What can the neurological manifestations of COVID-19 tell us: a meta-analysis

Background Covid-19 became a global pandemic in 2019. Studies have shown that coronavirus can cause neurological symptoms, but clinical studies on its neurological symptoms are limited. In this meta-analysis, we aimed to summarize the various neurological manifestations that occurred in COVID-19 patients and calculate the incidence of various neurological manifestations. At the same time, we further explored the mechanism of nervous system injury and prognosis in COVID-19 patients in combination with their nervous system manifestations. This study provides a reference for early clinical identification of COVID-19 nervous system injury in the future, so as to achieve early treatment and reduce neurological sequelae. Methods We systematically searched all published English literature related to the neurological manifestations of COVID-19 from January 1, 2020, to April 30, 2021, in Pubmed, Embase, and Cochrane Library. The keywords used were COVID-19 and terminology related to the nervous system performance. All included studies were selected by two independent reviewers using EndNote and NoteExpress software, any disagreement was resolved by consensus or by a third reviewer, and the selected data were then collected for meta-analysis using a random-effects model. Results A total of 168 articles (n = 292,693) were included in the study, and the meta-analysis showed that the most common neurological manifestations of COVID-19 were myalgia(33%; 95%CI 0.30–0.37; I2 = 99.17%), smell impairment(33%; 95%CI 0.28–0.38; I2 = 99.40%), taste dysfunction(33%; 95%CI 0.27–0.39; I2 = 99.09%), altered mental status(32%; 95%CI 0.22–0.43; I2 = 99.06%), headache(29%; 95%CI 0.25–0.33; I2 = 99.42%), encephalopathy(26%; 95%CI 0.16–0.38; I2 = 99.31%), alteration of consciousness(13%; 95%CI 0.08–0.19; I2 = 98.10%), stroke(12%; 95%CI 0.08–0.16; I2 = 98.95%), dizziness(10%; 95%CI 0.08–0.13; I2 = 96.45%), vision impairment(6%; 95%CI 0.03–0.09; I2 = 86.82%), intracerebral haemorrhage(5%; 95%CI 0.03–0.09; I2 = 95.60%), seizure(4%; 95%CI 0.02 -0.05; I2 = 98.15%), encephalitis(2%; 95%CI 0.01–0.03; I2 = 90.36%), Guillan-Barré Syndrome (GBS) (1%; 95%CI 0.00–0.03; I2 = 89.48%). Conclusions Neurological symptoms are common and varied in Covid-19 infections, and a growing number of reports suggest that the prevalence of neurological symptoms may be increasing. In the future, the role of COVID-19 neurological symptoms in the progression of COVID-19 should be further studied, and its pathogenesis and assessment methods should be explored, to detect and treat early neurological complications of COVID-19 and reduce mortality.


Introduction
At the end of December 2019, an epidemic of COVID-19 occurred in Wuhan, Hubei Province, China. As the epidemic has spread, cases have been found in many countries. At the time of writing, it has spread to more than 200 countries and regions, with a total of 187,637,579 confirmed cases and 4,066,292 death. Coronavirus has caused two fatal outbreaks in the past, the first in China in 2003, with a 10 percent mortality rate; The second outbreak was Middle East Respiratory Syndrome (MERS) in Saudi Arabia in 2012, with a mortality rate of 35% [1,2]. Novel Coronavirus was named severe acute respiratory syndrome coronavirus Type 2 (SARS-COV-2). Sars-Cov-2 is a single-stranded RNA coronavirus, belonging to the same -coronavirus branch as SARS-COV and MERS-CoV. Its genome sequence is 89.1% similar to the nucleotide sequence of a group of SARS-like coronaviruses. Studies have found that the viral structure and receptor binding domain of SARS-COV-2 is similar to that of SARS-COV, and both of them enter human cells by binding to ACE2 receptors [3,4]. To evaluate the genetic variation of SARS-COV-2, 86 complete or nearly complete genomes were genetically analyzed, and many mutations and deletions were found in both coding and non-coding regions [5]. At present, the biological characteristics and genetic variation of SARS-COV-2 are still not very clear and deserve further study.
Since nervous system manifestations are an important part of COVID-19 patients' clinical manifestations, early detection and treatment of nerve injury in COVID-19 patients is of great significance for their prognosis and reduction of neurological sequelae. This study aims to further summarize the incidence and characteristics of various nervous system manifestations in COVID-19 patients, and explore the mechanism and prognosis of nervous system injury in combination with nervous system manifestations, so as to provide positive clinical significance for better epidemic prevention and control and clinical diagnosis and treatment in the future.

