Executive Functioning in Pediatric Solid Organ Transplant Recipients: A Meta-analytic Review

Abstract

Background.

Examining executive functioning (EF) posttransplant has become increasingly prevalent, as EF deficits are associated with poor disease-related outcomes and psychosocial functioning. The purpose of the current meta-analysis was to compare overall and domain-specific EF between healthy youth and those with a kidney, heart, or liver transplant, and identify moderating variables related to EF differences between these 2 groups.

Methods.

A literature search of PsycINFO, Pubmed, and Medline was conducted for eligible articles published until January 2019. Twenty studies met eligibility criteria and were included in the present meta-analysis.

Results.

Results from the random-effects model indicated a significant standardized mean difference in overall EF skills with transplant recipients demonstrating worse EF (g = 0.40; 95% confidence interval [CI], 0.29–0.50) than healthy youth. Specifically, transplanted youth had worse working memory (g = 0.33; 95% CI, 0.01–0.66), processing speed (g = 0.41; 95% CI, 0.19–0.62), attentional control (g = 0.53; 95% CI, 0.33–0.73), and metacognitive skills (g = 0.36; 95% CI, 0.18–0.54). Assessment type and time since transplantation were not significant moderators.

Conclusions.

Pediatric solid organ transplant recipients demonstrate worse overall EF skills and deficits in working memory, processing speed, attentional control, and metacognitive skills. Many children who have undergone solid organ transplantation will require additional support in medical and academic settings because of deficits in various EF domains.

INTRODUCTION

Survival rates of children undergoing solid organ transplantation have recently risen, with 5-year survival rates for children ages 1–17 years old being approximately 83% for heart, 97% for kidney, and 88% for liver recipients.¹ While successful transplants now allow most infants, children, and adolescents to age into adulthood,² transplant recipients often struggle posttransplant with cognitive and developmental deficits.³ One area potentially relevant to posttransplant difficulties is executive functioning (EF).⁴ Since the 1980s, more research has focused on whether transplanted youth have EF difficulties⁴ and how this may be relevant to posttransplant functioning,⁵ especially medication adherence. Higher EF, which can be considered a key component in self-management, has been associated with greater medication adherence across pediatric chronic illness samples,^6,7 including one study examining pediatric solid organ transplant recipients.⁸ However, the extant literature has yet to be aggregated to determine whether transplanted children have worse EF compared with their healthy peers and in what specific EF domains these deficits exist. Once this information is available, it can be used to further conceptual models of posttransplant functioning as well as create and implement developmentally appropriate clinical care and academic interventions.

EF is generally considered to be a multicomponent construct whose individual domains (ie, attentional control, processing speed, behavior regulation, metacognitive skills, and working memory) are needed to complete complex, higher-order, goal-directed tasks.^9–11 Attentional control (ie, the ability to focus awareness on target stimuli),¹² working memory (ie, the capacity to briefly store information), and processing speed (ie, the speed one can encode, transform, and retrieve information) are considered primary domains of EF.¹³ Behavior regulation (ie, the ability to regulate impulses, shift between cognitive sets, and control behavioral and emotional expression) and metacognitive skills (ie, the ability to organize, plan, self-monitor, and initiate tasks) are both multicomponent domains themselves.¹⁴ Although the primary domains are all considered indicators of the same latent variable of EF, they are only moderately correlated¹⁵ and are each closely tied to separate outcomes.¹⁶ It is possible that transplanted and healthy youth will differ in overall EF as well as in specific domains of EF.

