School outcomes in children registered in the Studies for Pediatric Liver Transplant (SPLIT) consortium

Abstract

School performance is an important aspect of functional outcomes for pediatric liver transplant (LT) recipients. This longitudinal analysis conducted through the Studies of Pediatric Liver Transplantation (SPLIT) research consortium examines several indicators of school function in these patients. Thirty-nine centers participated in data collection using a semi-structured questionnaire designed specifically for this study. The survey queried school attendance, performance and educational outcomes including the need for special educational services. Participants included 823/1133 (73%) of eligible patients, mean age 11.34±3.84, 53% female, median age at LT 4.6 (range 0.05-17.8) years, and mean interval from transplant was 5.42±2.79. Overall, 34% of patients were receiving special educational services and 20% had repeated a grade, with older participants more likely to have been held back (p=0.0007). Missing more than 10 days of school per year was reported by one third of the group with this level of absence being more common in older participants (p=0.0024) and children with shorter intervals from LT (<0.0001). Multivariate analysis revealed the following factors were associated with the need for special educational services; type of immunosuppression at six months post-LT, CSA (OR 1.8, CI:1.1-3.1), or other (OR 4.9, CI:1.4-17.6) versus tacrolimus, symptomatic CMV infection within 6 months of LT (OR 3.1:CI 1.6-6.1), and pre-transplant special educational services (OR 22.5, CI:8.6-58.4).

Advances in medical and surgical techniques in liver transplantation have enabled long-term survival for pediatric recipients [1] and allowed a shift in research toward examining the long-term functional outcomes of these children. One of the most important areas of function in children and adolescents is school performance, as it reflects their developmental status and prepares them for independent functioning in adulthood. Impaired cognitive development, below average school performance, and inconsistent attendance have all been documented in chronic childhood illness, including solid organ transplantation.[2, 3] It has also been suggested that teachers tend to have lower expectations for academic achievement of the chronically ill child.[4] It is likely that several mechanisms observed in the setting of chronic disease adversely affect cognitive function. Some of these include the impact of the illness and its treatment on the growing brain, particularly when disease onset is during infancy, and the impact of multiple hospitalizations on psychosocial development and behavior. Approximately half of the pediatric liver transplant population require the procedure during infancy which is a particularly vulnerable period of neurological development. Children with liver failure frequently experience hepatic encephalopathy and advanced malnutrition. Transplantation reverses these medical problems, but exposes the patients to potentially neurotoxic medications and is associated with the need for prolonged hospitalization. The end result of these insults can be expressed as poor school performance and hence the potential requirement for special educational resources later in childhood.[2]

Previous single-center studies evaluating the cognitive outcomes in children following liver transplantation have found variable prevalence rates for IQ delays (defined as an IQ < 70) ranging from 5 to 24% [5-8]. Kennard’s cohort of children and adolescents who were evaluated at various intervals post-liver transplant demonstrated the diagnosis of cognitive delay in 18% and learning disability in 26% of children.[5] These previous studies have included relatively small patient cohorts which have limited the investigators’ ability to assess for practice variables that may influence cognitive outcomes.

The Studies of Pediatric Liver Transplant (SPLIT) research consortium has allowed investigators a unique opportunity to survey outcomes over a large cross-section of patients. The School Attendance and Academic Performance Survey (SAAPS) is an annual survey administered through SPLIT and represents the largest accumulation of parent reported information on educational outcomes and school attendance in the pediatric post-liver transplant population. The primary objectives of this study were several fold. Our first goal was to detail the characteristics of school attendance after liver transplantation. Secondly, we wanted to quantify the number of liver transplant recipients requiring special educational assistance and describe the types of services they required. Finally, we sought to develop a model to identify variables that predicted the need for special educational services in this population.

Methods

The SPLIT data registry is a multi-centre data registry for pediatric liver transplant candidates and recipients and includes 45 centers in Canada and the United States. All SPLIT centers have individual Institutional Review Board approval and individual informed consent that is obtained from the parents or guardians. Coded 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 data collection includes detailed information regarding clinical status, laboratory values, medical and operative therapies and patient complications and outcomes.

