Dr Kazuhiro Haginoya at Department of Pediatric Neurology, Takuto Rehabilitation Center for Children, Yumoto Akiumachi, Taihaku-ku, Sendai 982-0241, Japan. E-mail: firstname.lastname@example.org
Aim To investigate the association between magnetic resonance imaging (MRI) patterns and motor function, epileptic episodes, and IQ or developmental quotient in patients born at term with spastic diplegia.
Method Eighty-six patients born at term with cerebral palsy (CP) and spastic diplegia (54 males, 32 females; median age 20y, range 7–42y) among 829 patients with CP underwent brain MRI between 1990 and 2008. The MRI and clinical findings were analysed retrospectively. Intellectual disability was classified according to the Enjoji developmental test or the Wechsler Intelligence Scale for Children (3rd edition).
Results The median ages at diagnosis of CP, assignment of Gross Motor Function Classification System (GMFCS) level, cognitive assessment, and MRI were 2 years (range 5mo–8y), 6 years (2y 8mo–19y), 6 years (1y 4mo–19y), and 7 years (10mo–30y) respectively. MRI included normal findings (41.9%), periventricular leukomalacia, hypomyelination, and porencephaly/periventricular venous infarction. The frequency of patients in GMFCS levels III to V and intellectual disability did not differ between those with normal and abnormal MRI findings. Patients with normal MRI findings had significantly fewer epileptic episodes than those with abnormal ones (p=0.001).
Interpretation Varied MRI findings, as well as the presence of severe motor dysfunction and intellectual disability (despite normal MRI), suggest that patients born at term with spastic diplegia had heterogeneous and unidentified pathophysiology.
• Almost 42% of patients born at term with spastic diplegia had normal MRI findings.
• The frequency of patients with severe motor dysfunction and intellectual disability did not differ between those with normal and abnormal MRI findings.
Spastic diplegia is a subgroup of cerebral palsy (CP) found both in preterm and term infants; however, the underlying causes differ. Preterm spastic diplegia is the result of periventricular leukomalacia (PVL),1 whereas only 12% of children born at term with spastic diplegia have PVL.2 In a study based on the data collected from the Surveillance of Cerebral Palsy in Europe collaboration, 45.7% of patients born at term with CP with a birthweight of 2500g or more had bilateral spastic palsy.3 The ratio of patients with CP with severe impairments peaked in the 1980s, fell during 1990 to 1999, and was little changed in the 1990s compared with the late 1970s.4 However, another recent population-based comparative analysis of children with CP born between 1990 and 1996, and between 1997 and 2003 showed an increase in those born at term with spastic diplegia from 9 (4/46) to 20% (5/25)5 respectively. The pathogenesis of spastic diplegia in those born at term is poorly understood. Although some clinical and neuroradiological studies have reported magnetic resonance imaging (MRI) findings other than PVL,1,6–9 most used small sample sizes and, thus, did not provide a broad perspective on the pathology of spastic diplegia in those born at term.
The present study investigated the relation between MRI and clinical findings in a large sample of children born at term with spastic diplegia.
Between 1 September 1990 and 31 August 2008, 829 patients with CP underwent MRI at the Takuto Rehabilitation Center for Children, Sendai, Japan. Of those, we enrolled 86 term patients (54 males, 32 females; median age 20y, range 7–42y) with clinical evidence of spastic diplegia. The inclusion criteria were a diagnosis of CP, a diagnosis of spastic diplegia, and gestational age of 37 weeks or more. Spastic diplegia was defined as a type of spastic quadriplegia affecting the lower more than the upper extremities.1,6,7 Brain MRI and clinical findings of the participants were retrospectively analysed. The present study was approved by the ethics committee of the Takuto Rehabilitation Center for Children (number 22-2).
