds ratios (OR) were estimated with 95% confidence intervals.
Propensity Matching:
To adjust for baseline differences between the groups, propensity scores were estimated from a multivariable logistic regression model with positive XM (CDC or Flow) as the outcome and all potential cofounders included as covariates, whether significantly different between groups or not.15 Multiple imputation was used for small percentages of missing values for renal function (0.8%), total bilirubin (2.7%), and ischemic time (1.8%). Propensity scores were utilized in two ways: first as a covariate in regression models assessing the relationships between XM status and outcome, and second to create matched XM-positive and XM-negative groups. One-to-one matching of XM-positive and XM-negative patients was performed using a nearest neighbor algorithm, with a maximum caliper of 0.01 for propensity score. When using Cox and logistic regression models for the analyses of matched cohorts, robust variance estimators were used. To account for variability in practice among transplant centers, sensitivity analyses were performed using mixed effects models adjusting for propensity score with a random intercept for institution. Statistical analyses were conducted using SAS version 9.4 and Stata version 16.
RESULTS
Baseline Demographics
A total of 8198 patients less than 21 years of age underwent heart transplant between 1999 and 2019. Of these, 3503 patients were excluded from the T-cell XM analysis (72 multi-organ transplants, 1568 with no T-cell XM data, and 1863 with negative/unknown T-cell CDC-XM results and unknown Flow-XM results), leaving 4695 transplants included in the analysis (Figure 1A).
Patients with a positive T-cell XM (CDC or Flow) were compared to those with a negative XM (Flow) with regards to baseline demographics, clinical status, and donor characteristics (Table 1). Patients with a positive T-cell XM differed from those with a negative XM with regard to cardiac diagnoses and were more likely to undergo HT for congenital heart disease or failure of a prior HT (p<0.001). Year of transplant also differed between the two groups, with those with a positive XM transplanted more frequently in the earlier part of the analyzed time period (p<0.001). Positive T-cell XM patients were also more likely to require extracorporeal membrane oxygenation (ECMO), ventilator, or ventricular assist device (VAD) support prior to transplant (p=0.002) and had longer ischemic times (3.7 vs. 3.6 hours, p=0.02). There were no significant differences in age at transplant, weight, sex, race, UNOS listing status, blood type, renal dysfunction, bilirubin, donor age, donor sex, and donor cause of death between the two groups. The positive XM groups were further divided based on T-cell CDC- and Flow-XM results (Figure 1A) and compared with regard to baseline demographics (Supplemental Table S1). As the likelihood of a XM-positive HT differed with respect to key pre-transplant variables known to impact post-HT outcomes, propensity scores were created to account for baseline differences. The multivariable logistic regression model used for propensity score is shown in Supplemental Table S2 and baseline characteristics across the total cohort and after propensity matching are shown in Supplemental Table S3.
TABLE 1.
Baseline characteristics between transplants with a positive T-cell crossmatch (XM) (CDC or Flow) and those with a negative T-cell Flow XM. The comparison for a positive B-cell XM and negative B-cell Flow XM conditional on a negative T-cell Flow XM is also shown.
Data represent median [interquartile range] or number (percent). Acronyms: ECMO extracorporeal membrane oxygenation, kg kilograms, VAD ventricular assist device, XM crossmatch
| T-Cell |
B-Cell |
|||||
|---|---|---|---|---|---|---|
| XM Positive (n = 547) | XM Negative (n = 4148) | p-Value | XM Positive (n = 309) | XM Negative (n = 3718) | p-Value | |
