Abstract
BACKGROUND: Traditional models for delirium based on the Diagnostic and Statistical Manual for Mental Disorders and its 1990 offspring, the Confusion Assessment Method (CAM), were not designed to distinguish behavioral and psychological symptoms of dementia from rapid cognitive decline. We examined a new diagnostic criterion for delirium plus exclusion of behavioral and psychological symptoms of dementia and recent inattention with a 25% decline in digit span forward (DSF). METHODS: This was a prospective, randomized controlled trial comparing management of prevalent delirium in general medical with that in geriatric medical wards in a 370-bed hospital north of Sydney. Inclusion criteria were age ≥65 years and prevalent delirium in the emergency department based on: CAM; proof that CAM elements were not better explained by behavioral and psychological symptoms of dementia; proof of recent inattention on DSF; evidence of cognitive decline not due to sedatives or antipsychotics in the emergency department. Measurements included the Instrumental Activities of Daily Living (IADL, 22-item), Selective IADL (8-item), Mini-Mental State Examination, DSF daily, Delirium Index daily, and Apathy Evaluation Scale. Pre-delirium scores from past cognitive tests and best scores were imputed after admission. Relative change (RC) was calculated as absolute change/test range and RC/MPC ratio was calculated as RC after admission/maximal possible change. RESULTS: A total of 130 subjects were recruited but 14 with subsyndromal delirium were excluded, leaving 116 subjects (mean age 83.6 years). Forty-eight percent had prior dementia. RC from pre-delirium to admission was 42% for the Mini-Mental State Examination, 41% for Selective IADL, 34% for 5-DSF, 54% for 6-DSF, and 37% for the Apathy Evaluation Scale. Improvements after admission (RC and RC/MPC ratios) were 32%/98% for 5-DSF, 54%/82% for 6-DSF, and 45%/80% for the Delirium Index. General medicine and geriatric medicine groups had similar outcomes. CONCLUSION: This delirium phenotype selects for a rapid high amplitude critical decline in attention, executive function, IADL, and apathy that recovers almost as rapidly.