Ketamine for treatment-resistant obsessive-compulsive disorder: Double-blind active-controlled crossover study

Obsessive-Compulsive Disorder (OCD) is characterised by intrusive thoughts and repetitive behaviours and causes substantial impairment in quality of life; many patients remain symptomatic despite standard pharmacological and psychological treatments. Ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, has robust evidence for rapid short-term benefit in treatment-resistant major depressive disorder and growing interest for other psychiatric indications. Previous clinical work in OCD is sparse: one small crossover trial of IV ketamine reported a 50% response rate versus 0% with placebo but suffered from carryover effects, and other attempts (including intranasal ketamine) have been inconclusive or underpowered. Key unanswered questions include the robustness and specificity of ketamine’s effect in OCD, optimal dosing, and the duration of benefit. Beaglehole and colleagues set out to evaluate the short-term efficacy and tolerability of intramuscular (IM) racemic ketamine at two doses (0.5 mg/kg and 1.0 mg/kg) compared with an active psychoactive control (fentanyl 50 µg) in a treatment-resistant OCD outpatient sample. The trial was designed as a randomised, double-blind, three-way within-subject active-controlled crossover to characterise symptom change over the first week after a single IM dose and to assess acute dissociative and cardiovascular effects.

This was a randomised, double-blind, psychoactive-controlled, three-way within-subject crossover study conducted in two community sites in New Zealand. Eligible participants were aged 18–50 years with DSM-5 OCD judged to be treatment-resistant—operationalised as failure to respond to at least two adequate pharmacological trials and one relevant psychotherapy—and a baseline Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) score >26. Participants were required to have a Montgomery–Åsberg Depression Rating Scale (MADRS) score ≤20 at screening. Exclusion criteria included serious medical illness, schizophrenia, bipolar disorder, current psychosis, significant suicidal ideation, recent substance dependence, pregnancy or lactation, and history of seizures or head injury. Those on stable medications or receiving ongoing psychotherapy could continue these treatments but were not to start new interventions during the trial. Study treatments were single IM injections in the deltoid of racemic ketamine 0.5 mg/kg, ketamine 1.0 mg/kg, and fentanyl 50 µg (active control). Doses were administered in three sessions separated by at least 1 week; investigators premedicated participants with oral ondansetron 4 mg 1 h before dosing to reduce nausea. Randomisation used a computer-generated balanced code; after two early dropouts related to poor tolerability of the 1.0 mg/kg dose, the randomisation schedule was amended so that 1.0 mg/kg ketamine would only be given following a prior 0.5 mg/kg exposure in this OCD cohort. Blinding was maintained for raters and participants, although the investigators did not systematically assess blinding success. Primary outcome measurement was the Y-BOCS administered pre-dose and at 1, 2, 24 and 168 h post-dose, with instructions to participants to consider the interval since the last scale administration when answering items intended for longer intervals. Tolerability and acute dissociation were assessed using the Clinician-Administered Dissociative States Scale (CADSS) pre-dose and at 30 and 60 min post-dose. Vital signs were recorded pre-dose and at 15, 30, 45, 60, 90 and 120 min post-dose; bladder symptoms were screened with the BPIC-SS. Cognitive screening (orientation and Trail Making tests) was performed before discharge, and participants remained under observation for at least 2 h. Safety oversight included a Data Safety and Monitoring Committee review of blinded safety data. Statistical analyses used repeated measures analysis of variance (ANOVA) with extraction of orthogonal polynomials for the dimensional factors dose and time to assess effects on total Y-BOCS and CADSS scores. Responder status was predefined as >50% reduction in Y-BOCS at 24 h. Missing Y-BOCS data (<1.5% of total) were interpolated from intact timepoints. The sample size calculation, based on prior crossover data, indicated that with 12 subjects per arm and alpha = 0.05 the study would have approximately 88% power to detect a 50% response in a ketamine arm versus none in control; the trial recruited 12 participants to the OCD cohort.

