Childhood outcomes after low‐grade intraventricular haemorrhage: A systematic review and meta‐analysis

To undertake a systematic review and meta‐analysis exploring school‐age neurodevelopmental outcomes of children after low‐grade intraventricular haemorrhage (IVH).

Globally, over 10% of infants are born preterm, making preterm birth the foremost cause of death and disability. 1,2ntraventricular haemorrhage (IVH) is a common neuropathological consequence of preterm birth. 3,4Unfortunately, despite many advances in neonatal care-which have dramatically reduced mortality for the population born preterm-the incidence of IVH has remained static. 4,5mongst infants born very preterm (less than 32 weeks' gestation) the incidence of IVH is reported to be in the region of 20%. 3,4,6The neurological sequelae of severe IVH (grade 3 or 4) are widely recognized. 6,7[6]8 There is now, however, increasing evidence that even low-grade IVH results in an increased risk of neurodevelopmental impairment at 2 to 3 years of age, above and beyond the risks posed by prematurity alone. 6,7arents have highlighted that the longer-term functional impact of brain injuries such as IVH is a priority area for further research. 9,10Although there is growing evidence of an impact of low-grade IVH on development in infancy, 7 this does not necessarily translate into an ongoing impact throughout childhood.Early developmental assessment is widely acknowledged to be poorly predictive of future childhood functioning. 11,12Additionally, neurodevelopmental impairment is often not fixed: it can deteriorate or improve over time, influenced by the neuroplasticity of the early brain and external environmental factors.As such, there are unanswered questions around the developmental trajectories of children with a history of low-grade IVH.This information is essential to inform developmental follow-up and provision of early intervention services, in addition to preparing affected families for the future.We therefore undertook a systematic review with the aim of exploring school-aged childhood outcomes after low-grade IVH.

M ET HOD
The systematic review was conducted as per our a priori registered protocol (CRD 42021278572) and in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.Observational comparative studies, published between 2000 and 2021, exploring neurodevelopmental outcomes of school-aged children after IVH grades 1 and 2, compared to those of children born preterm without IVH, were included (Table 1).Only cohort study designs were included because of the nature of the review question and with a view to seeking the highest level of evidence available.Therefore, intervention studies, cross-sectional studies, case-control studies, and case series were excluded.A composite outcome of neurodevelopmental impairment, however defined by study authors, was the primary study outcome.Secondary outcomes included cognitive impairment (including poor academic attainment and special educational needs), motor impairment (including cerebral palsy), speech and language deficits, behavioural outcomes, visual impairment, and hearing impairment.
An extensive search strategy was developed in Medline Ovid which consisted of 99 key terms and MeSH headings (Appendix S1).This was adapted and employed across 10 databases (Appendix S2).In addition to searching the published and grey literature, searches were augmented by hand searching the reference lists of included studies.All identified abstracts were screened for relevance independently, by two reviewers.The full texts of potentially relevant abstracts were retrieved and again reviewed for inclusion by two trained reviewers independently.This included an assessment of studies' risk of bias as a result of population selection, the comparability of the exposed and comparator groups (with IVH grade 1 or 2 and without IVH), and outcome assessment, using the Newcastle-Ottawa Scale for cohort studies. 13Disagreements were resolved through third reviewer arbitration.
Key data from included studies were extracted to a purpose-built Microsoft Excel spreadsheet by two reviewers independently.Results from studies were narratively synthesized and described for each developmental domain.Where possible, data from suitably clinically and contextually comparable studies were pooled in random effects meta-analyses using RevMan 5.4. 14Dichotomous study data were combined using the Mantel-Haenszel method and continuous data were combined using the inverse variance method.Heterogeneity was assessed using the I 2 statistic. 14Caution was advised in interpreting meta-analyses where there was substantial (>85%) heterogeneity across studies, and, where possible, heterogeneity was explored through subgroup analyses.
The 12 studies included over 2036 infants with IVH grade 1 to 2 and 53 067 comparator infants born preterm without IVH (five studies contained a partially overlapping population).Infants were mostly born at less than 33 weeks' gestation or had a birthweight of less than 1500 g.IVH grade 1 to 2 was mainly confirmed on ultrasound imaging (n = 11) and occasionally on magnetic resonance imaging (MRI) (n = 2).The Papile classification was used by most studies to classify the severity of IVH (n = 8) based on the most severe IVH observed on imaging (whether unilateral or bilateral).Images were reviewed and classified by neonatologists (n = 3), radiologists (n = 2), and sonographers (n = 1), although most studies did not specify who reviewed the imaging and whether imaging review was blinded.Studies used 30 different neurodevelopmental assessment tools to determine outcomes at varying time points.

Neurodevelopmental impairment
Four studies explored neurodevelopmental impairment at school age of infants born preterm with IVH grade 1 to 2. [15][16][17][18] These studies reported conflicting results and were not comparable (Table S1 and Table 2).

What this paper adds
• The functional impact of low-grade intraventricular haemorrhage (IVH) at school age is unknown.• Low-grade IVH is associated with a lower IQ at school age.• The risk of cerebral palsy is increased after lowgrade IVH.• Low-grade IVH is not associated with speech and language impairment.

