Magnetic Resonance Imaging-negative Acute Inflammatory Myelopathy following Severe Acute Respiratory Syndrome Coronavirus 2 Infection

Abstract

A 55-year-old woman was admitted to our hospital because of gait disturbance and urinary retention that acutely emerged 1 week after severe acute respiratory syndrome coronavirus 2 infection. Acute inflammatory myelopathy was clinically suspected, based on bilateral lower-limb weakness with an extensor plantar response and an elevated immunoglobulin G level in the cerebrospinal fluid. Whole-spine magnetic resonance imaging findings were normal. The central conduction time was extended, based on somatosensory evoked potentials. Her lower-limb weakness was partially ameliorated with immunosuppressive therapy. Postinfectious myelopathy is a rare neurological complication of coronavirus disease 2019 and can develop with normal radiological findings.

Introduction

Coronavirus disease 2019 (COVID-19) is caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and is primarily accompanied by respiratory symptoms. However, many studies have shown that COVID-19 can also affect multiple organ systems of the body, including the nervous system. In the early days of the COVID-19 pandemic in 2019, Mao et al. (1) reported neurological complications in one-third of hospitalized patients in Wuhan, China. Several large-scale studies subsequently revealed a variety of neurologic complications of COVID-19, such as stroke, encephalopathy, headache, acute disseminated encephalomyelitis (ADEM), anosmia, cranial neuropathy, Guillain-Barré syndrome, and skeletal muscle injury (2-6). However, among the various neurological complications, inflammatory disease of the central nervous system (CNS) is relatively infrequently reported (6).

Acute inflammatory myelopathy (AIM) is an inflammatory disease of the CNS. AIM is a heterogeneous group of spinal cord disorders, which include multiple sclerosis, neuromyelitis optica spectrum disorders, ADEM, and acute transverse myelitis (ATM) (7). It is generally diagnosed, based on the total findings of neurological symptoms, cerebrospinal fluid (CSF) analyses, and radiological evaluations. Magnetic resonance imaging (MRI) usually reveals abnormal spinal cord lesions in patients with AIM. However, initially negative findings on MRI of the spine can occur, despite coexisting symptoms and/or signs of myelopathy (8-10). AIM can develop in association with COVID-19 infection, and longitudinally extended spinal cord lesions typically occur in MRI in patients with AIM-associated COVID-19 (11,12).

We herein report a case of MRI-negative AIM after COVID-19 infection.

Case Report

A 55-year-old Japanese woman without any remarkable medical history was admitted to our hospital for gait disturbance, urinary retention, and constipation. Ten days before admission, she had developed a slight fever and upper respiratory tract symptoms. She had already received four COVID-19 vaccinations. However, she was diagnosed with COVID-19 based on positive results on reverse transcription polymerase chain reaction (RT-PCR) of a nasopharyngeal specimen. Her respiratory symptoms gradually improved over the next few days without antiviral medication. However, her fever relapsed three days before admission, followed by nausea, headache, and gait disturbance the next day. The day before admission, she was unable to walk by herself and had difficulty urinating. She also had constipation for several days.

On admission, she was afebrile and had normal vital signs. A physical examination showed no abnormalities, except for urinary retention, for which she needed catheterization. On a neurological examination, she was fully conscious, and her cranial nerves and upper extremities were not disturbed. She had significant muscular weakness in the lower limbs [Medical Research Council score: upper limbs: 5/5; lower limbs (proximal): 3/5 to 4/5; lower limbs (distal): 4/5 to 5/5], particularly the iliopsoas and hamstrings. She also had a decreased pinprick sensation, decreased deep sensations in the lower limbs (i.e. below the Th12 level), and a positive Romberg sign. Tendon reflexes were brisk only in the lower limbs, and plantar responses were bilaterally extensor. Muscular tonus was initially flaccid in the lower limbs but gradually became spastic after admission. Automatism was frequently observed in the lower limbs. Urinary retention and constipation occurred as autonomic dysfunctions.

