Delirium

Risk factors

 * Sepsis (65% had delirium in one prospective study)

Specific models
Several complex models have been developed for delirium prediction.

E-PRE-DELIRIC model
 This model incorporates age, admission type, MAP at time of ICU admission, BUN, and presence/absence of cognitive impairment, alcohol abuse, corticosteroids, and respiratory failure and has been externally validated. It is too complicated for simple calculation, but an E-PRE-DELIRIC online calculator is available.

PRE-DELIRIC model
This model incorporates age, APACHE-II score, BUN, morphine dose, admission type, and presence/absence of coma, sedatives, morphine use, metabolic acidosis, and infections and was subsequently recalibrated and then externally validated. It is too complicated for simple calculation, but a PRE-DELIRIC online calculator is available.

Diagnosis


Confusion Assessment Method - ICU (CAM-ICU)
This is a well-validated method but lacks sensitivity and specificity in the brain injured population.



Intensive Care Delirium Screening Checklist (ICDSC)
This method seems to have greater sensitivity and specificity in the brain injured population (77% sensitve and 97% specific in one study).

EEG
One study of 102 septic patients suggested that patients with beta activity on the EEG were unlikely to have delirium, while those with delta activity were more likely. A reduction in beta activity tended to occur upon the transition from non-delirium to delirium. The presence of reactivity on EEG is associated with a lower risk of mortality (HR 0.5, 95% CI 0.2-0.9).

No episodes of seizures or NCSE were observed in one study of 102 patients with 6723 hours of vEEG. However, periodic discharges were present in 15% of patients, more commonly in those with delirium.

Biomarkers
Elevated levels of IL-6, IL-8, IL-10, TNF-α, CRP, and S-100β are associated with longer duration and greater severity of delirium. IL-8 and S-100β levels were associated with higher mortality.

Ramelteon
One small trial suggests benefit in patients age 65-89 years, but it is not useful for prevention in a broad ICU population. Several retrospective studies have suggested that ramelteon, by preventing sleep deprivation, may reduce delirium in older patients. In a small clinical trial (n=77) of general medical patients, ramelteon 8 mg qhs was associated with a lower risk of delirium (RR 0.09, 95% CI 0.01-0.69) compared to placebo in patients age 65-89 years old. Similarly, in another small trial of 88 ICU patients, those given ramelteon 8 mg/d had a lower incidence of delirium, length of delirium, and a trend toward shorter ICU stays.

In the largest trial to date, 120 ICU patients admitted for elective pulmonary thromboendarterectomy were given ramelteon 8 mg/d or placebo starting the night prior to surgery and continuing for up to 6 nights after the procedure. They found no difference in the incidence of delirium or ICU length of stay.