Dose-Related Effects of Adjunctive Ketamine in Taiwanese Patients with Treatment-Resistant Depression

Ketamine has been shown in prior research to produce rapid antidepressant effects in patients with treatment-resistant depression, in contrast to the slow onset and modest remission rates associated with standard antidepressants. Despite replicated signals of efficacy, important practical questions remain unresolved, notably the dose–response relationship for R/S-ketamine when used as a 40 minute intravenous infusion and the generalisability of findings across racial and ethnic groups. There is particular interest in Han Chinese and Taiwanese populations because of a high prevalence of the Val66Met polymorphism of the brain-derived neurotrophic factor (BDNF) gene, in which the Met allele has been suggested by some preclinical and small clinical studies to attenuate ketamine’s antidepressant response. Su and colleagues therefore set out to characterise the dose-related antidepressant effects of adjunctive R/S-ketamine in Taiwanese patients with treatment-resistant major depressive disorder who were genotyped for the BDNF Val66Met polymorphism. The trial tested single infusions of saline (placebo), 0.2 mg/kg, and 0.5 mg/kg R/S-ketamine, with serial mood ratings and plasma pharmacokinetics, to determine whether antidepressant effects vary by dose, baseline depression severity, and BDNF genotype.

This was a randomised, double-blind, parallel-group, placebo-controlled trial conducted at Taipei Veterans General Hospital between 2012 and 2015. Of 74 patients enrolled, three were excluded for medical or remission-related reasons, leaving a sample of 71 patients with recurrent major depressive disorder who had failed at least two adequate antidepressant trials. Eligibility was assessed using the MINI and a semistructured review of treatment history; exclusion criteria included bipolar disorder, psychosis, substance dependence (other than nicotine), and mild depressive symptoms (HAMD thresholds were applied at screening and before entry). Ongoing antidepressant regimens were continued but no medication changes were permitted for two weeks before randomisation or during the study. Participants were randomised to a single 40 minute intravenous infusion of saline, 0.2 mg/kg R/S-ketamine, or 0.5 mg/kg R/S-ketamine. Vital signs (blood pressure, heart rate, pulse oximetry) were monitored repeatedly during and after infusion. Clinical assessments included the 17-item Hamilton Depression Rating Scale (HAMD) as the primary outcome and the Brief Psychiatric Rating Scale (BPRS) positive symptom subscale to monitor psychotomimetic effects; HAMD was administered in person at baseline and at 40, 80, 120, and 240 minutes after infusion on the test day, with subsequent telephone assessments up to day 28. Telephone HAMD administration was used based on prior evidence of good correlation with face-to-face interviews. Patients were discharged 240 minutes after infusion once acute effects had abated. Laboratory procedures comprised BDNF Val66Met genotyping and measurement of plasma ketamine and norketamine at baseline and 40, 80, 120, and 240 minutes after infusion using LC–MS/MS. The limit of quantitation was 10 ng/ml and assay precision and accuracy were reported. The primary statistical approach was mixed effects regression modelling with group (dose) as a between-subject factor, time as a within-subject factor, baseline HAMD as a predictor, and an unstructured variance–covariance matrix to account for repeated measures. Symptom-cluster analyses used previously defined HAMD factor groupings (emotional, atypical, insomnia-related). Responder status was prespecified as ≥50% reduction in HAMD on at least two daily measures between 24 and 96 hours (days 2–5) post-infusion; Fisher’s exact tests compared responder rates. Nonparametric mixed models were used for BPRS outcomes and log-transformed mixed models for plasma levels. Analyses were performed in SAS and significance was set at the 0.05 family-wise level.

