Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jun 1.
Published in final edited form as: Pediatr Transplant. 2020 Mar 25;24(4):e13702. doi: 10.1111/petr.13702

DETERMINANTS OF LENGTH OF STAY AFTER PEDIATRIC LIVER TRANSPLANTATION

Karina Covarrubias 1, Xun Luo 1, Allan Massie 1,2, Kathleen B Schwarz 3, Jacqueline Garonzik-Wang 1, Dorry L Segev 1,2, Douglas B Mogul 3
PMCID: PMC7260078  NIHMSID: NIHMS1576822  PMID: 32212292

Abstract

Background

We sought to identify factors that are associated with length of stay (LOS) following pediatric (<18 years) liver transplantation in order to provide personalized counseling and discharge planning for recipients and their families.

Methods

We identified 2,726 infants (≤24 months) and 3,210 children (>24 months) who underwent pediatric liver-only transplantation from 2002–2017 using the Scientific Registry of Transplant Recipients. We used multilevel multivariable negative binomial regression to analyze associations between LOS and recipient and donor characteristics and calculated the median LOS ratio (MLOSR) to quantify heterogeneity in LOS across centers.

Results

In infants, the median LOS (IQR) was 19 (13–32) days. Hospitalization prior to transplant (ICU ratio:1.461.591.70; non-ICU ratio:1.081.161.23), public insurance (ratio:1.031.091.15), and a segmental graft (ratio:1.081.151.22), were associated with a longer LOS; thus, we would expect a 1.59-fold longer LOS in an infant admitted to the ICU compared to a non-hospitalized infant with similar characteristics. In children, the median LOS (IQR) was 13 (9–21) days. Hospitalization prior to transplant (ICU ratio:1.491.621.77; non-ICU ratio:1.341.441.56) public insurance (ratio:1.021.071.13), a segmental graft (ratio:1.201.271.35), a living donor graft (ratio:1.271.381.51), and obesity (ratio:1.031.101.17) were associated with a longer LOS. The MLOSR was 1.25 in infants and 1.26 in children, meaning if an infant received a transplant at another center with a longer LOS, we would expect a 1.25-fold difference in LOS driven by center practices alone.

Conclusions

While center-level practices account for substantial variation in LOS, consideration of donor and recipient factors can help clinicians provide more personalized counseling for families of pediatric liver transplant candidates.

Keywords: pediatric, liver transplantation, length of stay, PELD/MELD

INTRODUCTION

Approximately 600 children with end-stage liver disease (ESLD) will undergo liver transplantation annually in the United States, with priority given to the sickest candidates as determined by the Pediatric End-Stage Liver Disease/Model for End-Stage Liver Disease (PELD/MELD) score 1. Understanding hospital length of stay (LOS) is necessary to provide more personalized counseling and discharge planning for pediatric liver transplant recipients and their families 2. From a more practical perspective, understanding LOS is important for families when considering how duration of hospitalization will affect other aspects of their lives, such as arranging childcare for siblings, requesting time off work, and coordinating travel between the transplant center and the patient’s home 3. Most importantly, a prolonged hospitalization places pediatric liver transplant recipient at risk for developing hospital acquired infections. Finally, from a financial perspective, a longer LOS incurs more costs; a longer LOS in adult liver transplant recipients has been tied to increasing costs and resource utilization 4.

A previous multi-center study from the pre-PELD/MELD era of pediatric liver transplantation found that recipients age less than 1 year at transplant, fulminant liver failure, public insurance, segmental grafts from a deceased donor, and a transplant before 1999 were associated with an increased length of stay 5. However, since the prior analysis, changes in allocation have improved waitlist mortality across all ages and more objectively prioritized the sickest candidates, which has led to an increased resource utilization for liver transplantation 6,7. Furthermore, immunosuppression and surgical techniques have markedly improved since the prior analysis, particularly with regards to the increased use and improved outcomes of living donor grafts and segmental grafts 7–9.

In order to provide more personalized patient counseling, the goals of this study were to (i) identify recipient and surgical characteristics that influence LOS following pediatric liver transplantation in the modern era and (ii) to quantify the variation in LOS at the center level.

METHODS

Data Source

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 US, submitted by the members of the Organ Procurement and Transplantation Network (OPTN) and has been described elsewhere 10. The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services, provides oversight to the activities of the OPTN and SRTR contractors. The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of, or interpretation by, the SRTR or the U.S. Government.

Study Population

We studied 6,621 pediatric (<18 years old) liver-only transplant recipients less than 18 years old from March 2002-December 2017 using the SRTR. We excluded recipients with missing height or weight (n = 368), PELD/MELD score (n = 12), and recipients with biologically implausible values in height- or length-for-age, weight-for-length, weight-for-age, or body mass index-for-age (BMI-for-age) (n = 403) as defined by the Centers for Disease Control (CDC) 11,12. We also excluded recipients listed as receiving a whole liver from a live donor (n = 15), recipients with a diagnosis that could not plausibly be concurrently managed as an outpatient and listed as status 1 based on OPTN listing requirements (n = 48), recipients who died prior to discharge (n = 228) and those without a reported post-transplant LOS (n = 26). Prior to analysis, recipients were stratified by age. We defined infants as recipients ≤24 months (n = 2,726) and we defined children as recipients >24 months (n = 3,210).

