Acute Cerebrovascular Events With COVID-19 Infection

Mandip S. Dhamoon, MD, DrPH; Alison Thaler, MD; Kapil Gururangan, MD; Amit Kohli, MD; Daniella Sisniega, MD; Danielle Wheelwright, RN; Connor Mensching, MS; Johanna T. Fifi, MD; Michael G. Fara, MD, PhD; Nathalie Jette, MD, MSc; Ella Cohen, BS; Priya Dave, BA; Aislyn C. DiRisio, BS; Jonathan Goldstein, BA; Emma M. Loebel, BA; Naomi A. Mayman, BS; Akarsh Sharma, MS, BS; Daniel S. Thomas, BBA; Ruben D. Vega Perez, MPH; Mark R. Weingarten, BA; Huei Hsun Wen, MD, MSCR; Stanley Tuhrim, MD; Laura K. Stein, MD


Stroke. 2021;52(1):48-56. 

In This Article


In this large, multihospital retrospective observational study located in the epicenter of the COVID-19 pandemic in the United States, we found that 38% of all admitted strokes occurred in COVID-19-positive patients from March 1, 2020 through April 30, 2020. The COVID-19-positive cohort had more severe strokes and a higher prevalence of cryptogenic stroke mechanism and lobar stroke location. COVID-19-positive patients with stroke had mild coagulopathy, but the majority had elevated inflammatory markers. Most significantly, outcomes were much worse in the COVID-19-positive cohort compared with the COVID-19-negative cohort, and 33.3% suffered in-hospital death. The number of stroke admissions overall were reduced during the pandemic.

Substantial evidence has suggested that infection with COVID-19 may predispose to venous and arterial thromboembolism, to a greater degree with worse disease severity.[13,14] However, overall IS incidence during the COVID-19 pandemic appeared to have decreased, perhaps because people avoided seeking health care for more minor symptoms or symptoms went unrecognized in hospitalized patients whose deficits may have been masked by other sequelae of critical illness or sedation.[9,15,16] Prior studies have provided evidence for a prevalence of acute cerebrovascular disease, most commonly acute IS, of 2% to 6%, with greater prevalence in critically ill patients with worsened COVID-19 disease severity.[14,17–20] However, one recent study estimated that the prevalence may be <1%.[10] We found that 1.9% of COVID-19 patients admitted to MSHS had acute stroke, and overall admissions for stroke were reduced during the pandemic. This may have been because of the selective presentation of patients to medical attention. If patients with less debilitating symptoms did not seek health care during the pandemic, this would decrease the number of admissions for acute cerebrovascular events and increase the observed severity of strokes and associated morbidity and mortality during this period.

Most studies describing the occurrence of acute cerebrovascular disease in COVID-19 patients have been relatively small retrospective cohorts or case series.[7,21–25] Consequently, data on the incidence, cause, and outcomes of COVID-19-associated cerebrovascular events are lacking. In addition to traditional cerebrovascular risk factors, risk factors for cerebrovascular disease in the setting of COVID-19 infection have been thought to include hypercoagulability, whether secondary to the systemic inflammatory response to infection, or a COVID-19-specific cytokine storm and endothelial inflammation.[26–30] Coagulopathy, often evidenced by elevated D-dimer, and a history of cerebrovascular disease have been found to be associated with increased disease severity and mortality.[26,31,32] We, too, found traditional cerebrovascular risk factors to be present in the COVID-19-positive cohort, although not more common compared with the COVID-19-negative cohort. Additionally, a major behavioral cerebrovascular risk factor, smoking, was significantly less common in the COVID-19-positive cohort, but it is possible that behavioral risk factors such as smoking may be inaccurately recorded in the medical record among those in poorer health on admission. We also found evidence of a mild coagulopathy among COVID-19 patients. While we did not examine the association between coagulopathy and mortality, outcomes were significantly worse in the COVID-19-positive cohort. Others, including colleagues in Italy and New York City, have reported poorer outcomes in their COVID-19 patients with stroke as well.[10,16,26]

