Spectrum of spinal cord involvement in COVID-19: A systematic review

1 Introduction

COVID-19 pandemic has now wreaked havoc across the globe for around 6 months since WHO declared it as a pandemic on March 11, 2020. Initially, it was considered to be primarily a respiratory pathogen. However, with time it has been understood as a virus with the potential to cause multi-system involvement. Among reports of various organ involvements in COVID-19 neurological manifestations have drawn significant attention [1]. It is now known that diverse central and peripheral nervous system features may appear following SARS-CoV-2 infection [2]. Neurological features may occasionally precede the typical constitutional or respiratory symptoms of COVID-19 [3]. Among the CNS features stroke is the most frequently reported manifestation while demyelination [4] and seizures [5] are also being increasingly documented. COVID-19 related PNS involvement has mostly been seen in the form of Guillain-Barre Syndrome (GBS) and myositis [6].

In recent times, several reports have come up describing spinal cord involvement in the context of COVID-19. Ranging from vascular to demyelinating pathology, various forms of spinal cord involvement have been documented. Additionally, there are reports revealing cord enhancement in association with polyradiculoneuropathy [4]. In sum, spinal cord involvement in COVID-19 has been steadily receiving attention for some time now [7].

Seemingly the inter-relationship between demyelinating disorders and COVID-19 has two facets. SARS-CoV-2 infection may lead to spinal cord as well as brain demyelination. On the other hand, patients with known primary demyelinating disorders may experience an exacerbation of pre-existing neurological features. In the present systematic review, we set out with the objective of organizing the so far available evidence of spinal cord demyelination in the setting COVID-19 in terms of clinical presentation, laboratory features, and imaging characteristics. The presented summary of COVID-19 associated myelitis has important prognostic as well as therapeutic implications from the perspective of clinical neurology.

2 Methodology

3 Result

Around 750 articles were initially identified from the aforementioned databases and 253 articles were identified from different pre-prints servers. Finally, a total of 712 articles were identified from different databases searched after removing the duplicates. Of the selected articles, 600 articles were excluded after screening titles and abstracts, leaving 112 articles for full-text review for possible inclusion in this study. Of these, 60 articles were excluded based on the inclusion and exclusion criteria for the study sample (e.g., excluded demyelination of only brain, pre-existing demyelination related articles), and few other articles were excluded for study types (e.g review papers, correspondence, viewpoints, commentaries). A total of 52 articles were finally selected for this study; 21 articles were included in the analysis, and the remaining 31 articles were synthesized narratively [Figure 1].

• Download figure
• Open in new tab

Figure-1.

Flow chart showing the algorithm used to identify the studies based on spinal involvement in COVID-19 that met inclusion criteria. Flow diagram template adopted from PRISMA

The selected study comprises of 21 case reports with 52% (n=11) male and 48% (n=10) female along with the mean age around (46.66±17.97). They had developed various COVID-19 related symptoms such as fever (n=11, 52%),cough (n=6, 29%), myalgia (n=6, 29%),vomiting (n=2, 10%),nausea (n=1, 5%)and many more. The neurological symptoms reported were weakness (n=14, 66.66%), sensory deficit (n=14, 66.66%) ataxia (n=1, 4.76%), autonomic dysfunction (n=8, 38.09%),cognitive impairment (n=2, 9.52%).

Various CSF parameters under biochemical analysis such as Glucose (mg/dl) (n=5, 24%), Protein (mg/dl) (n=13, 70%), Adenosine Deaminase (μ/L) (n=1, 5%), Lactate Dehydrogenase (unit/L) (n=1, 5%) were recorded with significant levels of elevation from the normal range. Moreover certain cell count parameters such as WBC (n=2, 10%), RBC (n=2, 10%) was significantly higher than the normal range. Cases with Lymphocytic pleocytosis (n=5, 24%) were also reported. SARS-COV-2 was detected in (n=5, 24%) cases using RT-PCR (n=4) and IgG positive (n=1). SARS-COV-2 detection came negative in (n=13, 62%) cases whereas SARS-COV-2 testing was not done in two cases and data was unavailable for one case. Autoimmune profiling for majority of cases came negative except two cases with positive oligoclonal band (OCB) and lupus antigen respectively [Table-1].

Neuroimaging findings were reported in 20 studies consisting of 20 cases and the remaining 1 case imaging data were not available. Among the imaging modalities, MRI of the spine was done for all the 18 cases whereas brain MRI was found for 16 cases. CT scan of the spine and brain were performed for 1 and 2 cases respectively. Interestingly, we found a majority of cases (n=9, 45%) were associated with isolated LETM whereas a combination of both LETM and patchy was 10% (n=2). Imaging data also reveals that 10% (n=2) cases were associated with isolated patchy involvement and 25% cases (n=5) were associated with isolated short segment involvement. Furthermore, we found co-involvement of both brain and spine for 30% (n=6) cases [Table-2 and Figure-2].

