Outcome of hemiplegic cerebral palsy born at term depends on its etiology

Brain & Development xxx (2015) xxx–xxx www.elsevier.com/locate/braindev

Original article

Outcome of hemiplegic cerebral palsy born at term depends on its etiology Yukihiro Kitai a,⇑, Kazuhiro Haginoya b, Satori Hirai a, Kayo Ohmura a, Kaeko Ogura a, Takehiko Inui b, Wakaba Endo b, Yukimune Okubo b, Mai Anzai b, Yusuke Takezawa b, Hiroshi Arai a b

a Department of Pediatric Neurology, Morinomiya Hospital, Japan Department of Pediatric Neurology, Takuto Rehabilitation Center for Children, Japan

Received 28 July 2015; received in revised form 14 September 2015; accepted 14 September 2015

Abstract Objectives: To elucidate the etiology and its relationship to the outcomes of hemiplegic cerebral palsy (HCP). Participants and methods: MR images and outcomes of 156 children with HCP born at term and older than three years were investigated in two major centers for cerebral palsy in Japan. Etiologies were classified into perinatal ischemic stroke (PIS), cerebral dysgenesis (CD), and others. PIS was divided into periventricular venous infarction (PVI) and two types of arterial infarction; middle cerebral artery infarction (MCAI) and deep gray matter infarction (DGMI). Initial signs and the time of presentation were investigated among the three types of PIS. As functional outcomes, laterality of paresis, age at initial walk, affected hand’s function, intellectual development, and occurrence of epilepsy were compared among all the four types. Results: Etiology: PIS was found in 106 children (68%), while CD accounted for 28 (18%). Among PIS, venous infarction was more common than arterial infarction (62:44). Outcomes: PVI revealed later presentation of motor asymmetry and more involvement of lower extremity as the initial sign among PIS groups. Only MCAI showed right-side predominance in laterality of paresis. DGMI related to better intellectual development and PVI showed lower occurrence of epilepsy, while there was no significant difference in affected hand’s function among the four groups. PIS groups showed significantly earlier attainment of independent walk, better intellectual development, and lower occurrence of epilepsy than CD. Conclusions: PVI was the most common cause of HCP born at term, and the etiology closely related to the initial signs of hemiplegia and overall outcomes. Ó 2015 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Keywords: Hemiplegic cerebral palsy; Presumed perinatal ischemic stroke; Periventricular venous infarction; Arterial infarction; Etiology; Outcome

1. Introduction Hemiplegic cerebral palsy (HCP) is a heterogeneous disease with various developmental outcomes, caused ⇑ Corresponding author at: Department of Pediatric Neurology, Morinomiya Hospital, 2-1-88, Morinomiya, Joto-ku, Osaka City, Osaka 536-0025, Japan. Tel.: +81 6 6969 0111; fax: +81 6 6969 8001. E-mail address: [email protected] (Y. Kitai).

by factors such as perinatal infarction, cerebral malformation, and infection. The most common cause of HCP is perinatal ischemic stroke (PIS), which is classified by the time of insult into fetal, neonatal, and ‘presumed perinatal’ [1,2]. Presumed perinatal ischemic stroke (PPIS) defines term-born children older than 28 days with normal neonatal neurological history, presenting with neurologic deficit or seizure referable to focal,

http://dx.doi.org/10.1016/j.braindev.2015.09.007 0387-7604/Ó 2015 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007

2

Y. Kitai et al. / Brain & Development xxx (2015) xxx–xxx

chronic infarction. Its diagnosis is made by CT or MRI performed after the neonatal period, revealing a lesion consistent with remote ischemic or hemorrhagic stroke [3]. It is presumed (but not certain) that the ischemic event occurs sometime between the 20th week of fetal life and the 28th postnatal day [2]. According to recent studies, not only arterial but also venous infarction of the periventricular region may cause HCP in termborn children [4,5] and PPIS [3]. The objectives of this study are to investigate the ratio of arterial and venous infarction in a population of HCP and to compare functional outcomes between them and with those of children with cerebral dysgenesis (CD). 2. Participants and methods

