A phase 2 trial of inhaled nitrous oxide for treatment-resistant major depression

In this randomized, controlled Phase II crossover trial, L. and colleagues reported several key observations about single 1-hour inhalations of nitrous oxide for severe treatment-resistant major depression. The trial compared inhaled nitrous oxide at 50% and 25% concentrations against placebo and assessed antidepressant effects and adverse events over time. First, both 50% and 25% nitrous oxide produced rapid antidepressant effects in this severely treatment-resistant population. The authors further observed that the magnitude of antidepressant benefit increased over time in some patients and persisted for up to 4 weeks in certain cases. Adverse events were substantially less common at the lower (25%) concentration. Second, categorical clinical outcomes measured three months after study initiation showed a high overall rate of improvement in this sample: 85% of patients had improved, 55% met criteria for treatment response, and 40% were in remission at study completion. By treatment arm, after placebo one of nine patients had a treatment response (11.1%) and one of nine was in remission (11.1%). After 25% nitrous oxide, three of nine patients had a treatment response (33.3%; relative risk (RR) 2.50; 95% CI, 0.43 to 16.30) and two of nine were in remission (22.2%; RR 1.82; 95% CI, 0.27 to 12.84). After 50% nitrous oxide, 5 of 12 patients had a treatment response (41.7%; RR 2.94; 95% CI, 0.57 to 18.02) and 5 of 12 were in remission (41.7%; RR 2.94; 95% CI, 0.57 to 18.02). Third, although 25% and 50% nitrous oxide showed roughly equivalent antidepressant efficacy overall, there was evidence of a dose–response relationship at the 2-week follow-up, and the lower dose had a markedly reduced rate of adverse side effects. The authors also note considerable interindividual variability in time courses of response; incorporating this variability into their statistical models did not eliminate the significance of treatment-related effects. Differences between categorical outcomes were assessed by Fisher's exact test, and RRs with 95% confidence intervals were calculated using the Koopman asymptotic score.