Children with solid organ transplants are at risk of experiencing EF deficits due to risk factors directly related to their health status and factors such as abnormal or limited interactions with their peers, school, and community. Specifically, pretransplant and posttransplant infections, hypoxia, malnutrition, and anesthesia-related complications may directly affect cognitive deficits, whereas being unable to attend school and interact with peers in normative environments may adversely impact EF as well.^17–19 As cognitive insults and reduced opportunities for enrichment may begin with illness onset, it is likely that observed EF deficits begin to develop before transplantation with further opportunity for impairments to develop across the surgical time period. Existing literature has been mixed in regards to EF deficits in transplanted youth, with some studies finding that transplanted children have worse EF than healthy peers,^20,21 whereas others have observed no significant differences.^22,23 However, initial evidence indicates that those with EF impairments have more barriers to adherence, lower rates of medication adherence, and more academic difficulties.^8,24

The extant literature on EF in transplanted youth consists of mostly single-center samples with small sample sizes, broad age ranges,⁵ examination of only 1 organ group, and have taken the form of systematic reviews. The present review is the first to use a statistical, meta-analytic approach to aggregate existing data, and quantify the difference in EF between transplanted and healthy children and adolescents.

Given the current gap in the literature and the importance of posttransplant outcomes (eg, medication adherence) that may be related to EF, the present meta-analysis was conducted. The goals were to (1) examine differences in overall and domain-specific EF between children who have undergone a solid organ transplant and healthy children and (2) identify moderating variables that could contribute to heterogeneity in observed differences between the 2 populations. It was hypothesized that transplanted children would have overall EF and domain-specific deficits compared with healthy children. Assessment type and time since transplantation were included as exploratory moderators.

MATERIALS AND METHODS

Selection of Studies

Relevant studies were identified using 3 methods. First, a literature search was conducted in PsycINFO, MedLine, and Pubmed for articles published through January 2019. A Boolean search was conducted using all keywords from the following 3 categories: (1) executive functioning, memory, processing speed, attention, metacognition, behavior regulation; (2) pediatric, adolescent, child(ren), youth; and (3) organ transplant(ation), kidney transplant(ation), liver transplant(ation), heart transplant(ation). Some keywords were excluded from the search (ie, bowel, intestine, stem cell) to limit the number of irrelevant articles. Second, an ancestry search was conducted by examining reference lists of relevant review articles and book chapters to identify additional articles. Third, reference lists of articles that met inclusion criteria were examined for additional potential studies.

Criteria for Inclusion and Exclusion

Specific inclusion and exclusion criteria were utilized. Studies needed to meet 5 criteria to be included in the present meta-analysis. First, studies needed to include a sample of pediatric solid organ transplant recipients (ie, kidney, heart, liver, or any combination of these organs). Second, participants in each study must have been 18 years or younger. Third, studies must have included a performance task or questionnaire of youth EF. Tasks must have been completed by the youth, whereas questionnaires could be self-report or parent proxy–report. Global measures examining overall EF as well as domains (eg, processing speed) of EF were accepted. Fourth, studies must have included a healthy comparison group or have utilized a standardized measure of EF for which normative data were available. Fifth, studies were required to be empirical articles, peer-reviewed, and published in English.

Studies were excluded if a portion of the participant sample was older than 18 years and if the required information needed to calculate effect sizes for participants 18 years and younger was indiscernible. Additionally, studies were excluded if they reported EF data from a combined group of transplant recipients and those awaiting transplant. Studies were excluded if they utilized a global mental development measure due to the inclusion of domains outside of the EF realm and the inability to differentiate effect size information for EF-specific areas. Given the need to access normative data if studies did not have a healthy comparison group, studies were excluded if they utilized a measure of EF that they created specifically for that study or was not standardized.

Meta-analytic Method

All meta-analytic procedures were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.²⁵

Effect Size Calculation

Standardized mean differences of EF were examined between transplanted and healthy youth. Means and SDs were used to compute Hedges unbiased g, a measure of effect size that includes an adjustment for small sample bias.²⁶ Corresponding confidence intervals (CIs) were also calculated. Effect sizes from studies that reported information relevant to executive dysfunction, or difficulties with EF, were multiplied by −1 so that positive effect sizes indicated better EF for healthy youth compared with transplanted youth across all studies. For studies that did not provide the necessary information to calculate an effect size, corresponding authors were contacted for more information. The Metafor package²⁷ was used within the statistical computing program R to aggregate effect sizes, run moderation analyses, examine heterogeneity, and plot the data. The interpretation of the magnitude of the overall effect size was based upon Cohen guidelines for d and g, where 0.20 was considered a small effect size, 0.50 was considered medium, and 0.80 was considered large.²⁸ Larger effect sizes indicate a larger observed difference in EF abilities between youth with solid organ transplants and those without.²⁹ Positive effect sizes indicated superior EF skills for healthy youth.