Cognitive
0][21][22] These studies used 17 different cognitive assessment tools and reported conflicting results for children born across a wide time-period (1983-2005).Two comparable studies highlighted a significant pooled mean difference in IQ scores of 4 points at 8 years of age between those with and without IVH grade 1 to 2 and heterogeneity was low (−4.23,95% confidence interval [CI] -7.53, −0.92,I 2 = 0%) (Figure 1). 21,22These studies were of low and moderate risk of bias.This was largely because of the risk of bias stemming from how the exposed and comparator cohorts were selected by Wy et al. 22 and in the ascertainment of low-grade IVH (Table S3).Although this difference was attenuated and not statistically significant on including only those with left-sided IVH grade 1 to 2 (−2.96, 95% CI -6.61, 0.7, I 2 = 0%) and equivocal on including only those with right-sided IVH grade 1 to 2 (−0.12, 95% CI -5.89, 5.66, I 2 = 50%) from Vollmer et al. 21in sensitivity analyses (Figure S2 and S3).Wy et al. 22 did not report outcomes by laterality.
Van de Bor and den Ouden 17 also highlighted a cognitive impact of IVH grade 1 to 2 with 22.5% having special educational needs at 5 years of age compared to 8.7% of comparators.These needs persisted at 9 and 14 years of age after adjusting for key confounders (adjusted OR [aOR] 2.1, 95% CI 1.01, 4.35). 17On the other hand Wy et al. 22 (after adjusting for covariates) and Hollebrandse et al. 19 which included children born more recently than Van de Bor and den Ouden 17 reported no increased risk of cognitive impairment.Hollebrandse et al. 19 also did not find any increased risk of impaired academic performance in reading, spelling, or arithmetic (OR 0.7, 95% CI 0.39, 1.27).

Inclusion criteria Exclusion criteria
Prospective Motor impairment, as defined by authors (including direct testing, clinical record review, and reporting) Visual-motor impairment, as defined by authors (on direct testing) Emotional-behavioural difficulty, as defined by authors (including direct testing, clinical record review, and parental reporting) Speech and language impairment, as defined by authors (on direct testing) Visual impairment, as defined by authors (including direct testing, clinical record review, and parental reporting) Hearing impairment, as defined by authors (including direct testing, clinical record review, and parental reporting) T A B L E 2 Overview of key findings for school age outcomes of infants with perinatal brain injury compared to those without brain injury.IVH grades 1 and 2 a Four studies [15][16][17][18] Not comparable Small studies with conflicting results Six studies 16,17,[19][20][21][22] Two comparable studies in meta-analysis Conflicting results

Motor and cerebral palsy
21]23 An increased combined crude risk of cerebral palsy at school age after IVH grade 1 to 2 was reported (OR 2.92, 95% CI 1.95, 4.37, I 2 = 41%) with moderate heterogeneity (Figure 2).There was clinical heterogeneity in terms of outcome assessment (timing and method) which may account for the moderate statistical heterogeneity.Two of these studies had a low risk of bias and one had a moderate risk of bias because of the poor comparability of the exposed and comparator cohort and high loss to follow-up in the absence of a clear explanation (Table S4).Sherlock et al. 16 highlighted similar rates of abnormal movement scores amongst those with previous IVH grade 1 to 2 (n = 17, 23.6%) compared to those without IVH at 8 years of age (n = 39, 22.5%). 16However, Hollebrandse et al. 19 reported a stepwise increase in the prevalence of motor dysfunction, defined as a low movement ABC score (below the 5th centile) or cerebral palsy, from those with no IVH (n = 81, 24%), to those with IVH grade 1 (n = 22, 28%) and those with IVH grade 2 (n = 20, 38%).This was statistically significant for those with IVH grade 2 (OR 1.93, 95% CI 1.05, 3.56) but not those with IVH grade 1 (OR 1.17, 95% CI 0.67, 2.03) (Table S1). 19Vollmer et al. 21

Speech and language
Four studies explored speech and language outcomes at school age after low-grade IVH. 16,17,19,22They consistently reported no significant difference in speech and language between infants born preterm with low-grade IVH and infants born preterm without IVH at school age (Table 2).Sherlock et al. 16 S1).Additionally Van de Bor and den Ouden 17 did not find any significant differences in the incidence of speech disability at 5 years of age after IVH grades 1 and 2 (n = 12, 26.6%) compared to those without IVH (n = 34, 15.7%) (Table S1).Wy et al. 22

Behaviour
Two studies explored behavioural outcomes at school age. 22,25They did not report any significant differences on the Child Behaviour Checklist between those with and F I G U R E 1 Forest plot reporting the mean difference in IQ score at 8 years of age after intraventricular haemorrhage (IVH) grades 1 and 2. Abbreviation: CI, confidence interval.

F I G U R E 2
Forest plot reporting the crude risk of cerebral palsy after intraventricular haemorrhage (IVH) grades 1 and 2. Abbreviation: CI, confidence interval.

