Linear Growth Patterns in Prepubertal Children Following Liver Transplantation
Abstract
Factors impacting linear growth following pediatric liver transplantation (LT) are not well understood. This longitudinal analysis examines predictors of linear growth impairment in prepubertal children included in Studies of Pediatric Liver Transplantation. In 1143 children with serial measurements, mean height scores increased from −1.55 at LT to −0.87 and −0.68 at 24 and 36 months post LT with minimal subsequent catch up growth observed until 60 months. Subgroup analysis of height measurements at 24 months (n = 696), 33.8% were below 10th percentile at 24 months post LT. Multivariate analysis revealed linear growth impairment more likely in patients with metabolic disease (OR 4.4, CI: 1.83–10.59) and >18 months of steroids exposure (OR 3.02, CI: 1.39–6.55). Higher percentiles for weight (OR 0.80, CI: 0.65–0.99) and height (OR 0.62, CI: 0.51–0.77) at LT decreased risk. Less linear catch up was observed in patients with metabolic disease, non-Biliary atresia cholestatic diseases and lower weight and higher height percentiles prior to LT. Prolonged steroid exposure and elevated calculated glomerular filtration rate and γ-Glutamyltransferase following LT were associated with less catch up growth. Linear growth impairment and incomplete linear catch up growth are common following LT and may improve by avoiding advanced growth failure before LT and steroid exposure minimization.
Introduction
Physical growth is an important indicator of overall health in children with chronic disease states, including those with liver diseases who require transplantation. Prior to transplant, the majority of children with chronic liver disease have significant nutritional growth failure and alterations of the growth hormone axis, both of which contribute to delayed linear growth (1–3). Malnutrition is due variously to inadequate nutrient intake, malabsorption of fat and fat-soluble vitamins, abnormal nutrient and intermediary nitrogen synthesis, storage and metabolism, and altered energy expenditure (2,4). The low serum levels of insulin-like growth factor I (IGF-I), insulin-like growth factor II (IGF-II) and insulin-like growth factor binding protein 3 (IGFBP-3) observed in the setting of adequate caloric intake and elevated growth hormone levels, suggest growth hormone resistance (1,3). Recognizing the important impact of growth failure in the pediatric candidates, UNOS elected to include this variable in the current liver allocation policy.
After successful liver transplantation (LT) most children display accelerated or ‘catch up’ growth, usually in the second postoperative year (5–9). Growth hormone levels and IGF-I levels return to normal as the rate of linear growth improves (10). However, this ‘catch up’ appears incomplete and the height distribution of this population is consistently lower than normal children (8,11). Single center reports suggest linear growth following LT is negatively correlated with the degree of malnutrition and growth delay before transplant, the need for retransplantation, and steroid usage (7,8,12). However, analysis of factors impacting linear growth after transplant to date are limited by relatively small sample sizes and a wide distribution of age at transplant, primary disease and outcome status.
The aims of this analysis were to describe the linear growth patterns of prepubertal children included in the Studies of Pediatric Liver Transplantation (SPLIT) registry and identify potentially modifiable predictors of linear growth impairment in this large, prospective, multi-center cohort.
Methods
The SPLIT registry was founded in 1995 as a prospective data repository for children receiving LT in the United States and Canada and comprises the largest detailed database of pediatric liver transplants. All SPLIT centers have individual Institutional Review Board approval and informed consent is obtained from parents and/or guardians. De-identified information is submitted to the SPLIT data coordinating center via a standardized web-based data entry system beginning at the time of listing for transplantation and includes clinical, laboratory, operative, medical treatment, complications and outcome data fields. Data reported to the SPLIT data-coordinating center as of June 1, 2007 were included in this study.
This study had two primary objectives: (1) to describe the distribution of height at defined intervals following transplantation and (2) to identify predictors of linear growth impairment and delayed catch up growth at 24 months following transplantation via multivariate analyses. The 24-month time point was chosen because the distribution of z-scores at yearly intervals following transplantation suggested that growth acceleration slows after this time point.
