A rapid positive influence of S-ketamine on the anxiety of patients in palliative care: a retrospective pilot study

Patients in palliative care commonly experience intertwined physical and psychological symptoms, and rapid-acting treatments for anxiety and depression are often needed because conventional antidepressants take weeks to act and benzodiazepines carry problematic side effects. Ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, has been shown to produce rapid antidepressant and anxiolytic effects; its S-enantiomer (S-ketamine, or esketamine) is used intravenously for refractory pain at 0.25 mg/kg over 45 minutes and has greater NMDA-blocking potency with reportedly fewer psychotomimetic effects than the racemate. Earlier research indicates ketamine’s potential to reduce psychological distress in palliative populations, but data specifically on the purified S-enantiomer in this setting are lacking. Falk and colleagues therefore conducted a retrospective pilot study of routine clinical data from a specialised palliative care unit to test whether an analgesic S-ketamine infusion is associated with short-term reductions in anxiety and depression compared with matched controls. The main hypothesis was that a single S-ketamine infusion would reduce psychological distress (measured as anxiety and depression on the STADI) from before to after administration, and that this effect would persist after adjustment for potential confounders such as pain, need for care, psychological support, days on antidepressants and concurrent medications.

This study was a retrospective analysis of routine data collected over one year (April 2016 to March 2017) from an inpatient specialised palliative care unit at a university hospital in Germany. Inclusion criteria for the S-ketamine group were age ≥ 18 years, receipt of an S-ketamine infusion for analgesia and available STADI data both before and after the infusion. Controls were patients aged ≥ 18 with STADI data at two measurement points; those whose first STADI was on the day of admission were excluded to avoid admission-related confounding. All patients received the same multidisciplinary palliative care apart from the S-ketamine treatment itself. The primary outcome was psychological distress measured with the State-Trait Anxiety Depression Inventory (STADI) state section, yielding T-scores for anxiety and depression (mean = 50, SD = 10; T > 60 classified as pathological). STADI state scores were taken at repeated routine assessments; for S-ketamine patients T1 was the last STADI before infusion and T2 the first after. Secondary outcomes included pain (numerical rating scale, NRS, within the Palliative Symptom Burden Score, PSBS) and an ordinal PSBS item for restlessness/anxiety as a measure of persistent psychotomimetic effects. Potential confounders were specified a priori: pain (NRS), need for care (AEDL score), minutes of psycho-oncological therapy received, days on antidepressants while on the ward, days in specialised palliative care, and medication intake of antidepressants, benzodiazepines and opioids. Statistical methods included calculation of reliability coefficients (Cronbach’s α) and test-retest correlations for key measures. A propensity score based on age and gender was used to match controls 1:1 to the S-ketamine group by nearest neighbour matching without replacement, producing a final matched control group of eight patients. The main inferential approach was multivariate analysis of variance (three-way mixed MANOVA) with STADI anxiety and depression T-scores as dependent variables and predictors group (S-ketamine vs control), time (T1 vs T2) and anxdep (anxiety vs depression). Follow-up two-way mixed ANOVAs examined anxiety and depression separately. Interval confounders were entered one at a time as covariates in MANCOVA/ANCOVA models, and nominal medication variables were analysed as separate predictors; phi coefficients assessed stability of medication intake between time points. For secondary outcomes, pain change from the morning before (Z1) to the morning after (Z2) was tested with two-way mixed ANOVA and restlessness/anxiety with Wilcoxon signed-rank tests. The study treated p < 0.05 as significant and did not correct for multiple comparisons; effect sizes r were reported with conventional thresholds (0.1 small, 0.3 medium, 0.5 large).

