Pilot Data on Salivary Oxytocin as a Biomarker of LSD Response in Patients with Major Depressive Disorder

Cazorla and colleagues frame the study around the need for objective biomarkers to characterise pharmacodynamic effects and predict response in LSD-assisted psychotherapy for major depressive disorder (MDD). They note that LSD acts mainly via partial agonism at 5-HT2A receptors and that this receptor-mediated cascade can include release of neuropeptides such as oxytocin, which is implicated in social bonding, stress modulation and psychological flexibility. Prior work in healthy volunteers found rises in plasma oxytocin 90–180 minutes after oral LSD, and preclinical rodent studies showed 5-HT2A-dependent oxytocin release; however, oxytocin dynamics have not been characterised in psychiatric patients, and salivary oxytocin (a minimally invasive option suitable for repeated sampling) had not been evaluated during LSD administration. This pilot observational study therefore aimed to describe salivary oxytocin trajectories across the acute pharmacodynamic window of a single LSD-assisted psychotherapy session in patients with treatment-resistant MDD. A secondary, exploratory aim was to record concurrent momentary ratings of subjective drug intensity, to compare temporal patterns between oxytocin release and perceived psychedelic effects. The investigators positioned this work as feasibility data to inform larger, controlled trials that could test whether oxytocin reactivity relates to clinical outcomes or patient stratification.

The study was a single-centre, observational pilot embedded within routine clinical care at the Psychedelic Program, Division of Addictology, Geneva University Hospitals, conducted under compassionate-use authorisations. Ethical approval and trial registration were obtained. Data collection ran from September 2024 to February 2025. Patients invited to participate had already secured individual compassionate-use approval for LSD-assisted psychotherapy and were approached to provide additional saliva samples for research; consenting patients received the usual clinical treatment without modification. Eligibility required DSM-5 major depressive disorder with treatment resistance (at least two failed antidepressant trials of different classes), age 18–55, and receipt of a first or second LSD treatment at 100 or 150 μg. Routine clinical exclusions (psychotic or bipolar disorder, imminent suicide risk, significant medical comorbidity, pregnancy/breastfeeding, systemic corticosteroid use) were applied. Twelve participants (5 female, 7 male; mean age 43.9, SD 8.79) consented. Two participants had insufficient saliva at multiple time points and their salivary data could not be analysed; sampling procedures were adapted thereafter so that ten participants contributed analysable salivary data. Subjective intensity ratings were available for 12 participants. On the treatment day patients received LSD orally at 09:00 under continuous nursing supervision in a therapeutic clinical setting. Saliva was collected with Salivette® synthetic swabs at four time points: baseline (−10 minutes), 60, 90 and 180 minutes post-dose. Samples were frozen then stored at −80 °C and shipped on dry ice for assay. Oxytocin concentrations (pg/ml) were measured by radioimmunoassay (RIA) in a single batch; samples underwent evaporation and reconstitution with assay buffer and an oxytocin-specific rabbit antibody. Reported assay performance included a detection limit of 0.1–0.5 pg/sample and intra- and inter-assay coefficients of variation below 10%. Momentary subjective drug intensity was rated on a 0–10 scale at the post-dose saliva time points (60, 90 and 180 minutes). For statistical analysis the investigators used a linear mixed model (LMM) with oxytocin concentration as the dependent variable, time (four categorical levels) as a fixed effect, participant-specific random intercepts to account for repeated measures and individual variability, and an autoregressive first-order (AR1) covariance structure for within-subject measurements. The model was estimated with restricted maximum likelihood (REML). A Friedman ANOVA, a non-parametric test appropriate for repeated ordinal data, assessed changes in perceived intensity across the three post-dose time points.

Twelve participants provided subjective intensity ratings; salivary oxytocin analyses were impeded for two participants due to insufficient sample volume, leaving ten with analysable saliva and with some missing data handled within the LMM framework. The LMM showed a significant main effect of time on salivary oxytocin, F(3, 19.8) = 14.9, p < 0.001, indicating oxytocin concentrations changed across the four measurement points (pre-LSD, 60, 90 and 180 minutes). The paper reports estimated marginal means and standard errors by time point in a figure (these numerical values are not presented in the extracted text). In the random-effects output, the participant-specific random intercept variance was estimated at 0.004 (SE = 0.01) and was not significantly different from zero (Z = 0.40, p = 0.690), with a 95% confidence interval reported as approximately [0.000, 0.587]. The AR1 parameter for within-subject covariance was 0.009 (SE = 0.011, 95% CI = [0.001, 0.097]) and likewise was not significantly different from zero (Z = 0.81, p = 0.418), indicating no detectable autocorrelation between adjacent time points in this sample. Subjective drug intensity varied significantly over time: the Friedman ANOVA yielded χ2(2, N = 12) = 21.273, p < 0.001. In the Discussion the investigators note that salivary oxytocin peaked at 90 minutes post-dose, paralleling the temporal course of perceived psychedelic intensity; the specific numerical peak values are shown in the figure but are not provided in the extracted text.

The investigators report this as the first characterisation of salivary oxytocin dynamics after acute LSD administration in patients with MDD and show that a non-invasive salivary sampling protocol can capture temporal oxytocin changes comparable to prior plasma-based observations in healthy volunteers. They highlight a significant modulation of salivary oxytocin over time, with an apparent peak at 90 minutes post-administration that parallels subjectively rated drug intensity. This pattern is consistent with earlier human plasma studies and with preclinical evidence implicating 5-HT2A-mediated oxytocin release; the authors also situate LSD's oxytocin effects relative to MDMA, which is known for more robust oxytocin release yet suggests overlapping neuroendocrine pathways across compounds. Cazorla and colleagues underscore the feasibility of using salivary oxytocin as a pharmacodynamic biomarker in psychedelic-assisted psychotherapy, emphasising the advantages of a minimally invasive measure for repeated sampling. They caution, however, that the pilot, open-label design and small sample size prevent definitive conclusions about oxytocin as a predictive biomarker for clinical response. Important limitations acknowledged include the absence of placebo or active control conditions, heterogeneous concurrent antidepressant use (patients continued usual medication), lack of control for menstrual cycle phase or hormonal contraception among female participants, and no longitudinal oxytocin follow-up to relate acute reactivity to clinical outcomes. The authors recommend that future, larger and rigorously controlled trials account for sex-hormone influences and antidepressant class effects, and include adequately powered correlational and longitudinal analyses to test whether baseline or acute oxytocin profiles predict reductions in depressive symptoms or improvements in mental flexibility. They conclude that the present findings provide preliminary support for further investigation of oxytocin dynamics in the context of LSD-assisted therapy for MDD.