Salivary Sex Hormone Responses to Intranasal Oxytocin in Autistic and Neurotypical Women

Background: Oxytocin may be of therapeutic use for enhancing social functioning in autism spectrum conditions (ASC). Given recent evidence that (i) response to oxytocin treatment in ASC may depend on baseline hormone levels and (ii) oxytocin treatment may alter short-term production of other socially-relevant hormones, further investigation of the relationships among endogenous hormone levels and oxytocin treatment is warranted.

Objectives: This study aimed to assess baseline levels of estradiol and testosterone in adult women with and without an ASC diagnosis and to test relationships between these sex hormones and psychological variables of relevance to autism. Changes in endogenous estradiol and testosterone levels following single intranasal administration of oxytocin or placebo—and whether such changes were dependent on ASC diagnosis—were also explored.

Methods: As part of a larger fMRI experiment with a cross-over design (Bethlehem et al., 2017), saliva samples were collected from 47 women (age range 18–50) at three timepoints: (1) baseline, (2) after intranasal administration of placebo or 24 IU oxytocin (Syntocinon, Novartis), and (3) at the end of scanning (approximately 2 hours after administration). Participants comprised 16 women with an autism-spectrum condition (ASC group) and 31 neurotypical women (NT group) who completed the Autism-Spectrum Quotient (AQ) and Empathy Quotient (EQ). Salivary estradiol and testosterone were quantified by an independent lab blinded to the research hypotheses using commercially-available enzyme-linked immunosorbent assay kits (Salimetrics). Baseline sex steroid levels were calculated as the mean of the two pre-administration saliva samples per participant. The ratio of baseline estradiol to testosterone was then determined. Post-administration changes in estradiol and testosterone levels were calculated as percent change relative to individual baseline.

Results: Baseline sex steroid levels did not differ between age- and IQ-matched ASC and NT women. Baseline estradiol to testosterone ratio was negatively correlated with AQ (r = -0.34, p=0.02) and positively correlated with EQ (r=0.35, p=0.02). For the overall population, both estradiol and testosterone levels showed small but significant decreases over time, consistent with diurnal patterns. Percentage change in estradiol and testosterone from time 1 to time 3 differed significantly between the ASC and NT groups (p<0.01). For estradiol, the mean change was +12% for the ASC group and -9.9% for the NT group (Tukey HSD, p = 0.01) for combined drug conditions. For testosterone, the mean change was +7.8% for the ASC group and -14.6% for the NT group (Tukey HSD, p = 0.001). The percentage change testosterone showed an even larger between-group difference under the oxytocin condition (+14.4% for the ASC group vs. -15.5% for the NT group, Tukey HSD, p = 0.013). Interestingly, in the overall population under the oxytocin condition, there was a trend of a positive relationship between AQ score and percent change testosterone (r= 0.36, p = 0.06).

Conclusions: These findings support a relationship between sex steroid levels in women and psychological traits of relevance to autism. Further, they present evidence of a difference in endogenous sex steroid response to intranasal oxytocin between neurotypical women and women with autism or higher levels of autism-spectrum traits.