ABSTRACT
OBJECTIVE: To assess diagnostic accuracy of 36-week anthropometric weight, length, and head circumference <10th and <3rd percentiles to predict preterm infant cognitive impairment. STUDY DESIGN: Cohort study of 898 preterm <30-week very-low-birth weight (<1500 g) infants. Anthropometric measures' accuracy to predict cognitive impairment (Bayley-III Cognitive Composite score) <80, 21-months corrected age (CA) and Wechsler Preschool and Primary Scale of Intelligence Quotient (intellectual outcomes) <70, 36-months CA, were determined using receiver operating characteristic (ROC) curves. RESULT: Thirty-six-week weight, length or head circumference <10th or <3rd percentile did not predict cognitive impairment; areas under ROC curves were <0.6. Sensitivities and specificities for 10th and 3rd percentile cut points were all poor, with most not exceeding 70%, whether the Fenton 2013 or INTERGROWTH 2015 growth charts were used. Brain injury and low maternal education were better predictors of cognitive impairment. CONCLUSION: Preterm infant 36-week anthropometric measurements are not accurate predictors of cognitive impairment.