Hospital use, institutionalisation and mortality associated with delirium

Design: Prospective, cohort study. Setting and Participants: 278 medical patients aged 75 years and over admitted acutely to a district general hospital in South Wales. Measurements: Patients were screened for delirium at presentation and on alternate days throughout their hospital stay. Assessments also included illness severity, preadmission cognition, co-morbidity and functional status. Patients were followed for 5 years to determine rates of institutionalization and mortality. Number of days in hospital in the 4 years prior to and 5 years after index admission were recorded. Results: Delirium was detected in 103 patients and excluded in 175. Median time to death was 162 days [Interquartile range (IQR) 21-556] for those with delirium compared to 1444 days (25% mortality 435 days, 75% mortality > 5 years) for those without (P < 0.001). After adjustment for multiple confounders, delirium was associated with an increased risk of death [hazard ratio range 2.0-3.5; P < 0.002]. Institutionalisation was higher in the first year following delirium (P=0.03). While those with delirium tended to be older with more preadmission cognitive impairment, greater functional dependency and more co-morbidity, they did not spend more days in hospital in the 4 years prior to index admission. Conclusions: Delirium is associated with high rates of institutionalization and an increased risk of death up to five years after index event. Prior to delirium, individuals seem to compensate for their vulnerability. The impact of delirium itself, directly or indirectly, may convert vulnerability into adverse outcome.