Delirium is heterogeneous and can vary by etiology.

We sought to determine how delirium subtyped by etiology affected six-month function and cognition.

Prospective cohort study.

Tertiary care, academic medical center.

A total of 228 hospitalized patients > 65 years old were admitted from the emergency department (ED).

The modified Brief Confusion Assessment Method was used to determine delirium in the ED. Delirium etiology was determined by three trained physician reviewers using a Delirium Etiology checklist. Pre-illness and six-month function and cognition were determined using the Older American Resources and Services Activities of Daily Living (OARS ADL) questionnaire and the short-form Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE). Multiple linear regression was performed to determine if delirium etiology subtypes were associated with six-month function and cognition adjusted for baseline OARS ADL and IQCODE. Two-factor interactions were incorporated to determine pre-illness function or cognition-modified relationships between delirium subtypes and six-month function and cognition.

In patients with poorer pre-illness function only, delirium secondary to metabolic disturbance (β coefficient = −2.9 points, 95%CI: −0.3 to −5.6) and organ dysfunction (β coefficient = −4.3 points, 95%CI: −7.2 to −1.4) was significantly associated with poorer six-month function. In patients with intact cognition only, delirium secondary to central nervous system insults was significantly associated with poorer cognition (β coefficient = 0.69, 95%CI: 0.19 to 1.20).

Delirium is heterogeneous and different etiologies may have different prognostic implications. Furthermore, the effect of these delirium etiologies on outcome may be dependent on the patient's pre-illness functional status and cognition.