Abstract
Background
Body weight is the traditional metric for matching donor and recipient size for pediatric heart transplantation (pHT). We hypothesized that mismatch in body mass index (BMI) or body surface area (BSA) rather than weight is better associated with outcomes of transplantation and therefore should be used for donor–recipient size matching.
Methods
Analysis of the United Network for Organ Sharing database limited to pHT recipients was performed. Donor and recipient mismatch groups were created for weight, BMI, and BSA ratios. Differences in recipient characteristics between each cohort and the impact of mismatch on outcomes were statistically analyzed.
Results
A total of 4,465 patients were included in the analysis of which 43% had congenital heart disease (CHD). There were significant differences in patient characteristics by matching, independent of the matching parameter. Multivariable regression analysis showed that a low donor–recipient BMI ratio (compared to normal) (CHD OR 1.70; non-CHD 2.78) was a predictor of one-year mortality (all P < .001) in both CHD and non-CHD cohorts. Low BMI ratio was also associated with worse long-term survival in non-CHD groups, but not in the CHD cohort. Weight and BSA ratio did not predict one year or long-term survival.
Conclusion
The use of low BMI donors compared to recipient may predict poor early and long-term survival and therefore should be avoided in pHT. The use of BMI matching may improve donor–recipient matching in pHT.
Introduction
Pediatric heart transplantation (pHT) remains an important treatment option for children with end-stage heart disease. Unfortunately, the annual demand for organs greatly exceeds the numbers donated, and the number of patients needing a transplant continues to grow. 1 In this setting, there is increasing focus on optimal utilization of donated organs as there is a significant rate of non-utilization of the donated organs. 2 Parameters for donor–recipient matching, therefore, continue to be important.
Body weight has been the traditional metric for matching donor and recipient size as suggested by the International Society for Heart and Lung Transplantation guidelines. 3 The guidelines only outline recommendations using donor weight relative to recipient weight, which is based on limited data. The guidelines also mention significant oversizing and undersizing of donors based on weight as potential reasons for primary graft dysfunction. 3 The data on the impact of weight mismatch on transplant outcomes are mixed with some studies suggesting lower donor weight as a risk factor for poor outcomes, while others show no relationship. 4 This potential inadequacy of data is reflected in center practices for listing criteria and acceptance of a donor. Most of the accepted donors were matched conservatively while only the sickest of the recipients used somewhat liberal matching by weight. 5
Certainly, other parameters for appropriate matching have been explored. In the study by Szugye et al, the model that included height, weight, gender, and age had a R2 of 0.97, while addition of a chest x-ray to the model increased the agreement to R2 of 0.99. The difference in performance is small, and overall, not clear that a model using total cardiac volume is superior to the model using BMI as we propose here.6,7
Height mismatch between donor and recipients has demonstrated a weak association with short-term mortality in underweight and obese patients but is not associated with long-term mortality. 8 Finally, there has been some investigation into body mass index (BMI) mismatch between donor and recipient heart transplant candidates in adults. Results have shown that inappropriate weight-matched donors versus appropriate weight-matched donors (using BMI) is associated with increased 30-day, 1-year, and 5-year mortality in nonobese patients. 8 However, research is lacking on how BMI or BSA in donor–recipient size matching could affect posttransplant survival in pediatric patients. Importantly, research has shown that cardiac dimensions in children have a linear correlation with height while a non-linear relationship with weight and BSA. 9
Therefore, the goal of the current study was to assess the various donor recipient parameters and impact on outcomes using a contemporary cohort. Specifically, we hypothesize that mismatch in compound measures such as BMI or BSA rather than weight is better associated with outcomes in pHT and may serve to improve decisions around donor matching.
