Introduction

The world at present is in combat with COVID-19, appropriately termed by Bill Gates as a “Once-in-a-Century” Pandemic Gates (2020). The dual challenge posed by this pandemic, namely saving lives as well as preventing horizontal transmission, has stretched clinicians to the extreme. With the passage of time, the situation has become more complicated given the unusual clinical manifestation of this viral infection. Several extra-pulmonary manifestations of COVID-19 have been brought to attention by treating physicians worldwide. Among these, neuro-invasive potential of SARS-CoV-2 has received significant attention. There is now sufficient body of evidence to support the idea that COVID-19 can have pure neurological presentations and on some occasions, preceding the typical respiratory manifestations (Zubair et al. 2019; Koralnik and Kenneth 2020; Lahiri et al. 2020a). Besides, a myriad of neurological consequences following typical clinical presentation of COVID-19 has also been documented across the globe. In sum, neuro COVID as a distinct topic of discussion has been steadily gaining attention over the last couple of months.

Among the central nervous system (CNS) manifestations, impaired consciousness/encephalopathy is a widely reported symptom of COVID-19 Zubair et al. (2019); Encephalopathy and in patients with COVID-19: ‘Causality or coincidence’ Journal of Medical Virology 2020d and various possible aetiologies underlying this symptom have been scrutinized in literature. It has been widely speculated that the virus gains access into the brain via olfactory bulb and the hypothesis gained substantial support from the observation that anosmia is a fairly consistent symptom of early COVID-19 (Zubair et al. 2019). An alternative proposed route for the virus to invade brain is hematogenous (Zubair et al. 2019). Given the potential of SARS-CoV-2 to invade CNS and also our previous experience with MERS and SARS-CoV, meningoencephalitis in COVID-19 is a duly anticipated clinical feature. Indeed, in the last couple of months, multiple reports of meningoencephalitis associated with COVID-19 have surfaced up. These cases not only have encompassed a wide range of clinical presentations but also have documented varied laboratory and imaging results. That said, it can be assumed that in current situation under-reporting of cases is a non-negligible issue, and therefore, these cases only represent tip of the iceberg. Nevertheless, an organized summary and critical review of these documents can reveal a wealth of information about the clinical, laboratory, and imaging features of meningoencephalitis in COVID-19.

In this background, we set out to develop a systematic review of meningoencephalitis cases in COVID-19 available in a wide array of databases. Our main objective is to summarize the clinical presentations, laboratory parameters including CSF abnormalities, and brain imaging features of SARS-CoV-2-associated meningoencephalitis. Such a documentation will not only work as a guide to clinicians dealing with COVID-19 patients but also will open up new avenues towards understanding the neuro-invasive potential and routes of this virus.

Methodology

Design

This systematic review was conducted by following the Preferred Reporting for Systematic Review and Meta-Analysis (PRISMA) consensus statement (CRD42020185571) (Lahiri et al. 2020b). Studies relevant to the confirmed cases of COVID-19 infection with confirmed or suspected association of meningoencephalitis were included.

Search strategy

In this systemic review four databases, Pubmed/Medline, NIH LitCovid, Embase, and Cochrane Library, were searched using pre specified searching strategies and this search was concluded on 10 June 2020. The search strategy consists of variation of keywords of relevant medical subject headings (MeSH) and key words, including “SARS-CoV-2,” “COVID-19,” “coronavirus,” “clinical symptoms,” “neurological impairments,” and “Meningoencephalitis.” Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV) and Middle East Respiratory Syndrome Coronavirus (MERS-CoV) were also included in our search strategy to capture related articles. We also hand searched additional COVID-19 specific articles using the reference list of the selected studies, relevant journal websites, and renowned pre-print servers (medRxiv, bioRxiv, pre-preints.org) from 2019 to current date for literature inclusion. To decrease publication bias, we invigilated the references of all studies potentially missed in electrical search. Content experts also searched the grey literature of any relevant articles.

Study selection criteria

All peer-reviewed, pre-print (not-peer-reviewed) including cohort, case–control studies and case reports which met the pre-specified inclusion and exclusion criteria were included in this study. The PRISMA flow chart for study selection for this current review has been depicted in Fig. 1.

