Neuromuscular & Autonomic Complications of COVID-19

COVID-19, first reported in December 2019 and declared a “Public Health Emergency of International Concern” in March 2020, has caused a recorded 3,857,563 deaths. The still-ongoing pandemic of COVID-19 caused by SARS-CoV-2 infection has also spawned an unprecedentedly large body of literature describing new onset or aggravation of extrapulmonary conditions, particularly neurologic disease, in temporal association with COVID-19. An analysis of publication trends in the last 15 months reveals an ever-growing number of papers describing, analyzing, and summarizing multiple aspects of COVID-19 and neuromuscular conditions (Figure).

At a glance, this number may suggest a causal relationship between COVID-19 and neuromuscular disease, but biases could overestimate the significance and erroneously indicate causality. Bibliometric analysis demonstrates that this “tsunami” of COVID-19 publications contains a high number of poor-quality studies and a low number of studies of higher evidence (eg, clinical trials, large-cohort data registries, or meta-analysis).^1,2 Most published articles related to COVID-19 and neuromuscular disorders are case series or reports. Only 25% of more than 2,000 papers published on COVID-19 in the first quarter of 2020 contained original data.³ Although case reports are important to raise awareness of rare and novel associations, they are, in most instances, insufficient to establish causality. To assess evidence of neuromuscular and autonomic complications of COVID-19, objective criteria are required. Criteria for assessing causality proposed by Bradford Hill in 1965 consist of 9 characteristics: strength, consistency, specificity, temporality, biologic gradient, plausibility, coherence, experiment, and analogy.^4,5 Not all can be applied in this setting; for example, experimental evidence and specificity are lacking for all conditions. For coherence, it has been argued that data from severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS) epidemics could be consulted, because these coronaviruses share a 50% to 80% homology with SARS-CoV-2.⁵ The extent to which neuromuscular conditions discussed in this review meet these criteria is summarized in the Table.

Guillain-Barré and Miller Fisher Syndromes

Guillain-Barré syndrome (GBS) and Miller-Fisher syndrome (MFS) were among the earliest neurologic complications reported in people with SARS-CoV-2 infection and COVID-19. Evidence for the criteria strength and consistency is weak, however. Although the incidence of GBS was reported to be 2.6 higher in the first wave of the pandemic in Italy,⁶ studies from the United Kingdom⁷ and Singapore⁸ reported a lower incidence of GBS during the pandemic. The occurrence of GBS within 2 to 4 weeks after SARS-CoV-2 infection does meet the criteria of temporality.⁹ The time interval between SARS-CoV-2 infection and onset of GBS varies and is sometimes impossible to determine because GBS has been observed after asymptomatic SARS-CoV-2 infection. In more than 80% of those affected, GBS symptoms co-occurred with COVID-19 symptoms, including the need for artificial ventilation, which may mask a clear delineation of the conditions.¹⁰ Regarding the criteria of a biologic gradient, data are lacking in that it is not known whether increased exposure, more severe disease course, or higher virus load predispose people infected with SARS-CoV-2 to GBS.

GBS after SARS-CoV-2 infection is biologically plausible, based on the conception of GBS as a postinfectious disorder in which molecular mimicry is essential. This mechanism, however, requires viral epitopes (ie, peptide or protein) with similarity to molecules expressed in the peripheral nervous system, allowing antibodies to the virus to cross-react with endogenous proteins. Data suggesting such cross-reaction could occur, are mixed. A genomic and proteomic analysis showed no significant similarity between SARS-CoV-2 and human proteins.⁷ Other analyses demonstrated shared oligopeptides between SARS-CoV-2 and 2 human heat-shock proteins¹¹ and up to 34 proteins that have an oligopeptide sequence shared by the SARS-CoV-2 spike glycoprotein.¹² Whether heat-shock proteins or any of the other proteins with homology to SARS-CoV-2 are relevant targets of aberrant immune responses in GBS is unknown, however. The analogy criterion might be strong for GBS because numerous viruses are commonly accepted as triggers for GBS including human herpes viruses, cytomegalovirus, varicella zoster and influenza.^13,14 Whether existing evidence is coherent is debatable. Using the suggestion that coherent data could be derived from experience with SARS and MERS, no case of GBS after either has been reported and only 1 case was reported after MERS. On a cautionary note, the overall number of infected individuals for SARS and MERS is low, thus these epidemics may not serve as good models to study rare complications. Experimental evidence for a relationship between SARS-CoV-2 and GBS or MFS is lacking. In contrast, this has been shown for other postinfectious molecular mimicry in GBS (eg, gangliosides targeted by autoantibodies that are generated by infection with Campylobacter jejuni).¹⁵