Only 1 effect size for each EF domain was included for each assessment type (ie, performance task or questionnaire) and timepoint per study.²⁶ If studies contained >1 effect size per domain, then a composite score was utilized, when possible. If a composite score was not available, the effect sizes were aggregated using a weighted mean procedure.²⁶ If studies measured a domain of EF at 2 separate time points, both effect sizes were included.

Given the risk of increased type 1 error and bias in CIs found in fixed-effects models, as well as the reason to believe that the studies might be heterogeneous in nature, a random-effects model was utilized.^30–32 A random-effects model takes into account the variability beyond just sampling error present in most studies³¹ and can allow for more generalizable findings.³² Additionally, random-effects models are often preferred because they incorporate variability into analyses and will reduce to the fixed-effects model if the variance component is 0.³²

Tests of Heterogeneity

Cochran Q and I² were used to assess heterogeneity. The Q statistic examines the null hypothesis that all studies are evaluating the same effect using a χ² test with k-1 df, whereas k is the number of studies. However, a significant Q, indicating heterogeneity in the distribution of effect sizes, tends to be underpowered, especially in meta-analyses with a small number of studies.³³ Therefore, the present meta-analysis also utilized I², which describes the percentage of variation across studies that is not due solely to sampling error.³³ In accordance with Higgins et al’s guidelines,³³ values of I² were interpreted as low (25%), moderate (50%), and high (75%) levels of heterogeneity.

Moderator Analyses

Moderation analyses were conducted to test for differences in overall mean effect size based on the following variables: (1) use of performance measures of EF (eg, Wechsler Intelligence Scale for Children) and self-reports or parent proxy–reports (eg, Behavior Rating Inventory of Executive Function) and (2) time since transplantation. All studies provided the necessary information regarding assessment type to be included in moderation analyses, but 5 studies (ie, 12 effect sizes) did not report time since transplantation and were, therefore, not included.

Moderation analyses were conducted using a random-effects model with maximum likelihood estimation.³⁴ Contrast coding was used for assessment type and time since transplantation was a continuous moderator.

Potential Sources of Bias

Multiple methods were used to evaluate publication bias in the present meta-analysis. First, Rosenberg fail-safe N+ was used to calculate the number of additional studies of average sample size with a null effect that would be needed to make the overall observed effect size nonsignificant.³⁵ This is used to examine the file-drawer problem or the potential for unpublished studies with nonsignificant results. Publication bias would be indicated if the fail-safe N+ was smaller than 5n + 10, with n being the number of studies included in the meta-analysis.^34,35 Second, a funnel plot of the standardized mean differences and standard errors (SEs) was constructed, and the Egger test was evaluated.³⁶ Lastly, it is possible that inclusion of multiple effects from one study could violate the assumption of independence in the effect size calculations.^26,37 The number of effect sizes contributed by each study, as measured via a continuous variable, was therefore examined to determine whether it impacted the overall mean effect size.

Quality Assessment

An assessment of included studies was completed by using an adapted version of the National Institute of Health Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies.³⁸ Five questions specifically designed for randomized control trials (RCTs) and longitudinal studies were eliminated, resulting in 9 criteria. Two authors independently rated each study on whether they addressed (yes) or did not address (no) the criterion or if it was not applicable or not reported (other). Disagreements between raters were resolved through discussion.

