Eur J Pediatr DOI 10.1007/s00431-007-0629-2

ORIGINAL PAPER

Influence of gestational age on the type of brain injury and neuromotor outcome in high-risk neonates

Christine Van den Broeck & Eveline Himpens & Piet Vanhaesebrouck & Patrick Calders & Ann Oostra

Received: 16 July 2007 /Accepted: 4 October 2007 # Springer-Verlag 2007

Abstract This study was an investigation of a possible periventricular leukomalacia, 24% intraventricular hemor- correlation between either the gestational age (GA) and rhage and 18% persistent flares. There was a significant type of brain injury or between the gestational age and type, correlation between the GA and type of brain injury (P< distribution and severity of cerebral palsy (CP). Four 0.001; Cramer’s V=0.76) and between the GA and type (P= hundred sixty-one children with a birthweight ≥1250 g 0.004; Cramer’s V=0.47) and distribution (P<0.001; and GA ≥30 weeks with a complicated neonatal period and/ Cramer’s V=0.55) of CP. There was no significant correla- or brain injury on serial cerebral ultrasound were selectively tion between the GA and severity of CP. The type of brain followed at the regional Center for Developmental Disor- injury detected by serial ultrasound during the neonatal ders. The children were divided into a preterm and term period, as well as the type and location of CP detected group. There were 40 children with cerebral palsy in the during later childhood, are all GA-dependent in at-risk preterm group and 38 children with cerebral palsy in the newborn infants with a birthweight of ≥1,250 g and term group. Various types of brain injury diagnosed by GA ≥30 weeks. echography were nosologically classified. The type, distri- bution and severity of cerebral palsy were also registered. Keywords Cerebral palsy. Gestational age . Type of brain The type of brain injury most frequently occurring in the injury . Distribution of cerebral palsy. Type of cerebral palsy term group was hypoxic-ischemic injury to the basal ganglia (39%), focal ischemia (18%), subcortical hemor- Abbreviations rhage (13%) and parasagittal cerebral injury (10%). In the CP cerebral palsy preterm group 39% of the children with cerebral palsy had GA gestational age PVL periventricular leukomalacia IVH intraventricular hemorrhage