The matched analytic sample comprised n = 8 patients treated with S-ketamine and n = 8 propensity-matched controls (originally 15 eligible controls before matching). At baseline the S-ketamine group had higher psychological distress (STADI T-values > 60) and greater symptom burden: moderate pain (NRS ≥ 3) versus mild pain in controls, higher AEDL scores indicating greater need for care, less use of psycho-oncological treatment, and universal opioid use in the S-ketamine group. Reliability of the STADI and AEDL scales was high (Cronbach’s α > 0.87; STADI α > 0.91). Test-retest correlations were strong for pain (r = 0.95; p < 0.001; n = 15) and restlessness/anxiety (ρ = 0.92; p < 0.001; n = 15). The three-way mixed MANOVA showed a significant multivariate interaction of group by time on anxiety and depression combined, F(1, 14) = 4.78; p = 0.046; r = 0.50. Univariate follow-up analyses indicated a pronounced effect on anxiety: a significant group-by-time interaction with large effect size, and pairwise comparisons demonstrated a significant reduction in STADI anxiety from T1 to T2 in the S-ketamine group (reported F(1, 14) = 19.89; p = 0.001; r = 0.77) but not in controls. For depression the group-by-time interaction was not statistically significant (e.g. reported F(1, 14) = 1.60; p = 0.23; r = 0.32), though effect sizes were in the small-to-medium range. Adjustment for interval-scaled confounders (pain, AEDL, psycho-oncological treatment minutes, days on antidepressants, days of palliative care) in separate MANCOVAs/ANCOVAs did not eliminate the significant group-by-time interaction for anxiety; large effect sizes persisted. Medication analyses treated benzodiazepine and antidepressant intake as between-subject variables; intake was stable between time points (phi for antidepressants φ = 0.76, p = 0.002; benzodiazepines φ = 0.59, p = 0.018). Including benzodiazepines or antidepressants in multivariate models did not account for the observed anxiety improvement, although patients on antidepressants had higher baseline anxiety scores (mean 65.89 vs 52.07). An alternative propensity matching that also balanced baseline STADI global scores resulted in similar findings: the anxiolytic effect of S-ketamine remained significant. Regarding secondary outcomes, there was no evidence of a persistent psychotomimetic effect as measured by the PSBS restlessness/anxiety item from Z1 to Z2 (Wilcoxon T = 0; z = -1.00; p = 0.32). Likewise, there was no significant prolonged analgesic effect from Z1 to Z2: the interaction of group and time on pain was non-significant, F(1, 14) = 0.11; p = 0.75; r = 0.09. Descriptive plots indicated that the reduction in global STADI scores was more pronounced when T2 occurred one day after S-ketamine rather than up to four days later. The authors note that 5 of 8 S-ketamine patients showed clinically relevant (>10 T-score points) reductions in global STADI.

Falk and colleagues interpret these retrospective data as preliminary evidence that a single intravenous S-ketamine infusion given for analgesia is associated with a rapid reduction in psychological distress in palliative care patients, with the principal effect on anxiety rather than depression. They position their findings alongside prior reports of racemic ketamine producing rapid anxiolytic and antidepressant effects, and note that, to their knowledge, this is the first analysis focusing on the purified S-enantiomer in a palliative inpatient population. The anxiolytic effect showed consistently large effect sizes across analyses and remained after accounting for measured confounders and concurrent psychotropic medication. The investigators discuss possible explanations for the prominent anxiolytic effect in this population, citing similar patterns in hospice case reports and feasibility studies in which anxiety improved rapidly while depression effects took longer to emerge. They report no evidence of persistent psychotomimetic side effects by the morning after infusion and no sustained analgesic benefit beyond the immediate period. The discussion addresses safety by referencing other studies reporting low rates of serious adverse events with ketamine and transient acute side effects that typically abate within 24 hours. Key limitations are emphasised: the retrospective, non-randomised design, small sample size and baseline differences between groups (S-ketamine patients had higher symptom burden and baseline STADI scores). The timing of assessments varied (T2 sometimes occurring up to four days after infusion), which may have attenuated measured effects. To mitigate some biases the authors used propensity matching (age/gender and an alternative approach including baseline STADI), tested instrument reliability and ensured statistical assumptions were met, but they acknowledge that prospective, randomised, placebo-controlled trials will be required to confirm efficacy, to clarify optimal dosing and routes of administration, and to compare S-enantiomer versus racemate. The study team suggests initial steps could include feasibility studies with qualitative components when sample sizes are likely small.

The study’s retrospective findings indicate that a single S-ketamine infusion administered for refractory pain was associated with a rapid and primarily anxiolytic effect in patients receiving specialised palliative care. Results align with existing ketamine literature but are constrained by non-randomisation, retrospective data collection and a small sample size. The authors conclude that prospective, ideally randomised studies are needed to establish efficacy, safety and clinical utility of S-ketamine for anxiety and depression in palliative care.