Neurological Complications of COVID-19: A Systematic Review of Literature

Background: To study the nature and frequency of occurrence of “signicant” neurological complications in coronavirus disease-2019 (COVID-19) via a systematic review of the literature. Methods: We screened all articles resulting from a search of PubMed, Cochrane, Google Scholar and Scopus, using the keywords "COVID-19 and CNS", "SARS-CoV-2 and CNS”, “COVID-19 and neurological manifestation”, “SARS2 and neurological manifestation” and “COVID-19 and Brain” looking for reports of signicant neurological manifestations that would potentially have an impact on the outcome. Results: Twenty-six articles met the inclusion criteria. The signicant neurological diagnoses reported were stroke, Guillain Barre Syndrome (GBS) and its variants, encephalitis, seizures, acute hemorrhagic necrotizing encephalopathy, acute disseminated encephalomyelitis (ADEM) and transverse myelitis. Although stroke, predominantly ischemic, was observed in ~ 6% of COVID-19 patients from Wuhan, China, mortality in this cohort was 38%. Of the 24 pooled patients with reports of etiology, 17 had large vessel occlusions. GBS occurred in 5/1200 (0.4%) of the COVID-19 cohort from Italy. One of the six reported encephalitis cases, the ADEM case and the report of transverse myelitis do not have data for conclusive diagnosis. Conclusion: The most frequent signicant neurological association with COVID-19 is stroke, predominantly ischemic. In a cohort from Wuhan, China, this was as frequent as ~ 6%, with a 38% mortality. Most common reported etiology is large vessel occlusion. Other reported signicant neurological complications are GBS/variants, encephalitis, seizures and acute hemorrhagic necrotizing encephalopathy. The reports of ADEM and transverse myelitis lacked diagnostically conclusive data.


Results
Of the twenty-six articles that quali ed for inclusion, four were retrospective cohort studies, six were case series describing 2-6 patients and sixteen were case reports of single patient descriptions. Of the four cohort studies, two studies used the same cohort of patients from Wuhan, China. One of these (Mao et al) [7] reviewed all neurological manifestations including stroke and minor manifestations such as anosmia, dysgeusia and the other (Li et al) [11] reported cerebrovascular complications only. Of the 26 articles (Table 1) included in this study for the discussion, data was pooled from 25 (excluding Mao et al due to assumption of overlapping cohort with Li et al) and generated 1520 patients. In all patients, COVID-19 was diagnosed based on nasopharyngeal reverse transcriptase polymerase chain reaction (RT-PCR) except two where diagnosis was made based on cerebrospinal uid RT-PCR [12,18]. Of the 1520, 56 patients quali ed as having signi cant neurological manifestations ( Table 1, Fig. 2). Thirty-one patients experienced a cerebrovascular complication, with 27 ischemic strokes [11,[19][20][21][22][23], 3 hemorrhagic strokes [11,15,23] and 1 cerebral venous sinus thrombosis [11]. There were 15 patients with GBS [13,14,[24][25][26][27][28][29][30] or its variants, 6 reported as encephalitis [12,18,[31][32][33], one with seizures [34], one with acute hemorrhagic necrotizing encephalopathy [16] one reported as transverse myelitis [35], and one reported as ADEM [36]. Nine of these cases were reported as fatal. The neurological diagnosis in these reported fatal cases were stroke -ischemic and hemorrhagic, encephalitis and GBS. CSF was reportedly tested by RT-PCR for SARS-CoV-2 in 13 and out of these 2 were positive [12,18].          , were more likely to present with severe disease (per American Thoracic Society guidelines for pneumonia) [37], and were more likely to have cardiovascular risk factors such as hypertension and diabetes. Further, they were more likely to have higher C reactive protein and D-dimer, re ecting a higher degree of in ammation and coagulability. The median duration from the rst symptoms of SARS-CoV to stroke was 10 days". Among patients with ischemic stroke, 5 had large vessel disease, 3 had small vessel disease, and 3 had cardioembolic phenomena [11]. All six patients in the Beyrouti et al series from UK "had large vessel occlusion with markedly elevated D-dimer levels (≥ 1000 µg/L). Three patients had multi-territory infarcts, two had concurrent venous thrombosis, and in two, ischemic strokes occurred despite therapeutic anticoagulation". All had severe disease per ATS criteria for pneumonia [37] and four had cardiovascular risk factors such as hypertension and diabetes,one was fatal [21].  [13,14,[24][25][26][27][28][29][30]. Of the total of 15 pooled cases of GBS/variants, 2 were reported as fatal.
Altogether, there have been six reports of encephalitis [12,18,[31][32][33], one of seizures [34], one of acute hemorrhagic necrotizing encephalopathy [16], one transverse myelitis and one of ADEM [35,36]. Of the six patients reported as having encephalitis, CSF pleocytosis was reported in 4/6, elevated CSF protein in 3/6, CSF RT-PCR was tested in 5/6 and was positive in 2/5 [12,18]. The one patient reported as having seizures had multiple generalized tonic-clonic seizures in the setting of fever, fatigue, and shortness of breath. MRI and CSF analysis (5 lymphocytes, normal glucose, and chemistry) were reported as normal [34]. One patient diagnosed with acute hemorrhagic necrotizing encephalopathy presented with impaired consciousness, brain MRI revealed bilaterally symmetric rim enhancing hemorrhagic lesions in thalami, medial temporal lobes and sub-insular regions hemorrhage. CSF was negative for HSV, VZV, WNV, but traumatic tap limited testing for SARS-CoV-2. CSF cell count and chemistry were not reported [16]. One patient presented with presumed transverse myelitis in the setting of COVID-19. The authors reported acute accid paralysis with bowel and bladder impairment and a sensory level at T10 and there is no con rmatory CSF or spine imaging to support the diagnosis of acute transverse myelitis [35,38]. One patient reported as ADEM [36] presented with fever, dysarthria, dysphagia and change in mentation. CSF showed normal glucose, protein and cell count. MRI was found to have extensive patchy lesions involving the bilateral frontoparietal, temporal, basal ganglia and thalami. With area of DWI and ADC changes and "minimal questionable enhancement". CSF for VZV, HSV 1, 2 and 6 and WNV pcr, Cryptococcal ag test was all negative. She was treated with hydroxychloroquine and intravenous immunoglobulins following which she improved in her clinical course.