| Age at transplant (years) | 7 [1, 14] | 7 [0, 14] | 0.76 | 8 [2, 14] | 6 [0, 14] | 0.070 |
|
| ||||||
| Weight (kg) | 21.4 [9.0, 46.9] | 20.9 [7.9, 50.3] | 0.71 | 24.0 [10.5, 49.0] | 20.4 [7.7, 50.5] | 0.14 |
|
| ||||||
| Female sex | 228 (42%) | 1833 (44%) | 0.27 | 134 (43%) | 1648 (44%) | 0.77 |
|
| ||||||
| Race | 0.96 | 0.54 | ||||
| White | 303 (55%) | 2292 (55%) | 183 (59%) | 2048 (55%) | ||
| Black | 117 (21%) | 860 (21%) | 61 (20%) | 765 (21%) | ||
| Hispanic | 93 (17%) | 714 (17%) | 48 (16%) | 654 (17%) | ||
| Other | 34 (6%) | 282 (7%) | 17 (6%) | 251 (7%) | ||
|
| ||||||
| UNOS listing status | 0.22 | 0.095 | ||||
| 1A | 347 (63%) | 2518 (61%) | 172 (56%) | 2277 (61%) | ||
| 1B / Old status 1 | 85 (16%) | 770 (19%) | 60 (19%) | 691 (19%) | ||
| 2 / inactive | 115 (21%) | 860 (21%) | 77 (25%) | 750 (20%) | ||
|
| ||||||
| Cardiac diagnosis | <0.001 | <0.001 | ||||
| Congenital heart disease | 306 (56%) | 1782 (43%) | 138 (45%) | 1584 (43%) | ||
| Cardiomyopathy | 178 (32%) | 1954 (47%) | 113 (37%) | 1782 (48%) | ||
| Myocarditis | 9 (2%) | 125 (3%) | 12 (4%) | 113 (3%) | ||
| Re-transplant | 52 (10%) | 239 (6%) | 45 (15%) | 192 (5%) | ||
| Other | 2 (<1%) | 48 (1%) | 1 (<1%) | 47 (1%) | ||
|
| ||||||
| Hemodynamic support | 0.002 | 0.85 | ||||
| ECMO | 36 (7%) | 193 (5%) | 14 (5%) | 170 (5%) | ||
| Ventilator | 82 (15%) | 505 (12%) | 38 (12%) | 444 (12%) | ||
| VAD | 112 (20%) | 714 (17%) | 59 (19%) | 644 (17%) | ||
| Medical therapy | 317 (58%) | 2736 (66%) | 198 (64%) | 2460 (66%) | ||
|
| ||||||
| Blood type | 0.40 | 0.45 | ||||
| O | 252 (46%) | 1908 (46%) | 132 (43%) | 1722 (46%) | ||
| A | 185 (34%) | 1511 (36%) | 123 (40%) | 1332 (36%) | ||
| B | 87 (16%) | 561 (14%) | 44 (14%) | 508 (14%) | ||
| AB | 23 (4%) | 168 (4%) | 10 (3%) | 156 (4%) | ||
|
| ||||||
| Renal dysfunction* | 0.67 | <0.001 | ||||
| Normal | 350 (64%) | 2720 (66%) | 164 (53%) | 2485 (67%) | ||
| Moderate | 165 (30%) | 1213 (29%) | 121 (39%) | 1050 (28%) | ||
| Severe | 32 (6%) | 215 (5%) | 24 (8%) | 183 (5%) | ||
|
| ||||||
| Total bilirubin | 0.43 | 0.21 | ||||
| <0.6 | 226 (41%) | 1813 (44%) | 124 (40%) | 1647 (44%) | ||
| 0.6–1.2 | 188 (34%) | 1416 (34%) | 119 (39%) | 1251 (34%) | ||
| >1.2 | 133 (24%) | 919 (22%) | 66 (21%) | 820 (22%) | ||
|
| ||||||
| Year of transplant | <0.001 | <0.001 | ||||
| 1999–2003 | 85 (16%) | 299 (7%) | 40 (13%) | 182 (5%) | ||
| 2004–2008 | 109 (20%) | 668 (16%) | 57 (18%) | 576 (15%) | ||
| 2009–2013 | 160 (29%) | 1166 (28%) | 97 (31%) | 1066 (29%) | ||
| 2014–2019 | 193 (35%) | 2015 (49%) | 115 (37%) | 1894 (51%) | ||
|
| ||||||
| Donor age (years) | 8 [1, 16] | 8 [1, 17] | 0.46 | 9 [1, 17] | 8 [1, 16] | 0.15 |
|
| ||||||
| Donor sex female | 237 (43%) | 1615 (39%) | 0.051 | 124 (40%) | 1437 (39%) | 0.63 |
|
| ||||||
| Ischemic time (hours) | 3.7 [3.1, 4.3] | 3.6 [2.9, 4.3] | 0.020 | 3.7 [3.0, 4.3] | 3.6 [2.9, 4.2] | 0.16 |
|
| ||||||
| Donor cause of death stroke | 38 (7%) | 269 (6%) | 0.65 | 21 (7%) | 236 (6%) | 0.72 |
Renal dysfunction is normal if age < 1 year and GFR ≥ 40, or age = 1 year and GFR ≥ 60, or age ≥ 2 years and GFR ≥ 90; moderate if age < 1 year and GFR 20–39, or age = 1 year and GFR 30–59, or age ≥ 2 years and GFR 40–89; severe if dialysis prior to transplant, or age < 1 year and GFR < 20, or age = 1 year and GFR < 30, or age ≥ 2 years and GFR < 40.
Type of Donor-specific Crossmatch Testing Performed at Time of Pediatric Heart Transplant
Over the course of the study from 1999 to 2019 there was a significant change in the type of donor-specific XM testing performed, with CDC-XM being more common (79% of transplants) in the first 5 years of the study period and Flow-XM becoming more common (87% of transplants) during the last 5 years of the study (Figure 2). Additionally, only 31% of heart transplants had a CDC-XM performed during the last 5 years of the study period.
Figure 2.
