COVID-19 infection and stroke in hospitalised patients: A retrospective analysis of incidence, stroke character, risk factors and mortality

Background and Aims: COVID-19 infection is associated with thrombosis, but the relationship with stroke is not fully understood.We report the incidence, characteristics of stroke, demographics, risk factors and outcomes in hospitalised COVID-19 patients. Methods: We conducted a multi-centre retrospective case-control study across five London hospitals. All PCR-positive COVID-19 cases were reviewed, and radiologically confirmed stroke patients identified, with incidence calculated from sites that recorded all COVID-19 patients. Two control groups, both age-and sex-matched, were used: COVID-19-positive without stroke;and COVID-19-negative with stroke. Pertinent clinical, biochemical and radiological data were collected from patient records. Multivariate analysis was performed to assess the covariate effect of markers of COVID-19 severity and traditional stroke risk factors. Results: Between 1st March and 31st May 2020, 35 patients had a confirmed diagnosis of COVID-19 up to two weeks before or after a radiologically confirmed stroke, with 14(41%) requiring intensive care unit (ITU) care and an inpatient mortality rate of 29%. The incidence of stroke was 1.7%. The median age was 66(±14.8) years, median NIHSS 10(±9) and 23(67.6%) of these strokes were ischaemic. Peak D-dimer during admission was associated with stroke (p=0.03), with no significant associations with any other biochemical or radiological marker on admission. Diabetes mellitus had a negative association with stroke (p=0.03). Cumulative hazard analysis using a composite outcome of ITU admission or death demonstrates concomitant COVID-19 and stroke conferring a greater hazard than stroke alone (p<0.01), but not COVID-19 alone (p=0.50). Conclusions: COVID-19 patients with stroke had poor outcomes. Peak D-dimer during admission was positively associated while diabetes was negatively associated with stroke.