IV low dose ketamine infusions for treatment resistant depression: Results from a five-year study at a free public clinic in an academic hospital

The investigators performed a retrospective cohort analysis of medical records from the Mood Disorders Research and Treatment Service at Providence Care Hospital, covering patients referred for IV low dose ketamine between October 2018 and September 2023. All outpatients aged 18–75 years who met DSM‑5 criteria for a primary diagnosis of major depressive disorder with TRD (defined here as failure of two or more adequate antidepressant trials) and who received IV ketamine at the clinic were included; no additional exclusion criteria were applied for the analysis. Routine clinic exclusion criteria for receiving ketamine (and therefore effectively applied to the treated cohort) were substance abuse, psychosis, a primary personality disorder diagnosis, uncontrolled hypertension, prior adverse reaction to ketamine, and pregnancy or breastfeeding. Patients were asked to discontinue benzodiazepines, naltrexone, lamotrigine, gabapentin and pregabalin prior to treatment under physician supervision; other psychiatric medications were generally continued. The clinic's standard acute treatment course comprised eight infusions given twice weekly over four weeks. Each infusion delivered 0.5 mg/kg ketamine intravenously over 40 minutes, with patients resting during administration and for 30 minutes after. Nursing staff recorded blood pressure, oxygen saturation, cardiac rhythm and common side effects every 10 minutes during sessions. Primary outcome assessment used the Clinician Global Impression–Severity (CGI‑S), with predefined thresholds for response (≥2‑point reduction or post‑treatment score of 3), partial response (1‑point reduction where post‑treatment CGI‑S remained >3) and remission (post‑treatment CGI‑S ≤2). Secondary outcomes included self‑reported Beck Depression Inventory‑II (BDI‑II) and clinician‑rated MADRS; item Q9 of BDI‑II and Q10 of MADRS were used to index SI. Depersonalization was assessed with the CADSS‑6. CGI‑S, BDI‑II and MADRS were collected up to 24 hours before the first infusion and up to 24 hours after the eighth infusion; CADSS‑6 was recorded up to 24 hours after the first and eighth sessions. Statistical analyses were performed in IBM SPSS (v24) at a two‑tailed significance threshold of p < 0.05. Both per‑protocol (completers) and intention‑to‑treat (ITT) approaches were used; for ITT missing data were imputed using multiple imputation (five iterations, linear regression imputation with pseudo‑random numbers generated by the Mersenne Twister algorithm). Pre‑to‑post changes on psychiatric scales were tested with paired t‑tests and effect sizes reported as Cohen's d with 95% confidence intervals. Proportions meeting response, partial response and remission criteria were reported. The impact of age, sex, marital status and concurrent psychiatric medications on outcomes was explored with one‑ and two‑way multivariate analysis of variance (MANOVA) using Wilk's Lambda and Tukey HSD for post‑hoc tests. Side effects and dropouts were summarised as proportions and categorised by reported severity (mild/medium/severe) according to clinical guidelines; the authors note that absolute vital sign values were not recorded in the dataset.

