Low-dose ketamine for treatment resistant depression in an academic clinical practice setting

Clinical depression is common and disabling, and a substantial minority of patients fail to respond to multiple standard antidepressant trials. Earlier research shows that a single sub‑anesthetic intravenous infusion of ketamine can produce rapid antidepressant effects in treatment‑resistant depression (TRD), with aggregated response rates around 61% in published randomised and open‑label studies. However, most of those prospective trials applied strict inclusion and exclusion criteria that limited psychiatric comorbidity, concurrent medications and suicide risk, raising concerns about the generalisability of their findings to routine clinical populations. This paper reports a retrospective analysis of outcomes following a single 0.5 mg/kg, 40‑minute intravenous ketamine infusion delivered in an outpatient academic clinical programme at the University of California, San Diego. Feifel and colleagues set out to examine the safety and efficacy of that initial infusion in a ‘‘naturalistic’’ TRD sample that included prevalent comorbidities and concurrent psychotropic medications, using self‑reported Beck Depression Inventory (BDI) scores at baseline, 1‑hour, 24‑hours and, when available, 7‑days post‑infusion as the primary outcome measures. The study aims to assess real‑world response and tolerability outside the constraints of typical clinical trials.

Patients were referred to the UCSD ketamine programme by themselves or by treating psychiatrists and were interviewed by the prescribing psychiatrist to determine whether the clinical risk–benefit balance favoured a trial of ketamine. Unlike prospective trials, there were no fixed a priori inclusion/exclusion criteria; candidacy was based on clinical judgement. Typical reasons supporting treatment were severe, functionally impairing depression or suicide risk after failure of multiple antidepressant trials (generally at least four), and some patients had previously received ECT or rTMS. Relative contraindications that did not automatically preclude treatment included a history of psychosis, uncontrolled hypertension, prior adverse reaction to ketamine, active medical conditions, substance use, pregnancy or breastfeeding. Patients provided written informed consent detailing risks and the non‑FDA‑approved status of the treatment. Infusions were delivered in a same‑day outpatient procedure centre. Ketamine was administered at 0.5 mg/kg based on actual body weight, infused intravenously over 40 minutes. Patients were allowed to continue concurrent psychotropic medications; morning benzodiazepines or opiates were held on the day of infusion. Monitoring included periodic vital signs, oxygen saturation and mental status during the infusion and for at least 30 minutes afterwards. Nurses and the attending psychiatrist assessed patients for ongoing mental status changes, orthostatic instability or other adverse effects prior to discharge. Patients completed the BDI prior to infusion, immediately before leaving the centre (≈1 hour post‑infusion), and were given a BDI to complete at 24 hours; a follow‑up visit typically occurred 24–72 hours after infusion. An additional 7‑day BDI was solicited from patients who reported substantial benefit at 24 hours. Feifel and colleagues conducted a retrospective chart review of the first 50 patients treated in the programme; charts were included if patients were ≥18 years old, had a primary diagnosis of Major Depressive Disorder (MDD) or Bipolar Disorder (BD), and had recorded pre‑infusion, 1‑hour and 24‑hour BDI scores. The institutional review board approved the review. One author abstracted de‑identified data using a predefined electronic abstraction sheet agreed by the study team. Extracted variables included demographics, psychiatric and medical history, prior treatments, concurrent medications, ketamine dose verification, BDI scores at the specified timepoints, nursing and physician notes, and safety data (blood pressure, heart rate, oxygenation). The primary outcome was response at 24 hours, defined as ≥50% reduction in BDI from baseline; remission was defined as BDI < 13. Changes in mean BDI were analysed with paired Student’s t‑tests (α < 0.05). Predefined hemodynamic thresholds (SBP increase >20 mmHg, DBP increase >10 mmHg, HR increase >25 bpm) were summarised with frequencies, and adverse events were identified from nursing documentation and medication administration records.