Patients
The patient sample for the study included children followed in the SPLIT registry between 1995 and 2007, receiving their first liver transplant between the ages of 6 months and 8 years and having height measurements at the time of transplant and at least one posttransplant visit before 10 years of age. Patients receiving a second transplant were excluded. The patient sample for the multivariate analyses was restricted to 696 of the 1143 children identified as having height measurements at 24 months (measurement window 21–30 months) following transplant. The upper age of 8 years at the time of transplantation and 10 years at the time of the last included height measurement was selected to focus the analysis on prepubertal growth. Since Tanner staging was not an element of SPLIT data collection until 2005, age was the only variable that could be used to estimate developmental stage. Children younger than 6 months at time of transplant were excluded because the majority experienced perinatal illness or prematurity that may have had an independent impact on growth during the first 2 years of life.
Growth data collection and analysis
SPLIT growth data collection includes height measurements collected by stadiometer for ambulatory children and crown to heel measurements for younger children. Heights are measured prior to transplant, at time of transplant, at 6, 12, 18 and 24 months following transplant, and annually thereafter. Height standard deviation scores (z-scores) were calculated using age and sex-specific levels for the normal population provided by 2000 Center for Disease Control and Prevention (CDC) growth charts, http://www.cdc.gov/growthcharts (13). When a height measurement was not collected at the time of transplant, but was available at a time point within 3 months prior to transplant for children of age <36 months at LT and within 6 months prior to transplant for children of age ≥36 months at LT, the z-score for the height measurement from that last follow-up was used as the baseline height values prior to transplantation.
For the purpose of this analysis, linear growth impairment was defined as z-score for height of less than −1.28 which is the tenth percentile for age and sex. This broader definition was chosen to allow modeling that would predict a wider spectrum of linear growth impairment. Catch up growth was defined as change in z-score from transplant to 24 months and was analyzed as a continuous variable. Factors analyzed as possible predictors of these two outcomes included 5 demographic and 22 medical variables routinely collected by SPLIT.
Statistics
Data are summarized using means and standard errors for continuous factors and using proportions for categorical factors. Risk factors were identified from a two-stage process using logistic regression for the growth impairment outcome and linear regression for the change in height z-score outcome. In stage 1, univariate analyses were performed using Kruskal–Wallis test for continuous factors and chi-square test for categorical factors. In stage 2, factors significant at the 0.15 level in the univariate analyses were included in the initial multivariate models. Final multivariate models were derived using stepwise backward elimination procedure. Model simplification continued until the reduced model yielded a significant worsening of fit according to the likelihood ratio criterion (p ≤ 0.05). All statistical analyses were performed using SAS for Windows, version 9.1 (SAS Institute Inc., Cary, NC).
Results
Study population
The study population included 1143 children. Patient characteristics are included in Table 1. Figure 1A displays the mean z-score for height at yearly intervals from transplant to 60 months for the group and by gender. Mean group standardized height scores increased from –1.55 at baseline/transplant to −0.87 at 24 months with a slope of 0.029 units/month. Slower growth was noted after 24 months resulting in a mean z-score of −0.68 at 36 months and a minimal change in height z-score was noted thereafter. Gender differences were observed in that girls have a steeper slope for a longer period. Figure 1B displays the mean change in z-score, ‘catch up growth’, from baseline to each follow-up time point for the group and by gender. It demonstrates that catch up growth slows from 24 to 36 months and plateaus thereafter. Figure 2A and B depict these same outcomes by primary diagnosis. Data are included from 43 patients with Alagille syndrome and 25 patients with familial cholestasis. The Alagille's group had the lowest average z-score for height at transplant and at 24 months and a pattern of catch up growth similar to patients with Biliary atresia during the first 36 months posttransplant.
| N | % | |
|---|---|---|
| Total | 1143 | 100.0 |
| Sex | ||
| Male | 529 | 46.3 |
| Female | 614 | 53.7 |
| Race | ||
| White | 623 | 54.5 |
| Black | 174 | 15.2 |
| Hispanic | 202 | 17.7 |
| Other | 130 | 11.4 |
| Missing | 14 | 1.2 |
| Age at transplant | ||
| 6–11.9 months | 416 | 36.4 |
| 1–2.9 years | 428 | 37.4 |
| 3–4.9 years | 153 | 13.4 |
| 5–7.9 years | 146 | 12.8 |
| Primary diagnosis | ||
| Biliary atresia | 604 | 52.8 |
| Other cholestatic | 143 | 12.5 |
| Fulminant liver failure | 104 | 9.1 |
| Metabolic disease | 150 | 13.1 |
| Other | 142 | 12.4 |
| Transplant year | ||
| 1995–2001 | 476 | 41.6 |
| 2002–2006 | 667 | 58.4 |
| Weight z-score at transplant (Mean ± SE) | 1087 | −1.22 ± 0.06 |
(A) Height z-score for study cohort at yearly intervals from transplant by gender (mean ± standard error). (B) Change in height z-score from transplant to yearly follow-up intervals after transplant by gender (mean ± standard error).