Inclusion and exclusion criteria
The inclusion criteria for this study are as follows: all English literature from 1 January 2020 solstices 30 April 2021 that has been published reporting neurological manifestations of COVID-19 patients. Studying types include case-control studies, cohort studies, cross-sectional studies, and case series. Only those subjects who were diagnosed with SARS-COV-2 infection by realtime reverse transcription-polymerase chain reaction or by high-throughput sequencing of swab samples were included in the study. Only those studies whose results included specific neurological manifestations were included.

Statistical analysis
Our study synthesizes the results of several similar studies to provide a quantitative summary. The extracted data details and instructions are presented in tables and figures. We used Stata (version 15.1) for the meta-analysis. I 2 was calculated to assess the level of heterogeneity and can be divided into four categories: maybe unimportant (0-40%), may represent moderate heterogeneity (30-60%), may represent significant heterogeneity (50-90%) and significant heterogeneity (75-100%). In this study, a fixed-effect model or random-effect model was used to calculate the comprehensive prevalence rate and 95% confidence interval. When there was no heterogeneity, the fixed effect model was selected; when there was heterogeneity, a random effect model was selected. Forest maps are used to visually indicate the magnitude of the effects that the study summarizes and their 95% confidence intervals. Funnel plot and Egger's test were used to assess the publication bias of all literature, and P < 0.05 was considered statistically significant.

Results
A total of 8812 studies were selected from the database, including 5789 in Pubmed, 2971 in Embase, and 52 in the Cochrane Library. After eliminating duplicate studies (n = 1916), 6896 studies met the preliminary screening criteria based on titles and abstracts. After excluding review, mata analysis, case report, age < 18, animal experiment and literature with inconsistent research contents (n = 6483), the remaining 413 articles were screened according to the inclusion and exclusion criteria. After the full-text screening, 168 articles (N = 292,693) were included in the meta-analysis. Figure 1                              shows the details and characteristics of the 16 included studies divided into severe and non-severe groups. Included studies had good methodological quality (Figs. 2, 3, 4, 5).
In addition, COVID-19 involvement in the peripheral nervous system includes other manifestations, such as cranial neuropathy and neuromuscular joint disease, etc. However, since only a few studies have reported such cases, and most of these studies were case reports, these less frequent COVID-19 peripheral nervous system manifestations were not included in this study. In the future, as more and more of these studies are reported, we can conduct further research.
Among 168 studies that were finally included in the meta-analysis, 123 studies discussed the incidence of headache, 106 studies discussed smell impairment, 86 studies discussed myalgia, and 80 studies discussed taste dysfunction. In our opinion, this also indirectly indicates that COVID-19 patients have symptoms of headache, smell impairment, myalgia and smell impairment earlier and have more cases, which is worthy of further study and discussion, to provide a diagnosis and treatment direction for early intervention in the future.
In 16 of the included studies, patients were divided into severe and non-severe groups for detailed analysis according to the severity of COVID-19 infection in patients. We found that the most studied neurological manifestations in this literature were headache, dizziness, seizure, myalgia, smell impairment and taste dysfunction. Our results showed that there was no significant difference in the incidence of headache between the two groups, both at 16%; the incidence of dizziness and seizure in the severe COVID-19 group was higher than that in the non-severe group (12% VS 9%, 3% VS 1%, respectively). The incidence of myalgia, smell impairment and taste dysfunction in the severe COVID-19 group was lower than that in the non-severe group, which was 21% VS 24%, 8% VS 13%, 9% VS 14%, respectively. However, considering that the severity of COVID-19 disease is not classified the same according to different prevalence periods of COVID-19 and its prevalence in different countries, our statistical data may be biased to a certain extent, and the statistical results may lack a certain scientific nature. At the same time, there are still limited studies on various neurologic characteristics of COVID-19 under different severity, which deserves more research and exploration.
We also found that altered mental status, encephalopathy, and alteration of consciousness were reported separately, but they may be related to each other. Through reading different literature, we found that different researchers have different definitions of the above three aspects. Josef Anrather et al. [181] argue that altered mental status (e.g., confusion, disorientation, emotional restlessness and lethargy) can be collectively referred to as encephalopathy. Emad Nader Eskandar et al. [130] suggested that if there was evidence of cognitive impairment (e.g., confusion, disorientation, agitation or delirium) or impaired arousal (e.g., lethargy or dullness), the patient can be included in the altered mental status cohort. In another study, altered mental status was defined as including personality, behavior, cognition, and consciousness changes; encephalopathy, encephalitis; catatonia, mania, anxiety or depression, etc. [131]. We recognize that there may be overlapped parts among altered mental status, encephalopathy and alteration of consciousness, but they can all indicate that the central nervous system of patients is damaged, so their clinical reference value is not affected.