Morinomiya Hospital and Takuto Rehabilitation Center for Children. 2.2. Neuroimaging analysis MR images of participants were assessed by two experienced pediatric neurologists, one per each hospital, unblinded to functional outcomes. They classified abnormality on MRI into PIS, CD, and others. PIS was divided into periventricular venous infarction (PVI) and arterial infarction, which consists of middle cerebral arterial infarction (MCAI) and localized ischemic lesions in deep gray matter (DGMI). Typical MR images of arterial and venous infarction are shown in Fig. 2. The diagnosis of PVI was based on Kirton’s criteria [3].

2.1. Participants 2.3. Initial signs and functional outcomes Participants were HCP children recruited from outpatients of Morinomiya Hospital and Takuto Rehabilitation Center for Children, both of which are regional center hospitals for children with cerebral palsy in Japan. Inclusion criteria were (a) born at term (>36 weeks of gestational age), (b) older than three years at the time of investigation, and (c) with precise perinatal information on chart and brain MRI after the neonatal period (>28 postnatal days). All participants were Japanese. We excluded those who were diagnosed with asymmetrical diplegia or triplegia, and those who had hemiplegia caused by insults after the neonatal period. Finally, we selected 156 children, 97 boys and 59 girls, out of 251 children with HCP. A complete description of the selection of participants at each hospital is provided in Fig. 1. Most of them received weekly or monthly rehabilitation according to the severity of paresis, but none of them received constraint-induced movement therapy, hand-arm bimanual intensive training, or botulinum-A toxin injection to the affected hand. This study was approved by the ethical committees of Morinomiya Hospital

Takuto Rehabilitation Center

(From 2006 to 2012)

(From 2007 to 2012)

All Cerebral Palsy (n=2049) Asymmetrical Cerebral Palsy (n=238)

All Cerebral Palsy (n=665) Asymmetrical Cerebral Palsy (n=85)

We surveyed the patient characteristics (the number, sex, perinatal history, and laterality of paresis) of each group. Among PIS (PVI, MCAI, and DGMI), the detailed information about initial signs of HCP and the time of presentation were also investigated. Initial signs included apparent symptoms or signs of stroke during the perinatal period, motor asymmetry of extremities, and seizures. Dominant extremities of initial signs of PPIS (upper, lower, and both) were classified based on the topography of motor asymmetry their parents had first noticed. As to functional outcomes, we compared the age at initial walk, the affected hand’s function, intellectual development at the time of investigation, and occurrence of epilepsy among the four groups (PVI, MCAI, DGMI, and CD). We classified the affected hand’s function into four grades according to its best performance: pinch, grasp-release, press (e.g. press a sheet of paper to stop it from moving when writing), and disuse (totally neglected). We evaluated intellectual development by three grades: normal or mildly retarded (DQ or IQ above 70), moderately retarded (DQ or IQ between 30 and 70), and severely retarded (DQ or IQ below 30). DQ was evaluated by the Kyoto Scale of Psychological Development, and IQ was assessed by the third edition of Wechsler intelligence scale for children.

Exclusion (n=62, 10) Asymmetrical diplegia or triplegia Insufficient information (MRI, clinical history)

Hemiplegic Cerebral Palsy with sufficient information (n=176)

Hemiplegic Cerebral Palsy with sufficient information (n=75)

Exclusion (n=73, 22) Preterm birth (< 37 weeks) Younger than 3 years

Participants of this study (n=103)

Participants of this study (n=53)

Hemiplegic Cerebral Palsy cases born at term and older than 3 years (n=156)

Fig. 1. Selection of participants.

