Frequently Asked Questions

In this VA cohort, intravenous ketamine was associated with modest improvement in depression symptoms for a subset of patients with highly treatment-resistant depression. At 6 weeks, the mean PHQ-9 improvement was 4.6 points (SD = 6.8; n = 164), 47% met the minimally clinically important difference threshold, 26% achieved response, and 15% achieved remission. At 12 weeks, the mean improvement was 4.4 points, with 44% reaching minimally clinically important difference, 25% response, and 12% remission; at 26 weeks, the mean improvement was 4.7 points, with 50% reaching minimally clinically important difference, 28% response, and 13% remission.

Most patients received repeated treatment over many months rather than a single short series. In the 12 months after the initial infusion, 96% received additional infusions, with a mean total of 18 treatments (SD = 13; median = 16). The mean time from first to last infusion was 207 days (SD = 145; median = 245).

The infusion schedule generally started more frequently and then tapered to maintenance-style treatment. The mean interval between infusions was 5 days during the month after the first infusion, increased to 12 to 17 days during months 2 to 4, and then widened to 23 to 28 days during months 5 to 12. The authors interpreted this pattern as consistent with an initial acute phase followed by longer-term maintenance treatments every 3 to 4 weeks.

Among patients who ever met improvement criteria, clinically meaningful improvement usually occurred after multiple infusions over several weeks rather than immediately. The mean time to minimally clinically important difference was 51 days and 9 infusions (median 37 days and 8 infusions), the mean time to response was 56 days and 9 infusions (median 37 days and 9 infusions), and the mean time to remission was 56 days and 10 infusions (median 37 days and 9 infusions).

Yes. In bivariate regression, the only variable associated with greater PHQ-9 improvement at 6 weeks was the number of infusions received during that period. Each additional infusion was associated with a mean 0.7-point greater improvement in 6-week PHQ-9 score (b2 = 0.7; P = .002).

The cohort consisted of 215 VA patients with severe, highly treatment-resistant depressive illness and substantial psychiatric comorbidity. Their mean baseline PHQ-9 score was 18.6, 98% had a past-year diagnosis of major depressive disorder, 22% had received inpatient psychiatric treatment in the prior year, 13% had received repetitive transcranial magnetic stimulation, and 18% had received electroconvulsive therapy. Comorbidity was common, including 70% with posttraumatic stress disorder, 50% with other anxiety disorders, 27% with alcohol use disorders, and 27% with other substance use disorders.

In this study, patient demographic and diagnostic characteristics were not associated with change in PHQ-9 score at 6 weeks. The authors found that the only measured factor linked to greater 6-week improvement was the number of infusions, not age, demographic characteristics, or psychiatric diagnoses such as posttraumatic stress disorder.

The dose tended to increase over the course of treatment. The mean ketamine dose across all infusions was 59 mg (SD = 25; median = 54 mg), with a mean of 45 mg at the first infusion (SD = 14; median = 43 mg) and 66 mg at the last infusion (SD = 32; median = 60 mg).

Intravenous ketamine was used relatively rarely in VA during fiscal year 2020. Nineteen VA facilities treated at least 1 patient, but only 7 facilities treated 10 or more patients, and 50% of all patients were treated at the 2 highest-volume facilities. The authors noted that only 215 patients received the treatment out of the much larger population of VA patients with depression.

This was an observational medical record study, so it cannot establish causality the way a randomized clinical trial can. Important limitations included no direct safety assessment, PHQ-9 data only on infusion dates, possible underestimation of longer-term response if patients who stopped treatment had more sustained benefit, reliance on medical record diagnoses rather than structured interviews, and uncertain generalizability outside VA settings. The authors also noted that validated electronic medical record measures for ketamine-related adverse outcomes were not available.