Ritwick Mondal,a Shramana Deb,b Gourav Shome,c Upasana Ganguly,a Durjoy Lahiri,a,d,⁎ and Julián Benito-Leóne,f,g,⁎⁎ Mult Scler Relat Disord. 2021 Jun; 51: 102917.Published online 2021 Mar 21. doi: 10.1016/j.msard.2021.102917 PMCID: PMC7981271 PMID: 33845350
Abstract
Background
: Spinal cord complications associated with coronavirus infectious disease of 2019 (COVID-19) are being widely reported. The purpose of this systematic review was to summarize so far available pieces of evidence documenting de novo novel severe acute respiratory syndrome coronavirus (SARS-CoV-2) mediated spinal cord demyelinating diseases. Indeed, the spinal demyelinating disorders that have been reported in those patients who have suffered from COVID-19 rather than on the people already living with diagnosed or undiagnosed primary demyelinating disorders.
Methods
: We used the existing PRISMA consensus statement. Data were collected from PubMed, NIH Litcovid, EMBASE and Cochrane library databases, as well as Pre-print servers (medRxiv, bioRxiv, and pre-preints.org), until September 10, 2020, using pre-specified searching strategies.
Results
: The 21 selected articles were all case reports and included 11 (52%) men and 10 (48%) women. The mean age was of 46.7 ± 18.0. The neurological manifestations included weakness, sensory deficit, autonomic dysfunction and ataxia. In most cases, elevated cerebrospinal fluid protein as well as lymphocytic pleocytosis were found. SARS-CoV-2 was detected in five (24%) patients, meanwhile in 13 (62%) patients, the testing was negative. Testing was not performed in two cases and, in one, data were unavailable. Nearly half of the cases (N = 9) were associated with isolated long extensive transverse myelitis (LETM), whereas a combination of both LETM and patchy involvement was found in two. Only five patients had isolated short segment involvement and two patchy involvement. Furthermore, concomitant demyelination of both brain and spine was reported in six patients. Concerning the prognosis, most of the patients improved and the mortality rate was low (N = 2, <10%).
Conclusion
: Spinal cord demyelination should be added to the plethora of immune mediated neurologic complications associated with COVID-19.
1. Introduction
The novel severe acute respiratory syndrome coronavirus (SARS-CoV-2) has now prevailed across the world since WHO declared it as a pandemic on March 11, 2020. Initially, SARS-CoV-2 was primarily considered as a respiratory pathogen. However, with time it has behaved as a virus with the potential to cause multi-system involvement. Apart from pulmonary, renal, gastrointestinal, and hematological complications, neurological manifestations have also been reported frequently (Roy et al., 2021). It is currently known that several central and peripheral nervous system disorders may appear following SARS-CoV-2 infection (Roy et al., 2021, Yavarpour-Bali and Ghasemi-Kasman, 2020). Neurological features may occasionally precede the typical constitutional or respiratory symptoms of the coronavirus infectious disease of 2019 (COVID-19) (Lahiri and Ardila, 2020). Among the severe central nervous system (CNS) manifestations, stroke is one of the most frequently reported (Roy et al., 2021, Yavarpour-Bali and Ghasemi-Kasman, 2020, Lahiri and Ardila, 2020, Asadi-Pooya, 2020), meanwhile demyelinating disorders and seizures are now being increasingly documented as well (Roy et al., 2021, Yavarpour-Bali and Ghasemi-Kasman, 2020, Lahiri and Ardila, 2020, Asadi-Pooya, 2020, Ellul et al., 2020, Nepal et al., 2020). COVID-19 related peripheral nervous system (PNS) involvement has mostly been seen in the form of Guillain-Barré syndrome (GBS) and myositis (Roy et al., 2021, Yavarpour-Bali and Ghasemi-Kasman, 2020, Lahiri and Ardila, 2020, Asadi-Pooya, 2020, Nepal et al., 2020). Further, spinal cord complications associated with COVID-19 are being widely reported (Sánchez-Raya and Sampol, 2020).
The inter-relationship between demyelinating disorders and COVID-19 has two dimensions. On the one hand, SARS-CoV-2 infection may lead to brain and spinal cord demyelination. On the other hand, patients with known primary demyelinating disorders may experience an exacerbation of pre-existing neurological features (Roy et al., 2021). The purpose of this systematic review was to summarize so far available pieces of evidence documenting de novo SARS-CoV-2 mediated spinal cord demyelinating diseases. Indeed, the spinal demyelinating disorders that have been reported in those patients who have suffered from COVID-19 rather than on the people already living with diagnosed or undiagnosed primary demyelinating disorders.