Methods
Data from the United Network for Organ Sharing (UNOS) were requested from the Organ Procurement and Transplantation Network (OPTN) in the form of Standard Transplant Analysis and Research (STAR) files. The study was exempt from the institutional review due to the de-identified nature of the dataset. Information was limited to pHT recipients with transplantations performed between 2007 and June 2021. Criteria for inclusion were patients who were <18 years old who underwent heart transplantation secondary to congenital heart disease (CHD) or cardiomyopathy (non-CHD). We calculated donor–recipient ratios for various parameters: weight, BMI, and BSA. Based on the distribution of ratios, we created a low ratio (25th percentile or lower), normal ratio (25th-75th percentile), and high (greater than 75th). This gave us the following categories and distribution:
Weight: Low ratio <1.04, normal ratio 1.04-1.56, high ratio >1.56 BMI: Low ratio <0.92, normal ratio 0.92-1.24, high ratio >1.24 BSA: Low ratio <1.02, normal ratio 1.02-1.37, high ratio >1.37
Different recipient characteristics were studied within each study group using univariate nonparametric tests (Kruskal-Wallis for continuous variables and χ2 for categorical variables). The characteristics chosen are variables that have been known to affect mortality in pHT. These include recipient age, gender, race, creatinine level, total bilirubin level, diabetes, pulmonary artery pressure mean (PAP mean), mechanical ventilation, ECMO use, BMI, weight, and BSA. Also included were donor age, BMI, weight, BSA, and ischemic time (Tables 1-6). Normal kidney and liver function were determined as creatinine and total bilirubin levels <1.0 mg/dL as these are normal, average levels for kids between 0 and 18 years old. Quantitative variables were presented as a median with interquartile ranges, while categorical variables were presented as a percentage.
Donor–Recipient Ratio of BMI in CHD.
Abbreviations: BMI, body mass index; BSA, body surface area; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; PAP, pulmonary artery pressure.
Bold values are statistically significant (p = <0.05).
Donor–Recipient Ratio of BSA in Non-CHD.
Abbreviations: BMI, body mass index; BSA, body surface area; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; PAP, pulmonary artery pressure.
Bold values are statistically significant (p = <0.05).
Considering collinearities between the weight, BSA, and BMI ratios, we performed individual logistic regression models for each weight, BSA, and BMI stratified by the etiology of CHD versus non-CHD. This provided us with 6 logistic regression models (Figures 1 and 2, Supplemental Figures 1-4). After initial stratification of CHD versus non-CHD, the variables used in each analysis were the same and included age, gender, race, creatinine, ECMO, mechanical ventilation, total bilirubin, donor age, and corresponding ratios of weight, BSA, and BMI. The regression model was a binary logit model (1-year death yes vs no) with no selection. The significance level was set at .05. Last, survival analysis was performed by Kaplan-Meier curves, which were again stratified into CHD and non-CHD cohorts (Figures 3 and 4). All analysis was done using SAS 9.4 software (SAS Inc).
Multivariate logistic regression analysis for CHD by BMI.
Multivariate logistic regression analysis for non-CHD by BMI.
Kaplan-Meier survival curve, showing long-term survival outcomes according to donor–recipient BMI ratio in non-CHD cohort (P ≤ .01).
Kaplan-Meier survival curve, showing long-term survival outcomes according to donor–recipient BMI ratio in CHD cohort (P = .15).
Results
A total of 4,465 pHT patients were included in the study in which 47% had congenital heart disease and 53% had cardiomyopathy. The distribution of patient characteristics and outcomes is further detailed below by each parameter for matching: weight, BMI, and BSA.
Body Mass Index
Patient characteristics for low (<0.92), normal (0.92-1.24), and high (>1.24) ratios of BMI in CHD and non-CHD cohorts are compared using univariate analysis (Tables 1 and 2). Recipients with a low D-R BMI ratio in the CHD cohort are more likely to be the youngest (2, IQR 0-9), have normal kidney function (0.4, IQR 0.3-0.6), be on ECMO (3%), and to have increased one-year mortality (19%) (P < .05). Comparatively, recipients with a low D-R BMI ratio in non-CHD cohort are less likely to be Caucasian (39% Caucasian), more likely to have normal to abnormal liver function (0.5, IQR 0.3-1.0), be on ECMO (5%), and have increased one-year mortality (8%) (P ≤ .05).
Donor–Recipient Ratio of BMI in Non-CHD.
Abbreviations: BMI, body mass index; BSA, body surface area; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; PAP, pulmonary artery pressure.
Bold values are statistically significant (p = <0.05).
Weight
Patient characteristics for low (<1.04), normal (1.04-1.56), and high (>1.56) donor–recipient ratio of weight in CHD and non-CHD cohorts are summarized in Tables 3 and 4. Recipients with high D-R weight ratios in the CHD cohort are more likely to be the youngest (2, IQR 0-9), have normal kidney function (0.4, IQR 0.3-0.6), and require ECMO (6%) (P < .05). Similarly, in the non-CHD cohort, recipients with a high D-R weight ratio are more likely to be younger (2, IQR 0-9), Caucasian (49%), and have normal kidney function (0.3, IQR 0.2-0.5) (P < .02).
