Preventative strategies to reduce infections of the coronavirus, SARS-CoV-2, including wearing masks (N95 most effective) and face shields, avoiding prolonged exposure to infected individuals, avoiding confined indoor spaces, quarantines, and vaccines and boosters, have proven successful. While disease-modifying therapeutic strategies for coronavirus-induced disease-2019 (COVID-19), including the antiviral remdesivir (1), are emerging, the contagiousness of the virus and the morbidity and mortality associated with COVID-19 continue to place an urgent demand on the development of new therapies. Presently worldwide, nearly 450 million persons have been infected worldwide, 6 million of whom have succumbed to the disease (https://coronavirus.jhu.edu/map.html). Severe cases of COIVD-19 are often associated with a “cytokine storm,” which may cause dysfunction of the lungs and other organs. Studies investigating pharmacological therapies targeting the immune and respiratory systems are ongoing.

The human immune and respiratory systems are regulated by the central nervous system (CNS), and therapeutic targeting of these neural pathways is potentially one approach to regulate important aspects of the pathologic sequelae of COVID-19. Neuromodulation directly targets the nervous system through the application of electric or magnetic fields to neuronal tissue, including peripheral nerve fibers and directly to the brain. Neuromodulation may therefore provide alternative therapeutic strategies for the neurobiological systems involved in the pathophysiology of SARS-CoV-2 infections.

This review provides an overview of the role of the nervous system in immune and respiratory control, current understanding of how neuromodulatory techniques may be effective in immune or pulmonary disorders, and provides a future perspective of its role in the treatment of COVID-19. The authors of this review formed the International Consortium on Neuromodulation for COVID-19 (www.covidneuromod.org) in April, 2020, in response to the pandemic with the understanding that neuromodulation, particularly vagus nerve stimulation (VNS) or sacral nerve modulation (SNM), may be able to block infection-induced inflammation through the cholinergic antiinflammatory pathway (CAP). Now including 30 members on three continents, this group was formed to advocate for and support the scientific study of neuromodulation technologies for treatment of excessive inflammation and pulmonary dysfunction in this and future pandemics when appropriate. The authors provide recommendations on clinical trial endpoints, as well as recommended publication guidelines in an effort to standardize technology characterizations.

In up to 30% of mortality caused by SARS-CoV-2, a severe systemic inflammatory response known as a “cytokine storm” is implicated in multi-system organ failure (2, 3). These cases are characterized by high concentrations of circulating proinflammatory cytokines, including TNF, IL-6, IL-7; the inflammatory chemokines CCL2, CCL3 and CXCL10 (4); and acute-phase proteins including C-reactive protein and ferritin are also associated with disease severity (4, 5). Clinically, the most typical phenotype of severe COVID-19 is acute respiratory distress syndrome (ARDS), which affects 3.4% of infected patients overall and 15.6–17.0% of severely infected patients (6). The COVID-19 hyperinflammatory response may induce disseminated intravascular coagulation through both endothelial cell activation and impairment of endogenous anticoagulant pathways (7–12). This disease pathophysiology may pose specific problems for individuals with preexisting cardiopulmonary conditions, including hypertension and chronic obstructive pulmonary disease.

The mammalian nervous and immune systems evolved to maintain homeostasis and protect the host against tissue injury and infection (13). Autonomic nerves monitor infection and inflammation and form homeostatic reflexes to regulate host responses (14, 15). An important neural conduit for monitoring and regulating immune responses is the vagus nerve (VN), the body's longest cranial nerve. Peripheral vagal fibers sense local cytokines and damage- and pathogen-associated molecular patterns (DAMPs, PAMPs), which activate neural regulatory reflexes (16, 17). The brain can respond to active inflammation via glucocorticoid release from the hypothalamic-pituitary-adrenal (HPA) axis, or through activation of the CAP (18). The CAP comprises efferent VN fibers innervating the spleen via the abdominal celiac ganglion and splenic nerve. The role of sympathetic nerves in regulating inflammation and pulmonary function are reviewed elsewhere (15, 19–21).

Electrical VNS was first shown to protect against excessive inflammation in 2000, when Borovikova et al. reported that VNS reduces systemic proinflammatory cytokine production and maintains blood pressure in a rodent model of endotoxic shock (22). Electrically or pharmacologically stimulating the CAP protects against excessive systemic inflammation, maintains tissue perfusion, and preserves vital organ function in models of arthritis, burns, colitis, hemorrhagic shock, polymicrobial intraabdominal sepsis, and stroke (23–34). Importantly, disruption of CAP signaling delays resolution of inflammation and bacterial clearance in experimental peritonitis (35). Specific CAP components also play key roles in regulation of circulation, microvascular contractility, and viral infection clearance in experimental models (36, 37).