Change in Crossmatch Technique Over Study Period
Percent of transplants with a CDC-XM (Green), Flow-XM (Blue) or both (Red) by study year.
Treated Rejection in the First Year after Heart Transplantation is More Common if the T-Cell CDC-XM is Positive
We first evaluated the association between XM status and risk of treated rejection in the first year. Six hundred ninety-five patients were excluded from analyses given unknown treated rejection status within the first year (i.e., not reported), leaving 4,000 recipients for analyses. In unadjusted analysis, the odds of treated rejection were 1.8 times higher in patients with a positive T-cell XM (OR 1.80 (1.46, 2.23); Table 2). This relationship remained significant when adjusting for propensity score to account for baseline differences between groups (OR 1.53 (1.23, 1.91); Figure 3A, Table 2). When analyzing the T-cell XM positive subgroups (CDC-XM positive, Flow-XM positive/CDC-XM negative, and Flow-XM positive/CDC-XM unknown), the T-cell CDC-XM positive group had 3 times higher odds of treated rejection within the first year than the XM negative group prior to propensity score adjustment (OR 3.07 (2.11, 4.46), Table 2). In propensity score adjusted cohorts, patients who were CDC-XM positive remained at higher odds of rejection within the first year (OR 2.29 (1.56, 3.37)) while patients who were Flow-XM positive did not have a significantly increased risk of rejection whether the CDC-XM result was negative or unknown (Table 2, Figure 3B). Odds ratios did not change significantly in mixed effects models with random intercepts for center (Supplemental Table S4). Additional analyses using propensity matched cohorts also yielded similar results (Table 2).
TABLE 2.
Risk of treated rejection within the first year post heart transplant
Odds ratio for treated rejection within the first year after heart transplant are presented first as an unadjusted analysis followed by analyses that (1) adjusted for propensity score or (2) matched on propensity score. For each comparison the groups are stratified by crossmatch results.
| Crossmatch Group | Number of Treated Rejection/Number of Patients | Unadjusted Odds Ratio (95% CI) | Adjusted for Propensity Score Odds Ratio (95% CI) | Matched on Propensity Score Odds Ratio (95% CI) | |
|---|---|---|---|---|---|
| T-Cell | XM Negative | 821/3571 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) |
| XM Positive | 150/429 | 1.80 (1.46, 2.23) | 1.53 (1.23, 1.91) | 1.60 (1.15, 2.22) | |
|
| |||||
| XM Negative | 821/3571 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) | |
| Flow Positive/ CDC Negative | 25/82 | 1.47 (0.91, 2.37) | 1.32 (0.82, 2.15) | 1.38 (0.68, 2.83) | |
| Flow Positive/ CDC Unknown | 70/232 | 1.45 (1.08, 1.94) | 1.30 (0.97, 1.75) | 1.21 (0.76, 1.92) | |
| CDC Positive | 55/115 | 3.07 (2.11, 4.46) | 2.29 (1.56, 3.37) | 3.00 (1.52, 5.94) | |
|
| |||||
| B-cell | XM Negative | 715/3215 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) |
| XM Positive | 76/266 | 1.40 (1.06, 1.85) | 1.12 (0.84, 1.50) | 1.26 (0.83, 1.91) | |
|
| |||||
| XM Negative | 715/3215 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) | |
| Flow Positive/ CDC Negative | 12/53 | 1.02 (0.54, 1.96) | 0.63 (0.31, 1.27) | 0.94 (0.46, 1.90) | |
| Flow Positive/ CDC Unknown | 48/170 | 1.38 (0.98, 1.94) | 1.18 (0.83, 1.67) | 1.24 (0.80, 1.93) | |
| CDC Positive | 16/43 | 2.07 (1.11, 3.87) | 1.66 (0.88, 3.15) | 1.86 (0.94, 3.67) | |
FIGURE 3.
Treated Rejection within the First Year after Heart Transplant
Percent treated rejection within the first year after heart transplant for propensity matched cohorts (A) T-Cell XM positive vs. XM negative, (B) T-cell XM subgroups, (C) B-Cell XM positive vs. XM negative, and (D) B-cell XM subgroups. Black bars represent 95% confidence intervals. P values are comparisons to the reference group (Negative crossmatch “XM-“).