Donor–Recipient Ratio of Weight in CHD.
Abbreviations: BMI, body mass index; BSA, body surface area; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; PAP, pulmonary artery pressure.
Bold values are statistically significant (p = <0.05).
Donor–Recipient Ratio of Weight in Non-CHD.
Abbreviations: BMI, body mass index; BSA, body surface area; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; PAP, pulmonary artery pressure.
Bold values are statistically significant (p = <0.05).
Body Surface Area
Patient characteristics for low (<1.02), normal (1.02-1.37), and high (>1.37) donor–recipient BSA ratio in the CHD and non-CHD cohorts are summarized in Tables 5 and 6. Recipients with high D-R BSA ratios were more likely to be the youngest (1, IQR 0-6), have normal kidney function (0.4, IQR 0.3-0.5), have normal to abnormal liver function (0.5, IQR 0.3-1.1), and require ECMO (7%) (P < .05). All study groups in the CHD cohort were more likely to be male (59%) and Caucasian. The recipients with high donor–recipient BSA ratios in the non-CHD cohort are more likely to be the youngest (1, IQR 0-7), less likely to be male (47%) and Caucasian (49%), have normal kidney function (0.3, IQR 0.2-0.5), and have normal liver function (0.5, IQR 0.3-0.9).
Donor–Recipient Ratio of BSA in CHD.
Abbreviations: BMI, body mass index; BSA, body surface area; CHD, congenital heart disease; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; PAP, pulmonary artery pressure.
Bold values are statistically significant (p = <0.05).
Correlation Matrix and Risk Factors for 1-Year Mortality
Some variables in our study may be thought to be collinear variables, such as BMI, BSA, and weight ratios. However, in our Supplemental Table 1, we have created a correlation matrix assessment. Based on the correlation matrix, we identified that between “weight ratio high” and “BSA ratio high” there exists a strong negative collinearity (correlation coefficient of −0.82). For the BMI ratio, the collinearity coefficient ranged from −0.467 to 0.39 which does show the existence of some collinearity. For this reason, we performed the following logistic regression models for each BMI, BSA, and weight stratified by CHD versus non-CHD and adjusting for the same risk factors in each analysis. Each multivariate logistic regression model for the CHD cohort demonstrated that the presence of kidney dysfunction (weight: OR 1.14, CI 1.01-1.28; BMI: OR 1.14, CI 1.01-1.29; BSA: OR 1.14, CI 1.01-1.29), use of ECMO (weight: OR 3.57, CI 2.33-5.46; BMI: OR 3.81, CI 2.48-5.85; BSA: OR 3.58, CI 2.33-5.46), use of a mechanical ventilator (weight: OR1.43, CI 1.04-1.97; BMI: OR 1.42, CI 1.03-1.95; BSA: OR 1.42, CI 1.03-1.95), and presence of liver dysfunction (weight: OR 1.09, CI 1.04-1.13; BMI: OR 1.08, CI 1.04-1.13; BSA: OR 1.09, CI 1.04-1.13) were predictors of one-year mortality (P ≤ .05) (Supplemental Figures 1 and 2). Furthermore, the BMI logistic regression model noted that a low D-R BMI ratio was also a predictor of 1-year mortality (OR 1.56, CI 1.16-2.09) (Figure 1). In the non-CHD cohort, each multivariate logistic regression model demonstrated that the use of ECMO (weight: OR 2.30, CI 1.05-5.04; BMI: OR 2.31, CI 1.06-5.03; BSA: OR 2.31, CI 1.06-5.05), use of a mechanical ventilator (weight: OR 2.07, CI 1.17-3.67; BMI: OR 2.04, CI 1.15-3.61; BSA: OR 2.06, CI 1.16-3.64), and liver dysfunction (weight: OR 1.10, CI 1.04-1.17; BMI: OR 1.10, CI 1.04-1.17; BSA: OR 1.10, CI 1.04-1.17) were predictors of 1-year mortality (P ≤ .05) (Supplemental Figures 3 and 4). Specifically, the BMI logistic regression model noted that a low D-R BMI ratio was also a predictor of 1-year mortality (OR 2.26, CI 1.33-3.84) (Figure 2). All other variables in all 6 analyses showed no statistical significance.