Twelve participants were randomised and 10 completed all three dosing sessions; two participants dropped out after a single dose because they could not tolerate the acute effects of 1.0 mg/kg ketamine. The completer sample comprised seven females and three males, all New Zealand European, with a mean age of 33 years (range 23–49). Baseline mean (SD) Y-BOCS was 29.9 (3.9). Psychiatric comorbidity was common (predominantly major depressive disorder and generalized anxiety disorder), and participants had a mean (SD) of 3.9 (1.7) failed antidepressant trials prior to enrolment. Y-BOCS scores declined after dosing for all treatments but with greater reductions following both ketamine doses compared with fentanyl. A linear dose effect was observed across treatments (dose [linear], F(1, 9) = 6.5, p = 0.031). Symptom change was maximal at 1–2 h post-dose (time [quadratic] and time [cubic], F(1, 9) > 30, p < 0.001), and a steady separation of scores persisted out to 168 h, most clearly for the 0.5 mg/kg ketamine condition (dose [quadratic] × time [linear], F(1, 9) = 6.1, p = 0.036). Post hoc ANOVAs at 24 h showed significant post–pre differences between ketamine 0.5 mg/kg and fentanyl (F(1, 9) = 8.0, p = 0.020) and between ketamine 1.0 mg/kg and fentanyl (F(1, 9) = 8.5, p = 0.017), but no significant difference between the two ketamine doses (F(1, 9) = 0.7, p = 0.443). Responder proportions (>50% Y-BOCS reduction at 24 h) were 10% after fentanyl, 60% after ketamine 0.5 mg/kg and 18% after ketamine 1.0 mg/kg. On safety and tolerability, all patients reported dissociative symptoms after ketamine, typically beginning 3–5 min post-injection with peak intensity around 15–30 min and gradual resolution thereafter. CADSS scores peaked at 30 min and were highest after 1.0 mg/kg (drug [linear], F(1, 9) = 19.28, p = 0.002; overall drug effect F(1, 9) = 20.45, p = 0.001). Cardiovascular changes were transient: at 30 min mean systolic blood pressure changes (SE) were reported as -12 (255), 1 (131) and 12 (277) mmHg for fentanyl, ketamine 0.5 mg/kg and ketamine 1.0 mg/kg respectively, with diastolic changes of -5 (64), 7 (69) and 8 (46) mmHg; blood pressure largely normalised by 60 min. The most common adverse effects after ketamine were dissociation, blurred vision, lightheadedness or sedation and perioral numbness; the most common effect after fentanyl was brief mild drowsiness. No serious adverse events occurred, but two participants withdrew due to intolerable dissociative effects at 1.0 mg/kg.

Beaglehole and colleagues report that single IM doses of ketamine (0.5 mg/kg and 1.0 mg/kg) produced greater short-term reductions in OCD symptoms than an active fentanyl control in a small, treatment-resistant outpatient sample. Symptom improvements emerged as early as 1 h, were evident at 2 and 24 h, and showed some residual separation out to 168 h though effects were not robustly long-lived; this pattern is consistent with prior observations that ketamine’s clinical benefits are relatively short-term for most disorders. The study confirmed dose-related acute dissociative and transient cardiovascular effects, and two participants dropped out after receiving 1.0 mg/kg because they could not tolerate dissociation. The investigators note that the 0.5 mg/kg dose may offer an advantage in tolerability and observed a higher 24-h responder rate after 0.5 mg/kg than after 1.0 mg/kg, though the 24-h Y-BOCS difference between ketamine doses was not statistically significant. Clinical experience in the trial suggested that prior exposure to a lower ketamine dose improved tolerability of a subsequent higher dose, and the authors therefore recommend that future OCD studies consider starting at lower doses and using strategies to reduce dissociation—such as slower IV infusion or oral administration—to improve tolerability. The authors acknowledge important limitations that temper interpretation. The small sample size and short follow-up restrict conclusions about longer-term efficacy and safety. Maintaining blinding was challenging: fentanyl at the chosen dose did not produce dissociative effects similar to ketamine, and the prominent CADSS differences suggest the active control did not fully preserve the blind, leaving open the possibility of expectation effects contributing to the observed benefit. Because the trial used a crossover design in which participants received all treatments, comparisons by experience across conditions could further compromise blinding. Finally, the study focused on acute single-dose effects; the durability of benefit, optimal dosing regimen (including repeated dosing), relapse rates and longer-term safety remain unknown. In conclusion, the authors interpret their findings as preliminary evidence that IM ketamine can produce short-term symptomatic benefit in treatment-resistant OCD, but they call for larger, longer-term and carefully controlled studies to determine optimal dosing, tolerability strategies, and whether repeated or maintenance regimens can provide sustained clinical benefit.