2.4. Statistical analysis The mean ages at initial walk were analyzed using the Kaplan–Meier estimator and compared among groups by log-rank test. The ages of first clinical symptoms were compared by Kruskal–Wallis and Mann–Whitney nonparametric test. Difference in laterality, initial symptoms, degree of the affected hand’s function, intellectual development, and occurrence of epilepsy were analyzed by chi-squared test. P < 0.05 was considered significant.

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007

Y. Kitai et al. / Brain & Development xxx (2015) xxx–xxx

3

A㧕 㧕 MCAI

B㧕 DGMI

C㧕 DGMI

D㧕 PVI

E㧕 PVI

F㧕 PVI

Arterial

Venous

Fig. 2. Typical MR images of arterial and venous infarction. MCAI usually involves cortical and subcortical MCA territory (A). LLS infarction shows greater damage to the basal ganglia than periventricular white matter, resulting in ‘cranial triangle’ on coronal image (dot-lined triangle; B). Unilateral small thalamic lesion suggests localized infarction of a penetrating artery distributed to the thalamus (thin arrow; C). The features of PVI are smooth enlargement of lateral ventricle (dot-lined curve; D) with relative sparing of cortex and basal ganglia, a ‘caudal triangle’ on coronal image (dot-lined triangle; E), and a ‘callosal notch’ on sagittal image (arrowheads; F).

3. Results

3.3. Initial signs of PIS (Table 2)

3.1. Neuroimaging analysis

A total of 17 children out of 106 with PIS presented with apparent signs of infarction during the perinatal period (non-PPIS). Five showed abnormal fetal echography, and twelve showed acute symptoms and imaging abnormality of infarction during the neonatal period. The rest of the PIS cases (89 children, 84%) revealed no apparent signs of perinatal stroke, and met the criteria of PPIS. Among PPIS, 55 (62%) were PVI and 34 (38%) were arterial infarction (23 MCAI and 11 DGMI) (Table 2-A). For PPIS, a total of 81 children (91%) showed motor asymmetry as the initial symptom. The ages at first clinical symptom were significantly different among the three types of PPIS (p < 0.01 by Kruskal–Wallis test), and children with MCAI developed motor abnormality significantly earlier than those with PVI (p < 0.01 by Mann–Whitney U test). As the initial sign, motor asymmetry including lower extremity was significantly more common in PVI (27/51, p < 0.05) and significantly less common in MCAI (5/20, p < 0.05) among the three groups (PVI, MCAI, and DGMI) (Table 2-B).

The number of cases and detailed information of each group (PVI, MCAI, DGMI, CD, and others) are described in Table 1. PIS (PVI, MCAI, and DGMI) was found in 106 children. Among PIS, PVI was more common than arterial infarction (62:44), and MCAI was twice as common as DGMI (30:14). The main lesion of DGMI was located to the basal ganglia in 10 children and to the thalamus in 4. CD was found in 28 children, equivalent to one-quarter of PIS. CD consisted of polymicrogyria (n = 13), hemimegalencephaly (5), focal cortical dysplasia (4), schizencephaly (4), pachygyria (1), and unclassified malformation (1). Others (22 children) included asymmetrical lesions from perinatal asphyxia, neurocutaneous syndrome such as SturgeWeber syndrome and tuberous sclerosis, tumor, neonatal viral or bacterial infection, and neonatal intracranial hemorrhage of unknown origin. Two children showed a normal MR image but clinically developed hemiplegia. 3.2. Gender and laterality of paresis (Table 1)

3.4. Functional outcomes (Figs. 3 and 4) PVI showed significant male predominance (p < 0.01), while arterial infarction and CD did not show significant difference between the sexes. HCP caused by MCAI showed significant right-side predominance (L:R = 7:23, p < 0.05), while the other groups showed no significant disproportion in laterality.

