BackgroundObsessive compulsive disorder (OCD) and autistic spectrum disorder (ASD) are each characterised by compulsive behaviours. Existing data suggest up to 20% of treatment seeking individuals with OCD demonstrate comorbid autistic traits [1].The impact of autistic traits on OCD is poorly understood. In this study, we investigated the frequency, phenomenology and demographic characteristics associated with the presence of comorbid autistic traits in adult patients attending a specialist clinic for OCD. MethodologyConsecutive patients attending a UK specialist OCD clinic with a clinical diagnosis of DSM IV OCD were invited to consent. Participants completed the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) Checklist, Brown Assessment of Beliefs Scale (BABS) and the Autism Quotient (AQ) [2], a 50-item self-rated screen for autistic traits (AT). A total AQ score of 26/50 was chosen as the ‘cut-off’ to determine the likelihood of ASD. Previous validation studies using the AQ have shown that a score of >=26 predicts a 50% likelihood of DSM-IV ASD [3]. Patients were rated for severity of OCD and Obsessive Compulsive Personality Disorder with the Y-BOCS and Compulsive Personality Assessment Scale (CPAS), respectively. Gender and age of onset of OCD was also collected. Participants who scored >=26 on the AQ (OCD+AT) were seen by an experienced clinician for a structured one-to-one assessment for ASD (including Autistic Disorder, Asperger’s Syndrome, Pervasive Development Disorder Not Otherwise Specified) based on DSM-IV criteria. ResultsOut of a total sample of 135 OCD outpatients, complete data was collected for 67. Twenty-nine of these (43%) scored >=26 on the AQ and of these 21 (31% of sample) were diagnosed with an ASD (Female 48%) following in-depth clinician assessment.In the OCD+AQ group the gender-ratio was roughly equal, (Female 49%, whereas amongst OCD–AQ group, females predominated (58%). Compared to normative data [3], in the complete OCD sample the mean AQ total score (mean=24.19, SD=8.17) fell between that reported for a healthy population sample without mental disorder (mean=16.4, SD=6.3) and another sample diagnosed with high-functioning ASD (mean=35.8, SD=6.5). In the OCD sample, the AQ total score frequencies approximated a bimodal distribution, with the OCD+AT subgroup clustering around a mean total AQ score of 31.69 (SD=4.21). In the OCD+AT group, the AQ items endorsed most frequently occurred within symptom-domains representing impaired ‘attentional switching’ (mean 8.45, SD 1.55), ‘social skills’ (mean 6.86, SD 1.98) ‘communication skills’ (mean 5.38, SD 1.27) ‘attention to detail (mean 6.03, SD=1.92) and ‘imagination’ (mean 5.03, SD1.84 ).No significant differences between the OCD+AT and OCD-AT groups were observed in the age of OCD onset (mean=14.54, SD=8.76 versus 14.10SD=8.03) or the total Y-BOCS score (19.56,SD=8.78 versus 17.74, SD=8.18). However, the OCD+AT group scored significantly higher on the CPAS (mean 17.26, SD=7.24 versus 11.37, SD=7.2: P=0.002) and the BABS (mean 8.30, SD=5.95 versus 4.84, SD=4.42, P=0.014). The AQ total score correlated moderately with the CPAS and BABS (Pearson’s correlation (2 tailed) respectively 0.41 (P=0.001);0.30 (P=0.013). DiscussionOur study sample was limited by small size and the results should be considered preliminary. We found a high prevalence of autistic traits and ASD (around one third) in the treatment-seeking OCD population. These patients were no more severely symptomatic with OCD than the patients scoring low on the AQ, suggesting this does not simply represent a more severe form of OCD.Although the autistic traits that were the most strongly endorsed involved problems shifting attention, suggesting overlap with OCD, other autistic traits not usually associated with OCD were also frequently endorsed, including poor social, communication and imaginative skills, providing support for a true overlap with ASD. The presence of autistic traits was associated with OCPD and poor insight, which carries implications for treatment. The finding of an equivalent gender ratio merits further exploration.