Fig. 1
figure 1

The PRISMA flow chart for study selection

Inclusion criteria

Studies met the following inclusion criteria were included if (i) conducted for the patients infected with COVID-19 with or suspected meningoencephalitis; (ii) studies registering neurological manifestations of COVID-19 patients were included with encephalitis like symptoms; (iii) parallel studies to look into the detailed distribution and incidences of meningoencephalitis in previous outbreaks, i.e. SARS-CoV, MERS-CoV, and various other coronaviruses, were compared with current pandemic in the discussion section; and (iv) published in the English language. Studies without complete information but met our inclusion criteria were included in the narrative review.

Exclusion criteria

Studies excluded if COVID-19 was not confirmed among patients and written in languages other than English. We also excluded review papers, viewpoints, commentaries, and studies where no information related to neurological symptoms or meningoencephalitis was reported.

Data extraction

Prior to the screening process, team of three reviewers (GS, DB, and SD) participated in calibration and screening exercises. First two reviewers (GS and DB) subsequently screened independently the titles and abstracts of all identified citations, and the third reviewer (SD) verified those citations and screened papers by (GS and DB). Other two reviewers (UG and SA) then retrieved and screened independently the full texts of all citations deemed eligible by the reviewer (SD) on the team and analysed those data. Another reviewer (RM) independently verified these extracted full texts for eligibility towards analysis and designed the overall study structure. The corresponding author (DL) had resolved disagreements whenever necessary and took final decisions regarding the study. Throughout the screening and data extraction process, the reviewers used piloted forms. In addition to the relevant clinical data, the reviewers also extracted data on the following characteristics: study characteristics (i.e. study identifier, study design, setting, timeframe), population characteristics, comparator characteristics, outcomes (qualitative and/or quantitative), clinical factors (definition and measurement methods), reported funding sources and conflict of interests; and study limitations. The Newcastle–Ottawa scale was used to assess the selection procedure, the comparability, and the outcomes of each reviewed study.

Statistical analysis

Quantitative data were presented using means ± standard deviation (SDs). Qualitative data were presented as percentage value. Unit discordance among the variables was resolved by converting the variables to a standard unit of measurement. Further, one sample t test was performed to find out significant difference between mean value and respective reference value (highest value of the reference range was taken) of each parameter. A value of “p” < 0.05 was considered as statistically significant. All statistical analyses were performed and analyzed using GraphPad prism (Version 6, San Diego, CA, USA). A meta analysis was planned to analyze the association of the demographic findings, co-morbidities, symptoms, diagnostic parameters, and outcomes with imaging findings but was later omitted due to lack of sufficient data.

Results

Study characteristics (Table 1)

Table 1 Included studies

A total of around 61 cases were included in the study from 25 different articles coming from both 3 database search and preprint servers. Complete data were available for 54 (n = 54) cases, and laboratory findings data was unavailable in 7 cases from Varatharaj et al. (2020). The studies of 54 cases comprised of case reports and original article.

Demography and clinical symptomatology (Tables 2 and 3)

Table 2 Demographic characteristics of patients hospitalized with COVID-19
Table 3 Clinical signs and symptoms

The selected study comprised of 54 cases with the mean age around (50.8 ± 19.09) and 35 (65%) male with 19 (35%) female cases were reported for analysis with full data availability. SARS-CoV-2 was primarily detected in nasopharyngeal/oropharyngeal/bronchoalveolar swab (92.6%), CSF (5.55%), and antibody detected in two patients (3.7%) only. Majority of the reported cases had no travel history (3.7%) or contact with COVID-19 positive individuals (7.4%). They had various symptoms ranging from fever, cough, headache, etc. Potent neurological symptoms reported were tonic–clonic seizures (5.55%), disorientation to time and place (22.22%), nuchal rigidity (9.25%), limb ataxia (11.11%), etc. Various kinds of co-morbidities that existed in patients were diabetes mellitus (18.5%), hypertension (37%), Obesity(3.7%), and coronary artery disease (3.7%) with two special cases of Alzheimer’s disease and autism. The patients were hospitalized on (8.88 ± 4.094) days (mean ± SD) from the onset of the symptoms. Their average days of hospitalization were (13.83 ± 7.901) days (mean ± SD).

