Acute Ischemic Stroke During the Convalescent Phase of Asymptomatic COVID-2019 Infection in Men

Key Points

Question  Is the risk of acute ischemic stroke (AIS) elevated in patients in the convalescent phase of an asymptomatic COVID-19 infection?

Findings  In this case series of 18 male adults aged 50 years or younger who presented with AIS during the convalescent phase of an asymptomatic COVID-19 infection confirmed by a positive SARS-CoV-2 serological (antibodies) test result, the median onset of stroke was 2 months after the diagnosis of COVID-19.

Meaning  Results of this study suggest a persistent increased risk of AIS in individuals with asymptomatic COVID-19 months after serological diagnosis, warranting stroke units to be on alert and use SARS-CoV-2 serological testing.

Importance  Acute ischemic stroke (AIS) is a known neurological complication in patients with respiratory symptoms of COVID-19 infection. However, AIS has not been described as a late sequelae in patients without respiratory symptoms of COVID-19.

Objective  To assess AIS experienced by adults 50 years or younger in the convalescent phase of asymptomatic COVID-19 infection.

Design, Setting, and Participants  This case series prospectively identified consecutive male patients who received care for AIS from public health hospitals in Singapore between May 21, 2020, and October 14, 2020. All of these patients had laboratory-confirmed asymptomatic COVID-19 infection based on a positive SARS-CoV-2 serological (antibodies) test result. These patients were individuals from South Asian countries (India and Bangladesh) who were working in Singapore and living in dormitories. The total number of COVID-19 cases (54 485) in the worker dormitory population was the population at risk. Patients with ongoing respiratory symptoms or positive SARS-CoV-2 serological test results confirmed through reverse transcriptase–polymerase chain reaction nasopharyngeal swabs were excluded.

Main Outcomes and Measures  Clinical course, imaging, and laboratory findings were retrieved from the electronic medical records of each participating hospital. The incidence rate of AIS in the case series was compared with that of a historical age-, sex-, and ethnicity-matched national cohort.

Results  A total of 18 male patients, with a median (range) age of 41 (35-50) years and South Asian ethnicity, were included. The median (range) time from a positive serological test result to AIS was 54.5 (0-130) days. The median (range) National Institutes of Health Stroke Scale score was 5 (1-25). Ten patients (56%) presented with a large vessel occlusion, of whom 6 patients underwent intravenous thrombolysis and/or endovascular therapy. Only 3 patients (17%) had a possible cardiac source of embolus. The estimated annual incidence rate of AIS was 82.6 cases per 100 000 people in this study compared with 38.2 cases per 100 000 people in the historical age-, sex-, and ethnicity-matched cohort (rate ratio, 2.16; 95% CI, 1.36-3.48; P < .001).

Conclusions and Relevance  This case series suggests that the risk for AIS is higher in adults 50 years or younger during the convalescent period of a COVID-19 infection without respiratory symptoms. Acute ischemic stroke could be part of the next wave of complications of COVID-19, and stroke units should be on alert and use serological testing, especially in younger patients or in the absence of traditional risk factors.

As the COVID-19 pandemic progresses, many asymptomatic or mildly symptomatic cases of COVID-19 infection have been identified, either by contact tracing^1-3 or through surveillance programs, in various risk-stratified population groups.^4-6 At the same time, there have been reports of symptoms emerging or persisting long after the resolution of the original acute infection, which have been described as long-haul symptoms of COVID-19.⁷

Acute ischemic stroke (AIS) is a known neurological complication in patients with acute COVID-19 infection.^8,9 The mechanism of AIS that is associated with COVID-19 has been postulated to be secondary to an associated coagulopathy¹⁰ either by antiphospholipid antibodies¹¹ or endotheliopathy.¹² This theory has been observed in critically ill patients and in younger patients presenting with a large vessel occlusion.¹³ However, it is unknown whether patients who had an asymptomatic or a minimally symptomatic COVID-19 infection are similarly at risk for AIS as those patients who had overt acute respiratory COVID-19 illness.