• Download figure
• Open in new tab

Figure-2.

Summarized level and types of spinal cord involvements in COVID-19 as per included case reports (reference number corresponds to Table-2)

Treatment was mainly carried out with corticosteroids (n=18, 86%) like methyprednisolone (n=14, 67%), dexamethasone (n=2, 10%), oral prednisolone (n=1, 5%) etc. Intravenous immunoglobulin was given to (n=5, 24%) cases. Sessions of therapeutic plasma exchange were given to (n=8, 38%) patients. Antivirals drugs such as acyclovir, lopinavir/ritonavir, ganciclovir, favipiravir were administered to (n=7, 33.33%) patients with the most common drug of choice being Acyclovir in (n=5, 24%) cases. Hydroxychloroquine was given to three patients (14.28%). Two cases were reported where anti-epileptic drug was given. A majority of the patients recovered/discharged (n=15, 71%) with mortality of two cases (n=2, 10%) and four patients (n=4, 19%) were still under treatment during the period of study.

4 Discussion

During the course of this ongoing pandemic, several immunological features of SARS-CoV-2, affecting different organ systems have been observed by clinicians and researchers across the globe. In the context of neurological manifestations of COVID-19, immunological mechanisms are known to play important roles in giving rise to diverse clinical presentations affecting both CNS as well as PNS [29]. Quite antithetical to how SARS-CoV-2 affects PNS (mostly in the form of GBS) [29], data regarding spinal cord involvement are scarce. Acute or post-infectious myelitis is not uncommon with several viruses ascribable to its cause, such as Epstein-Barr virus, cytomegalovirus, measles virus, or rhinovirus. In addition, infection-related spinal cord demyelination is known to be associated with brain involvement as well, giving rise to the clinical picture of ADEM.

Spinal cord involvement in COVID-19 might be an under-recognized neurological complication of this novel infectious disease. In the present systematic review, we have found spinal cord involvement in 21 patients with COVID-19 along with their laboratory features and imaging abnormalities. We found 38.1% (n=8) of patients with para-infectious acute myelitis and 14.3% (n=3) with post-infectious acute myelitis. Weakness (66.7%), sensory deficit (66.7%), autonomic dysfunction including sphincter dysfunction (38.1%), and ataxia (4.8%) were the most frequent neurological manifestations at onset.

