Background/Aims: Traditionally, incident (first-ever) stroke case fatality studies focus on hospitalised cases identified through admissions data rather than mortality. We aimed to study incident case distribution before stroke death and explore hospital admission and comorbidity patterns linked to fatal incident strokes.
Method: Using linked morbidity and mortality data from WA, with a 15-year lookback period, incident strokes between 2000-2020 were identified. Fatal cases were stratified into 1) SH: hospitalised due to stroke and died within 28 days, 2) OH: non-stroke hospitalisation with stroke as the underlying cause-of-death (UCoD) within 28 days, 3) COM: death from stroke without hospitalisations in the preceding 28 days. Descriptive statistics were used to characterise the cohort.
Results: Among 45,488 incident strokes (median age=73.6 years, males=51%), overall crude case fatality was 22% (18% male/27% female), with 60% in SH, 9% in OH, and 31% in COM. Females represented 59% of incident stroke deaths and were, on average, 4.8(95%CI: 4.3-5.2) years older than males. In-hospital mortality was 81% in SH, and the UCoD was acute stroke in 56% of patients. In OH, 55% died in-hospital with a median stay of 4 days (median stay=2 days for those discharged before death). Frequent principal diagnoses were injuries (21%), other cerebrovascular diseases (13%), and pneumonitis (6%). In COM, 10% were hospitalised in the year before death, with a median 307-days since last admission.
Conclusion: A high proportion of stroke deaths occur without recent hospitalisation, demonstrating a need for better identification of patients at high-risk in primary-care and other pre-hospital settings.