Neurological manifestations in patients with COVID‐19: A systematic review and meta‐analysis

Abstract Introduction The intensification of coronavirus disease 2019 (COVID‐19) complications, severe symptoms, and high mortality rate has led researchers to focus on this significant issue. While respiratory and cardiac complications have been described as high‐risk manifestations in patients with COVID‐19, neurological complications can also enhance mortality. This study aimed to evaluate the prevalence of neurological complications arises from SARS‐CoV‐2 and assess the mortality rate from neurological complications. Material and Methods Literature review was conducted by searching in PubMed/Medline, Web of Sciences, and Embase. After performing search strategies with relevant terms, a number of articles were excluded, including review articles, systematic review or meta‐analysis, duplicate publication of same researchers, congress abstracts, animal studies, case reports, case series, and articles reporting a history of neurological features prior to COVID‐19 infection. After retrieving the data, statistical analysis was performed using the STATA Version 14 software. Results From 4455 retrieved publications, 20 articles were selected for further analysis. Among 18,258 included patients, 2791 showed neurological symptoms, which were classified into different groups. Headache, confusion, and fatigue were reported as the most non‐specific neurological features in confirmed COVID‐19 patients. Psychiatric symptoms, CNS disorders, cerebrovascular disorders, CNS inflammatory disorders, PNS disorders, neuromuscular disorders, etc., were defined as specific neurological manifestations. The pooled prevalence of neurological manifestations and mortality rate of COVID‐19 patients with neurological features were estimated to be 23.0% (95% CI: 17.8–29.2) and 29.1% (95% CI: 20.3–39.8), respectively. Conclusion Neurological manifestations may commonly happen in patients with COVID‐19. This study reported a high prevalence of neurological complications and mortality rates in COVID‐19 patients. Therefore, patients with COVID‐19 who indicated neurological symptoms should be taken seriously and should receive early treatment to prevent undesirable events.


| INTRODUC TI ON
The coronavirus disease 2019 (COVID- 19), an infection induced by severe acute respiratory syndrome coronavirus 2 (SARS-COV-2), originated in Wuhan, China in December 2019. This disease spread rapidly around the world and caused the deaths of millions of people. 1 In 2002 and 2012, the outbreaks of other members of this betacoronavirus, namely SARS-CoV and MERS-CoV, were, respectively, recorded. 2 According to the WHO report, more than 100 million cases were infected with SARS-CoV-2, and about four million deaths from COVID-19 were reported.

Conforming to the Centers for Disease Control and Prevention
(CDC), COVID-19 is accompanied by three main symptoms, including common (fever, dry cough, headache, and fatigue), less common (pain, sore throat, diarrhea, and loss of taste and appetite), and severe (shortness of breath, chest pain, and dysfunction in speech).
The disease is not limited to the respiratory system, but it encompasses a wide range of cardiovascular and neurological complications. 3 Search of genomes and coronavirus-related sequences in the cerebrospinal fluid (CSF) verifies this assumption. 4 The intensified complications, severe symptoms, and high mortality rate of COVID-19 have forced special attention to this disease. While respiratory and cardiac complications are considered high risks in patients with COVID-19, neurological complications have also been demonstrated to enhance mortality rate. 5 SARS-CoV-2, the same as SARS-CoV and MERS-CoV viruses, has several routes of entry to central nervous system (CNS) and peripheral nervous system (PNS), including olfactory pathway and gut-brain axis as the neural pathways 6 and blood-brain barrier (BBB) as hematogenous pathway. 7,8 These accessibilities are provided by the expansion and expression of angiotensin-converting enzyme 2 (ACE2) on nerve tissues and capillary endothelial tissues, 9 which are also present in brain endothelial cells and small intestine. 10 This notion corroborates earlier investigations that stated reduced RNA tracking in CSF may be the indication of direct SARS-CoV-2 invasion to nervous tissue in the cell-to-cell spread pattern. 11,12 However, inaccessibility to virus in the normal range of detection via the analysis of CSF samples and also the presence of hemoproteins product may relate to the lack of identifying SARS-CoV-2 in these samples. 13 The first step in brain damage caused by the SARS-COV-2 is triggering the inflammatory response of cytokines (e.g., IL-6, IL-12, and TNFα) after binding the virus to ACE2 and its proliferation in the CNS. 5,10 Neurological complications that affect CNS, PNS, and musculoskeletal system 8,13 entail impaired consciousness, stroke, anosmia, ataxia, and seizures, acute necrotizing encephalopathy, meningoencephalitis, and Guillain-Barre syndrome (GBS), recognized as PNS-related symptoms, as well as Miller Fisher syndrome and hypogeusia that happen as a result of SARS-CoV-2 entry to peripheral neurons. 14 Collectively, the direct invasion of virus, immune-induced complications, and cytokine storms are the main pathogenesis of the SARS-CoV-2. It is surmised that the immune system role is central to the development of neurological complications caused by this virus. 15,16 The present study aimed to evaluate the prevalence of neurological manifestations and mortality rate in patients with COVID-19.