Long-Term Survival
When assessing long-term survival rates, the Kaplan-Meier curves demonstrate that a low donor–recipient BMI ratio was associated with worse long-term survival in non-CHD groups (P < .01), but not in the CHD cohort (P = .15). The CHD group associated with a low donor–recipient BMI ratio does have decreased survival for the first 8 years posttransplantation, but then develops the same survival rate as groups with normal and high BMI ratios (Figures 3 and 4). Using survival analysis, unlike BMI, weight, and BSA were nonpredictive of mortality.
Discussion
In this study, we have comprehensively evaluated the association between mortality and donor–recipient mismatch using weight, BMI, and BSA in pHT patients. We found that there is more frequent mismatch in recipients with a higher weight, BMI, and BSA donor–recipient ratio. Secondly, we found that low D-R BMI ratio (eg, recipients with a normal BMI with a low donor BMI) were likely to have higher one year and long-term mortality in the non-CHD cohort along with increased 1-year mortality in the CHD cohort. Other measures such as weight and BSA were not predictors of outcomes.
The current guidelines for donor–recipient matching are based on weight.2,3 Our findings suggest that incorporation of a compound measure such as BMI for matching may help further optimize outcomes. This statement is further supported by data collected in adult studies as there have been no pediatric studies evaluating donor–recipient BMI size mismatch and outcomes in survival. For example, Bergenfeldt et al 8 found that inappropriate weight-matched donors (defined as donor weight <70% of recipient weight and measured as BMI) as opposed to appropriate weight-matched donors were associated with increased 30-day, 1-year, and 5-year mortality in nonobese recipients. Interestingly, they found no significant association between inappropriate weight-matched donors and mortality in obese recipients. Our study did not separate patients into obese or non-obese recipients.
Furthermore, Barac et al 10 found that using a low or normal donor–recipient BMI ratio in a female-to-male heart transplant correlated with an increased mortality. Their study then evaluated if using an increased donor–recipient BMI ratio in a female-to-male heart transplant would decrease the mortality risk. They found that a BMI difference between donor and recipient that is 1.5 kg/m2 or higher can eliminate the risk of transplanting female hearts into male recipients. One theory is that female hearts are smaller relative to male hearts and may not have the functional capacity to support a male recipient. This further supports their finding that once the female donor BMI is higher than the male recipient BMI by 1.5 kg/m2 the survival matches that of male-to-male transplants. Similarly, in our study, a low donor–recipient BMI ratio was associated with decreased short term and overall survival. Reed et al found similar findings in their study in that male recipients of female donors had decreased overall survival but attributed this to differences in predicted heart mass. As this was completed in the adult population, it would be intriguing to see whether predicted heart mass had a similar impact on pHT. Furthermore, as shown in these previous 2 studies, gender mismatch can significantly impact survival posttransplantation in the adult population. Our study did not specifically evaluate the impact of gender mismatch in the pediatric population. 11
Many studies have evaluated the effect of donor–recipient weight mismatch in pHT and there appears to be conflicting data. Kanani et al investigated the effect of donor–recipient weight mismatch on mortality and found that survival from operation to final follow-up was lowest in the group with the smallest weight mismatch. 1 However, Riggs et al looked at how donor body weight parameters are currently utilized in cardiac transplantation and its influence on waitlist outcomes, such as posttransplant survival. They found that variable donor–recipient weight ratios had similar posttransplant survival. 2 This is similar to our findings that weight, irrespective of the ratio or patient category (congenital, noncongenital) was not predictive of outcome. Thus, although it serves as a reasonable starting point for donor selection, it can be further refined by the use of BMI which also accounts for weight to height distribution.
Our results show that the CHD cohort with a low donor–recipient BMI ratio did show increased mortality for the first 8 years posttransplant, but then developed the same survival rate as groups with normal and high BMI ratios. This finding is most likely due to the overall inferior outcomes of patients with CHD. It is plausible that BMI, which corrects the weight for height is a better reflection of the CHD cohort, which in general can have lower linear growth for weight. 12 Thus, a donor matched on weight to a CHD recipient may in fact be taller and have a lower BMI. This may represent a true mismatch of consequences on the performance of the graft. Studies have shown the variable impact of weight on pHT outcomes and found that weight and BSA did not predict 1-year or long-term survival. Thus, consideration should be given to BMI during donor selection to optimize matching and potentially improving outcomes.