Laboratory findings

Hematological parameters (Table 4)

Table 4 Clinical and laboratory diagnostic parameters

Major cell counts were observed within various patients. Recorded WBC counts in cases (n = 13) were (10.12 ± 8.33) × 109 (mean ± SD) with a p = 0.734 and lymphocyte counts observed for six patients (n = 6) were (3.177 ± 3.06) × 109 (mean ± SD). Elevated levels of C-reactive protein (CRP) among patients (n = 13) around (102 ± 129.8; mean ± SD) (mg/L) were reported with a p = 0.064. Moreover, highly significant levels of D-dimer (ng/ml) in patients (n = 10) were observed with a significant p = 0.0186 (p < 0.05).

CSF parameters (Table 4)

Various CSF parameters under biochemical analysis such as glucose (mg/dl) (n = 16) and protein (mg/dl) (n = 18) were recorded where significant levels of protein (p = 0.00457) with elevation of (73.61 ± 56.31) mg/dl from the normal range of (15–45) mg/dl was observed. Moreover, certain cell count parameters such as lymphocyte (n = 6) was significantly (p = 0.009) higher than the normal range. CSF IgG levels (mg/L) were also found to be significantly (p = 0.0001) raised in five patients (n = 5).

EEG findings (Table 5)

Table 5 EEG findings

EEG findings were reported in 9 cases, and in the remaining 20 cases this parameter was not available for analysis. Bernard Valnet et al. reported EEG from one patient as abundant bursts of anterior low-medium voltage irregular spike-and waves superimposed on an irregularly slowed theta background (Bernard-Valnet et al. 2020), whereas Sohal et al. documented that the patient had six left temporal seizures, left temporal sharp waves which were epileptogenic in nature (Sohal and Mossammat 2020). Pilotto et al. showed generalized slowing with decreased reactivity to acoustic stimuli (Pilotto et al. 2020), and bilateral slowed activity without seizures was reported by Chaumont et al. (2020). Generalized slowing with no epileptic discharges was also documented by Duong et al. (2020). McAbee et al. only reported frontal intermittent delta activity (McAbee et al. 2020). On the other hand Adelaide Panariello et al. reported theta activity at 6 Hz, unstable and non reactive to visual stimuli without significant asymmetries (Panariello et al. 2020). Continuous monitoring with amplitude-integrated electroencephalography (EEG) by Charcon Aguilar et al. for 36 h revealed a continuous background pattern with sleep–wake cycles in the absence of electrical and clinical seizures for a patient (Chacón-Aguilar et al. 2020). Interestingly Morassi et al. reported that on day 4 the EEG showed a normal background in the alpha range (8 Hz) associated with recurrent sharp slow waves over the left temporal region, which occasionally were seen also on the right homologous regions. On day 10, a new EEG excluded non-convulsive status epilepticus, while showing persistence of sharp slow waves, mainly over the left hemispheric regions for the same patient (Morassi et al. 2020).

Neuroimaging findings (Table 6)

Table 6 Neuroimaging findings

Neuroimaging findings were reported in 20 studies consisting of 26 cases, and in the remaining 35 cases brain scans were not available. Among the imaging modalities, CT scan and MRI were done in 10 and 16 cases respectively. Vessel imaging was available in 1 case (CTA). In majority of the cases neuroimaging finding did not reveal any abnormality (13/26, 50%). At least one radiological abnormality was reported in 14 cases (14/26, 53.84%). Among the abnormal MRI findings, T2/T2 FLAIR hyperintensity in MRI brain was the most frequently reported (6/16, 37.5%). Noteworthy, temporal lobe was the commonest site of involvement in MRI brain (6 cases). Stroke-like findings manifested by hypodensity on CT scan (3 cases) and diffusion restriction in MRI (2 cases) were also observed. Beaded appearance on CT angiography was reported in 1 case associated with extradural haemorrhage.

