Presented at the Neonatal Society 2012 Autumn Meeting.
Jary S1, Whitelaw A1, Walloeb L2, Thoresen M1,2
1 Neonatal Neuroscience, School of Clinical Science, University of Bristol, Bristol, UK
2 Department of Physiology, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway
Background: Clinical trials of neuroprotection for infants with moderate or severe neonatal encephalopathy (NE) have used death or severe disability as trial outcomes with Bayley Scales of Infant Development-2 Index (1) scores <70 as part of criteria for severe disability (2). Bayley-3 (3) has now replaced Bayley-2 and, in preterm infants, is reported to give higher scores than Bayley-2 (4,5). Term infants with NE have a different spectrum of neurodevelopment (6,7) as all have brain injury. We hypothesised that fewer infants would have scores <70 using Bayley-3 than with Bayley-2. If so, this could confuse interpretation of new neonatal trials of NE and historical comparisons. Our aims were to prospectively compare Bayley-2 with corresponding scores in Bayley-3; to investigate the cut-off thresholds for moderate and severe disability in both tests and to derive conversion algorithms to predict Bayley-2 scores from Bayley-3 scores in a well-defined cohort of infants following NE.
Methods: Study participants were 61 children born 2007-2010 who fulfilled the entry criteria for therapeutic hypothermia (TH) and for whom complete scores for both Bayley-2 and Bayley-3 were available. Assessment was at median age 18.3 (16.8-19.7) months in a single session by one assessor proficient in the administration of both versions of the test. Bayley-3 scales were administered with any extra items specific to Bayley-2 interspersed as judged appropriate for the individual child to maximise motivation and minimise fatigue. Items with differences in administration were presented in accordance with Bayley-3 instructions but scored according to instructions of each version of the test. Inter-scorer reliability, investigated in 10 infants from video recordings, was 97%. Thresholds for disability were defined as moderate/severe <70; mild 70-84; none >85. Local ethical approval and parental consent were obtained.
Results: Median Bayley-3 Cognitive Composite (CC) score (100 [65-125]) was 9 points higher than median Bayley-2 Mental Developmental Index (MDI) score (91 [37-121]). Median Bayey-3 Motor Composite score (100 [58-124]) was 17 points higher than median Bayley-2 Psychomotor Developmental Index (PDI) score (83 [29-107]). Of the 10 children with Bayley-2 MDI <70, only 3 had Bayley-3 CC scores <70 and 7/10 had scores <85. Of the 11 children with Bayley-2 PDI <70, 3 had Bayley-3 MC scores <70 and 11/11 had scores <85. The relationship between different scores was explored using correlation analysis and score comparisons with the strongest correlations were used to derive regression equations from linear regression analysis. Estimated Bayley-2 Index scores derived from the equations were found to restore the number of infants classified with moderate/severe (<70) and mild (<85) degrees of developmental delay to similar proportions found using observed Bayley-2 scores.
Conclusion: Fewer children had scores <70 using Bayley-3 than with Bayley-2. This finding prohibits the direct comparison of Bayley-3 and Bayley-2 scores in infants with NE. Linear regression equations for cognitive and motor development can facilitate more direct comparison of Bayley-2 and Bayley-3 outcomes in cooled infants surviving NE.
Corresponding author: sally.jary@bristol.ac.uk
References
1. The Bayley Scales of Infant Development 2nd Edition1993 Psych Corp
2. Azzopardi DV. et al (2009) NEnglJMed 361(14)
3. The Bayley Scales of Infant and Toddler Development 3rd Edition 2006 Harcourt Assessment
4. Vohr BR. et al (2012) J Pediatr 161(2)
5. Moore T.et al (2012) J Pediatr 160(4)
6. Murray DM. et al (2010) DMCN 52(2)
7. van Handel M. et al (2007) Eur J Pediatr 166(7)