One of the largest studies comparing stroke between COVID-19 and non-COVID-19 patients during the pandemic came from Italy.[26] Of their 111 patients with cerebrovascular disease, a similar percentage to ours tested positive for COVID-19 (38.7%) and they found significant elevations in inflammatory markers; however, they reported a lower percentage of ICH (7.0%) compared with ours (15.2%). Notably, they did not describe patterns of treatment with anticoagulation, which represents a significant difference from our study given that 45.7% of our COVID-19-positive patients with stroke were treated with anticoagulation. The higher rate of treatment with anticoagulation among our COVID-19-patients with stroke may reflect an overall difference in approach to prevention of thrombotic events in COVID-19-positive patients, which may also explain the higher prevalence of ICH in our population. We report a mortality rate among COVID-19-positive patients with stroke of 33.3%, which is similar to the Italian cohort. However, that study did not comment on cause of stroke or provide details on stroke location and medical complications among COVID-19-positive patients with stroke.

A smaller study from New York City compared COVID-19 patients with stroke to concurrent and historical controls.[10] In this study of 32 COVID-19-positive patients with stroke, cryptogenic stroke occurred in 65.6% (compared with 52% observed in our study), and there were elevations in troponin and inflammatory markers in the COVID-19-positive patients. A larger percentage (78.1%) were treated with systemic anticoagulation. Notably, the in-hospital death rate was higher in their study (63.6%). This study also had limited detail about stroke location and medical complications among COVID-19-positive patients with stroke.

Results from whole blood analysis and thromboelastography with SARS-CoV-2 have revealed an underlying state of hypercoagulability, distinct from disseminated intravascular coagulation, associated with elevated levels of fibrinogen, Factor VIII, Protein C, von Willebrand Factor, C-reactive protein, and D-dimer.[33] Case reports have revealed a number of individuals found to have antibody profiles similar to those of the antiphospholipid syndrome with elevated anticardiolipin and β-2 glycoprotein I antibodies.[34] Each of these findings suggests dysregulation in the setting of a cytokine-mediated inflammatory response. A generalized inflammatory state may cause clotting factor dysfunction that is nonspecific and may be seen in other disease entities such as influenza.[35] Viral translation through ACE-2 receptors expressed in vessel walls may cause endotheliitis.[36] Along with this mechanism, metalloproteinase activation and procoagulant gene expression may contribute to thrombus formation and may explain the phenomenon of intramural thrombi and large vessel occlusions, which seem to be more common in infected patients of varying disease severity.[7,36] In addition, the SARS-CoV strain that caused the 2003 SARS epidemic has been implicated in increased stroke risk, possibly through the transcription of procoagulant genes.[37]

There are several strengths of this study, including the large sample size from a multihospital health system caring for patients in the epicenter of the US pandemic. We used an existing infrastructure of data collection that we expanded with rigorous and standardized data collection and end point adjudication by vascular neurologists. Limitations of this study include lack of data on cerebrovascular events occurring outside of the MSHS inpatient setting. Furthermore, this was an observational study, and we cannot comment on efficacy of particular treatments for COVID-19-related cerebrovascular disease. Regarding the comparison of stroke admissions with stroke admissions during all of 2019, this may represent an overestimation of expected number of stroke admissions since it includes winter months; the other comparison group involving admissions from March 1 through April 30, 2019 may represent a more accurate comparison. Last, the small sample sizes in our cohorts may have limited our ability to detect differences that actually exist. Also, due to the small number of events, we could not reliably run multivariable models due to the risk of overfitting the model, which would have resulted in inaccurate and biased results.

Future research will include ongoing collection of clinical and diagnostic characteristics of patients admitted to our health system with cerebrovascular disease and COVID-19, allowing longitudinal outcome collection, in-depth neuroimaging, and granular neurocognitive testing to elucidate the long-term effects and outcomes of COVID-19 infection after stroke.