Associations between neurologic manifestations and mechanisms of nervous system injury in patients with COVID-19
ACE2 has been proved to be a functional receptor of SARS-CoV-2, which binds to the ACE2 receptor through its Spike (S) protein C-terminal domain (CTD) [182]. The expression profile of ACE2 is very extensive, and it is expressed in various regions of the human brain, such as the ventricle, motor cortex and posterior cingulum gyrus, middle temporal gyrus, substantial nigra, olfactory bulb, contralateral medulla oblongata, nucleus solitaries, vagus nerve, neurons, astrocytes, microglia and oligodendroglia, etc. [183,184]. Therefore, the nervous system is at risk of SARS-COV-2 infection. The mechanism of nervous system injury in COVID-19 patients includes cross-neuronal hypothesis, homogenous transmission and BBB transmission, hypoxia, inflammatory response and hypercoagulability and immune mechanism as well.
In the following sections, we will analyze and discuss the relationship between the neurological manifestations we have studied and the corresponding mechanisms of neurological damage.

Altered mental status and alteration of consciousness
The altered mental status and alteration of consciousness of COVID-19 patients may be a systemic consequence of the over-activation of the body's immune response or may be caused by the direct invasion of the nerves by SARS-COV2 [185,186]. Hypoxia, organ dysfunction, the need for large doses of sedatives, and prolonged isolation may also contribute to consciousness changes [181].

Headache and dizziness
Headaches and dizziness can be caused by a variety of causes. One reason may be the large and rapid increase of inflammatory cytokines, including IL-1, IL-6, and TNFα, after infection. It can also cause pain if the virus invades the nervous system directly and damages the nerve. Hypoxia leads to the accumulation of acid in brain cells, swelling, and interstitial edema, resulting in cerebral vasodilation or cerebral blood flow obstruction, etc., which is also considered as one of the potential mechanisms of causing headache [185].

Myalgia
Because there are ACE2 receptors in skeletal muscle [187], direct cytotoxicity caused by the interaction between SARS-CoV-2 and ACE2 in skeletal muscle should be considered. Secondly, muscle injury may be related to the harmful immune response mediated by infection and the increase of pro-inflammatory cytokines [8]. Elevated concentrations of C-reactive protein and D-dimer in the blood may induce the expression of a strong immune response, leading to direct immune-mediated nerve and muscle injury [188]. At the same time, prolonged mechanical ventilation or the use of neuromuscular blockers in patients with severe COVID-19 infection is also closely related to the progression of critical myopathy [189].

Smell impairment and taste dysfunction
Studies have shown that the anosmia of most COVID-19 patients has nothing to do with rhinorrhea or nasal obstruction, and may be related to the transmission of the virus through olfactory nerve epithelial cells and further invasion of the olfactory bulb and central nervous system [43,190]. It has been reported that higher levels of angiotensin can induce the apoptosis effect of neural stem cells [191], reducing the number of nerve cells migrating to the olfactory bulb, resulting in reduced replacement of  19:363 new neurons in the olfactory bulb, and interference with the smell. Animal experiments have shown that SARS-CoV-2 may invade the brain retrograde along with taste and trigeminal nerve pathways in the early stage of infection, causing lesions such as taste disturbance [192].