Treatment and outcome (Tables 7 and 8)

Table 7 Treatment and outcome
Table 8 Drug dosages with duration

Treatment were carried out with various drugs such as Hydroxycholoroquine (74%), Azithromycin (29.41%), Antiretroviral drugs (Lopinavir/Ritonavir/Darunavir/Cobisitat) (55.55%), and Favipiravir (14.81%) or used in multidrug therapy (81.48%). Various procedures such as Plasmapheresis (11.11%) and invasive ventilation (26%) were used along with many other therapies on the patients. Frequently used multidrug therapy includes Hydroxycholorquine along with Antivirals (57.40%). Majority of the patients recovered (72%) with a mortality of (17%), and six patients were still under treatment during the period of study.

Discussion

In the index paper we attempted to summarize the clinical features, laboratory parameters, EEG findings, and imaging abnormalities in patients presenting with confirmed SARS-CoV-2 infection and clinical/confirmed meningo-encephalitis. Of note, a fair number of cases with meningo-encephalitic presentation of COVID-19 have already surfaced up even in the background of possible under-reporting because of the ongoing pandemic. We found description of 61 such cases in the available literature so far among which 54 were included in the final analysis as detailed data for the remaining 7 cases were not available. To the best of our knowledge, this is the first ever attempt to explore meningoencephalitis in COVID-19 by means of systematic review of both peer reviewed as well as pre-print data. The present study is important in terms of advancing our knowledge and understanding related to the neuro-invasive potential of SARS-CoV-2 which primarily is considered a respiratory pathogen.

Coronaviruses have been detected in both cerebrum and cerebrospinal fluid of individuals with seizures, encephalitis, and encephalomyelitis(Varatharaj et al. 2020).According to different case reports of MERS coronavirus, neurological manifestations such as meningoencephalitis, hyporeflexia, and ataxia have been found (Bohmwald et al. 2018). It was also shown that neuronal infection leads to death in hACE2 (full form) transegnic mice infected with SARS-CoV (Kim et al. 2017 Jul). Accumulative evidences reflect that virus has been detected in cerebrum (Netland et al. 2008) as well as in CSF in case of SARS-CoV infection (Ding et al. 2004). Other coronaviruses such as HCoV-OC43 under the genera of Betacoronaviridae is found to be associated with viral encephalitis (Lau et al. 2004) and acute disseminated encephalomyelitis (Morfopoulou et al. 2020). Similarly, CoV-NL63 under the genera of Alphacoronaviridae causes acute encephalitis with self-limiting CNS infection (Yeh et al. 2004). Viral encephalitis caused by coronavirus infection was reported among paediatric population with respiratory tract infection (Schattner et al. 1032). Overall, the predisposition of various coronaviruses towards neurotropism seems evident and very much happening which is generating a major concern for virus mediated encephalitis.

Analysis of the demographic parameters reveals that mean age of the patients with COVID-19-related meningoencephalitis was (50.8 ± 19.09) years with males being more commonly affected (70%). Mean latency between symptom onset and hospital admission was close to 9 days, while mean number of days of hospital stay was approximately 14 days. Majority of the documented patients had co-morbidities with diabetes mellitus and hypertension being most frequent. Among the general symptoms, fever was by far the commonest followed by cough. Notably, anosmia and ageusia (commonly considered as markers of neuro-invasion) were reported in uncommonly. In contrast, diarrhoea was relatively common (7.40%) among the general symptoms. Respiratory distress was reported in 2 (3.7%) of the documented cases.

Various neurological symptoms were reported in meningoencephalitis associated with SARS-CoV-2 infection. Confusion or disorientation to time and place or altered mental status was the most frequently reported symptom accounting for 22.22% of the cases. It may therefore be assumed that in places reporting high incidence of COVID-19, confusion or disorientation may prompt investigation for detection of SARS-CoV-2. While the classic Kernig’s and Brudzinski’s signs were not frequently documented, nuchal rigidity has been more commonly (9.25%) observed. Limb ataxia was reported in around 11% of the cases including the case reported by one of the authors of the present review. Other less reported neurological symptoms include tonic clonic seizures, slurred speech, and cognitive difficulties such as verbal and motor perseveration. SARS-CoV-2 detection in CSF was reported in only 3 cases, whereas majority (92.6%) of the patients were diagnosed based on nasopharyngeal swab/oropharyngeal swab/bronchoalveolar lavage testing.

