NEONATAL SOCIETY ABSTRACTS
Predicting motor outcome and death in infants with central grey matter damage associated with hypoxic-ischaemic encephalopathy
Presented at the Neonatal Society 2009 Autumn Meeting (programme).
Martinez-Biarge M1,2, Diez J2, Kapellou O1, Allsop J3, Rutherford MA3, Cowan FM1
1 Dept of Paediatrics, Imperial College, Hammersmith Hospital, DuCane Rd, London W12 OHS
2 Department of Neonatology, La Paz University Hospital, Madrid, Spain
3 Robert Steiner MR Unit, Imaging Sciences, Imperial College, Hammersmith Hospital, DuCane Rd, London W12 OHS
Background: Damage to the central grey matter is the hallmark of acute perinatal hypoxic ischaemia in term infants(1). This frequently leads to motor difficulties, cerebral palsy (CP) and sometimes death. The precision with which these outcomes can be determined from neonatal imaging has not been fully explored.
Aim: To evaluate, in term newborns with HIE and basal ganglia/thalamic (BGT) lesions, the relation between the severity of early MR imaging findings to death and 2 year motor outcome
Methods: From 1993-2007, all inborn and referred term infants (≥35 weeks gestation) with (1) fetal compromise (abnormal CTG, meconium stained liquor, and/or a sentinel event) (2) poor condition at birth (Apgar scores <6 at 5 min, cord pH <7.1, or need for major resuscitation) (3) neonatal encephalopathy and (4) BGT injury seen on MR scans obtained within 6 postnatal weeks were included in this study. Antenatal and perinatal data were obtained. MR brain images were assessed for anatomical development, evidence of more prolonged or sub-acute problems, of long-standing established or unusual patterns of injury. Abnormality in the BGT, white matter (WM), internal capsule (PLIC) and cortex were classified by severity(1). Cerebellar, hippocampal, corpus callosal and brainstem abnormality were also documented Motor outcome was assessed using a standarised neurological examination(2) and Griffith’s developmental assessment where appropriate. The severity of cerebral palsy (CP) was determined using the Gross Motor Function Classification System (GMFCS)(3).
Results: 175 infants fulfilled the entry criteria, all had BGT lesions (median age at scan 10 days (range 1-42).
Perinatal factors: 54% were male, median GA was 40 (35-42) weeks, median birthweight was 3.3kg (1.92-4.64, 13%<10th centile), 5 infants (3%) were a twin. A sentinel event occurred in 27% (n=45). Median Apgar scores were 1 (0-6) and 4 (0-9) at 1 and 5 minutes; mean pH was 6.87; 82% required major resuscitation.
MRI: BGT lesions were mild (16%), moderate (21%), severe (37%) or very severe (26%) and definite PLIC abnormality was seen in 78%. Some WM injury was seen in 96%, mostly moderate (45%, including any sub-cortical change) or severe (37%). Cortical involvement was seen in 90%. Brainstem injury was seen in 68%, and in 90% of those with severe BGT lesions.
Outcome: 28% died; 71% of survivors had CP, mostly severe (GMFCS levels IV/V). Of the 37 infants without CP, 14 had minor neurological abnormality but all walked independently by 2 years. Of the children with CP only 9 were able to walk, and walking was delayed in 8 of the 9.
MRI and outcome: the severity of BGT/PLIC lesions was strongly associated with the severity of motor impairment (Spearman´s rank correlation=0.773; p<0.001). No infant with mild BGT had severe CP and only 8% had moderate impairment (GMFCS levels II/III). 95% of infants with severe or very severe BGT/PLIC lesions had severe motor impairment (p<0.001, of a spastic dystonic quadriplegic type in all but 3 who had athetoid CP. Moderate/severe BGT+PLIC lesions had a sensitivity of 0.87 and PPV of 0.91 for the prediction of CP. The association of WM, cortex and brainstem injury with motor impairment was less strong and only BGT injury correlated significantly in a logistic regression model. Of the 49 infants who died, 23 died neonatally (20 after withdrawal of intensive care), 17 died in year 1 and 9 died later. The presence and severity of brainstem injury was associated with death in the neonatal period and later: there were no deaths among infants without brainstem injury, whereas 55% of infants with severe brainstem injury died (p <0.001).
Conclusion: In infants with neonatal HIE and BGT injury, the severity of the BGT lesions is the best predictor of the severity and type of motor outcome; brainstem injury is significantly associated with death.
1. Okereafor A et al. Pediatrics 2008; 121; 906-914
2. Haataja L et al. J Pediatr 1999; 135:153-61
3. Palisano RJ et al. Physical Therapy 2000; 80: 974-985