Hearing and vision
Three studies explored hearing outcomes after low-grade IVH and four explored visual outcomes. 17,20,21,26These studies were not comparable and the outcome of hearing or visual impairment was often too rare to enable inferential analysis.However, Kaur et al. 26 20 No significant difference in visuomotor integration scores after IVH grade 1 to 2 was reported. 20,21

DISCUS SION
This review adds to the growing evidence that low-grade IVH is not a benign pathology by reporting on its association with specific persisting neurodevelopmental impairments at school age.Evidence of an increased composite risk of neurodevelopmental impairment after low-grade IVH was conflicting.There was however evidence of an impact of low-grade IVH on IQ at school age-although the impact of this on school performance and the functional importance of a 4-point reduction in IQ was less clear.There was also a demonstrably increased risk of cerebral palsy after low-grade IVH and some suggestion of an impact on motor performance outside of cerebral palsy.Promisingly, studies consistently reported no increased risk of speech and language impairment after low-grade IVH and no increased risk of behavioural impairment was found.There was a suggestion of increased risk of visual impairment after IVH grade 1 to 2, although few studies addressed this.

Strengths and limitations
This systematic review included a sensitive search strategy, covering an extensive time-period, and a rigorous review and appraisal process was followed.Data synthesis and review findings were limited by the heterogeneity of included studies.The 12 included studies employed 30 different neurodevelopmental assessment tools at varying time points.Data extraction was limited by the presentation of results within studies.For example, disaggregated outcome data for those with IVH grade 1, IVH grade 2, and no IVH were typically not presented.Additionally, outcomes were often not presented by laterality of injury.Most studies used ultrasound imaging to diagnose low-grade IVH despite concerns about interrater variability in the classification of IVH severity on ultrasound.Additionally using ultrasound imaging over MRI poses a risk of not detecting additional confounding lesions such as white matter injuries which may affect outcomes. 4,8Although we required the study to have a non-brain injured preterm comparator group for inclusion, many studies did not account for other key confounders including comorbidities such as retinopathy of prematurity and necrotizing enterocolitis.

Context of literature
A previous review reported an increased adjusted risk of moderate to severe neurodevelopmental impairment after IVH grade 1 to 2 at 2 to 3 years of age (aOR 1.35, 95% CI 1.05, 1.75). 7Unfortunately it was not possible to determine whether this risk persists to school age.Previous reviews of this topic have also been limited by heterogeneity across studies. 6However, we were able to report that the cognitive impact associated with low-grade IVH at age 2 to 3 years of age (an increased risk of Bayley Scale of Infant Development scores <70) may persist at school age in the form of reduced IQ scores at 8 years of age. 21,22However, it was unclear whether this translated into a functional impact on school performance. 17,19,22nsurprisingly, the previously reported increased risk of cerebral palsy at 2 to 3 years after low-grade IVH was also seen at school age. 727][28][29][30][31][32][33][34] There are many postulated mechanisms through which grade 1 to 2 IVH may affect neurodevelopment including through direct (primary) damage to the germinal matrix or through secondary damage by impacting adjacent brain development (e.g.cerebral white matter injury). 8,35After 24 weeks' gestation the germinal matrix plays a key role in the development and migration of glial precursor cells and GABAergic interneurons to the cerebral cortex and thalamus. 4,8,36Disruption to these processes by IVH and its deleterious effects could impact neurodevelopment especially motor and cognitive development. 4,35,36

Implications
This review provides further evidence that low-grade IVH is unlikely to be a benign pathology (as previously thought) by highlighting that it has a measurable impact on children's cognitive and motor development at school age.These findings, including the outstanding gaps in our knowledge, should inform counselling of parents of infants born preterm with low-grade IVH on the neonatal unit.The findings also provide further evidence to support the importance of ongoing quality improvement initiatives to prevent IVH amongst infants born preterm on the neonatal unit in addition to reiterating the importance of neurodevelopmental follow-up and support for this population.Further studies which are adequately powered to explore school-aged outcomes after low-grade IVH, which adjust for key confounders (including white matter injury), and report outcomes by laterality and specific grade of IVH, are needed to more fully understand the implications of low-grade IVH for affected children.Additionally, use of a core outcome set and more transparent reporting would reduce heterogeneity across studies enabling powerful meta-analyses which could efficiently address these outstanding questions.
Executive dysfunction, as defined by authors (direct testing) Low numeracy, as defined by authors (by direct testing or educational achievement tests) Low literacy, as defined by authors (by direct testing or educational achievement tests) Special educational needs as defined by authors (school or parental report) Primary outcome(s)Neurodevelopmental impairment, as defined by authors (including direct testing, clinical record review, and parental interview/survey)Secondary outcome(s)Any cognitive impairment, as defined by authors (direct testing) Mild cognitive impairment (IQ or developmental quotient 1-2 SD below the mean) Moderate-severe cognitive impairment (IQ or developmental quotient more than 2 SD below the mean) 22ported no significant mean difference in Child Behaviour Checklist internalizing (0.91, 95% CI −0.76, 2.58), externalizing (−0.14, 95% CI −2.19, 1.90), or total raw scores (0.56, 95% CI −4.60, 5.72) between infants born preterm with low-grade IVH and infants born preterm without IVH at 8 years of age.22