Towards a predictive model for post-stroke delirium

Primary objective: To assess predisposing and precipitating risk factors and create a predictive model for post-stroke delirium. Research design: A prospective observational study in a cohort of consecutive patients with ischemic stroke or intracerebral haematoma admitted within 24 hours of stroke onset. Methods: Patients were assessed daily for delirium during the first week by means of DSM-IV criteria and risk factors were recorded. Results: One hundred patients completed a 7-day evaluation (47 women and 53 men, median age 77 years). An episode of delirium was detected in 43 patients (43%). Using multivariate logistic regression, a predictive statistical model was developed that utilized independent risk factors: age (OR = 1.08; 95% CI = 1.02-1.15); intracerebral haemorrhage (OR = 6.11; 95% CI = 1.62-22.98), lesion volume > 40 ccm (OR = 3.99; 95% CI = 1.29-12.39) and either elevated gamma-glytamyl transferase (OR = 4.88; 95% CI = 1.45-16.35) and elevated serum bilirubin (OR = 3.70; 95% CI = 1.32-10.38) or maximum sequential organ failure assessment score >2 (OR = 3.33; 95% CI = 1.06-10.45) with acceptable sensitivity and specificity (69.0% and 80.7%). In ischemic strokes, total anterior circulation infarctions were more frequently associated with delirium (73.3% developed delirium) compared with the remainder of the groups combined (p = 0.004; OR = 6.66; 95% CI = 1.85-24.01). Conclusion: Higher age, metabolic disturbances, intracerebral haemorrhage and larger ischemic hemispheric strokes increase the risk of post-stroke delirium.