Keywords
moyamoya disease - stroke - pediatric - neuropsychology - cognitive dysfunction
Introduction
Moyamoya arteriopathy is a progressive cerebral arteriopathy characterized by bilateral
internal carotid artery stenosis and compensatory formation of collateral vessel networks.[1] Moyamoya is associated with up to 10% of childhood strokes and transient ischemic
attacks, and roughly one-third of children with moyamoya who experience stroke has
recurrent stroke within 1 year.[2]
[3] Moyamoya can occur in association with systemic conditions such as neurofibromatosis
type 1, sickle cell disease, or trisomy 21, termed moyamoya syndrome (MMS), or without
an associated condition, termed moyamoya disease (MMD).[4]
[5] Moyamoya is known to cause neuropsychological impairment in anywhere from 13 to
69% of affected children.[6]
[7]
[8] Prior studies are limited by exclusion or underrepresentation of children with MMS
and often do not examine domains such as social and emotional well-being and adaptive
functioning.
Currently, clinical management decisions (e.g., when to consider revascularization
surgery) tend to focus on stroke or transient ischemic attack as the only relevant
outcomes. However, one recent study of children with moyamoya found prominent executive
dysfunction associated with altered cerebral hemodynamics in the right parietal and
white matter regions, even in the absence of stroke.[9] As the first step in understanding how and when to incorporate risk of poor neuropsychological
outcome in clinical treatment decision-making, it is important to first characterize
long-term neuropsychological outcomes in children with moyamoya. Therefore, we sought
to characterize the full complement of neuropsychological impairment in children with
both MMD and MMS and to identify associations between clinical variables and cognitive
outcomes. These data will allow us to begin to understand the role neuropsychological
outcomes should play in treatment decision-making and to lay the foundation for future
treatment studies that include cognitive and behavioral function as relevant outcome
measures.
Methods
This study was approved by the Johns Hopkins Medicine Institutional Review Board and
carried out under a waiver of consent. Patients with moyamoya seen at our center from
2003 to 2021 were identified retrospectively using International Classification of
Disease codes.[5]
[10] Patients who underwent neuropsychological testing in multiple cognitive domains
were included. Medical and neuropsychological testing records were reviewed with attention
to demographic characteristics, comorbidities, occurrence of clinical and radiographic
stroke, and neuropsychological test scores.
Radiographic Data
The magnetic resonance image (MRI) most proximate to the date of neuropsychological
evaluation (either before or after) was analyzed. Clinical stroke was defined as an
acute-onset neurologic deficit with associated acute infarct on MRI. Patients were
classified as having radiographic infarct if there was any evidence of previous infarct
on MRI.
Neuropsychological Measures
Standardized, age-normed neuropsychological measures were administered as part of
clinical care. Overall intellectual functioning was assessed using either the Wechsler
Intelligence Scale for Children, Fourth Edition (WISC-IV)[11] or Fifth Edition (WISC-V)[12] full-scale intelligence quotient (FSIQ) or the Differential Abilities Scale, Second
Edition (DAS-II)[13] general cognitive ability (GCA) score based on the child's age. The WISC-IV, WISC-V,
and DAS-II FSIQ/GCA are strongly correlated (r = 0.84–0.86).[11]
[12]
[13] Scores are presented as standard scores (mean 100, standard deviation 15).
Verbal intelligence was assessed using WISC-IV/V verbal comprehension index or the
DAS-II verbal reasoning composite, which are highly correlated (r = 0.73–0.85).[11]
[12]
[13]
Nonverbal intelligence was assessed using the WISC-IV/V perceptual/fluid reasoning
index or the DAS-II nonverbal reasoning composite, which are moderately correlated
(r = 0.66–0.71).[11]
[12]
[13] For all above-mentioned WISC and DAS indices, higher standard scores denote better
performance.
