Delirium Subtypes and Prognosis in Non-Intubated Critically Ill Patients: Subanalysis of an RCT

Introduction: Delirium subtypes are associated with clinical outcomes in critically ill patients. Previous delirium studies focused on intubated patients receiving sedatives, potentially prohibiting precise assessment of delirium. We hypothesized that, in non-intubated critically ill patients, delirium subtypes and their prognoses could be more accurately and clearly described.


Methods: We performed a subanalysis of 393 non-intubated critically ill patients from the DEX-HD trial, which compared dexmedetomidine and haloperidol for sedating nocturnal hyperactive delirium in those who were emergently admitted to two high-dependency units (HDUs) of a tertiary care hospital. Delirium was assessed twice daily with RASS and CAM-ICU over the first 7 days of HDU stay, and categorized into four groups: none, hyperactive, hypoactive, and mixed. Outcomes included in-hospital mortality, changes in functional status, and HDU length of stay. Functional status was assessed using a scale to measure the patient’s need for medical/nursing care used across Japan (turning in bed, transferring, oral care, eating, dressing, understanding care instructions, and dangerous behavior) with a maximum score of 12 points (higher scores indicate greater dependency). Delirium trajectories over the 7 days were also analyzed.


Results: Mean age was 78.7 (SD 14.9) years, and 52.2% were male, with median SOFA-SF score of 3 (IQR 2 – 4). Delirium occurred in 39.9% of the patients, with hyperactive accounting for 17.8%, hypoactive for 28.7%, and mixed for 53.5% of delirium cases. In-hospital mortality rates were 1.7%, 0.0%, 13.3%, and 6.0%, respectively. Median change in overall functional status was -1 (IQR -3 - 0), 0 (IQR -2 - 1), 0 (IQR -2 - 1), and 1 (IQR -1 - 3) (P < 0.001). Median HDU length of stay was 3 days (IQR 1–5), 4 (IQR 3–9), 4 (IQR 2–8), and 5 (IQR 2–9) (P < 0.001). The prevalence of delirium declined sharply in hyperactive, gradually in hypoactive, and prolonged in mixed.


Conclusions: In non-intubated critically ill patients in HDUs, mixed-type delirium was the most prevalent and may be associated with poorer prognosis. These findings support the importance of delirium subtyping and suggest that mixed-delirium may require prophylaxis, and targeted interventions.