Cytokines and chemokines profile in encephalitis patients: A meta-analysis.

BACKGROUND: Encephalitis is caused by autoimmune or infectious agents marked by brain inflammation. Investigations have reported altered concentrations of the cytokines in encephalitis. This study was conducted to determine the relationship between encephalitis and alterations of cytokine levels in cerebrospinal fluid (CSF) and serum.
METHODS: We found possibly suitable studies by searching PubMed, Embase, Scopus, and Web of Science, systematically from inception to August 2021. 23 articles were included in the meta-analysis. To investigate sources of heterogeneity, subgroup analysis and sensitivity analysis were conducted. The protocol of the study has been registered in PROSPERO with a registration ID of CRD42021289298.
RESULTS: A total of 23 met our eligibility criteria to be included in the meta-analysis. A total of 12 cytokines were included in the meta-analysis of CSF concentration. Moreover, 5 cytokines were also included in the serum/plasma concentration meta-analysis. According to the analyses, patients with encephalitis had higher CSF amounts of IL-6, IL-8, IL-10, CXCL10, and TNF-α than healthy controls. The alteration in the concentration of IL-2, IL-4, IL-17, CCL2, CXCL9, CXCL13, and IFN-γ was not significant. In addition, the serum/plasma levels of the TNF-α were increased in encephalitis patients, but serum/plasma concentration of the IL-6, IL-10, CXCL10, and CXCL13 remained unchanged.
CONCLUSIONS: This meta-analysis provides evidence for higher CSF concentrations of IL-6, IL-8, IL-10, CXCL10, and TNF-α in encephalitis patients compared to controls. The diagnostic and prognostic value of these cytokines and chemokines should be investigated in future studies.

1. Introduction

Encephalitis is defined as inflammation of active brain tissues that can be caused by autoimmune reactions or infections. Viral pathogens are the main culprit behind infectious encephalitis, including tick-borne encephalitis, Japanese encephalitis, and herpes simplex virus (HSV)-caused encephalitis, among others. More than 100 different pathogens and toxins have been identified as agents contributing to encephalitis pathogenesis [1]. More recently, a growing number of studies have investigated biomarkers for diagnosis and therapeutic targeting of autoimmune encephalitides, such as Anti-N-Methyl-D-Aspartate Receptor (NMDAR) [2-5].

Although the known number of etiologies for encephalitis is on the rise, it is still hard to establish the cause of encephalitis cases in clinical practice, and the underlying mechanisms are complex, not fully understood, and vary according to the cause [6-8]. According to Khetsuriani et al., From 1988 to 1997, about 18,680 cases of encephalitis were hospitalized in the United States each year; with the underlying cause for most (59.5%) not being identified [1]. Patients with encephalitis present with diverse neurological symptoms, including localizing and non-localizing symptoms such as altered mental status and seizures, alongside fever, headache, and meningeal signs [9].

Diagnosing encephalitis and its causative agents is a crucial factor in guiding treatment. A late diagnosis or an incorrect one can lead to negative clinical outcomes. The time of diagnosis might affect the prognosis of the disease vastly. Patients who are diagnosed sooner have a much better prognosis than others [10]. It is of utmost importance to have fast, reliable methods of diagnosing encephalitis in patients in order to improve the clinical outcome and prevent the patients’ conditions from getting worse. Currently, the diagnosis of encephalitis is confirmed via neuroimaging methods, mainly brain magnetic resonance imaging (MRI), and measuring cerebrospinal fluid (CSF) biomarkers of central nervous system infection and autoantibodies and interpreting electroencephalograms (EEG) in order to detect specific patterns in brain waves [11, 12]. Some studies have described increased levels of inflammatory biomarkers, such as cytokines and chemokines, in patients with encephalitis [2, 4, 13]. Cytokines and chemokines are essential agents contributing to the regulation of the innate and adaptive immune system and inflammatory reactions. Cytokines are small, low-weight polypeptides secreted by immune cells that play an important role in intercellular interactions. Different cells can secret the same cytokine, and a single cytokine can act on various cell types. They are produced and released in a cascade, leading to increased or decreased inflammatory response based on the cytokine secreted, the target tissue, and the interaction between various biochemical agents [14-18]. Different pro-inflammatory cytokines, such as interleukin (IL)-1, IL-6, IL-12, IL-18, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α and anti-inflammatory ones, such as IL-4, IL-10, IL-13, and IL-19 which are secreted from immune cells interact with body cells to mediate the immune responses in the body and thus have the most optimum inflammatory response [19]. Chemokines are a group of cytokines that are released mainly by leukocytes to induce chemotaxis in damaged tissues and attract white blood cells. In addition, chemokines could be secreted by tissue-resident cells, such as neurons and glial cells in the brain [20-22]. The term chemotaxis is used to refer to a situation in which cells adjust their movement according to the presence of specific agents in the environment. The reason might be a foreign agent such as a bacterium, fungus, virus, or simply a foreign body. The alteration in the concentration of the different chemokines, such as CXCL10, CXCL-13, CCL-4, CCL-17, CCL-20, and cytokines, including IL-2, IL-4, IL-6, IL-9, IL-10, and IFN-γ have been observed in encephalitis patients [23-30].

