Background:
Infections in older people are a major cause of mortality and morbidity and may be important modifiers of outcomes for people with dementia. We explored mortality in individuals with dementia admitted to hospital for infections alongside deprivation.

Methods:
A retrospective cohort analysis was created using secondary care data from one of Europe’s largest mental health and dementia care providers in South London (UK). People with recorded dementia between January 2008 and March 2017 were linked to national hospitalisation and mortality data and classified as either admitted for infections, admitted for other causes or no admissions around first recorded dementia diagnosis and followed to death or a census date on 31st March 2019. Multivariate Cox proportional hazards regression with date of first dementia diagnosis as index date were applied.

Results:
In total 14,375 people diagnosed with dementia were included, of whom 3,480 (24%) were admitted for infections, 3604 (25%) admitted for other causes and 7,291 (51%) not admitted to hospital. Individuals with dementia admitted for infections had lower MMSE scores, higher rates of co-morbidity and ranked significantly higher on deprivation indices compared to those with no admissions. Adjusted multivariate Cox regression models indicated that people with dementia admitted to hospital for infections had significantly higher mortality rates compared to no admissions (Hazard ratio (HR): 2.3; 95% confidence interval (CI): 2.1 - 2.5) and to admissions for other causes (HR: 1.5; 95% CI: 1.4 - 1.6).

Conclusions:
In people with dementia, admissions for infections are critical events associated with substantial increased mortality risk. Addressing care patterns both before and after admission with a focus on infection could improve outcomes and is a key health research area, with optimisation of care a direct priority.