We summarize the functional outcomes of each group in Fig. 3. At the time of investigation, all children with PIS could walk independently, while 7 out of 28 children (25%) with CD could not. As shown in Fig. 4, the estimated age of initial walk was not significantly different

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007

4

Y. Kitai et al. / Brain & Development xxx (2015) xxx–xxx

Table 1 Characteristics of each group. n (%)

Sex M:F

Age at investigation mean, SD (years)

Gestational age mean, SD (weeks)

Laterality L:R

62 (40%) 30 (19%) 14 (9%)

43:19** 14:16 5:9

9.7, 4.6 12.4, 6.4 11.1, 5.6

39, 1.3 40, 1.3 39, 1.2

32:30 7:23* 5:9

CD

28 (18%)

18:10

8.6, 3.8

39, 1.1

11:17

Others

22 (14%)

17:5

9.3, 5.4

39, 1.1

14:8

PIS

Venous Arterial

PVI MCAI DGMI

PIS; perinatal ischemic stroke, PVI; periventricular venous infarction, MCAI; middle cerebral artery infarction, DGMI; deep gray matter infarction, CD; cerebral dysgenesis. PVI revealed significant male predominance. Right-sided paresis was significantly predominant in MCAI. * p < 0.05. ** p < 0.01.

between PVI (19 months) and arterial infarction (18 months), but both of these were significantly earlier than that in CD (48 months, p < 0.01). As for the affected hand’s function, 47 children (35%) could pinch and 106 (79%) could grasp-release. There was no significant difference among the groups concerning the rates of pinch, grasp-release, and press, but disuse was found only in CD. Normal intellectual development or mild retardation was observed in 76% of PVI, 70% of MCAI, all of DGMI, and 36% of CD. The ratio of normal development and mild retardation was significantly higher in DGMI (p < 0.01) and significantly lower in CD (p < 0.01) among the four groups. Among PVI patients, the only child with severe mental retardation had extensive PVI lesion, which was found during fetal life (nonPPIS), and revealed globally poor outcomes; she got to walk independently at the age of 83 months, couldn’t grasp-release in her affected hand, and had a history of infantile spasms. The overall occurrence rate of epilepsy was 35%. It was significantly higher in CD (60%, p < 0.01) and significantly lower in PVI (23%, p < 0.01) among the four groups. 4. Discussion 4.1. Etiology of HCP Germinal matrix-intraventricular hemorrhage (GMH-IVH) is one of the most common complications in very premature infants, and intraparenchymal hemorrhage (grade IV IVH) is now considered not as a simple extension of IVH but as a hemorrhagic conversion of venous infarction (periventricular hemorrhagic infarction, PVHI) [6–9]. A similar event might occur in utero and present as PVI in term-born children [4,5,10]. In this study, 68% (106/156) of HCP children born at term were PIS. This result is compatible with that of Wu’s population-based cohort study on congenital

hemiparesis, although she reported that perinatal arterial infarction was the most common cause in term infants [1]. Golomb also reported that all of 22 PPIS cases were of arterial origin [11]. Even in a more recent study on PPIS by Kirton, PVI was clearly less common than arterial infarction (venous:arterial = 12:47) [3]. The present result showed that PVI was more common than arterial infarction (MCAI and DGMI), and accounted for 40% of HCP children born at term. PVI also prevailed over arterial infarction in PPIS (venous: arterial = 55:34). To the best of our knowledge, this is the first report revealing that venous infarction was the most common etiology of HCP born at term, and of PPIS. One reason for the difference between Kirton’s and the present study may be the difference in genetic predisposition. A recent survey of the genetic etiology of PIS in Japan revealed that 16.4% of children with PIS had mutations of the COL4A1 gene, while such data have yet to be reported from other countries [12]. It is reasonable that those with COL4A1 mutation are predisposed to GMH-IVH and consequent PVI in utero. Therefore, high prevalence of COL4A1 mutation could explain the predominance of venous infarction as a cause of HCP and of PPIS in Japan. Another explanation is the difference in recruitment pathway: infants with PPIS detected by neonatologists vs. children with HCP recruited from regional centers for cerebral palsy. Because paresis of PVI tended to present later than that of MCAI, neonatologists might have overlooked candidates for imaging study. In addition, it might be difficult to detect PVI in countries where MRI is not as prevalent as in Japan because some PVI patients had too small white matter lesions to be detected by ultrasound or even by CT. As for DGMI, we found that thalamic artery infarction caused HCP as well as lateral lenticulostriate artery infarction. This means that insufficiency of posterior cerebral artery circulation also causes HCP. Among a relatively few studies on infarction localized to deep gray matter in children, Abels reported five thalamic artery

