16939
Parsing Heterogeneity of Early ASD Phenotype: Stability and Change

Thursday, May 15, 2014
Atrium Ballroom (Marriott Marquis Atlanta)
S. H. Kim1, S. Macari2, C. A. Saulnier3, A. M. Steiner4, T. R. Goldsmith5, J. Koller6, K. D. Tsatsanis2 and K. Chawarska2, (1)40 Temple St., Suite 7D, Yale University School of Medicine, New Haven, CT, (2)Child Study Center, Yale University School of Medicine, New Haven, CT, (3)Marcus Autism Center, Children's Healthcare of Atlanta and Emory University School of Medicine, Atlanta, GA, (4)Yale University, New Haven, CT, (5)Department of Pediatrics, University of New Mexico, Albuquerque, NM, (6)The Hebrew University of Jerusalem, Jerusalem, Israel
Background: Past studies have consistently found high stability in the diagnosis of Autism Spectrum Disorders (ASD) during toddler years. However, diagnostic stability has not been examined in current cohorts of toddlers for whom intensive interventions are readily accessible. Moreover, due to tremendous heterogeneity observed within ASD early on, knowledge of short-term outcomes is still limited.

Objectives: (1) To replicate and extend previous findings on stability of early diagnosis using toddlers from a large clinic-referred sample (born between 2006-2012); (2) To identify more homogeneous subgroups within ASD based on a constellation of key clinical features in the second year of life; (3) To examine short-term outcomes in the identified subgroups at 1-2-year follow-up.

Methods:  One hundred toddlers referred for a differential diagnosis of ASD at 14-27 months (Time 1) and followed until 30-49 months (Time 2) were evaluated with regard to autism symptoms, verbal and nonverbal skills, and adaptive functioning.  At Time 1, we identified homogeneous subgroups within ASD using Hierarchical Clustering (HC) analysis.  Developmental outcomes of the identified clusters were compared using generalized linear mixed models. All but one child received early intervention (Mean=16.9 hours/week; SD=9.16).

Results: 94% of children given an ASD diagnosis at Time 1 maintained the ASD diagnosis at Time 2. HC analysis identified 4 clusters at Time 1. Toddlers in clusters 1 and 2 had higher nonverbal and verbal skills and adaptive communication, daily living, and social functioning than toddlers in clusters 3 and 4.  Clusters 1 and 2 differed from one another by levels of autism symptoms and nonverbal skills. Clusters 3 and 4 differed in nonverbal, verbal, adaptive social, and daily living skills. Within each cluster, toddlers demonstrated high stability of nonverbal and daily living skills over time.  However, changes over time were found within clusters in severity of autism symptoms, verbal, communication, and socialization skills. Specifically, autism severity in cluster 2 increased over time (Cohen’s d=-1.1) whereas in the three remaining clusters, it remained stable. Verbal and communication skills significantly improved in three clusters (d’s=1-1.8). However, cluster 4 showed only a modest increase in verbal skills (d=0.5) and a decrease (d=-0.6) in adaptive communication skills. Finally, cluster 1 showed improvement (d=0.6) in adaptive social skills, whereas cluster 4 showed worsening of skills (d=-1.4); adaptive socialization for the other two clusters remained stable over time. The clusters differed in the amount of intervention they received between Time 1 and Time 2 with cluster 4 showing the most hours of intervention.

Conclusions: High stability of early ASD diagnosis was replicated with the current cohort of toddlers. However, considerable heterogeneity in clinical presentation at 14-27 months resulted in four subgroups whose short-term outcomes also varied. Short-term prognosis was better for children with intact cognitive and adaptive skills early on despite pronounced autism symptoms, especially in terms of progress in verbal and adaptive communication functioning. Identifying subgroups within ASD during toddler years based on multiple clinical features will be crucial for predicting outcomes and programming treatment.