In the city-state of Singapore, COVID-19 has been confirmed in 57 889 individuals as of October 14, 2020.¹⁴ Most of the infections have been localized to clusters of workers from South Asian countries (India and Bangladesh) who were living in dormitories,^14,15 accounting for 94% (54 485 cases) of all COVID-19 cases in Singapore. This clustering was primarily associated with the proximity of the inhabitants in residential complexes. Active surveillance of close contacts has identified many cases of COVID-19 through serological tests. In this case series, we assessed AIS that occurred in a series of men aged 50 years or younger in the convalescent phase of asymptomatic COVID-19 infection.

In this case series, all patients who experienced AIS and were under the care of public health care institutions in Singapore from May 21 to October 14, 2020 (a total of 21 weeks), were identified prospectively. These patients were admitted or referred to neurology units for the care of their AIS.¹⁶ This study was approved by the Singhealth Centralised Institutional Review Board, which granted a waiver of informed consent because of the observational nature of the study. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.¹⁷

We retrieved clinical course, imaging, and laboratory data from the electronic medical records of each participating hospital. Acute ischemic stroke was confirmed by neuroimaging using either computed tomography angiography and computed tomography of the brain or magnetic resonance imaging and magnetic resonance angiography of the brain. Included patients had asymptomatic or no respiratory symptoms of COVID-19, which was confirmed by a positive SARS-CoV-2 serological (antibodies) test result. Serological testing was performed using either the Architect SARS-CoV-2 IgG assay (Abbott Diagnostics) or the Elecsys Anti-SARS-CoV-2 assay (Roche Diagnostics), which are immunoassays designed to detect the nucleocapsid antibody of SARS-CoV-2. All patients were managed by their respective neurologists, and the tests performed were according to the physician’s discretion. Patients were excluded if they had ongoing respiratory symptoms or positive SARS-CoV-2 test results confirmed through reverse transcriptase–polymerase chain reaction (RT-PCR) nasopharyngeal swabs.

We calculated the annual incidence rate of AIS for this study population and compared it with the annual incidence rate of a historical, national ischemic stroke cohort (obtained from the Singapore Ministry of Health) that was matched by age, sex, and ethnicity (Indian and Bangladeshi).¹⁸ The population at risk in the study cohort was the total number of confirmed COVID-19 cases (n = 54 485) within the dormitory population in Singapore.

Rate ratio (95% CI) and significance were calculated with the statistical test described by Rothman,¹⁹ in which a 2-sided, unpaired P < .05 was used to indicate statistical significance. Stata release 16 (StataCorp LLC) was used for the statistical analysis.

A total of 18 consecutive male patients, with a median (range) age of 41 (35-50) years (Table 1), presented with AIS as the initial but delayed manifestation of COVID-19. Seventeen patients were asymptomatic for acute respiratory illness but were diagnosed with COVID-19 (by a positive SARS-CoV-2 serological test result) before their AIS. One patient (6%) was tested during an acute hospital stay for AIS. One patient (6%) reported mild diarrhea during the time of isolation but had no respiratory symptoms or anosmia. All patients were tested (and had negative results) at least once for COVID-19 using RT-PCR swabs because they were either in direct contact with individuals with COVID-19 infection or had stayed in the same dormitory facilities as others with COVID-19 infection.

All patients survived with no evidence of respiratory symptoms during their AIS hospitalization. All patients had negative nasopharyngeal and pharyngeal RT-PCR swab results for COVID-19 during their acute hospitalization for AIS.

The median (range) time from positive serological result to AIS was 54.5 (0-130) days. Chest radiographs were unremarkable in all patients, and 12 patients (67%) had no known preexisting risk factors of AIS (ie, hypertension and hyperlipidemia). The spectrum and severity of stroke varied among the 18 cases, with a median (range) National Institutes of Health Stroke Scale score of 5 (1-25) (Table 2). Six patients (33%) had intravenous thrombolysis and/or endovascular therapy. Ten patients (56%) had a large vessel occlusion, including 7 patients (39%) who had an anterior circulation large vessel occlusion that was detected on presentation through neuroradiological imaging, and thrombectomy was performed in 5 of 10 patients (50%).