Seventy-one patients entered the randomized treatment groups. Demographic and clinical characteristics were similar across groups; the sample was predominately female, aged about 40–50 years, and most patients were receiving combination pharmacotherapy (antidepressant plus a second-generation antipsychotic in 51%). Psychiatric comorbidity was common (for example, dysthymia and generalized anxiety disorder each in roughly half of patients), and 56.3% had two or more comorbid diagnoses. A subset of 48 subjects underwent concurrent neuroimaging (results reported elsewhere). The authors note that a majority of participants carried at least one copy of the BDNF Met allele. In the mixed-effects model for HAMD scores there was a significant interaction of group, time, and baseline HAMD (F(22, 64.9) = 1.90, p = 0.02), and a significant group by time interaction (F(22, 64.9) = 1.74, p = 0.04), indicating that dose effects depended on time and baseline severity. Overall, the 0.5 mg/kg dose reduced HAMD scores significantly more than placebo (Tukey-adjusted p = 0.008), whereas the 0.2 mg/kg dose did not differ significantly from placebo (adjusted p = 0.20); the two ketamine doses did not differ significantly from each other in the overall comparison (adj. p = 0.37). Post hoc testing showed that the 0.5 mg/kg dose separated from placebo at mean baseline severity (adjusted p = 0.01) and at high baseline severity (adjusted p = 0.02); at high baseline severity the 0.5 mg/kg dose was also more effective than 0.2 mg/kg (adjusted p = 0.05). No significant ketamine effects were observed at low baseline HAMD severity. Responder rates (≥50% HAMD reduction on at least two daily measures between days 2–5) were 45.8% in the 0.5 mg/kg group, 39.1% in the 0.2 mg/kg group, and 12.5% in the placebo group. Fisher’s exact test showed a significant difference across groups (p = 0.03) and a significant linear dose trend (p = 0.01). Post hoc comparisons found higher responder rates for 0.5 mg/kg versus placebo (p = 0.01) and for 0.2 mg/kg versus placebo (p = 0.05); the two ketamine doses did not differ from one another (p = 0.77). Analysis of HAMD symptom clusters revealed significant dose-by-baseline-severity interactions for the emotional cluster (F(2,65) = 3.21, p = 0.05) and a group–time–baseline interaction for the atypical cluster (F(22,65) = 2.10, p = 0.01). Improvements were most evident on emotional and atypical symptoms and were more pronounced at higher baseline severity levels; there were no significant between-group differences for the insomnia-related cluster. Plasma ketamine and norketamine concentrations were dose-related and varied over time, with significantly higher levels in the 0.5 mg/kg group than the 0.2 mg/kg group (ketamine F(1, 40.7) = 28.5, p < 0.0001; norketamine F(1, 44.4) = 22.8, p < 0.0001). The authors report no significant differences in plasma levels by responder status. Notably, peak ketamine concentrations observed in this sample were lower than those reported in some prior studies of predominantly Caucasian patients. Ketamine produced dose-related haemodynamic changes: systolic and diastolic blood pressure rose significantly with dose (SBP F(8, 68) = 3.25, p = 0.004; DBP F(8, 68) = 2.65, p = 0.01), with the largest increases at 40 minutes and a return to baseline by two hours. Heart rate increased in a dose-related manner (F(8, 59) = 3.12, p = 0.005), peaking around two hours. On safety and tolerability, there were no unexpected serious adverse events; BPRS positive symptom scores did not show significant dose-related psychotomimetic effects (p-values ~0.08–0.34 depending on contrast), nausea occurred in 6 patients (8.5%), and one infusion was stopped because of transient behavioural effects that resolved without intervention.

Su and colleagues interpret their findings as the first demonstration, in their view, of dose-related antidepressant efficacy for R/S-ketamine in a genotyped Chinese (Taiwanese) sample in which most participants carried at least one Met allele of the BDNF Val66Met polymorphism. The principal conclusion is that 0.5 mg/kg produced greater reduction in HAMD scores than placebo, and that dose-related benefits were moderated by baseline depression severity: effects were minimal in patients with relatively mild depression and increased as baseline severity increased. The 0.2 mg/kg dose did not reduce HAMD scores significantly versus placebo in the main analyses, leading the authors to suggest that the lower dose may lack sufficient efficacy despite potential tolerability advantages. The discussion highlights that ketamine’s rapid antidepressant effects in this trial were most apparent for emotional and atypical symptom clusters, whereas insomnia-related symptoms did not improve; the authors note that heterogeneous effects on sleep-related items have been observed in prior work and may have clinical relevance, for example in relation to persistent suicidality. The authors consider whether the antidepressant potency of R/S-ketamine might be lower than that of S-ketamine, citing prior reports of greater potency for the S-isomer at NMDA receptors; this could account for differences between the present 0.2 mg/kg R/S-ketamine result and prior positive findings with 0.2 mg/kg S-ketamine. Regarding BDNF genotype, the investigators did not find clear evidence that high rates of the Met allele abolished ketamine efficacy, and they interpret this as evidence that any reduction in ketamine responsiveness associated with the Met allele was insufficient to obscure the drug’s antidepressant effects in this sample. The authors also note that mean peak plasma ketamine concentrations in this study were lower than those reported in some Caucasian samples (reported mean ≈ 115 ng/ml here vs 150–200 ng/ml elsewhere), with unclear reasons; norketamine levels did not suggest simple increases in metabolism, and plasma ketamine levels did not correlate with clinical response within dose groups. The absence of measurements of the 2R,6R-hydroxynorketamine metabolite is acknowledged as a limitation given emerging interest in its potential role. Other limitations discussed include the inclusion of patients with relatively mild depression, which may complicate interpretation of clinical significance and responses that meet remission by definition, and the small number of subjects with the Val/Val genotype limiting genotype-specific inference. The authors conclude that their results support continued use of 0.5 mg/kg R/S-ketamine in treatment-resistant depression and justify further work to test whether the 0.2 mg/kg dose is effective in patients without the Met allele and whether higher R/S-ketamine doses (to achieve target plasma levels) would increase response rates.