Delirium and incident nursing home admission among people with and without dementia

Delirium is a common, preventable hospital complication and a recognised risk factor for cognitive decline, but its association with both the risk and timing of nursing home admission remains uncertain, particularly in relation to dementia.

We conducted a matched cohort study of hospitalised UK Biobank participants with linked hospital records (1997-2022). Episodes of delirium were identified using ICD-10 codes and matched 1:1 to non-delirium control episodes within no dementia (n = 13 004) and dementia (n = 1790) groups. Matching was performed on age, sex, hospital frailty risk score, primary diagnosis, length of stay and intensive care admission. Incident nursing home admissions were identified from hospital discharge records. Fine-Gray models estimated adjusted subdistribution hazard ratios (SHRs), while accelerated failure-time models quantified time to admission.

Delirium was associated with a higher risk of nursing home admission in participants without dementia (SHR 1.23, 95% CI 1.14-1.34) and with pre-existing dementia (SHR 1.22, 95% CI 1.07-1.38). Delirium also precipitated admission by ~40% in both groups, corresponding to 1.6 years earlier (95% CI 1.1-1.9) at a survival probability of 0.90 in those without dementia and 1.3 years earlier (95% CI 0.6-1.8) at a survival probability of 0.70 in those with dementia.

Hospital-recorded delirium was associated with both if and when older adults entered nursing home care. In contrast to previous evidence, this association was present regardless of pre-existing dementia. These findings highlight delirium prevention as a potential strategy to delay-or avert-transitions to long-term care.