2. Methods
2.1. Design
This systematic review was conducted by following the Preferred Reporting for Systematic Review and Meta-Analysis (PRISMA) consensus statement (CRD42020201843) [https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020201843]. Studies relevant to cases of COVID-19 with suspected or confirmed spinal cord demyelinating diseases were included.
2.2. Search strategy
In this systemic review, data were collected from PubMed, NIH Litcovid, EMBASE and Cochrane library databases, until September 10, 2020, using pre-specified searching strategies. The search strategy consisted of a variation of keywords of relevant medical subject headings (MeSH) and keywords, including “SARS-CoV-2”, “COVID-19”, “coronavirus”, “demyelinating disorders”, “multiple sclerosis” and “encephalomyelitis”. Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV), Middle East Respiratory Syndrome Coronavirus (MERS-CoV) and various other human coronaviruses were also included in our search strategy to capture related articles. We also hand-searched additional COVID-19 specific articles using the reference list of the selected studies, relevant journal websites, and renowned pre-print servers (medRxiv, bioRxiv, and pre-preints.org) from 2019 to the current date for literature inclusion. To decrease publication bias, we invigilated the references of all studies potentially missed in the electrical search. Content experts also searched the gray literature of any relevant articles.
2.3. Study selection criteria
All peer-reviewed, pre-print (not-peer-reviewed) including cohort, case-control studies, and case reports, which met the pre-specified inclusion and exclusion criteria, were included in this study.
2.4. Inclusion criteria
Studies meeting the following inclusion criteria were included: (i) Studies regarding COVID-19 positive patients with suspected or confirmed spinal cord demyelinating diseases; (ii) COVID-19 studies revealing possible association with multiple sclerosis (MS) or related neuroimmune disorders with confirmed or suspected spinal involvement; and (iii) Studies published in English. Simultaneously, parallel search was conducted to have a comparative as well as a retrospective outlook into the distribution of cases with similar neurological manifestations in previous outbreaks, i.e., Severe Acute Respiratory Syndrome (SARS-CoV), Middle East Respiratory Syndrome Coronavirus (MERS-CoV) and various other human coronaviruses.
2.5. Exclusion criteria
We excluded studies if COVID-19 had not been confirmed and those written in languages other than English. We also excluded review papers, viewpoints, commentaries, and studies where information related to neurological manifestations or spinal demyelination was not reported.
2.6. Data extraction
Before the screening process, a team of two reviewers (GS and SD) participated in the calibration and screening exercises. The first reviewer (GS) subsequently screened independently the titles and abstracts of all identified citations, meanwhile the second reviewer (SD) verified those citations and screened papers by the first reviewer (GS). Another reviewer (UG) then retrieved and screened independently the full texts of all citations deemed eligible by the reviewer (SD) and analyzed those data. The other two reviewers (RM and DL) independently verified these extracted full texts for eligibility towards analysis and designed the overall study structure. JBL resolved disagreements whenever necessary and took final decisions regarding the study. Throughout the screening and data extraction process, the reviewers used piloted forms. In addition to the relevant clinical data, the reviewers also extracted data on the following features: study characteristics (i.e., study identifier, study design, setting, and timeframe); outcomes (qualitative and/or quantitative); clinical factors (definition and measurement methods); and study limitations. The Newcastle-Ottawa scale was used to assess the selection procedure, the comparability, and the outcomes of each reviewed study.
2.7. Statistical analysis
Both qualitative and quantitative data were expressed in percentages. Unit discordance among the variables was resolved by converting the variables to a standard unit of measurement. A p value < 0.05 was considered as statistically significant, but it could not be calculated due to insufficient data. A meta-analysis was planned to analyze the association of the demographic findings, symptoms, biochemical parameters, and outcomes, but was later omitted due to lack of sufficient data.
3. Results
Around 750 articles were initially identified from the databases and 253 from different pre-prints servers. Finally, 712 articles were selected after removing the duplicates. Of these selected articles, 600 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 and few other articles were excluded for study types (e.g., review papers, correspondence, viewpoints, or commentaries). Fifty-two articles were finally selected for this study. Of these 52, 21 articles were included in the analysis (Abdelhady et al., 2020, AlKetbi et al., 2020, Chakraborty et al., 2020, Chow et al., 2020, Domingues et al., 2020, Kaur et al., 2020, Lisnic et al., Maideniuc and Memon, 2021, McCuddy et al., 2020, Munz et al., 2020, Novi et al., 2020, Otluoglu et al., 2020, Sarma and Bilello, 2020, Sotoca and Rodríguez-Álvarez, 2020, Utukuri et al., 2020, Valiuddin et al., 2020, Zachariadis et al., 2020, Zanin et al., 2020, Zhao et al., 2021, Zhou et al., 2020, Zoghi et al., 2020), and the remaining 31 were synthesized narratively (See, Fig. 1 and Table 1 ).