The School Attendance and Performance Survey (SAAPS) is a semi-structured questionnaire that was specifically designed for all school age children, 6-18 years who participate in the SPLIT project, see Appendix. The survey is completed by the child’s primary caregiver during all annual post-transplant assessments. The survey contains three domains: 1) school attendance; 2) school performance and educational outcomes; 3) parental concerns regarding development and behavior. In the section regarding school performance and educational outcomes, parents are asked to indicate the specific types of special educational support the child receives including individual educational plans (IEP) and 504 Plans. [9]

Potential risk factors for the need for special educational services were assessed and included both pre-transplant and post-transplant variables. Pre-transplant variables included; recipient’s age at transplant, interval since transplant, gender, race, highest parental education, primary diagnosis and requirement of special education pre-transplant. Transplant variables included; graft type, era of transplant (≤ 2001 vs ≥ 2002), patient’s status at transplant, Pediatric End-Stage Liver Disease (PELD) score [10] at transplant, log INR, total bilirubin and albumin at transplant, height and weight z score at transplant, growth failure at transplant (≤ −2SD for height or weight), nutritional intake at listing and wait time for transplant. Since growth failure at transplant had a lower significance value (p=0.048) in univariate analysis than either height (p=0.051) or weight (p=0.17) z score at transplant the combined variable was selected for the model. Post-transplant variables included; retransplantation within the first month, steroid use at transplant, use of poly- or monoclonal antibodies at transplant, type of immunosuppressant at six months post-transplant, rejection within the first six months, biliary tract complication within the first six months, vascular complication within the first six months, hospital days following transplantation, interval from transplant, glucose intolerance within the first six months, symptomatic EBV, CMV or PTLD within the first six months, patient’s current age group (6-11, 12-14, or 15-18 years of age) and change in weight z-score at six months post-transplant. Change in height z-score at six months was not included since a prior analysis of post-transplant growth revealed limited variability in this parameter at six months. [11] Likewise growth failure at six months was not included because it was highly correlated with growth failure at transplant. Data from later time points following transplant could not be included since some patients (n=40) had only completed 9-17 months of follow-up.

Patients

Eligible patients were school age (between 6 and 18 years) and had survived liver transplantation by at least nine months. The SAAPS was completed between June 01, 2005 and March 31, 2008 and all patients considered eligible for this analysis were maintaining routine follow-up at their transplant center as evidenced by a completed SPLIT long-term follow-up form recorded during the study period. If parents had completed the SAAPS twice during the study period, only data from the last form filed was included. Although SPLIT includes 45 centers, only 39 chose to participate in SAAPS data collection.

Statistics

Descriptive data was summarized, comparing survey participants and non-participants, with means, medians, standard deviations and standard errors for continuous factors and proportions for categorical factors. Educational outcomes of the survey participants were analyzed for the total sample and by subsets based upon the age of the child at survey, 6-11 years, 12-14 years and 14-18 years, and time interval from transplant, 9-17 months, 18-35 months and ≥ 36 months. Univariate analyses of variables associated with the requirement for special education were performed using the Kruskal-Wallis test for continuous factors and chi-square test for categorical factors. Variables significant at the 0.10 level in the univariate logistic regression analysis were included in the multivariate model. Final multivariate models were derived using stepwise backward elimination process. Model simplification continued until the reduced model yielded significance (p < 0.05). All statistical analyses were performed using SAS for Windows, version 9.2 (SAS Institute Inc., NC).

Results

During the study period, 1133 patients were eligible for participation, of which 823 participated (72.6%) and 310 were non-participants (27.4%). Table 1 compares selected demographic and medical variables between participants and non-participants. Of note, participants had a lower mean calculated PELD score at the time of transplant (11.5±14.5 versus 13.8±13.8, p=0.0087). Participating patients had a mean age at survey of 11.34±3.84 years, a median age at LT of 4.6 (range 0.05-17.8) years, and a mean interval from transplant of 5.42±2.79 years. Overall, 95.6% of children had attended school during the 12 months prior to completion of the SAAPS. Outcomes for the entire group and subset analysis are demonstrated in Tables 2 through 4. Table 2 details the number of days of school that were missed due to illness or doctor’ visits. Note that 32.8% of children missed greater than 10 days of school per year. Older participants (p=0.0024) and children with shorter intervals from transplant (p<0.0001) were more likely to miss greater than 10 days of school in the preceding year, Table 3 and 4.

