Design
This exploratory study was part of a larger placebo-controlled, double-blinded, randomised crossover trial investigating a range of social functions in adolescent females with eating disorders (ED). Participants were tested twice, approximately seven days apart [M (SD) = 7.95 (6.24)].
Participants
Forty-eight adolescent females aged 16 and 17 years were recruited via social media campaigns. Four participants were excluded from the final analyses for this study due to their BMI exceeding 30kg/m2 6.
To confirm eligibility, participants completed an online Qualtrics screening survey. The inclusion criteria required that participants were females aged 16 to 17 years, able to travel to the Children’s Clinical Research Facility at St. Mary’s Hospital, London. Exclusion criteria included insufficient English language skills and uncorrected hearing or sight impairment; regular or accomplished solo singers; pregnancy or breastfeeding; or recent experience of a significant life event. It was noted whether participants took an oral contraceptive pill due to the potential influence of menstrual cycle on endogenous oxytocin concentrations.
Materials
Intranasal sprays containing 24 international units (IU) of Syntocinon® or vehicle were supplied by Victoria Apotheke, Switzerland and over-the-counter urine pregnancy tests were provided.
Self-Report Questionnaires
A bespoke questionnaire via Qualtrics collected information on participant gender, age, ethnicity, parental education level, type of social support, date since last period, time since most recent meal, caffeine consumption, hours slept the previous night, and handedness.
A digital Visual Analog Scale (VAS) scale with a sliding marker from 0 to 10 anchored by “not at all” and “as much as I can imagine” asked how stressed the participant felt at baseline and post stressor. Other measures included in the parent study were not analysed here.
The 9-item Revised Children’s Anxiety and Depression scale – Social Phobia subscale29 determined social phobia symptomatology. For each statement, participants selected one of four responses: ‘Never’, ‘Sometimes’, ‘Often’, and ‘Always’. Possible scores for the social phobia subscale range from 0 to 27, with scores between 0 and 15 considered a ‘normal range’ for this population. Other measures included in the parent study were not analysed here.
Shortened ‘Sing-a-Song' Stress Test
Singing to an audience is a reliable method for inducing social-evaluative threat in participants. A shortened Sing-a-Song Stress Test that required a 50 second singing performance, was used30. After choosing a song, participants stood and performed it in front of the research team and the person accompanying them. As part of the sham story, and to increase stress, the researchers were seen to be scoring the participant’s performance. Nursery rhymes, humming and children’s songs were not allowed as these were considered less stressful to perform.
Procedures
The study was sponsored by Imperial College London and approvals were obtained from the Leicestershire South Research Ethics Committee (reference: 22/EM/0044; IRAS project ID: 297695). All methods were carried out in accordance with relevant guidelines/regulations. Testing was carried out in the Clinical Research Facility at St Mary’s Hospital, London.
Each visit lasted approximately one hour, and participants were instructed to bring a social support who remained present throughout the visit. This recommendation is based on evidence suggesting that the presence of a social support figure may enhance oxytocin’s effects31. Participants and their social support were compensated through a £50 and £20 online voucher, respectively, upon the completion of both visits.
Participants were told that the study was investigating whether oxytocin, an agent that may reduce anxiety and increase sociability, impacts social behaviour in adolescent females with and without ED. To minimise biases, both researchers and participants were blinded to whether subjects met the criteria for an ED.
At the start of the first visit, participants verbally confirmed eligibility and provided informed consent electronically. Experimenters recorded the blind weights and heights of participants and noted whether they had consumed alcohol in the last 12 hours, smoked or vaped, or had any conditions causing nasal congestion on the day. After participants cleared their nasal passages using a tissue, with a closed mouth, they forcefully exhaled through each nostril while keeping the other nostril closed, and self-reported which nostril felt more, and less, blocked. The decision on left or right nasal dominance was then confirmed using the Glatzel Mirror test20. Participants proceeded to self-administer either placebo or 24IU oxytocin by closing one nostril with their finger, tilting their head 32 and inhaling once, with a 30 second interval to allow for maximum absorption in either the dominant or non-dominant nostril. The selection of the nostril was pre-determined using a pseudo-randomised alternating design.
Oxytocin’s peak effects typically manifest between 20 and 90 minutes following intranasal administration21. Following a series of additional tasks detailed in the parent study, participants underwent the SSSTs, approximately 50 minutes post oxytocin administration. Subsequently, the stress VAS was administered before concluding the testing session.
Data
Data were analysed using Statistical Package for the Social Sciences (SPSS) software version 29. A p-value < 0.05 was considered statistically significant. No missing data were observed in this study6. Two-tailed t-tests were employed to investigate whether there were differences in stress scores between baseline and post-stressor readings in both the oxytocin and placebo conditions. A RM-ANOVA was performed for stress scores in the oxytocin and placebo conditions with baseline stress scores and RCADS social phobia subscale as covariates. An ANOVA was performed for stress scores within the oxytocin condition, with nostril dominance as the independent factor and RCADS social phobia score and baseline stress scores as covariates.