Parent-reported executive functioning was assessed using either the Behavior Rating
Inventory of Executive Functioning, Second Edition (BRIEF-2)[14] global executive composite (GEC), behavioral regulation index, and metacognition
index scores or BRIEF-Preschool (BRIEF-P)[15] GEC and emergent metacognition index. BRIEF versions are well correlated (r = 0.89–0.91).[14]
[15] Scores are presented as T-scores (mean 50, standard deviation 10). Higher scores
denote greater impairment.
Parent-reported adaptive functioning was assessed using the Adaptive Behavior Assessment
System, Third Edition (ABAS-3).[16] The measure provides an overall adaptive score (general adaptive composite) as well
as three subscales: conceptual, social, and practical skills. Higher standard scores
indicate better adaptive functioning.
Social, emotional, and behavioral functioning were assessed using the composite scores
from the Behavioral Assessment System for Children, Second (BASC-2)[17] or Third (BASC-3)[18] Edition. Indices associated with the BASC include externalizing symptoms (misconduct,
aggression, defiance), internalizing symptoms (mood, anxiety), behavioral symptoms
(inattention, hyperactivity), and adaptive skills (daily living and social skills).
If a child was administered the Conners Comprehensive Behavior Rating Scale (CBRS)[19] in lieu of the BASC, the analogous scores were used. The BASC and CBRS subscales
are moderately correlated (r = 0.68–0.73).[17]
[18]
[19] Scores are presented as T-scores; higher scores indicate greater impairment.
Visual-motor integration was assessed using either the Beery–Buktenica Developmental
Tests of Visuomotor Integration, Sixth Edition[20] or the DAS-II copying subtest.[13] Higher standard scores indicate better performance.
Statistical Analysis
Statistical analysis was conducted using Stata 17.0 (StataCorp, College Station, Texas,
United States). Normality of cognitive assessment distributions was assessed using
the Shapiro–Wilk's test. Neuropsychological assessment scores were compared to age-based
population means using one-sample t-tests. Scores between the two groups (with and without clinical stroke) were compared
using t-tests. Scores among syndromic groups were compared using one-way analysis of variance.
A two-tailed p-value of <0.05 was considered statistically significant.
Results
Of 83 children with moyamoya assessed at our center, 13 had neuropsychological testing
across multiple domains. This included four children (31%) with MMD and nine (69%)
with MMS (three sickle cell disease, five neurofibromatosis type 1, and one hemophilia
A). Median age at testing was 7.8 years (range: 5.3–16.5 years). MRI took place within
a median of 43 (IQR: 19–195) days of neuropsychological testing. Six patients (46%)
had a history of clinical ischemic stroke before testing, and 10 (77%) had prior infarct
(including silent stroke) on MRI, including 4 (31%) with left hemispheric, 1 (8%)
with right hemispheric, and 5 (38%) with bilateral infarcts. Seven (54%) had undergone
surgical revascularization before testing.
Compared to age-based normative means, children in this cohort had significantly lower
overall intelligence (p = 0.003), verbal intelligence (p = 0.009), and nonverbal intelligence (p = 0.003) scores ([Fig. 1]). These children also had significantly elevated (impaired) BRIEF GEC (p = 0.005) and metacognition index (p = 0.002) scores. They demonstrated impairment in visual-motor integration (p < 0.001) and overall adaptive functioning (p = 0.015). There were no significant differences in social, emotional, or behavioral
functioning.
Fig. 1 Boxplots depicting cohort scores for test of intelligence (WISC-IV/V, DAS-II [A]), adaptive functioning (ABAS-3 [B]), executive functioning (BRIEF-2/P [C]), and social/emotional well-being (BASC-2/3, CBRS [D]). Arrow indicates direction of impairment. Dotted line represents age-appropriate
population mean. Comparison to age-based population means using one-sample t-tests. *p < 0.05, **p < 0.005. v indicates that lower t-scores indicate impairment for the adaptive skills score of the BASC-2/3. ABAS-3,
Adaptive Behavior Assessment System, Third Edition; BASC-2/3, Behavioral Assessment
System for Children, Second/Third Edition; CBRS, Conners Comprehensive Behavior Rating
Scale; DAS-II, Differential Abilities Scale, Second Edition; WISC-IV/V, Wechsler Intelligence
Scale for Children, Fourth/Fifth Edition.