: flowchart showing the algorithm used to identify the studies included in this review.
Flow diagram template was adopted from PRISMA.
Table 1
Included Articles.
| Authors name | Article name | Number of cases |
|---|---|---|
| Abdelhady et al. (Abdelhady et al., 2020) (Qatar) | Acute flaccid myelitis in COVID-19 | 1 |
| AlKetbi et al. (AlKetbi et al., 2020) (UAE) | Acute myelitis as a neurological as a neurological complication of COVID-19: A case report and MRI findings | 1 |
| Chakraborty et al. (Chakraborty et al., 2020) (India) | COVID-19 associated acute transverse myelitis: rare entity | 1 |
| Chow et al. (Chow et al., 2020) (Australia) | Acute transverse myelitis in COVID-19 infection | 1 |
| Domingues et al. (Domingues et al., 2020) (Brazil) | First case of SARS-CoV-2 sequencing in CSF of a patient with suspected demyelinating disease | 1 |
| Kaur et al. (Kaur et al., 2020) (USA) | Transverse myelitis in a child with COVID-19 | 1 |
| Lisnic et al. (Lisnic et al., 02 August 2020) (Moldova) | Acute transverse myelitis in a HIV positive patient with COVID-19 | 1 |
| Maideniuc and Memon (Maideniuc and Memon, 2021) (USA) | Acute necrotizing myelitis and acute motor axonal neuropathy in COVID-19 patient | 1 |
| Mc Cuddy et al. (McCuddy et al., 2020) (USA) | Acute demyelinating encephalomyelitis (ADEM) in COVID-19 infection: A case series | 1 |
| Munz et al. (Munz et al., 2020) (Germany) | Acute transverse myelitis after COVID-19 Pneumonia | 1 |
| Novi et al. (Novi et al., 2020) (Italy) | Acute disseminated encephalomyelitis after SARS-CoV-2 infection | 1 |
| Otluoglu et al. (Otluoglu et al., 2020) (Turkey) | Encephalomyelitis associated with COVID-19 infection: case report | 1 |
| Sarma et al. (Sarma and Bilello, 2020) (USA) | A case report of acute transverse myelitis following novel Coronavirus | 1 |
| Sotoca et al. (Sotoca and Rodríguez-Álvarez, 2020) (Spain) | COVID-19 associated acute necrotizing myelitis | 1 |
| Utukuri et al. (Utukuri et al., 2020) (USA) | Possible acute disseminated encephalomyelitis related to severe acute respiratory syndrome coronavirus 2 infection | 1 |
| Valiuddin et al. (Valiuddin et al., 2020) (USA) | Acute transverse myelitis associated SARS-CoV-2: A case report | 1 |
| Zachariadis et al. (Zachariadis et al., 2020) (Switzerland) | Transverse myelitis related to COVID-19 infection | 1 |
| Zanin et al. (Zanin et al., 2020) (Italy) | SARS-CoV-2 can induce brain and spine demyelinating lesions | 1 |
| Zhao et al. (Zhao et al., 2021) (China) | Acute myelitis after SARS-CoV-2 infection: a case report | 1 |
| Zhou et al. (Zhou et al., 2020) (USA) | Myelin oligodendrocyte glycoprotein antibody- associated optic neuritis and myelitis in COVID-19 | 1 |
| Zoghi et al. (Zoghi et al., 2020) (Iran) | A case of possible atypical demyelinating event of the central nervous system following COVID-19 | 1 |
The 21 selected articles were all case reports and included 11 (52%) men and 10 (48%) women. The mean age was of 46.7 ± 18.0. Fever (N = 11, 52%), cough (N = 6, 29%), myalgia (N = 6, 29%), vomiting (N = 2, 10%) and nausea (N = 1, 5%) were among the most frequent COVID-19 related features (See Table 2 ). The neurological manifestations included weakness (N = 14, 66.7%), sensory deficit (N = 14, 66.7%) autonomic dysfunction (N = 8, 3.1%), and ataxia (N = 1, 4.8%) (See Table 2).