An extensive evaluation of the etiological factors in AIS was performed for all patients (Table 2). Three patients (17%) had a postulated cardiogenic cause of stroke: left ventricular thrombus and patent foramen ovale. Four patients (22%) had elevated homocysteine levels. In particular, 1 patient had high levels of homocysteine (93 μmol/L), suggesting a possible genetic origin of hyperhomocysteinemia (Table 3). All other patients had normal cardiac function without an alternative cardiac source of embolism and no atrial fibrillation detected on at least 48 hours of telemetry. Two patients (11%) were deemed to have had a small vessel occlusion as the subtype of the stroke, whereas 3 patients (17%) were considered to have had cardioembolism in accordance with the Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification²⁰ (Table 2). Eight patients (44%) were diagnosed with a large vessel occlusion but had no obvious atherosclerosis, and the cause of stroke was classified as undetermined.

A screen for underlying coagulopathy was performed for all patients, and increased levels of dimerized plasmin fragment D (D-dimer) were detected in 3 patients (17%; Table 4). Two patients (11%) had positive results for lupus anticoagulant antibodies but negative results for other antiphospholipid antibodies (Table 5). Although not universally tested, von Willebrand factor antigen levels were elevated in 2 of 3 patients tested (Table 5).

The annual incidence rate of AIS in this all-male cohort with asymptomatic COVID-19 infection was 82.6 cases per 100 000 people. This rate was calculated on the basis of 18 patients who experienced AIS and a population at risk of 54 485 people over a study period of 21 weeks. The historical annual age-, sex-, and ethnicity-matched incidence rate of AIS was 38.2 cases per 100 000 people. This rate was calculated on the basis of 34 age-, sex-, and ethnicity-matched patients who experienced AIS from the 2018 national data and a similarly matched population at risk of 89 069 people over the same period.²¹ Therefore, the annual incidence rate in this all-male cohort was significantly higher compared with the annual incidence rate of the historical age-, sex-, and ethnicity-matched cohort (rate ratio, 2.16; 95% CI, 1.36-3.48; P < .001).

This case series highlights that adults 50 years or younger with asymptomatic or pauci-symptomatic COVID-19 infection diagnosed by positive SARS-CoV-2 serological test result may present with ischemic stroke during the convalescent phase of the infection. This finding underscores the value of SARS-CoV-2 serological testing during the etiological workup of patients who experienced AIS, given that a negative result from an RT-PCR test is expected during this period. Identifying antibodies to SARS-CoV-2 before AIS has public health implications, along with defining the complications associated with this emerging infection.

The patients in this series presented with AIS as the initial manifestation of COVID-19 infection a median of 54.5 days after the diagnosis was confirmed by a positive SARS-CoV-2 serological test result. This time from diagnosis to AIS was much longer compared with the time (estimated median of 16 days) for patients who presented with stroke and acute respiratory COVID-19 infection.^8,22 In particular, 1 patient presented with a large vessel occlusion at 130 days after a positive SARS-CoV-2 serological test result, suggesting that the prothrombotic nature of COVID-19 may persist long after an acute infection is resolved.¹⁰ Moreover, the presence of these antibodies in younger adults coupled with the lack of traditional cardiovascular risk factors suggests an etiological association. The presence of SARS-CoV-2 antibodies in combination with other stroke risk factors also may hasten the manifestation of ischemic stroke.²³ In addition, with an estimated production and persistence of COVID-19 antibodies about 2 weeks (although ranging from 1 to 6 weeks) after the initial COVID-19 infection,²⁴ coagulopathy may likely be observed for months after the initial exposure in patients with a subclinical COVID-19 infection.²⁵

In most of the cohort (56%), large vessel occlusions were detected on initial computed tomography or magnetic resonance angiography imaging, and 7 patients (39%) had an anterior circulation large vessel occlusion. This percentage was higher than an estimated 13% of patients in Hong Kong who had an ischemic stroke and an anterior circulation large vessel occlusion.²⁶ Even in patients without a large vessel occlusion, the infarct pattern suggested an embolic phenomenon from a proximal source. Further evaluation did not reveal the origin of any proximal thrombus or any atherosclerotic-related lesions. These observations suggest an embolic or a prothrombotic phenomenon as the cause of AIS. Despite assessing a thorough stroke workup, we were unable to find an underlying mechanism except the unifying positive SARS-CoV-2 serological test results in all 13 patients (72%) whose stroke had an undetermined classification. The percentage (72%) of strokes classified as having an undetermined cause is higher than the estimated 20% to 25% cryptogenic stroke observed in Asian patients.²⁷ In particular, a thorough cardiac workup showed that 1 patient with a left ventricular thrombus had no medical history or a medical reason to have a left ventricular thrombus. Again, the lack of an obvious source of thrombus formation suggested an etiological association with COVID-19.