(A) Mean height z-score for study cohort at yearly intervals from transplant by diagnosis. (B) Mean change in height z-score from transplant to yearly follow-up intervals after transplant by diagnosis.
Predictors of growth failure
Data from 696 children with height measurements at 24 months following transplant were available for analysis of predictors of linear growth impairment and change in z-score at 24 months posttransplant. Linear growth impairment defined as a height z-score of less than the tenth percentile at 2 years follow-up was observed in 235 children (33.8%) and 168 (24.1%) had a height z-score of less than the fifth percentile, see Figure 3. Univariate analysis to predict growth impairment identified factors with a significance level below 0.15, which were included in a multivariate logistic regression model to predict growth impairment at 2 years follow-up. The factors were race, parental marital status at LT, transplant era 1995–2001 versus 2001–2005, primary diagnosis of Biliary atresia versus other diagnoses, parenteral nutritional support at LT versus tube or by mouth, z-score for height, z-score for weight, calculated glomerular filtration rate (cGFR) (14) and total bilirubin at LT, prednisone exposure during 24 months post transplant, number of hospitalizations, number of rejection episodes, history of biliary complications, albumin, cGFR and γ-Glutamyltransferase (GGTP) at 12 months after LT and history of diabetes up to 24 months after LT. Full univariate analysis results are available in Table S1 in Supplementary Material. The multivariate analyses included 361 patients with complete data for all variables selected. Table 2 compares patients with complete versus incomplete data. Distribution of transplant year was the only significant difference between the groups. The model revealed that patients with higher weight and height z-scores at LT, lower total bilirubin at LT, and lower GGTP at 12 months after LT were less likely to be growth impaired at the 2 year follow-up time point (Table 3). Patients that had metabolic disease as the indication for transplant were four times (OR 4.4; CI 1.83–10.59) more likely to be growth impaired as children with Biliary atresia and patients that received prednisone for more than 18 months were three times (OR 3.02;CI 1.39–6.55) more likely to be growth impaired.
Distribution of patients with growth failure at yearly posttransplant intervals.
| Complete data | Incomplete data | p-Value | |||
|---|---|---|---|---|---|
| N | % (column percent) | N | % (column percent) | ||
| Total | 361 | 100.0 | 335 | 100.0 | |
| Sex | |||||
| Male | 159 | 44.0 | 160 | 47.8 | 0.3610 |
| Female | 202 | 56.0 | 175 | 52.2 | |
| Race | |||||
| White | 205 | 56.8 | 175 | 52.2 | 0.7640 |
| Black | 56 | 15.5 | 52 | 15.5 | |
| Hispanic | 61 | 16.9 | 60 | 17.9 | |
| Other | 39 | 10.8 | 42 | 12.5 | |
| Age at transplant | |||||
| 6–11.9 months | 127 | 35.2 | 121 | 36.1 | 0.8780 |
| 1–2.9 years | 142 | 39.3 | 123 | 36.7 | |
| 3–4.9 years | 45 | 12.5 | 47 | 14.0 | |
| 5–7.9 years | 47 | 13.0 | 44 | 13.1 | |
