Plasma oxytocin concentrations following MDMA or intranasal oxytocin in humans

Ecstasy (MDMA) produces characteristic prosocial effects such as increased empathy and friendliness, which has motivated interest in its potential therapeutic use and in the neurobiology of social behaviour. Previous animal and human studies have implicated the neuropeptide oxytocin in affiliative behaviour, and animal work has specifically shown that MDMA increases oxytocin release and that blocking oxytocin receptors attenuates MDMA's prosocial effects. However, the evidence for MDMA-induced oxytocin release in humans was limited and confounded by naturalistic designs that lacked controlled dosing and timed blood sampling. Dumont and colleagues set out to test whether oral MDMA acutely increases peripheral oxytocin concentrations in humans and whether changes in oxytocin relate to MDMA's subjective prosocial effects. To address this question they carried out a double-blind, randomised, placebo-controlled crossover study in regular ecstasy users, measuring plasma MDMA and oxytocin concentrations alongside repeated ratings of prosocial mood.

The study used a double-blind, randomised, placebo-controlled crossover design. Each volunteer attended two study days separated by a 7-day washout and received either a single oral 100 mg capsule of MDMA or matched placebo on each occasion. Pre-study screening included medical history, physical and ECG examination, urine drug screening and standard blood tests; drug use was not allowed from 14 days before the first study day until completion. A practice session preceded the first study day to familiarise participants with procedures. Fifteen healthy volunteers (12 male, 3 female), aged 18–24 years and regular ecstasy users, were recruited. One participant did not refrain from drug use after the first study day and was denied further participation; the data from that participant's MDMA day were included. Two participants experienced mild transient anxiety after MDMA, producing partially missing data. The extracted text does not clearly state the final analytic sample for every outcome; some analyses report degrees of freedom consistent with a smaller sample (see Results). Blood sampling and assays: an indwelling catheter was used for repeated blood draws. The extracted schedule for oxytocin sampling in the methods section is unclear, but plasma MDMA was sampled at baseline and at 15, 60, 105, 150, 240 and 300 minutes after dosing. Reported assay methods were HPLC–diode array detection for MDMA, and extraction on C18 Sep-Pak columns followed by an in-house radioimmunoassay (125I-labelled oxytocin) for oxytocin; the oxytocin assay had within- and between-assay CVs of 2.2% and 6.6% at 7.2 pmol/l, an analytical range of 1–90 pmol/l and a sensitivity of 1 pmol/l. Average recovery was reported as 78–96%. Subjective prosocial effects were assessed at baseline and at 15, 60, 105, 150, 240 and 300 minutes using two items from the Bond and Lader Visual Analogue Mood Rating Scale that target antagonistic/amicable and withdrawn/gregarious dimensions. For statistical analysis, the investigators used mixed-model ANOVA (drug and time as fixed factors, subject as a random factor) for subjective measures. Because of variability in oxytocin measurements across timepoints, the total oxytocin response was summarised as area under the curve (AUC) and compared between drug conditions with paired t-tests. Relationships between subjective measures and biochemical concentrations were examined using a summary-statistics approach: for each subject, correlations between subjective ratings and concurrent oxytocin or MDMA concentrations were computed, and Wilcoxon signed-rank tests (exact p-values) were used to assess whether correlations differed from zero and whether correlations with oxytocin differed in strength from those with MDMA.