The sustained prothrombotic mechanism in the convalescent phase of COVID-19 infection is currently uncertain and is an area of active research. In addition, the mechanism of stroke in patients with asymptomatic or mildly symptomatic COVID-19 infection likely differs from the mechanism in critically ill patients in the intensive care unit who have active respiratory COVID-19 infection. Overt inflammatory response and cytokine storm seen in critically ill patients with COVID-19 infection are factors in stroke through possible endothelial injury,²⁸ in which elevated D-dimer levels have been found to be associated with arterial thrombotic events in patients with active COVID-19 infection.^29,30 However, a recent coagulation study of patients 4 months after the resolution of respiratory COVID-19 infection demonstrated normal prothrombin time, fibrinogen level, D-dimer level, and von Willebrand factor antigen compared with levels in healthy control patients.³¹ This finding was consistent with the results of this case series, in which most patients had normal levels of these coagulation markers and only 3 patients had substantially elevated D-dimer levels. Sustained elevated levels of plasma factor VIII and plasminogen activator inhibitor 1 were found in the convalescent phase, partially explaining the hypercoagulable and hypofibrinolytic states.³¹ Tests for these coagulation markers were not routinely performed in clinical care and were not performed in this study cohort.

Overall, this case series revealed substantial heterogeneity in inflammatory and coagulation values across a spectrum of patients who experienced AIS in the convalescent period of asymptomatic COVID-19 infection. This finding suggests that the inciting prothrombotic mechanism of AIS in COVID-19 has yet to be defined in the convalescent phase and that current clinical inflammatory and coagulation markers are insufficient to identify individuals who are at risk for AIS and may only indicate the severity of the current stroke. Moreover, the variability of the tests performed by clinicians was also observed. Hence, to identify clinically useful and externally validated biomarkers for detecting arterial thrombosis, we propose testing a consistent panel of coagulation and inflammatory markers, which include plasma factor VIII or plasminogen activator inhibitor 1, across cohorts of patients with COVID-19 infection; these markers may be used prospectively and then subsequently repeated in the convalescent phase.

Antiphospholipid antibodies have been associated with AIS in patients with active respiratory COVID-19 infection.^11,29 However, in the present case series, only 2 patients had lupus anticoagulant antibodies. Moreover, both of these patients had no other antiphospholipid antibodies, such as anticardiolipin IgM and IgG or anti–β2-glycoprotein I IgM. Although 1 patient had a large vessel occlusion seen on initial cerebral vascular imaging, 9 other patients with a large vessel occlusion did not have antiphospholipid antibodies. These results suggest that the presence of antiphospholipid antibodies is not commonly associated with the post–COVID-19 procoagulant state and may only be a partial risk factor in developing a large vessel occlusion in patients in the convalescent phase of COVID-19. Therefore, it is uncertain whether SARS-CoV-2 antibodies are factors in the persistent antiphospholipid response in the convalescent phase that is associated with AIS.

The observed AIS incidence rate (82.6 cases per 100 000 people) in this case series was 2.16 times higher than the national historical stroke incidence rate, resulting in an annual excess of 44 cases per 100 000 people. With the ongoing COVID-19 pandemic affecting more than 50 million persons worldwide at the time of this writing, the additional cases may translate to thousands of cases of strokes worldwide. This novel observation warrants confirmation in other locations with high volumes of unexplained AIS in younger adults with COVID-19 infection. Furthermore, future studies need to examine the association of COVID-19 with the increased risk of strokes in the older population, which may be potentially higher.

A previous study has shown that serological assays exhibit diagnostic accuracy for COVID-19 only after 14 days of symptom onset, allowing appropriate antibody seroconversion in the host.³² The present report suggests another suitable case-use criterion for COVID-19 serological tests, taking into consideration its natural history and clinical course of infection. Given that these tests are designed to be qualitative in nature, future research may identify the association between quantitative antibody titers and the severity of stroke.