Hematologic findings were unremarkable, including the levels of vitamins, folate, angiotensin-converting enzyme (ACE), and autoantibodies associated with various connective tissue disorders, such as antinuclear antibodies, Sjögren syndrome-related antigens A and B, double-stranded deoxyribonucleic acid (DNA), and neutrophil cytoplasmic antibodies. Autoantibodies against aquaporin-4 and myelin oligodendrocyte glycoprotein, which are associated with demyelinating CNS disorders, were also negative. Blood coagulation data were normal, including D-dimer levels, prothrombin time, and activated partial thromboplastin time. There were no signs suggesting the presence of cancer, such as anemia or an elevated inflammatory reaction. Workup for the detection of bacterial and viral infection, including syphilis, hepatitis B and C, human T-cell leukemia virus type 1, and human immunodeficiency virus revealed no abnormalities, except for COVID-19, based on the positive RT-PCR results for a nasopharyngeal specimen. A CSF analysis revealed mild mononuclear cell pleocytosis (7 /μL), an elevated immunoglobulin G (IgG) index (0.78, normal range: ≤0.73), and normal protein and glucose levels. Oligoclonal IgG bands were positive. Herpes viral DNA was not detected. RT-PCR of SARS-CoV-2 was not examined in the CSF. MRI of the brain (1.5 T, plain) and whole spine (1.5 T, contrast-enhanced) revealed no significant abnormalities (Fig. 1).

Figure 1.

T2-weighted MRI findings on admission. Axial section of the brain (a) and sagittal sections of the cervical cord (b) and thoracic cord (c). MRI: magnetic resonance imaging

In a neurophysiological analysis, a nerve conduction study (NCS) showed normal findings. However, an extended central conduction time (CCT) was observed in a lower extremity somatosensory evoked potentials (SSEPs) study (Fig. 2) performed as described elsewhere (13). Thus, the patient was diagnosed with the MRI-negative ATM type of AIM associated with COVID-19 infection, based on the abnormal findings in clinical presentation, laboratory data, CSF, and SSEPs.

Figure 2.

The components of the tibial nerve SSEP, examined as reported elsewhere (13). Electroencephalogram disc electrodes are placed at the ipsilateral popliteal fossa (PFi), ipsilateral medial epicondyle (K), contralateral iliac crest (ICc), ipsilateral greater trochanter (GTi), spinous processes of the first lumbar vertebrae (L1S), spinous process of the second cervical vertebrae (C2S), 2 cm before Cz (i.e., Cz´) and the central electrode in the International 10-20 system contralateral to the stimulation (i.e., Cc). The central conduction time is significantly extended, measured as the interval between the peak of the N21 potential (N21) in the L1S-ICc lead and the onset or peak of the P38 potential (P38o and P38, respectively) in the Cz´-Cc lead. SSEP: somatosensory evoked potential

The patient was treated with 2 cycles of intravenous methylprednisolone pulse (IVMP) therapy (1 g/day from hospital days 5-7 and 12-16), followed by the oral administration of prednisolone (35 mg/day) (Fig. 3). Sensory disturbance of the lower limbs, urinary retention, and constipation ameliorated considerably within three weeks after the initiation of steroids. However, the improvement of the lower-limb weakness was insufficient. On hospital day 21, a re-examination of MRI findings for the cervical and thoracic spine (1.5 T, plain) remained unremarkable. In a neurophysiological analysis, extended CCT also persisted in an SSEP study.

Figure 3.

The clinical course. Immunotherapy was effective for sensory disturbance and bladder bowel dysfunction, which were considerably ameliorated after two cycles of IVMP. Lower-limb weakness was insufficiently improved, even after additional treatment with IVIg. IVMP: intravenous methylprednisolone pulse, PSL: prednisolone, IVIg: intravenous immunoglobulin

The patient was still unable to walk on her own and under gait training with a walker. Therefore, intravenous immunoglobulin therapy (IVIg; 400 mg/kg/day from hospital days 25-29) was introduced, and oral prednisolone was tapered to 30 mg/day after completion of IVIg therapy. Two weeks after IVIg therapy, no abnormalities were identified on a CSF analysis, including the IgG index. However, lower-limb weakness and gait difficulty persisted. On hospital day 39, the patient was transferred to a rehabilitation hospital to continue gait training. Unfortunately, even after two months of gait rehabilitation, she remained wheelchair-bound.

Discussion

We encountered a 55-year-old woman who developed AIM after COVID-19 infection. Among various etiologies of spinal cord disorders included in AIM, ATM associated with COVID-19 infection was suspected because the patient had no brain lesions and because screening for comorbid connective tissue disorders and infectious diseases was negative, except for the recent COVID-19 infection. Clinical symptoms including spastic paraplegia, decreased sensation of the lower limbs and bladder bowel disturbance, rapidly progressed and were compatible with a diagnosis of ATM. However, MRI findings in the brain and spine were unremarkable. A CSF analysis and SSEP study were useful for the detection of biological and physiological abnormalities of the CNS, respectively.

