Preliminary evidence that serum interleukin-6 is a candidate biomarker of response to esketamine in treatment-resistant depression

Previous research had suggested that elevated serum interleukin-6 (IL-6) might predict response to ketamine in treatment-resistant depression (TRD), but the evidence was still limited and the biology was not fully understood. The paper frames ketamine and esketamine as rapid-acting antidepressants that may also influence inflammatory pathways, including cytokines such as IL-6, yet it notes that prior findings on cytokine changes after ketamine have been complex and time-dependent rather than straightforward. The authors also highlight that esketamine, while pharmacologically related to ketamine, has a different delivery route and safety profile, and that its possible immunological effects have not been as well characterised. Colizzi and colleagues set out to examine whether baseline IL-6 could predict clinical response to intranasal esketamine in adults with TRD. More specifically, they aimed to test whether IL-6 levels were associated with change in depressive symptoms and disability over treatment, and whether IL-6 itself changed during the course of therapy. The study was positioned as an early mechanistic Phase II trial intended to explore IL-6 as a candidate biomarker for biomarker-guided, personalised treatment in TRD.

The study was a 24-week, open-label, non-comparator, investigator-initiated Phase II trial conducted at Udine University Hospital between March 2023 and October 2024. Fourteen adults with TRD were recruited. Eligibility required TRD defined conservatively, with inadequate response to at least two antidepressant trials and confirmed adherence; the authors state that they used stricter duration thresholds than the minimum sometimes accepted in recent consensus guidance. Exclusion criteria included aneurysmal vascular disease or arteriovenous malformation, intracerebral haemorrhage, severe hepatic impairment, and other significant individual risk factors unless clinical benefit was judged to outweigh risk. Esketamine was given according to approved prescribing information. It was administered twice weekly for 4 weeks, followed by a tapering maintenance schedule. Dosing differed by age: participants younger than 65 years started at 56 mg and could increase to 84 mg twice weekly during induction; those aged 65 years or older started at 28 mg and could increase to 56 mg. During weeks 5-8, dosing was reduced to once weekly, and from weeks 9-24 it was given every 2-3 weeks depending on response, tolerability, and protocol phase. Treatment was individualized when needed. Clinical assessments were performed at baseline, week 8, and week 24. The primary outcome was depressive symptom severity measured with the Montgomery-Åsberg Depression Rating Scale (MADRS). Functional disability was assessed with the World Health Organization Disability Assessment Schedule (WHODAS). Serum IL-6 was measured at the same time points from fasting morning blood samples collected before esketamine administration, using an electrochemiluminescence immunoassay. The assay’s stated analytical sensitivity and quantitation limits were reported in the text. The main analysis used linear mixed-effects models to examine behavioural change over time as a function of IL-6, with the IL-6 × time interaction as the main inferential term. Missing data were handled by imputation, and sensitivity analyses added both time-invariant covariates (demographic and baseline clinical variables) and time-varying covariates (physiological measures and concomitant treatment dose) to test robustness. Sample size adequacy was explored through a simulation-based sensitivity analysis. Analyses were performed in R.

Fourteen patients were enrolled, with a mean age of 51.9 years and 57.1% male. Psychiatric comorbidity included generalised anxiety disorder in two participants, and attention-deficit/hyperactivity disorder and histrionic personality disorder in one participant each. Medical comorbidities included type 2 diabetes, hypertension, hypercholesterolaemia, and Hashimoto’s thyroiditis. The text also notes neurological conditions in two participants. Four adverse events were recorded clinically: one episode of bradycardia in week 5, somnolence/sedation in the first two administrations in one patient, headache after the first administration in another, and anxiety during treatment in one participant, which was considered likely unrelated to esketamine. One participant was hospitalised psychiatrically after week 8, also judged unrelated to esketamine. Depressive symptoms improved significantly over time. Mean MADRS scores fell from 30.9 at baseline to 19.5 at week 8 and 18.7 at week 24, with a significant overall time effect (F2,26.0 = 21.4, p < 0.001). The authors report a large effect size and substantial between-person variability. The proportion of responders, defined as at least 50% reduction in MADRS, increased from 14.3% at week 8 to 35.7% at week 24, although one participant lost responder status between those time points. At baseline, IL-6 was not significantly correlated with MADRS score. However, higher IL-6 levels were associated with a faster reduction in depressive symptoms over time, as shown by a significant IL-6 × time interaction in the MADRS mixed model (F2,25.4 = 4.59, p = 0.020). Conversely, lower IL-6 tended to be associated with greater symptom severity over time (F1,28.8 = 4.24, p = 0.049). The authors describe the effect sizes as large and medium respectively, while noting wide confidence intervals because of the small sample. IL-6 itself did not decline significantly over time in the full sample, whether analysed with complete data or with interpolated values for two missing week-8 measurements. For WHODAS, disability also improved over time, and the pattern broadly mirrored the MADRS results. The IL-6 × time interaction was significant (F2,17.3 = 5.0, p = 0.019), and lower IL-6 was significantly associated with greater disability (F1,19.6 = 6.9, p = 0.016). Sensitivity analyses generally supported the main findings. Adding baseline demographic and clinical covariates did not remove the significance of the IL-6 × time interaction for MADRS, and results for WHODAS remained broadly similar. When phase-specific mean daily esketamine dose was added as a time-varying covariate, the IL-6 × time interaction and time effects remained significant; the main effect of IL-6 became only marginal for MADRS but remained significant for WHODAS. The authors report that dose itself was not significantly associated with the outcomes and that coefficient changes were small overall.

The authors interpret the findings as additional evidence that IL-6 may be associated with response to NMDA antagonist-based antidepressant treatment, including esketamine. They argue that the results fit with a broader view of depression as involving inflammatory mechanisms in at least some patients, and they present the IL-6 findings as consistent with the idea that ketamine and esketamine may have immunomodulatory as well as antidepressant effects. They outline several possible biological pathways, including NMDA receptor blockade, downstream effects on glutamatergic signalling and brain-derived neurotrophic factor, and interactions with microglia, astrocytes, and macrophages. Colizzi and colleagues place their findings in the context of earlier research suggesting that ketamine can alter cytokines and that IL-6 has been associated with treatment resistance. They also note that the relationship between IL-6 and depressive illness is probably not a simple marker of inflammation alone, but may reflect broader metabolic or “somatic maintenance” processes linked to stress and immune allocation. On this view, IL-6 changes during treatment might represent a shift in neuroimmune and metabolic priorities alongside symptomatic improvement. The authors emphasise several limitations. The sample was very small, the study was single-arm and open-label, and there was no randomisation or blinding, which limits causal inference and makes it difficult to establish whether IL-6 is specific to esketamine rather than reflecting nonspecific improvement. They also note the limited ability to adjust for inflammatory confounders, the single-centre design, and restricted generalisability because the sample was relatively old. They characterise the work as exploratory and hypothesis-generating. In terms of implications, the authors suggest that if these findings are replicated in larger, better controlled studies, IL-6 could become a candidate biomarker for treatment selection and patient stratification in TRD. They present this as part of a broader move towards more personalised, biologically informed care.