This study aims to systematically review studies measuring the levels of interleukins and chemokines in encephalitis, which may help identify and/or develop diagnostic signatures based on these biomarkers. Moreover, these signatures might contribute to a better understanding of the underlying pathophysiology.

2. Materials and methods

2.1. Search strategy and databases

Potentially eligible studies were found by conducting a systematic search in PubMed, Embase, Scopus, and Web of Science with no date and type of study limits. The comprehensive search string is available as S1 Table. The search has been updated until August 17th, 2021. The reference list of the retrieved studies has been screened to find further related studies. The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) has been used to report the results. The completed PRISMA checklist is available as S1 Checklist. The protocol of the study has been registered in the PROSPERO (https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021289298) with the registration ID of CRD42021289298, and the file is available as S1 Protocol.

2.2. Selection criteria

The studies providing the concentration of the cytokines and chemokines in the plasma/serum or CSF have been considered to be eligible for the review. Studies were included if they met the following inclusion criteria: 1) original studies on human subjects, 2) diagnosis of any type of encephalitis, meningoencephalitis, and encephalomyelitis, 3) measurement of the cytokine concentration in the serum, plasma, or CSF of patients and controls (healthy controls or patients without encephalitis and other infectious diseases), and 4) adequate data for calculation of standardized mean difference (SMD). Meanwhile, the studies were excluded if they had any of the following criteria: 1) reviews, book chapters, case reports, meeting abstracts, 2) studies containing animal subjects, 3) in vivo and in vitro studies, 4) studies on gene expression of the cytokines but not their levels, 5) studies that did not have control groups 6) All cancer patients and patients with paraneoplastic encephalitis, because of the existence of cancerous tissue or remote neoplasia that might dysregulate cytokines profile, 7) concurrent complications such as pulmonary edema which could affect the concentration of cytokines, and 8) no measurement of the cytokines in the plasma, serum, or CSF. Screening and eligibility assessment were performed independently by two authors (ASK and MTPF), and discrepancies and conflicts were resolved by discussion. The third author (PS) was consulted for conflict resolution in case of disagreement.

2.3. Data extraction

Two authors have extracted the data independently (ASK and MTPF). The following data were extracted from the studies: the first author’s name, publication year, location, diagnostic criteria for encephalitis, inclusion and exclusion criteria of patients and controls, sample size, demographic information (e.g., age and sex), duration of the hospitalization, assay type, sampling sources, type of the cytokine, cytokine concentration, mean and standard deviation (SD) in patients and controls, measurement unit, and WBC levels in blood and CSF, where available. Discrepancies were resolved by discussion and agreement. When enough data was not available in the paper, we contacted the corresponding author to request further data. In case the reported concentration was median and interquartile range (IQR) or range, we used the transformation reported by Wan et al. [31], Luo et al. [32], and Shi et al. [33] to calculate mean and SD.

2.4. Quality assessment

The overall quality of the included studies was critically appraised independently by two raters (ASK and MTPF) using the Joanna Briggs Institute’s (JBI) checklist for the analytical cross-sectional studies. It has eight questions for quality assessment of the studies [34, 35]. The JBI tool assesses studies based on their inclusion criteria, study subjects and setting, measurement, confounding factors and dealing with them, outcomes, and statistical methods. The list of questions and their detailed definition is available in Table 2. The incongruences in the quality assessment were settled by consulting with the third author (PS).

Table 2

Assessment of the quality of the included studies using Joanna Briggs Institute’s (JBI) checklist.

Study	Criteria and corresponding scores								Total	%
	#1	#2	#3	#4	#5	#6	#7	#8
Tsai, M. L. 1996 [30]	0	0	0	1	NR	NR	1	1	3	37.5
Singh, A. 2000 [29]	1	1	1	1	NR	NR	1	1	6	75
Lin, T. Y. 2002 [142]	1	1	0	0	1	1	1	1	6	75
Wang, S. M. 2003 [28]	1	1	1	1	NR	NR	1	1	6	75
Leake, J. A. 2004 [149]	1	1	1	1	NR	NR	1	1	6	75
Babu, G. N. 2006 [175]	1	1	0	0	NR	NR	1	1	4	50
Lepej, S. Z. 2007 [147]	1	1	1	1	NR	NR	1	1	6	75
Wang, S. M. 2008 [127]	1	1	1	1	NR	NR	1	1	6	75
Zajkowska, J. 2011 [179]	1	1	1	1	NR	NR	1	1	6	75
Ygberg, S. 2016 [125]	1	1	1	1	NR	NR	1	1	6	75
Singh, S. K. 2017 [131]	1	1	1	0	NR	NR	1	1	5	62.5
Koper, O. M. 2018 [152]	1	1	1	1	NR	NR	1	1	6	75
Ai, P 2018 [178]	1	1	0	1	NR	NR	1	1	5	62.5
Maric, L. S. 2018 [180]	1	1	1	1	NR	NR	1	1	6	75
Liu, B. 2018 [141]	1	1	1	1	NR	NR	1	1	6	75
Liu, J. 2018 [140]	1	1	0	1	NR	NR	1	1	5	62.5
Lin, Y. T. 2019 [25]	1	1	1	1	NR	NR	1	1	6	75
Peng, Y. 2019 [134]	1	1	1	1	1	1	1	1	8	100
Jiang, X. Y. 2020 [156]	1	1	0	1	NR	NR	1	1	5	62.5
Li, J. 2020 [144]	1	1	0	1	1	1	1	1	7	87.5
Ygberg, S. 2020 [126]	1	1	1	1	NR	NR	1	1	6	75
Zou, C. 2020 [184]	1	1	1	1	NR	NR	1	1	6	75
Xie, J. 2021 [183]	1	1	0	1	NR	NR	0	1	4	50