XM+: crossmatch positive, XM-: crossmatch negative, CDC+: CDC-crossmatch positive, CDC-: CDC-crossmatch negative, Flow+: Flow-crossmatch positive, CDCnr: CDC no result (CDC-crossmatch unknown)
Time to Allograft Loss or Death is Shorter after T-cell CDC-XM Positive Heart Transplantation
In unadjusted analysis, a positive T-cell XM was significantly associated with a higher risk of death or graft loss (HR 1.47 (1.27, 1.70)) and these results remained significant after adjusting for propensity score (HR 1.29 (1.11, 1.50)) and when matching on propensity score (Table 3, Figure 4A). When stratifying by XM subgroup, a positive CDC-XM was significantly associated with death or transplant loss (HR 1.90 (1.53, 2.36)), while Flow-XM positive/CDC-XM negative transplants had no significantly higher risk of death/allograft loss compared to XM negative transplant (HR 1.26 (0.90, 1.76)) (Table 3). After adjusting for propensity score, only CDC-XM positive transplants were associated with higher risk of death or allograft loss (HR 1.50 (1.19, 1.88)), while the Flow-XM positive subgroups were not significantly associated with allograft loss (Table 3). Hazard ratios were similar in mixed effects models with random intercepts for transplant center (Supplemental Table S4). Similar results were found in survival analysis using the propensity score matched cohorts (Table 3, Figure 4B).
TABLE 3.
Time from transplant to death or allograft loss by crossmatch result
Hazard ratios for death or allograft loss is presented first as an unadjusted analysis followed by analyses that (1) adjusted for propensity score or (2) matched on propensity score. For each comparison the groups are stratified by crossmatch results.
| Crossmatch Group | Number of Deaths & Allograft Losses/Number of Patients | Unadjusted Hazard Ratio (95% CI) | Adjusted for Propensity Score Hazard Ratio (95% CI) | Matched on Propensity Score Hazard Ratio (95% CI) | |
|---|---|---|---|---|---|
| T-Cell | XM Negative | 1095/4148 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) |
| XM Positive | 214/547 | 1.47 (1.27, 1.70) | 1.29 (1.11, 1.50) | 1.34 (1.09, 1.64) | |
|
| |||||
| XM Negative | 1095/4148 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) | |
| Flow Positive/ CDC Negative | 35/94 | 1.26 (0.90, 1.76) | 1.22 (0.87, 1.70) | 1.16 (0.80, 1.68) | |
| Flow Positive/ CDC Unknown | 92/288 | 1.28 (1.04, 1.59) | 1.17 (0.94, 1.45) | 1.14 (0.88, 1.47) | |
| CDC Positive | 87/165 | 1.90 (1.53, 2.36) | 1.50 (1.19, 1.88) | 1.77 (1.36, 2.31) | |
|
| |||||
| B-Cell | XM Negative | 917/3718 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) |
| XM Positive | 105/309 | 1.19 (0.98, 1.46) | 1.06 (0.86, 1.31) | 1.02 (0.77, 1.34) | |
|
| |||||
| XM Negative | 917/3718 | 1.00 (Ref) | 1.00 (Ref) | 1.00 (Ref) | |
| Flow Positive/ CDC Negative | 27/61 | 1.24 (0.84, 1.83) | 1.01 (0.68, 1.50) | 1.06 (0.67, 1.67) | |
| Flow Positive/ CDC Unknown | 64/200 | 1.24 (0.97, 1.60) | 1.14 (0.88, 1.47) | 1.06 (0.77, 1.46) | |
| CDC Positive | 14/48 | 0.95 (0.56, 1.61) | 0.85 (0.50, 1.40) | 0.82 (0.46, 1.45) | |
FIGURE 4.
Time from transplant to death or allograft loss by crossmatch result
Time to death or allograft loss after transplant for propensity matched cohorts (A) T-Cell XM negative vs. XM positive, (B) T-cell XM subgroups, (C) B-Cell XM negative vs. XM-positive, and (D) B-Cell XM positive subgroups. P values are comparisons to the reference group (Negative crossmatch “XM-“).
XM+: crossmatch positive, XM-: crossmatch negative, CDC+: CDC-crossmatch positive, CDC-: CDC-crossmatch negative, Flow+: Flow-crossmatch positive, CDCnr: CDC no result (CDC-crossmatch unknown)
Given the 20 year span of the study and change in crossmatch technique and immunosuppression regimens over time, we were also interested in the impact of transplant era on results. Year of transplant was included in the propensity models (Supplemental Table S2) and was therefore already indirectly accounted for when examining the relationship between XM status and outcomes. Furthermore, when adjusting for transplant year in addition to propensity score in both models, a positive CDC-XM remained significantly associated with higher risk of both treated rejection in the first post-transplant year and time to allograft loss or death (Supplemental Table S5).
