Screening for autism in extremely preterm infants: potential pitfalls

Abstract

<h3>Objectives</h3> Recent studies using screening tests have raised concern of a high prevalence of autism in extremely preterm children. However, the specificity of screening in infancy may be confounded by the high prevalence of neuro-developmental sequelae in this population. The authors have investigated neuro-developmental correlates of positive autism screening in extremely preterm infants. <h3>Method</h3> All babies born &lt;27 weeks gestational age in England in 2006 were recruited to the EPICure 2 study. Parents of 559 (55%) survivors completed the modified checklist for autism in toddlers (M-CHAT) to screen for autistic features, the Parent Report of Children9s Abilities—Revised to screen for cognitive disability and questionnaire items to assess motor and sensory impairment at 2 years corrected age. <h3>Results</h3> Based on parent report, cognitive impairment was present in 340 (61%), motor impairment in 71 (13%), visual impairment in 10 (2%) and hearing impairment in 6 (1%) survivors. 216 (41%) screened positive for autistic features. Males were twice (OR 1.9; 95% CI 1.3 to 2.7) as likely to screen positive than females. Risk of a positive M-CHAT screen was significantly associated with motor (41.7; 12.9 to 135), cognitive (5.3; 3.5 to 8.0) and sensory impairment (p&lt;0.005). Among 200 children without neuro-developmental disability, 33 (16.5%) had positive M-CHAT screens (3.6; 2.6 to 5.0). <h3>Conclusions</h3> A high proportion of extremely preterm infants screened positive for autistic features. Positive screens were significantly associated with neuro-developmental impairment potentially yielding a high false-positive rate. Caution should be observed when screening for autism in this population and results should be interpreted in light of other clinical findings. Follow-up at school age is necessary to establish autism spectrum disorder diagnoses.