With regard to neurological manifestations associated with COVID-19, several large-scale cohorts from different countries (1-6) and a wide variety of neurological syndromes have been reported. Among these, the most common diseases were stroke, encephalopathy, and headache (1-3,6). Compared to these diseases, AIM associated with COVID-19 seems to occur less frequently, as it has primarily been described in case reports (11,12). However, the accurate rate of occurrence of AIM remains unknown.

In 33 patients with COVID-19-associated myelopathy with MRI evidence of abnormalities in the spinal cord (i.e. MRI-positive COVID-19-associated myelopathy) (11), the ratio of men to women was nearly 1:1, the mean age was 47 years old, and the common neurological manifestations included lower-limb weakness, sensory deficit, and bowel and bladder dysfunction. Most patients presented with viral infection symptoms, such as cough, a fever, and myalgia. The mean period between the onset of COVID-19 infection until the first symptom of myelopathy ranged from two days to three weeks. Serological autoimmune autoantibody findings were usually negative. In a CSF analysis, lymphocytic pleocytosis and an elevated protein level were generally observed. On neuroimaging, longitudinally extensive transverse myelitis (LETM; extending over three or more vertebral segments) was predominantly observed on MRI of the spine and involved the entire spinal cord in some patients. Concomitant brain lesions were found in approximately 20% of patients. The therapeutic effect of immunosuppressive therapy varied; however, about half of the patients showed moderate to significant improvement in the lower motor limb function. The characteristics of patients with MRI-positive COVID-19-associated myelopathy are shown in Table 1.

Table 1.

Characteristics of Patients with MRI-positive COVID-19-associated Myelopathy.

Item	Patients with MRI-positive myelopathy	Our case
Mean age (y, ±standard deviation)	47±17.7	55
Male-to-female ratio	Close to 1:1	Female
Latency period between viral infection and onset of myelopathy	2 days to 3 weeks	1 week
MRI findings for the spine	LETM mainly involving cervical and/or thoracic region	Negative
Concomitant brain lesion	Positive in approximately 20% of patients	Negative
Treatment	IVMP alone or in combination with PLEX or IVIg	IVMP+IVIG
Therapeutic response	Moderate to significant improvement in lower limb motor function in about half of the patients	Partial

IVIg: intravenous immunoglobulin, IVMP: intravenous methylprednisolone pulse, LETM: longitudinally extensive transverse myelitis, MRI: magnetic resonance imaging, PLEX: plasma exchange

Patients with MRI-positive COVID-19-associated myelopathy typically have LETM (11,12,14,15); however, a small number of cases with suspected myelopathy show no abnormal findings on MRI of the spine (16-20), as in our patient. To our knowledge, there have been nine reported cases of MRI-negative COVID-19-associated myelopathy, including our patient (16-20), as summarized in Table 2. Concomitant brain lesions were not observed in these patients. The ratio of men to women was 4:5. All patients were ≥50 years old and developed paraplegia alone or with sensory deficit and/or bowel bladder dysfunction. Viral infection symptoms of COVID-19 ranged from none to mild, except for one patient who had pneumonia requiring hospitalization and oxygen administration via a nasal cannula. The latency period of the AIM onset from COVID-19 infection was within three weeks in four patients, two months or longer in four patients, and unknown in one patient. A CSF analysis showed mild pleocytosis and/or an elevated protein level, and a NCS and/or electromyogram revealed normal or denervation in the lower limbs in all examined patients. In addition to the CSF and electrophysiological findings, our patient had positive oligoclonal IgG bands, an elevated IgG index in CSF, and extended CCT on SSEP, which further supported the diagnosis of ATM. All patients were treated with IVMP in combination with IVIg or PLEX, and their response to treatment was none or partial, as in our patient. An older age at the onset, an occasional longer latency period between COVID-19 infection and the AIM onset, infrequent concomitant brain lesions, and a poor therapeutic response seem to be the characteristics of MRI-negative patients, compared to patients with MRI-positive COVID-19-associated myelopathy.

Table 2.

Clinical Presentation of Patients with MRI-negative COVID-19-associated Myelopathy.