Limitations
Accessibility to the UNOS registry made this retrospective study possible, which provided a limitation to our study as our study is limited by the variables associated with the UNOS database. Some of the variables that may impact outcomes, especially 1-year mortality, such as organ preservation, postoperative course details, medication compliance, and details around rejection are not fully accounted for. Nonetheless, the UNOS data have been previously used to guide our understanding of risk factors in pHT. Using this dataset does create a bias as these are already transplanted patients and we assume, they were matched on weight. One of the most important limitations of this retrospective study is the current bias in donor selection, and the possibility that the practice is conservative. Therefore, there is a possibility of either underestimation of the effect or of a type II error if the boundaries of matching were pushed further on either end of the current matching profiles.
Another potential limitation is that studies that have looked at the impact of various parameters such as weight or BMI mismatch, often use different cutoff values for low, normal, and high ratios. These cutoffs are often decided a priori on minimal clinical or statistical justification. We therefore chose to let the actual distribution of the data (ratios) decide the cutoff values using 25th percentile, 25th to 75th percentile, and >75th percentile as a cutoff. Ultimately, the ratio between donor BMI and recipient BMI is easy to calculate and see whether it fits with the low, normal, or high ratio categories to apply our findings.
Conclusion
In conclusion, our results indicate that donor–recipient mismatch based on BMI rather than weight alone may be a better predictor of outcomes in pHT. This is especially true for a low donor–recipient BMI ratio which was associated with an increased short-term mortality in CHD and non-CHD patients, and with an increased long-term mortality in the non-CHD patients. The current ISHLT guidelines for donor selection based on weight could potentially be strengthened further with the use of BMI to ensure adequate matching and potentially improve outcomes.
Supplemental Material
sj-docx-1-pch-10.1177_21501351221127284 - Supplemental material for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation
Supplemental material, sj-docx-1-pch-10.1177_21501351221127284 for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation by Laura K Lowrey, Jaimin Trivedi, Karthik Ramakrishnan, Pranava Sinha and Shriprasad R Deshpande in World Journal for Pediatric and Congenital Heart Surgery
Supplemental Material
sj-docx-2-pch-10.1177_21501351221127284 - Supplemental material for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation
Supplemental material, sj-docx-2-pch-10.1177_21501351221127284 for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation by Laura K Lowrey, Jaimin Trivedi, Karthik Ramakrishnan, Pranava Sinha and Shriprasad R Deshpande in World Journal for Pediatric and Congenital Heart Surgery
Supplemental Material
sj-docx-3-pch-10.1177_21501351221127284 - Supplemental material for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation
Supplemental material, sj-docx-3-pch-10.1177_21501351221127284 for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation by Laura K Lowrey, Jaimin Trivedi, Karthik Ramakrishnan, Pranava Sinha and Shriprasad R Deshpande in World Journal for Pediatric and Congenital Heart Surgery
Supplemental Material
sj-docx-4-pch-10.1177_21501351221127284 - Supplemental material for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation
Supplemental material, sj-docx-4-pch-10.1177_21501351221127284 for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation by Laura K Lowrey, Jaimin Trivedi, Karthik Ramakrishnan, Pranava Sinha and Shriprasad R Deshpande in World Journal for Pediatric and Congenital Heart Surgery
Supplemental Material
sj-docx-5-pch-10.1177_21501351221127284 - Supplemental material for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation
Supplemental material, sj-docx-5-pch-10.1177_21501351221127284 for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation by Laura K Lowrey, Jaimin Trivedi, Karthik Ramakrishnan, Pranava Sinha and Shriprasad R Deshpande in World Journal for Pediatric and Congenital Heart Surgery
Supplemental Material
sj-docx-6-pch-10.1177_21501351221127284 - Supplemental material for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation
Supplemental material, sj-docx-6-pch-10.1177_21501351221127284 for Influence of Body Mass Index in Donor–Recipient Size Mismatch in Pediatric Heart Transplantation by Laura K Lowrey, Jaimin Trivedi, Karthik Ramakrishnan, Pranava Sinha and Shriprasad R Deshpande in World Journal for Pediatric and Congenital Heart Surgery
Footnotes
Acknowledgments
This work was supported in part by Health Resources and Services Administration contract 2342005-370011C. The content is the responsibility of the authors 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.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