Impact of a Positive B-Cell Crossmatch on Heart Transplant Outcomes
We then sought to understand the impact of an isolated positive B-cell XM on HT outcome. Given that a positive B-cell XM is expected in the setting of a positive T-cell XM, only patients with a negative T-cell Flow-XM were included in the B-cell analyses to more clearly delineate the impact of Class II anti-HLA antibodies detected by the B-cell XM. Of patients with a negative T-cell Flow-XM, 142 patients were excluded due to either no B-cell XM data (n=93) or negative/unknown B-cell CDC XM data and unknown B-cell Flow-XM results (n=49), leaving 4027 patients in the B-cell analysis (Figure 1B). Patients with a positive B-cell CDC- or Flow-XM were compared to those with a negative XM with regards to baseline demographics, clinical status, and donor characteristics (Table 1, Supplemental Table S6). Similar to T-cell results, patients with a positive B-cell XM differed from those with a negative XM with regard to cardiac diagnosis (p < 0.001) and year of transplant (p <0.001). Patients with a positive B-cell XM were also more likely to have moderate or severe renal dysfunction, compared to those with a negative XM (p<0.001). To account for baseline differences in groups, propensity scores were created for the B-cell XM analysis. The multivariable logistic regression model used for B-cell propensity score is shown in Supplemental Table S7 and baseline characteristics across the total cohort and after propensity matching are shown in Supplemental Table S8.
Patients with a positive B-cell XM had 1.4 times higher odds of treated rejection within the first year (OR 1.40 (1.06, 1.85)), and positive CDC-XM patients were 2 times more likely to have treated rejection compared to XM negative patients (OR 2.07 (1.11, 3.87)), similar to T-cell analyses. However, these findings did not remain significant when either adjusting for or matching on propensity score (Figure 3C, 3D, Table 2). There was also no significant association between B-cell XM result and risk of death or graft loss in the unadjusted (HR 1.19 (0.98, 1.46)) or propensity matched cohorts (Table 3, Figure 4C). Subdividing by type of XM and XM result did not show any difference in the risk of allograft loss in unadjusted or propensity matched analysis (Table 3, Figure 4D).
DISCUSSION
Given the high prevalence of anti-HLA sensitization within the pediatric HT population, there is a compelling need to understand the association of a positive donor-specific XM with post-HT outcomes. Anti-HLA antibody detection methods have evolved, leading to increased sensitivity for detection as well as the likelihood of false positive results and results with unclear clinical implications. In this study, we found that a positive T-cell XM was associated with higher odds of rejection in the first year after HT and a shorter time to allograft loss or death. When comparing propensity adjusted outcomes by XM technique, we found that a positive T-cell CDC-XM remained associated with rejection and allograft loss while a positive T-cell Flow-XM or positive B-cell XM (CDC or Flow) result alone was not associated with higher rates of rejection or allograft loss. The use of propensity score adjustment is unique to our study of pediatric HT outcomes and increases confidence in the findings.
The association of a positive T-cell CDC-XM with worse HT outcomes is consistent with multiple other studies that have demonstrated higher rates of rejection and/or allograft loss in children after a XM positive HT.4,6–11 To attenuate the impact of a positive XM, many centers utilize enhanced immunosuppressive regimens with anti-thymocyte globulin (ATG) induction, peri- and post-operative plasmapheresis, IVIG, and tacrolimus/mycophenolate mofetil/prednisone maintenance therapy. These augmented regimens have been associated with improved survival after XM positive HT, though rejection has remained a major obstacle in small single-center studies.6,7,16 Although augmented immunosuppression regimens may attenuate the rejection and mortality risk associated with a positive XM, the overall impact of these regimens on patient morbidity must also be considered. Augmented immunosuppression is associated with higher risks of infection, malignancy, medication side effects, and longer/more frequent hospitalizations.7,17
The reported successes of XM positive HT in multiple single-center studies led to the development of the CTOTC-04 multicenter prospective cohort study. This study compared outcomes in sensitized pediatric HT patients with a positive CDC-XM to CDC-XM negative/non-sensitized patients.17,18 Immunosuppression was standardized with ATG induction and tacrolimus/mycophenolate mofetil maintenance. Sensitized patients with a positive CDC-XM additionally underwent perioperative volume exchange, postoperative plasmapheresis and IVIG, and addition of corticosteroids to maintenance therapy.18 CDC-XM positive patients had inferior freedom from antibody mediated and acute cellular rejection, however, there was no difference in the primary composite endpoint of death, re-transplantation, or rejection with hemodynamic compromise across groups. While this finding may suggest that an augmented immunosuppressive regimen can abrogate the effects of a positive CDC-XM on overall outcome, the study only included 11 CDC-XM positive patients and was unfortunately underpowered to detect a difference amongst groups. Furthermore, given the authors did find higher rates of rejection in the CDC-XM positive group, it will be important to ascertain the impact of a positive CDC-XM and contemporary treatment strategies on longer-term graft/patient survival.