Standardized Weight and Z-score calculations

Z-scores were calculated using SRTR data and the STATA command zanthro 13. The decisions to stratify our study population by age reflects several key differences between the two age groups, including calculation of Z-scores 11,14,15, indications for liver transplantation 16, and differences in absolute LOS. In infants, Z-scores were calculated using World Health Organization (WHO) WFL growth standards (2006) and recipient gender, length, and weight at time of transplant. In infants, listed height was assumed to be measured as recumbent length. Infants were stratified into the following categories based on Z-score: underweight (Z < −2), normal (−2 ≤ Z ≤ 2), and overweight (Z > 2). In children, Z-scores were calculated using CDC BMI-for-age growth charts (2000), gender, and recipient BMI-for-age at time of transplant and subsequently stratified into the following categories: underweight (Z<−1.64), normal (−1.64 ≤ Z ≤ 1.04), overweight (1.04 < Z ≤ 1.64), and obese (Z > 1.64).

Length of Stay

LOS was defined as the number of hospitalized days after liver transplant. Univariate analysis was performed using a negative binomial regression. Multivariate analysis was performed using a multi-level multivariable negative binomial regression to estimate the association of recipient and surgical factors with LOS following transplant in infants (≤24 months) and children (>24 months), accounting for center-level variation. Our results are reported as a ratio, such that an adjusted ratio for obese children of 1.09 would be interpreted as 1.09-fold longer LOS for obese children compared to normal weight children with similar characteristics.

Model selection was performed in infants and children separately given marked differences in absolute LOS. Variables that were judged to be clinically important or with a P value < 0.20 on univariate analysis were candidates for multivariable analysis. Multivariable models were compared using the Akaike Information Criterion (AIC) to mitigate the risk of over-fitting. In infants, we adjusted for gender, underlying disease (biliary atresia vs. non biliary atresia), lab PELD score (PELD < 15 [reference], 15–22, 23–28, 29–34, and PELD > 34), hospitalization at transplant (non-hospitalized [reference], hospitalized non-intensive care unit (ICU), and ICU), race (white vs non-white), and graft type (whole graft [reference], segmental graft (i.e., deceased donor split graft), and living donor graft). In children, we adjusted for age (2–12 [reference] vs. 13–17), gender, biologic PELD/MELD score (PELD/MELD < 15 [reference], 15–22, 23–28, 29–34, and PELD/MELD > 34), hospitalization at transplant (non-hospitalized [reference], hospitalized non-ICU, and ICU), race (white vs non-white), and graft type (whole graft [reference], segmental graft, and living donor graft); we also adjusted for underlying disease with biliary atresia as the reference, acute liver failure (ALF; listed as acute hepatic necrosis in SRTR, autoimmune, malignancy, and other.

Estimation of Median LOS Ratio

We sought to characterize the variation in LOS across centers. From the multilevel negative binomial regression, we used the center-level variance to estimate the median LOS ratio (MLOSR). The center level MLOSR indicates the extent in heterogeneity of LOS across centers and has been previously used to quantify differences in transplantation rate by region 17,18. MLOSR can be interpreted in the context of between center ratios of LOS. Between center ratios of LOS are the pairwise comparisons in LOS ratio between each pair of transplant centers. For each center pair, the center with the higher LOS ratio is compared with the center with the lower LOS ratio, such that the between center ratio is always greater than one. The median of these between center ratios is the MLOSR. A larger MLOSR indicates greater differences between centers. Specifically for LOS, MLOSR indicates the extent to which a recipient’s center determines their LOS, or the expected increase in LOS should a given recipient be hospitalized at a center with a longer LOS.

Statistical Analysis

All statistical tests used a two-sided α of 0.05. Differences in LOS were compared using Kruskal-Wallis test. Confidence intervals were reported using the method of Louis and Zeiger 19. All analyses were performed using STATA 14.2 for Macintosh (College Station, TX, USA). This study was approved by the Institutional Review Board of the Johns Hopkins University School of Medicine.

RESULTS

Recipient Characteristics

We studied 2,726 infants and 3,210 children who underwent liver transplantation between March 2002 and December 2017. Infants who underwent liver transplantation had a mean age of 11 months, were more likely to be female (52.2%), more likely to be white (48.5%) than black (15.9%), Hispanic (24.3%), or other (11.3%), and more likely to have private insurance (53.8%) than public (43.9%) or other types of insurance (2.3%) (Table 1). Infants were more likely to have a normal weight for length (81.1%) than be overweight (10.1%) or underweight (8.8%). The most common diagnosis in infants was biliary atresia (62.0%) followed by other diagnoses (24.2%), malignancy (7.3%), acute liver failure (6.3%), and autoimmune conditions (0.2%). The average biologic PELD score at transplantation was 17, and the average allocation PELD score was 26 with 28.6% of infants transplanted as status 1. Infants were less likely to be hospitalized (23.2%) or in the ICU (19.5%) than not hospitalized (57.3%) at time of transplantation. Finally, infants were more likely to receive a whole graft (48.9%) than a segmental (35.1%) or living donor graft (16.0%). Children who underwent liver transplantation had a mean age of 9 years, were less likely to be female (49.4%), more likely to be white (55.6%) than black (15.7%), Hispanic (20.8%), or other (7.9%), and more likely to have public insurance (50.2%) than private (45.2%) or other types of insurance (4.6%). Children were more likely to have a normal BMI-for-age (63.1%) than be overweight (16.3%), obese (14.9%), or underweight (5.7%). Children were less likely to be diagnosed with biliary atresia (21.3%), ALF (18.0%), autoimmune conditions (8.8%), or malignancy (11.8%) than to carry other diagnoses (40.1%). The average biologic PELD/MELD score at transplantation was 11, and the average allocation PELD/MELD score was 21 with 33.6% of children transplanted as status 1. Children were less likely to be hospitalized (11.1%) or in the ICU (22.0%) than not hospitalized (66.9%) at time of transplantation. Finally, children were more likely to receive a whole graft (70.5%) than a segmental (21.0%) or living donor graft (8.5%).

