Oxytocin in response to MDMA provocation test in patients with arginine vasopressin deficiency (central diabetes insipidus): a single-centre, case-control study with nested, randomised, double-blind, placebo-controlled crossover trial

Arginine vasopressin and oxytocin are closely related nine–amino-acid neuropeptides produced in neighbouring hypothalamic nuclei (SON/PVN) and released from the posterior pituitary. Disruption of the hypothalamic–pituitary axis from causes such as inflammation, tumours, or head trauma can produce arginine vasopressin deficiency (central diabetes insipidus), a condition clinically characterised by polyuria and polydipsia and commonly treated with desmopressin. Despite adequate desmopressin therapy, many patients continue to report socioemotional and psychological symptoms (heightened anxiety, alexithymia, depressed mood) that reduce quality of life, and the anatomical proximity of oxytocin-producing neurons raises the possibility that an accompanying oxytocin deficiency could contribute to these residual symptoms. Previous attempts to detect oxytocin deficiency using basal peripheral measurements have been inconclusive, and no standard provocation test for oxytocin has been established. Atila and colleagues therefore used 3,4-methylenedioxymethamphetamine (MDMA), a psychoactive compound known to evoke robust central and peripheral oxytocin release and prosocial subjective effects, as a biochemical and psychoactive provocation test. The study aimed to compare MDMA-induced oxytocin responses and associated subjective, emotional-recognition, and empathy effects in patients with arginine vasopressin deficiency (central diabetes insipidus) versus matched healthy controls, testing the hypothesis that patients would show a blunted oxytocin response and attenuated prosocial subjective effects in response to MDMA. By doing so, the investigators sought to determine whether an oxytocin deficiency could be demonstrated as a clinically relevant consequence of posterior pituitary dysfunction.

This single-centre case–control study included a nested, randomised, double-blind, placebo-controlled crossover trial of MDMA versus placebo for each participant. The work took place at University Hospital Basel, Switzerland, with regulatory and ethics approvals and MDMA authorised for research use. Adult patients (18–65 years) with a confirmed diagnosis of arginine vasopressin deficiency (central diabetes insipidus) were recruited and matched 1:1 to healthy controls by age (±3 years), sex, BMI (±2 kg/m²), menopausal status, and hormonal contraceptive use. Participants were screened to exclude significant somatic or psychiatric comorbidities, substance use, and medications that could interfere with study drugs; full criteria are provided in the appendix (not reproduced here). Randomisation used blocks of four to determine whether each participant received MDMA (100 mg orally, given as four 25 mg GMP-prepared capsules) or matched placebo in the first experimental session; allocation was concealed and the MDMA/placebo order was masked to participants and outcome assessors. Each participant completed two 7-h main visits separated by at least a 14-day washout. Visits involved fasting, a standardised breakfast and lunch schedule, continuous medical supervision, an intravenous catheter for repeated blood sampling, and urine drug screening. Female participants were tested during the follicular phase and pregnancy screening was performed before visits. Primary biological sampling consisted of plasma oxytocin measured at 0, 90, 120, 150, 180, and 300 min after drug administration; copeptin (a surrogate for vasopressin) was measured at 0 and 120 min; and sodium and prolactin were also assayed. Oxytocin was extracted from EDTA plasma and quantified by ELISA; all hormone assays were run blinded in a single batch. Pharmacokinetics of MDMA and MDA were measured by HPLC–MS/MS. Subjective effects were assessed repeatedly (0–360 min) using 10-point visual analogue scales for 20 items known to be sensitive to MDMA effects, and acute anxiety was measured with the STAI-State subscale at 0 and 180 min. At the expected oxytocin peak (around 150 min), participants completed the Multifaceted Empathy Test and a Facial Emotion Recognition Task. The primary outcome was the net incremental area under the curve (AUC) of plasma oxytocin between 0 and 300 min after MDMA versus placebo in patients and controls. The study was powered on expert assumptions and prior data to detect a clinically meaningful blunting of the response in patients, yielding a planned sample size of 15 per group. The AUC was analysed using a restricted maximum-likelihood linear mixed-effects model with fixed effects for treatment, participant group, and their interaction, and baseline oxytocin as a covariate; a random intercept accounted for within-participant correlation. Bootstrapping (1,000 replicates) was used to estimate percentage differences and 95% CIs. Secondary and exploratory endpoints (psychological questionnaires, subjective-effect AUCs, emotion tasks, safety measures) were described using appropriate summary statistics, and correlations between oxytocin changes and subjective effects were explored. No imputation was planned; all enrolled participants completed the study and were included in analyses.