Our study examined outcomes in over 500 T-cell XM positive patients and found that a positive XM was associated with rejection and allograft loss. When teasing apart the impact of the type of XM on outcomes, worse allograft survival and higher rejection rates were only observed in the T-cell CDC-XM positive subgroup. Patients who were T-cell Flow-XM positive and CDC-XM negative/unknown did not have worse allograft survival or higher rejection rates. These data are interesting given studies have shown that pre-transplant anti-HLA antibodies (particularly donor-specific antibodies) are associated with worse short-term outcomes and higher rejection rates even in the presence of a negative CDC-XM.11,19 The authors of these studies suggested that a CDC-XM may lack sensitivity to detect clinically important degrees of anti-HLA sensitization. However, our data suggest that a positive Flow-XM alone does not result in differences in the key clinical outcomes of rejection or allograft loss and are similar to findings in an adult cohort.20 It is possible that the use of antibody-directed treatment strategies, such as plasmapheresis, ATG, rituximab, IVIG, and eculizumab may mitigate the impact of antibodies contributing to a positive Flow-XM. The OPTN database does not collect data on all antibody-directed therapies (including plasmapheresis and IVIG), limiting our ability to describe how treatment strategies may modify outcomes in the presence of a positive XM. However, patients who were Flow-XM positive and CDC negative/unknown were more likely to receive rituximab induction and maintenance steroids compared to the XM negative group suggesting these patients may have received an augmented treatment regimen that abrogated some effect of the positive Flow-XM (Supplemental Table S9).
These data highlight that the shift away from performing a CDC-XM may result in loss of important prognostic information and suggest that additional strategies are needed to risk stratify Flow-XM positive patients. Flow-based assays provide semi-quantitative measurements which may allow for risk stratification, however, given a lack of standardization, each HT program relies on their local HLA lab to develop cutoffs for acceptable risk.21–24 These quantitative data are not captured in national databases, making it difficult to perform sufficiently powered studies to identify universal cutoffs and understand the impact of a positive Flow-XM on post-HT outcomes. In exploratory analyses in CTOTC-04, similar outcomes were observed when comparing transplants across a positive CDC-XM and those from patients with donor-specific antibodies present at a mean fluorescence intensity >8,000, suggesting that antibody strength captured by these assays may be relevant for risk stratificaton.17 Additionally, while Flow-based assays detect antibody binding, the ability of these antibodies to mediate complement fixation is not tested by this method. Addition of C1q assays and HLA epitope mapping may provide additive biologic information to risk stratify patients with positive Flow-XM results.25–29
Although there are many advantages to using data from national registries such as OPTN, this study was a retrospective analysis with multiple limitations. These limitations include that XM data were reported from different labs using variable techniques and that further granularity of the XM results were unavailable (i.e., unable to differentiate between weakly vs. strongly positive Flow-XM and labs may have different cutoffs for what was considered a positive XM). Similarly, a large proportion of pediatric transplant recipients were excluded from these analyses (n = 3503) with the majority of patients excluded due to absent/incomplete T-cell XM data. These patients were excluded in order to more clearly delineate the impact of a positive CDC vs. Flow-XM results on transplant outcomes, however, we do not know how these patients with incomplete XM data may have impacted outcomes. The OPTN database does not capture information about pre-HT desensitization and the impact of immunosuppression/treatment on outcomes cannot be ascertained with this study design. There is a selection bias in that center-based practices around HLA compatibility are certain to have influenced which transplants occurred. Similarly, there may be risk of classification bias concerning the outcome of “treated rejection” as centers may define “treated rejection” differently. We have, however, accounted for center-specific differences, in part, by including a mixed effects models with random intercept for institution (Supplemental Table S4).
In conclusion, our data demonstrate that a positive T-cell CDC-XM is associated with a higher risk of rejection and allograft loss/death after pediatric HT. However, a positive T-cell Flow-XM alone or isolated positive B-cell CDC- or Flow-XM did not result in higher risk of early rejection or allograft loss. These data highlight differences in methods used to detect donor-specific anti-HLA antibodies and emphasize the need to better understand how a positive Flow-XM affects post-HT outcomes. In our opinion, a positive Flow-XM alone should not preclude acceptance of a donor organ. Further studies are necessary to delineate how the breadth/strength of pre-transplant anti-HLA antibodies detected by multiplex bead assays and the strength of a positive Flow-XM can be utilized to make more informed donor acceptance decisions. Unless replaced with validated Flow-XM cutoffs, the shift away from CDC-XM testing will result in a loss of important prognostic information. Finally, the ability of current approaches to pre- and post-HT desensitization and immunosuppression to mitigate post-HT risks requires further study.
Supplementary Material
ACKNOWLEDGMENTS
CM was supported in part by the NIH training grant T32-HL007572. CM& RW were supported by the Furber ACT Fellowship Fund & Cardiac Transplant Research and Education Fund. This work was supported in part by HRSA contract 234-2005-370011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
Abbreviations:
- ATG
anti-thymocyte globulin
- CDC
complement-dependent cytotoxicity
- ECMO
Extracorporeal Membrane Oxygenation
- HLA
human leukocyte antigen
- HR
Hazard ratio
- HT
heart transplant
- OPTN
Organ Procurement and Transplantation Network
- OR
Odds Ratio
- VAD
Ventricular Assist Device
- XM
Crossmatch
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.
