Intracranial microhemorrhages in the setting of COVID-19: what we know so far

Abstract

Background

Diffuse intracranial susceptibility abnormalities have recently been described among patients with coronavirus disease 2019 (COVID-19), although prior studies have consisted of case reports and/or series. This brief literature review seeks to compile and catalogue the available data to elucidate characteristic features of such findings.

Materials and methods

Scientific articles and studies on intracranial microhemorrhages in the setting of COVID-19 were searched on PubMed, Google Scholar, and the Cochrane Library. Included studies described intracranial microbleed(s) on magnetic resonance imaging in patients with COVID-19. If multiple patients were described, only patients with intracranial microhemorrhage on magnetic resonance imaging were included for analysis. Patient demographics, severity of illness (e.g. intensive care unit admission and/or intubation), time from diagnosis of COVID-19 to magnetic resonance imaging, and location(s) of any observed microhemorrhages were noted.

Results

A total of 39 patients with suspected intracranial microhemorrhages have been described in prior studies. The average age of patients was 64.7 years; 21.9% were women. The average time between COVID-19 diagnosis and magnetic resonance imaging was 31.7 days. All patients in the cohort were admitted to critical care and were either intubated or treated with extracorporeal membrane oxygenation during their clinical course. Microhemorrhages were most commonly located in the subcortical/juxtacortical white matter and corpus callosum.

Conclusions

Intracranial microbleeds are a well-documented finding in patients with severe COVID-19, and are most commonly callosal and subcortical/juxtacortical in location.

Introduction

As coronavirus disease 2019 (COVID-19) spreads around the world, the clinical and radiological features of the illness have continued to be defined. Although COVID-19, caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is predominantly an acute respiratory disease, it can affect numerous other systems, culminating in multiorgan dysfunction. ¹ A variety of neurovascular complications, in particular, have been documented in a number of prior studies, even in patients without virus in the cerebrospinal fluid (CSF). ² Neurological manifestations of COVID-19 include stroke, intracranial hemorrhage, seizure, and encephalopathy.^3–6 In fact, neurological abnormalities may be present in over 36% of infected patients, and may be the only observed symptom. ⁷

A number of case reports and limited case series have recently reported the presence of intracranial microhemorrhages among patients infected with COVID-19. However, the pathophysiological mechanism by which such hemorrhages develop remains uncertain. The purpose of this literature review is to compile and interpret the available data regarding intracranial microbleeds in patients with COVID-19, and to present the most frequently hypothesized physiological pathways by which such abnormalities are thought to occur.

Materials and methods

Search strategy and study selection

A broad literature search was conducted for any studies that discussed intracranial microhemorrhages in the setting of COVID-19. Searches on PubMed, Google Scholar, and the Cochrane Library were completed on 11 July 2020. Keywords ‘COVID-19’, ‘coronavirus’, ‘nCov’, ‘2019-nCoV’, ‘intracranial’, ‘hemorrhage’, ‘neurologic’, ‘microhemorrhage’, and ‘radiology’ were employed. Additional studies were retrieved from the archives of the authors and by screening the reference list of all included studies. Included studies required a description of intracranial susceptibility foci and/or presumed microhemorrhages in one or more patients with a laboratory-confirmed diagnosis of COVID-19. Only patients with brain magnetic resonance imaging (MRI) were included. If a case series included patients both with and without MRI, only the patients with MRI were included in the results. Included studies were not restricted by language, publication date, or clinical setting.

Data extraction

A single neuroradiology reviewer (JCB) reviewed all available studies. Relevant data were extracted from each study, including: (a) patient demographics; (b) severity of illness (e.g. whether the patient was admitted to the intensive care unit (ICU) and/or intubated); (c) time from diagnosis of COVID-19 to MRI; and (d) location(s) of any observed microbleeds. Only cases with microhemorrhages were extracted, as the purpose of this study was not to find the prevalence of such abnormalities in COIVD-19 patients. All findings were catalogued in an attempt to assess for similarities between cases.

