Introduction:
Ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, is increasingly utilized for analgosedation in mechanically ventilated intensive care unit (ICU) patients due to its analgesic and sedative properties. Its effects are dose dependent, with lower infusion rates (0.1–0.4 mg/kg/hr) providing analgesia and higher rates (0.4–1.0 mg/kg/hr) producing sedation. At higher doses, ketamine may cause dissociative or psychiatric effects that can resemble ICU delirium. However, existing data on the relationship between ketamine and delirium are limited and conflicting, often without adequate adjustment for confounders or consideration of dosing variability. This study aimed to evaluate the association between maximum daily ketamine infusion rate and ICU delirium. A secondary objective assessed the relationship between cumulative ketamine dose and delirium.
Methods:
This multicenter retrospective cohort study included adult medical ICU patients requiring mechanical ventilation and received ketamine infusions for sedation between January 1, 2013 and June 30, 2025 at two tertiary care hospitals. Patients were excluded if ketamine was used for non-sedation indications, administered for less than 24 hours, or if continuous neuromuscular blockade was used. Demographic and clinical data were collected, including age, race, sex, Sequential Organ Failure Assessment (SOFA) score, comorbidities, history of substance use or psychiatric illness, and sedation-related variables. Additional data included ketamine infusion rates, cumulative ketamine dose, duration of ketamine therapy, timing of ketamine initiation, concomitant sedative or analgesic use, Richmond Agitation-Sedation Scale (RASS) scores, and Confusion Assessment Method for the ICU (CAM-ICU) scores.
The primary outcome was ICU delirium, defined by a positive CAM-ICU at any point during ICU stay. A modified Poisson regression evaluated linear associations of delirium with the primary exposure variable, highest daily ketamine rate, and secondary exposure variable, cumulative ketamine dose. Covariates included SOFA score, cumulative morphine milligram equivalents (MME), cumulative benzodiazepines, and ICU length of stay (LOS). Logistic spline models and multivariate logistic regression were also conducted for both exposure variables. Other secondary outcomes, including mortality and LOS were analyzed using descriptive statistics, Mann Whitney U, or Chi Square as appropriate.
Results:
Of 837 screened patients, 114 met inclusion criteria and 42 (36.8%) developed delirium. Baseline characteristics were similar between groups, with a median age of 57–58 years. There was no difference between those who developed delirium and those who did not with regard to the maximum daily ketamine infusion rate (0.6 vs 0.43 mg/kg/hr, p=0.08) and cumulative doses (3440.6 vs 2461.4 mg, p=0.10). Patients with delirium had longer mechanical ventilation duration (210.7 vs 140.9 hours, p<0.01), longer ICU length of stay (364.7 vs 268.6 hours, p=0.04), and higher opioid exposure (5126.8 vs 2911.8 morphine milligram equivalents, p=0.85). Benzodiazepine use and mortality were not significantly different.
In the modified Poisson regression, there was no association between maximum daily ketamine infusion rate and delirium (Relative risk ratio (RR) 1.15, p=0.21) or cumulative ketamine dose and delirium (RR 1.0, p=0.97). Duration of mechanical ventilation was independently associated with delirium in both regression models (RR 1.0, p<0.01). No association between maximum daily ketamine infusion rate or cumulative ketamine dose with delirium was seen in logistic spline or multivariate regression models.
Conclusion:
Neither maximum daily ketamine infusion rate nor cumulative dose was associated with ICU delirium after adjustment for confounders. As seen in previous literature, duration of mechanical ventilation was associated with delirium. These findings suggest ketamine infusion rate and dosing may not significantly influence delirium risk, though prospective studies are needed to confirm these results.