Chronic Inflammatory Demyelinating Polyradiculoneuropathy

Chronic inflammatory demyelinating polyradiculoneuro-pathy (CIDP) is a chronic progressive or relapsing inflammatory autoimmune neuropathy. It typically presents as subacute evolving symmetric neurologic deficits, distributed distally and proximally. CIDP variants include distal acquired demyelinating symmetric (DADS), multifocal acquired demyelinating sensory and motor neuropathy (MADSAM, or Lewis-Sumner syndrome), and pure motor or sensory variants (see Chronic Inflammatory Demyelinating Polyradiculoneuropathy in this issue).¹⁶ Although post-COVID-19 CIDP is plausible, the frequency of reports is low such that strength, consistency, and biologic gradient is lacking. The general plausibility of COVID-19 causing CIDP derives from the pathogenic concept of CIDP as an autoimmune condition triggered by bacterial or viral infections. In contrast to GBS, however, the spectrum of infections preceding CIDP is much less known. In a cohort study of 92 people with CIDP, approximately one-third could identify an infection within 6 weeks before CIDP onset, and of those individuals, 60% remembered a nonspecific upper respiratory tract infection.¹⁹ Thus, neither evidence from analogy, nor coherence can be invoked. In summary it is very unlikely that CIDP is triggered or exacerbated by infection with SARS-CoV-2 or COVID-19.

Myasthenia Gravis

Several case reports from Italy, Germany, and the US describe onset of ocular or generalized myasthenia gravis (MG) 5 to 10 days after COVID-19, which may lay within the range of a temporally plausible timeframe. The concept of postinfectious MG, however, is not well developed. Considering there is a background incidence for MG of 2 to 3 per 100,000 per year (see Myasthenia Gravis in this issue),²⁰ a much higher number of postCOVID-19 cases of MG than have been reported would be expected to fulfill the causality criteria of strength, consistency, and biologic gradient. 

Neuralgic Amyotrophy

Neuralgic amyotrophy (ie, Parsonage Turner syndrome) is an idiopathic inflammatory neuropathy of the upper limbs that usually affects the upper part of the brachial plexus.²¹ Therefore, a brachial plexus neuritis preceded by SARS-CoV-2 infection appears principally plausible. Infections with DNA and RNA viruses, including hepatitis E, parvovirus B19, HIV, herpes viruses, and West Nile virus can precede neuralgic amyotrophy supporting an analogous autoimmune pathophysiologic mechanism. A few reported cases of neuralgic amyotrophy occurred approximately 2 weeks after people had COVID-19, suggesting temporality.²² Like MG, however, the incidence of neuralgic amyotrophy is estimated as 1 to 3 per 100,000 per year,²³ making the reported cases within the error margin of any statistical evidence. Hence, the causality criteria strength, consistency, and biologic gradient are absent.

Myalgia, Myositis, and Rhabdomyolysis

Myalgias are considered among the most common and early neurologic symptoms of COVID-19, affecting up to 50% of all patients.²⁴ In approximately half of these individuals, myalgias improve within a few days, similar to symptoms of fever and cough. The proportion of individuals who had COVID-19 (hospitalized or not) who complain about myalgia decreases by 6 months after illness to 2% to 4%.^25,26

Approximately one-third of people with COVID-19 have an elevated serum CK level,²⁴ and these individuals had a higher likelihood of death from COVID-19 (odds ratio [OR], 2.1 when CK>185 U/l),²⁷ but this association was not found in a comparable study.²⁸ Additionally, much higher likelihood of COVID-19-related mortality is seen with other prognostically relevant laboratory parameters (eg, OR, 45.43 with elevated lactate dehydrogenase).²⁷ Elevated CK also is not specific for COVID-19 and occurs in severe influenza.²⁹ Whether dexamethasone improves this risk is unclear because data from trials has not reported changes in CK levels during treatment.