Results

Included studies

A total of five relevant studies were identified, all of which were case reports/series or observational studies.^8–10 It should be noted that the study by Kremer et al. described two different patterns of intracranial hemorrhage: extensive and isolated white matter microhemorrhages and non-confluent white matter lesions with associated hemorrhage. ¹¹ Both of these patterns were included together in the results of the current review, although ‘micro’ hemorrhages were not explicitly defined in the latter category. One case series, by Franceshi et al., did note microhemorrhages in one of 10 patients with neurovascular manifestations of COVID-19. However, this was not included in the final analysis as the descriptions of the other patients did not report the presence or absence of susceptibility-weighted foci. In addition, the case report of a patient with acute hemorrhagic necrotizing encephalopathy by Poyiadji et al. was not included as it described a distinct intracranial process. ¹²

Patient characteristics and imaging findings

Relevant extracted data are detailed in Table 1. Prior studies have described a total of 42 patients with MRI evidence of intracranial microbleeds in COVID-19. Genders were reported in 35 patients: eight of 35 (22.9%) patients were women. The average age (when reported) was 64.6 years (standard deviation (SD) 2.3). The average time between COVID-19 diagnosis and MRI (when reported) was 30.8 days (SD 4.2). The most common located of suspected microbleeds was in the subcortical/juxtacortical white matter and corpus callosum. The internal capsules, cerebellar peduncles, basal ganglia, septum pellucidum, and cerebellum were less frequently involved (Figures 1 and 2). One study also noted a blooming signal within the subarachnoid space, potentially representing subarachnoid hemorrhage. For patients in whom the clinical status had been reported, all of the included patients were admitted to critical care and were either intubated or treated with extracorporeal membrane oxygenation (ECMO) at some point of their clinical course.

Table 1.

Catalogue of prior studies describing suspected microbleeds in patients with COVID-19.

Authors	No. of patients	Average age (years)	Gender	Average time COVID-19 diagnosis to MRI (days)	ICU	Intubated and/or ECMO	Microhemorrhage locations
Fitsiori et al. 20	9	67.7	7 M, 2 F	28.8	9/9	9/9	CC: 8/9 SWM/JWM: 7/9 IC: 5/9 CP: 5/9 BG: 2/9 SP: 1/9 Cerebellum: 1/9
Nicholson et al. 21	2	59.5	2 M	N/A	2/2	2/2	1 Patient: Predominantly SWM/JWM 1 Patient: Subarachnoid and subpial spaces
Kremer et al. 11	20	64	15 M, 5 F	33	20/20	N/A	11/20: Non-confluent FLAIR hyperintense white matter lesions ± enhancement, with associated hemorrhage 9/20: Extensive and isolated white matter microhemorrhages
Radmanesh et al. 26	7	N/A	N/A	N/A	7/7	7/7	Predominantly CC (particularly splenium) and/or JWM
Franceschi et al. 3	1	62	F	N/A	1/1	1/1	Cerebral hemispheres, cerebellum
Soldatelli et al. 9	1	67	M	25	1/1	1/1	CC, JWM
Vattoth et al. 8	1	66	F	N/A	1/1	1/1	JWM, IC
Nicolas-Jilwan et al. 10	1	59	M	11	1/1	1/1	JWM

ICU: intensive care unit; ECMO: extracorporeal membrane oxygenation; FLAIR: Fluid-attenuated inversion recovery; MRI: magnetic resonance imaging; CC: corpus callosum; SWM: subcortical white matter; JWM: juxtacortical white matter; IC: internal capsule; CP: cerebellar peduncle; BG: basal ganglia; SP: septum pellucidum.

Figure 1.

Artistic illustration showing the most frequent sites of intracranial microbleeds in coronavirus disease 2019 (COVID-19). Foci were most commonly noted in the juxtacortical white matter and corpus callosum, with less frequent involvement of the basal ganglia and internal capsules. Involvement of the cerebellum and cerebellar peduncles has also been noted (not shown).

Figure 2.

Example of microhemorrhages in a 53-year-old woman with a history of coronavirus disease 2019 (COVID-19). Axial susceptibility weighted images demonstrate numerous blooming foci, particularly within the juxtacortical white matter. No abnormal foci were seen in the corpus callosum.

Discussion

Intracranial microhemorrhage in patients infected with COVID-19 is a newly documented phenomenon. The results of this study indicate that microbleeds are often characteristically located in the corpus callosum and subcortical white matter, with additional involvement of the internal capsules and cerebellar peduncles. The proclivity of microbleeds to these locations differentiates their MRI appearance from unrelated pathologies such as hypertensive encephalopathy and amyloid angiopathy. ¹³

The clinical significance of these microhemorrhages remains undetermined. In the absence of COVID-19, microhemorrhages are often clinically silent, but can be associated with cognitive dysfunction and an increased risk of dementia.^14–16 The unique distribution of microhemorrhages seen in COVID-19 suggests that the impact on cognition would be similarly characteristic, although this requires further studying with neuropsychometric testing. Nevertheless, there is early evidence that COVID-19 patients with microbleeds have poorer clinical statuses and outcomes. ¹¹ More profound abnormalities in the corpus callosum in patients with COVID-19, such as infarctions, have raised the possibility of disconnection syndrome, although this hypothesis remains untested. ¹⁷ Nevertheless, isolated microhemorrhages – without more substantial imaging abnormalities – seem unlikely to cause any sort of similar clinical manifestation.