Molecular and functional mapping of the CAP have revealed that VNS inhibits release of pro-inflammatory mediators via acetylcholine-releasing T lymphocytes expressing choline acetyltransferase (ChAT⁺) (38). Local acetylcholine secretion in spleen inactivates cytokine-producing macrophages through surface α7 nicotinic acetylcholine receptor subunits (37). Pro-inflammatory cytokine production in spleen contributes significantly to overall systemic inflammation (39). Elimination of these cellular or molecular components via genetic deletion or surgical disruption abolishes CAP activity and results in significantly elevated serum cytokine levels (40).

In addition to providing neural regulation of inflammation, the VN also provides the dominant autonomic innervation of the airways, carrying fibers that are responsible for both bronchoconstriction and relaxation through distinct sets of pulmonary nerve branches (41–43). Low-intensity VNS relaxes airway smooth muscle in animal models of bronchoconstriction by activating the HPA axis to release catecholamines from the adrenal cortex (41, 44, 45). These in turn bind to β-receptors on bronchial smooth muscle causing relaxation. Measured concentrations of epinephrine and norepinephrine were sufficient to relax smooth muscle, but insufficient to cause significant cardiac effects (46, 47).

The sacral nerve (SN) is part of the parasympathetic nervous system that also holds therapeutic potential for pathological inflammatory processes. Both afferent and efferent SN fibers innervate the pelvic organs including colon, rectum, urinary bladder and genital organs. Sacral nerve modulation (SNM) has been used to treat diseases of the pelvic organs, including pain, urinary retention, overactive bladder and fecal incontinence (48, 49). The mechanisms of action of SNM are not completely understood. Limited findings suggest that SNM ameliorates overactive bladder by blocking C fiber activities or overactivity (50). In treating urinary retention, SNM seems to stimulate relaxation of pelvic floor muscles and the urethra, facilitating micturition (51, 52). The mechanism involved in SNM-induced improvement in fecal incontinence is also poorly understood, with most research findings suggesting that SNM stimulates afferent fibers from the rectum, pelvic floor and anal sphincter, resulting in reduced activation of C fibers during rectal filling and inhibiting inputs from the rectum to the pontine center (53). SNM in patients with fecal incontinence may activate somatic afferent fibers to inhibit colonic motility and enhance internal anal sphincter pressure (54).

Recently, SNM has been shown to reduce intestinal inflammation in rodent models of inflammatory bowel diseases, suppress a number of proinflammatory cytokines including TNF, and enhance anti-inflammatory cytokines such as IL-10 in both intestinal tissues and in circulation (55, 56). The anti-inflammatory effect of SNM on the colon is mediated via the spinal afferent and vagal efferent pathways evidenced by the following factors (57): (1) SNM activates neurons in the nucleus tractus solitaries (57, 58); (2) SNM increases vagal efferent activity as assessed by the spectral analysis of heart rate variability and plasma level of pancreatic polypeptide (55); (3) bilateral subdiaphragmatic vagotomy almost completely abolishes the anti-inflammatory effect of SNM on the colon (57); and (4) SNM induces the release of acetylcholine in the colon (55). These findings suggest that SNM shares a similar vagal efferent pathway as VNS. Indeed, a comparative study has shown that the anti-inflammatory effect of SNM was similar to that of VNS and the combination of SNM and VNS did not show a synergistic effect in comparison with SNM or VNS alone (59, 60).

Together, these findings suggest that VNS or SNM enhance vagal efferent activity and have systemic anti-inflammatory effects. Several VNS technologies are being investigated, some for use in COVID-19, as summarized below.

Electrical signaling between the CNS and individual organ systems is essential to organ function. Homeostasis is continually monitored and regulated through a bi-directional feedback loop that senses the physiological status of organ function, central nervous system processing of the sensory information, and appropriate feed-back to the organs. Therapeutic neuromodulation aims to restore a diseased organ system by interfering with or augmenting its neural control mechanisms via targeted application of external energy sources, typically electrical or magnetic. Neuromodulation technologies have been developed to treat neuropsychiatric disorders and conditions where the nervous system is known to play an important role, including epilepsy, depression and pain (61, 62). Of note, VNS utilizing a surgically implanted electrical pulse generator is an established clinical therapy for the treatment of medically refractory epilepsy and depression. Since 1988, more than 100,000 patients have received this FDA-approved and CE-marked therapy, and long-term follow-up studies demonstrate overall safety of the approach (63).

