Delirium is a common syndrome in intensive care unit (ICU) patients. Those experiencing delirium may suddenly feel confused, have trouble thinking clearly, struggle to pay attention, or see and hear things that are not real. Delirium is associated with worse long-term outcomes such as cognitive impairment, depression, and PTSD (post-traumatic stress disorder). This study examines whether an investigational medical-grade ketone supplement drink (ketone monoester [brand name: Ultrapure Ketone Monoester]) is safe and feasible to use in ICU patients, and to look for signals that it might reduce delirium or shorten its duration compared to a volume-, taste-, and calorie-matched placebo.

Delirium is a prevalent neuropsychiatric syndrome characterized by an acute disturbance in attention, cognition, and consciousness. It is associated with significant morbidity, mortality, and healthcare expenditures. Recent research has provided evidence supporting the connection between brain metabolism and delirium. During states of increased systemic inflammation, such as sepsis or trauma, the brain experiences a mismatch between energy supply and demand, which is commonly associated with delirium, especially in those with preexisting cognitive impairment.

In critically ill patients, mitochondrial dysfunction occurs in the setting of systemic inflammation, contributing to increased blood-brain barrier permeability and neuroinflammation. The downstream consequence of this is microglial activation, which amplifies the inflammatory response through the release of pro-inflammatory cytokines. The resultant mitochondrial dysfunction leads to impaired oxidative phosphorylation, decreased adenosine triphosphate (ATP) production, and increased reactive oxygen species production. In response to systemic inflammation, microglia transition to a pro-inflammatory phenotype characterized by increased aerobic glycolysis. This metabolic reprogramming depletes glucose availability for neurons and exacerbates the cerebral energy deficit. Emerging evidence suggests that activated microglia compete with neurons for metabolic substrates during inflammation. Activated microglia exhibit metabolic flexibility, shifting toward increased glycolysis to meet their heightened energy and biosynthetic demands. This competition for nutrients exacerbates the neuronal energy deficit and increases metabolic stress. The investigators hypothesize that this brain energy deficit contributes to the cognitive and neurological symptoms characteristic of delirium.

Ketones, such as β-hydroxybutyrate, are the brain's secondary source of energy when glucose is not available. After transport across the blood-brain barrier, β-hydroxybutyrate is metabolized to acetyl-CoA (acetyl coenzyme A), thereby directly entering the tricarboxylic acid cycle, bypassing the glycolytic bottleneck, to produce ATP. In addition to serving as a substrate for ATP production, ketones support mitochondrial function, limit oxidative stress, and reduce neuroinflammation. Ketones confer a two-fold therapeutic advantage in the setting of nutrient competition. Not only do they support neuronal oxidative phosphorylation by bypassing impaired glycolysis, but they also promote anti-inflammatory microglial phenotypes, inhibit inflammasome activation, and support metabolic reprogramming. This dual effect further reduces microglial glucose demand, enhancing neuronal substrate availability.

The investigators propose a prospective, randomized, placebo-controlled pilot study of exogenous ketone ester supplement administration in 40 critically ill patients to assess the safety and feasibility of this novel intervention and to generate preliminary data on its efficacy in reducing ICU delirium, as measured by delirium and coma free days (DCFDs). Exogenous ketones have been shown to support brain energetics and reduce neuroinflammation, directly targeting pathways implicated in the development of delirium. By reducing the duration of delirium or preventing its onset, this research has the potential to improve long-term cognitive outcomes for ICU survivors. The investigators propose enrolling adult patients at the time of ICU admission, with randomization to either an enteral ketone ester treatment group or a taste, volume, and calorie-matched dextrose-containing placebo. The study drug or placebo will be administered at the time of enrollment, within 24 hours of ICU admission, and every six hours thereafter for up to 7 days until ICU discharge, or death, whichever occurs first. Ketone administration will be continued after the diagnosis of delirium. In accordance with prior studies, the initial dose of β-hydroxybutyrate will be 25 g; however, subsequent doses will be titrated to maintain serum β-hydroxybutyrate levels between 1.5 and 3.5 mM, with protocolized monitoring of vital signs, serum pH, glucose levels, and adverse gastrointestinal effects. Delirium will be assessed using the Confusion Assessment Method for the ICU (CAM-ICU) delirium screening tool twice daily for a period of 7 days.

This pilot study will assess the feasibility, safety, and tolerability of oral exogenous ketone supplementation in critically ill patients. The goal is to demonstrate that ketone administration is well-tolerated, with no significant safety concerns, consistent with prior evidence that oral ketones can be administered safely, even in vulnerable patient populations. Successful completion of this aim will establish a safety profile for ketone use in the ICU, which is essential before adopting this novel therapy for critically ill patients. The investigators hypothesize that patients receiving ketones will have more DCFDs compared to those receiving a placebo. The investigators will also perform an exploratory analysis of the biological impact of ketone therapy by examining biomarkers associated with delirium and ketone metabolism through serial measurement of serum levels of peripheral inflammatory mediators, metabolic stress assays, β-hydroxybutyrate levels, and markers of central nervous system (CNS) injury.