TABLE 1.

Demographics and Clinical Characteristics of Pediatric Liver Transplant Recipients

Characteristics Infants (N = 2,726) Children (N = 3,210)
Age, mean (SD) 11 (6) months 9 (5) years
Female, % 52.2 49.4
Standardized Weight, %
 Underweight 8.8 5.7
 Normal 81.1 63.1
 Overweight 10.1 16.3
 Obese -- 14.9
Race, %
 White 48.5 55.6
 Black 15.9 15.7
 Hispanic 24.3 20.8
 Other 11.3 7.9
Biologic PELD/MELD, mean (SD) 17 (14) 11 ± 14
Allocation PELD/MELD, mean (SD)* 26 (13) 21 ± 14
Status 1, % 28.6 33.6
Cause of ESLD, %
 Biliary Atresia 62.0 21.3
 Acute Liver Failure 6.3 18.0
 Autoimmune 0.2 8.8
 Malignancy 7.3 11.8
 Other 24.2 40.1
Hospitalization Status, %
 Not Hospitalized 57.3 66.9
 Hospitalized 23.2 11.1
 ICU 19.5 22.0
Insurance, %
 Private 53.8 45.2
 Public 43.9 50.2
 Other 2.3 4.6
Graft Type, %
 Whole 48.9 70.5
 Segmental 35.1 21.0
 Living 16.0 8.5
Open in a new tab

Abbreviations: PELD/MELD, pediatric end-stage liver disease/model for end-stage liver disease; ESLD, end-stage liver disease.

*

Recipients transplanted as status 1 do not have an allocation PELD/MELD Score

Length of Stay Analysis in Infants

In infants, the median (IQR) LOS was 19 (13–32) days. The median LOS was longer for infants who were hospitalized (20 days) or in the ICU (28 days) compared to non-hospitalized infants (17 days, P < 0.001), infants with public insurance (20 days) compared to private insurance (18 days, P < 0.001), and infants of segmental grafts (22 days) compared to infants who received living donor (18 days) or whole grafts (18 days, P < 0.001) (Table 2). Similarly, median LOS was longer for infants with increasing PELD scores (P < 0.001). There was no difference in median LOS between underweight (19 days), normal weight-for-length (19 days), and overweight infants (20 days, P = 0.33), or between white (19 days) and non-white infants (20 days, P = 0.09).

TABLE 2.

LOS and Adjusted LOSR by Recipient and Surgical Characteristics in Infants

Median LOS (IQR) Days P* Adjusted LOSR P†
Standardized Weight 0.33
 Underweight 19 (12–31) 1.001.101.20 0.04
 Normal 19 (13–31.5) --
 Overweight 20 (13–36.5) 0.961.041.14 0.39
Gender 0.52
 Male 19 (13–31) --
 Female 19 (13–32) 0.920.961.01 0.14
Race 0.09
 White 19 (12–31) --
 Non-white 20 (13–32) 1.001.051.11 0.06
Biologic PELD < 0.001
 <15 17 (11–27) --
 15–22 19 (13–31) 1.061.141.22 < 0.001
 23–28 21 (14–35) 1.111.201.30 < 0.001
 29–34 22 (14–36) 0.981.071.16 0.12
 >34 24 (15.5–40) 0.991.091.20 0.07
Cause of ESLD 0.02
 Biliary Atresia 19 (13–31) --
 Non-Biliary Atresia 20 (13–33) 0.971.031.08 0.39
Hospitalization Status < 0.001
 Not Hospitalized 17 (12–27) --
 Hospitalized 20 (14–33) 1.081.161.23 < 0.001
 ICU 28 (17–47) 1.461.591.70 < 0.001
Insurance < 0.001
 Private 18 (12–39) --
 Public 20 (13–34) 1.031.091.15 0.003
 Other/Missing 17 (13–27) 0.810.961.14 0.66
Graft Type < 0.001
 Whole 18 (11–31) --
 Segmental 22 (15–34) 1.081.151.22 < 0.001
 Living Donor 18 (13–28) 0.860.931.01 0.09
Open in a new tab
*

P-values are testing for differences in LOS between groups using a Kruskal-Wallis test.

†

P-values are testing for statistical significance in the multivariable negative binomial regression.

Abbreviations: LOS, length of stay; LOSR, length of stay ratio; IQR, interquartile range; PELD, pediatric end-stage liver disease; ESLD, end-stage liver disease.

