Predictors of Response to Ketamine in Treatment Resistant Major Depressive Disorder and Bipolar Disorder

Major Depressive Disorder (MDD) is a common, chronic and disabling condition that often responds slowly to conventional antidepressants, with clinically meaningful improvement commonly delayed by 4–8 weeks. A substantial proportion of patients—roughly one-third—fail to remit despite multimodal treatment, creating clinical urgency to find faster-acting interventions. Subanaesthetic intravenous ketamine (commonly 0.5 mg/kg over 40 minutes) has shown rapid antidepressant and anti-suicidal effects in treatment‑resistant depression (TRD) within hours to days, with reported response rates after a single infusion of about 54% at 6 hours, 71% at day 1 and 54% at day 3, and a mean time to relapse of roughly 17 days. Enthusiasm for ketamine is tempered by concerns about tolerability, longer-term safety, diversion risk and the costs and logistics of IV administration. Rong and colleagues set out to identify pretreatment clinical and biological predictors of response to IV ketamine among adults with TRD as part of MDD or bipolar disorder (BD). Focusing deliberately on IV infusion studies—the most commonly reported delivery route—the review aimed to summarise variables measured before treatment that are associated with subsequent antidepressant response, with a view to informing patient selection, mechanistic hypotheses and future predictive work.

The investigators performed a narrative literature review. Electronic searches of PubMed/MEDLINE, ClinicalTrials.gov and Scopus were conducted from database inception to January 2018, using the term ketamine cross-referenced with depression, major depressive disorder, bipolar disorder, predictors, and response and/or remission. Search results were screened by title and abstract; bibliographies of retrieved articles were hand-searched to identify additional reports. Predefined inclusion criteria were: adult human subjects aged 18–65 with MDD or BD and/or TRD (DSM-defined); an intervention trial of IV ketamine infusion; and assessment of depression severity before and after infusion with validated rating scales. Exclusion criteria comprised preclinical studies and duplicated datasets. The initial search returned 582 records; after exclusion of non-human/preclinical work and screening, 12 clinical studies in adults with MDD or BD/TRD met inclusion and were included in the review (nine from electronic search, three from manual search). The authors extracted variables reported as pretreatment predictors across clinical, biochemical, polysomnographic, neurochemical, neuroimaging, genetic and cognitive domains.

Twelve clinical studies met the inclusion criteria. The review found no sociodemographic factors that robustly predicted ketamine response. Clinical predictors: A positive family history of alcohol use disorder (FHP) was repeatedly associated with greater antidepressant response following IV ketamine, as measured by changes in MADRS and HDRS scores; some data suggested this could extend to sustained benefit up to four weeks in FHP subjects. Higher baseline body mass index (BMI) was correlated with greater early symptomatic improvement at 230 minutes and one day post-infusion in a pooled sample (n = 108 in a post-hoc analysis), although the BMI–response relationship was not always evident at seven days and could reflect higher absolute ketamine dosing in heavier individuals. A prior history of suicide attempts moderated outcome such that individuals without prior attempts had greater improvement at seven days but not at earlier time points. Peripheral biomarkers: Low pretreatment plasma adiponectin (an anti-inflammatory, insulin‑sensitising adipokine) predicted a more rapid antidepressant response in one study, suggesting a role for metabolic–inflammatory status. A small study (n = 20) reported that higher circulating vitamin B12 levels were associated with ketamine response (defined as ≥50% HDRS reduction at day 7), consistent with prior associations between B12 and antidepressant response. Polysomnography: In a study of 30 MDD subjects, a low delta sleep ratio (DSR)—the ratio of slow‑wave activity between the first two non‑REM sleep episodes—was associated with greater improvement in depressive symptoms after ketamine, whereas high pretreatment DSR predicted less favourable response. Neurochemistry and neuroimaging: Proton magnetic resonance spectroscopy (1H‑MRS) in 14 MDD subjects showed that pretreatment Glx/glutamate ratio (a surrogate for glutamine relative to glutamate) was inversely correlated with symptomatic improvement after 230 minutes; absolute GABA and glutamate concentrations did not predict change. An ACC (anterior cingulate cortex) activation study (MDD n = 11 versus controls n = 11) found exaggerated ACC activity in patients, and higher pretreatment ACC responsiveness to emotional stimuli correlated positively with rapid antidepressant response to ketamine. Genetics: The BDNF Val66Met polymorphism was examined; Met allele carriers experienced a smaller mean change in HDRS from baseline to endpoint than Val/Val carriers, implying greater likelihood of response in Val/Val individuals at rs6265. Cognitive function: Baseline neurocognitive testing in 25 TRD subjects indicated that lower baseline cognitive performance—particularly reduced processing speed—was associated with increased likelihood of antidepressant response at 24 hours (response defined as ≥50% MADRS reduction). The authors note hypotheses that ketamine’s pro-cognitive effects may contribute to its anti‑suicidal benefits. Overall, the most replicated predictors across studies were higher BMI and a positive family history of alcohol use disorder, but no single pretreatment variable demonstrated robust, consistent predictive power across the literature.

Rong and colleagues interpret the literature as showing preliminary but heterogeneous evidence for a set of pretreatment variables that may moderate response to IV ketamine in TRD. They propose that predictors cluster around mechanistic domains consistent with ketamine’s putative actions—glutamatergic signalling, inflammation/metabolic processes, and neuroplasticity—citing findings such as associations with adiponectin, Glx/glutamate ratios, ACC reactivity and BDNF Val66Met genotype. The authors emphasise that the current evidence remains insufficient to support precision selection of patients for ketamine on the basis of any single baseline measure. They note that treatment response variability likely reflects heterogeneity in pathophysiology, genetics and clinical features, and therefore a multimodal predictor set combining phenotypic and biological variables will probably be necessary. Machine learning and other data-driven, agnostic computational approaches are highlighted as promising strategies to refine predictive phenotypes. Important limitations acknowledged in the review include small sample sizes in many studies, heterogeneity of study designs, lack of randomisation in several reports, prevalence of open-label trials, and limited or absent healthy control comparisons. The authors also draw attention to gaps in long‑term clinical data, unresolved safety and tolerability concerns with repeated ketamine exposure, and the logistical and cost issues inherent to IV administration. They suggest future work should examine subdomain‑specific predictors (for example, cognitive or reward‑related outcomes), and extend to functional and patient‑reported outcomes such as workplace functioning.

The review identified several putative pretreatment predictors of antidepressant response to IV ketamine in TRD, with the most replicated findings being higher BMI and a positive family history of alcohol use disorder. However, the evidence is incomplete and a coherent phenotypic or biotypic profile predicting benefit remains elusive. Emerging signals point toward metabolic–inflammatory alterations and cognitive impairment as relevant domains. The authors conclude that iterative, agnostic, data‑driven methods are likely required to produce actionable baseline predictive information for ketamine treatment.