Long-Segment Nonfocal Peripheral Neuropathies After COVID-19 Infection: A Case Report of Magnetic Resonance Neurography Findings

Introduction

Neurologic complications associated with COVID-19 may occur in more than 80% of patients, according to recent findings [6,12]. Although many studies have focused on central nervous system manifestations, peripheral nerve complications may be similarly prevalent and, in many cases, severe [6]. Several reports have described cranial neuropathies and Guillain-Barré syndrome following COVID-19, but few have described other types of peripheral neuropathies [20]. Of those reports that do, it is unclear whether such neuropathies result from para- or postinfectious immune responses, direct viral infection of nerves, or prolonged prone positioning for those with respiratory complications [1,3,7,13,16–19]. Magnetic resonance (MR) neurography is often used as an adjunct to the neurologic examination and electrodiagnostic testing (EDX) to evaluate peripheral neuropathies and may aid in the recognition of COVID-19–associated neuropathies.

In this article, we describe MR neurography findings in 2 patients who developed severe peripheral neuropathies while recovering from COVID-19.

Case Reports

For the reporting of these cases, we received Institutional Review Board approval and obtained written informed consent. Medical records were retrospectively reviewed of 2 patients with peripheral neuropathies and recently confirmed COVID-19 referred for MR neurography between March 1, 2020, and August 24, 2020. Magnetic resonance imaging acquired at 3 Tesla (Signa Premier, GE Healthcare, Waukesha, WI) was reviewed by a musculoskeletal radiologist with more than 7 years of peripheral nerve imaging experience (D.B.S.).

Patient 1, a 50-year-old man with diabetes, hypertension, and asthma, was hospitalized with acute respiratory failure 5 days after developing a cough and fever. He tested positive for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by pharyngeal swab real-time reverse-transcriptase polymerase chain reaction (RT-PCR) and was intubated for 3 weeks in the supine position. He awoke with the inability to plantarflex or dorsiflex his left foot or toes. He also had severe neuropathic pain and hyperesthesia in the sciatic distribution. He was discharged 1 week later to inpatient rehabilitation and given pregabalin 150 mg 3 times daily. Approximately 1.5 months postdischarge, he had significant recovery in the tibial division of his sciatic nerve by examination; however, he reported persistent pain and had severe deficits in the peroneal nerve distribution. He also had diminished sensation to light touch distal to the knee. EDX at that time confirmed a severe, proximal sciatic neuropathy.

MR neurography of the left pelvis and thigh 3 months postonset revealed diffuse, long-segment, T₂-weighted signal hyperintensity of the sciatic nerve throughout the thigh, with otherwise maintained nerve fascicular architecture and size (Fig. 1). Prominent T₂-signal hyperintensity of all hamstring musculature was compatible with active denervation.

Fig. 1.

Magnetic resonance neurography of the pelvis and left thigh in patient 1 (50-year-old man). Coronal (a) and axial (b) T₂-weighted Dixon water images through the pelvis (a) and left thigh (b) demonstrate long-segment signal hyperintensity of the left sciatic nerve (solid arrows) and denervation edema pattern of the hamstring musculature (bracket, b). Compare with the normal signal intensity of the right sciatic nerve (dashed arrows, a).

Patient 2, a 55-year-old man with a history of paroxysmal atrial fibrillation was admitted with acute respiratory failure and diagnosed with COVID-19 by pharyngeal swab RT-PCR. He was intubated for 7 weeks in the supine position. After extubation, severe, patchy, left upper extremity weakness, numbness, and pain were noted. EDX 5 days later suggested a left brachial plexopathy with complete muscle denervation in the distribution of the musculocutaneous, median, and ulnar nerves and partial radial nerve denervation. He was treated with gabapentin 1800 mg daily, with some improvement in pain and numbness but persistent weakness. Repeat EDX 7 weeks later revealed marked sensorimotor abnormalities in the left median and ulnar nerve distributions and, to a lesser degree, the radial, axillary, musculocutaneous, and suprascapular nerve distributions. Neurologic examination revealed patchy weakness throughout the left upper extremity, with most severe involvement (no movement observed) of the abductor pollicis brevis and interosseous muscles of the hand.

MR neurography of the left brachial plexus and arm 3 months postonset demonstrated diffuse, T₂-weighted signal hyperintensity of the plexus (Fig. 2). Within the arm, the ulnar, median, and musculocutaneous nerves were all hyperintense. Findings of active muscle denervation in the distribution of multiple nerves within the arm and forearm were noted.

Fig. 2.

Magnetic resonance neurography of the left brachial plexus in patient 2 (55-year-old man). Coronal T₂-weighted, post-gadolinium inversion recovery image (a) demonstrates prominent signal hyperintensity of nerves throughout the left brachial plexus. Oblique sagittal T₂-weighted Dixon water image (b), orthogonal to the infraclavicular plexus, as delineated by the dashed line in A, confirms signal hyperintensity of the terminal branches in short axis. Axial T₂-weighted Dixon water image (c) through the proximal arm demonstrates denervation edema pattern of the biceps muscle (bracket) and signal hyperintensity of the median (solid arrow), ulnar (dashed arrow), musculocutaneous (arrowhead), and radial (dotted arrow) nerves.