Seventy‑one outpatients treated with IV low dose ketamine between October 2018 and September 2023 met inclusion for analysis. Using the three clinical scales (CGI‑S, BDI‑II and MADRS), patients showed a statistically significant reduction in depressive severity from baseline to treatment endpoint, with large effect sizes reported (exact effect‑size values and confidence intervals are presented in the paper's tables). On the primary outcome (CGI‑S) 54.93% of patients met the study definition of treatment response, 23.94% achieved remission, and 14.09% had a partial response. The MANOVA found no statistically significant effect of the examined sociodemographic or clinical variables on treatment response, either singly or in combination (non‑significant Wilk's Lambda and between‑subjects effects). Suicide ideation, as measured by BDI‑II item Q9 and MADRS item Q10, declined significantly from baseline to endpoint. In the per‑protocol analysis the BDI‑II Q9 change showed a medium effect size while MADRS Q10 showed a large effect; in the ITT analysis the BDI‑II Q9 change became non‑significant but MADRS Q10 retained a statistically significant medium effect. Seventy‑three point nine one percent of the sample endorsed SI at baseline; among those, 55.88% showed a reduction in SI and 44.12% had no SI (score of 0 on the respective items) at treatment endpoint. Adverse effects were common but generally mild and transient. After the first infusion 78.3% of patients experienced depersonalization symptoms and between 4% and 30% reported other common side effects (changes in heart rate or blood pressure, anxiety, light‑headedness, drowsiness, dizziness, blurred vision, nausea). By treatment endpoint the frequency of depersonalization fell to 26.1%, and 12% of patients had elevations in heart rate, nausea or increased blood pressure; other symptoms such as anxiety, light‑headedness, drowsiness, dizziness and blurred vision remained reported around 30%. Reported side effects typically resolved within 15–20 minutes after infusion cessation. CADSS‑6 scores for depersonalization showed a statistically significant reduction from baseline to endpoint with a large effect size in per‑protocol analyses and a medium effect in ITT analyses. Eleven point two seven percent of patients dropped out before completing the eight‑infusion acute course; reasons cited were personal (scheduling, relocation, loss of interest) or perceived lack of efficacy, not symptom exacerbation or intolerable side effects.

Gutierrez and colleagues interpret their findings as supporting prior literature that IV low dose ketamine is a rapid, effective and generally well‑tolerated treatment option for depressive symptoms and SI in outpatients with MDD‑TRD in a public clinic setting. Over half of the treated patients met the study's response criterion by the end of the acute eight‑infusion course, consistent with previously reported response ranges of about 35%–60% in other studies. The authors note that differences between this and other reports may reflect heterogeneity in dose regimens (this cohort received eight fixed 0.5 mg/kg infusions), outcome instruments, degrees of prior treatment resistance and the naturalistic versus controlled trial contexts. In terms of tolerability, the observed side‑effect profile—predominantly transient dissociation/depersonalization and short‑lived somatic symptoms—aligns with systematic reviews and controlled trials, and most side effects resolved within minutes after infusion. The investigators highlight the absence of symptomatic worsening in their sample, contrasting this with modest rates of worsening reported with some antidepressant treatments. They also discuss broader concerns acknowledged in the literature: inconsistent TRD definitions that limit generalisability, limited long‑term data on efficacy and safety (including potential for misuse, tolerance, withdrawal and organ toxicity), restricted accessibility and public funding, and practical challenges of maintenance parenteral administration that motivate study of alternative routes (intranasal, oral, sublingual) that themselves carry risks for non‑monitored use. The paper explicitly acknowledges several limitations. As a retrospective open‑label chart review it is subject to selection bias, potential under‑reporting of adverse events and variable clinical practice across providers. The authors did not record the exact number or classes of prior failed antidepressant trials, which constrains characterisation of TRD severity. Side effects were reported categorically relative to patient baseline rather than as numeric vital sign changes, limiting granularity. The cohort comprised outpatients only and the analysis focused on the short‑term acute course (eight infusions over four weeks), so long‑term effectiveness and safety remain unassessed. Gutierrez and colleagues further note that practice variability and assessment differences across clinicians could introduce heterogeneity, although they suggest their sample size and use of multiple assessment scales mitigate some of these concerns. The authors call for further real‑world observational work and controlled trials to clarify short‑ and long‑term applications, tolerability and safety, and to better characterise risks such as misuse and organ toxicity.

The authors conclude that, within this single‑site retrospective real‑world cohort, IV low dose ketamine administered as eight 0.5 mg/kg infusions over four weeks was associated with rapid, clinically meaningful reductions in depressive symptoms and suicide ideation and was generally well tolerated. They caution that additional research—both observational and controlled—is needed to define long‑term effectiveness, maintenance strategies, safety, and misuse potential, and to address accessibility and scalability of ketamine treatment for MDD‑TRD.