Christine Van den Broeck and Eveline Himpens equally contributed to this article. : * : C. Van den Broeck E. Himpens ( ) P. Calders Introduction Rehabilitation Sciences and Physiotherapy Ghent, University College Arteveldehogeschool-Ghent University, Campus Heymans (UZ) 1B3, De Pintelaan 185, Cerebral palsy (CP) is one of the most common causes of 9000 Ghent, Belgium motor disability in childhood. CP describes a group of dis- e-mail: [email protected] orders of movement and posture causing activity limitation A. Oostra that are attributed to non-progressive disturbances occurring Center for Developmental Disorders, in the developing fetal or infant brain. Abnormal motor Ghent, Belgium behavior is the core feature of CP [3]. Barkovich [2] has shown that the brain region most sus- P. Vanhaesebrouck Department of Neonatology, University Hospital Ghent, ceptible to hypoxic/ischemic damage changes as the infant Ghent, Belgium matures. Preterm and full-term infants may thus experience Eur J Pediatr ischemia in different regions of the brain. Vohr [23] found spastic cerebral palsy if they had at least two of the following that the prevalence of any type of cerebral palsy and the criteria: abnormal posture or movement, increased tone or degree of severity of CP differed significantly for two groups . An abnormal pattern of posture and/or move- of different gestational age. Therefore, probable relation- ments with involuntary, uncontrolled, recurring, occassion- ships between either the gestational age (GA) and type of ally stereotyped movements was defined as dyskinetic brain injury or between the GA and major characteristics cerebral palsy. Two types of dyskinetic cerebral palsy were of CP (i.e., type, distribution and severity) are described discerned, i.e., dystonic and choreo-athetotic cerebral palsy. [6, 10–14, 21]. Dystonic CP is described as hypokinesia (reduced activity Registers of childhood impairments and large population- and stiff movements) and hypertonia. Hyperkinesia (in- based studies monitor trends in the rates of cerebral palsy creased activity with stormy movements) and hypotonia are according to gestational age, but focus particularly on the classified as the choreo-athetotic form of dyskinetic CP. Two prevalence of CP as an assessor of the quality of perinatal mixed forms of CP are also identified, i.e., dystonia with care [6, 9, 21]. Other studies focus on the type, distribution spasticity and with spasticity. Lastly, is a and severity of cerebral palsy of term and pretem infants [1, primary disorder of coordination. Children with ataxia show 3, 6, 7, 9, 12, 21, 26]. The aim of this study was to an abnormal pattern of posture and/or movements with loss investigate a possible relationship either between the GA of muscular coordination so that movements are performed and type of brain injury, defined by serial echography, or with abnormal force, rhythm and accuracy. between GA and characteristics of CP. Different types of distribution of CP were nosologically defined. is a unilateral impairment. Limbs on one side of the body are involved. Diparesis, triparesis and Patients and methods quadriparesis are bilateral motor involvements. When only the lower limbs are affected, CP is classified as diparetic. During an 8-year period (January 1995 to December 2002), When four limbs are affected CP is classified as a quadri- 461 children with a birthweight ≥1,250 g and a GA ≥30 . Triparesis is used when one arm is not or only who had completed weeks with a complicated neonatal period minimally affected in a quadriparetic child [25]. and/or brain injury consulted the regional Center for Devel- The degree of severity of CP was classified as mild, opmental Disorders. The children were assessed at the moderate and severe. A child with a mild cerebral palsy corrected age of 4, 7, 12, 18 and 24 months. In the Department performs gross motor skills independently, but speed, bal- of Neonatology of the University Hospital of Ghent, an ance and coordination are reduced. CP was defined as echography was performed on day 1, 3, 7, 14 and 90 (or at the moderate if the child reached a form of independent walking time of discharge). For children with GA ≥35 weeks major with or without walking aids. When the child had no basic ultrasound findings were usually confirmed by MRI. antigravity postural control and/or could not walk, CP was The children were divided into two groups: a preterm group classified as severe. Nowadays it is common practice to use (276 children) with a gestational age (GA) between 30 and the GMFCS for the degree of severity of CP. Bax et al. [4] 37 weeks and a term group (185 children) with GA ≥37 weeks. describe a strong correspondence (K=0.583; P<0.001) be- Each child was subjected to a detailed physical and neu- tween the classification of mild-moderate-severe and the rological examination. Children were classified conforming GMFCS. to the definition proposed by the Surveillance of Cerebral Different types of brain injury detected by cerebral Palsy in Europe [3, 20]. Children were classified as having echography were identified.

Fig. 1 Relative frequency of term preterm type of brain injury in the term and preterm group (PVL= 40 periventricular leukomalacia; 35 IVH=intraventricular hemor- rhage; bas=injury to basal gan- 30 glia; para=parasagittal cerebral 25 injury; focal=focal ischemia; haemor=subcortical 20 hemorrhage) 15 10

relative frequency (%) 5 0 PVL flares IVH bas para focal haemor Eur J Pediatr

Fig. 2 Relative frequency of term preterm type of cerebral palsy in the term and preterm group (spast= 80 spastic; dyst=dystonic; ath=choreo-athetosis; dyst+ spast=dystonia with spasticity; 60 ath+spast=athetosis with spasticity) 40

