The effect of poststroke delirium on short-term outcomes of elderly patients undergoing rehabilitation.

OBJECTIVES: Delirium is a common poststroke complication, but its prevalence and effect in rehabilitation settings is unknown. We retrospectively assessed the prevalence of delirium in elderly patients undergoing poststroke rehabilitation and its association with short-term outcomes.
METHODS: All patients (aged ≥65 years) admitted to the Department of Rehabilitation between November 2007 and October 2011 after a recent stroke were screened for delirium. Delirium was diagnosed using the confusion assessment method. Multiple logistic regressions were used to evaluate the association between delirium, institutionalization, and inhospital death, while multiple linear regressions were used for the association between delirium and functional recovery, defined in 3 different ways which include (1) measuring the relative functional gain of the Barthel index (BI-RFG); (2) the change in Barthel index (BI) walking subscore from admission to discharge; and (3) the change in Tinetti score from admission to discharge.
RESULTS: In all, 58 (33%) patients of the total 176 patients were consecutively admitted to our department with delirium. After adjustment for potential confounders, poststroke delirium (PSD) was an independent predictor of institutionalization (odds ratio [OR] = 7.23; 95% confidence interval [CI] = 4.79 to 10.91; P ≤ .0003) and inhospital death (OR = 4.26; 95% CI = 1.15 to 15.81; P = .03); PSD was not a predictor of functional recovery at discharge, neither using the BI-RFG (P = .96) nor using the change from admission to discharge of both the BI walking subscore (P = .57) and the Tinetti score (P = .61) as outcome measures.
CONCLUSIONS: In elderly patients undergoing poststroke rehabilitation, delirium is an independent predictor of institutionalization and inhospital death, but it does not affect functional recovery.