In an adjusted model, hospitalized infants had a 1.16-fold longer LOS (ratio:1.081.161.23, P < 0.001) and hospitalized infants in the ICU had a 1.59-fold longer LOS (ratio:1.461.591.70, P < 0.001) compared to non-hospitalized infants (Table 2). Infants who received segmental grafts had a 1.15-fold longer LOS (ratio:1.081.151.22, P < 0.001), while there was no difference in LOS for infants who received live donor grafts (ratio:0.860.931.01, P = 0.09) and infants who received whole grafts. Infants with public insurance had a 1.09-fold longer LOS (ratio:1.031.091.15, P = 0.003) compared to infants with private insurance. Underweight infants had a 1.10-fold longer LOS (ratio:1.001.101.20; P = 0.04) while the LOS of overweight infants did not differ (ratio:0.961.041.14; P = 0.39) from normal weight infants. Increasing PELD scores did not have a consistent association with LOS. Similarly, there was no difference in LOS between infants transplanted for reasons other than biliary atresia (ratio:0.971.031.08, P = 0.39) and those transplanted for biliary atresia or between female (ratio:0.920.961.01, P = 0.14) and male infants.

Length of Stay Analysis in Children

In children, the median (IQR) LOS was 13 (9–21) days. Median LOS was longer for children who were hospitalized (14 days) or in the ICU (17 days) compared to non-hospitalized children (12 days, P < 0.001) and children who received segmental (16 days) or living donor grafts (16 days) compared to children who received whole grafts (12 days, P < 0.001) (Table 3). Median LOS was also longer for underweight (14 days) and obese (14 days) children compared to normal BMI-for-age children (13 days, P < 0.001) and for non-white (14 days) children compared to white children (13 days, P = 0.04). Median LOS was also longer for children transplanted for ALF (15 days) but shorter for children transplanted for autoimmune conditions (10 days) compared to children transplanted for biliary atresia (13 days, P < 0.001). Similarly, median LOS was longer for recipients with increasing PELD/MELD scores (P < 0.001). Median LOS was shorter for children aged 13–17 (12 days) compared to children aged 2–12 (14 days, P < 0.001).

TABLE 3.

LOS and Adjusted LOSR by Recipient and Surgical Characteristics in Children

Median LOS (IQR) Days P* Adjusted LOSR P†
Age (years) <0.001
 2–12 14 (9–22) --
 13–17 12 (8–20) 0.920.971.03 0.32
Standardized Weight <0.001
 Underweight 14 (10–23) 0.961.061.17 0.23
 Normal 13 (9–21) --
 Overweight 13 (9–19) 0.870.920.98 0.01
 Obese 14 (10–25) 1.031.101.17 0.006
Gender 0.01
 Male 13 (9–22) --
 Female 13 (9–21) 0.930.971.01 0.17
Race 0.04
 White 13 (9–21) --
 Non-white 14 (9–22) 0.991.041.09 0.15
Biologic PELD/MELD <0.001
 <15 13 (9–20) --
 15–22 13(8–21) 0.941.001.08 0.92
 23–28 15 (10–23) 0.880.961.05 0.39
 29–34 15 (11–28) 0.991.101.21 0.064
 >34 16 (11–25) 0.981.091.21 0.128
Cause of ESLD <0.001
 Biliary Atresia 13 (9–20) --
 Acute Liver Failure 15 (10–25) 0.780.860.94 0.001
 Autoimmune 10 (8–17) 0.820.911.00 0.051
 Malignancy 13 (9–22) 1.001.091.18 0.055
 Other 13 (9–22) 1.031.091.16 0.006
Hospitalization Status <0.001
 Not Hospitalized 12 (9–19) --
 Hospitalized 14 (9–24) 1.341.441.56 < 0.001
 ICU 17 (12–28) 1.491.621.77 < 0.001
Insurance <0.001
 Private 13 (9–20) --
 Public 13 (9–22) 1.021.071.13 0.008
 Other/Missing 15 (10–25) 1.091.211.36 < 0.001
Graft Type <0.001
 Whole 12 (8–19) --
 Segmental 16 (11–26) 1.201.271.35 < 0.001
 Living Donor 16 (10–26) 1.271.381.51 < 0.001
Open in a new tab
*

P-values are testing for differences in LOS between groups using a Kruskal-Wallis test.

†

P-values are testing for statistical significance in the multivariable negative binomial regression.

Abbreviations: LOS, length of stay; LOSR, length of stay ratio; IQR, interquartile range; PELD/MELD, pediatric end-stage liver disease/model for end-stage liver disease; ESLD, end-stage liver disease.

In an adjusted model, hospitalized children had a 1.44-fold longer LOS (ratio:1.341.441.56, P < 0.001) and hospitalized children in the ICU had a 1.62-fold longer LOS (ratio:1.491.621.77, P < 0.001) compared to non-hospitalized children (Table 3). Children who received segmental grafts had a 1.27-fold longer LOS (ratio:1.201.271.35, P < 0.001) and children who received living donor grafts had a 1.38-fold longer LOS (ratio:1.271.381.51, P < 0.001) compared to children who received whole grafts. Children transplanted for acute ALF had a shorter LOS (ratio:0.780.860.94, P < 0.001), while children transplanted for autoimmune conditions (ratio:0.820.911.00, P = 0.051) or malignancy (ratio:1.001.091.18, P = 0.055) had similar LOS compared to children transplanted for biliary atresia. Children transplanted for other causes of ESLD had a longer LOS (ratio:1.031.091.16, P = 0.006) compared to children transplanted for biliary atresia. Children with public insurance had a 1.07-fold longer LOS (ratio:1.021.071.13, P = 0.008) compared to children with private insurance. Underweight children had a similar LOS (ratio:0.920.971.03, P = 0.32) compared to children with a normal BMI-for-age. Overweight children had a shorter LOS (ratio:0.870.920.98, P = 0.01) while obese children had a longer LOS (ratio:1.031.101.17, P = 0.006) compared to normal BMI-for-age children. Increasing PELD/MELD score categories demonstrated no differences in LOS. Similarly, there was no difference in LOS between female (ratio:0.930.971.01, P = 0.17) and male children or between children ages 13–17 (ratio:0.920.971.03, P = 0.32) and children ages 2–12.