RESULTS

Description of Studies and Study Participants

Study Search and Inclusion

Figure 1 depicts the screening process for studies and the specific exclusionary criteria. Of the 1329 articles initially identified through the literature search, 1202 were excluded due to irrelevant subject matter based on titles and abstracts or duplications. A total of 127 full-text articles were retrieved and thoroughly searched, with 19 studies meeting inclusion criteria. Through a search of relevant review articles and reference lists of included articles, one additional article met inclusion criteria and was added to the meta-analysis. Therefore, a total of 20 studies^{8,20,22–24,40–54} and 53 effect sizes were included in the present meta-analysis. Study characteristics are described in Table 1. Two authors of the present meta-analysis independently abstracted information to check for accuracy and completeness, and all disagreements were solved by consensus before analyses.

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram depicting the search and identification of included studies. Reproduced with permission from Moher et al.³⁹

TABLE 1.

Study characteristics of included articles

Study	N of Tx group	N of healthy group	Organ	Assess. type	Domain	Reference group	Years since Tx	Age of Tx group, y	% male of Tx group	% Caucasian of Tx group
Afshar et al40	40	–	L	Task	Working memory	Norm. Data	8.71	11.2	45	–
	40	–	L	Task	Processing speed	Norm. Data	8.71	11.2
Crowe et al20	3	–	L	Quest.	Global EF	Norm. Data	1.25	7.8	25	–
	3	–	L	Quest.	Global EF	Norm. Data	3.52	10.09
	4	–	L	Task	Processing speed	Norm. Data	1.25	7.8
	4	–	L	Task	Processing speed	Norm. Data	3.52	10.09
	4	–	L	Task	Working memory	Norm. Data	1.25	7.8
	4	–	L	Task	Working memory	Norm. Data	3.52	10.09
	3	–	L	Task	Attentional control	Norm. Data	1.25	7.8
	3	–	L	Task	Attentional control	Norm. Data	3.52	10.09
Ee et al22	13	6	L	Task	Behav. regulation	Recruited	10.89	13.08	46	–
	13	6	L	Quest.	Behav. regulation	Recruited	10.89	13.08
	13	6	L	Task	Metacog. skills	Recruited	10.89	13.08
	13	6	L	Quest.	Metacog. skills	Recruited	10.89	13.08
	13	6	L	Task	Working memory	Recruited	10.89	13.08
	13	6	L	Task	Processing speed	Recruited	10.89	13.08
	13	6	L	Task	Attentional control	Recruited	10.89	13.08
Fennell et al41	20	18	K	Task	Working memory	Recruited	.08	–	–	–
	20	18	K	Task	Working memory	Recruited	1.00	–
	20	18	K	Task	Attentional control	Recruited	.08	–
	20	18	K	Task	Attentional control	Recruited	1.0	–
Gold et al26	13	–	L	Task	Working memory	Norm. Data	–	4.45	39	46
	13	–	L	Task	Processing speed	Norm. Data	–	4.45
Gutierrez-Colina 20178	41	–	L, H, K	Quest.	Behav. regulation	Norm. Data	9.08	16.21	56.4	56.4
	41	–	L, H, K	Quest.	Metacog. skills	Norm. Data	9.08	16.21
Haavisto et al42	84	–	L, H, K	Task	Attentional control	Norm. Data	6.67	11.44	49.4	–
Johnson and Warady43	9	9	K	Quest.	Behav. regulation	Recruited	–	11.00	78	58
	9	9	K	Quest.	Metacog. skills	Recruited	–	11.00
	9	9	K	Task	Working memory	Recruited	–	11.00
	9	9	K	Task	Processing speed	Recruited	–	11.00
Kaller et al44	59	–	L	Task	Attentional control	Norm. Data	6.83	10.25	48	–
	59	–	L	Task	Working memory	Norm. Data	6.83	10.25
Kaller et al45	64	64	L	Task	Working memory	Recruited	8.92	11.90	54.7	–
	64	64	L	Task	Processing speed	Recruited	8.92	11.90
Krishnamurthy et al46	22	–	H	Quest.	Attentional controla	Norm. Data	–	15.00	64	64
	22	–	H	Quest.	Attentional controlb	Norm. Data	–	15.00
Lee et al47	43	–	K, L	Task	Attentional control	Norm. Data	9.8	12.9	46.5	–
Qvist et al48	32	–	K	Quest.	Attentional control	Norm. Data	7	9.6	69	–
Rasbury et al49	14	14	K	Task	Working memory	Recruited	.08	–	36	–
	14	14	K	Task	Attentional control	Recruited	.08	–
Sorensen et al50	133	–	L	Quest.	Behav. regulationa	Norm. Data	4.87	6.26	42	58
	130	–	L	Quest.	Metacog. skillsa	Norm. Data	4.87	6.26
	134	–	L	Task	Processing speed	Norm. Data	4.87	6.26
Sorensen et al24	92	–	L	Quest.	Behav. regulationa	Norm. Data	6.87	8.49	48	59.8
	91	–	L	Quest.	Metacog. skillsa	Norm. Data	6.87	8.49
	92	–	L	Task	Processing speed	Norm. Data	6.87	8.49
	92	–	L	Task	Working memory	Norm. Data	6.87	8.49
Stein et al51	20	11	H	Task	Working memory	Recruited	5.4	13.9	70	60
	20	11	H	Task	Processing speed	Recruited	5.4	13.9
Törnqvist et al52	133	133	L	Quest.	Attentional control	Recruited	5.7	–	55.6	–
Wray et al53	65	45	H	Task	Processing speed	Recruited	.83	9.40	40	–
Yssaad-Fesselier et al54	13	–	L	Task	Working memory	Norm data	9.2	10.5	30.7	–
	13	–	L	Task	Processing speed	Norm data	9.2	10.5
Weighted mean (SD)	64.55 (66.60)						5.58 (2.63)	10.15 (2.82)	45.26 (15.21)	58.65 (3.23)
Range	3–226						0.08–10.89	4.45–16.21	25–78	46–64