Encephalopathy and encephalitis
Patients with COVID-19 may develop many types of encephalopathy. The combination of SARS-CoV-2 virus and ACE2 receptor increases peripheral vascular resistance, leading to a significant and rapid increase in blood pressure, triggering a cascade reaction that leads to the destruction of blood-brain barrier integrity, cerebral hyperperfusion, and cerebral edema, leading to the development of hypertensive encephalopathy [193,194]. Hypoxia, insufficient energy supply, or overall insufficiency after severe brain injury can progress to hypoxicischemic encephalopathy. Steatosis, liver function damage, or cardiovascular dysfunction after COVID-19 infection may induce the occurrence of hepatic encephalopathy, which may be related to direct liver damage of SARS-CoV-2 or the side effects of some drugs [195,196].
In addition, recent studies have found an increased incidence of acute renal damage after SARS-CoV-2, and the pathophysiological mechanism remains unclear, which may also be caused by an inflammatory response, ACE/ ACE2 imbalance, or dysfunction of other organs (such as the heart). When patients have renal insufficiency, the body cannot completely discharge toxins and regulate the concentration of cytokines normally, and the brain homeostasis will be destroyed, leading to uremic encephalopathy [197]. When SARS-CoV-2 enters the human body through ACE2, the human body immediately generates an immune response. When the virus invades the brain, it will cause immune damage and lead to the attack of encephalitis [198]. Encephalitis may also be associated with blood-brain barrier disruption and cytokine surges [199].

Stroke and intracerebral hemorrhage
SARS-CoV-2 directly invades the nerve through ACE2 receptors and deregulation of blood pressure is the potential mechanism of COVID-19 stroke [200,201]. Continuous hypoxia in COVID-19 patients will eventually lead to disorders in neurotransmitter metabolism and mitochondrial failure, causing irreversible nerve damage and increasing the risk of stroke [202]. Studies have shown that increased inflammatory markers such as IL-2, IL-6, macrophage inflammatory protein 1-α or coagulation dysfunction in patients with severe COVID-19 may increase the likelihood of stroke compared with moderate patients [200]. After the SARS-CoV-2 virus binds to the ACE2 receptor, it can lead to increased blood pressure, which may lead to hypertensive crisis in severe cases and increase the risk of intracranial hemorrhage [188].

Seizure
The process of SARS-CoV-2 virus replication in host cells disrupts neuronal function and manifests as seizures, convulsions, and loss of consciousness [203]. The increase of pro-inflammatory mediators may contribute to epileptic seizures, which in turn enhance the production of cytokines such as IL-1B and TNFα [204]. After the destruction of the blood-brain barrier, inflammatory factors and viral particles flood into the central nervous system, leading to brain damage that can lead to epileptic seizures [188].

Guillan-Barré Syndrome (GBS)
Guillain-Barre syndrome (GBS) is an increasing occurrence in patients with COVID-19, a typical viral autoimmune nervous system disease caused by a strong immune response, which includes a range of polyneuropathy characterized by acute motor weakness, mild to moderate paresthesia, cranial nerve involvement, and muscle or nerve root pain [205]. At present, large sample studies on GBS are still lacking, and most researchers express their views in the form of case reports.

Biomarkers reflecting nervous system damage in COVID-19 patients
A prospective study showed that total Tau, GFAP, and NFL protein levels in cerebrospinal fluid were elevated in 63, 37, and 16% of patients, respectively, and NFL protein was associated with disease severity, duration of intensive care, and level of consciousness [206]. Multiple studies on COVID-19 patients with ischemic stroke have shown that the neutrophil-lymphocyte ratio (NLR) is increased in 90% of the patients [207], CRP is increased in over 90% of the patients [208], and serum ferritin is also increased [209], in which serum ferritin can also predict the degree of nerve injury in patients with acute ischemic stroke [210]. There was a significant correlation between the decrease of interleukin-6 level and the improvement of olfaction and taste function in COVID-19 patients [211]. Elevated serum nerve filament light chain (SNFL) levels in critically ill patients with COVID-19 are closely associated with poor prognosis [212].