Center-Level Variation and Median Incidence Rate Ratio

There were 69 centers who performed liver transplants in infants during the study period. In an adjusted model, the center specific LOS varied from 13 to 35 days (Figure 1) and the MLOSR was 1.25. A MLOSR of 1.25 would indicate that two infants with similar characteristics transplanted at different centers would expect a 1.25-fold difference in LOS. There were 96 centers who performed liver transplants in children during the study period. In an adjusted model, the center specific LOS varied from 8 to 21 days (Figure 2) and the MLOSR was 1.26.

Figure 1:

Figure 1:

Open in a new tab

Distribution of estimated adjusted* LOS for each of the 69 centers in our study that performed liver transplants in infants under 24 months of age from 2002–2016.The red line denotes the median LOS.

*Adjusted for standardized weight, gender, race, biologic PELD categories, cause of ESLD, hospitalization status, insurance, and graft type. Abbreviations: LOS, length of stay; PELD, pediatric end-stage liver disease; ESLD, end-stage liver disease.

Figure 2:

Figure 2:

Open in a new tab

Distribution of estimated adjusted* LOS for each of the 96 centers in our study that performed liver transplants in children over 24 months of age from 2002–2016. The red line denotes the median LOS.

*Adjusted for age, standardized weight, gender, race, biologic PELD/MELD categories, cause of ESLD, hospitalization status, insurance, and graft type. Abbreviations: LOS, length of stay; PELD/MELD, pediatric end-stage liver disease/model for end-stage liver disease; ESLD, end-stage liver disease.

DISCUSSION

In this national study of 2,726 infants and 3,210 children who underwent liver transplantation from 2002–2016, we found that hospitalization at transplant, both ICU and non-ICU hospitalization, public insurance, and receipt of a deceased donor segmental graft were associated with a longer LOS in both infants and children. In infants, underweight and PELD scores of 15–22 and 23–28 were associated with a longer LOS. However, in children receipt of a living donor graft, obesity, malignancy, and other causes of ESLD (i.e., causes other than biliary atresia, ALF, autoimmune conditions, and malignancy) were associated with a longer LOS. In children, overweight and a diagnosis of ALF were associated with a shorter LOS. We found that infants had a longer LOS than children. Furthermore, we found that center-level practices account for substantial variation in LOS for both infants and children.

In this analysis, we stratified our data based on age given the anticipated different trajectories for hospitalization of infants and children. At the same time, we provided unadjusted median LOS associated with each variable such that physicians and families can better understand how these variables will influence LOS in a specific patient. For example, infants with biliary atresia had a median LOS of 19 days whereas children with biliary atresia had a median LOS of 13 days. Similar to findings by Bucuvalas et al., who examined length of stay for pediatric liver transplant recipients in the pre-PELD/MELD era (i.e., 1995–2003), we found that infants and children that received segmental grafts experienced a longer LOS 5. Given the persistently high waitlist mortality rates in pediatric candidates under 1, there is increasing interest in the transplant community to ensure maximal usage of all viable grafts through the use of segmental grafts, despite the likelihood of higher rates of vascular and biliary complications 7. While recent studies have suggested that outcomes such as graft failure and death have improved for recipients of segmental grafts and may be equivalent with receipt of a whole graft, vascular and biliary complications remain high, even in centers that employ frequent use of segmental grafts and may be a driver of the longer LOS seen in these recipients 7,20,21. Greater use of these segmental grafts, and the corresponding complications, would likely lead to greater healthcare costs secondary to increase LOS and this impact should be considered when advocating for their use. Similarly, while research to characterize the ideal outcome for pediatric candidates has looked at long-term outcomes, additional LOS over a baseline “best-case scenario” reflects the existence of barriers (e.g., poor nutrition, vascular complications) towards an ideal short-term outcome 22,23.

Similarly, our study also found that infants and children with public insurance had a longer LOS. The relationship between public insurance and worse post-transplant outcomes has been previously reported but poorly characterized 21,24. Public insurance in the United States has classically been associated with lower socioeconomic status and while non-white race in our model was not associated with a longer LOS, it is important to note that approximately 70% of black infants and 65% of black children in our cohort had public insurance, compared to only 50% of non-black infants and 40% of non-black children. Several mechanisms could explain greater LOS for individuals on public insurance including greater disease severity at transplant that is not adequately accounted for by PELD/MELD, fewer immediate resources (e.g., available social support, ability to pay uncovered healthcare costs), and lower health literacy that is necessary to assure a smooth transition home. Although our study can not identify the specific cause of greater LOS in this group, it is likely the case that discharge coordinators and educators should initiate insurance authorization for medications and education even earlier for people on public insurance in order to minimize unnecessary hospital days.

We found that infants who received living donor partial grafts experienced a similar LOS to infants who receive whole grafts. This association is consistent with other studies reporting lower rates of graft failure in infants receiving partial grafts from living donors and the increased surgical complexity of liver transplantation in infants, such that in infants, a sub-population of pediatric liver transplant recipients with high rates of vascular and biliary complications, the benefit of live donor liver transplantation is seen even in the post-operative hospital LOS 25,26.