MRI was performed using a Toshiba MRIT 50A scanner (0.5T; Toshiba, Tokyo, Japan) with 10mm multislice axial and sagittal T1-weighted spin-echo images (repetition time/echo time [TR/TE] 375/14) and axial T2-weighted spin-echo images (TR/TE 2000/120), or a Shimadzu-Marconi Magnex Eclipse scanner (1.5T; Shimadzu-Marconi, Kyoto, Japan) with 5mm multislice axial and sagittal T1-weighted images (TR/TE 500/20) and axial T2-weighted images (TR/TE 5000/72). Two authors (YN, IS-S) with no knowledge of clinical information independently evaluated the MRI findings. The MRI findings were categorized as shown in Table I and are illustrated in Figure 1. Hypomyelination was defined as a delayed or arrested normal myelinating process resulting in a persistent immature pattern of myelination in the cerebral white matter. Porencephaly was defined as fluid-filled cavities in the cerebral hemisphere connecting to the lateral ventricle. In the present study, there were no patients with porencephaly connected to the hemisphere surface. Periventricular venous infarction (PVI) was defined to the loss of periventricular tissues after venous infarction caused by subependymal haemorrhage.10 Because porencephaly and PVI are probably based on the same pathophysiology, we combined them into the one category. Cerebellar atrophy was defined as a small cerebellum with prominent fissures between shrunken folia,11 whereas cerebellar hypoplasia was defined as a small cerebellum with normally sized fissures and folia.11 When MRI revealed mixed abnormalities, the diagnosis was based on the most prevalent and dominant findings as the cause of the spastic diplegia. Furthermore, patients with abnormalities in more than one radiological category were analysed. Disagreements in diagnosis were resolved by the third author (YK) and second author (AO).
Table I. Magnetic resonance imaging classification and its distribution among patients born at term with spastic diplegia
No abnormality detected
Symmetric reduction of the peri-trigonal white matter volume with a consecutive enlargement of the posterior horns, scalloped ventricular contours, periventricular gliosis and secondary atrophy of the posterior part of the body and the splenium of the corpus callosum
Includes cortical dysplasia, schizencephaly, polymicrogyria, pachygyria, and heterotopias
Delayed or arrested normal myelinating process resulting in persistent immature pattern of myelination in the cerebral white matter
Shrinkage of cerebellar vermis and hemisphere
Diffuse cortical atrophy
Diffuse reduction of the volume of cerebral gyri
Enlargement of the bilateral lateral ventricles
Enlargement of bilateral lateral ventricles except the findings of periventricular leukomalacia
Ischemic changes in the watershed area between anterior and medial cerebral artery or between medial and posterior cerebral artery
Porencephaly/periventricular venous infarction
Fluid-filled cavities in the cerebral hemisphere connecting to the lateral ventricle/loss of periventricular tissues after venous infarction caused by subependymal haemorrhage
Thin corpus callosum
Thinning of the corpus callosum except the findings of periventricular leukomalacia and porencephary
Unable to be classified into one of the above groups
The patients’ age at diagnosis of CP, assignment of Gross Motor Function Classification System (GMFCS)12 level, and cognitive assessment as well as the history of asphyxia were retrospectively analysed from the medical records.
Motor-deficit severity was classified into levels I to V according to the GMFCS; the levels were interpreted from the latest clinical records. We found it difficult to distinguish between GMFCS levels I and II clinically; thus, those levels were combined in the statistical analysis.
The children’s developmental quotients (DQ) or intelligence quotients (IQ) were assessed using either the Enjoji developmental test, which evaluates physical abilities of the whole body, skilled hand motor activities, behaviour, interpersonal skills, speech ability, and language comprehension,13 or the Wechsler Intelligence Scale for Children (3rd edition).14 In cases of serial rating studies, we used the latest result to ensure an exact classification. The patients’ DQ/IQs were graded and classified into three groups: normal, DQ/IQ≥70; mild, 50?DQ/IQ<70; severe, DQ/IQ<50.
Epilepsy was defined as the occurrence of at least two unprovoked epileptic seizures; neonatal seizures and febrile convulsions were not included.15 The epilepsy outcome at the time of the present study was classified as good when the patient was seizure-free for more than 2 years, poor when the patient had more than two seizures per month despite appropriate treatment, and moderate if the patient was between the two classifications.
Relationships between normal or abnormal MRI findings and GMFCS level, epilepsy, and intellectual disability are shown in Table II.