Table 2
Demographic and clinical features.
| Authors | Sex, age (yrs) | Comorbidities | Neurological signs/symptoms | CSF characteristics | Clinical diagnosis | Clinical features of COVID-19 | Latency (days) | Treatment | Outcome | |
|---|---|---|---|---|---|---|---|---|---|---|
| SARS-Cov2 detection | Abnormal parameters | |||||||||
| Abdelhady et al. (Abdelhady et al., 2020) | M, 52 | Type 2 diabetes mellitus and G6PD deficiency | Lower abdominal pain, inability to pass urine, and flaccid paralysis | Negative | Increased WBC and proteins | Acute flaccid myelitis | Fever | Not mentioned | Steroids and acyclovir | Death |
| AlKetbi et al. (AlKetbi et al., 2020) | M, 32 | Not mentioned | Flaccid paralysis | Not mentioned | Not mentioned | Acute myelitis | Fever with flu-like symptoms | 1 | Methylprednisolone, acyclovir, and andenoxaparin | In treatment |
| Chakraborty et al. (Chakraborty et al., 2020) | F, 59 | Obesity | Acute, progressive, ascending flaccid paraplegia, urinary retention, and constipation. No sensation below Th10 level | Negative | Increased protein and adenosine deaminase | Acute transverse myelitis | Fever | Not mentioned | Methylprednisolone, antipyretics, and supportive care | Death |
| Chow et al. (Chow et al., 2020) | M, 60 | Arterial hypertension and hypercholesterolemia | Paraparesis with constipation, urinary retention, hyperreflexia, reduced proprioception of lower limbs, and patchy paresthesia to the level of umbilicus | Negative | Increased WBC and proteins | Acute transverse myelitis | Fever, dysgeusia, anosmia and cough | 2 | Methylprednisolone | Discharged |
| Domingues et al. (Domingues et al., 2020) | F, 42 | Similar clinical picture 3 years ago with spontaneous full recovery. | Paresthesia of upper left limb progressing to left hemithorax and hemiface | Positive | Normal | Clinically isolated syndrome | Mild respiratory symptoms | Not mentioned | Not mentioned | Discharged with full recovery |
| Kaur et al. (Kaur et al., 2020) | F, 3 | Not mentioned | Flaccid quadriparesis and neurogenic respiratory failure requiring intubation. Complete quadriplegia after 12 h. | Negative | Increased RBC, WBC mainly neutrophilic and protein | LETM | None | 21 | Methylprednisolone, IVIG, plasma exchange and rituximab | Discharged |
| Lisnic et al. (Lisnic et al.,) | M, 27 | HIV infection for 1 year, on anti-retroviral therapy | Spastic tetraparesis,Th7 superficial and C7 deep sensory level disturbance | Negative | Normal | Acute transverse myelitis | Mild fever | Not mentioned | Methylprednisolone and plasma exchange | Still in treatment |
| Maideniuc and Memon (Maideniuc and Memon, 2021) | F, 61 | Arterial Hypertension, hyperlipidemia, hypothyroidism, and history of nasopharyngeal and uterine cancer | Tingling sensation in fingers and toes. Lost sensation from chest down and progressive spastic paraparesis. Bowel and bladder retention. Sensory level at C3. | Negative | Increased RBC, protein, and glucose | Acute necrotizing myelitis with acute motor axonal neuropathy | Runny nose and chills | 4 | Methylprednisolone and plasma exchange | Discharged with rehabilitation facility |
| Mc Cuddy et al. (McCuddy et al., 2020) | F,40(patient 1) | Type 2 diabetes mellitus, arterial hypertension, obesity and 30 weeks pregnant | Paraplegia and significant, symmetric weakness in the upper extremity. | Negative | Increased protein and glucose | ADEM | Cough, chest pain, fever and shortness of breath | Not mentioned | Dexamethasone, HCQ, Zinc, and convalescent plasma therapy | Improving. |
| Munz et al. (Munz et al., 2020) | M, 60 | Arterial hypertension, mild fatty liver and ureterolithiasis | Bladder dysfunction, progressive spastic paraparesis, and hypesthesia below Th9 level | Negative | Abnormal lymphocytic pleocytosis and increased protein | Acute transverse myelitis | Respiratory features | 2 | Acyclovir, ceftriaxone, and methylprednisolone | Dischargedwith steroid taper scheme |
| Novi et al. (Novi et al., 2020) | F, 64 | Vitiligo, arterial hypertension, and monoclonal gammopathy | Irritability, headache, bilateral pupillary defect, visual loss, right abdominal sensory level deficit, and left sided lower limb hyperreflexia with Babinski sign | Positive | Lymphocytic pleocytosis, increased protein and positive OCB | Suspected ADEM | Flu-like symptoms, anosmia and ageusia | 25 | Methylprednisolone, oral prednisolone, and IVIG | Discharged with oral prednisolone tapering |