Examination methods reflecting nervous system damage in COVID-19 patients
In our study, stroke incidence was 12%, intracerebral hemorrhage 5%, encephalopathy 26% and encephalitis 2%. Studies have shown that patients with COVID-19 with acute neuroimaging abnormalities are more likely to have an ischemic stroke. Neuroimaging features of these patients are not invariable but are predominated by acute ischemic infarction, intracranial hemorrhage, and leukoencephalopathy [81]. COVID-19 is an independent risk factor for acute ischemic stroke and a valid indicator of poor prognosis. Meningitis and encephalitis are not very common [213,214]. In one study, continuous EEG monitoring was performed in 11 of 16 patients with COVID-19, most of whom presented with nonspecific encephalopathy [215]. At the same time, some studies also found significant structural changes in the olfactory nerve, olfactory bulb, olfactory cortex and other olfactory pathways in MRI examination of COVID-19 patients, suggesting that SARS-CoV-2 may enter the central nervous system through the olfactory bulb mediated crossneuronal pathway [94,95]. The relationship between COVID-19 imaging changes and neurological symptoms requires more research.

Prognosis of nervous system injury in COVID-19 patients
Nervous system damage is closely associated with the mortality of SARS-COV-2 infection, and whether the neurological symptoms are reversible is not yet clear. Patients with COVID-19 who require ICU admission for neurological problems or develop neurological dysfunction in the ICU have significantly increased mortality [216].
In one animal study, 4 out of 14 infected mice developed significant respiratory distress and neurological symptoms 2 days after infection, and only the mice showing neurological symptoms died, suggesting that neurological involvement may be a cause of death [217]. Some studies have confirmed that recovery from acute SARS-CoV-2 infection does not completely clear the virus, and has been found to have a higher potential risk for long-term residual neuropsychiatric and neuroscientist impairments, including depression, obsessive-compulsive disorder, psychosis, Parkinson's disease and Alzheimer's disease [218]. An experimental animal study has also shown that coronavirus can persist in the central nervous system of its host [219]. The study results of Helms et al. showed that 36% of severe COVID-19 survivors developed Dysfunction Syndrome [220], and some studies also reported that patients with severe dysfunction after the acute phase had significant recovery after active neurological rehabilitation [221]. In a prospective study, 68.33% of patients developed neurological symptoms during infection with SARS-CoV-2, and 50% recovered 3 months after infection but still had neurological symptoms [213]. Imaging results of 60 patients after recovery of COVID-19 neurological symptoms suggest that the microstructure and functional brain integrity of the brain may be damaged during the rehabilitation phase, which may require long-term neurological and neuroimaging follow-up [214].

Advantages and disadvantages of this study
Our meta-analysis included 168 articles (n = 292,693) detailing the various neurologic symptoms common in COVID-19 patients and providing a comprehensive view of the neurologic symptoms of COVID-19. The results were comprehensive. We started the discussion of the clinical manifestations and analyzed the evidence, possible mechanism and prognosis of the nervous system injury caused by SARS-CoV-2. We also discussed the biomarkers and examination methods of nervous system injury caused by SARS-COV-2, providing some valuable suggestions for early identification, monitoring, screening, diagnosis and follow-up of nervous system injury and poor prognosis in patients with COVID-19 and potential targets for future clinical intervention strategies. This study also has some limitations. Firstly, most of the included literature was retrospective studies, which may cause some potential bias. Secondly, this study failed to provide an analysis of the correlation between various neurological manifestations and disease severity and mortality. The question of which neurological manifestations are the most insidious and which are the most difficult to recover from is currently unanswered, and more findings will be needed in the future. Finally, the high degree of heterogeneity in our study may be due to differences in patient race selection, disease severity, comorbidities, only a few studies specifically assessed neurological symptoms, differences in the number of patients in different studies, or differences in publication bias and study methodology.

Conclusion
In conclusion, our study suggests that nervous system expression in COVID-19 is diverse and pervasive but easily underestimated. Therefore, the long-term pathophysiological results of SARS-COV-2 neuropathy should be of serious concern to us. At the same time, summarizing the entire symptom spectrum, biomarkers and examination methods of the disease is conducive to predicting the severity of neurological impairment and providing better suggestions for preventing misdiagnosis, early diagnosis, preventing disease transmission, early intervention treatment, evaluation of therapeutic effect and follow-up. In the future, more clinical and experimental studies should be carried out to provide strong evidence to support clinical practice, and further, explore the role of nervous system symptoms in the incidence of COVID-19 and its possible mechanism, to reduce mortality, reduce disease severity, control disease progression and prevent possible long-term nervous system complications.