Diffusion-weighted imaging changes in cerebral watershed distribution following neonatal encephalopathy are not invariably associated with an adverse outcome.

Aim Patterns of injury in term-born infants with neonatal encephalopathy following hypoxia-ischaemia are seen earlier and are more conspicuous on diffusion-weighted magnetic resonance imaging (DW-MRI) than on conventional imaging. Although the prognostic value of DW-MRI in infants with basal ganglia and thalamic damage has been established, data in infants in whom there is extensive injury in a watershed distribution are limited. The aim of this study was to assess cognitive and functional motor outcome in a cohort of infants with changes in a predominantly watershed distribution injury on neonatal cerebral MRI, including DWI. Method DW-MRI findings in infants with neonatal encephalopathy following hypoxia-ischaemia were evaluated retrospectively. Twenty-two infants in whom DWI changes exhibited a predominantly watershed distribution were enrolled in the study (10 males, 12 females; mean birthweight 3337g, 2830-3900g; mean gestational age 40.5wks, 37.9-42.1wks). Follow-up MRI data at the age of 3months (n=15) and over the age of 18months (n=7) were analysed. In survivors, neurodevelopmental outcome was assessed with the Griffiths Mental Development Scales at the age of at least 18months. Amplitude-integrated electroencephalography was used to score background patterns and the occurrence of epileptiform activity. Results DW-MRI revealed abnormalities that were bilateral in all infants and symmetrical in 10. The posterior regions were more severely affected in five infants and the anterior regions in three. Watershed injury occurred in isolation in 10 out of 22 infants and was associated with involvement of the basal ganglia and thalami in the other 12, of whom seven died. Cystic evolution, seen on MRI at age 3months, occurred in three of the 15 surviving infants. Neurodevelopmental assessment of the surviving infants was performed at a median age of 35months (range 18-48mo). Of the five survivors with basal ganglia and thalamic involvement, two developed cerebral palsy, one had a developmental quotient of less than 85, and two had a normal outcome. Of the 10 infants with isolated watershed injury, nine had an early normal motor and cognitive outcome. In all infants with a favourable outcome, background recovery was seen on amplitude integrated EEG within 48hours after birth. Conclusion Extensive DWI changes in a watershed distribution in term-born neonates are not invariably associated with adverse sequelae, even in the presence of cystic evolution. Associated lesions of the basal ganglia and thalami are a better predictor of adverse sequelae than the extent and severity of the watershed abnormalities seen on DW-MRI.