| Otluoglu et al. (Otluoglu et al., 2020) | M, 48 | None | Progressive headache. | Positive | Increased glucose | Viral encephalitis and myelitis | Persistent cough, fatigue, myalgia, and anosmia | Not mentioned | HCQ, favipiravir, acyclovir, methylprednisolone, levetiracetam, piperacillin and tazobactam | Still under treatment with stable condition |
| Sarma et al. (Sarma and Bilello, 2020) | F, 28 | Hypothyroidism | Persisting lumbosacral back pain without radiation. Paresthesia in lower limbs with total loss of sensation upwards until tip of tongue, urinary retention, decreased sensation below Th5 level and Lhermitte’s sign | Not mentioned | Increased WBC and protein | Transverse myelitis | Cough, low grade fever, low back pain, myalgia, rhinorrhea, nausea and vomiting | Not mentioned | Prednisolone and plasma exchange | Discharged with steroid tapering |
| Sotoca et al. (Sotoca and Rodríguez-Álvarez, 2020) | F, 69 | Not mentioned | Irradiated cervical pain, imbalance, motor weakness, numbness in left hand, right facial and left hand hypesthesia, left hand weakness and hyperreflexia | Negative | Lymphocytic pleocytosis | Acute necrotizing myelitis | Fever and dry cough | 1 | Methylprednisolone and plasma exchange | Discharged with oral prednisolone |
| Utukuri et al. (Utukuri et al., 2020) | M, 44 | Not mentioned | Paraparesis and urinary retention | Positive | Increased WBC, mainly lymphocytes | ADEM | Lethargy | Not mentioned | Methylprednisolone and IVIG | Dischargedwith rehabilitation facility |
| Valiuddin et al. (Valiuddin et al., 2020) | F, 61 | Not mentioned | Progressing paraparesis with bilateral extensor plantar responses, constipation, difficulty in voiding and upper limb weakness | Negative | Increased protein | Acute transverse myelitis | Generalized weakness, rhinorrhea and chills | 4 | Methylprednisolone and plasma exchange | Still in treatment with rehabilitation facility |
| Zachariadis et al. (Zachariadis et al., 2020) | M, 63 | Obesity | Paraplegia with total anesthesia below Th10 and sphincter dysfunction | Negative | Increased WBC and protein | Transverse myelitis | Rhinorrhea, odynophagia, myalgia and fever | 5 | IVIG and corticosteroids | Transferred to neurorehabilitation center |
| Zanin et al. (Zanin et al., 2020) | F, 54 | Anterior communicating artery aneurysm 20 years ago, treated surgically | Found unconscious at home | Negative | Normal | Spinal demyelinating lesion | Anosmia and ageusia. | Not mentioned | Dexamethasone, antiepileptic, antiretroviral drugs, and HCQ | Discharged but with rehabilitation facility |
| Zhao et al. (Zhao et al., 2021) | M, 66 | Not mentioned | Paraparesis with urinary and bowel incontinence and sensory level at Th10. Decreased tendon reflexes of both lower limbs | Not done | Not done | Acute myelitis | Fever and fatigue | 6 | Lopinavir/Ritonavir, ganciclovir, dexamethasone, moxifloxacin, meropenem, glutathione and mecobalamin | Transferred to rehabilitation center |
| Zhou et al. (Zhou et al., 2020) | M,26 | Not mentioned | Numbness in lower limbs, neck discomfort, and sequential vision loss | Negative | Increased WBC | MOG-Ab associated optic neuritis with myelitis | Dry cough | Not mentioned | Methylprednisolone followed by prednisolone taper | Complete resolution |
| Zoghi et al. (Zoghi et al., 2020) | M, 21 | Not mentioned | Tetraparesis with absent Babinski’s sign and impaired sensation below Th8 level | IgG positive | Increased protein and WBC | LETM | Fever, cough, sore throat, loss of appetite, vomiting, malaise | 17 | Plasma exchange, vancomycin, meropenem and acyclovir | Discharged but with lower limbs paresis |
ADEM = Acute disseminated encephalomyelitis; CSF = Cerebrospinal fluid; F = Female; G6PD = Glucose-6-phosphate dehydrogenase; HCQ = Hydroxychloroquine; HIV = Human immune deficiency virus; IgG = Immunoglobulin; IVIG =Intravenous immunoglobulin; LETM = Long extensive transverse myelitis; M = Male; MOG-Ab = Myelin oligodendrocyte glycoprotein antibody; OCB = Oligoclonal bands; RBC = Red Blood Cells. WBC = White Blood Cells.