^aParent report.

^bChild self-report.

Assess, assessment; Behav, behavioral; EF, executive functioning; H, heart; K, kidney; L, liver; Metacog, metacognitive; Norm, normative; N, number of participants; Quest, questionnaire; Tx, transplant.

Study Characteristics

Among the included studies, 2 effect sizes measured overall EF, 13 effect sizes measured attentional control, 14 effect sizes measured working memory, 12 effect sizes measured processing speed, 6 effect sizes measured metacognitive skills, and 6 effect sizes measured behavioral regulation skills.

Regarding assessment type, 36 effect sizes were from performance tasks that children completed, and 17 were from self-report or parent proxy–report questionnaires.

Eighteen (90%) of the studies were cross-sectional and only examined EF at 1 time point, whereas 2 studies measured EF twice. Eleven studies did not provide measurements of EF for a healthy comparison group. For these studies, normative data was used.

The dates of publication for included articles ranged from 1983 to 2018. Fifteen (75%) of the studies were published after 2005.

Participant Demographics

EF of 886 transplanted and 399 healthy children were included across studies, with sample sizes ranging from 3 to 226. A total of 479 parents of transplanted children completed measures of their child’s EF. Table 1 displays detailed demographic characteristics of the transplant group for each study.

Of the 20 studies, 9 utilized a healthy comparison group. A total of 148 parents completed measures of their healthy child’s EF. Mean age of healthy participants ranged from 8.2 to 13.4 years (M = 10.91, SD = 1.80) and of the 4 studies that reported on sex, the average percentage of male healthy children was 50% (SD = 15%). One study reported information about the race of the healthy sample, with 58% of participants identifying as Caucasian.

Aggregated Main Effects

A forest plot of the effect sizes and 95% CIs for each study is shown in Figure 2. Results from the random-effects model indicated that transplanted children had worse overall EF skills compared with healthy youth (g = 0.40, P < 0.001; 95% CI, 0.29–0.50). The magnitude of this effect was small to medium.²⁸

FIGURE 2.

Forest plot of all effect sizes. Specifications regarding timepoints were only included for studies in which EF was assessed multiple times. See Table 1 for time since transplant for all studies. 1M, 1-mo posttransplant; 1Y, 1-y posttransplant; 3Y, 3-y posttransplant; AC, attentional control; BR, behavior regulation; CI, confidence interval; EF, executive functioning; GEF, global executive functioning; MS, metacognitive skills; PS, processing speed; WM, working memory.

