Autism is a pervasive developmental disorder with a strong genetic component, which is supported by results from twin studies (e.g., Bailey et al, 1995) and genome-wide association studies (see Abrahams & Geschwind, 2008). There is believed to be a type of autism, regressive autism, which accounts for roughly 20% of autism cases (e.g., Lord et al, 2004, based on parent report from the ADI-R), wherein an infant attains, and then loses, language and/or social skills. With rare exception (see genetic studies by Molloy et al. 2005), whether regression in autism also has a genetic component has yet to be investigated.
Objectives:
To investigate a genetic component in regressive autism by determining whether observed concordance rates for regression between sibling pairs with autism is greater than would be expected by chance.
Methods:
The data for our analyses came from the Autism Genetic Resource Exchange (AGRE) database. It included families who have two biological siblings who met criteria for an Autism Spectrum Disorder (i.e., Autism, Not Quite Autism or Broad Spectrum) based on the ADI-R (2003), parent report. The total number of sibling pairs were 451 (total subjects = 902), separated into three subject groups: (1) 24 monozygotic (MZ) twin pairs, (2) 49 same gender dizygotic (DZ) twin pairs, (3) 378 same gender non-twin pairs. Individuals were considered as having regressed if their parents said “yes” to questions 11 (language regression) or 25 (social regression) on the ADI-R. Observed concordance rates (i.e., both siblings regressed or both siblings did not regress) were compared to rates that would be expected by chance, using actual regression rates and binomial statistics. Analyses were conducted on four different regression categories: (A) either language or social, (B) only language, (C) only social, or (D) both types.
Results:
Across all 902 subjects, the mean rates of regression were 28.6%, 17.4%, 21.4% and 10.1% for the four different regression categories, A, B, C and D, respectively. Of 12 analyses (3 subject groups x 4 regression categories), only MZ twins showed concordance rates that were significantly greater than chance. Among MZ twins, greater-than-chance twin concordances were observed exclusively for category A (observed concordance rate = 75.0%, chance concordance rate = 51.1%, p = 0.02) and C (observed concordance rate = 83.3%, chance concordance rate = 58.4%, p = 0.01). The corresponding concordance rates for DZ twins (category A = 67.3%, category C = 67.3%), and non-twin siblings (category A = 61.4%, category C = 69.6%) did not exceed what would be predicted by chance.
Conclusions:
Although the higher concordance rate observed in our small sample of MZ twins suggests a possible component of genetic influence on regression, the fact that neither same gender DZ twins nor non-twin siblings exhibited concordance rates that differed significantly from chance suggests that genetic effects are likely limited and complex. Regression as measured by the ADI-R may be substantially influenced by non-genetic factors, and efforts to study inherited influences on regressive autism may require alternate methods and approaches to refine characterization of this phenotype.
See more of: Clinical Phenotype
See more of: Symptoms, Diagnosis & Phenotype