Background: The patient clinical phenotypes at particularly high risk for early cardiac complications after a recent acute ischaemic stroke (AIS), that is, stroke-heart syndrome (SHS), remain poorly defined. We utilised hierarchical cluster analysis to identify specific phenotypic profiles associated with this risk.
Methods: We gathered data on patients with AIS from the Virtual International Stroke Trials Archive, a global repository of clinical trial data. We examined cardiac complications within 30 days post-stroke, including acute coronary syndrome, heart failure, arrhythmias and cardiorespiratory arrest. We employed hierarchical cluster analysis to define distinct phenotypic risk profiles. The incidence/risk of SHS and 90-day mortality were compared across these profiles.
Results: We included 12,482 patients (mean age 69 ± 12 years; 55% male), yielding five phenotypes: Profile 1 ('elderly and AF'), Profile 2 ('young and smoker'), Profile 3 ('young'), Profile 4 ('cardiac comorbidities') and Profile 5 ('hypertension with atherosclerotic comorbidities'). Profiles 4 and 1 exhibited the highest risk for SHS (adjusted HR (95% CI): 2.01 (1.70-2.38) and 1.26 (1.05-1.51), respectively, compared to Profile 3), while Profiles 5 and 2 showed moderate risk and Profile 3 had the lowest risk. Although Profiles 1 and 4 were at the highest risk for most SHS presentations, Profile 5 had the highest risk for cardiorespiratory arrest (adjusted HR (95% CI): 2.99 (1.22-7.34)). The 90-day mortality risk was stratified by phenotype, with the highest risk observed in Profiles 5, and 4.
Conclusions: Hierarchical cluster analysis effectively identified phenotypes with the highest risk of SHS and early mortality in patients with AIS.
Keywords: Hierarchical cluster analysis; ischaemic stroke; stroke-heart syndrome.