Functional mapping of the CAP has provided a foundation for clinical trials to evaluate the efficacy of VNS to treat inflammatory diseases. Several studies, albeit with a limited number of patients included, show encouraging data on use of electrical VNS for the treatment of chronic inflammation, e.g., rheumatoid arthritis and Crohn's disease (64–66). VNS significantly reduced whole blood TNF, IL-6, and IL-1β concentrations, and decreased scores on a validated disease activity composite score (DAS28-CRP score) (65). Likewise, non-invasive cervical VNS was tested in two clinical trials for treating acute asthma patients presenting to the emergency department who were unresponsive to 1 h of standard of care therapy (67–69). In both studies, VNS significantly improved two measures of airway patency, FEV₁ and perceived work of breathing, after 15, 30 and 60 min (compared with a matched control group), with no reported serious adverse events. These data suggest a possible, novel treatment for COVID-19 patients suffering from acute respiratory dysfunction. However, care must be exercised in choosing stimulation parameters because high intensity cervical VNS, such as is used to treat epilepsy and depression, is capable of inducing bronchoconstriction, which is reversed upon cessation of stimulation (70).

In light of the available mechanistic experimental data on the capacity of neuromodulation techniques to attenuate excessive acute and chronic inflammation, and induce bronchodilation; the clinical safety record of electrical VNS; and the encouraging reports on reduced inflammation with chronic implantation of an electrical vagus nerve stimulator in clinical trials, it is reasonable to consider whether neuromodulation devices may be useful tools to reduce the morbidity and mortality of COVID-19 patients (71–79). A significant challenge for applications in acute-onset excessive inflammation is the delivery method of VNS, as it is currently impractical to implant highly symptomatic COVID-19 patients with an active device that is only needed for acute treatment. However, less invasive neuromodulation techniques, either percutaneous or transcutaneous, may be utilized (Tables 1–4). These methods may be technically suitable for acute use, but their clinical feasibility and efficacy in COVID-19 patients are just beginning to be evaluated. Results from two recent COVID-19 studies have been published, results are pending from several ongoing clinical trials, and a number have been discontinued (discussed below).

The exact mechanisms of VNS are not fully understood. In addition to activating efferent CAP fibers to the spleen, VNS is believed to proportionally activate afferent fibers (40, 125). Vagal afferents carry sensory information to the dorsal medullary complex, in particular to the nucleus of the solitary tract (NTS). Activation of the NTS plays an important role in modulating cardiopulmonary function (126–128). Therapeutic mechanisms in epilepsy and depression may involve regulating the serotonergic and/or noradrenergic projections from the raphe and LC, respectively, which in turn modulates activity in the hypothalamus, thalamus, insular cortex, and cerebellum, hippocampus, amygdala, and posterior cingulate gyrus (129, 130). These structures are associated with the regulation of mood and emotion, seizure activity, anxiety, intestinal activity, satiety, and pain perception (131–134). Other potential, centrally mediated mechanisms for VNS include elevation of gamma aminobutyric acid (GABA) levels in the brain stem (135), desynchronization of cortical and thalamocortical network activity (125), suppression of cortical spreading depression (136), altering expression of brain-derived neurotrophic factor and other growth factors and increasing norepinephrine in prefrontal cortex (137), and altering brain waste clearance dynamics (138).

There are currently six studies listed on ClinicalTrials.gov assessing the effects of VNS on immune and/or pulmonary function in COVID-19 (Table 5). Two of the trials used gammaCore, and thus targeted the cervical vagus nerve transcutaneously. Four trials employed several different auricular vagus nerve stimulation devices. Two studies are completed, three are open to enrollment or active but not recruiting, and one terminated early due to low recruitment rates.

The use of cervical non-invasive VNS in COVID-19 was first published in a case report of two COVID-19-positive subjects, which provided preliminary observations that transcutaneous cervical VNS reduced the use of opioid and cough suppressant medications and relieved symptoms of chest tightness and shortness of breath (139). Two randomized control trials of gammaCore in inpatients with COVID-19 and respiratory distress have since been initiated: SAVIOR-1, performed in Valencia, Spain, was a prospective, randomized controlled trial (RCT) of VNS vs. standard of care, which evaluated a total of 90 patients, 45 in each arm. Improvement in respiratory function and proinflammatory cytokine levels were measured following VNS (NCT04368156) (140). SAVIOR-2, performed at Allegheny Health Network, was scheduled to study 60 patients and also followed respiratory parameters and cytokines (NCT04382391). This study closed to enrollment after 21 subjects due to low case burdens resulting in low recruitment rates.