Table 1. Characteristics of Participants versus Non-participants.

	Survey Participants				p-value
	Yes (n=823)		No (n=310)
	N	%	N	%
Sex					p = 0.4080
Male	389	47.3	138	44.5
Female	434	52.7	172	55.5
Race
White	514	62.5	175	56.5	p = 0.0777
Black	128	15.6	54	17.4
Hispanic	103	12.5	38	12.3
Asian/Pacific	34	4.1	21	6.8
Aboriginal	9	1.1	9	2.9
Other	31	3.7	10	3.2
Primary Diagnosis
Biliary Atresia	307	37.3	115	37.1	p = 0.8880
Other  cholestatic/metabolic	268	32.6	102	32.9
Fulminant liver failure	115	14.0	49	15.8
Cirrhosis	57	6.9	18	5.8
Other	76	9.2	26	8.4
Age at Transplant
0-6 mos	49	6.0	20	6.5	p = 0.2272
6-12 mos	123	14.9	62	20.0
1-5 yrs	258	31.3	91	29.4
5-13 yrs	294	35.7	96	31.0
13-17 yrs	99	12.0	41	13.2
Primary Payor
Medicaid	278	33.8	114	36.8	p = 0.4566
Provincial gov’t	74	9.0	30	9.7
HMO/managed care	162	19.7	67	21.6
Private insurance	220	26.7	65	21.0
Military	12	1.5	3	1.0
PELD
Mean	11.5±14.5		13.8±13.8		p = 0.0087
Wait time for transplant (mos)
Mean	6.2±11.7		5.6±11.2		p = 0.0847

Table 2. School Attendance.

School Attendance		N	%
Attended school in last 12 months *	No	36	4.4
	Yes	774	95.6
Missed > 10 days of school	No	512	67.2
	Yes	250	32.8
School Days Missed	Total	762	100.0
0-4		277	36.4
5-10		235	30.8
11-20		116	15.2
21-30		47	6.2
31+		87	11.4

^*13 patients with missing data

Table 4. Educational Outcomes by Time Interval from Transplant.

	Time interval from transplant						Total (N=823)		p-value
	9-17 mos (N=71)		18-35 mos (N=128)		≥ 36 mos (N=624)
	N*	%	N*	%	N*	%	N**	%
Currently receiving special education	22	35.5	37	32.2	202	34.1	769	33.9	p = 0.988
Testing for IEP	26	38.8	40	32.5	224	36.7	800	36.3	p = 0.891
History of 504 Plan	14	20.9	12	9.8	61	10.1	796	10.9	p = 0.030
Repeated grade	16	23.2	25	20.7	118	19.4	797	19.9	p = 0.451
Attended school in last 12 months	62	89.9	116	94.3	596	96.4	810	95.6	p = 0.034
Missed school > 10 days	41	67.2	45	38.8	164	28.0	762	32.8	p<0.0001

^N*Number of patients with outcome among total evaluable

^%Percent of patients with outcome among total evaluable

^N**Number evaluable for each outcome

Table 3. Educational Outcomes by Age Group.

	Age Groups						Total (N=823)		p-value
	6-11 yr (N=509)		12-14 yr (N=125)		15-18 yr (N=189)
	N*	%	N*	%	N*	%	N**	%
Currently receiving special education	163	34.1	38	32.2	60	34.7	769	33.9	p = 0.962
Testing for IEP	175	35.4	42	34.4	73	39.9	800	36.3	p = 0.328
History of 504 Plan	44	8.9	14	11.5	29	16.2	796	10.9	p = 0.008
Repeated grade	80	16.2	29	24.0	50	27.3	797	19.9	p = 0.001
Attended school in last 12 months	480	95.6	119	96.7	175	94.6	810	95.6	p = 0.689
Missed school > 10 days	140	29.6	36	30.8	74	43.0	762	32.8	p = 0.002

^N*Number of patients with outcome among total evaluable

^%Percent of patients with outcome among total evaluable

^N**Number evaluable for each outcome

Table 3 demonstrates the educational outcomes for the entire cohort. Overall, 33.9% of participants were receiving special educational services at the time of survey. This percentage ranged from 32.2% to 35.5% depending upon age and interval from transplant, Table 3 and 4. A history of having had testing for an individualized education plan (IEP) was reported for 36.3%, and a history of receiving classroom accommodations via a 504 Plan was reported for 10.9%. A 504 plan was reported more often for older age participants (p=0.0075) and those with the shortest interval from transplant (p=0.0297). The number of participants who repeated a grade level was 19.9% with the older participants more likely to have repeated a grade level (p=0.0007). Parents reported that their child had previously been given a diagnosis of learning disability or mental retardation in 17.4% and 5.2%, respectively.