Table II. Correlation between normal or abnormal magnetic resonance imaging findings and Gross Motor Function Classification System (GMFCS) level, epilepsy, and intellectual disability
The association between factors in the children’s clinical profile (motor function, epilepsy, and intellectual disability) and MRI findings was investigated. Statistical analyses were conducted using a χ2 test. A p value <0.05 was deemed statistically significant.
The median ages at diagnosis of CP, assignment of GMFCS level, cognitive assessment, and MRI were 2 years (range 5mo–8y), 6 years (2y 8mo–19y), 6 years (1y 4mo–19y), and 7 years (10mo–30y) respectively. Seventeen patients experienced birth asphyxia, including those with normal MRI (n=4), PVL (n=5), diffuse cortical atrophy (n=2), hypomyelination (n=2), porencephaly/PVI (n=1), enlargement of the bilateral lateral ventricles (n=1), thin corpus callosum (n=1), and border-zone infarction (n=1).
The MRI results of the 86 patients born at term with spastic diplegia (Table I) revealed normal findings in 36 (41.9%) patients (Fig. 1a) and abnormal findings in 50 (58.1%). We examined 60 patients and 26 patients by 0.5T and 1.5T MRI respectively. The proportion of normal findings did not significantly differ between 0.5T and 1.5T MRI (p=1.0).
Of the abnormal MRI findings, PVL was observed in 12 patients (Fig. 1b), malformation was found in three, hemispheric polymicrogyria of the left hemisphere in one (Fig. 1c), bilateral parieto-occipital schizencephaly in one (Fig. 1d), and cerebral and cerebellar malformation in one (Fig. 1e-1, e-2). Hypomyelination was observed in five (Fig. 1f), cerebellar atrophy in two (Fig. 1g), diffuse cortical atrophy in four (Fig. 1h), enlargement of the bilateral lateral ventricles in 12 (Fig. 1i), border-zone infarction in three (Fig. 1j), porencephaly (Fig. 1k)/PVI (Fig. 1l) in six, and thin corpus callosum in one (Fig. 1m). Two patients were unclassified according to the above system.
The associations between MRI findings and motor-deficit severity, epilepsy, and cognitive impairment are shown in Table II. Of the 86 patients, 61 (69%) were classified in GMFCS levels I or II. Of the 36 patients with normal MRI, 30 (83%) were graded in GMFCS levels I or II, whereas 30 of 50 patients (62%) with abnormal MRI findings were independently ambulatory (GMFCS levels I or II). These results indicate the frequency of patients with motor dysfunction did not differ between those with normal and abnormal MRI findings (p=0.053). The present study identified 25 patients in GMFCS levels III to V. Their MRI findings included normal findings (n=6), PVL (n=3), malformation (n=2), hypomyelination (n=3), cerebellar atrophy (n=2), diffuse cortical atrophy (n=2), enlargement of the bilateral lateral ventricles (n=2), border-zone infarction (n=1), porencephaly/PVI (n=2), thin corpus callosum (n=1), and unclassified (n=1). They contained five patients with abnormalities in more than one radiological category: one with cerebellar hypoplasia and pachygyria plus thin corpus callosum and enlargement of bilateral lateral ventricles; one with diffuse cortical atrophy plus enlargement of bilateral lateral ventricles; two with hypomyelination plus thin corpus callosum; and one with hypomyelination plus thin corpus callosum plus enlargement of bilateral lateral ventricles plus diffuse cortical atrophy (Fig. 1n-1, n-2). The first category was evaluated with the most dominant findings adopted as the MRI classification in Table II.
Of the 86 patients, 24 (27.9%) had a history of epilepsy, including three of the 36 patients (8%) with normal MRI and 21 of the 50 patients (42%) with abnormal MRI findings. Thus, the frequency of epilepsy was significantly lower in patients with normal MRI findings than in those with abnormal ones (p=0.001). Among the 24 patients with epilepsy, 14 (58%) exhibited good control, four (17%) had moderate control, and six (25%) exhibited poor control. The MRI findings of the six patients with poorly controlled epilepsy were as follows: porencephaly/PVI (n=2), PVL (n=1), malformation (n=1), hypomyelination (n=1), and enlargement of the bilateral lateral ventricles (n=1). MRI of one patient who developed West syndrome revealed that hypomyelination overlapped diffuse cortical atrophy and enlargement of the bilateral lateral ventricles; the patient had severe intellectual disability, a poor epilepsy outcome, and was classified in GMFCS level V.
