Rapid and longer-term antidepressant effects of repeated-dose intravenous ketamine for patients with unipolar and bipolar depression

Mood disorders carry high personal and societal burden, and available antidepressant treatments typically take more than two weeks to produce clinically meaningful benefit, which prolongs the period of elevated suicidal risk. Electroconvulsive therapy produces faster antidepressant and antisuicidal effects but is limited by invasiveness and cognitive adverse effects. Earlier research implicates glutamatergic mechanisms and has shown that a single subanaesthetic IV infusion of ketamine (0.5 mg/kg) produces rapid antidepressant and antisuicidal effects in unipolar and bipolar depression, but these benefits frequently wane within days to around two weeks. This study aimed to examine whether a repeated‑dose ketamine protocol (six IV infusions of 0.5 mg/kg over 40 minutes, given on a Monday–Wednesday–Friday schedule across 12 days) would produce rapid and sustained antidepressant and antisuicidal effects in Chinese patients with unipolar or bipolar depression. Zheng and colleagues hypothesised that the regimen would be tolerable and that early therapeutic effects would be maintained across subsequent infusions.

This was an open‑label clinical study conducted at the Affiliated Brain Hospital of Guangzhou Medical University between November 2016 and December 2017. Participants received six IV ketamine infusions (0.5 mg/kg diluted in saline, administered over 40 minutes) on a Monday–Wednesday–Friday schedule across 12 days. Infusions were given with continuous telemetry monitoring of vital signs; a study psychiatrist was present and anaesthesiology practitioners were available as needed. Patients fasted overnight before infusion and remained under observation for at least 30 minutes after each infusion. Participants were aged 18–65 years and had a DSM‑5 diagnosis of non‑psychotic major depressive disorder (MDD) or bipolar disorder (BD) confirmed by treating psychiatrists and a checklist. Key inclusion criteria were a HAMD‑17 score ≥17 at screening, treatment resistance (insufficient response to ≥2 adequate antidepressant trials during the current episode) and suicidal ideation (SSI‑part 1 score ≥2). The study permitted comorbid anxiety disorders if not the primary cause of illness in the previous year. Major exclusions included current or past substance abuse/dependence, schizophrenia or other serious mental disorders, positive urine toxicology (e.g. methamphetamine), pregnancy or breastfeeding, unstable medical illness, significant neurological disease, or inability to cooperate. Participants continued stable doses of existing psychotropic medications but did not receive nonpharmacological interventions such as ECT. The primary outcome was change in depressive symptoms measured by the 10‑item Montgomery‑Åsberg Depression Rating Scale (MADRS). Response was defined as ≥50% reduction from baseline MADRS and remission as MADRS ≤10. Secondary outcomes included suicidal ideation (SSI‑part 1), anxiety (HAMA), psychotomimetic symptoms (four‑item positive symptom subscale of the BPRS) and dissociative symptoms (CADSS). Assessments were performed at baseline (t0), 4 hours after each infusion (t+4h), 24 hours after each infusion (t+24h), and at a naturalistic follow‑up approximately 14 days after the sixth infusion. Five trained raters completed the scales and demonstrated inter‑rater reliability >0.9. Statistical analyses used an intention‑to‑treat or modified ITT approach with SPSS 24. Normality was tested with Kolmogorov‑Smirnov. Between‑group comparisons of baseline characteristics used chi‑square, Fisher’s exact, t‑tests or Mann–Whitney U tests as appropriate. Logistic regression examined predictors of response at t+24h after the sixth infusion, with variables from univariate analyses entered as predictors and Bonferroni correction applied. Time to response and remission were analysed by Kaplan–Meier methods. Changes in continuous measures over time were tested with paired t‑tests or Wilcoxon signed‑rank tests; group comparisons across time used linear mixed models. Significance was set at P < 0.05 (two‑tailed).