Besides, neuroimaging data revealed that half of the patients presented with longitudinally extensive transverse myelitis (LETM) followed by patchy (11.1%) and short (27.7%) segment involvement. LETM is usually observed in neuromyelitis optica spectrum disorder (NMOSD), infectious myelitis, lupus-related demyelination, and, occasionally, in MS. In India, tuberculosis is a known cause of LETM [30]. Of interest, half of the reported cases of myelitis in COVID-19 present with long-segment spinal cord involvement. From a clinician’s perspective, this observation basically extends the differential diagnosis of LETM to include COVID-19 related spinal cord demyelination. Due to the small number of available cases, we could not establish a definite pattern of brain involvement. Noteworthy, one-third of our patients revealed co-existent demyelination in varied areas across the brain (including the brainstem) in addition to the spinal cord. Laboratory parameters revealed elevated CSF protein and lymphocytic pleocytosis among reported cases of acute myelitis. These observations are also important from the perspective of a clinical neurologist. Concerning the prognosis, most of the patients came up with a complete disease resolution and the mortality rate was low (<10%). Noteworthy, microbes including Mycobacterial pneumoniae, Epstein Barr Virus (EBV), cytomegalovirus (CMV), rhinovirus, and measles are implicated in post infectious acute myelitis [31-33]. Reports have already suggested severe cytokine storm with significant elevation of IL-6, IL-7, IL-8, IL-9, TNF-alpha, IFN-gamma cytokines, and depletion of CD8⁺T_c cells and NK cells; i.e. lymphocytopenia [34].This event raises a possible hypothesis that the infectious organisms are targeted by the immune system which may also invade the CNS tissue including the spine due to structural similarities of microbial components and neuronal receptors. Different strains of coronaviruses such as HCoV-OC43 and MERS-CoV have been found to initiate several immunopathogenic responses which further cause the progression of demyelinating events in the central nervous system (CNS) [35]. The presence of HCoV RNA was found in the CNS of patients with demyelinating disorders, which suggests a possible association between coronaviruses and demyelination [36]. In a previous study, nucleic acids of HCoV-229E have been detected in the CNS tissue in four out of eleven multiple sclerosis (MS) patients, which suggest neurotropism of this species of coronavirus [37]. Moreover, coronavirus-like particles under electron microscopy were found from autopsies of two MS patients [38]. Using in situ hybridization technique with cloned coronavirus cDNA probes, Murray et.al detected coronavirus RNA in the demyelinating plaques in twelve out of twenty-two MS patients. Significant amounts of coronavirus antigen and RNA were observed in active demyelinating plaques from two patients with rapidly progressive MS [39]. A 3-year-old girl who was admitted with lower respiratory tract infection and acute flaccid paralysis showed co-infection with HCoV-229E and HCoV-OC43, which was detected by real-time PCR analysis of nasal swab samples [40]. Further, a 15-year old boy, who presented with acute disseminated encephalomyelitis (ADEM), tested positive for HCoV-OC43 in cerebrospinal fluid (CSF) using RT-PCR [41]. In a series of 4 patients affected with neurological complications due to MERS-CoV, one of them met a few criteria for diagnosis of the ADEM [42]. Furthermore, RNA recombination demonstrated that the S gene of the coronavirus mouse hepatitis virus (MHV) is related to certain molecular aspects of demyelination, which indicates the potential role of viral envelope S glycoproteins in autoimmune-induced demyelination [43].According to Kim et al, an experimental strain of CoVs JHM, even in the absences of T and B cells, developed an autoimmune demyelinating disorder in a mouse model suggesting that the formation of anti-JHM antibodies is sufficient to cause the demyelination. They also showed a decrement of such anti-JHM antibodies mediated demyelination by 90% and 76% in F_cR_r^-/- and complement depleting agents like cobra venom respectively. Overall observation suggests that direct autoimmune antibody formation against JHM structural protein is enough to start the cascade of demyelination [44]. Mutations in the spike glycoprotein of human coronavirus OC43 (HCoV-OC43) modulated the disease from chronic encephalitis to flaccid paralysis and demyelination in BALB/c mice [45]. Taken together, it seems that viral-mediated demyelination is very much evidential and concerning in times of a pandemic. Several cases of viral encephalitis have been found as another presentation of CNS involvement in context to acute COVID-19 indication. Interestingly, a few case reports, included in our study have also revealed the presence of SARS-CoV-2 RNA in CSF and brain MRI - findings consistent with meningoencephalitis, along with a simultaneous association of post or para-infectious acute myelitis. Wu et al hypothesized that SARS-CoV-2 may induce neuronal injury via hypoxic and immune mediated pathways. The presence of demyelination, as well as SARS-CoV-2 viral particles and genome sequence in the brain, has also been detected in autopsy studies [46, 47]. Besides, SARS-CoV-2 binds strongly with ACE2 receptors that have been distributed in kidney, lung, CNS, skeletal muscle, and the visceral tissues [ref-https://v15.proteinatlas.org/ENSG00000130234-ACE2/tissue]. Viral replication and subsequent turnover rate of ACE2 activation in CNS including spinal tissue may trigger systemic inflammatory response resulting in the increased permeability of blood brain barrier (BBB) and immune-mediated inflammation of the CNS [48]. Several other studies demonstrated that pathogenesis of severe viral infections is closely associated with the development of viral-induced severe inflammatory response syndrome (SIRS) like immune disorders [49] for SARS-CoV-2 infections. The pro-inflammatory state induced by cytokine storm mainly sustained by IL-6, IL-1, TNF-alpha may be the responsible for the activation of glial cells, which may also trigger the onset of demyelination [50].A possible hypothesis that comes up through such findings is the production of autoantibodies against glial cells triggered by the viral infection as a para or post-infectious phenomenon. Considering case reports and case series throughout this pandemic reporting multiple incidences of GBS [51] or GBS relate disorders (eg.-AMAN, AMSAN, and AIDP) [52] in COVID-19 along with profound documentation of viral encephalitis and encephalopathy; this systemic review elaborates unfathomable possibilities behind the onset of spinal cord demyelinating disorders, which needs to be addressed so that neurologists can be aware of the plethora of autoimmune neurologic complications involving CNS and can promptly be recognized and treated to minimize permanent neurologic disabilities and subsequent disease burden.

There are some implicit limitations in the present endeavor. Given the notable asymmetry between the total number of affected cases and reported cases of infection-related myelitis, it can be assumed that cases are presently under-reported which may be due to several reasons. Therefore, the present systematic review is based on a small number of cases even after an extensive search of available literature, both peer-reviewed, and pre-print. Also, several of the available reports do not describe the timeline of events in an organized manner making interpretation difficult. Laboratory features have also not been mentioned in detail in a few of the cases. In addition, there is considerable heterogeneity in the available data that may be considered a hindrance in advanced analysis. Despite these shortcomings, the present organized review will act as a preliminary guide for clinicians while dealing with suspected spinal cord demyelination in the context of SARS-CoV-2 infection.