Biochemical analysis of cerebrospinal fluid (CSF) revealed that glucose was within the normal range in most of the cases (N = 12, 57.1%), whereas one case was found with elevated glucose, another one with glucose below the normal range and the value of the remaining seven cases were not mentioned (See Table 2). The mean range of glucose (mg/dl) in CSF was 64.5 ± 16.4. In only three cases, both blood and CSF glucose were mentioned (blood glucose = 135 mg/dl, CSF glucose = 73 mg/dl, ratio = 0.5 (Valiuddin et al., 2020); and blood glucose = 110 mg/dl, CSF glucose = 34 mg/dl, ratio = 0.30 (Zhao et al., 2021)). In the report by Otluoglu et al. (Otluoglu et al., 2020), the CSF glucose was 5 mmol/L (90 mg/dl), while the corresponding value of blood glucose was 105 mmol/L (1890 mg/dl), which seems to be erroneous. Apart from this, CSF protein was found to be increased in most of the cases (N = 12, 57.1%), whereas in six cases (N = 6, 28.6%), it was below the normal range and the value was not mentioned in the remaining three cases (See Table 2). Median value of CSF protein (mg/dl) was 59 (IQR 36–79.1). Certain cell count parameters, such as white blood count (N = 10, 47.6%) and red blood count (N = 2, 10%) were found to be elevated (See Table 2). In addition, lymphocytic pleocytosis (N = 4, 19.0%) was also reported (See Table 2). SARS-CoV-2 was detected in the CSF in five (24%) patients, by means of qualitative real-time reverse-transcriptase–polymerase-chain-reaction assay (RT-PCR) in four and IgG positivity in one (See Table 2). SARS-CoV-2 detection was negative in 13 (62%) patients, whereas SARS-CoV-2 testing was not performed in two cases and data was unavailable in one case (See Table 2). Autoimmune profiling was negative in most cases, except two cases with positive oligoclonal bands and lupus antigen, respectively (See Tables 2 and and33 ).
Table 3
Neuroimaging features and autoimmune profiling.
| Authors | Spinal Involvement | Level of involvement | Brain involvement | Autoimmune profiling | ||
|---|---|---|---|---|---|---|
| LETM | Patchy | Short | ||||
| Abdelhady et al. (Abdelhady et al., 2020) | + | – | – | Hyperintense signal in ventral horns of gray matter in upper and mid thoracic spinal cord on T2WI | Normal | ANCA and ANA were negative |
| AlKetbi et al. (AlKetbi et al., 2020) | – | + | – | Extensive diffuse hyperintense signal involving predominantly the gray matter of the cervical, dorsal, and lumbar regions of the spinal cordon T2WI | Not done | Anti-LA, ANCA, RF, anti-cardiolipin, and anti –beta gp were negative |
| Chakraborty et al. (Chakraborty et al., 2020) | – | – | + | Dorsal spine hyperintensity at T6 – T7 on T2WI | Not done | Not done |
| Chow et al. (Chow et al., 2020) | + | – | – | T2 signal increased centrally in spinal cord from T7-T10 | Normal | Anti-myelin associated gp IgM, anti-MOG and anti-NMO IgG were negative |
| Domingues et al. (Domingues et al., 2020) | – | – | + | Hyperintense signal in cervical spine on T2WI and STIR, indicating small left lateral ventral lesion with mass effect (0.4 cm) (with Gd contrast) | Normal | Anti SSA and anti SSB were negative |
| Kaur et al. (Kaur et al., 2020) | + | – | – | Swelling of cervical spinal cord, involving most of the transverse aspect of spinal cord, from lower medulla to mid thoracic level | Normal | Negative rheumatoid assessment. Anti-AQP-4 and anti-MOG were negative |
| Lisnic et al. (Lisnic et al.,) | + | – | – | Extensive C4-T5 lesion in posterior column on T2WI and right lateral column with Gd contrast | Normal | ANA, ANCA, anti-AQP-4 and anti-MOG were negative |
| Maideniuc and Memon (Maideniuc and Memon, 2021) | + | – | – | Hyperintense signal at C3—C4 level on T1WI | Normal | Lupus antibody was positive. Anti-MOG and anti-AQP-4 were negative |
| McCuddy et al. (McCuddy et al., 2020) | – | – | – | Normal | Multiple T2WI hyperintense lesions with restricted diffusion, involving corpus callosum, bilateral cerebral white matter, right pons and in the bilateral ventral medulla | Autoimmune profiling of cerebrospinal fluid was negative |
| Munz et al. (Munz et al., 2020) | – | – | + | Patchy hyperintense signal at T9-T10 and T3-T5 | Normal | Anti-neuronal, anti-MOG and anti-AQP4 were negative |
| Novi et al. (Novi et al., 2020) | – | – | + | Hyperintense spindle like single lesion al T8 level on T2WI | Multiple T1WI post Gd enhancing lesions in brain. Follow-up MRI showed reduction in number of lesions | Anti-AQ4 and anti-MOG were negative |