Reference No.	16	17	18, 19	20	-
Number of patients	1	5	1	1	1 (our case)
Age (y), sex	63, M	58-69, M3/F2	60, F	54, F	55, F
Latency period between viral infection and onset of myelopathy	2 weeks	3 weeks (1 patient); 2-6 months (3 patients); Not described (1 patient)	2 months	2 weeks	1 week
Neurological manifestations	Paraplegia; Sensory deficit below Th10 level	In lower limbs: Weakness (5/5 patients); Muscle cramp (2/5 patients); Sensory deficit (4/5 patients); BBD (1/5 patients)	Paraplegia; BBD	Paraplegia; Sensory deficit below Th9 level; Urinary retention	Paraplegia; Sensory deficit below Th12 level; BBD
Concomitant brain lesion	Negative	Negative	Negative	Negative	Negative
Treatment	IVMP+IVIg	IVMP (2 patients); IVMP+IVIg (1 patient); Not described (2 patients)	IVMP+PLEX	IVMP+IVIg	IVMP+IVIg
Therapeutic response	Partial	None (1 patient); Partial (2 patients); Not described (2 patients)	Partial	Partial	Partial

BBD: bladder bowel dysfunction, CSF: cerebrospinal fluid, F: female, IVIg: intravenous immunoglobulin, IVMP: intravenous methylprednisolone pulse, M: male, OCB: oligoclonal IgG band, PLEX: plasma exchange, Th: thoracic spine, VEP: visual evoked potential

The exact mechanism underlying the development of COVID-19-associated myelopathy is unclear, but it has been suggested that SARS-CoV-2 can damage the spinal cord through ACE-2 receptors or through either cytokine storm or a post-infectious inflammatory or immune-mediated mechanism (11). It is also unclear why some patients with COVID-19-associated myelopathy have negative findings on MRI. However, this may be due to the timing of imaging (i.e. late imaging may not reveal a transient lesion, or early imaging may not reveal an evolving lesion), insensitivity of MRI for detection of COVID-19-associated inflammation, or a functional disturbance mediated by binding of a certain intrathecal IgG-induced via COVID-19 infection (21). Repeated examinations and the careful evaluation of sagittal and axial images on MRI may aid in the detection of radiological abnormalities. One limitation in our patient is that the time interval between initial and follow-up MRI was only three weeks, so MRI abnormalities, such as atrophy of the spinal cord, might have been detectable on a re-examination several months later, had one been conducted.

Abrams et al. (17) speculated that small-vessel strokes of the spinal cord, based on endothelial damage and coagulopathies due to COVID-19 infection, may underlie the lack of apparent radiologic findings and lead to a poor response to immunotherapeutic agents. The limited response to multiple immunotherapies in our patient is consistent with this hypothesis of underlying strokes. Hypercoagulability induced by systemic and focal inflammation has been implicated in COVID-19-associated strokes (22). However, hematological changes based on coagulopathy, such as elevated D-dimer levels, were not found in our patient. Regarding endothelial damage, Varga et al. showed endothelial SARS-CoV-2 invasion affecting blood vessels in pathology specimens (23), suggesting that SARS-CoV-2 infection facilitates the induction of endotheliitis. Endothelial dysfunction leads to impairment of the microcirculatory function with subsequent organ ischemia, which may lead to spinal cord infarction. A greater understanding of the mechanism underlying the development of MRI-negative myelopathy, even months after COVID-19 infection, is needed to improve the poor therapeutic outcomes.

The findings in our patient and in a literature review revealed several clinical characteristics of MRI-negative COVID-19-associated myelopathy, including an older age, possible late manifestation of myelopathy, and a poor response to treatment. In a previous review, MRI findings of the spinal cord were found to be normal in up to 10% of patients with COVID-19-associated myelopathy (12), which makes confirmation of the diagnosis difficult in some cases. Repeated MRI may subsequently detect culprit abnormal lesions; however, detailed CSF analyses, such as an examination of the IgG index and/or electrophysiological evaluation by SSEP, should be considered as a basis for the early initiation of immunotherapy. Given the poor treatment outcome with standard immunotherapy for AIM, more aggressive immunotherapy regimens or another treatment strategy should be considered in combination with comprehensive multidisciplinary rehabilitation. Further studies are needed to determine the mechanism underlying the development of MRI-negative COVID-19-associated myelopathy.

The authors state that they have no Conflict of Interest (COI).