| Primary diagnosis | |||||
| Biliary atresia | 189 | 52.4 | 188 | 56.1 | 0.1562 |
| Other cholestatic | 48 | 13.3 | 44 | 13.1 | |
| Fulminant liver failure | 38 | 10.5 | 21 | 6.3 | |
| Metabolic disease | 40 | 11.1 | 48 | 14.3 | |
| Other | 46 | 12.7 | 34 | 10.1 | |
| Transplant year | |||||
| 1995–2001 | 162 | 44.9 | 188 | 56.1 | 0.0030 |
| 2002–2005 | 199 | 55.1 | 147 | 43.9 | |
| Weight z-score at transplant (Mean ± SE) | 361 | −1.30 ± 0.09 | 302 | −1.23 ± 0.11 | 0.5039 |
| cGFR at transplant (mL/min/1.73 m2) (Mean ± SE) | 361 | 148.10 ± 3.28 | 261 | 149.18 ± 4.03 | 0.8220 |
| cGFR at 12 months (mL/min/1.73 m2) (Mean ± SE) | 361 | 138.50 ± 2.71 | 248 | 140.85 ± 3.86 | 0.9447 |
| Factor | Comparison group | Reference group | OR | 95% CI | p-Value |
|---|---|---|---|---|---|
| Primary disease (overall p = 0.0146) | Other cholestatic disease | Biliary atresia | 1.40 | (0.64, 3.04) | 0.4028 |
| Fulminant liver failure | 1.10 | (0.38, 3.17) | 0.8556 | ||
| Metabolic disease | 4.40 | (1.83, 10.59) | 0.0009 | ||
| Other | 2.16 | (0.95, 4.91) | 0.0675 | ||
| Prednisone use up to 24 months (overall p = 0.0046) | 6–17.9 months | <6 months | 1.42 | (0.70, 2.90) | 0.3350 |
| 18+ months | 3.02 | (1.39, 6.55) | 0.0053 | ||
| Weight z-score at transplant | Continuous | 0.80 | (0.65, 0.99) | 0.0385 | |
| Height z-score at transplant | Continuous | 0.62 | (0.51, 0.77) | <0.0001 | |
| Total Bilirubin at transplant | Continuous | 1.05 | (1.02, 1.07) | 0.0010 | |
| γ-Glutamyltransferase at 12 months (GGTP) | Continuous | 1.0021 | (1.000, 1.004) | 0.0328 | |
- 1The odds ratio and corresponding confidence interval for a 50 unit increase in GGTP is 1.114 and (1.017, 1.221). Note that numbers for a 100 unit increase are 1.242 (1.043, 1.491).
Change in height z-score from transplant to 24 months posttransplant was analyzed as a continuous variable. The mean change in z-score at 24 months for the group was 0.71 ± 1.48. Factors with a significance level below 0.15 in the univariate analyses included gender, age at LT, primary diagnosis of Biliary atresia versus other diagnoses, hospitalization status at LT (p = 0.120), primary immunosuppression, z-score for height, z-score for weight, cGFR and total bilirubin at LT, prednisone exposure during 24 months posttransplant, number of hospitalizations, history of vascular complications, and cGFR and GGTP at 12 months after LT. See Table S2 in Supplementary Material for details of univariate analysis. The multivariate analyses included 422 patients with complete data for these variables. Table 4 compares patients with complete versus incomplete data, with no significant differences noted. The model revealed better catch up growth in patients with Biliary atresia, patients receiving steroids for less than 18 months and patients with a lower cGFR and GGTP at 12 months after LT (Table 5). Standardized height and weight at transplant had opposing effects on catch up growth. Patients with smaller z-scores for weight had less catch up growth, whereas patients with smaller z-scores for height had more catch up growth.
