Abstract
BACKGROUND: delirium affects up to 40% of older hospitalised patients, but there has been no systematic review focussing on risk factors for incident delirium in older medical inpatients. We aimed to synthesise data on risk factors for incident delirium and where possible conduct meta-analysis of these. METHODS: PubMed and Web of Science databases were searched (January 1987-August 2013). Studies were quality rated using the Newcastle-Ottawa Scale. We used the Mantel-Haenszel and inverse variance method to estimate the pooled odds ratio (OR) or mean difference for individual risk factors. RESULTS: eleven articles met inclusion criteria and were included for review. Total study population 2338 (411 patients with delirium/1927 controls). The commonest factors significantly associated with delirium were dementia, older age, co-morbid illness, severity of medical illness, infection, 'high-risk' medication use, diminished activities of daily living, immobility, sensory impairment, urinary catheterisation, urea and electrolyte imbalance and malnutrition. In pooled analyses, dementia (OR 6.62; 95% CI (confidence interval) 4.30, 10.19), illness severity (APACHE II) (MD (mean difference) 3.91; 95% CI 2.22, 5.59), visual impairment (OR 1.89; 95% CI 1.03, 3.47), urinary catheterisation (OR 3.16; 95% CI 1.26, 7.92), low albumin level (MD -3.14; 95% CI -5.99, -0.29) and length of hospital stay (OR 4.85; 95% CI 2.20, 7.50) were statistically significantly associated with delirium. CONCLUSION: we identified risk factors consistently associated with incident delirium following admission. These factors help to highlight older acute medical inpatients at risk of developing delirium during their hospital stay.