Impact of COVID-19 Infection on the Outcome of Patients With Ischemic Stroke

Joan Martí-Fàbregas, MD, PhD; Daniel Guisado-Alonso, MD; Raquel Delgado-Mederos, MD, PhD; Alejandro Martínez-Domeño, MD; Luis Prats-Sánchez, MD, PhD; Marina Guasch-Jiménez, MD; Pere Cardona, MD; Ana Núñez-Guillén, MD; Manuel Requena, MD; Marta Rubiera, MD, PhD; Marta Olivé, MD; Alejandro Bustamante, MD; Meritxell Gomis, MD, PhD; Sergio Amaro, MD, PhD; Laura Llull, MD, PhD; Xavier Ustrell, MD; Gislaine Castilho de Oliveira, RN; Laia Seró, MD; Manuel Gomez-Choco, MD, PhD; Luis Mena, MD; Joaquín Serena, MD, PhD; Saima Bashir Viturro, MD; Francisco Purroy, MD, PhD; Mikel Vicente, MD; Ana Rodríguez-Campello, MD, PhD; Angel Ois, MD, PhD; Esther Catena, MD; Maria Carmen Garcia-Carreira, MD; Oriol Barrachina, MD; Ernest Palomeras, MD, PhD; Jerzky Krupinski, MD, PhD; Marta Almeria, MD; Josep Zaragoza, MD; Patricia Esteve, MD; Dolores Cocho, MD, PhD; Antia Moreira, MD; Cecile van Eendenburg, MD; Javier Emilio Codas, MD; Natalia Pérez de la Ossa, MD, PhD; Mercè Salvat, RN; Pol Camps-Renom, MD, PhD


Stroke. 2021;52(12):3908-3917. 

In This Article


In our multicenter and prospective study of consecutive patients with acute ischemic stroke conducted in Catalonia during the COVID-19 outbreak, about 1 out of 8 patients had a concomitant infection by SARS-CoV-2. Compared with patients without COVID-19 infection, those with ischemic stroke and concomitant COVID-19 infection had a more severe neurological deficit at admission and at 72 hours and higher mortality (3.1-fold). Although we found an increased probability of an unfavourable functional outcome by the shift analysis, this was not confirmed when this variable was dichotomized. We found also a higher risk of neurological deterioration during the acute stage in the COVID-19 group. Although we found some interference with routine acute stroke management in patients with COVID-19, these logistic difficulties were not associated with outcome except for a worse functional outcome in those patients who were admitted to the ICU and a greater probability of death in patients not admitted to the Stroke Unit, both observations probably confounded by the presence of COVID-19.

The strengths of our study include the prospective design, the high number of patients with stroke with COVID-19 infection, and the comparison with a group without COVID-19 infection who were attended during the same time period. We obtained the data at the peak of the outbreak in our country. Moreover, the multi-center design including comprehensive and primary stroke centers covering most of the territory of Catalonia, supports the generalizability of our findings.

Stroke is one of the possible neurological complications of COVID-19 and although intracerebral hemorrhage and cerebral venous thrombosis have also been reported, most patients suffer from ischemic stroke. Stroke may precede, be coincidental or be diagnosed after the COVID-19 infection. Reported cases of ischemic stroke are included in retrospective registries,[4–7] or in small series.[13–15] In a multicentre study[6] of 3 hospitals from New York, 32 (0.9%) of 3556 admitted patients had imaging proven ischemic stroke. A Spanish single-center study[5] reported that 11 out of 841 (1.3%) patients with COVID-19 infection had an ischemic stroke. We observed a remarkable decrease of the proportion of patients with ischemic stroke and COVID-19 infection that dropped from 41.7% at mid-March to 0% at mid-May as reflected in Figure I (Data Supplement) thus showing a clear decline as the first wave of the pandemic waned.

We hypothesized that patients with ischemic stroke and COVID-19 infection would have worse prognosis. Despite presenting similar demographic, risk factor profile and time from the onset of symptoms, stroke severity measured by the NIHSS score was higher in the patients with COVID-19. Moreover, NIHSS score remained higher at 72 hours, thus reflecting the persistence of a worse neurological deficit. A recent retrospective study[6] found also higher baseline NIHSS score in patients with COVID-19 infection compared with patients non-COVID-19 (median NIHSS score: 19 versus 8). In comparison, the severity of stroke in our cohort, was lower (median NIHSS score: 8 versus 6).

We observed a nonsignificant lower proportion of favorable functional outcome and a significant higher mortality rate in patients with COVID-19 than in patients without. Functional outcome has not been evaluated in previous studies. We report a difference in the proportion of favourable outcome between groups and a higher risk of increasing 1 point in the mRS score in the shift analysis in the COVID-19 group compared with the non-COVID 19 group. It is difficult to determine if these differences were due to worse stroke severity in the setting of a COVID-19 infection or due to difficulties in stroke care and stroke rehabilitation in patients diagnosed of COVID-19. Based on our data, we believe that both possibilities should be considered. However, it is important to highlight that the difference on functional outcome observed between groups was not significant in the logistic regression multivariable analysis after adjusting by other well-known predictors of poor outcome such as older age,[16,17] baseline NIHSS score, and also ICU admission. In addition to comorbidities such as cardiovascular diseases, older patients with COVID-19 infection present molecular differences[18] that explain this life-threatening evolution. Prior stroke also predicted a worse outcome. This agrees with other prognostic studies.[1,18–20] A recent study reported that a history of stroke in patients with COVID-19 infection was independently associated with an increase in severe events and poorer outcomes after a propensity-matched analysis.[19] The authors attributed these findings to more aggressive inflammatory responses, and more underlying coagulation disorders in patients with prior stroke. Finally, admission to the ICU also worsened the outcome. We may speculate that admission to the ICU is reserved for the most critically ill patients with life-threatening complications.