Viruses are known to trigger myositis, making myositis after COVID-19 plausible.³⁰ Although direct infection of muscles by viruses is rare, because muscle fibers express the angiotensin-converting enzyme 2 (ACE2) receptor through which SARS-COV-2 enters cells, COVID-19 may be an exception. This hypothesis, however, needs confirmation and therefore Hill’s criterion of analogy does not apply. Only a few cases of myositis have been reported after COVID-19, and these diagnoses were predominantly based only on nonspecific MRI changes.³¹ A small case series reported 5 people who had dermatomyositis with COVID-19 and responded to corticosteroids or intravenous immunoglobulin (IVIG).³² Fatigue and muscle weakness, but not myalgia, are commonly present in patients 6 months after COVID-19.^26,33 From the 9 Bradford Hill criteria, only plausibility and temporality are supported, whereas strength, consistency, specificity, biologic gradient, coherence, and analogy are not.

Rhabdomyolysis is a clinical and biochemical syndrome caused by acute skeletal muscle necrosis. With rhabdomyolysis, clinically significant myoglobinuria may occur and leads to renal failure in 15% to 33% of cases.³⁴ Rhabdomyolysis has many causes, including substance abuse, trauma, extreme overexertion, epileptic seizures, and less frequently, viral infections. Rhabdomyolysis has been described in MERS and SARS, fulfilling criteria for analogy, and coherence may apply. Virally mediated rhabdomyolysis is thought to be caused by direct viral invasion of muscle, and as noted, muscle cells do express the ACE2 receptor through which SARS-CoV-2 infects the host, making SARS-COV-2-induced rhabdomyolysis plausible. Strength and consistency are supported by numerous case reports of rhabdomyolysis during or after COVID-19 infection as well as 2 retrospective studies that reported an incidence ranging from 2.2% to 17% in persons hospitalized with COVID-19.^35,36 This incidence increases to up to 50% of those in the intensive care unit (ICU),³⁷ supporting a biologic gradient. Male sex, obesity, hypertension, diabetes mellitus, and chronic kidney disease are risk factors for rhabdomyolysis.

ICU-acquired weakness (ICUAW)

The term ICU-acquired weakness (ICUAW) is used to describe polyneuropathy and/or myopathy that occurs in persons who are critically ill during admission to the ICU. ICUAW after COVID-19 is biologically plausible, considering the high rates of intensive care, sepsis, and prolonged ventilation with COVID-19, which are all risk factors for ICUAW. People who have recovered from COVID-19 frequently complain about muscle weakness, as long as 6 months after the disease,²⁶ which may point to a relevant proportion of individuals who develop ICUAW. Weakness after COVID-19 may also occur in analogy to other viral diseases (eg, influenza requiring prolonged stays in the ICU), but the criterion coherence cannot be applied because data regarding the frequency of ICUAW after critical illness due to SARS, MERS, or COVID-19 are unavailable. A prospective study from Finland reported a general incidence of critical illness-related polyneuropathy/myopathy of approximately 10% in COVID-19 cases, which is more frequent than is seen with non-COVID-19 causes of ICU stays, supporting a strong association of the ICUAW and COVID-19. Consistency is yet not clear, however, because only the Finnish study evaluated ICUAW.³⁸

Autonomic Nerve Failure

Dysfunction of the autonomic nervous system has also been suggested to be among extrapulmonary manifestations of COVID-19 and postacute sequelae of SARS-CoV-2 infection (PASC) (also termed long COVID). Autonomic dysfunction has also been described in SARS³⁹ and other viruses, supporting the criteria analogy and coherence. Plausibility, however, seems questionable, because direct infection of autonomic nerves has not been demonstrated, and autonomic dysfunction in other postviral neuropathic conditions usually occurs with both sensory and motor fiber dysfunction (eg, GBS). Smaller case series have been reported that show altered sudomotor function,⁴⁰ and postural tachycardia in people with COVID-19 during illness and recovery phase,⁴¹ supporting temporality, but these are too small to demonstrate strength and consistency of such an association.

Summary

Taken together—owing to the limitations that the Bradford Hill criteria may bear—currently, rhabdomyolysis and ICUAW seem probable to be causally linked to COVID-19, whereas for the other conditions discussed here, evidence is much lower. To further prove or exclude causality, cohort studies are warranted. In addition, experimental evidence derived from preclinical studies would be highly desirable.