Diagnosing Autism in Fragile X Syndrome with the Revised ADOS Algorithms and Severity Score

Background: Fragile X syndrome (FXS) is associated with an increased risk of autism, with prevalence rates ranging from ~25-50%. A key step in studying the phenotypic overlap of autism and FXS (and potential role of FMR1 in autism symptomatology) will be to improve the sensitivity and specificity of autism diagnostic instruments. Revised algorithms for Modules 1-3 of The Autism Diagnostic Observation Schedule (ADOS; Lord et al. 2000), a gold standard measure for autism diagnosis, were recently developed with this goal in mind (Gotham 2007, 2008). Additionally, a standardized severity metric based on the revised algorithm scores has also been developed (Gotham, Pickles & Lord, 2009). Importantly, emerging research comparing autism diagnosis using the previous Western Psychological Services (WPS) algorithms to the new algorithms has revealed a shift in classification of autism spectrum disorders (ASD) in individuals with FXS (Harris et al., 2009).

Objectives: This study compared autism classifications derived from the new severity metric with those based on the original WPS version of the ADOS algorithm and the new research algorithms in children with FXS with and without autism, as well as children with autism only. We also examined the association between autism classification and severity scores to determine the utility of considering such scores to characterize autism symptomatology in FXS.

Methods: The new algorithm classifications for Modules 2 and 3 were examined for a group of 68 boys and 25 girls with FXS, and 14 boys with ASD only (no FXS). Participants ranged in age from 4-15 years. Raw scores of the new algorithm were mapped onto the calibrated severity scores provided in Gotham et al. (2009). The classification of each individual based on the calibrated severity scores was then compared to classifications from the new algorithm, and the original WPS algorithm. We examined changes in classification across groups and frequency of severity scores within groups. 

Results: The new severity index classification (SIC) resulted in substantial changes in autism classification among males with FXS. Twenty-seven percent (12/45) previously assigned a “spectrum” diagnosis using the old algorithm, were re-classified as autistic using the new SIC.. The frequency of the severity scores was also most varied in males with FXS. Twenty-five percent of males with FXS (31% in Module 2) received a score of 6 (the autism cut-off score for the SIC). This mirrors the percentage of diagnostic change seen using the new algorithms and the SIC classifications. We found no significant classification changes or frequency differences in girls with FXS or children with ASD only.

Conclusions: Findings indicate a considerable increase in autism classification in boys with FXS using the new algorithms and severity index. Such changes may be related to the severity score cut off. Future work may profit from considering the severity metric in tandem with classifications to better capture the heterogeneity of autism in FXS, particularly in treatment and intervention studies.