Is Prenatal Testosterone Exposure Associated with Early Vocabulary Development? A Prospective Cohort Study

Thursday, May 17, 2012: 11:00 AM
Osgoode Ballroom East (Sheraton Centre Toronto)
10:30 AM
L. Hollier1,2, E. Mattes3, M. T. Maybery2, J. A. Keelan4, M. Hickey5 and A. Whitehouse2,3, (1)Telethon Institute for Child Health Research, University of Western Australia, Perth, Australia, (2)School of Psychology, University of Western Australia, Perth, Australia, (3)Telethon Institute for Child Health Research, Perth, Australia, (4)School of Women’s and Infant’s Health, University of Western Australia, Perth, Australia, (5)Department of Obstetrics and Gynaecology, University of Melbourne, Melbourne, Australia
Background: Prenatal exposure to androgens is known to affect fetal brain maturation and later neurocognitive function.  Baron-Cohen (2002) proposed that exposure to enhanced concentrations of prenatal androgens, may play a causal role in autism.  Communication difficulties are at the core of ASD, and any aetiological theory must incorporate a plausible explanation of this symptom.  However, research on the effects of prenatal androgen exposure has been limited by indirect measures of androgens (e.g. 2D:4D ratio) and small unrepresentative samples.

Objectives: The aim of this study is to investigate whether circulating fetal testosterone concentrations are associated with language development in early childhood, using umbilical cord serum testosterone as a surrogate measure of prenatal testosterone exposure in a large unselected birth cohort. We hypothesised that higher fetal testosterone concentrations will be inversely related to spontaneous language ability in early childhood.

Methods: Umbilical cord serum samples taken immediately after delivery in a subset of the Western Australian Pregnancy Cohort Study (N=373; M=197, F=176) were assayed for testosterone by liquid chromatography-mass spectrometry. Expressive vocabulary was measured at two years of age using the Language Development Survey (LDS; Rescorla, 1989). A range of sociodemographic variables posited to influence neurocognitive outcomes were also investigated.

Results: Cord blood testosterone levels were significantly correlated with spontaneous language at two years in males (r=-.19, p=.007) but not females (r=.08). Multivariate regression analyses found that, when a range of sociodemographic variables were controlled for, testosterone concentrations significantly predicted spontaneous language development in males, such that higher concentrations were associated with lower LDS scores (β=-.159, p=.025).

Conclusions: These findings indicate that higher circulating fetal testosterone concentrations at birth may be associated with slower language development in early childhood among males. This relationship did not extend to females.  The significant association in males may have implications for developmental disorders that involve poor language development, such as autism.

| More