Predictors of utilization of special educational services

Variables with a significance level of ≤ 0.10 in univariate analysis are detailed in Table 5. Logistic regression analysis with stepwise backward selection procedure was performed on 562 patients with complete data for these variables, Table 6. The use of cyclosporine, (OR=1.83, 95% CI=1.08, 3.10; p=0.0239) and non-calcinuerin inhibitor based immunosuppressant regimes (OR=4.88, 95% CI = 1.35, 17.61; p=0.0154) compared to tacrolimus at six months post-transplant were associated with increased utilization of special educational services. Of note, all but 14 patients received either cyclosporine or tacrolimus at this early time point. Symptomatic CMV infection within the first six months post-transplantation (OR = 3.10, 95% CI 1.57, 6.09; p=0.0011) and a history of special educational services in the pre-transplant period (OR=22.46, 95% CI 8.64, 58.42; p<0.0001) were also both associated with post-transplant special educational support.

Table 5. Univariate Analysis of Risk Factors for Currently Receiving Special Education Services, Factors p ≤ 0.10.

Factor	Comparison Group	Reference Group	Odds Ratio	p-value
Primary diagnosis	Other Cholestatic or Metabolic Fulminant liver failure Cirrhosis Other	Biliary atresia	1.54 1.01 0.63 1.27	p = 0.0448
Growth Deficit at Tx	Yes	No	1.39	p = 0.0477
Immunosuppressio n at Tx	Cyclosporine Other	Tacrolimus	1.72 0.63	p = 0.0057
Immuno at 6 months post tx	Cyclosporine Other	Tacrolimus	1.59 5.49	p = 0.0019
Nutrition Intake at Listing	Tube IV	Mouth	1.75 1.38	p = 0.0289
Early use of monoclonal or polyclonal antibodies	Yes	No	1.41	p = 0.0846
Symptomatic CMV within 6 mths of Tx	Yes	No	2.15	p = 0.0051
PreTx Special Edu Needs	Yes	No	13.97	p <0.0001
Parent Highest Edu	College or Above	Less than College	0.70	p = 0.0234
Year of Tx	>=2002	<=2001	0.72	p = 0.0354
Initial Hospital Stay	Continuous Predictor		1.01	p = 0.0008
Log INR	Continuous Predictor		0.75	p = 0.0985
Change in weight z score at 6 months post Tx	Continuous Predictor		0.87	p = 0.0600

Table 6. Multivariate Analysis of Risk Factors For Currently Receiving Special Education.

Factor	Comparison Group	Reference Group	Odds Ratio	95% CI	p-value
Immunosuppression at 6 months post transplant (overall p=0.0061)	Cyclosporine Other	Tacrolimus	1.83 4.88	(1.08, 3.10) (1.35, 17.61)	p = 0.0239 p = 0.0154
Symptomatic CMV within 6 months of transplant	Yes	No	3.10	(1.57, 6.09)	p = 0.0011
Pre-transplant Special Education Needs (overall p<0.0001)	Yes No education	No	22.46 1.70	(8.64, 58.42) (1.08, 2.66)	p <0.0001 p = 0.0211

Sample size was limited to 562 patients who had complete data for all the variables included in the model.

Discussion

As the population of children that achieve long-term survival following liver transplantation increases it is becoming more important to understand and optimize their functional outcomes.[12] Previous literature has been hampered by single-center design and small sample sizes [5, 7, 8,13] and has focused on intelligence quotient measurements or registration in school.[5, 7, 8, 13, 14] This study greatly expands upon the previous literature by examining functional outcomes, such as school attendance and utilization of special educational services, in a large cohort representing a broader spectrum of children post-liver transplant in a multi-center design.