Our study found that children with ALF experienced a shorter LOS, as opposed to the previously reported longer LOS 5. Though historically associated with poor outcomes, outcomes for pediatric patients with ALF have improved, both in medical management alone and after liver transplantation 27. Furthermore, a recent study of children with ESLD due to chronic disease suggests that these children may be further disadvantaged by the presence of frailty, a marker of morbidity not adequately captured by conventional laboratory tests 28. Frailty was shown to be present in 24% of their cohort of children and represents a possible mechanism by which children with chronic liver disease would experience a longer LOS compared to children with ALF. Children with ALF are likely to be previously healthy prior to referral for liver transplantation and may be more likely to withstand the acute stress of liver transplantation 27. Finally, our analysis is subject to survival bias due to both the increased waitlist mortality and post-transplant mortality of children with ALF compared to those with chronic liver disease 29.

In contrast to the prior analysis, we analyzed the effects of hospitalization status, PELD/MELD scores at time of transplant, and standardized weight, both insufficient and excess weight. We found that both ICU and non-ICU hospitalization at time of transplant was strongly associated with a longer LOS, suggesting that both ICU and non-ICU hospitalization at time of transplant are accurate reflections of disease acuity. We also found PELD scores in infants and PELD/MELD scores in children did not have consistent associations with a longer LOS following liver transplant. Our findings are consistent with a prior study demonstrating that PELD/MELD scores in pediatric recipients are poor predictors of post-transplant mortality, given that PELD/MELD scores were intended to determine the 90-day mortality of candidates awaiting liver transplant and not transplant recipients 30. In addition to the frailty recently described in children with chronic ESLD, another possible contributing factor to the poor associations between PELD/MELD scores and post-transplant outcomes in pediatric liver transplant recipients is the widespread use of nonstandard exception requests in pediatric liver transplant candidates, quoted as 40% in a recent study, further suggesting the calculated biologic PELD/MELD alone is not predictive of post-transplant outcomes 31.

While we found an association between underweight infants and a longer LOS as seen in other studies of underweight pediatric patients, we did not find this association in underweight children but did find an association with excess weight 32,33. We found that overweight children had a shorter LOS, while obese children had a longer LOS than their normal weight counterparts. In other studies, pediatric patients undergoing operative intervention at the extremes of weight have a longer LOS compared to their normal weight counterparts, though this was not always inclusive of overweight children 34–36. We found that overweight children experienced a shorter LOS, which may be due to residual confounding, such that overweight may confer a benefit to withstand the stressors of operative intervention in children with ESLD not accounted for by the covariates in our model. Although our findings should be interpreted cautiously given the challenges in using weight as a marker of malnutrition for individuals with ESLD and potential ascites, our findings in underweight and obese recipients suggest that multidisciplinary teams should focus their efforts on optimizing the nutritional status of underweight recipients in a clinical setting, aiming to correct any nutritional deficiencies prior to operative intervention and improve post-operative outcomes 37–38. Underweight infants in our study experienced a longer LOS, which may be due to a more gradual return to enteral nutrition in this young age group. However, obese children generally do not receive such interventions as their excess weight is seen as a marker of nutritional homeostasis, though this may not be the case 39–41. Furthermore, obese children are at increased risk for perioperative events that prolong LOS, such as adverse respiratory events 42.

The concept of value in healthcare is inevitably tied to outcomes relative to costs, with the ultimate goal of reducing costs without compromising quality 43. Our findings demonstrate that center level practices are associated with substantial variation in LOS. In infants, the differences in LOS across centers were as impactful as non-ICU hospitalization. In children, the differences in LOS across centers were as impactful as the receipt of a segmental graft. We did not directly analyze differences in quality, but the substantial differences in LOS across centers represent a potentially modifiable factor to reduce costs. Just as research has shown that transplant center volume effects waitlist mortality, additional analyses should be undertaken to evaluate center-level variables on post-transplant LOS including volume, center experience with segmental grafts, and regional variation in access to organs 43.

We acknowledge several limitations of our study. First, we utilized a national registry of transplant outcomes (SRTR) and given that LOS is a physician reported outcome, there can be errors in reporting. Secondly, while we were able to adjust for hospitalization at transplant, we were unable to analyze differences (if any) in recipients with different lengths of pre-transplant ICU or hospital stays.

In summary, we have characterized the associations between LOS and recipient and surgical characteristics, and how they differ between infants and children. These results may have important implications for physicians aiming to provide more personalized patient and family counseling and aid in discharge planning.

ACKNOWLEDGEMENTS/FUNDING

The data reported here have been supplied by the Minneapolis Medical Research Foundation (MMRF) as the contractor for the SRTR. The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of, or interpretation by, the SRTR or the U.S. Government. Dr. Covarrubias is supported by a Doris Duke Clinical Research Mentorship Grant. Dr. Massie is supported by grant number K01DK101677 from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). Dr. Garonzik-Wang is supported by grant number K23DK115908 from the NIDDK and a Doris Duke Clinical Scientist Development Award. Dr. Segev is supported by grant number K24DK101828 and R01DK111233 from the NIDDK. Dr. Mogul is supported by grant number K08HS023876 from the Agency for Healthcare Research and Quality.