| Otluoglu et al. (Otluoglu et al., 2020) | – | – | + | Confined hyperintense lesion at upper cervical spine | Hyperintense lesions both in the posterior medial cortical surface of the temporal lobe | Not done |
| Sarma et al. (Sarma and Bilello, 2020) | + | – | – | Elongated signal changes involving medulla and throughout the spinal cord to conus medullaris | Not done | ANA was negative |
| Sotoca et al. (Sotoca and Rodríguez-Álvarez, 2020) | + | – | – | Hyperintense signal lesion extending from medulla oblongata to C7 | Normal | Neuronal surface antibody was ruled out |
| Utukuri et al. (Utukuri et al., 2020) | + | + | – | Hyperintense signal lesions throughout cervical and thoracic spinal cord and conus medullaris | Hyperintense lesions within left posterior parietal lobe and periventricular region on FLAIR | Cardiolipin antibody immunoglobulin M was mildly elevated |
| Valiuddin et al. (Valiuddin et al., 2020) | – | + | – | Extensive ill-defined patchy hyperintense signal throughout the central aspect of spinal cord on STIR | Not done | Anti-GFAP, mGLUR, NMDA-R and anti-MOG were negative |
| Zachariadis et al.[24] | – | – | – | Normal | Normal | ANCA, ANA, anti-MOG, anti-SSB, anti-SSA, RF, GFAP, and Beta 2 glycoprotein 1 were negative |
| Zanin et al. (Zanin et al., 2020) | + | + | – | Hyperintense intra-medullary signal lesions on T2WI at bulb-medullary junction, and from C3 to Th6 | T2W1 hyperintense lesions in periventricular white matter without restriction of diffusion nor contrast enhancement | Not done |
| Zhou et al. (Zhao et al., 2021) | + | – | – | Spinal signal enhancement at lower cervical and upper thoracic segment with mild central cord thickening | Bilateral optic nerve thickening up to the perichiasmal segments on T1WI | AQP4 was negative. Anti-MOG IgG was elevated |
| Zoghi et al. (Zhou et al., 2020) | + | – | – | LETM of upper cervical with intramedullary lesion | Corticospinal tract lesions in internal capsule extending to cerebral peduncles and pons on T2-FLAIR. Heterogeneous marble patterned hyperintensity in splenium of corpus callosum without diffusion weighted restrictions or contrast enhancement. | Anti-NMDAR, anti-MOG, anti-AQP4, anti-phospholipid, HLA B5, and ACE were negative |
ACE = angiotensin converting enzyme; ANA = antinuclear antibody; ANCA = antineutrophil cytoplasmic antibodies; Anti-SSA = anti-Sjögren’s syndrome-related antigen A; Anti-SSB = anti-Sjögren’s syndrome-related antigen B; AQP-4 = aquaporin-4; FLAIR = fluid-attenuated inversion recovery; Gd = gadolinium; GFAP = glial fibrillary acidic protein; HLA = human leukocyte antigen; LA = lupus antigen; m-GLUR = metabotropic glutamate receptors; LETM = long extensive transverse myelitis; NMDA-R = N-methyl-D-aspartate receptor; NMO = neuromyelitis optica; RF = rheumatoid factor; STIR = short T1 inversion recovery; T2WI = T2-weighted image.
Neuroimaging findings were reported in all except one patient. Among the neuroimaging modalities, MRI of the spine was done in 20 patients, whereas MRI of the brain in 16. Computed tomography (CT) scan of the spine and brain were done in one and four patients, respectively. We found that nearly half of the cases (N = 9, 45%) were associated with isolated long extensive transverse myelitis (LETM), whereas a combination of both LETM and patchy involvement was found in two (See Table 3). Only five patients had isolated short segment involvement and two patchy involvement (See Table 3). Furthermore, concomitant demyelination of both brain (including the brainstem) and spine was reported in six patients (Table 3).
Treatment was mainly carried out with corticosteroids (N = 18, 86%), including intravenous methylprednisolone (N = 14, 67%), intravenous dexamethasone (N = 2, 10%), and oral prednisolone (N = 1, 5%). Intravenous immunoglobulin was administered to five (24%) patients, meanwhile plasma exchange sessions to eight (38%). Antiviral drugs, such as acyclovir, lopinavir/ritonavir, ganciclovir and favipiravir, were administered to seven patients (33.3%), being acyclovir the most common drug of choice (N = 5, 24%). Hydroxychloroquine was administered to three (14.3%) patients. Two patients also received anti-epileptic drugs. Most of the patients recovered although they were still under treatment at the time of discharge. Specifically, approximately one-third of the discharged patients were transferred to a rehabilitation center (N = 6, 28.6%) for further improvement. Only two (10%) patients died during the period of study.