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007

Y. Kitai et al. / Brain & Development xxx (2015) xxx–xxx

5

Table 2 Clinical presentation of perinatal ischemic stroke. (A) Period of detection of PIS Non-PPIS

PVI (n = 62) MCAI (n = 30) DGMI (n = 14)

PPIS

Fetal

Neonatal

Presumed perinatal

5 0 0

2 7 3

55 23 11

(B) Initial signs of PPIS Age at first presentation (months)

Clinical symptoms Dominant extremity of motor asymmetry

PVI (n = 55) MCAI (n = 23) DGMI (n = 11)

11.6 5.9** 12.9

Upper

Lower or both

24 15 6

27* 5* 4

Seizure

Others or unknown

1 0 1

3 3 0

PPIS, presumed perinatal ischemic stroke. Age at the presentation of initial signs in MCAI (5.9 months) was significantly earlier than that in PVI (11.6 months). Motor asymmetry of lower extremity as the initial sign was detected significantly more in PVI (27/51) and significantly less in MCAI (5/20). * p < 0.05. ** p < 0.01.

infarctions and eight lateral lenticulostriate artery infarctions out of 28 neonatal perforator strokes in 24 infants [13]. 4.2. Laterality of hemiplegia and initial signs of PPIS As for laterality of paresis in perinatal arterial stroke, our result in MCAI showed almost the same tendency of right-sided predominance as in previous reports [14–16]. This may be due to the hemodynamic system of the fetus, where ductus arteriosus and left common carotid artery provide an easier route for emboli to the left hemisphere than to the right. As shown in Table 2-B, most children with PPIS showed motor asymmetry as the first clinical symptom, but the topography of paresis and the age of first parental concern differed according to the lesions. Compared with MCAI, most of which showed motor asymmetry in the upper extremities, children with PVI showed asymmetry in lower as much as upper extremities, supposedly because periventricular lesion affects descending fibers from the lower extremity regions of the motor cortex. Most children with MCAI presented with asymmetrical hand use when they started reaching for various objects (mean age of 5.9 months), while a certain number of PVI patients revealed asymmetrical lower limb activity after they started standing or walking (mean age of 11.6 months). Although we did not investigate the delay of diagnosis, Kirton reported that the interval from the first parental concern to the final diagnosis in PVI was longer than in MCAI, and he insisted that PVI should be better known among primary care pediatric physicians [16].

4.3. Functional outcomes Functional outcomes of HCP clearly differed according to its etiology. In PIS groups (PVI, MCAI, DGMI), the mean ages at initial walk were significantly younger, intellectual development was significantly better, and occurrence of epilepsy was significantly lower than in CD. This may be because cerebral dysgenesis often exists widely or bilaterally, although motor paresis presents only on one side. We considered that severe mental retardation would partially contribute to an inability to walk independently. Among the four groups, intellectual development was significantly better in DGMI, and the occurrence of epilepsy was significantly lower in PVI. The absence of direct cortical involvement in DGMI and PVI could explain these results. These correlations between cortical involvement and non-motor outcomes parallel the findings in Kirton’s report about PPIS [3]. As for the measurement of the affected hand’s function, we adopted movement itself (pinch, grasp-release, and press) because it will reflect the fundamental neural function more directly than such activity-related functional measures as Assisted Hand Assessment (AHA) or Manual Ability Classification System (MACS). But we could not show significant difference in the functional level among all the four groups, though MCAI is assumed to affect hand function more than PVI considering the location of the lesion. The severest condition, disuse, was found only in CD, which may partly because severe mental retardation would contribute to neglect of the affected hand. Concerning habilitation, occupational therapy or constraint-induced movement therapy for upper limb