REFERENCES
- 1.Kobayashi D, Du W, L’Ecuyer TJ. Predictors of cardiac allograft vasculopathy in pediatric heart transplant recipients. Pediatr Transplant. 2013;17(5):436–440. doi: 10.1111/petr.12095 [DOI] [PubMed] [Google Scholar]
- 2.Feingold B, Bowman P, Zeevi A, et al. Survival in Allosensitized Children After Listing for Cardiac Transplantation. Journal of Heart and Lung Transplantation. 2007;26(6):565–571. doi: 10.1016/j.healun.2007.03.015 [DOI] [PubMed] [Google Scholar]
- 3.Frandsen EL, Albers EL. Management of the sensitized pediatric heart transplant patient. Transl Pediatr. 2019;8(4):302–313. doi: 10.21037/tp.2019.07.13 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mahle WT, Tresler MA, Edens RE, et al. Allosensitization and outcomes in pediatric heart transplantation. Journal of Heart and Lung Transplantation. 2011;30(11):1221–1227. doi: 10.1016/j.healun.2011.06.005 [DOI] [PubMed] [Google Scholar]
- 5.Feingold B, Webber SA, Bryce CL, et al. Comparison of listing strategies for allosensitized heart transplant candidates requiring transplant at high urgency: A decision model analysis. American Journal of Transplantation. 2015;15(2):427–435. doi: 10.1111/ajt.13071 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Holt DB, Lublin DM, Phelan DL, et al. Mortality and Morbidity in Pre-sensitized Pediatric Heart Transplant Recipients With a Positive Donor Crossmatch Utilizing Peri-operative Plasmapheresis and Cytolytic Therapy. Journal of Heart and Lung Transplantation. 2007;26(9):876–882. doi: 10.1016/j.healun.2007.07.011 [DOI] [PubMed] [Google Scholar]
- 7.Daly KP, Chandler SF, Almond CS, et al. Antibody depletion for the treatment of crossmatch-positive pediatric heart transplant recipients. Pediatr Transplant. 2013;17(7):661–669. doi: 10.1111/petr.12131 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Thrush PT, Pahl E, Naftel DC, et al. A multi-institutional evaluation of antibody-mediated rejection utilizing the Pediatric Heart Transplant Study database: Incidence, therapies and outcomes. Journal of Heart and Lung Transplantation. 2016;35(12):1497–1504. doi: 10.1016/j.healun.2016.06.014 [DOI] [PubMed] [Google Scholar]
- 9.Rossano JW, Morales DLS, Zafar F, et al. Impact of antibodies against human leukocyte antigens on long-term outcome in pediatric heart transplant patients: An analysis of the United Network for Organ Sharing database. Journal of Thoracic and Cardiovascular Surgery. 2010;140(3):694–699.e2. doi: 10.1016/j.jtcvs.2010.04.009 [DOI] [PubMed] [Google Scholar]
- 10.Wright EJ, Fiser WP, Edens RE, et al. Cardiac Transplant Outcomes in Pediatric Patients with Pre-formed Anti-Human Leukocyte Antigen Antibodies and/or Positive Retrospective Crossmatch. Journal of Heart and Lung Transplantation. 2007;26(11):1163–1169. doi: 10.1016/j.healun.2007.07.042 [DOI] [PubMed] [Google Scholar]
- 11.Lavee J, Kormos RL, Duquesnoy RJ, et al. Influence of panel-reactive antibody and lymphocytotoxic crossmatch on survival after heart transplantation. Journal of Heart and Lung Transplantation. 1991;10(6):921–929. [PubMed] [Google Scholar]
- 12.Tait BD, Susal C, Gebel HM, et al. Consensus Guidelines on the Testing and Clinical Management Issues Associated With HLA and Non-HLA Antibodies in Transplantation. Transplantation. 2013;95(1):19–47. [DOI] [PubMed] [Google Scholar]
- 13.Kobashigawa J, Colvin M, Potena L, et al. The management of antibodies in heart transplantation: An ISHLT consensus document. Journal of Heart and Lung Transplantation. 2018;37(5):537–547. doi: 10.1016/j.healun.2018.01.1291 [DOI] [PubMed] [Google Scholar]
- 14.Vaidya S, Partlow D, Susskind B, Noor M, Barnes T, Gugliuzza K. Prediction of crossmatch outcome of highly sensitized patients by single and/or multiple antigen bead luminex assay. Transplantation. 2006;82(11):1524–1528. doi: 10.1097/01.tp.0000246311.43634.0a [DOI] [PubMed] [Google Scholar]
- 15.D’Agostino RB. Propensity scores in cardiovascular research. Circulation. 2007;115(17):2340–2343. doi: 10.1161/CIRCULATIONAHA.105.594952 [DOI] [PubMed] [Google Scholar]
