A Population-Based Twin Study of Autism in California

Background:

Autism has been referred to as the most heritable of the neurodevelopmental disorders. To date only three “population based” twin studies have been conducted each with very small samples of twin pairs (36 MZ and 30 DZ total). These studies reported a DZ concordance estimate of 0. This is clearly implausible, since the estimated sib recurrence risk is much higher. Currently we have no reliable heritability estimates and need to better understand twin concordance.

Objectives:

To estimate the probandwise concordance rates for autism in MZ and DZ twin pairs.

Methods:

Twin pairs for whom at least one child in the pair was receiving services for autism from the Department of Developmental Services (DDS) were identified.  DDS operates a system of 21 Regional Centers (RC) that coordinate services for persons with autism, mental retardation, and other developmental disabilities. The electronic DDS client files were linked by California Autism and Developmental Disabilities Research and Epidemiology (CADDRE) staff to California live birth records, which provided information on twin status. In this study we recruited twin pairs born in California from the birth years 1987 to 2004.

Assessments were carried out by two sites: AGRE and Stanford. Zygosity  was determined by genotyping.

Probandwise concordance rates in the twin pairs were calculated for two different definitions of autism. In the narrow definition, only twin individuals meeting criteria for autism on the ADI-R and autism on the ADOS are considered affected. The broader definition is based on the criteria published by Risi et al. (2006), which requires that criteria for autism spectrum disorder are met on the ADOS and criteria for autism in the domain of social reciprocity, and either the communication or restricted, repetitive behavior domain are met on the ADI-R.

Results:

We have identified 1122 twin pairs receiving DDS/RC services for autism.  Of these 1007 twin pairs fulfill our eligibility criteria. A total of 533 families expressed interest in participating in the study and agreed to be contacted by Stanford team members. Ten families turned out to be ineligible based on our exclusion criteria and 130 families decided not to participate because of the time commitment involved. So far assessments have been completed on 210 pairs of twins.

Concordance rates in MZ twins for the narrow and broad definition are higher (0.57 and 0.81 respectively) than for DZ pairs (0.13 and 0.19). Concordance rates for sex-discordant DZ twin pairs (0.08 for both the narrow and broad definition) are lower than concordance rates for both male-male DZ (0.19 and 0.30) and female-female DZ pairs (0.27 and 0.38). The risk for a co-twin to be affected is much higher if one of the affected twin individuals is female.

Conclusions:  Autism concordance rates for DZ twin pairs are higher than previously published concordance rates of 0 for the narrow, and 0.1 for the broad definition. Further, the difference between MZ and DZ pairs (4.3-fold) is lower than previously reported.