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007

6

Y. Kitai et al. / Brain & Development xxx (2015) xxx–xxx

A

B

Ambulatory function

100%

100% 7

80% 60% 30

20%

0%

0%

C

DGMI

CD

80%

8

PVI

MCAI Pinch

**

1 14

4

6

25

unable to walk

Intellectual development

100%

28

##

D 100%

6

9

3 4

13

40%

20%

MCAI

4

14 21

able to walk

8

14

40%

PVI

9

80% 60%

62

Affected hand's function

Grasp

DGMI Press

CD

Disuse

Complication of epilepsy

**

##

3

14

80%

8

13 17

60%

12

14 40%

47

60% 40%

21

11

48 17

20%

10

20%

11

0%

0% PVI

MCAI

normal or mild MR

DGMI moderate MR

CD

PVI

MCAI epilepsy (-)

severe MR

DGMI

CD

epilepsy (+)

Fig. 3. Functional outcomes of each group. MR: mental retardation. **: significantly better among the four groups (p < 0.01). ##: significantly poorer among the four groups (p < 0.01).

4.4. Limitation of this study Time to initial walk by the type of HCP

Ratio of unable-to-walk

1.00

Arterial Venous CD

0.80

0.60

0.40

0.20

0.00 0

30

60 90 Age (months)

120

150

Fig. 4. Time to initial walk by the type of hemiplegic cerebral palsy. The mean ages at initial walk in venous infarction (PVI, 19 months) and in arterial infarction (MCAI and DGMI, 18 months) did not differ significantly (p = 0.41), and both of them were significantly earlier than that in cerebral dysgenesis (48 months) by log-rank test (p < 0.001).

function should be of great importance for children with PIS even if they can obtain good gross motor function. This is because 63% of children with PIS could not achieve selected finger movements, although all of them could walk independently at nearly the same age as normally developing children. Children with CD showed poor outcomes on the whole, and were considered to need multimodal interventions such as medication for epilepsy, habilitation for ambulatory and upper limb function, and education for intellectual development.

First, this is a retrospective cross-sectional study at regional center hospitals specialized in pediatric neurorehabilitation, so there should be selection bias. According to Husson’s prospective study of 73 neonates with arterial ischemic stroke, 54 children (74%) had not developed hemiplegia at 2 years of age [17]. Such very mildly impaired children with no need of habilitation and who are too severely impaired to come to rehabilitation hospitals were excluded from this study. Secondly, we couldn’t re-evaluate the hand function by validated functional measures such as AHA and MACS, owing to the study design of retrospective chart review. More validated evaluation of hand function in a prospective study would be the future issue. 5. Conclusions Etiological classification by MRI is of great diagnostic and prognostic value to provide an appropriate treatment for each child with HCP because the overall outcome apart from hand function clearly differed between PIS and CD. PIS was related to earlier attainment of independent walk, better intellectual development, and lower occurrence of epilepsy than CD. Infarction limited to deep gray matter was related to excellent intellectual development, and PVI showed lower occurrence of epilepsy. Issues for future study are more validated evaluation of hand function and the establishment of customized rehabilitation, medication, and education for each type of HCP.