Discussion
From the analysis of the included studies, the most frequent signi cant neurologic manifestation in COVID-19 patients is a stroke with a frequency of ~ 6%. Li et al. studied exclusively cerebrovascular disease in COVID-19 patients and reported a mortality of 38% in their cohort with stroke as a complication [8,11]. Out of 11 ischemic stroke cases, 5 had large vessel disease, 3 had small vessel disease and 3 had cardioembolic phenomena [11]. Two subsequent series, one from UK and the other from NY reported all large vessel strokes [21,22]. Pooling all the stroke patients where TOAST classi cation [39] was discussed, 17/24 had large vessel occlusions, 4 had small vessel disease and 3 had cardio-embolic etiology. One possible mechanism for thrombosis leading to ischemic stroke in patients with COVID-19 is the cytokine storm caused by SARS-CoV-2 viremia, leading to severe endothelial dysfunction and progressive microvascular thrombosis. Rise in D-dimer and the prolonged prothrombin time seen clinically support this phenomenon [40]. Activation of the coagulation cascade leading to disseminated intravascular coagulation can signi cantly contribute to the multiorgan involvement in patients with COVID-19, resulting in acute ischemic stroke, cerebral venous sinus thrombosis or intracerebral hemorrhage [8,40,41]. Moreover, the adhesion of SARS-CoV-2 to ACE2 receptors gains particular importance in the cases of intracerebral hemorrhage due to the inactivation of the receptor and subsequent dysfunction in blood pressure regulation [8,11,15,20,40]. It is also possible that some cases of ischemic stroke in COVID-19 patients have a cardio-embolic source from virus-related cardiac injury [11]. The large vessel occlusion (LVO) strokes have been shown to be associated with cardioembolic etiology [42].
There have been six reports described as encephalitis, but one of these lacks basic CSF data, and encephalitis was presumed due to the detection of viral DNA by RT-PCR in CSF [18]. However, mere detection of SARS-CoV-2 DNA by RT-PCR in the CSF is not necessarily equivalent to CNS infection. A traumatic lumbar puncture could have contaminated the CSF sample with the patient's blood, which contained the genetic material of SARS-CoV-2. More conclusive evidence of SARS-CoV-2 associated meningitis/encephalitis is depicted by a case report of a comatose patient with positive RT-PCR for SARS-CoV-2 in the CSF and neuroimaging features indicative of right lateral ventriculitis and encephalitis mainly on the right mesial lobe and hippocampus. Nasopharyngeal swab testing, however, was negative for SARS-CoV-2 [12]. Viral neuroinvasion and subsequent neuronal injury have also been proposed to contribute to the pathogenesis of these severe in ammatory complications of the central nervous system in coronavirus diseases [7,41]. New-onset seizures were reported in one patient with SARS-CoV-2 without accompanying evidence of encephalitis [34]. Patients with severe SARS-CoV-2 illness may have hypoxia, multiorgan failure, metabolic and electrolyte derangements that may predispose to encephalopathy and seizures [43]. Pro-in ammatory cytokines release interleukins 2, 6, 7, and 10, tumor necrosis factor α, and the granulocyte colony-stimulating factor, causing neuronal hyperexcitability through glutamate receptor activation leading to seizure [34].
The report of transverse myelitis has no objective data (CSF or spinal imaging) to con rm a diagnosis; acute accid paraparesis with sensory level at T10 and bowel and bladder involvement was the basis of diagnosis in this case. Therefore, the diagnosis is possible but not con rmed based on presented data [35]. One patient with possible acute disseminated encephalomyelitis (ADEM) did not have de nite evidence of in ammation and/or demyelination and did not ful ll the criteria for ADEM [36,44].
In addition to the central nervous system manifestations described above, the peripheral nervous system may also be affected and is particularly vulnerable to immune-mediated diseases. GBS occurred in 5/1200 (0.4%) patients with COVID-19 in one of the studies [13]. There were six other case report of GBS, one Miller Fisher syndrome, and two polyneuritis cranialis [14,[24][25][26][27][28][29][30]. Interestingly, neurotropism of human coronaviruses has been suggested by in vitro and in vivo studies that showed that certain strains of the viruses could persist in the human CNS by targeting oligodendrocytic and neuroglial cell lines [41].
As neurologists practicing during this pandemic, recognition of the extra-pulmonary consequences of COVID-19 infection will optimize these patients' management. A subpopulation potentially even more susceptible to neuroin ammation are those patients who already have immune-mediated neurological disorders and whose immunosuppressive therapy raises their risk of COVID-19 infection. Other populations of special interest are neurologically frail patients with previous diagnoses of stroke, traumatic brain injury, treatment-refractory epilepsy or severe neurodegenerative disorders. These patients are prone to pulmonary infections and may reside in nursing homes that increase the risk of infection with SARS-CoV-2.

