Prenatal Influences on Autism Spectrum Disorders: Negative Evidence From a Twin Study

Background: A recent meta-analysis revealed several prenatal factors linked to later autism development (Gardener, Spiegelman, & Buka, 2009). General prenatal optimality may also play a role. Cotwins do not share all prenatal factors, making a twin study a viable and informative design to study the prenatal influences on autism. Dichorionic twins have especially differentiated prenatal experiences. During fetal development, dermal ridges develop concurrently with the brain, so disturbances in one system may reflect atypical development of the other. Also, physical asymmetries are thought to represent poor developmental canalization. Finally, the sexual dimorphism of the digit ratio between the second and fourth fingers is widely attributed to prenatal sex hormone exposure, making it a potential marker of the “extreme male brain” theory of autism.

Objectives: Older studies link chorionicity, dermatoglyphics, asymmetry, and digit ratio to autism or other psychiatric conditions. We sought to replicate and extend these prior investigations to determine whether autism concordance in twins varies by chorionicity and whether autism status is related to a-b ridge count, a-b asymmetry, and digit ratio.

Methods: Fifty-four pairs (23 monozygotic) of proband-ascertained twins (i.e., at least one twin is autistic) and participated, with a mean age of 8:5. We classified participants as being on the autism spectrum based on the Autism Diagnostic Observation Schedule and the Social Communication Questionnaire. We determined chorionicity via birth records, measured dermatoglyphics from palm prints, and scanned participants’ hands to calculate digit ratio.

Results: As expected, monozygotic twins are more concordant for autism than are dizygotic twins (69% vs. 23%). Monochorionic twins are no more likely to be concordant than monozygotic-dichorionic twins (53% vs. 67%). Autistic and non-autistic participants did not differ in average a-b ridge count, t(98) = -0.21, p = 0.84; asymmetry, t(88) = 0.13, p = 0.90; or digit ratio, t(90) = -0.83, p = 0.41. Digit ratio did show the expected sex effect, t(90) = 2.18, uncorrected p = 0.02. Zygosity is unrelated to cotwin similarity of a-b ridge count, t(45) = 0.03, p = 0.98 or asymmetry, t(35) = 0.96, p = 0.34, but digit ratios are more similar for monozygotic twins  than dizygotic same sex twins, t(40) = -2.48, uncorrected p = 0.02.

Conclusions: The current study failed to replicate prior findings linking lower a-b ridge counts and digit ratios with autism and did not find a relationship between dermatoglyphic asymmetry and autism. Collectively, these null results suggest that any nonshared prenatal environmental factors that are differentiated by chorionicity or that affect a-b ridge count, asymmetry, and digit ratio are not major players in the development of autism.