19163
Detection of Early ASD Risk in Preterm Infants
During the recent years, there has been significant advance in delineating the early markers of Autism Spectrum Disorders (ASD). The Autism Observation Scale for Infants (AOSI) is an observational measure designed to detect and monitor these early abnormalities in infants aged 6-18 months. Most of the research using the AOSI included high-risk siblings of children with ASD, therefore, it's generalization in other high-risk populations is still unclear.
Objectives:
The objectives of this this study were: a) to examine the prevalence of early social-communication abnormalities, as assessed with the AOSI, in a sample of preterm infants at 8 and 12 months corrected age (CA). b) to examine the association between the AOSI at ages 8 and 12 months and the ADOS at 18 months. c) To identify a cut-off criterion with optimal sensitivity and specificity in identifying preterm infants at risk for ASD, as assessed with the ADOS at 18 months.
Methods:
This study included 101 singleton infants (45% girls), born at 34 weeks gestation or earlier (M = 31.28; SD = 2.57; range = 24-34 weeks) at Hadassah Medical Centers in Jerusalem, Israel. Infants were assessed at 8 and 12 months CA with the AOSI and the Mullen; and at 18 months CA with the ADOS-T and the Mullen. A total score of ≥ 7 (counting the number of items coded one or above) and a total score of ≥ 9 are considered to be AOSI scores of concern.
Results:
Using the original cut-off criteria of the AOSI, eighteen infants scored above the cut-off at 8 months CA and six infants scored above the cut-off at 12 months CA (see Table 1). At 18 months CA, eight infants' ADOS-T algorithm scores were in the moderate to severe concern range. Using the original cut-off of ≥ 7 markers and total score ≥ 9 in the AOSI, sensitivity and specificity were 0.5 and 0.90 at 8 months, and 0.5 & 0.95 at 12 months, respectively (see table 2). Receiver operating characteristic analyses (ROC) analyses yielded lower cut-offs of 5 for number of items and 7 for total scores with more optimal prediction on ADOS classification at 18 months with 0.87 sensitivity and 0.73 specificity. Seven out of 8 children who scored positive on the ADOS have been identified at risk by the AOSI using these cut-offs. About half of the false positives were infants with other developmental difficulties.
Conclusions:
Although there were alarmingly high rates of social-communication abnormalities at 8 months, by 12 months most of those who scores at risk “grew out” of it and had much less abnormalities at 12 months. However, 8% of the children scored at risk for ASD by the ADOS-T at 18 months CA. Lower cut-off criteria on the AOSI yielded better predictive validity than those suggested for siblings. The trajectories and the phenotype of ASD in preterm population may be different from other high-risk populations and requires more research and follow-up.