Study Limitations
The results of this study were in uenced mainly by two large cohorts (Li et al. and Toscano et al.) [11,13], one of these was from China that looked at cerebrovascular disease alone and the other from Italy that looked at GBS alone. These results may not be applied widely due to genetic and environmental differences individual to regional geography. As discussed above, the diagnosis of possible encephalitis in one patient and possible transverse myelitis in one patient were not con rmed as these reports lacked conclusive data. Further, we can infer an association of these neurological complications with COVID-19; however, we cannot presume causality. The high mortality reported in stroke with the COVID-19 cohort cannot be presumed to be related to stroke; it may re ect the severity of underlying illness.

Conclusion
Although most SARS-CoV-2 infections predominantly involve the respiratory system, the brain is not immune to injury, and when neurologic complications associated with the virus occur, they can be severe or even fatal. The most frequent signi cant neurological reported complication associated with COVID-19 is stroke, mainly ischemic, with a frequency of ~ 6% and 38% mortality. The most common etiology reported is large vessel occlusions. Neurologists should also be aware of less common neurological associations of COVID-19, such as GBS/variants, encephalitis, seizures, and acute hemorrhagic necrotizing encephalopathy. Reports of transverse myelitis and ADEM lacked data for a conclusive diagnosis. Overall, of all cases with neurological complications, 2 had CSF RT-PCR positivity for SARS-CoV-2 of the 13 patients that were tested. Observational data is key to gaining a better understanding of the disease's natural history and plays a central hypothesis-forming role in pathophysiology and treatment trials. Further studies, however, are needed to identify subpopulations who may be at increased risk and to provide evidence-based treatment paradigms.

Declarations
Ethics approval and consent to participate: Not applicable.
Consent for publication: Not applicable.
Availability of data and materials: Data was extracted from the articles published in Google Scholar, PUBMED, Scopus. This will be provided on request.