Background Infants with post-haemorrhagic ventricular dilatation (PHVD) have a high risk of severe disability and parenchymal infarction increases this risk. Existing cranial ultrasound (CUS) markers of neurodevelopmental outcome are based on categorical features.

Objective To investigate to what extent quantitative CUS measurements correlated with severity of developmental outcome and the need for ventriculoperitoneal (VP) shunt at 2 years of age.

Design 69 premature infants with PHVD had lateral ventricle area, intraventricular echodensity and parenchymal lesion dimensions measured at the start of treatment for PHVD. Outcome measures were the Bayley Scales of Infant Development-II and functional ability at 2 years of age. Bayley developmental quotients (DQ) were used in preference to index scores to enable inclusion of severely disabled children.

Results Quantitative CUS measurements of parenchymal lesion area correlated significantly with later mental and motor DQ. Intraventricular echodensity area correlated with motor DQ in infants with grade 4 intraventricular haemorrhage (IVH). Neither ventricular area nor ventricular width correlated with DQ in grade 3 IVH. Infants who ultimately required a VP shunt had a significantly larger intraventricular echodensity area.

Conclusions CUS measurement of parenchymal lesions in infants with PHVD can increase the precision of predicting severe mental and motor disability, but ventricular size at the start of treatment is not predictive of outcome in infants with PHVD following grade 3 IVH.