Genetic analyses of the obsessive compulsive syndrome (OCS) gender and informant effects

Background: The diagnoses of Obsessive Compulsive disorder in children remains a difficult target. Recent concerns that the prevalence of OCD in children may higher than previously thought, has fueled the need to find specific and sensitive screening instruments. Recently, our group developed and tested a highly sensitive and specific screening approach to identify children at risk for OCD. We tested 11 items from the CBCL and via factor analyses of OCD and non-OCD aged matched samples of children, found an 8-item Obsessive Compulsive Scale (OCS) of the CBCL to be have high internal consistency, to be highly sensitive and specific for clinically diagnosed OCD (Nelson E, Hanna GL, Hudziak J et al.). Evaluating the Child Behavior Checklist as a Screening instrument for Obsessive-Compulsive Disorder in Children and Adolescents. In press, Pediatrics, 2001). The purpose of this study was to test the OCS in a large general population twin sample to determine if the scale would have high internal consistency, estimate cut points to be used to identify an appropriate number of children to be clinically tested for OCD, and to determine the genetic and environmental influences on the OCS. Method: CBCL data on 4246 10-year-old twin pairs from the Netherlands Twin Registry was analyzed in order to determine the genetic and environmental contributions to OCS. These data were also used to determine the factor structure and the internal consistency of the scale in a much larger and general population sample than we had used in our previous study. Results: Data were factor analyzed and a similar one factor OCS scale emerged on our data. Model fitting tested for genetic and environmental contributions were tested. Although the full model was fitted, the best fitting model was one that indicated significant additive genetic influences, 59% (80% C.I., 57-60), and unique environmental influences, 41% (80% C.I., 39-43). There were no gender effects and no evidence of sibling interaction. The fit statistics for the AE, no gender effects model were: chi2 =46.947, with 13 degrees of freedom, with the AIC of 20.947. Using an analyses of raw scores only, we were able to determine that all children who had a raw score of 4 or greater on this scale should be tested for clinical OCD (a raw score of 4 or greater identified the top 5% of our sample of 8492 10-year-old children. Because it is estimated that OCD may have a prevalence of 2.5% or greater in the general population, we concluded that this would be a conservative cut point to be used for screening). Discussion: We have demonstrated that the OCS, an 8-item scale that is collected via parent report as part of the 118-item CBCL, is influenced by genetic factors (59%), unique environmental factors (41%), and is not affected by differences in gender or informant patterns. Our data do not support gender differences reported by others, but do support the concept of using the OCS to screen for clinical OCS because of the ease of administration of the CBCL.