The pathophysiological basis by which such intracranial hemorrhages occur remains uncertain. There is a growing body of evidence that COVID-19 patients are susceptible to thromboembolic events. Platelet activation, endothelial dysfunction, and stasis have all been hypothesized to be contributing factors. ¹⁸ Inflammation, too, may play a role: both the severe inflammatory response that affects patients in the later stages of COVID-19 may potentially result in thrombotic microangiopathy and/or disseminated intravascular coagulation (DIC). ¹⁸ The resultant thromboemboli occur frequently in patients with COVID-19, with one study documenting pulmonary emboli in 30% of affected patients. ¹⁹ In addition, some of the prior studies noted imaging evidence of microthrombi. Fitsiori et al. noted that some of the T2* foci had a linear shape, raising the possibility that the abnormalities represented tiny intravascular thrombi. ²⁰ Nicholson et al. described a patient in which hyperdense cortical veins were observed, possibly representing slow flow or thrombosis within such vessels. ²¹

Alternatively, the observed abnormalities may represent true punctate foci of hemorrhage. The seemingly favored hypothesized mechanism for this pathway is that of endothelial damage related to the interaction of viral particles with angiotensin-converting enzyme 2 (ACE2) receptors on the vascular endothelium. According to this model, the subsequent endothelial damage may result in rupture of the cerebral capillaries with associated intracranial hemorrhage. ²² This theory is further evidenced by frank intracranial hemorrhage seen on computed tomography (CT) imaging as described in multiple patients by Nicholson et al. ²¹ In addition, a post-mortem examination completed on a patient with COVID-19 found hemorrhagic lesions throughout the cerebral white matter. However, the conductors of the post-mortem examination were unable to state the mechanism by which the lesions developed. ²³ It is also worth noting that the same biomechanism seems to play a role in hemorrhage elsewhere in the body, including the lungs. ²⁴

Finally, it is possible that at least some of the observed microhemorrhages are associated with critical illness, but are not specific to COVID-19. As per the study by Fanou et al., intracranial microbleeds in critically ill patients also have a proclivity for the corpus callosum and subcortical, with rarer involvement of the cerebellar peduncles and internal capsules, similar to those described in COVID-19. ²⁵ Radmanesh et al. posited that the observed microhemorrhages were related to hypoxemia, as the distribution was similar to that seen in delayed posthypoxic leukoencephalopathy. ²⁶

To address this, Lersy et al. compared the clinical characteristics of patients with COVID-19 that had either (a) a normal brain MRI or (b) extensive white matter microhemorrhages. ²⁷ The authors found that the patients with intracranial microhemorrhages were intubated for a longer period of time on average (24 days compared to 8 days, respectively; P=0.0002), required higher levels of fractional inspired oxygen (Fio₂) (100% vs. 75%, respectively; P = 0.03), and were more likely to be on ECMO (five patients versus no patients, respectively; P=0.04). These findings would seem to indicate that at least some of the observed abnormalities are related to microhemorrhages of critical illness, rather than necessarily being specific to COVID-19. Nevertheless, it remains possible that intracranial microhemorrhages only occur in patients with extremely severe clinical manifestations of the illness.

Furthermore, some or all of these mechanisms may contribute to the observed findings. That is, thrombosis, hemorrhage, and critical illness pathways could each play a role in the development of intracranial T2* foci. As the body of literature on COVID-19 continues to grow, a more exact explanatory model for the intracranial manifestations of the virus will hopefully be developed. Until then, it may be best to consider the proposed mechanisms as possibly representing interrelated processes that often result in numerous intracranial susceptibility weighted foci.

No study is without limitations. The current study is limited by the relative novelty of COVID-19, with only a few prior studies available for review. In addition, multiple other intracranial abnormalities have been described in patients with COVID-19, which are out of the scope of this article, but may be useful to describe in future reviews as such findings may help elucidate the pathophysiology of these microbleeds. Finally, because the heterogeneity of prior case series and case reports prevented statistical analysis the conclusions of the current study are based on a relatively subjective interpretation of prior results.

Conclusions

Intracranial microhemorrhages are a common finding on MRI in patients with severe COVID-19, and are most commonly observed in the corpus callosum and subcortical/juxtacortical white matter. The microhemorrhages reported in this case series all occurred in critically ill patients with severe respiratory distress requiring intubation. The most commonly hypothesized pathophysiological mechanisms for these abnormalities are hemorrhage, microthrombi, critical illness and/or hypoxemia, or a combination of each.