Regarding mortality, our results agree with recently published studies[6,21] that reported that in patients with stroke, infection with COVID-19 was associated with a higher case-fatality. Remarkably, the multivariable analysis in our cohort showed that the exposition to COVID-19 was independently associated with mortality after adjusting by age, stroke severity and admission to the stroke unit. Several aspects related to the COVID-19 infection may explain our observation including the respiratory distress and the multiorgan failure observed in some patients.[22] In addition, besides the cardiovascular risk, aging[18,22] may explain the life-threatening evolution. Obesity has been described as an independent variable associated with increased mortality in patients with COVID-19,[23] although in our study it was not associated with a worse functional or vital outcome. Its prognostic significance could be influenced by the presence of other coexisting prognostic variables. Furthermore, it should be noted that in most cases the body mass index was not available and we relied on the assessment of the local investigator. As detailed in Table IV in the Data Supplement, about 50% of the deaths in the COVID-19 group was due to COVID-19 complications, while deaths due to stroke complications or recurrent stroke were more common in the non-COVID-19 group. If we put together, the results of functional and vital outcome, our study suggests that, despite the high mortality in patients with COVID-19 infection, survivors will have a similar likelihood of a favorable functional outcome than patients with ischemic stroke who do not have concomitant COVID-19 infection. In the survival curves, the curves markedly separate in the first days but then remain parallel until 3 months.

COVID-19 infection may possibly act as a trigger of conventional stroke causes.[24] It is possible that COVID-19 infection increases the risk of stroke, similarly to other viral infections.[16] In agreement with this assumption, patients from our cohort were typical of a stroke cohort, with a mean age of 72 years, predominance of men (60%) and a high proportion of patients with varied vascular risk factors. The higher proportion of strokes of unusual cause observed in the COVID-19 group may be explained by the multiple different COVID-19 related stroke pathogenesis described.[24] Among others, cytokine storm, prothrombotic state, antiphospholipid syndrome, other coagulopathies, myocardial injury, cardiac arrhythmias, and endothelial infection, they all have been proposed.[24–27] Finally, considering that the number of thrombectomies, a surrogate marker of LVO, was comparable in both groups, we did not find the previously reported[14] excess of young patients with large-vessel occlusion.

The pandemic has compromised seriously the ability of the health systems to care for patients with stroke and other acute medical emergencies, and these obstacles to the routine stroke care pathways may explain in part worse clinical outcomes. This has been reported by several authors in Spain[8,9] and other countries.[11,16,28,29] In our cohort, we found that compared to patients without COVID-19 infection, the COVID-19 group were less often admitted to the stroke unit, and had a longer hospital stay. However, assessment by a neurologist was performed at admission in about 80% of patients in both groups, and the proportion of patients treated with IVT or with MT as well as in-hospital times were also comparable. Remarkably, when analyzing the primary and secondary outcomes in patients with COVID-19 according to the site of admission (Stroke Unit versus COVID unit) we did not find significant differences. Although there is no doubt that the pandemic notably stressed the health care system, this affected non-COVID-19 patients as well as patients with COVID-19, with some differences between groups that had no influence on vital or functional outcome as assessed by the multivariable analysis.

Our study has some limitations. Patients were recruited within the first 48 hours of evolution, and therefore, the onset of the stroke had to be known. This prevented the inclusion of patients whose stroke was discovered after days or weeks in the ICU, and who often were intubated or under the effect of sedative medication. The stroke of these patients could be more severe and have different pathogenesis. Additionally, this is a hospital-based study and, therefore, the frequency and severity of the stroke associated with the COVID-19 infection may be biased towards more severe cases, assuming that the milder ones may not have sought medical attention or may not have gone to the hospital. Our classification of patients as COVID-19 or non-COVID-19 could be incorrect in some patients and lead to under diagnosis. In most patients a PCR test was performed, which can give false negatives results.[30,9] Our study design does not allow us to reliably distinguish between asymptomatic carriers and uninfected patients. This is important as recent studies suggest that 17% to 20% of SARS-CoV-2 infected patients are completely free of symptoms.[31] These patients are able to transmit the infection although less than symptomatic patients. However, we would like to emphasize that the results of the sensitivity analysis of patients excluding those in whom a PCR was not available were consistent with those reached in the analysis of the whole sample. The frequency of neurological worsening was measured in within the first 72 hours of hospital admission. The frequency and reasons for neurological and general worsening beyond 72 hours are not detailed in our study. Evidently, the higher proportion of ICU admissions and the higher frequency of intubation in patients in the COVID-19 group indicate that severe complications associated with the COVID infection were determinants of the worse prognosis at 3 months in this group. Finally, despite adjusting our results according to the admission of the patients to the Stroke Unit, we did not have information of the specific treatments that received patients not admitted to the Stroke Unit. This was probably different across participating centers and could have influenced poor outcomes in some sites due to suboptimal stroke care. However, when analyzing stroke outcomes of patients with COVID-19 in our study according to the admission site (Stroke Unit versus COVID unit) there were no significant differences. Further studies will help to elucidate whether changes in the standard stroke care of these patients influence functional outcome and mortality beyond COVID-19 infection.