The SPLIT cohort of 823 participants represents the most comprehensive analysis to date of school outcomes following pediatric liver transplantation. Three quarters of eligible SPLIT patients participated in the survey. Participants were representative of the SPLIT database in terms of demographics, but participants may have been slightly healthier at transplant as evidenced by their lower PELD score. However, the clinical relevance of this small difference is questionable.[15] Therefore, we believe the data collected on this patient cohort is representative of the larger group.

In general, school attendance was excellent with almost 96% of school age children attending school over the past 12 months. Of note, 2% reported being home schooled and only 1% reported being unable to attend school for medical reasons. Although almost all were attending school, one third of these patients missed more than 2 weeks of school (10 school days) in the proceeding school year and more than 11% missed greater than 6 weeks. This is compared to national statistics on chronic school absence which reveals that approximately 10% of children in primary grades miss 12-18 days per year and only 5% miss 18 days or more (http://www.nccp.org/publications/pdf/text_771.pdf). As expected, it was more likely for children to miss school within the first eighteen months post-transplant. Also, older adolescent patients (15-18 years) were more likely to have missed school regardless of interval from transplant. Reasons for missed days require further investigation, as attendance likely influences school performance and academic achievement. Likewise, the relationship between missed school days and cognitive function were not explored in this study, but will be addressed in a longitudinal analysis of cognitive function that is currently in progress.

A diagnosis of learning disability was reported in 17.4% of participants. The expected normative population rate of learning disability is 8% [16], therefore the post-pediatric liver transplant rate is more than twice what is expected. Furthermore, one-third (33.9%) were receiving special education services at the time of survey. This suggests that patients were experiencing a broader range of academic difficulties requiring special education supports than those fitting narrowly in the category of learning disabilities and/or that parents were under-reporting (and perhaps lacking awareness of) their children’s learning disabilities. The pediatric liver transplant population is not dissimilar to other pediatric groups with chronic diseases as many of these have also been associated with neurocognitive impairment and school achievement issues. These findings have been attributed to multiple factors including, a disruption of development, chronic effects of the specific condition on central nervous system growth or treatment specific effects. [3, 17] Some well documented groups include childhood cancer survivors, chronic renal disease and juvenile diabetes. Among childhood cancer survivors, special education use is reported in 23% compared to 8% in their siblings.[18] Children with chronic end-stage renal disease are at increased risk for neurocognitive impairment [17], and poor metabolic control in patients with insulin dependent diabetes is related to weaker academic performance compared to siblings or matched controls.[19]

A high prevalence of special education requirements in the liver transplant population is supported by previous single centre studies examining neurocognitive outcomes and documenting mean Full-Scale score IQ (FSIQ) ranging from 84 to 94.[5, 6, 8, 20] Since FSIQ more than one standard deviation below the mean (FSIQ<86) usually requires a modified learning environment, [5] the finding of a high proportion of the SPLIT cohort using special education services is consistent with prior IQ results. Although other chronic pediatric diseases have diminished school achievement, the prevalence of special education utilization in pediatric liver transplant patients is higher and indicates a need for further studies to assess the potential influencing factors.

Other potential indicators of school problems included development of a 504 Plan for classroom accommodations/modifications and grade retention. Over one-third (36.3%) of participants required an IEP. An IEP identifies a student’s specific learning expectations and outlines how a school will address these expectations through appropriate accommodations, program modifications and /or alternative programs as well as specific instructional and assessment strategies. Among participants, only 10.9% overall had a 504 Plan, although the rate was higher for participants closer to the time of their transplant.

Parents reported that their child had been retained at least one grade in 20% of the sample, and this was more likely in the older participants. This prevalence is lower than that of the requirement for special education but that may be due to different factors such as the policy for grade retention varying between jurisdictions. Interestingly grade retention is more common in the older age group possibly because the requirements for senior matriculation are more rigorous than in earlier grades.