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007

Y. Kitai et al. / Brain & Development xxx (2015) xxx–xxx

PVI is supposed to be the most common cause of HCP requiring rehabilitation, accounting for twothirds of PPIS. However, the early detection of PVI seems to be difficult because of mild clinical symptoms. A reliable screening system is necessary for applying early intervention in order to improve affected hand’s function. References [1] Wu YW, Lindan CE, Henning LH, Yoshida CK, Fullerton HJ, Ferriero DM, et al. Neuroimaging abnormalities in infants with congenital hemiparesis. Pediatr Neurol 2006;35:191–6. [2] Raju TN, Nelson KB, Ferriero D, Lynch JK. Ischemic perinatal stroke: summary of a workshop sponsored by the National Institute of Child Health and Human Development and the National Institute of Neurological Disorders and Stroke. Pediatrics 2007;120:609–16. [3] Kirton A, Deveber G, Pontigon AM, Macgregor D, Shroff M. Presumed perinatal ischemic stroke: vascular classification predicts outcomes. Ann Neurol 2008;63:436–43. [4] Takanashi J, Barkovich AJ, Ferriero DM, Suzuki H, Kohno Y. Widening spectrum of congenital hemiplegia: periventricular venous infarction in term neonates. Neurology 2003;61:531–3. [5] Takanashi J, Tada H, Barkovich AJ, Kohno Y. Magnetic resonance imaging confirms periventricular venous infarction in a term-born child with congenital hemiplegia. Dev Med Child Neurol 2005;47:706–8. [6] Takashima S, Mito T, Ando Y. Pathogenesis of periventricular white matter hemorrhages in preterm infants. Brain Dev 1986;8:25–30.

7

[7] Gould SJ, Howard S, Hope PL, Reynolds EO. Periventricular intraparenchymal cerebral haemorrhage in preterm infants: the role of venous infarction. J Pathol 1987;151:197–202. [8] Volpe JJ. Brain injury in the premature infant–from pathogenesis to prevention. Brain Dev 1997;19:519–34. [9] de Vries LS, Roelants-van Rijn AM, Rademaker KJ, Van Haastert IC, Beek FJ, Groenendaal F. Unilateral parenchymal haemorrhagic infarction in the preterm infant. Eur J Paediatr Neurol 2001;5:139–49. [10] Ghi T, Simonazzi G, Perolo A, Savelli L, Sandri F, Bernardi B, et al. Outcome of antenatally diagnosed intracranial hemorrhage: case series and review of the literature. Ultrasound Obstet Gynecol 2003;22:121–30. [11] Golomb MR, MacGregor DL, Domi T, Armstrong DC, McCrindle BW, Mayank S, et al. Presumed pre- or perinatal arterial ischemic stroke: risk factors and outcomes. Ann Neurol 2001;50:163–8. [12] Yoneda Y, Haginoya K, Kato M, Osaka H, Yokochi K, Arai H, et al. Phenotypic spectrum of COL4A1 mutations: porencephaly to schizencephaly. Ann Neurol 2013;73:48–57. [13] Abels L, Lequin M, Govaert P. Sonographic templates of newborn perforator stroke. Pediatr Radiol 2006;36:663–9. [14] Lee J, Croen LA, Backstrand KH, Yoshida CK, Henning LH, Lindan C, et al. Maternal and infant characteristics associated with perinatal arterial stroke in the infant. JAMA 2005;293:723–9. [15] Nelson KB, Lynch JK. Stroke in newborn infants. Lancet Neurol 2004;3:150–8. [16] Kirton A, Shroff M, Pontigon AM, deVeber G. Risk factors and presentations of periventricular venous infarction vs arterial presumed perinatal ischemic stroke. Arch Neurol 2010;67:842–8. [17] Husson B, Hertz-Pannier L, Renaud C, Allard D, Presles E, Landrieu P, et al. Motor outcomes after neonatal arterial ischemic stroke related to early MRI data in a prospective study. Pediatrics 2010;126:912–8.

Please cite this article in press as: Kitai Y et al. Outcome of hemiplegic cerebral palsy born at term depends on its etiology. Brain Dev (2015), http://dx.doi.org/10.1016/j.braindev.2015.09.007