Risk stratification models for stroke in patients hospitalized with COVID-19 infection: An American heart association COVID-19 CVD registry study

ABSTRACT

Introduction: Coronavirus Disease 2019 (COVID-19) is associated with an increased risk of stroke and worse stroke outcomes. A clinical score that can identify high-risk patients could enable closer monitoring and targeted preventative strategies. Methods: We used data from the AHA's COVID-19 CVD Registry to create a clinical score to predict the risk of stroke among patients hospitalized with COVID-19. We included patients aged >18 years who were hospitalized with COVID-19 at 122 centers from March 2020-March 2021. To build our score, we used demographics, preexisting comorbidities, home medications, and vital sign and lab values at admission. The outcome was a cerebrovascular event, defined as any ischemic or hemorrhagic stroke, TIA, or cerebral vein thrombosis. We used two separate analytical approaches to build the score. First, we used Cox regression with cross validation techniques to identify factors associated with the outcome in both univariable (p<0.10) and multivariable analyses (p<0.05), then assigned points for each variable based on corresponding coefficients. Second, we used regularized Cox regression, XGBoost, and Random Forest machine learning techniques to create an estimator using all available covariates. We used Harrel's C-statistic to measure discriminatory performance. Results: Among 21,420 patients hospitalized with COVID-19 (mean age 61 years, 54% men), 312 (1.5%) had a cerebrovascular event. Using traditional Cox regression, we created and internally validated a risk stratification score (CANDLE) (Fig) with a C-statistic of 0.66 (95% CI, 0.60-0.72). The machine learning estimator had similar discriminatory performance, with a C-statistic of 0.69 (95% CI, 0.65-0.72). For ischemic stroke or TIA, CANDLE's C-statistic was 0.67 (95% 0.59-0.76). Conclusion: We developed an easy-to-use clinical score, with similar performance to a machine learning estimator, to help stratify stroke risk among patients hospitalized with COVID-19.