20

relative frequency (%) 0 spast dyst ath dyst + spast ath + spast ataxia

Periventricular leukomalacia (PVL) is defined as peri- For the children with CP a highly significant (P<0.001) ventricular areas of increased echogenicity evolving into correlation between the GA and the type of brain injury was small or extensive periventricular cystic lesions involving found (Cramer’s V=0.76). The most common brain injury occipital and fronto-parietal white matter. Flares are defined in the term group with CP was an injury to the basal ganglia as diffuse periventricular echodensities persistent beyond (39%), followed by focal ischemia (18%), subcortical hem- 7 days of age. Deep grey matter injury involves the thal- orrhage (13%) and parasagittal injury (10%). The diagnosis amus and basal ganglia. Parasagittal cerebral injury refers to for all four children with parasagittal injury was confirmed a lesion of the cerebral cortex and subcortical white matter by MRI. In the preterm group 39% had periventricular with a characteristic distribution. Although the injury is usually leukomalacia, 24% had an intraventricular hemorrhage and symmetrical, it may be more striking in one hemisphere as 18% of these prematurely born children had persistent flares well. Focal ischemia refers to necrosis in the territory of one or on neonatal ultrasound scans. For two preterm infants we more large cerebral blood vessels. Hemorrhage can be sub- had no neonatal ultrasound scan (Fig. 1). cortical or intraventricular. Intraventricular hemorrhage (IVH) A significant correlation (P<0.01; Cramer’s V=0.47) is a hemorrhage into the germinal matrix and the ventricles. was also found between the GA and type of CP. In the term The subcortical hemorrhages are situated in the vascular group 31% of the children had a spastic type of CP, 29% boundery (watershed) zones. For term infants these are the had dystonia, and 24% had the mixed form of dystonia with regions between the anterior and middle cerebral arteries and spasticity. In the preterm group 72% of the children with between the middle and the posterior cerebral arteries. cerebral palsy had the spastic type of CP; 12% had the All analyses were performed using SPSS 12 for Windows. mixed form of dystonia with spasticity (Fig. 2). Categorical measures were analyzed with chi2-analysis. The The Cramer’s V correlation between gestational age and Cramer’s V test was used to examine the strength of the distribution of cerebral palsy was 0.55 (P<0.001). In the association between two categorical variables. The level for term group 37% of the children were classified as having statistical significance was set at P<0.05. hemiparesis and 58% as having quadriparesis. For the pre- term group 47% of the children had diparesis, 40% quadri- paresis and 10% hemiparesis (Fig. 3). Results There was no significant correlation between the GA and severity of CP. For the term group 47% of the children had CP was unequivocally present in 20% (38/185) of the term a mild, 18% a moderate and 34% a severe degree of CP. For group and in 14% (40/276) of the preterm group. The overall the preterm group 45% of the children had a mild, 35% a prevalence of CP was 17%. moderate and 20% a severe degree of CP.

Fig. 3 Relative frequency of term preterm distribution of cerebral palsy in the term and preterm group

60 50 40 30 20 10 0

relative frequency (%) Hemiparesis diparesis triparesis quadriparesis Eur J Pediatr