A left medial brachial cutaneous nerve biopsy 4.5 months post-COVID-19 onset demonstrated marked loss of large myelinated axons, consistent with a nonspecific neuropathic process (Figs. 3 and 4). There was a moderate increase in epineurial, perineurial, and endoneurial fibrosis with no inflammatory infiltrate detected. SARS-CoV-2 was not detected in the nerve by electromicroscopy, in situ hybridization, or immunohistochemical staining for viral antigens.

Fig. 3.

Biopsied nerve demonstrates marked loss of myelinated axons and increased perineurial and endoneurial fibrosis (thin arrow: endoneurial fibrosis, thick arrow: epineurial fibrosis, arrowhead: myelinated axon). Epon thick section, toluidine blue stain, ×40.

Fig. 4.

Biopsied nerve demonstrates marked loss of large and small myelinated axons (thick arrow: single myelinated axon in field, thin arrows: endoneurial collagen). Transmission electron microscopy (TEM) image. Magnification indicated by 5-μm scale bar.

Discussion

We report MR neurography findings in 2 patients with nonfocal peripheral nerve abnormalities extending over long segments (entire thigh; plexus to elbow) that developed within 3 and 7 weeks for patients 1 and 2, respectively, of the COVID-19 diagnosis. Both patients experienced severe clinical courses of COVID-19 that required prolonged intubation, and neuropathies were realized upon awakening. Clinical deficits affected lower (patient 1) and upper (patient 2) extremity nerves and were severe, with motor function loss ensuing for months and unresolved at the time of this report. No other major neurologic complications were observed in either patient.

The imaging finding of peripheral nerve T2-weighted signal hyperintensiy is non-specific and common to most types of neuropathies, including common entrapment neuropathies such as carpal and cubital tunnel syndrome, and in cases of nerve compression [8]. In both patients, however, T₂ signal hyperintensity extended over long segments. The observation of long-segment involvement is more characteristic of an inflammatory or demyelinating etiology, as seen in multifocal motor neuropathy and chronic inflammatory demyelinating polyneuropathy [5], rather than from focal compressive or entrapment-type neuropathies.

While several reports have proposed nerve compression from prolonged prone positioning as a cause of COVID-19–associated neuropathy in intubated patients [13,18], both patients in our report remained intubated in the supine position only. In addition, the MR findings of long-segment T₂ signal hyperintensity in our report, while nonspecific, are more suggestive of an inflammatory than a compressive or entrapment-type etiology, in which abnormalities would be shorter segment (ie, more focal). Both patients in this report, as well as many patients in prior reports of COVID-19-associated neuropathies, experienced lower extremity neuropathies or mononeuropathies [13,18]. The Proning Severe ARDS Patients (PROSEVA) trial, which evaluated the effects of prone positioning over multiple 16-hour sessions in 237 patients with acute respiratory distress syndrome, also did not describe peripheral neuropathy as a complication [9,13].

Other neuropathies with presumed inflammatory or immune-mediated etiologies have been associated with SARS-CoV-2. These include Guillain-Barré syndrome, in which SARS-CoV-2 is thought to trigger an immune-mediated pathological response by molecular mimicry or failure of tolerance to self-antigens [4], as well as Parsonage-Turner syndrome (aka neuralgic amyotrophy). In Parsonage-Turner syndrome, long-segment T₂ signal hyperintensity of affected nerves is detected, but in addition “hourglass-like” constrictions are typically identified in affected nerves.

Other possible etiologies include ischemia (COVID-19 is associated with higher thrombosis rates) and direct viral infiltration of the nerve as seen with neurotropic viruses such as herpesviruses (eg, cytomegalovirus) and rabies [2,14,15]. SARS-CoV-2 has been shown to directly infect brain cells, including astrocytes, endothelial cells, and neurons, but infection of peripheral nerves has yet to be demonstrated [10]. Patient 2 in this series underwent peripheral nerve biopsy to determine whether viral infiltrate or signs of inflammation were present, both of which, in theory, would be amenable to medical therapy. While biopsy found neither of these, it was performed 4.5 months after initial infection, and results cannot be considered conclusive.

Critical illness polyneuropathy is another possible cause of neuropathy in COVID-19 patients [11]. This syndrome is characterized by acute-onset, sensory-motor polyneuropathy resulting in flaccid and symmetrical weakness of the extremities (and respiratory muscles), as well as sensory symptoms [11,21]. Critical illness polyneuropathy seems an unlikely diagnosis for our cases, in which neuropathy involved only a single limb, and for 1 case a single nerve distribution.

This case report has several limitations. It includes only 2 patients, and larger studies are needed to evaluate the incidence, timeline, and characteristics of peripheral neuropathies that develop following COVID-19 diagnosis. Peripheral nerve histology was additionally available in only a single patient, and further research is needed to discern the etiology of post-COVID-19 neuropathies. Furthermore, MR neurography in this study was performed 3 months postneuropathy onset. In future studies, sequential imaging can be obtained to evaluate acute and chronic imaging features of these neuropathies.