Differences in EF skills between transplanted and healthy children were also observed within specific domains. Results demonstrated that transplanted children had worse working memory (g = 0.33; P = 0.04; 95% CI, 0.01–0.66), processing speed (g = 0.40, P < 0.001, 95% CI, 0.19–0.62), attentional control (g = 0.53, P < 0.001, 95% CI, 0.33–0.73), and metacognitive skills (g = 0.36, P < 0.001, 95% CI, 0.18–0.54) compared with their healthy peers. Effect sizes were small to medium for working memory, processing speed, and metacognitive skills and medium for attentional control. Results indicated no significant difference in behavior regulation skills between transplanted and healthy children.

Moderator Analyses

Moderate heterogeneity was present in overall EF analyses (Q[52] = 109.03, P < 0.001; I² = 50.65%). Therefore, moderation analyses were performed to examine sources of heterogeneity between the studies. Assessment type (β = 0.004, SE = 0.06, z = 0.06, P = 0.95, 95% CI, −0.11 to 0.12) and time since transplantation (β = 0.03, SE = 0.02, z = 1.67, P = 0.09, 95% CI, −0.01 to 0.07) failed to emerge as significant moderators.

Potential Sources of Bias

Rosenberg fail-safe N+ indicated that 1641 additional effects from studies of average sample sizes would be needed to render the overall observed effect size nonsignificant. Results indicated that the fail-safe N+ was robust, and, therefore, the overall effect was likely not subject to the file-drawer problem.

Next, a visual inspection of the funnel plot in Figure 3 revealed a somewhat asymmetrical shape, indicating potential publication bias or true heterogeneity of the sample. Results from the Egger test also suggested potential publication bias (z = 2.69, P = 0.01).

FIGURE 3.

Funnel plot of standardized mean differences and standard errors.

Lastly, moderation analyses revealed that the number of effects per study was not a significant moderator (β = 0.02, SE = 0.03, z = 0.58, P = 0.56), which indicates that the number of effect sizes per study was unlikely to contribute to any potential bias.

Quality Assessment

Figure 4 displays the results of the quality assessment of the included studies. All studies clearly defined the research objective (n = 20, 100%) and most reported detailed information on the study population (n = 19, 95%), whereas over half reported participation rates of eligible participants (n = 12, 60%) and specific inclusion and exclusion criteria (n = 13, 65%). Studies generally used validated, task-based measurements of EF (n = 16, 80%). Few studies reported justification for their sample size (n = 6, 30%) or assessed for, or controlled for, confounding variables in analyses (n = 8, 40%), and the majority of studies (n = 18, 90%) did not assess EF before transplant or multiple times posttransplant. Given that the majority of studies were cross-sectional, the low endorsement of the questions about multiple assessments of EF did not warrant exclusion of any studies.

FIGURE 4.

Quality assessment of included studies. Other denotes not applicable or not reported. EF, executive functioning. Adapted checklist from National Institute of Health. Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies.³⁸

DISCUSSION

Although EF is related to posttransplant outcomes, such as medication adherence⁸ and academic performance,⁵⁵ no previous meta-analyses have examined whether transplanted youth have EF deficits relative to their healthy peers. Even less was known about which EF domains were particularly difficult for pediatric transplant recipients. The primary aim of the present meta-analysis was to fill this gap in the literature. As hypothesized, results indicated that transplanted youth exhibit mild to moderate EF deficits compared with their healthy peers, which is consistent with previous systematic reviews.⁵⁶ The present meta-analysis extended previous findings by demonstrating that transplanted youth have specific deficits in working memory, attentional control, processing speed, and metacognitive skills.