| MATERIAL S AND ME THODS
The present systematic review was conducted by relying on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statements. 17

| Search strategy
Our systematic search was performed using databases viz PubMed/ Medline, Web of Sciences, and Embase from January 1, 2019 to March 7, 2021. The terms employed in the search strategies were as follows: "Central nervous system" or "brain" or "neurologic" or "CNS "and "COVID-19" or "severe acute respiratory syndrome coronavirus 2" or "novel coronavirus" or "SARS-CoV-2" or "SARS2" or"2019-nCoV" or "nCoV disease" or "coronavirus disease-19" or "coronavirus disease 2019" or "2019 novel coronavirus" or "Wuhan coronavirus" or "Wuhan seafood market pneumonia virus" or "Wuhan pneumonia". There were no language restrictions; however, for non-English articles, online translation systems were utilized. Besides, no limitations were considered for the location and type of articles.

| Inclusion and exclusion criteria
neurological features and were positive for SARS-CoV-2. Reverse transcription-polymerase chain reaction (RT-PCR) was the main diagnosis method of confirming COVID-19, but multiple studies added serological tests and computed tomography (CT) scan, as well.
Regarding nervous system involvement, we took into consideration all reported neurological findings, comprising CNS and PNS symptoms, CNS inflammatory features, and cerebrovascular attributes, as well as neuromuscular, psychiatric, and non-specific symptoms.
All the methods used for the diagnosis of neurological symptoms were viewed as qualified. The exclusion of articles was based on any of the following papers: Review articles, systematic review or meta-analysis, duplicate publication of same researchers, congress abstracts, and articles reporting a history of neurological features prior to COVID-19 infection. Likewise, animal studies, case reports, and case series were excluded.

| Data extraction and quality assessment
The data retrieved for each article were as follows: first author's name, location, publication time, sample size, number of COVID-19 patients, gender, median age, COVID-19 diagnosis method, neurological manifestations, diagnosis method of neurological findings, and number of deaths. Four authors extracted the information from the full-text of 20 selected studies, independently. Inconsistencies between reviewers were resolved by consultation. Table 1 shows the extracted data.

| Data synthesis and analysis
Analysis of data was performed by using STATA (version 14, IC; Stata Corporation, College Station, TX, USA), and the prevalence of neurological complications was estimated with confidence intervals (CIs) of 95%. The pooled frequency was calculated by a random effect model with 95% CI. Publication bias was also assessed by Begg's and Egger's tests, and p value <0.05 was considered an indication of statistically significant publication bias.

