Comparison of characteristics, management and outcomes in hospital-onset and community-onset stroke: a multi-centre registry-based cohort study of acute stroke

OBJECTIVE: Hospital-onset stroke (HOS) is associated with poorer outcomes than community-onset stroke (COS). Previous studies have variably documented patient characteristics and outcome measures; here, we compare in detail characteristics, management and outcomes of HOS and COS. METHODS: A total of 1656 men (mean age ± SD = 73.1 years ± 13.2) and 1653 women (79.3 years ± 13.0), with data prospectively collected (2014–2016) from the Sentinel Stroke National Audit Programme, were admitted with acute stroke in four UK hyperacute stroke units (HASU). Associations between variables were examined by chi-squared tests and multivariable logistic regression (COS as reference). RESULTS: There were 272 HOS and 3037 COS patients with mean ages of 80.2 years ± 12.5 and 76.4 years ± SD13.5 and equal sex distribution. Compared to COS, HOS had higher proportions ≥ 80 years (64.0% vs 46.4%), congestive heart failure (16.9% vs 4.9%), atrial fibrillation (25.0% vs 19.7%) and pre-stroke disability (9.6% vs 5.1%), and similar history of stroke, hypertension, diabetes, stroke type and severity of stroke. After age, sex and co-morbidities adjustments, HOS had greater risk of pneumonia: OR (95%CI) = 1.9 (1.3–2.6); malnutrition: OR = 2.2 (1.7–2.9); immediate thrombolysis complications: OR = 5.3 (1.5–18.2); length of stay on HASU > 3 weeks: OR = 2.5 (1.8–3.4); post-stroke disability: OR = 1.8 (1.4–2.4); and in-hospital mortality: OR = 1.8 (1.2–2.4), as well as greater support at discharge including palliative care: OR = 1.9 (1.3–2.8); nursing care: OR = 2.0 (1.3–4.0), help for daily living activities: OR = 1.6 (1.1–2.2); and joint-care planning: OR = 1.5 (1.1–1.9). CONCLUSIONS: This detailed analysis of underlying differences in subject characteristics between patients with HOS or COS and adverse consequences provides further insights into understanding poorer outcomes associated with HOS.