It is likely that these deficits are related to risk factors before, during, and after transplantation. Before transplantation, these youth are likely to experience fewer stimulating or normative environmental interactions, which can affect their neurological and psychosocial development. More time spent in hospitals and out of school may also affect their developmental progression.¹⁷ From a medical perspective, pretransplant infections, abnormal metabolic function, and malnutrition are also common and are all known to negatively affect EF and cognitive development.¹⁷ During the perioperative phase, incidents such as seizures, cardiac arrest, alterations in blood flow and oxygenation, loss of cerebrovascular autoregulation, and anesthesia-related issues are also risk factors for later cognitive functioning deficits.^18,19,57 Posttransplantation infections, medication toxicity related to immunosuppressants, rehospitalizations, seizures, and strokes may all affect the child’s ability to maintain normal environmental interactions or may cause neurological insults.^17,57–59 Therefore, it is likely that these unique stressors are related to EF deficits.

The lack of a significant difference between transplanted and healthy children’s behavior regulation may be due to the heterogeneous nature of the domain or the possibility that it remains uniquely intact throughout the transplantation process. Alternatively, youth with end-stage organ disease may have learned to cope with and accommodate their behavioral regulation difficulties before transplantation, resulting in few differences between them and their same-aged, healthy peers. Additionally, both moderators failed to emerge as significant. The relative stability of EF over time⁵⁹ may explain why time since transplantation was not a significant moderator, as the ability to perform EF tasks shortly after transplant may not be significantly different from years later. Regarding assessment type, it is possible that due to the strict inclusion criteria of only using standardized measures, the assessments utilized were valid and reliable in examining EF. This finding is notable, as it suggests that future studies examining EF in transplanted children can utilize short questionnaires (eg, the Behavior Rating Inventory of Executive Function¹⁴), instead of longer batteries or tasks. It is likely that additional moderators (eg, organ type, time since diagnosis of underlying disease, number of hypoxic events) outside of the scope of the current meta-analysis are present, given that unexplained heterogeneity remains.

Results were varied regarding reporting and publication bias. The fail-safe N+ indicated that 1641 studies would be needed to render the overall effect size nonsignificant, suggesting the effect is likely not affected by the file-drawer problem. Although a visual inspection of the funnel plot and significant Egger’s test suggest potential bias, the high Q and I² values reveal that the findings may be more suggestive of true heterogeneity within the sample of studies included. This heterogeneity can be considered a potential limitation of the current research examining EF in transplanted youth.

Results of the quality assessment demonstrated a wide range in studies’ methodological rigor and the information reported regarding participants. The majority of studies did not report longitudinal EF data, which limits causal interpretations. Some studies also lacked detailed descriptions of participants, did not provide justification for sample sizes, and did not account for potentially confounding variables. Despite these limitations, all studies used validated measures of EF and provided enough information regarding recruitment and methods to provide meaningful information for the present meta-analysis. Future studies examining EF in pediatric transplant recipients should consider these quality criteria.

Implications

The findings of the present meta-analysis can be utilized to provide developmentally appropriate clinical care in medical settings.⁵⁹ Immediately upon discharge from the hospital posttransplant, adhering to a strict medication protocol is a top priority for transplant patients and their families.² The results of the present meta-analysis indicate that transplanted youth may need extra support in organizing and remembering to take their medications as prescribed due to metacognitive skills deficits. Adolescents may also have difficulty independently planning ahead, ordering more medications when needed, making physician and laboratory appointments in advance, and attending to details in long, printed medication information sheets. Clinicians can help these patients by creating a behavioral plan to increase the number and frequency of reminders to take their medications, make appointments, and fill prescriptions. Phone alarms, visual aids in youths’ bedrooms and bathrooms, and extra reminders from parents may also be beneficial. Scaffolding could also be incorporated into these plans so that children and adolescents slowly become more independent in their disease management.⁸ Physicians may also consider this information when communicating with transplant recipients about their medication protocols and the importance of adherence, as processing speed difficulties may impact patients’ ability to comprehend medication information presented verbally to them during clinic appointments. Visual reminders, including brief handouts or pamphlets, may be useful in supplementing conversations between physicians and transplanted youth. Additionally, these findings should be considered while planning for the transition of pediatric patients to adult healthcare settings, where they may receive fewer reminders.

