BACKGROUND

Levels of steroid hormones in the first three months of life, a period referred to as 'mini-puberty', are one of the earliest physiological differences between typical males and females postnatally. Autistic traits also show consistent typical sex differences in later infancy, after the 18 month of life. Previous studies have shown testosterone is associated with later levels of autistic traits. Studies testing if postnatal testosterone levels are associated with autistic traits have reported null results. No studies to date have investigated mini-puberty longitudinally or tested for interactions with baseline sex differences or familial likelihood of autism.

METHODS

The 'Cambridge Human Imaging and Longitudinal Development Study' (CHILD) is a prospective enriched cohort study in Cambridge, UK. It includes physiological measurements in early infancy, as well as neurodevelopmental follow-ups over the first two years of life. A subset of the cohort also includes children with a family history of autism (a diagnosed parent or sibling). Salivary testosterone levels were assessed at two time-points, just after the 2 and 6 month of life. Autistic traits were measured using the Quantitative Checklist of Autism in Toddlers (Q-CHAT) when the children were 18 months of age.

RESULTS

Salivary testosterone levels were significantly higher during 'mini-puberty' in the 2 and 3 month of life, compared to after the 6 month of life, in both males and females. There was no significant sex difference at either time-point. Log-transformed testosterone levels were not associated with autistic traits (Q-CHAT). There was no interaction effect with infant sex, autism family history or baseline testosterone levels after mini-puberty (at >6 months of age).

CONCLUSION

Both male and female infants have elevated levels of salivary testosterone during mini-puberty but in this relatively small sample this was not associated with their later autistic traits at 18 months or their family history of autism. This suggests that rather than postnatal testosterone levels are more relevant for understanding the causes of autism. Future studies should test these relationships in larger samples.