4. Discussion
During this ongoing pandemic, clinicians and researchers worldwide have observed several immunological manifestations of SARS-CoV-2, affecting different organ systems. In the case of neurological manifestations of COVID-19, immunological mechanisms are known to play an important role in giving rise to diverse clinical presentations affecting both CNS and PNS (Roy et al., 2021). Contrary to how SARS-CoV-2 affects PNS (mostly in the form of GBS spectrum) (Abu-Rumeileh et al., 2020, Ghosh et al., 2020, Zhao et al., 2020, Finsterer et al., 2020, Gutiérrez-Ortiz et al., 2020), data regarding spinal cord involvement are scarce.
Spinal cord demyelination in COVID-19 might be an under-recognized neurological complication. Thus, in the present systematic review, we have only found 21 patients with SARS-CoV-2 mediated spinal cord complications. Eight (38.1%) of these patients had para-infectious acute myelitis and three (14.3%) post-infectious acute myelitis. Weakness, sensory deficit, autonomic dysfunction, including sphincter dysfunction and ataxia were the most frequent neurological manifestations. Most cases showed elevated cerebrospinal fluid protein as well as lymphocytic pleocytosis. With respect to neuroimaging data, almost half of the patients presented with LETM, followed by short segment and patchy involvement. LETM is usually observed in neuromyelitis optica spectrum disorder, infectious myelitis, lupus-related demyelination, and, occasionally, in MS. In India, tuberculosis is a known cause of LETM (Sahu et al., 2014). From a clinician’s perspective, the above observations extend the differential diagnosis of LETM to include COVID-19 related spinal cord demyelination. Concerning the prognosis, most of the patients could be discharged and the mortality rate was low (<10%).
Pathogenetic mechanisms for the development of spinal cord demyelinating diseases following SARS-CoV-2 infection may be either mediated by direct neurotropism or by aberrant immune mediated injury, the latter being the most likely.
There are several pieces of evidence to support an aberrant immune mediated injury. SARS-CoV-2 infection is known to cause a cytokine storm (Mehta et al., 2020). The pro-inflammatory state, induced by cytokine storm, mainly sustained by IL-6, IL-1,and TNF-alpha, may be the responsible for the activation of glial cells, which may also trigger the onset of demyelination (Schett et al., 2020). In line with this, 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 CNS (Khateb et al., 2020). Furthermore, the S gene of the coronavirus mouse hepatitis virus is related to certain molecular aspects of demyelination, which indicates the potential role of viral envelope S glycoproteins in immune mediated demyelination (Sarma et al., 2000). Kim and Perlman (2005) demonstrated that an experimental strain of coronavirus JHM, even in the absence of T and B cells, developed an autoimmune demyelinating disorder in a mouse model, which suggests that the formation of anti-JHM antibodies is sufficient to cause the demyelination. In addition, last but not the least, dramatic response to intravenous immunoglobulin or plasma exchange sessions in several cases of SARS-CoV-2 mediated spinal cord demyelinating diseases points towards an underlying immune driven process.
Direct neurotropism as a pathogenic process of demyelination did not lie far behind. The presence of HCoV RNA in the CNS of patients with demyelinating disorders, suggests a possible association between coronaviruses and demyelination (Arbour et al., 2000). In a previous study, nucleic acids of HCoV-229E have been detected in the CNS tissue in four out of eleven MS patients, which suggests neurotropism of this species of coronavirus (Stewart et al., 1992). Moreover, coronavirus-like particles have been found from autopsies of two MS patients (Burks et al., 1980). Using in situ hybridization technique with cloned coronavirus cDNA probes, Murray et al. (Murray et al., 1992) detected coronavirus RNA in the demyelinating plaques in 12 of their 22 MS patients. In addition, significant amounts of coronavirus antigen and RNA were observed in active demyelinating plaques from two patients with rapidly progressive MS (Murray et al., 1992). Further, a 15-year-old boy, who presented with acute demyelinating encephalomyelitis (ADEM), tested positive for HCoV-OC43 in CSF using RT-PCR (Yeh et al., 2004). In a series of four patients affected with neurological complications due to MERS-CoV, one of them met some criteria for diagnosis of the ADEM (Arabi et al., 2015). Interestingly, in four of the cases included in our systematic review, SARS-CoV-2 RNA could be detected in CSF.
There are some implicit limitations in the present review. Given the notable asymmetry between the total number of affected cases and reported cases of SARS-CoV-2 mediated spinal cord demyelinating diseases, it can be assumed that cases are currently under-reported. The current 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 a 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 SARS-CoV-2 mediated spinal cord demyelinating diseases.