Disentangling the association of depression on the anti-fatigue effects of ketamine

Fatigue and depression are tightly linked clinically: both impair daily functioning and are frequently correlated, yet evidence also suggests they can be distinct constructs because fatigue sometimes persists after depressive remission. Amotivation—reduced initiative, interest and goal-directed behaviour—is common to both fatigue and depression and may share biological mechanisms with fatigue in a subset of patients. Distinguishing fatigue from depression in treatment studies is challenging due to measurement limits, and better understanding of which depressive symptom domains relate to fatigue could help separate shared from distinct pathways. Saligan and colleagues used data from prior ketamine trials to examine whether ketamine’s previously observed rapid anti-fatigue effects are independent of its antidepressant action. Specifically, the study tested whether changes in overall depression or in particular depressive symptom domains mediate ketamine’s effect on fatigue after a single intravenous dose. The work aims to clarify whether ketamine has a unique anti-fatigue effect or whether observed fatigue improvement is explained by reductions in depressive symptoms such as amotivation and depressed mood.

This is a secondary analysis of a double-blind, randomized, placebo-controlled, crossover clinical trial (NCT00088699) conducted at the NIH Clinical Center. Participants were adults aged 18–65 with treatment-resistant Major Depressive Disorder (DSM-IV), recurrent episodes without psychotic features, age of onset ≤40 years, current episode ≥4 weeks, MADRS ≥20 at screening and prior to each infusion, and failure to respond to at least one adequate antidepressant trial. Subjects were unmedicated or tapered off psychotropic medication prior to randomization; standard exclusions included recent substance dependence/abuse and unstable medical conditions. The analytic sample comprised 39 participants (the 35 from a prior publication plus four additional participants collected subsequently). Each participant received two experimental infusions separated by two weeks: ketamine hydrochloride 0.5 mg/kg intravenously over 40 minutes, and placebo, in a randomized order. Outcomes were assessed at multiple timepoints: baseline (−60 minutes) and 40, 80, 120, 230 minutes, and Day 1, Day 2, Day 3 post-infusion; Day 1 (24 hours) was selected as the primary timepoint of interest as it corresponds to peak ketamine response in these data. Fatigue was measured with the clinician-administered seven-item NIH-Brief Fatigue Inventory (NIH-BFI), which yields a 0–34 total score and was developed for use in depressed samples. Depression severity was measured with the 10-item Montgomery–Åsberg Depression Rating Scale (MADRS; 0–60 total). An exploratory factor analysis (EFA) of multiple depression scales identified depression subscales used here: Depressed Mood, Tension, Negative Cognition, Impaired Sleep, Suicidal Thoughts, Reduced Appetite and Amotivation; anhedonia was excluded due to incomplete data. Analyses used general linear mixed models with a repeated categorical time effect for each drug nested within subject, an unstructured covariance matrix and a random intercept per subject; degrees of freedom were estimated with the Kenward–Roger approximation. The investigators tested mediation by first confirming ketamine’s effect on MADRS, then on NIH-BFI, and finally entering MADRS total (or EFA symptom subscales) as putative mediators to see whether the ketamine effect on fatigue remained statistically significant. Baseline values were modelled, and the ketamine effect was estimated via a baseline–Day 1 contrast; infusion order was entered as a covariate. Cohen’s d effect sizes were calculated from the model-estimated contrasts. Both the treatment and mediator variables were mean-centred prior to analysis. Given the exploratory secondary nature of the work, significance was assessed at α = .05 without adjustment for multiple comparisons.

The analytic sample consisted of 39 participants with MDD; mean age was 36.26 (±10.06) years and 59% were female. Consistent with prior reports, ketamine produced a statistically significant improvement in fatigue at Day 1 compared with placebo as measured by the NIH-BFI, with a large effect size (Cohen’s d = 0.95, 95% CI: 0.45–1.45). However, when MADRS total score was entered into the model as a mediator, the unique effect of ketamine on NIH-BFI was no longer evident (d = 0.06, 95% CI: −0.42–0.54), indicating complete mediation by overall depressive symptom change. The reverse mediation was partial: NIH-BFI partially mediated ketamine’s effect on MADRS total score but accounted for only a small proportion of the antidepressant effect. The authors report effect-size comparisons indicating a larger total/mediated effect (d = 1.16, 95% CI: 0.69–1.64) versus a smaller remaining effect (d = 0.79, 95% CI: 0.29–1.29), as presented in the Results. When EFA-derived depressive symptom domains were entered as putative mediators, most domains reduced the ketamine–fatigue effect only slightly. Two exceptions—Depressed Mood and Amotivation—accounted for most of the ketamine effect on NIH-BFI, meaning improvements in these specific depressive symptom domains largely explained the observed anti-fatigue response. The extracted text does not provide exact numerical mediation estimates for each subscale beyond this qualitative statement. Anhedonia was not analysed because roughly half the sample lacked those ratings.

Saligan and colleagues interpret the findings to mean that ketamine produces rapid improvements in fatigue among treatment-resistant MDD patients, but these improvements appear to be driven by ketamine’s broader antidepressant effects rather than by a unique anti-fatigue action. In particular, reductions in depressed mood and amotivation, both depressive symptom domains identified by EFA, fully explained the apparent anti-fatigue effects at 24 hours. The authors note that the anti-fatigue effect was independent of ketamine’s effect on sleep disturbance in these analyses, which they find notable given established links between sleep impairment and fatigue in other populations. They caution that the relationships observed here may depend on the sample selection—these were patients with MDD—and might differ in a cohort of patients with clinically significant fatigue but no depression. The single-dose ketamine regimen has a short-lived pharmacodynamic profile, so results could differ with repeated dosing or with treatments that produce more sustained antidepressant responses. Measurement limitations receive emphasis: NIH-BFI was developed and validated within depressed samples and used MADRS-derived items in its development, so it may insufficiently separate fatigue from depressive symptoms in this setting. The investigators therefore recommend future studies use fatigue instruments demonstrably distinct from depression or supplement self-report with behavioural tasks that probe cognitive and motivational aspects of fatigue, and consider measures that capture diurnal variability. Finally, they acknowledge the small sample size and the consequent limits on generalisability to all MDD patients.

In this secondary analysis of a randomized crossover trial in treatment-resistant MDD, ketamine did not demonstrate a unique anti-fatigue effect independent of its antidepressant action when fatigue was measured with the NIH-BFI at 24 hours. Improvements in fatigue were largely explained by reductions in depressed mood and amotivation. The authors suggest that ketamine may be useful for affective forms of fatigue, but highlight the need for studies enrolling non-depressed fatigued patients and for use of fatigue measures that more cleanly separate fatigue from depression and capture relevant temporal dynamics.