| Complete data | Incomplete data | p-Value | |||
|---|---|---|---|---|---|
| N | % (column percent) | N | % (column percent) | ||
| Total | 422 | 100.0 | 274 | 100.0 | |
| Sex | |||||
| Male | 191 | 45.3 | 128 | 46.7 | 0.7067 |
| Female | 231 | 54.7 | 146 | 53.3 | |
| Race | |||||
| White | 226 | 53.6 | 154 | 56.2 | 0.6931 |
| Black | 67 | 15.9 | 41 | 15.0 | |
| Hispanic | 79 | 18.7 | 42 | 15.3 | |
| Other | 48 | 11.4 | 33 | 12.0 | |
| Age at transplant | |||||
| 6–11.9 months | 144 | 34.1 | 104 | 38.0 | 0.7460 |
| 1–2.9 years | 164 | 38.9 | 101 | 36.9 | |
| 3–4.9 years | 56 | 13.3 | 36 | 13.1 | |
| 5–7.9 years | 58 | 13.7 | 33 | 12.0 | |
| Primary diagnosis | |||||
| Biliary atresia | 221 | 52.4 | 156 | 56.9 | 0.2014 |
| Other cholestatic | 55 | 13.0 | 37 | 13.5 | |
| Fulminant liver failure | 43 | 10.2 | 16 | 5.8 | |
| Metabolic disease | 50 | 11.8 | 38 | 13.9 | |
| Other | 53 | 12.6 | 27 | 9.9 | |
| Transplant year | |||||
| 1995–2001 | 201 | 47.6 | 149 | 54.4 | 0.0819 |
| 2002–2005 | 221 | 52.4 | 125 | 45.6 | |
| Weight z-score at transplant (Mean ± SE) | 422 | −1.26 ± 0.09 | 241 | −1.28 ± 0.12 | 0.9486 |
| cGFR at transplant (mL/min/1.73 m2) (Mean ± SE) | 422 | 147.60 ± 2.97 | 200 | 150.57 ± 4.83 | 0.8488 |
| cGFR at 12 months (mL/min/1.73 m2) (Mean ± SE) | 422 | 136.02 ± 2.45 | 187 | 147.20 ± 4.74 | 0.0523 |
| Factor | Comparison group | Reference group | Parameter estimate | 95% CI | p-Value |
|---|---|---|---|---|---|
| Primary disease (overall p = 0.0167) | Other cholestatic | Biliary atresia | −0.40 | (−0.70, 0.10) | 0.0085 |
| Fulminant liver failure | −0.15 | (−0.49, 0.20) | 0.4045 | ||
| Metabolic disease | −0.35 | (−0.66,−0.04) | 0.0282 | ||
| Other | −0.09 | (−0.40, 0.21) | 0.5380 | ||
| Prednisone use up to 24 months (overall p = 0.0015) | 6–17.9 months | <6 months | −0.12 | (−0.35, 0.12) | 0.3331 |
| 18+ months | −0.48 | (−0.74, −0.22) | 0.0004 | ||
| Weight z-score at transplant | Continuous | 0.20 | (0.12, 0.28) | <0.0001 | |
| Height z-score at transplant | Continuous | −0.71 | (−0.79, −0.64) | <0.0001 | |
| cGFR at 12 months | Continuous | 0.003 | (0.001, 0.005) | 0.0021 | |
| γ-Glutamyltransferase at 12 months (GGTP) | Continuous | −0.000591 | (−0.00102, −0.00016) | 0.0068 | |
- 1The parameter estimate and corresponding confidence interval for a 50 unit increase in GGTP are −0.0295 and (−0.0508, −0.0082). Estimates for a 100 unit increase are −0.0590 and (−0.1016, −0.0164).
The compound (additive) effect of growth failure at transplant and steroid exposure posttransplant was also evaluated. Among the patients that were growth impaired at transplant (ht < 10th percentile), the risk of being growth impaired at 24 months posttransplant increases with increased exposure to steroids (37.7% among those receiving steroids for <6 months, 52.6% for those receiving steroids for 6–17 months, and 66.2% for >18 months of steroid use). A similar relationship was observed in patients with heights above the 10th percentile at transplant, although the incidence of posttransplant linear growth impairment is much lower in this group (12.2% among those receiving steroids for <6 months, 9.5% for those receiving steroids for 6–17 months and 22.6% for >18 months of steroid use). Hence, as expected, the statistical test of interaction between linear growth impairment at transplant and steroid use was not significant (chi-square p-value = 0.18). Overall, patients that are height impaired at transplant and receive steroids for >18 months are 14.1 times more likely to have linear growth impairment at 24 months compared to patients with height >10th percentile at transplant and receiving steroids for <6 months. The test of interaction between linear growth impairment at transplant and prednisone use for the outcome change in height z-score at 24 months was also not significant (p-value = 0.39). The gain in height z-score from transplant to 24 months declined with increasing use of steroids. The change in height z-score was 1.7 standard deviations in height-impaired patients receiving steroids for <6 months compared with 0.9 standard deviations in patients exposed to steroids for over 18 months.
