ABSTRACT
BACKGROUND: Very preterm infants (<30 weeks of gestation) are at increased risk of specific language impairment and systematic developmental follow-up is essential for the provision of targeted early intervention. AIMS: To define the predictive value of early language testing and stability of language development, and perinatal and demographic risk factors for the diagnosis of SLI at 5 years, in a cohort of preterm infants. STUDY DESIGN: We used a retrospective hospital based cohort study. SUBJECTS: Preterm infants <30 weeks of gestation, were cared for in NICU at RPAH, between 2004 and 2007, and prospectively enrolled in developmental follow-up. Standardised developmental assessment was done at 3 years utilising the Bayley Scales of Infant and Toddler Development-III and the Wechsler Preschool and Primary Scale of Intelligence-III was done at 5 years. OUTCOME MEASURES: Predictive value and stability of early language testing were assessed with respect to SLI at 5 years, using measures of diagnostic accuracy and kappa values. Multivariate logistic regression was performed during the distribution of perinatal and demographic risk factors for SLI. RESULTS: One-in-five met diagnostic criteria for SLI (19%, n=24). Limited diagnostic accuracy was found with early expressive language and the stability of language scores demonstrated only fair agreement (Cohen's κ .383). Multilingual status and extreme gestational age at 24-25 weeks were associated with a six-fold increased risk of SLI (OR 6.09, 95% CI 1.89-19.56; OR 6.09, 95% CI 1.28-29.0). CONCLUSION: We defined a high incidence of SLI among our cohort, but only a limited diagnostic accuracy of early language testing. Multilingual status and extreme prematurity were independent risk factors for SLI. It remains imperative to perform continued developmental assessments beyond pre-school age to identify language impairment with greater accuracy.