Ketones offer a promising novel therapeutic option for delirium. By targeting the underlying neurometabolic and neuroinflammatory changes associated with delirium, they support energy production, decrease oxidative stress, and modulate inflammation. Patients with preexisting cognitive impairment, such as those with mild cognitive impairment or Alzheimer's dementia, exhibit a baseline brain energy gap due to impaired cerebral glucose metabolism. This chronic energy deficit increases the vulnerability of the aging brain to delirium. Furthermore, the neurometabolic consequences of delirium in those with preexisting cognitive impairment exacerbate the brain energy gap, accelerating cognitive decline. The safety, tolerability, and rapid induction of ketosis following oral administration of ketone esters, in addition to the aforementioned beneficial effects, suggest this may be a therapy that could be initiated upon ICU admission as a potential preventative measure in those patients at risk. Ketones have the potential to transform delirium management and improve patient care; however, this clinical trial is required to evaluate the safety and efficacy of oral ketone ester supplementation in reducing the incidence, severity, and duration of delirium in critically ill patients.

Inclusion criteria:

1. Adult patients (≥18 years old) admitted to the medical intensive care unit.
2. Current ICU admission with anticipated ICU stay ≥24 hours.
3. Enteral access in place, planned enteral access placement, or PO intake appropriate, and the ability to receive enteral dosing within 24 hours of enrollment.
4. Ability to complete delirium assessments (CAM-ICU feasible) at time of enrollment.

Exclusion criteria:

1. Severe metabolic acidosis at screening: blood gas pH <7.20 or bicarbonate < 8 mmol/L.
2. Diabetic ketoacidosis as an ICU admission diagnosis or hyperketonemia from any ketoacidosis state.
3. Hypoglycemia as an ICU admission diagnosis or glucose <60 mg/dL.
4. Patients with a history of type 1 diabetes mellitus.
5. Hemoglobin <7.0.
6. Fulminant hepatic failure or AST/ALT > 5× ULN or total bilirubin > 3 mg/dL.
7. Refractory shock (defined as norepinephrine dose ≥20 µg/min or use of a second vasopressor agent).
8. Pregnancy (positive urine/serum hCG at screening or known pregnancy).
9. Uncontrolled ileus or gastrointestinal condition, such as an upper gastrointestinal bleed, preventing enteral dosing.
10. SGLT2 inhibitor use within the prior 7 days.
11. ADH/ALDH inhibitors (e.g., fomepizole, disulfiram) use in the prior 7 days or planned.
12. Severe dementia or neurodegenerative disease, defined as either impairment that prevents the patient from living independently at baseline or IQCODE >4.5, measured using a patient's qualified surrogate. This exclusion also pertains to mental illnesses requiring long-term institutionalization, acquired or congenital intellectual disability, severe neuromuscular disorders, Parkinson's disease, and Huntington's disease. It also excludes patients with severe deficits due to structural brain diseases such as stroke, intracranial hemorrhage, cranial trauma, malignancy, anoxic brain injury, or cerebral edema.
13. Benzodiazepine dependency or alcohol dependency based on the medical team's decision to institute a specific treatment plan involving benzodiazepines or barbiturates (either as continuous infusions or intermittent intravenous boluses) for this dependency.
14. Active seizures during this ICU admission being treated with intravenous benzodiazepines.
15. Expected death within 24 hours of enrollment or lack of commitment to aggressive treatment by family/medical team (e.g., likely to withdraw life support measures within 24 hours of screening).
16. Admission to ICU only for post-operative monitoring or frequent neurologic assessments.
17. Incarcerated status.
18. Inability to obtain informed consent within 24 hours from the time all inclusion criteria were met: Attending physician refusal.
19. Inability to obtain informed consent within 24 hours from the time all inclusion criteria were met: Patient and/or surrogate refusal.
20. Inability to obtain informed consent within 24 hours from the time all inclusion criteria were met: Patient unable to consent and no surrogate available.
21. Current enrollment in a study that does not allow co-enrollment.

De-identified individual participant data (IPD) for all randomized participants that underlie reported results, including: baseline demographics and clinical characteristics; eligibility variables; treatment assignment and dosing history; serum β-hydroxybutyrate and other protocol safety labs (e.g., glucose, acid-base measures); delirium/coma assessments (CAM-ICU, CAM-ICU-7) and derived delirium- and coma-free days through study day 7; key clinical outcomes (e.g., ICU/hospital length of stay, mortality); adverse events, dose holds/reductions/discontinuations; and research biomarker assay results. A data dictionary will be provided. Direct identifiers and free-text clinical notes will not be shared.

IPD and supporting documents will be available beginning 9 months after publication of the primary results and will remain available for 36 months following publication.

De-identified IPD and supporting documents will be shared with qualified researchers who submit a methodologically sound proposal. Access requires approval by the study team, execution of a data use agreement, and IRB/ethics approval as applicable. Data will be provided via secure electronic transfer. Requestors must agree not to attempt re-identification and to use the data only for the approved purpose.