Sample and tolerability: fifteen volunteers completed at least one study day. One participant was excluded from further participation after failing to abstain from drug use; the MDMA-day data from that person were retained. Two participants experienced mild anxiety after MDMA that resolved within 60 minutes, producing partially missing data for them. MDMA pharmacokinetics: mean maximum plasma MDMA concentrations (Cmax) were reported as 222.7 mg/l (SEM 9.8 mg/l) occurring at 105 minutes post-dose. Concentrations declined minimally over time and averaged 174.6 mg/l (SEM 10.3 mg/l) at 300 minutes. Oxytocin response: oxytocin AUC (transformed to AUC data) was significantly greater following MDMA than placebo, paired t-test t(12) = 4.27, MSE = 1125.78, p = 0.001 (this degree of freedom indicates the oxytocin AUC comparison used 13 participants). Mean plasma oxytocin rose from 0.8 pmol/l (SEM 0.3 pmol/l) at baseline to an average maximum of 34.3 pmol/l (SEM 7.2 pmol/l) at about 110 minutes after dosing, then declined to 4.0 pmol/l (SEM 0.8 pmol/l) at 300 minutes. No order (treatment sequence) effects were reported. Subjective prosocial effects and associations: subjective amicability showed a significant drug (treatment) effect, F(1,165) = 9.7, p = 0.002. Subjective gregariousness showed a significant time effect, F(6,162) = 2.6, p = 0.018. Both amicability and gregariousness demonstrated significant treatment-by-time interactions (amicability: F(6,164) = 3.5, p = 0.003; gregariousness: F(6,162) = 4.0, p < 0.001). Correlation analyses using per-subject correlations between subjective ratings and biochemical levels showed median correlations with oxytocin of 0.37 for amicability (p = 0.001) and 0.29 for gregariousness (p = 0.049). Amicability also correlated with MDMA concentrations (median r = 0.23, p = 0.049), whereas gregariousness did not correlate with MDMA (median correlation reported as 0). Wilcoxon signed-rank tests indicated that correlations of both subjective measures with oxytocin were significantly stronger than their correlations with MDMA (amicability: p = 0.013; gregariousness: p = 0.030). Timing note: the reported peak in subjective prosocial effects occurred around 60 minutes, whereas measured peak plasma oxytocin concentrations were reported at about 110 minutes; the authors note this temporal offset in the Discussion and recommend denser oxytocin sampling in the interval where subjective effects were most pronounced.

Dumont and colleagues interpret their findings as evidence that oral MDMA robustly elevates peripheral oxytocin concentrations in humans and increases subjective prosocial feelings, with changes in prosocial mood showing stronger correlations with oxytocin than with plasma MDMA levels. They contrast these controlled findings with a previous naturalistic study that reported a small non-significant oxytocin increase, attributing the discrepancy to lack of controlled dosing and sampling in the earlier work. The authors relate their human findings to prior animal studies in which MDMA increased oxytocin and increased social interaction, effects that were blocked by oxytocin receptor antagonists and by 5-HT1A antagonism, proposing a plausible mechanism whereby oxytocin reduces amygdala-mediated social anxiety and thereby facilitates social behaviour. They emphasise, however, that their results are exploratory and cannot establish causation. Several limitations are acknowledged. First, oxytocin was measured peripherally in blood rather than in cerebrospinal fluid; the relationship between peripheral and central oxytocin dynamics remains undefined, and the observed delay between peak subjective effects and peak plasma oxytocin complicates interpretation. Second, prosocial effects were assessed with subjective ratings rather than objective social behaviour measures; the authors recommend using behavioural tasks such as the Trust Game or Dictator Game in future studies. Third, fixed MDMA dosing may have increased variance in oxytocin responses; weight-adjusted dosing could reduce this variability. Fourth, the timing of oxytocin sampling omitted the 20–95 minute window during which subjective effects were most pronounced, so denser sampling in that interval is advised. Finally, the design cannot determine whether oxytocin mediates MDMA's prosocial effects; the authors suggest that interaction studies using an oxytocin receptor antagonist (for example, atosiban) would be required, while noting some practical issues around receptor antagonism. In conclusion, the investigators present convergent human evidence that MDMA elevates peripheral oxytocin and enhances self-reported prosocial feelings, and they suggest these findings are consistent with a role for oxytocin in MDMA's social effects and relevant to conditions characterised by impaired social functioning, while stressing that further mechanistic and behavioural studies are needed.