Identifying and describing at-risk populations is key to allowing clinicians to better counsel and assist families as well as to determine modifiable practice variables that influence school outcomes. The multivariate analysis examined the relationship between utilization of special education and multiple variables. The most striking predictor was the pre-transplant requirement for special education, OR 22.46 (<0.0001). This finding suggests that most neurocognitive deficits resulting in special education utilization originate prior to transplant. Although many pediatric liver transplant patients are very young and are not yet in school prior to transplant, these children also incurred a higher risk of using special education supports post-transplant (OR 1.70). Thus, factors related to disease and treatment prior to transplant appear to have the largest impact. In this analysis, age at LT and measures of nutritional status prior to LT were not associated with increased utilization of services. These factors have previously been associated with lower neurocognitive function in studies that have included individual patient testing. [8, 21] Reasons for this difference are not immediately apparent since this cohort did include an adequate number of children less than 12 months of age at transplant (n=166) and a large number with pre-transplant growth failure (n=239). Further studies which include individual patient testing to identify risk factors for lower cognitive outcomes coupled with more detailed anthropometric analysis to quantify malnutrition are ongoing within several pediatric liver disease research consortia.

Symptomatic CMV disease in the first six months following LT was associated with an OR of 3.1 for special education utilization. The finding of this association is novel and has not been previously described in this patient population. Congenital CMV is a well recognized cause of permanent neurological injury. Even asymptomatic congenital CMV is associated with increased rates of school failure and trends towards below-average intelligence and language development scores.[22] Although the adverse neurological effects have most commonly been associated with congenital infection, immunosuppressed patients are a population also noted to be at risk for neurological sequela.[23] This suggestion that CMV infection may have detrimental effects on long-term intellectual outcomes in pediatric liver transplant recipients warrants further study.

The multivariate model suggesting the use of cyclosporine in the early post-transplant period was associated with increased risk, OR 1.83, which was not influenced by the era of transplant. The association was even greater when comparing other immunosuppressive regimes to tacrolimus (OR 4.88), but this very small group of patients may have possessed some other confounding variable not assessed in this analysis, such as a post-transplant seizure disorder or pre-transplant neurological injury. Both cyclosporine and tacrolimus have been associated with transient neurotoxicity [24, 25] and it is generally accepted that both drugs pose a similar risk of neurological side-effects.[26] Thus, this observation should likewise be confirmed in more detailed assessments of long-term neurocognitive function.

Although the study represents the largest cohort of post-liver transplant children with prospectively collected clinical variables, the results must be interpreted within the confines of their limitations. This is a parent reported questionnaire and is not validated through school records or concurrent neuropsychological assessment. Whether or not a parent reported participation in special education depends upon their perception of what special education is. The incidence of special education requirements in this population appears to exceed that seen in other chronic disorders, but whether a child receives this type of support may be a function not only of their academic ability and performance, but also a reflection of social support and parental influence. Special educational services are regulated by a complex range of federal, state and local laws with requirements varying somewhat by state in the U.S. and by province in Canada. These criteria have been established to regulate the expenditure of these costly services, with services being granted to only those who have clearly demonstrated the need based on individual patient testing and classroom reports. In the U.S it may be somewhat easier to qualify for these services on the basis of a 504 plan, which by federal mandate provides children with chronic diseases and disabling conditions appropriate modifications within their educational program to accommodate their special needs. However, these accommodations may be implemented in a regular classroom and do not require children be placed in a special educational program and only 10.9% of the patients in this report utilized a 504 plan. Therefore, although criteria may have varied within the different geographic areas studied, it would appear that these services were granted by objective criteria and not solely on the basis of the history of having received a liver transplant. Thus, we believe the prevalence of special educational services in this report is not an over estimation of actual need. In fact, smaller single center studies that have included direct patient testing have suggested that special education needs are actually under-recognized by both families and the education system.[5, 20]

In summary, this study provides important information as it is the largest study examining the educational outcomes of post-pediatric liver transplant patients. Through this large cohort some novel variables, including immunosuppression, CMV and pre-transplant special education have been identified. But it is clear that further investigation is required to research educational attainment of children post-liver transplantation. If at least one-third of children post-liver transplantation are using special education services, then close monitoring and judicious neuropsychological and educational assessment are likely to be key to obtaining effective interventions. Evidence suggests that the most successful time for intervention is not when deficits are detected in the classroom but prior and in anticipation of academic performance deficits.[2] Understanding risk factors which identify the patient at high risk for lower school performance would assist clinicians and educators in designing pro-active programs to minimize academic performance deficits and maximize classroom success.

Abbreviations

IQ

Intelligence quotient

IEP

Individualized education plan

CMV

Cytomegalovirus

SAAPS

School Attendance and academic Performance Survey

Appendix