Of the 86 patients, 52 (60.5%) had an intellectual disability. In the group of 36 patients with normal MRI findings, 19 (51%) exhibited intellectual disability, including 12 (63%) with severe disability and seven (37%) with mild disability. Similarly, 33 of the 50 (66%) patients with abnormal MRI findings had intellectually disability; 25 (76%) severely and eight (24%) mildly. These results indicate that the MRI findings were not correlated with the frequency of intellectual disability in patients born at term with spastic diplegia (p=0.266).
Other characteristic findings
Two patients showed both cerebellar hemispheric and vermian atrophy. One had no history of asphyxia, the other had no perinatal record. They were referred to our hospital because of delayed motor development and crouched posture, and were diagnosed with spastic diplegia at the ages of 2 and 9 years 7 months respectively. Both had tendon-lengthening surgery around the age of 10 years. One had intellectual disability and epilepsy. One was non-ambulatory, whereas the other one was ambulatory with crutches.
Five patients (three males, two females) whose diagnosis of spastic diplegia was made between 2 and 3 years of age had hypomyelination. Their final MRI was studied at 3, 7, 8, 10, and 11 years respectively. MRI results showed severe hypomyelination with reduced white matter volume, whereas the thalamus, brainstem, and cerebellum showed preserved myelination. Three patients had a history of birth asphyxia. Two of these showed ventricular dilatation and cortical atrophy. Four patients had severe intellectual disability. Three had epilepsy. Three patients were non-ambulatory, whereas two were able to walk independently. Auditory brainstem response, which correspond to the electrophysiological activity of the auditory system throughout the brainstem,16 was abnormal in three patients in whom the upper brainstem components were not detected; two of these patients were siblings (one male, one female), suggesting an autosomal recessive trait. The PLP1, GJC2, and myelin basic protein gene analyses were normal in one patient. No patient showed clinical deterioration.
Eight patients had asymmetric MRI abnormalities. The MRI findings in these patients were porencephaly/PVI (n=5), malformation (hemispheric polymicrogyria; n=1), cerebellar atrophy (n=1), and enlargement of the bilateral lateral ventricles (n=1). Five patients with porencephaly/PVI and one patient with malformation had clinical asymmetry, which was dominant in the contralateral extremities.
Eight patients had abnormalities in more than one radiological category, as follows: enlargement of bilateral lateral ventricles plus thin corpus callosum (n=1) or diffuse cortical atrophy (n=1); malformation, which was cerebellar hypoplasia and pachygyria plus thin corpus callosum and enlargement of bilateral lateral ventricles (n=1); hypomyelination plus thin corpus callosum (n=2) plus enlargement of bilateral lateral ventricles (n=1) and diffuse cortical atrophy (n=1); and diffuse cortical atrophy plus enlargement of bilateral lateral ventricles (n=1). They included five patients in GMFCS levels III to V, four patients with epilepsy, and six patients with intellectual disability.
The results of the present study show that the MRI findings were more varied in those born at term with spastic diplegia than those born preterm with spastic diplegia, suggesting that the pathogenesis of spastic diplegia in those born at term is heterogeneous. A recent meta-analysis involving four studies of patients with spastic CP identified 51 term-born patients with spastic diplegia.17 The MRI findings were classified as normal in 24 (47%), PVL in 13 (25.5%), brain maldevelopment in 6 (11.8%), grey matter lesions in four (7.8%), and miscellaneous in four (7.8%).17 In our study, 36 patients (41.9%) were found to have normal MRI findings, which is consistent with previous reports (Table III). Our study demonstrated that patients born at term with spastic diplegia who had normal MRI findings showed significantly fewer epileptic episodes than those with abnormal MRI findings. Patients born at term with spastic diplegia and normal MRI findings tended to show less severe motor dysfunction than those with abnormal MRI findings, although it was not significant. However, the percentage of patients with intellectual disability did not differ between the normal (51%) and the abnormal (66%) MRI groups.