Discussion higher metabolic rate of subcortical nuclei compared with cerebral hemispheres explains the preponderance of sub- A large cohort of more than 450 preterm and term high-risk cortical damage [16]. newborn infants with a GA ≥30 weeks and a birthweight Secondly, a significant correlation was found between ≥1,250 g was systematically selected for standardized the GA and type and distribution of cerebral palsy. The developmental follow-up because of either an unfavorable existence of periventricular leukomalacia is the strongest course or because of the detection of significant cerebral and most independent risk factor for the subsequent de- lesions on serial cerebral ultrasound during their NICU stay. velopment of CP in preterm infants. The grade of PVL is Firstly, a significant correlation was found between fetal significantly correlated with the clinical type and severity of maturation, i.e., GA, and type of brain injury. As mentioned cerebral palsy [5, 7, 8, 12, 14, 15, 17]. Most of these earlier, the most susceptible region to hypoxic/ischemic children develop spastic di- or quadriplegia [1, 6, 21]. damage changes as the infant matures. Preterm infants and Preterm infants with major intracranial hemorrhage develop full-term infants experience ischemia in different regions of spastic , or hemiplegia [9]. Term infants the brain. Classically, this has been assumed to result from sustaining an acute, near-total intrauterine asphyxia tend to the changing location of the intervascular boundery zones present with athetosis in addition to pyramidal symptoms, [2]. This could be a possible explanation for the correlation severe seizure disorders and mental retardation. These patients between GA and type of brain damage. The most common are described as having dyskinetic cerebral palsy [2, 16]. brain injuries in preterm infants are periventricular leuko- Term infants subjected to prolonged partial asphyxial epi- malacia and intracranial hemorrhage. The pathogenesis of sodes most often have lesions of the cerebral cortex in a periventricular leukomalacia relates to three major factors. watershed type of distribution. They have pyramidal-spastic The first two of these, an incomplete state of development signs of cerebral palsy. of the vascular supply to the cerebral white matter and a We conclude that for high-risk infants with a birthweight maturation-dependent impairment in regulation of cerebral ≥1,250 g and GA ≥30 weeks, GA is highly predictive of the blood flow, underlie a propensity for ischemic injury to preferential type of brain injury in accordance with the cerebral white matter. The third major pathogenic factor is maturational cerebrovascular redistribution. The type and the maturation-dependent vulnerability of the oligodendrog- distribution of CP are also related to GA, indirectly due to lial precursor cell that represents the major cellular target in specific injuries of maturation-dependent susceptible brain PVL [24]. The germinal zones are most active between regions. approximately 8 and 28 weeks of gestation. Intrauterine exposure to infection and fetal inflammation are also related to an increased risk for PVL and CP [4, 13, 22]. The most common site of intraventricular hemorrhage in the preterm References infant is the subependymal germinal matrix region. The germinal matrix area is highly vascularized. The integrity of 1. Ancel PY, Livinec F, Larroque B, Marret S, Arnaud C, Pierrat V, ’ the capillaries in the germinal matrix is tenuous because of Dehan M, N Guyen S, Escande B, Burguet A, Thiriez G, Picaud JC, Andre M, Bréart G, Kaminski M (2006) Cerebral palsy among the lack of supportive tissue. Germinal matrix hemorrhage very preterm children in relation to gestational age and neonatal may be unilateral or bilateral and occurs in isolation in most ultrasound abnormalities: the EPIPAGE cohort study. Pediatrics preterm infants [18]. In term infants the most common 117:828–835 injuries are parasagittal cerebral injury, injuries to basal 2. Barkovich J (2005) Pediatric Neuroimaging. 4th edn. Raven Press, New York ganglia and focal ischemia. In term infants the vascular 3. Bax M, Goldstein M, Rosenbaum P, Leviton A, Paneth N, Dan B, boundary zones lie in the regions between the anterior and Jacobsson B, Damiano D, Executive committee for the definition middle cerebral arteries and between the middle and the of cerebral palsy (2005) Proposed definition and classification of – posterior cerebral arteries. The distribution of tenuous arte- cerebral palsy, April 2005. Dev Med Child Neurol 47:571 576 4. Bax M, Tydeman C, Flodmark O (2006) Clinical and MRI cor- rial supply has been termed parasagittal. This zone is the relates of cerebral palsy-the European cerebral palsy study. JAMA primary location of ischemic injury in prolonged partial 296:1602–1608 asphyxia [19]. An acute, near-total intrauterine asphyxia at 5. De vries LS, Eken P, Groenendaal F, van Haastert IC, Meiners LC the end of labor shows a consistent pattern of injury in (1993) Correlation between degree of periventricular leukomala- cia diagnosed using cranial ultrasound and MRI later in infancy in subcortical brain nuclei, including the thalamus, basal gan- children with cerebral palsy. Neuropediatrics 24:263–268 glia and brainstem in contrast with the relative sparing of 6. Drummond PM, Colver AF (2002) Analysis by gestational age of the cerebral cortex and white matter. The distribution of cerebral palsy in singleton births in north-east England 1970-94. – injury in these infants reflects the hierarchy of metabolic Paediatr Perinat Epidemiol 16:172 180 7. Han TR, Bang MS, Lim JY, Yoon BH, Kim IW (2002) Risk needs that are unmet after severe, sudden disruption of factors for cerebral palsy in preterm infants. Am J Phys Med substrate supply as occurs in an acute, severe asphyxia. The Rehabil 81:297–303 Eur J Pediatr