The results of the present meta-analysis should also be considered in nonmedical settings such as schools. Previous research has found that EF difficulties are related to poor academic outcomes in typically developing children and those with attention difficulties.^16,55 It is therefore possible that transplanted youth may require additional support relative to their peers. For issues with working memory, transplanted children may benefit from additional verbal reminders from teachers, handouts of important information, and help setting up an academic planner to write down their assignments. Previous research has also found that multiple repetitions and providing immediate, direct feedback decreased the differences in selected learning tasks between children with chronic kidney disease and healthy controls.¹¹ Implementing some of these interventions may provide transplanted children with the academic support they need to be successful in school settings.

Although providing accommodations for EF deficits may be beneficial, families and clinicians may also want to consider interventions directly targeted at improving EF. Recently, there has been an increase in the number of RCTs examining the effects of computer training programs on EF in school-aged children.^60–62 Results have indicated that improvements in EF tend to be largest when baseline skills are lower, when children are in middle-to-late childhood (ie, ages 9–13), and when certain domains of EF are targeted (eg, working memory).⁶³ Although fewer RCTs have examined improvements in EF through the implementation of physical exercise and mental training programs (eg, martial arts, yoga, mindfulness training), preliminary results are promising.⁶³ Although more research is needed within the pediatric transplant population, it is likely that improving EF skills through training programs would likely decrease disease-related difficulties, such as carrying out tasks related to medication regimens (eg, remembering to refill prescriptions).⁶⁴ Future research is needed to determine when (ie, how long after transplantation), where (ie, hospital, school, or home setting), and how (ie, computer training versus exercise training programs) EF training programs could be implemented for youth posttransplantation.

Limitations

The present meta-analysis has limitations that should be considered when interpreting the results. First, despite having a comparable number of studies to other pediatric meta-analyses,^65,66 the relatively small number may limit statistical power, particularly for moderation analyses. Despite strict inclusion and exclusion criteria, there was a substantial amount of heterogeneity in the included studies that was unexplained by moderation analyses. Another limitation is that the present meta-analysis could not account for pretransplant EF, and therefore, it should not be implied that the transplantation process in isolation causes EF deficits. It seems more likely that EF deficits measured posttransplantation are an accumulation of effects from end-stage organ disease to the time of measurement. Few studies published at the time of the literature search had examined pretransplant EF, and those that did were heterogeneous across EF domains assessed, making it difficult to aggregate this information in a meaningful and scientifically sound way. It is recommended that future research examine potential changes in EF from the pretransplantation to posttransplantation periods when more studies have been published. The current study could also not evaluate domain-specific EF deficits within organ groups or underlying disease groups due to the sample size; it is possible that disease-specific sequelae may also affect EF posttransplant. It is recommended that future meta-analyses evaluate the relation between pretransplant and posttransplant EF, as well as within-organ group deficits in specific EF domains when more empirical articles become available. Although all studies met the strict inclusion criteria, it should be noted that publication years varied widely and could have contributed to heterogeneity between the studies. Specifically, medical advances in the past 40 years have contributed to different transplantation and postoperative experiences in the 1980s compared with transplantations within the past 5 years. These varied experiences may have had differential effects on EF. Lastly, although results indicated minimal to moderate differences in EF between youth with and without transplants, definitive information regarding the relationship between effect sizes and clinical value and outcomes are unknown.²⁹

CONCLUSIONS

The present meta-analysis contributes to the literature by aggregating data on EF in youth who have undergone a solid organ transplantation. Results suggest that these children have mild to moderate impairments in overall EF as well as domain-specific deficits relative to their peers and, therefore, may need additional support in clinical and academic settings. Given transplanted youths’ potential deficits that may not be apparent outside of formal assessment, it is recommended that healthcare team members incorporate screening of EF into posttransplant evaluations to identify those at risk of negative outcomes and to tailor treatments and interventions accordingly. Additionally, EF evaluations posttransplant may identify appropriate candidates for training programs that could help improve EF skills and decrease differences between their healthy peers.