Clinical tools used in young infants born very preterm to predict motor and cognitive delay (not cerebral palsy): a systematic review.

Abstract

This systematic review evaluates the accuracy of clinical tools used at a corrected age of 6months or younger to predict motor and cognitive delay (not cerebral palsy) at 24months' corrected age, in infants born very preterm. Six databases were searched. Quality was evaluated using the Quality Assessment of Diagnostic Accuracy Studies tool. Predictive analysis included calculation of sensitivity and specificity, inspection of summary receiver operating characteristics curves, and bivariate meta-analysis. Six assessments were identified in 10 studies of 992 infants. Overall prevalence of motor delay was 13.8% and cognitive delay was 11.7%. Methodological quality was variable for patient selection, reference standard, flow, and timing. All studies had a low risk of bias for the index test. General Movement Assessment (GMA) predicted motor and cognitive outcomes with good accuracy for mild, moderate, and severe delays (fidgety age: pooled diagnostic odds ratio=12.3 [5.9-29.8]; hierarchical summary receiver operating characteristics curve=0.733). The Hammersmith Infant Neurological Examination (HINE) demonstrated excellent predictive accuracy for severe motor delay (3mo and 6mo; sensitivity 93% [68-100%], specificity 100% [96-100%]) but showed limited ability to predict milder delays. In the population of infants born very preterm, few assessment tools used at 6 months or younger corrected age have proven predictive accuracy for cognitive and motor delay at 24 months' corrected age. Only the GMA and HINE demonstrated useful predictive validity. General movements have predictive validity for both motor and cognitive dysfunction in infants born very preterm. The Hammersmith Infant Neurological Examination showed the highest predictive accuracy for severe motor delay. The General Movement Assessment was the best tool to predict mild-to-moderate motor and cognitive delays.