8. Hamrick SE, Miller SP, Leonard C, Glidden DV, Goldstein R, 17. Ryoon HT, Moon B, Young LJ, Hyeon B, In Wom K (2002) Risk Ramaswamy V, Piecuch R, Ferriero DM (2004) Trends in severe factors of cerebral palsy in preterm infants. Am J Phys Med brain injury and neurodevelopmental outcome in premature Rehabil 81:297–303 newborn infants: the role of cystic periventricular leukomalacia. 18. Shankaran S (2003) Hemorrhagic lesions of the central nervous J Pediatr 145:593–599 system. In: Stevenson DK, Benitz WE , Sunshine P (eds) Fetal and 9. Himmelmann K, Hagberg G, Beckung E, Hagberg B, Uvebrant P neonatal brain injury. University Press, Cambridge, pp 175–188 (2005) The changing panorama of cerebral palsy in Sweden. IX. 19. Sunshine P (2003) Perinatal asphyxia: an overview. In: Stevenson Prevalence and origin in the birth-year period 1995–1998. Acta DK, Benitz WE, Sunshine P (eds) Fetal and neonatal brain injury. Paediatr 94:287–294 University Press, Cambridge, pp 1–29 10. Joseph KS, Allen AC, Lutfi S, Murphy-Kaulbeck L, Vincer MJ, 20. Surveillance of cerebral palsy in Europe: a collaboration of cerebral Wood E (2003) Does the risk of cerebral palsy increase or palsy surveys and registers (2000) Surveillance of Cerebral Palsy in decrease with increasing gestational age? BMC Pregnanacy Europe (SCPE). Dev Med Child Neurol 42:816–824 Childbirth 23:1–8 21. Topp M, Uldall P, Greisen G (2001) Cerebral palsy births in 11. Krägeloh-Mann I, Petersen D, Hagberg G, Vollmar B, Hagberg B, eastern Denmark, 1987-90: implications for neonatal care. Pediatr Michalelis R (1995) Bilateral spastic cerebral palsy-MRI pathol- Perinat Epidemiol 15:271–277 ogy and origin. Analysis from a representive series of 56 cases. 22. Vermeulen GM, Bruinse HW, Gerards LJ, de Vries LS (2001) Dev Med Child Neurol 37:379–397 Perinatal risk factors for cranial untrasound abnormalities in 12. Krägeloh-Mann I, Horber V ( 2007) The role of magnetic neonates born after spontaneous labour before 34 weeks. Eur J resonance of cerebral palsy: a systematic review. Dev Med Child Obstet Gynecol Reprod Biol 94:290–295 Neurol 49:144–151 23. Vohr BR, Wright LL, Poole WK, McDonald SA (2005) 13. Nelson KB (2002) The epidemiology of cerebral palsy in term Neurodevelopmental outcomes of extremely low birth weight infants. Ment Retard Dev Disabil Res Rev 8:146–150 infants <32 weeks’ gestation between 1993 and 1998. Pediatrics 14. Okumura A, Kato T, Kuno K, Hayakawa F, Watanabe K (1997) 116:635–643 MRI findings in patients with spastic cerebral palsy. II: correlation 24. Volpe JJ (2001) Hypoxic-ishemic encephalopathy. In: neurology with type of cerebral palsy. Dev Med Child Neurol 39:369–372 of the newborn. 4th ed. W.B. Saunders CO, Philadelphia 15. Okumura A, Hayakawa F, Kato K, Kuno K, Watanabe K (1997) 25. Wichers MJ, van der Schouw YT, Moons KG, Stam HJ, van MRI findings in patients with spastic cerebral palsy. I: correlation Nieuwenhuizen O (2001) Prevalence of cerebral palsy in The with gestaional age at birth. Dev Med Child Neurol 39:363–368 Netherlands (1977-1988). Eur J Epidemiol 17:527–532 16. Pasternak JF, Gorey MT (1998) The syndrome of acute near- 26. Wu YW, Croen LA, Shah SJ, Newman TB, Najjar DV (2006) total intrauterine asphyxia in the term infant. Pediatr Neurol Cerebral palsy in a term population: Risk factors and neuro- 18:391–398 imaging findings. Pediatrics 118:690–697