Among the clinical signs pertaining to general survey, tachycardia and reduced oxygen saturation were common findings. Analysis of blood parameters reveal raised CRP, D-dimer, and LDH were prominent findings, while occurrence of lymphopenia or leucopenia did not reach statistical significance. CSF parameters often give valuable insight to the treating clinicians when faced with cases of meningoencephalitis. CSF parameters in COVID-19-associated meningoencephalitis raised protein and increased lymphocyte count as the most notable observations. Therefore, it can be assumed that brain infection in the background of SARS-CoV-2 infection may lead to CSF pleocytosis with lymphocytic predominance accompanied by significant level of rise in CSF protein, both pointing towards intra-thecal inflammation. In addition, while there was documented dip in CSF IgG levels, the IgG index did not deviate much from the normal limits, which may be a reflection of hypoglobulinemia resulting from lymphopenia (Li et al. 2016).

Among the abnormal EEG findings documented in COVID-19 encephalitis, focal epileptiform discharges and generalized slowing are most notable. While focal discharges may signify the occurrence of focal convulsions, generalized slowing is commensurate with the high frequency of altered mental status observed thus far. In short, EEG findings in COVID-19-related meningoencephalitis are mostly non-specific and may indicate ensuing encephalopathy in majority of the cases.

Majority of the patient’s brain scans were found devoid of any abnormality. However, among those with abnormal brain scans, a wide range of structural lesions have been observed. The commonest abnormality reported was T2/FLAIR hyper-intensity affecting the temporal lobes. Among the viral causes of temporal lobar hyperintensity, HSV is the forerunner followed by CMV, while anti-NMDAR related autoimmune encephalitis frequently presents temporal lobar changes (Prasad, 1982; Sureka & Jakkani, 2012). In addition, stroke-like presentations have also been noted which are characterized by restricted diffusion in DWI.

It must also be admitted that at least 2 cases of encephalitis compiled in the present series refer to underlying autoimmune pathogenesis mechanisms. While the case reported by Panariello et al. (2020) has mentioned about the detection of anti NMDAR antibody in association with encephalitis, Pilotto et al. (2020) has described a steroid responsive encephalitis again pointing towards immune pathogenesis. Therefore, it may be speculated that several cases of meningoencephalitis related to COVID-19 may not actually reflect direct viral invasion of CNS. Post-/para-infectious immune pathologies might come into play in some of the clinical presentations. As the reporting of cases would start rising, similar reports are supposed to surface up more frequently.

As far as the treatment is concerned, more than 74% patients have received therapy with HCQ, while significant proportion of patients received azithromycin, favipiravir, and acyclovir. In fact, almost all the patients were on multi-drug regimen as was documented in this review. Reported mortality rate in COVID-19 encephalitis is 17%. Majority (72%) of the affected patients experienced recovery and around 1/9th were still under treatment at the time of drafting this review.

Conclusion

In this systematic review we observe that quite a few reports of meningoencephalitis associated with COVID-19 are already available in recent literature even in the face of possible under-reporting in contrast to the earlier epidemics of MERS and SARS-CoV. Various clinical presentations have surfaced up among which confusion or disorientation is the most frequent. Among the laboratory parameters, raised CRP and D-dimer are most prominent findings. Analysis of CSF parameters reveals that high protein and lymphocytic pleocytosis are the most observable abnormalities. Focal epileptiform discharges and slowing of background are common EEG abnormalities observed in COVID-19-related meningoencephalitis. A myriad of radiological abnormalities have been observed of which temporal lobar hyper-intensity and diffusion restriction are common. This review has attempted to provide a basic outline of COVID-19-related meningoencephalitis from a clinician’s perspective. Hopefully with increment in reporting of neurological manifestations of COVID-19, a clearer picture will become available towards the future.