Of 185 screened patients, 97 received at least one ketamine infusion and 84 (86.6%) completed all six infusions. The sample included both unipolar (n = 77) and bipolar (n = 20) depressed patients; the extracted text does not provide the full breakdown of other baseline characteristics here. Analyses were performed on the ITT or modified ITT samples. Primary outcomes: The mean time to response was 10.5 days (95% CI: 8.5–12.4) and to remission 14.2 days (95% CI: 12.2–16.2). At t+24h after the first infusion the response rate was 14.4% (95% CI: 7.3–21.6) and remission 9.3% (95% CI: 3.4–15.2). After the sixth infusion (t+24h), response increased to 68.0% (95% CI: 58.6–77.5) and remission to 50.5% (95% CI: 40.4–60.6). At the naturalistic follow‑up approximately two weeks after the sixth infusion, the relapse rate among responders was 25.8% (17/66). Predictors: In logistic regression, response at the post‑ketamine assessment was positively associated with an early response at t+24h after the first infusion (OR = 8.94, p = 0.046), personal income ≥4000 yuan/month (OR = 3.04, p = 0.035), and no history of psychiatric hospitalisation (OR = 3.41, p = 0.014). After Bonferroni correction only the association with no history of psychiatric hospitalisation remained statistically significant. Symptom change: Large and statistically significant reductions were observed at t+4h after the first infusion across the sample: MADRS decreased by 6.8 points (95% CI: 5.4–8.2, P < 0.001), SSI‑part 1 by 1.8 points (95% CI: 1.3–2.3, P < 0.001) and HAMA by 4.6 points (95% CI: 3.3–5.9, P < 0.001). These improvements were generally maintained across subsequent infusions. The nonresponder subgroup nevertheless showed a rapid and sustained reduction in suicidal ideation throughout treatment; reductions in MADRS and HAMA among nonresponders were less consistent and in some cases were only apparent after multiple infusions. Effect sizes for differences between responders and nonresponders reached a maximum of 2.3 for MADRS and 1.3 for HAMA; SSI effect sizes were smaller (maximum ES = 0.64) and significant at fewer time points. Adverse events and tolerability: During the first infusion there was no significant increase in psychotomimetic symptoms measured by the BPRS from t0 to t+4h (4.4 ± 1.1 vs 4.2 ± 0.6, P = 0.08); a small but significant decrease in BPRS was observed from t0 to t+24h (4.4 ± 1.1 vs 4.1 ± 0.4, P = 0.02). Dissociative symptoms as measured by CADSS showed a mild, marginally significant increase at t+4h after the first infusion (0.3 ± 1.4 vs 0.7 ± 2.1, P = 0.05) which returned to baseline by t+24h. Differences between responders and nonresponders in psychotomimetic symptoms were only found at t+4h after the third infusion. Overall, six infusions were described as safe and well tolerated in this cohort.

Zheng and colleagues report that six repeated IV ketamine infusions produced rapid, robust antidepressant and antisuicidal effects in Chinese patients with unipolar or bipolar depression, and that these effects were generally maintained over the course of the infusion series. The study team emphasises that improvement in suicidal ideation occurred rapidly after the first infusion and was evident across the sample, including participants who did not meet response criteria on depressive symptoms. Anxiety symptoms also improved. Early clinical response to the first infusion predicted sustained response across the series, while a history of psychiatric hospitalisation predicted poorer outcome; higher personal income was also associated with better response in this sample. The authors position their findings against prior single‑infusion studies and smaller repeated‑infusion reports, noting that serial infusions substantially increased response and remission rates compared with the effect observed after a single infusion and appeared to confer greater short‑term durability. They also reiterate that dissociative and psychotomimetic effects were generally mild and transient. However, the investigators acknowledge several important limitations: the open‑label design limits causal inference and leaves results susceptible to expectancy and non‑specific effects; the follow‑up period was short (about two weeks), restricting conclusions about longer‑term durability and time to relapse; and participants continued real‑world psychotropic medications, which may confound attribution of effects to ketamine alone. The authors further note safety concerns around the potential for abuse and possible neurotoxicity with prolonged ketamine exposure and call for systematic research to determine optimal dosing and treatment frequency. In closing, the study team recommends randomised controlled trials with larger samples and longer follow‑up to confirm and extend these findings and to clarify the risk–benefit profile of repeated ketamine infusions for unipolar and bipolar depression.