Fifteen patients with arginine vasopressin deficiency (central diabetes insipidus) and 15 matched healthy controls were enrolled between Feb 1, 2021, and May 1, 2022, and all completed the protocol. Median ages were 34 years (IQR 25–46) in patients and 35 years (26–48) in controls; each group comprised eight female and seven male participants. In the patient group, eight (53%) had isolated posterior pituitary dysfunction and seven (47%) had combined pituitary dysfunction. Baseline psychological questionnaires showed higher trait anxiety and other adverse psychological features in patients: median STAI-Trait was 41 points (IQR 34–48) in patients versus 28 points (24–31) in controls. The primary biochemical finding was a marked difference in the plasma oxytocin response to MDMA. In healthy controls, MDMA produced an approximately eight‑fold increase in oxytocin, with a median AUC of 102,095 pg/mL (IQR 41,782–129,565). By contrast, patients showed only a minimal increase, with a median AUC of 6,446 pg/mL (IQR 1,291–11,577). Under placebo, AUCs were close to zero or slightly negative in both groups (controls 2,175 pg/mL, patients −1,343 pg/mL). The net incremental AUC difference (MDMA condition, adjusted for baseline oxytocin) was 85,678 pg/mL (95% CI 63,356–108,000), p<0.0001, corresponding to an 82% higher AUC in healthy controls than in patients (95% CI 70–186) according to bootstrap estimates. Subjective effects mirrored the biochemical results. Baseline state anxiety (STAI-State) was similar between groups, but at the expected peak oxytocin timepoint MDMA reduced state anxiety in controls to a median of 24 points (IQR 22–27) while patients’ scores remained essentially unchanged at 30 points (29–37). On visual analogue scales, patients showed absent or substantially lower MDMA-induced changes across many items (any effect, good/bad effect, liking, feeling high, stimulation, fear reduction, satisfaction, happiness, trust, talkative, openness). Moderate correlations were observed between maximal changes in some subjective effects and peak oxytocin concentrations. In emotion tasks, healthy controls under MDMA showed reduced recognition of negative emotions (anger, sadness, fear) with preserved recognition of happy/neutral faces, and increased explicit and implicit emotional empathy for positive stimuli. Patients showed a different pattern: MDMA impaired recognition of happy and sad faces, produced lower or contradirectional changes in positive-empathy ratings, and did not affect cognitive empathy in either group. Safety data indicated that MDMA moderately increased blood pressure, heart rate, and body temperature with no major differences between groups. Acute adverse-effects counts (66-item checklist) were similar: mean acute complaints 6.6 (SD 2.5) for patients and 6.3 (3.5) for controls; subacute (3-day) means were 3.6 (2.5) and 3.6 (2.7), respectively. The most frequent acute complaints included headache, tiredness, and lack of appetite. Transient mild hypokalaemia occurred in two (13%) controls and four (27%) patients; no new hyponatraemia episodes occurred during sessions, although one patient had mild hyponatraemia at baseline related to a slight desmopressin overdose. No serious adverse events were reported.

Atila and colleagues present three principal conclusions. First, MDMA produced the anticipated strong increase in peripheral oxytocin in healthy controls but elicited no notable oxytocin increase in patients with arginine vasopressin deficiency (central diabetes insipidus). Second, the blunted oxytocin response in patients was accompanied by markedly attenuated MDMA-induced subjective prosocial, empathic, and anxiolytic effects. Third, patients showed higher baseline levels of anxiety, alexithymia, and depressive symptoms despite the exclusion of active psychiatric disorders, suggesting residual socioemotional morbidity. The investigators position these results against previous work that relied mainly on basal peripheral oxytocin measures, which have proved unreliable; by using a provocation paradigm with MDMA they argue that a functional oxytocin deficiency can be unmasked in posterior pituitary dysfunction. The pattern of reduced oxytocin release and diminished central socioemotional effects supports a role for oxytocin in the characteristic behavioural and emotional actions of MDMA, and implies that oxytocin deficiency could contribute to the neuropsychiatric sequelae observed in some patients with hypothalamic–pituitary injury or disease. The authors also note consistency with prior observations in craniopharyngioma patients and animal oxytocin‑knockout models linking oxytocin absence to impaired social memory and increased anxiety. The paper acknowledges several limitations stated by the authors. It was a single-centre trial with a small sample size that limited subgroup analyses and generalisability; blinding could not be applied across the patient versus control grouping (that is, the trial was not blinded with respect to both groups), although drug/placebo assignment was masked. Concerns about assay accuracy for oxytocin measurements are noted, but all samples were processed and assayed in one blinded batch and similar relative increases after MDMA have been reported across assay types in prior studies. Selection bias is possible despite strict eligibility criteria excluding active psychiatric illness, so the sample may not fully represent the broader population of patients with arginine vasopressin deficiency. Concluding remarks by the authors characterise the findings as evidence for a clinically relevant oxytocin deficiency in patients with arginine vasopressin deficiency (central diabetes insipidus), suggesting a potential new hypothalamic–pituitary disease entity. They propose that future research should investigate whether oxytocin replacement (for example intranasal oxytocin) can ameliorate residual socioemotional symptoms in this patient group and recommend further studies in larger populations to evaluate therapeutic implications.