Discussion
This study is a comprehensive analysis of linear growth after LT in children derived from the largest cumulative data set of pediatric liver transplant recipients. These data provide valuable, clinically relevant information to healthcare providers caring for pediatric liver transplant patients. The pattern of growth observed during the initial 5 years after transplant confirms earlier single center studies indicating catch up growth is incomplete. The risk factor analysis suggests that while certain risk factors such as age and cause of liver disease may not be modifiable, there are factors such as pretransplant nutritional status and prolonged use of steroids after transplant that deserve detailed attention.
The observation that catch up growth in children following LT slows after the third to fourth postoperative year has been previously described (5,8). Clear reasons for this phenomenon have not been identified. The cohort of patients in this analysis were transplanted in an era during which pre and posttransplant treatment strategies were fairly stable and included careful attention to nutritional support and minimization of medications implicated in linear growth arrest. Yet, this plateau in catch up growth leaves the mean transplant group adjusted height z-score below the population mean even after 5 years. A recent study that examined final height in older liver transplant recipients found that 50% of these patients reach a final height that is lower than –1.3 standard deviations of their genetic potential (15). These findings further affirm the observation that catch up growth is incomplete in many of these patients even in long-term follow-up.
We performed a risk analysis that included a wide array of factors that could contribute to the two interrelated outcomes: linear growth impairment (using a broad definition of height z-score <10th percentile at 24 months) and catch up growth. Our goal was to focus on a time point just prior to the observed plateau in the slope of catch up growth and perform detailed multivariate modeling to determine the factors having the most important impact on linear growth. One third of the cohort was classified as growth impaired at 2 years. The determinants of linear growth impairment were not entirely surprising. Patients with an original diagnosis of Biliary atresia were less likely to be growth impaired than those with metabolic diseases. The duration of steroid exposure was longer in growth impaired patients, which was expected based on previous single center reports (5,7,8,12). A lower height z-score at transplant was associated with a moderately increased risk of linear growth impairment. Weight and bilirubin at transplant and GGTP at 12 months following transplant were statistically significant, but considering the associated odds ratios, these factors had less clinical impact.
The strongest predictors of catch up growth were weight and height z-score at transplant. There was a strong direct relationship between weight prior to transplant and posttransplant catch up growth with patients with lower weight percentiles exhibiting less growth acceleration. This observation is possibly related to the need for patients to recover from more severe malnutrition before catch up growth is achievable. Conversely, patients with lower height percentiles at transplant exhibited more linear growth acceleration during the first 24 months. Previous reports examining the relationships between pre and posttransplant growth have been inconclusive with some authors demonstrating pretransplant growth failure to have a positive impact (5,8) and others demonstrating a negative impact (12). This study suggests both observations may be valid. Children with more severe growth arrest prior to transplant have the most to recover and without other limitations the acceleration of their posttransplant linear growth may be more pronounced than that of patients with closer to normal growth patterns prior to transplant. However, even with an above average degree of catch up growth following transplant, patients with the lowest height percentiles at transplant would be less likely to achieve normal percentiles (i.e. z-score ≥ 0) posttransplant. Thus, catch up growth occurs, but is incomplete.
We were particularly interested in examining the impact of liver injury on posttransplant growth. Bilirubin, albumin and liver enzyme levels at 12 months following transplantation were included as markers of posttransplant graft function and/or injury. Of these, only GGTP at 12 months posttransplant was significant in the two growth outcomes models with the association with diminished catch up growth being much stronger than with linear growth impairment. GGTP is a sensitive, but nonspecific marker of bile duct injury. The group of patients in this analysis with elevated GGTP levels was likely a heterogeneous group including those with a history of rejection, partial biliary tract obstruction and other less common causes of bile duct injury. Both variables, rejection episodes and biliary obstruction within the first 12 months were not significant in the model predicting linear growth impairment and were excluded following univariate analysis from the model for catch up growth. The data collected for these variables were not fully detailed to allow us to identify patients with chronic rejection and chronic biliary obstruction, and combining patients with transient and chronic graft injury in these groups may have masked their impact on the outcomes. Although it is not possible from these analyses to identify which type of graft injury has the largest influence, these findings support the concept that ongoing graft injury may limit linear growth in long-term follow-up.