Of the first 50 charts reviewed, 41 met inclusion criteria; eight were excluded for missing baseline or 24‑hour BDI and one for age <18. No serious adverse events were identified in excluded charts. The analysed sample had a mean age of 48.6 ± 12.7 years. Thirty‑four patients (82.9%) had a primary diagnosis of MDD and seven (17.1%) had BD. Anxiety disorders were the most common psychiatric comorbidity (24 patients, 58.5%). Hypertension was the most common cardiovascular comorbidity (10 patients, 24.4%). Concurrent psychotropic medication use was common: patients were taking an average of 2.4 psychotropic agents at infusion and seven patients (17.1%) were taking no psychotropic medications at the time of ketamine. Efficacy outcomes showed rapid and substantial reductions in self‑reported depressive symptoms. The mean baseline BDI was 32.6 ± 8.4. At 1 hour post‑infusion the mean BDI fell to 16.8 ± 12.1 (49.5% reduction; p < 0.001) and at 24 hours to 16.0 ± 11.4 (51% reduction; p < 0.001). Twenty‑four hours after infusion, 22 of 41 patients (53.7%) met the predefined response criterion (≥50% reduction) and 17 (41.5%) met remission (BDI < 13). One‑hour response and remission rates were 51% and 44%, respectively. Patients with BD and MDD showed comparable mean percentage reductions at 24 hours (49% and 51%). Among the 22 responders at 24 hours, 16 had a recorded 7‑day BDI and 12 of those 16 (75%) remained responders at 7 days; when excluding charts missing 7‑day scores, the authors report a 7‑day response rate of 46% (16/35), comparable with prior prospective studies. Hemodynamic effects were generally modest and transient. Seven patients (17.1%) experienced an increase in systolic blood pressure >20 mmHg, 12 patients (29.3%) had an increase in diastolic blood pressure >10 mmHg, and one patient had a heart‑rate increase >25 bpm. No patient required antihypertensive intervention and no oxygen desaturation events (SaO2 < 93%) occurred. Adverse events recorded in the immediate peri‑infusion period included nausea requiring antiemetic in three patients (7.3%) and one episode of vomiting. One patient with significant baseline anxiety required stopping the infusion at 10 minutes and received 1 mg oral lorazepam; the infusion was restarted after symptom control and the patient met remission at 24 hours. Five patients had small (<20%) increases in BDI at 24 hours that were not considered clinically significant by patients or clinicians. Subjective experiences during infusion were described in all charts as a ‘‘psychedelic’’ phenomenon varying in intensity and content (perceptual changes, dissociation, emotional insight, relaxation or wellbeing). Among 18 charts that included detailed emotional descriptions, all but one characterised the experience as pleasant. Select short patient comments documented included: "It was like I was in a biological membrane…very pleasurable" and "it melted my armour (of depression)."

Feifel and colleagues interpret their findings as supporting a rapid antidepressant effect of a single low‑dose ketamine infusion in a real‑world outpatient TRD population, with generally favourable tolerability. The 24‑hour response rate of 53.7% is somewhat lower than the pooled 61% reported in prior prospective studies, but the authors note methodological differences that complicate direct comparison: this study used the self‑rated BDI rather than clinician‑administered scales (MADRS, HAM‑D), allowed concurrent medications, and was delivered in a naturalistic clinical setting rather than a tightly controlled trial. Those factors might reduce or modulate measured efficacy, and expectancy or selection biases could also influence open‑label or retrospective results. The discussion contrasts these findings with other open‑label series that enrolled less selected populations and reported lower response and greater immediate adverse effects; setting may be important, as the investigators observed poorer outcomes when ketamine was administered in noisy, chaotic ECT or PACU environments versus a dedicated outpatient procedure centre. Exploratory post‑hoc analyses in this cohort did not reveal clear differences in response or remission by gender, comorbid anxiety or medication use, but the authors emphasise the study was not powered for subgroup comparisons. Hemodynamic increases were common but transient and did not necessitate intervention. Limitations are acknowledged in detail: the retrospective design raises risks of selection and information bias and limits adverse‑event ascertainment to what was documented in charts; missing data (notably missing BDI scores) may bias estimates; lack of a control group prevents definitive attribution of effects to ketamine versus placebo or regression to the mean, although large effects in randomised trials elsewhere mitigate this concern; and the sample may be unrepresentative because patients self‑paid for treatment and the cohort was drawn from a single academic site with all patients assessed by a single prescribing psychiatrist. Taken together, the authors conclude that, within these constraints, a single 0.5 mg/kg ketamine infusion produced rapid antidepressant effects and was generally well tolerated in this naturalistic outpatient TRD sample, but they underline the need to interpret findings in light of the study's methodological limitations.