Introduction
The extracorporeal membrane oxygenation (ECMO) has widely been used and developed in the past 50 years, it is recognized as a life saving therapy. However with increasing survival rates, the incidence of complications both short and long term have also increased [
1,
2]. The long-term follow-up of neuropsychological complications in neonates undergoing ECMO is a major concern, and has attracted increasing research in this field [
3]. What is clearly known is that, neonates undergoing ECMO have increased incidence of abnormal neuroimaging findings with a reported incidence between 10 -60 %, while it is still unreliable to indicate the long-term prognosis of patients with the central nervous system (CNS) complications only by imaging findings, which need long-term follow-up of neuropsychological development [
4,
5].
In addition, an analysis of extracorporeal life support organization (ELSO) of neonates that received ECMO showed the acute neurological complications. However, long-term disability of those neonates was not reported, and follow-up time was mainly shorter than 5 years [
6]. A number of factors contribute to the poor prognosis of ECMO-treated neonates, including confounding factors such as the severity of disease, potential developmental abnormalities, the length of time on ECMO, and the irrational use of anticoagulant and procoagulant medications during the onboarding process.
Thus, neonates, who have survived an extremely critical condition requiring ECMO, need structured neuropsychological follow-up. The complications of this serious disease may not appear immediately, while they may gradually develop after several years until adolescence, because the children’s health and development are expected to improve. Follow-up programs should concentrate on routine pediatric care, pay attention to the potential and acquired conditions of specific diseases, structured ECMO-neuropsychological care, including school performance, and parental education and support [
7,
8].
Meanwhile, previous studies have shown that neuropsychological development would be significantly affected after premature puberty or major congenital abnormalities in full-term infants. It was reported that prepubertal and adolescent survivors were significantly affected in the areas of exercise, intelligence, hearing, vision, attention, and memory. However, follow-up of such patients has mainly been limited to single- or multi-center observational studies, and there is no systematic evaluation standard for these studies at present [
9‐
11].
The present systematic review and meta-analysis aimed to assess the morbidity of long-term follow-up of neuropsychological complications in neonates receiving ECMO.
Methods
Search strategy and selection criteria
The literature search was performed by two expert researchers. The PubMed, Web of Science, Cochrane, and EMBASE databases were searched for retrieving studies published in the recent 10 years (until June 10, 2022). References of full-text articles were screened for identification of relevant studies. No language restriction was applied. The complete search strategy is presented in the
Supplementary file. This systematic review and meta-analysis was performed in accordance with the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guideline [
12] and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [
13]. Two researchers independently assessed the title and abstract of retrieved studies. Significant disagreements were resolved through discussion. The full-text of potentially eligible studies was retrieved and independently assessed for eligibility.
The title and abstract of retrieved studies were initially screened, and full-text was reviewed with consideration of the following inclusion criteria: (a) randomized controlled trials (RCTs) and quasi-RCTs or observational studies; (b) neonates who received ECMO; (c) studies that reported independent outcomes for each mode of ECMO; (d) studies that concentrated on the long-term follow-up of neuropsychological complications in newborns who underwent ECMO; (e) the follow-up period was more than 3 years. The exclusion criteria were as follows: (a) case reports, reviews, conference abstracts, animal experiments, systematic reviews, meta-analyses, etc; (b) duplicated studies; (c) studies that did not include outcomes of interest; (d) follow-up studies with data from the same center performed at different times; (e) neonates with congenital craniocerebral dysplasia. The corresponding authors were contacted to request for additional data, if necessary.
Data collection
The data were independently extracted by the two researchers from our team using a structured data extraction table. Disagreements were resolved through discussion. If necessary, it was attempted to contact the corresponding authors twice to obtain missing data or to clarify uncertainty. The data collection covered the characteristics of the study (study type, country, year of publication, number of ECMO centers, number of participants, gestational age, and weight of newborns), ECMO-related characteristics (mode and duration of ECMO), and long-term follow-up of neuropsychological complications in neonates. Complications were represented broadly as per the ELSO reporting guidelines. Authors were contacted for providing additional data, if necessary.
Assessment of risk of bias and certainty of evidence
The quality of the included studies was assessed using the Newcastle-Ottawa Scale (NOS) score. The possibility of publication bias was examined using the Begg's test. The statistical heterogeneity was evaluated using the
I2 statistics, the Chi-square test, and visual inspection of the forest plots. Heterogeneity refers to differences between the different studies included. When heterogeneity exists between studies, the results of the merger may be unreliable or the merger itself may be inappropriate. The guidelines have classified heterogeneity as follows: low-heterogeneity (
I2 = 25–49%), moderate-heterogeneity (
I2 = 50–74%), and high-heterogeneity (
I2 ≥ 75%) [
14]. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach was used to assess the certainty of evidence (GRADEpro app available online at:
https://www.gradepro.org [date of access was June 15, 2022] [
15].
Statistical analysis
Statistical analysis was performed using RevMan 5.3 (Cochrane Collaboration organization, London, UK) and Stata/SE 12.0 (StataCorp LLC, College Station, TX, USA) software. A random-effects model was used to report results. All the statistical tests were two-sided, and P < 0.05 was considered statistically significant. The sensitivity analysis was utilized to identify sources of heterogeneity. Subgroup analysis was used to perform meta-analysis.
Discussion
In this systematic review and meta-analysis, we identified 10 studies (1199 patients) that described the long-term neuropsychological complications in neonates undergoing ECMO. The evidence mainly consisted of observational studies with a heterogeneous study design to assess neuropsychological development. Meta-analysis of the 10 studies (1199 patients) included 9 aspects of long-term neuropsychological complications in neonates undergoing ECMO. Reports on outcomes other than morbidity were scarce and inadequate for meta-analysis.