- 16.Pollock-BarZiv SM, Den Hollander N, Ngan BY, Kantor P, McCrindle B, Dipchand AI. Pediatric heart transplantation in human leukocyte antigen-sensitized patients: Evolving management and assessment of intermediate-term outcomes in a high-risk population. Circulation. 2007;116(11 SUPPL. 1):172–178. doi: 10.1161/CIRCULATIONAHA.107.709022 [DOI] [PubMed] [Google Scholar]
- 17.Webber S, Zeevi A, Mason K, et al. Pediatric heart transplantation across a positive crossmatch: First year results from the CTOTC-04 multi-institutional study. American Journal of Transplantation. 2018;18(9):2148–2162. doi: 10.1111/ajt.14876 [DOI] [PubMed] [Google Scholar]
- 18.Zuckerman WA, Zeevi A, Mason KL, et al. Study rationale, design, and pretransplantation alloantibody status: A first report of Clinical Trials in Organ Transplantation in Children-04 (CTOTC-04) in pediatric heart transplantation. American Journal of Transplantation. 2018;18(9):2135–2147. doi: 10.1111/ajt.14695 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kobashigawa JA, Sabad A, Drinkwater D, et al. Pretransplant panel reactive-antibody screens. Are they truly a marker for poor outcome after cardiac transplantation? Circulation. 1996;94(9 Suppl):II294–7. [PubMed] [Google Scholar]
- 20.Keeshan BC, O’Connor MJ, Lin KY, et al. Value of a flow cytometry cross-match in the setting of a negative complement-dependent cytotoxicity cross-match in heart transplant recipients. Clin Transplant. 2017;31(10). doi: 10.1111/ctr.13064 [DOI] [PubMed] [Google Scholar]
- 21.Batal I, Zeevi A, Lunz JG, et al. Antihuman leukocyte antigen-specific antibody strength determined by complement-dependent or solid-phase assays can predict positive donor-specific crossmatches. Arch Pathol Lab Med. 2010;134(10):1534–1540. doi: 10.5858/2009-0581-oa.1 [DOI] [PubMed] [Google Scholar]
- 22.Peräsaari JP, Jaatinen T, Merenmies J. Donor-specific HLA antibodies in predicting crossmatch outcome: Comparison of three different laboratory techniques. Transpl Immunol. 2018;46(September 2017):23–28. doi: 10.1016/j.trim.2017.11.002 [DOI] [PubMed] [Google Scholar]
- 23.Tambur AR, Herrera ND, Haarberg KMK, et al. Assessing Antibody Strength: Comparison of MFI, C1q, and Titer Information. American Journal of Transplantation. 2015;15(9):2421–2430. doi: 10.1111/ajt.13295 [DOI] [PubMed] [Google Scholar]
- 24.Morris GP, Phelan DL, Jendrisak MD, Mohanakumar T. Virtual crossmatch by identification of donor-specific anti-human leukocyte antigen antibodies by solid-phase immunoassay: A 30-month analysis in living donor kidney transplantation. Hum Immunol. 2010;71(3):268–273. doi: 10.1016/j.humimm.2010.01.003 [DOI] [PubMed] [Google Scholar]
- 25.Albers EL, Friedland-Little JM, Hong BJ, Kemna MS, Warner P, Law YM. Human leukocyte antigen eplet mismatching is associated with increased risk of graft loss and rejection after pediatric heart transplant. Pediatr Transplant. 2022;26(1):1–10. doi: 10.1111/petr.14126 [DOI] [PubMed] [Google Scholar]
- 26.Sullivan PM, Warner P, Kemna MS, et al. HLA molecular epitope mismatching and long-term graft loss in pediatric heart transplant recipients. Journal of Heart and Lung Transplantation. 2015;34(7):950–957. doi: 10.1016/j.healun.2014.12.017 [DOI] [PubMed] [Google Scholar]
- 27.Chin C, Chen G, Sequeria F, et al. Clinical usefulness of a novel C1q assay to detect immunoglobulin G antibodies capable of fixing complement in sensitized pediatric heart transplant patients. Journal of Heart and Lung Transplantation. 2011;30(2):158–163. doi: 10.1016/j.healun.2010.08.020 [DOI] [PubMed] [Google Scholar]
- 28.Zeevi A, Lunz J, Feingold B, et al. Persistent strong anti-HLA antibody at high titer is complement binding and associated with increased risk of antibody-mediated rejection in heart transplant recipients. The Journal of Heart and Lung Transplantation. 2013;32(1):98–105. doi: 10.1016/j.healun.2012.09.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ellison M, Mangiola M, Marrari M, et al. Immunologic risk stratification of pediatric heart transplant patients by combining HLA-EMMA and PIRCHE-II. Front Immunol. 2023;14. doi: 10.3389/fimmu.2023.1110292 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.