Renal function after transplantation might also be expected to impact linear growth. There was considerable variability in renal function within the patient cohort with 12% of the patients having a cGFR of <90 mL/min/1.73 m2 at 12 months posttransplant (data not shown). Since cGFR is thought to underestimate renal insufficiency in patients with early renal injury the percentage with renal impairment may actually have been higher (16). Unfortunately, direct measurement of GFR is not standard practice at the majority of centers participating in the SPLIT registry and thus this data was not available for inclusion in the analysis. The cGFR at 12 months posttransplant was strongly related to the amount of catch up growth, but not to the risk of linear growth impairment. Patients with compromised renal function did exhibit less catch up growth, but ultimately their level of height deficit prior to transplant played a bigger role in determining whether they were growth impaired at 24 months.
The observation that pretransplant height and weight are important risk factors for posttransplant linear growth impairment supports the current UNOS liver allocation policy, which advantages children with growth failure. Considering the high prevalence of linear growth impairment in long-term follow-up after LT, strategies to maximize posttransplant growth should be a high priority in the transplant community. Allocation policies should strive to provide children with suitable liver grafts before they have advanced linear growth failure. Such policies would require that the emphasis on growth failure in the allocation model be determined not only by its impact on waiting list mortality, but also by its impact on posttransplant growth potential.
What else can we do to improve posttransplant growth while awaiting transplantation? Specific comment is warranted with respect to two potentially modifiable risk factors for linear growth impairment after liver transplant, pretransplant nutrition and posttransplant steroid use. Lower weight percentiles at transplant predicted less catch up growth and a slightly higher risk of linear growth impairment. This observation suggests nutritional support to preserve fat and muscle mass prior to transplant may have a beneficial impact on posttransplant growth even if linear growth has stalled (2,17). Holt et al. have demonstrated that improved pretransplant nutrition and growth could be accomplished in a majority of patients through specialized aggressive nutritional rehabilitation (18). Similarly, the use of posttransplant steroids needs review and further study. The intervals of steroid exposure were selected to reflect standard practices among the treating physicians. During the past 10 years, typical immunosuppressive protocols for pediatric liver recipients have included elimination of steroid exposure between the 6th and 18th month. Only one third of this cohort received steroids for less than 6 months. The observed association between growth failure and prolonged steroid exposure in this study lend strong support to the trend toward steroid minimization, specifically steroid withdrawal within the first 3–6 months or steroid free regimes in pediatric transplantation (19,20). Steroid exposure suppresses linear growth in prepubertal children by diminishing the pulsatile release of growth hormone and by directly inhibiting new bone formation. These effects may be reduced by switching patients to an alternate day schedule and by overall dose reduction. It is acknowledged that many of the children in this study who were receiving steroids beyond the 18 months follow-up likely had chronic graft dysfunction that may have contributed to delayed linear growth as discussed above.
This analysis shares limitations that are common to many large, registry studies. Although data is gathered in a standardized, prospective fashion, compliance with data collection and entry is not complete. Data elements not routinely collected at all centers such as serum albumin levels at 1 year posttransplant were missing for a large number of patients. This reduced the number of patients included in the multivariate model. Comparison of patients with complete versus incomplete data suggests that the primary difference was era of transplant. Patients transplanted during the most recent era had more complete data. Since era of transplant was not significant in either model, this difference is unlikely to bias the results.
In summary, analysis of this large, multi-center pediatric liver recipient cohort revealed catch up growth slows and is incomplete after the second and third posttransplant year. It suggests infants less than 12 months of age and those with Biliary atresia have the most accelerated post transplant growth. Patients with lower z-scores for height prior to transplant exhibit the best catch up growth, but still have the highest risk for linear growth impairment at 2 years posttransplant. One important potentially modifiable risk factor is pretransplant nutrition and growth. Aggressive nutritional support with correction of nutritional deficits while awaiting LT may improve outcomes. Finally, steroid elimination by early withdrawal or steroid-free regimens is a key strategy to improve catch up growth and avoid linear growth impairment.
Acknowledgments
The authors are indebted to the SPLIT Research Group (https://web.emmes.com/study/lvr/) comprised of principal investigators and transplant coordinators for their commitment to the registry and for continuing to submit quality data.
This project was supported by grant number U01 DK061693 of the National Institute of Diabetes and Digestive and Kidney Diseases at the National Institutes of Health.