Of the included studies, 80% (8/10) were from the Netherlands where there were several high-volume ECMO centers and a sounder follow-up system [
27,
28] and , followed by Australia and the United States accounting for 10% (1/10) of the total rate, respectively. There were 5 (50%, 5/10) single-center studies, and the remaining (50%, 5/10) were multi-center studies. Although most of the studies were from the same country, there was no overlap in data from each of our studies.
In order to evaluate the neuropsychological complications in neonates undergoing ECMO, different institutions use different assessment methods, including the Pediatric Quality of Life Inventory (PedsQL), the Movement Assessment Battery for Children (MABC), the Beery-Buktenica Developmental Test of Visual-Motor Integration (Beery VMI), the Wechsler Intelligence Scale for Children (WISC-III-NL), Subtest Digit Span of the Wechsler Adult Intelligence Scale, the Revised Amsterdam Intelligence Test (RAKIT), the Bourdon-Vos test, the Bayley Scales of Infant and Toddler Development, the Child Behavior Checklist (CBCL), the Trail Making Test, the Stroop Color-Word test, the subtests Rebus Learning and Auditory Comprehension of the Kaufman Intelligence Test, the Rey Auditory Verbal Learning Test (RAVLT), and the Rey Complex Figure Test (RCFT). Most of the evaluation methods are generally recognized, and they are mainly used in the assessment of children's neuropsychological growth and development. Although the scoring rules of different scales are different, the main contents of their evaluation are consistent, including motor, language and social adaptability. These can be reflected in the nine items studied in our study. Some of them have been rarely reported and are even subjective, inevitably leading to unreliable results were excluded [
16‐
29].
As mentioned above, long-term neuropsychological complications of neonates undergoing ECMO included intelligence, learning, motor activity, hearing, vision, cognition, behavior, and attention. These complications have serious effects on children's neuropsychological development and even on their quality of life [
30]. Moreover, neuropsychological complications become more severe over time. These neuropsychological complications are a burden not only for the individual but even for society. For such children, we need to take more careful care until they become adults, which requires more investment in money and time. Even as adults, we still need to give some people special care. However, these efforts are worth it, because these people will also create different values for the society. More importantly, through the ECMO technology, more people have the chance to survive, so that they have the right to life.
The morbidity rates of different complications were statistically determined in the present study. Although a series of analyses were performed on neuropsychological complications in neonate undergoing ECMO in the included studies, some of studies did not provide insufficient evidences, which led to difficulties to find the reasons in data analysis, and restricted the generalization of more detailed results. Studies on the neuropsychological complications of neonatal CDH in neonate undergoing ECMO accounted for the majority. The pooled morbidity of hearing in neonatal CDH was 11.1% (95%CI: -0.05-0.27, I2: 68.1%, P=0.08) and the pooled morbidity of motor activity in neonatal CDH was 22.5% (95%CI: 0.10-0.35, I2: 0%, P=0.90). Meanwhile, we could summarize the morbidity of complications of VA-ECMO and the morbidity of motor activity in different follow-up periods. It can be seen from the results of VA-ECMO that the general trend is that with the extension of time on ECMO, the incidence of neonatal motor activity complications is increasing.
Meanwhile, the long-term prognosis in neonates undergoing ECMO is associated with the occurrence of short-term complications. According to the 2017 ESLO report, from 2013 to 2017, the CNS bleeding occurred in 12% of ECMO neonates who received respiratory support, and 39% of them survived. The CNS infarction occurred in 4% of neonates, and 33% of whom survived [
31]. However, in contrast, for adult patients undergoing VA-ECMO, previous meta-analyses have estimated a cumulative rate of 13.3% for all reported neurological complications and 5.9-7.8% for ischemic and/or hemorrhagic stroke [
32,
33].
Numerous factors influence neuropsychological complications in neonates undergoing ECMO. It is essential to determine how to collect detailed data before, during, and after ECMO for further analysis of neuropsychological complications in neonates. These detailed data should not only include the basic diseases, but also the feeding of newborns, intravenous administration of nutrients and nutrient components, drugs, and growth and development of neonates. Moreover, as a complex interplay among different factors associated with the underlying disease, pharmacological and non-pharmacological treatments and “iatrogenesis” are likely to determine a child’s neuropsychological outcome [
34].
There are some strengths in this study. On the one hand, this review is a rigorous search, independent screening, as well as being the detailed reporting of methods and findings. On the other hand, this meta-analysis is one of the few studies that quantitatively analyzed the long-term neuropsychological complications associated with ECMO in neonates, which is more intuitive and clearer. In addition, due to data limitations, the long-term neuropsychological complications of neonatal CDH and neonates undergoing VA-ECMO were analyzed separately. Sensitivity analysis was used to analyze the heterogeneity of data, which could lay a foundation for the future research.
However, this study has several limitations. First, only observational studies were involved in the quantitative analysis. Second, the number of participants in each study who were included in the systematic review was uneven. Although the overall number was sufficient, the number of participants who could be included in each study was very small. Third, the included studies used a variety of outcome measures at different time points and could not be combined in the meta-analysis. Because the results of the reports that used the same test were different, the level of details was also different. As a number of the included studies did not provide more detailed data, it was impossible to extract the basic and effective data for analysis, which made it difficult to summarize the results and to identify the neuropsychological complications of a certain type of neonatal diseases. Children with severe congenital and acquired illness requiring prolonged hospitalization and intensive care likely suffer morbidity in the domains described, regardless of ECMO status. However, what cannot be gleaned from this study is the degree to which ECMO is a cause of these complications independently.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.