New-onset cognitive impairment after delirium among hospitalised older Ghanaians: A Pilot Study

Background Delirium is a potentially modifiable risk factor for dementia. However, post-delirium cognition has hardly been studied in Africa, where the burden of both disorders is rising disproportionately. Objective We estimated the incidence of new-onset cognitive impairment (NCI) 6 months after hospitalisation with non-stroke delirium among older Ghanaians. Methods This was a prospective observational pilot study conducted among 60 previously cognitively unimpaired (retrospective proxy assessment) adults aged ≥60 years, hospitalised with non-stroke prevalent delirium in a Tertiary Hospital. Delirium was confirmed and rated using the Confusion Assessment Method (CAM) and Severity Scale (CAM-S) within 24–72 hours of admission and reassessed twice weekly during hospitalisation, and at 6-week, 3- and 6-month follow-up visits. NCI was defined as Identification and Interventions for Dementia in Elderly Africans Study Cognitive Test (IDEA) score <8/15 + CAM-negative at 6 months. Demographic, clinical and psychosocial data were collected at baseline. We calculated the cumulative incidence function for NCI, accounting for competing mortality risk and right censoring due to dropout. We analysed predictors of NCI using Fisher’s exact test and Firth regression, and mortality using Cox proportional hazards regression. Results The cumulative incidence of NCI was 22.4% (95% CI: 8.0–36.8%). Mortality was 42.3% (95% CI: 26.2–58.3%). NCI was associated with advanced age (≥80 years; p=.001). Mortality was associated with CAM-S at enrolment (HR: 1.153; 95% CI: 1.020–1.303), mean CAM-S during hospitalisation (HR: 1.423; 95% CI: 1.261–1.605), hyperactive delirium (HR: 2.632; 95% CI: 1.119–6.188) and frailty at 6 weeks (HR: 1.817; 95% CI: 1.001–3.297). Conclusion In this pilot cohort of older Ghanaians hospitalized with delirium, NCI and mortality were common by 6 months. Longitudinal studies of post-delirium cognitive outcomes in diverse African cohorts are feasible and necessary, but future research should incorporate design, sampling and analytical measures to address survivor bias.