| RE SULTS
A total of 4455 studies were collected from three databases. After the removal of duplicates, articles were screened for title and abstract, and 119 papers met the criteria. Thereafter, the full-text of these studies, which all reported the neurological symptoms of COVID-19 infection, were evaluated. Following the exclusion of irrelevant studies, only 20 articles were identified as qualified for final extraction and analysis ( Figure 1). These papers were appropriate for systematic review and entered into data extraction. Eighteen of these clinical studies used various methods, such as brain CT, MRI, EGG, and CSF for the diagnosis of neurological manifestations . 8,10,11,18-32 However, two other studies performed only clinical neurological exams. 4,33 In all 20 studies, RT-PCR was examined using nasopharyngeal swab, but nine 8,10,18,[23][24][25][26]29,30 and three 22,27,28 studies employed chest CT scan and ELISA, respectively, to confirm COVID-19. Ten articles were originated from Europe, five from the USA, two from China, and others from Brazil, Pakistan, Egypt, and Chile. The information of these 20 studies is summarized in Table 1.
A variety of neurological findings were reviewed and classified into different groups ( Table 2). Among 20 articles, 18,258 cases were found by RT-PCR test to be positive for SARS-CoV-2 (COVID- 19), and 2791 of patients had various neurological manifestations. The overall frequency of neurological symptoms in COVID-19 patients was 15.28%. Additionally, 12 articles reported mortality. Among the 2791 patients, 582 succumbed in hospital; therefore, the total mortality rate was 20.85%.

| Non-specific neurological features in confirmed COVID-19 patients
Since the emerging COVID-19 infection, a noticeable large number of studies have focused their attention on this outstanding issue; thus, various symptoms and signs have been reported for this disease.
Based on the literature, headache, 34 confusion, 35 and fatigue 36,37 are the most common symptoms. In addition, altered mental status is realized as a common initial presentation in COVID-19 patients. 38 Our results are in line with these findings. Among 20 included studies we reviewed, 13, 5, 3, 1, 6, and 6 studies reported headache, dizziness, weakness, fatigue, confusion, and altered mental status, respectively, as the most frequent symptoms. Five studies reported the observation of neuropathy features in 87 of 977 examined patients. 11,19,23,26,29 Two articles stated 45.76% (81/177) with sleep disorder, 8,33 and one of these studies mentioned daytime sleepiness in 36.17% (34/94). 33 The relative frequency of reduced reflex symptoms was 50% (2/4), standing as the most common symptoms.
However, it could not be considered a notable manifestation due to the inadequate number of cases. 27 Hemiparesis and hemiplegia were detected in 14.81% (4/27), as reported by one study. 25

| Characteristics of PNS and psychiatric symptoms in COVID-19 patients
Nine studies reported 160 anosmia/hyposmia cases out of 768  26 and psychomotor agitation 23 were also reported by three articles.

| Mortality of COVID-19 patients who developed nervous system symptoms
In total, 12 studies reported the number of deaths as the outcome.
The pooled mortality of COVID-19 patients with neurological features was estimated to be 29.1% (95% CI: 20.3-39.8). Figure 4 displays the meta-analysis of the overall mortality rate. As indicated in Figure 5, no evidence of publication bias was found (p = 0.54 and p = 0.57 for Begg's rank correlation analysis and Egger's weighted regression analysis, respectively).

| Risk of bias assessment
According to the quality assessment of the 20 included studies, all of them had a low risk of bias. We summarized the results of the critical appraisal (JBI checklist) of included studies in Table S1. There is a direct relationship between headache and inflammation caused by COVID-19 and high levels of IL-6 as a marker of pain. 47 One of the significant findings of the present study was the mortality rate (29.1%) due to neurological complications of COVID-19.

| DISCUSS ION
Other complications of COVID-19 have been described as respiratory and non-respiratory complications, which there is no doubt that respiratory complications have more severe outcome, as stated by Vakili and co-workers. 48 However, non-respiratory complica- cohort study. In-hospital death of patients with COVID-19 is also related to sex (i.e., males), diabetes mellitus, history of chronic pulmonary disease, increasing age, D-dimer, body mass index, sequential organ failure assessment scores, and so forth. The number of neurological manifestations such as stroke and altered mentation (p < 0.05) in COVID-19 patients may be a high-risk factor for hospitalrelated mortality. 31,50 Considering the reduced mortality rate of patients who were admitted in ICU, dexamethasone showed effective outcomes due to its inhibitory activity against immune response, which is essential for the inflammatory conditions. However, these corticosteroids induce neurological manifestations in long-term administration. 51 The mortality rate was higher in patients with the in vitro symptoms of COVID-19 who suffered from neurological complications. This finding agrees with the study conducted by Frontera et al. 19 who highlighted that patients with stroke are more prone to death than those who have stroke but not COVID-19, which is attributed to prolonged hospitalization.