Background and purpose: Mortality following infections in dementia has not yet been comprehensively explored. The aim of this cohort study was to investigate the short- and long-term mortality following infections in dementia. Methods: Follow-up was from 1 January 2000 or the 65-year birthday until death, immigration, or 31 December 2015. Exposure was incident dementia and a first infection. The outcome was all-cause mortality. Mortality rate ratios (MRRs) were calculated using Poisson regression in 4 exposure groups (dementia yes/no, infection yes/no) by sex, infection site, and time since infection. Results: 1,496,436 people were followed with 12,739,135 person-years. MRR in dementia/infection was 6.52 (95% confidence interval: 6.43–6.60) and was increased for infections of all sites. Increased mortality was short term (30 days) and long term (10 years). Conclusions: Increased mortality in people with dementia identifies them as a particularly vulnerable group that needs clinical attention.