ABBREVIATIONS

ALF

acute liver failure

ESLD

end-stage liver disease

LOS

length of stay

MELD

Model for End-stage Liver Disease

MLOSR

median length of stay ratio

PELD

Pediatric End-stage Liver Disease

SRTR

Scientific Registry of Transplant Recipients

Footnotes

DISCLOSURE

The authors declare no conflicts of interest.

REFERENCES

  • 1.Kim WR, Lake JR, Smith JM, et al. OPTN/SRTR 2016 Annual Data Report: Liver. American Journal of Transplantation. 2018;18 Suppl 1:172–253. [DOI] [PubMed] [Google Scholar]
  • 2.McAdams-DeMarco MA, King EA, Luo X, et al. Frailty, Length of Stay, and Mortality in Kidney Transplant Recipients: A National Registry and Prospective Cohort Study. Ann Surg. 2017;266(6):1084–1090. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Shellmer D, Brosig C, Wray J. The start of the transplant journey: referral for pediatric solid organ transplantation. Pediatr Transplant. 2014;18(2):125–133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Buchanan P, Dzebisashvili N, Lentine KL, Axelrod DA, Schnitzler MA, Salvalaggio PR. Liver transplantation cost in the model for end-stage liver disease era: looking beyond the transplant admission. Liver Transpl. 2009;15(10):1270–1277. [DOI] [PubMed] [Google Scholar]
  • 5.Bucuvalas JC, Zeng L, Anand R. Predictors of length of stay for pediatric liver transplant recipients. Liver Transplantation. 2004;10(8):1011–1017. [DOI] [PubMed] [Google Scholar]
  • 6.Wiesner RH. Evolving trends in liver transplantation: listing and liver donor allocation. Clin Liver Dis. 2014;18(3):519–527. [DOI] [PubMed] [Google Scholar]
  • 7.Hsu EK, Mazariegos GV. Global lessons in graft type and pediatric liver allocation: A path toward improving outcomes and eliminating wait-list mortality. Liver Transpl. 2017;23(1):86–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Abt PL, Rapaport-Kelz R, Desai NM, et al. Survival among pediatric liver transplant recipients: impact of segmental grafts. Liver Transpl. 2004;10(10):1287–1293. [DOI] [PubMed] [Google Scholar]
  • 9.Wiesner RH. Evolving Trends in Liver Transplantation. Listing and Liver Donor Allocation; In. Vol 182014:519–527. [DOI] [PubMed] [Google Scholar]
  • 10.Massie AB, Kucirka L, Segev DL. Big Data in Organ Transplantation: Registries and Administrative Claims. [DOI] [PMC free article] [PubMed]
  • 11.Prevention CfDCa. A SAS Program for the 2000 CDC Growth Charts (ages 0 to <20 years). 2016; https://www.cdc.gov/nccdphp/dnpao/growthcharts/resources/sas.htm. Accessed February 19, 2018, 2018.
  • 12.Prevention CfDCa. Modified Z-scores in the CDC Growth Charts. 2016; https://www.cdc.gov/nccdphp/dnpa/growthcharts/resources/BIV-cutoffs.pdf. Accessed February 19, 2018, 2018.
  • 13.Newton EHJ, Cox NJ, Editors A, et al. Standardizing anthropometric measures in children and adolescents with functions for egen: Update. The Stata Journal. 2013;13(2):366–378. [Google Scholar]
  • 14.Frieden TR, Director Harold Jaffe MW, Stephens JW, et al. Use of World Health Organization and CDC Growth Charts for Children Aged 0–59 Months in the United States. MMWR. 2009;5959(9). [Google Scholar]
  • 15.Greer FR, Bhatia JJS. CDC: Use WHO growth charts for children under 2. AAP News. 2010;31(11):1–1. [Google Scholar]
  • 16.Sundaram SS, Mack CL, Feldman AG, Sokol RJ. Biliary atresia: Indications and timing of liver transplantation and optimization of pretransplant care. Liver Transplantation. 2017;23:96–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Austin PC, Stryhn H, Leckie G, Merlo J. Measures of clustering and heterogeneity in multilevel Poisson regression analyses of rates/count data. Stat Med. 2018;37(4):572–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Zhou S, Massie AB, Luo X, et al. Geographic disparity in kidney transplantation under KAS. Am J Transplant. 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Louis TA, Zeger SL. Effective communication of standard errors and confidence intervals. Biostatistics. 2009;10(1):1–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Yersiz H, Renz JF, Farmer DG, Hisatake GM, McDiarmid SV, Busuttil RW. One hundred in situ split-liver transplantations: a single-center experience. Ann Surg. 2003;238(4):496–505; discussion 506–497. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Mogul DB, Luo X, Bowring MG, et al. Fifteen-Year Trends in Pediatric Liver Transplants: Split, Whole Deceased, and Living Donor Grafts. J Pediatr. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Ng VL, Alonso EM, Bucuvalas JC et al. Health status of children alive 10 years after pediatric liver transplantation performed in the US and Canada: report of the studies of pediatric liver transplantation experience. J Pediatr. 2012;160(5):820–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ng VL, Mazariegos GV, Kelly B et al. Barriers to ideal outcomes after pediatric liver transplantation. Pediatr Transpl. 2019;23(6) [DOI] [PubMed] [Google Scholar]