Table III. Patients born at term with spastic diplegia who had normal magnetic resonance imaging (MRI) and characteristics of spastic diplegia
Number born at term with spastic diplegia
Number born at term with spastic diplegia and normal MRI findings (%)
Characteristics of spastic diplegia (n born at term/preterm) (%)
Spastic diplegia (term/preterm)
Spastic diplegia (term/preterm) with severe motor impairment
Spastic diplegia (term/preterm) with epilepsy
Spastic diplegia (term/preterm) with intellectual disability
Previous studies of patients with preterm spastic diplegia have found that the severity of motor disability was correlated with the degree of white matter reduction.6 However, an Australian population-based cohort study of CP by Robinson et al.18 reported that spastic diplegia was the most common clinical motor impairment in patients with normal MRI findings (12/25). These findings suggest the existence of unknown pathophysiological processes. In our study, six patients were classified in GMFCS levels III to V despite their normal MRI findings. Additionally, the proportion of cerebellar atrophy and hypomyelination was high in the patients in GMFCS levels III to V.
Individuals with CP and epilepsy have been reported to have structural brain abnormalities.19 Our results are consistent with these earlier reports. Previous studies have reported that 18 to 39% of patients with spastic diplegia develop epilepsy.19–22 Among them, a comparative analysis of the frequency of epilepsy in those born preterm and at term with spastic diplegia revealed that epilepsy was more common in those born at term.22 The predominance of deep white matter lesions in preterm infants is thought to be one of the factors explaining why preterm patients with spastic diplegia are less likely to develop epilepsy.22,23 We found that the frequency of epilepsy was significantly higher in patients with abnormal MRI findings than in those with normal ones.
Surprisingly, there was no difference in the frequency of intellectual disability between groups with normal and abnormal MRI findings. To the best of our knowledge, there are no reports on the relationship between severity of cognitive deficits and MRI findings in patients born at term with spastic diplegia. Reports comparing cognitive abnormalities in preterm and term-born children with spastic diplegia with PVL showed that visuoperceptual defects were common in preterm children.24 Veelken et al.22 reported that one-quarter of their children born at term with diplegia had severe intellectual disability (IQ<50), in contrast to 5% among preterm children with diplegia. In our study, the proportion of patients with intellectual disability, 60.5% (52/86) of all patients born at term with spastic diplegia, might be relatively high. Over 50% of patients born at term with spastic diplegia with normal MRI had intellectual disability. ‘Normal MRI findings’ may reflect as-yet undetected abnormalities by widely used MRI scanners. Furthermore, the use of advanced brain imaging techniques with an examination of genetic or early prenatal factors25 may be helpful in elucidating the nature of spastic diplegia in those born at term.
PVL and porencephaly/PVI are categorized as preterm-type brain injuries because they are the result of vascular system immaturity, where vascular watershed areas are situated in the periventricular white matter.26 This type of injury typically causes spastic diplegia based on the somatotopic organization of descending corticospinal tracts.26 In the present study, PVL and porencephaly/PVI were observed in 18 patients with spastic diplegia; all but one, whose perinatal medical record was not available, showed no evidence of perinatal asphyxia, suggesting that PVL and porencephaly/PVI in the term infants occurred in utero early in the third trimester. Although five patients with porencephaly/PVI and one patient with malformation had asymmetric brain lesion and more severe contralateral palsy, they showed diplegia rather than hemiplegia, suggesting the presence of undetected lesions in the pyramidal tract.
In conclusion, the present study, which to our knowledge has used the largest reported sample size of patients born at term with spastic diplegia, found that (1) 41.9% of patients had normal MRI findings, (2) there was no difference in the incidence of severe motor dysfunction and intellectual disability between patients with abnormal and normal brain MRI findings, and (3) patients with no brain abnormalities exhibited significantly fewer epileptic episodes than those with abnormal MRI findings. Spastic diplegia in patients born at term with no brain abnormalities may be the result of unidentified pathophysiology that requires further clarification. The investigation of possible genetic or early prenatal factors may provide a better understanding of the pathogenesis of spastic diplegia in those born at term.