The need for universal COVID-19 serological screening in younger adults without respiratory symptoms who experienced a stroke is debatable, and further studies are required to define the subgroups in whom and the duration in which a prothrombotic or persistent inflammatory state is particularly pronounced. The implications for patients who have recovered from a symptomatic COVID-19 infection and their risk of stroke during the convalescent phase are uncertain. Even if patients with asymptomatic COVID-19 infection were identified using serological tests, we may still be unable to ascertain the therapeutic implications for thrombotic events in these patients. The recent halting of recruitment for 3 trials of critically ill patients with COVID-19 because of lack of improved outcomes³³ only demonstrates the lack of evidence for anticoagulation in COVID-19. In addition, the accuracy and limitations of existing serological methods for SARS-CoV-2 antibody detection have been openly debated and are an ongoing area of research and refinement.^34,35 Nevertheless, with an uncertain incidence of asymptomatic COVID-19 infection in the population, only comprehensive population-based serological testing may reveal the extent of the seroconversion rate, enabling the estimation of the true association of COVID-19 with AIS.

Given the paucity of risk factors for stroke apart from the presence of SARS-CoV-2 antibodies, AIS could be part of the next wave of complications in the COVID-19 pandemic. Stroke units should be on high alert and administer serological testing, especially in younger adults or in the absence of traditional risk factors.

This study has limitations. First, it was purely observational in nature, and coagulopathy testing was dependent on physician discretion and the availability of clinical resources at each health care institution during the pandemic. This limitation was exemplified by the minimal testing for von Willebrand factor³⁶ and other blood clotting factors, which are known coagulopathy risk factors in patients with COVID-19 infection. Second, all of the patients in this cohort were men from the South Asia region; most patients with COVID-19 infection in Singapore were living in the foreign worker dormitories with overwhelmingly male inhabitants. Hence, the findings related to AIS after COVID-19 infection may be generalizable only to a male South Asian population. In addition, the annual incidence rate was based on the single case series and should be interpreted with caution. A larger, population-based incidence rate is needed to verify the findings.

This case series described 18 adults 50 years or younger who presented with AIS in the convalescent period of COVID-19 infection. The findings suggest an increased risk of AIS for these patients months after a serological diagnosis. Stroke may be the next wave of complications from COVID-19. Hence, stroke units should be on alert and use serological testing, especially in younger patients or in the absence of traditional risk factors.

Accepted for Publication: March 4, 2021.

Published: April 22, 2021. doi:10.1001/jamanetworkopen.2021.7498

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Tu TM et al. JAMA Network Open.

Corresponding Author: Tian Ming Tu, MRCP, Department of Neurology, National Neuroscience Institute, 11 Jalan Tan Tock Seng, Singapore 308433 (tu.tian.ming@singhealth.com.sg).

Author Contributions: Dr Tu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Tu, Seet, Umapathi.

Acquisition, analysis, or interpretation of data: Tu, Seet, Tham, Koh, Chiew, De Leon, Chua, Hui, S. S. Y. Tan, Vasoo, B. Y.-Q. Tan, Tambyah, Yeo.

Drafting of the manuscript: Tu, Seet, Koh, De Leon, Chua, B. Y.-Q. Tan, Tambyah.

Critical revision of the manuscript for important intellectual content: Tu, Tham, Chiew, Hui, S. S. Y. Tan, Vasoo, B. Y.-Q. Tan, Umapathi, Tambyah, Yeo.

Statistical analysis: Tu, De Leon, B. Y.-Q. Tan.

Administrative, technical, or material support: Tham, Koh, Chiew, Hui, S. S. Y. Tan, B. Y.-Q. Tan, Umapathi, Tambyah, Yeo.

Supervision: Umapathi, Yeo.

Conflict of Interest Disclosures: Dr Tambyah reported receiving grants paid to the institution from Roche, Shionogi, Arcturus, and Johnson and Johnson as well as honorarium paid to the institution from AJ Biologics. Dr Yeo reported receiving personal fees from Stryker and grants from the National Medical Research Council of Singapore outside the submitted work. No other disclosures were reported.