We found no significant differences in overall intelligence, verbal intelligence,
or visual-motor integration scores when comparing children with versus without history
of clinical stroke (p > 0.05, [Fig. 2A]). However, children with any evidence of infarct on MRI had significantly lower
overall intelligence (p = 0.032) compared with children without radiographic evidence of infarct ([Fig. 2B]). There was a trend toward decreased visual-motor integration abilities in children
with radiographic infarcts (mean standard score 72.7, SD 16.1, p = 0.069). There were no significant differences in overall intelligence (p = 0.568), verbal intelligence (p = 0.507) or visual-motor integration (p = 0.586) among children in the MMD, sickle cell disease, and neurofibromatosis type
1 groups.
Fig. 2 Boxplots depicting differences in general intelligence (WISC-IV/V full-scale intelligence
quotient or DAS-II general cognitive ability), verbal intelligence (WISC-IV/V verbal
comprehension index or DAS-II verbal reasoning), and visual-motor integration (Beery
VMI or DAS-II copying) standard scores between patients with and without history of
clinical ischemic stroke (A) and any evidence of ischemic injury on MRI (B). Standard scores compared using Student's t-tests. Standard score of 100 represents population mean. *p < 0.05. Beery VMI, Beery–Buktenica Developmental Tests of Visuomotor Integration,
Sixth Edition; DAS-II, Differential Abilities Scale, Second Edition; MRI, magnetic
resonance image; WISC-IV/V, Wechsler Intelligence Scale for Children, Fourth/Fifth
Edition.
Discussion
In this single-center, retrospective cohort study, we found that children with moyamoya
had impairments in overall intelligence, visual-motor integration, executive functioning,
and adaptive functioning. Children with any infarct on MRI (including clinically silent
stroke) had significantly lower general intelligence when compared with those without
infarcts. This suggests that subclinical ischemic injury can affect neurocognitive
functioning in children with moyamoya. While limited by small sample size, we did
not detect differences in neuropsychological test scores among children in different
syndromic groups.
Notably, we found prominent impairment in executive functioning, consistent with previous
studies in children with moyamoya.[9]
[21] Chronic hypoperfusion in an anterior watershed distribution could contribute to
executive dysfunction by limiting blood supply to the prefrontal cortex, even in the
absence of stroke. Choi et al have reported that impaired cerebrovascular reactivity
in the right parietal and subcortical white matter regions is associated with lower
executive functioning.[9] Future studies investigating regional cortical thickness and/or using diffusion
tensor imaging to interrogate white matter integrity may further elucidate the causal
link between altered cerebral hemodynamics and changes in cognition.
Our study has some limitations. First, our small sample size limited statistical power
for comparisons. As our cohort is retrospective, selection of patients for specific
neuropsychological tests is subject to referral bias and may not be representative
of all children with moyamoya. In particular, referral for neuropsychological testing
may have been driven by neurocognitive symptoms resulting in a more severely affected
cohort. Importantly, as most children in this sample underwent neuropsychological
testing and MRI on different days, the possibility of silent strokes occurring in
the time between testing and MRI cannot be excluded.
Conclusion
In conclusion, our study showed that impairment across multiple neuropsychological
domains is common among children with moyamoya—even those without a history of clinical
stroke. Specifically, we found notable deficits in intelligence, executive functioning,
and visual-motor integration. Practitioners who care for these children should be
aware of the prevalence of neurodevelopmental differences in this group, as screening
and identification of challenges may allow for implementation of resources and accommodations
that help children with moyamoya succeed. Finally, our results support future research
that assesses the impact of different therapeutic strategies on long-term neuropsychological
function in addition to stroke risk.