The remaining studies employed devices targeting the auricular branch of the vagus nerve either in the cymba concha or the tragus. The results of the first study suggest that auricular VNS results in infrequent mechanical ventilation and a high rate of survival, though it should be noted that this trial in 51 subjects was a single-arm, uncontrolled, open-label, observational trial (141). Studies of two additional auricular VNS devices are actively recruiting, and results are not yet available. A French study is assessing auricular nerve stimulation via four percutaneously placed leads on each tragus and is being compared to sham (no needle placement; NCT04341415). Enrollment has been halted due to a low case load (142).

COVID-19 is a deadly disease, the most serious effects of which appear to be hyperactivation of an acute inflammatory response (“cytokine storm”) and compromised lung function. To date, pharmacological strategies have not been generally successful in modulating the cytokine response and clinical sequellae to SARS-CoV-2, with the exception of dexamethasone, a non-specific anti-inflammatory corticosteroid that reduces mortality by one-third in severe COVID patients with acute respiratory distress (143). As described here, there is substantial preclinical and clinical evidence that neuromodulation strategies may have a salutary impact on the host responses to the virus, and clinical trials to assess these strategies are warranted. Given the unabated SARS-CoV-2 pandemic, speed to execute these trials and deliver rigorous results is of essence.

The authors of this paper have formed the International Consortium on Neuromodulation for COVID-19 (ICNC; www.covidneuromod.org) with the mission to provide global leadership for the rapid advancement and clinical adoption of neuromodulation technologies to treat emerging infectious diseases. The ICNC envisions a global infrastructure capable of the rapid deployment and clinical validation of neuromodulation technologies for pandemics, epidemics, and other emerging public health crises for which they may be efficacious. Further, beyond applications for immunological and respiratory control in response to infection, neuromodulation has multiple potential therapeutic applications for the post-acute recovery phase. For example, a recent report has indicated that >87% of COVID-19 patients continue to experience at least one symptom following acute recovery (144), the most frequent of which was fatigue. Noninvasive brain stimulation approaches have reliable benefit for fatigue (145), as well as for psychological symptoms such as depression and anxiety that are common in the recovery period (146–148).

ICNC members have developed recommended endpoints for neuromodulation clinical in COVID-19, as well as recommended publishing guidelines to describe the technology tested.

While prospective, RCTs are the gold standard for such clinical trials and sham control arms are desirable, they might not always be necessary if objective endpoints are used to measure outcomes (149). Such endpoints will vary based on hypothesized mechanism of action of the technology, as well as the patient population being studied: wearable and handheld neuromodulation technologies can be trialed in hospitalized patients (non-ICU or ICU), as well as ambulatory outpatients pre- or post-hospitalization. Nevertheless, some key data elements and outcome measures are recommended, in the anticipation that post-study meta-analyses can illuminate overarching trends not apparent in individual studies (Table 6). Upon publication of study results, the ICNC members recommend careful descriptions of the neuromodulation technology (150).

We strongly believe that multi-national and academic-industry collaborative and coordinated efforts to tackle COVID-19 are necessary to advance our knowledge and provide the urgently needed therapeutic options. Our present collaboration and this review on neuromodulation will hopefully serve as an example/roadmap for other scientific areas.

All authors wrote sections of the manuscript, contributed to manuscript revision, and read and approved the submitted version.

CC is an equity holder of Convergent Medical Technologies, Inc., reports personal fees from electroCore Inc., and Spark Biomedical, and has issued and pending patents related to vagus nerve stimulation to control bleeding and inflammation. MB reports personal fees from Soterix Medical, grants from Google X, personal fees from Halo Neuroscience, outside the submitted work; in addition, he has a patent Brain Stimulation issued. MFr reports pending patent applications describing injectable electrode structures for Neuronoff, Inc., and is an employee and equity holder of Neuronoff, Inc. He further reports issued patents for selective Vagus nerve stimulation and selective electrical vagal block for regulation of autonomic functions such as heart rate and blood pressure for Boston Scientific. MFu reports personal fees from Axon Therapies, CVRx, Daxor, Edwards Life Sciences, Galvani, and Respicardia, outside the submitted work. SH is an employee and equity holder of Cala Health. NK reports a patent devices and methods for reducing inflammation using electrical stimulation pending. BS reports personal fees from electroCore, and reports patents issued and pending related to transcutaneous cervical vagus nerve stimulation. PS is an employee and equity holder of electroCore, and reports patents issued and pending related to transcutaneous cervical vagus nerve stimulation. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.