  • 24.Thammana RV, Knechtle SJ, Romero R, Heffron TG, Daniels CT, Patzer RE. Racial and socioeconomic disparities in pediatric and young adult liver transplant outcomes. Liver Transpl. 2014;20(1):100–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Roberts JP, Hulbert-Shearon TE, Merion RM, Wolfe RA, Port FK. Influence of graft type on outcomes after pediatric liver transplantation. Am J Transplant. 2004;4(3):373–377. [DOI] [PubMed] [Google Scholar]
  • 26.Kitajima T, Sakamoto S, Sasaki K, et al. Living donor liver transplantation for post-Kasai biliary atresia: Analysis of pretransplant predictors of outcomes in infants. Liver Transpl. 2017;23(9):1199–1209. [DOI] [PubMed] [Google Scholar]
  • 27.Kulkarni S, Perez C, Pichardo C, et al. Use of Pediatric Health Information System database to study the trends in the incidence, management, etiology, and outcomes due to pediatric acute liver failure in the United States from 2008 to 2013. Pediatric Transplantation. 2015:888–895. [DOI] [PubMed] [Google Scholar]
  • 28.Lurz E, Quammie C, Englesbe M, et al. Frailty in Children with Liver Disease: A Prospective Multicenter Study. J Pediatr. 2018;194:109–115 e104. [DOI] [PubMed] [Google Scholar]
  • 29.Baliga P, Alvarez S, Lindblad A, Zeng L, Studies of Pediatric Liver Transplantation Research G. Posttransplant survival in pediatric fulminant hepatic failure: the SPLIT experience. Liver Transpl. 2004;10(11):1364–1371. [DOI] [PubMed] [Google Scholar]
  • 30.Rana A, Pallister ZS, Guiteau JJ, et al. Survival Outcomes Following Pediatric Liver Transplantation (Pedi-SOFT) Score: A Novel Predictive Index. American Journal of Transplantation. 2015;15(7):1855–1863. [DOI] [PubMed] [Google Scholar]
  • 31.Braun HJ, Perito ER, Dodge JL, Rhee S, Roberts JP. Nonstandard Exception Requests Impact Outcomes for Pediatric Liver Transplant Candidates. Am J Transplant. 2016;16(11):3181–3191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Tzioumis E, Adair LS. Childhood dual burden of under- and over-nutrition in low- and middle-income countries: a critical review. Food Nutr Bull. 2014;35(2):230–243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.O’Byrne ML, Kim S, Hornik CP, et al. Effect of Obesity and Underweight Status on Perioperative Outcomes of Congenital Heart Operations in Children, Adolescents, and Young Adults: An Analysis of Data From the Society of Thoracic Surgeons Database. Circulation. 2017;136(8):704–718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Gleich SJ, Olson MD, Sprung J, et al. Perioperative outcomes of severely obese children undergoing tonsillectomy. In. Vol 222012:1171–1178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Becker PJ, Carney LN, Corkins MR, et al. Consensus Statement of the Academy of Nutrition and Dietetics/American Society for Parenteral and Enteral Nutrition: Indicators Recommmended for the Identification and Documentation of Pediatric Malnutrition (Undernutrition). Journal of the Academy of Nutrition and Dietetics. 2014;114(12):1988–2000. [DOI] [PubMed] [Google Scholar]
  • 36.Nafiu OO, Reynolds PI, Bamgbade OA, Tremper KK, Welch K, Kasa-Vubu JZ. Childhood body mass index and perioperative complications. Paediatr Anaesth. 2007;17(5):426–430. [DOI] [PubMed] [Google Scholar]
  • 37.Nightingale S, Ng VL. Optimizing Nutritional Management in Children with Chronic Liver Disease. In. Vol 56: Elsevier; 2009:1161–1183. [DOI] [PubMed] [Google Scholar]
  • 38.Canada NL, Mullins L, Pearo B, Spoede E. Optimizing Perioperative Nutrition in Pediatric Populations. Nutrition in Clinical Practice. 2016;31(1):49–58. [DOI] [PubMed] [Google Scholar]
  • 39.Turer CB, Lin H, Flores G. Prevalence of Vitamin D Deficiency Among Overweight and Obese US Children. PEDIATRICS. 2013;131(1):152–161. [DOI] [PubMed] [Google Scholar]
  • 40.Pinhas-Hamiel O, Doron-Panush N, Reichman B, Nitzan-Kaluski D, Shalitin S, Geva-Lerner L. Obese children and adolescents: a risk group for low vitamin B12 concentration. Arch Pediatr Adolesc Med. 2006;160(9):933–936. [DOI] [PubMed] [Google Scholar]
  • 41.Gunanti IR, Marks GC, Al-Mamun A, Long KZ. Low Serum Vitamin B-12 and Folate Concentrations and Low Thiamin and Riboflavin Intakes Are Inversely Associated with Greater Adiposity in Mexican American Children. Journal of nutrition. 2014;144(12):2027–2033. [DOI] [PubMed] [Google Scholar]
  • 42.Chidambaran V, Tewari A, Mahmoud M. Anesthetic and pharmacologic considerations in perioperative care of obese children. J Clin Anesth. 2018;45:39–50. [DOI] [PubMed] [Google Scholar]
  • 43.Porter ME. What is value in health care? N Engl J Med. 2010;363(26):2477–2481. [DOI] [PubMed] [Google Scholar]
  • 44.Rana A, Pallister Z, Halazun K, Cotton R, et al. Pediatric liver transplant center volume and the likelihood of transplantation. Pediatrics. 2015;136(1):e99–107. [DOI] [PubMed] [Google Scholar]

RESOURCES