#1: Were the criteria for inclusion in the sample clearly defined? #2: Were the study subjects and the setting described in detail? #3: Was the exposure measured in a valid and reliable way? #4: Were objective, standard criteria used for measurement of the condition? #5: Were confounding factors identified? #6: Were strategies to deal with confounding factors stated? #7: Were the outcomes measured in a valid and reliable way? #8: Was appropriate statistical analysis used?

Abbreviations: NR, not reported

0: The criterion is not fulfilled or reported by the study

1: The criterion is met by the study

2.5. Data synthesis and meta-analysis

The mean and SD of cytokine levels have been gathered as continuous data in encephalitis patients and the control group. When studies have reported concentrations of cytokines in more than one encephalitis group, if the types of encephalitis were similar (all groups had infectious encephalitis or all groups had autoimmune encephalitis), we pooled the mean and SD of the concentrations in those groups. In case median and IQR or range had been reported instead of mean and SD, they were converted to mean and SD. The between-group Hedges’ g standardized mean difference (SMD) was calculated based on sample size, mean, and SD to measure the effect size (ES) of the studies. Hedges’ g is similar to Cohen’s d with an adjustment for small samples. SMD and its 95% confidence interval (CI) were used to represent the difference in cytokines between the encephalitis group and controls. The ES of 0.2, 0.5, and 0.8 demonstrate a small, moderate, and large effect, respectively [36]. Cochran’s Q test was used to assess heterogeneity, and a P-value of 0.10 was considered as the existence of heterogeneity. Moreover, I2 was used for a more precise estimation of the heterogeneity, and I < 25%, 25–75%, and >75% are deemed as low, moderate, and high heterogeneity, respectively. Random effect model analysis using the DerSimonian and Laird method was used for meta-analysis, and the P-value equal to or less than 0.05 was considered significant [37]. Subgroup analysis was deployed based on whether the type of encephalitis is infectious or autoimmune. The overall effect for each subgroup is reported when at least two studies exist in the subgroup. In addition, the comparison of the subgroups was performed when each subgroup had at least two studies. Meta-regression was also performed based on the mean age of the participants of each study if it was possible (more than two studies with reported age were present in the meta-analysis; studies that did not report the mean or median age of the participants were omitted from meta-regression). Funnel plot asymmetry and Egger’s test were used for evaluation of the publication bias in the included studies. To further assess the source of the heterogeneity, sensitivity analysis was used to determine potentially influential cases. Each time one study was removed, and the effect size was recalculated to optimize the robustness of the combined effect estimate and examine its influence on the pooled SMD. All statistical analysis has been performed using the “meta” package of R software (version 4.1.1) [38].

3. Results

3.1. Study selection

The search was conducted using EMBASE, Scopus, PubMed, and Web of Science Databases for related articles, which yielded a total of 8,685 records. After searching for duplicates and removing them, 3,054 results remained for screening, of which 159 articles remained after title/abstract screening. In the process of full-text revision, 88 articles did not meet the eligibility criteria and were excluded for different reasons, as outlined in Fig 1 based on PRISMA guidelines (_((((xxx))))_)[39-103]. In some of the excluded articles, the study was conducted to compare cytokines levels in other illnesses or different stages of encephalitis, all lacking control groups (_((((xxx))))_)[44, 45, 47, 50, 53, 54, 64–66, 69, 71, 72, 74–76, 80–86, 89, 90, 94, 95, 100–103]. There were articles whose results could not be included in this review because of being in vivo, in vitro, case reports, review articles, and conference abstracts (_((((xxx))))_)[39, 40, 42, 46, 49, 51, 52, 55, 57, 59, 61, 63, 67, 68, 73, 78, 79, 93, 97–99]. In 12 studies, there were neither qualitative reports of cytokine levels nor were there quantitative data of cytokine concentrations; therefore, they did not meet our inclusion criteria [41, 48, 56, 58, 60, 62, 70, 77, 87, 88, 91, 92]. Two studies measured the efficacy and side effects of vaccines [43, 96]. The full-text article could not be found in 21 studies (_((((xxx))))_)[80, 104–123]. Based on the inclusion and exclusion criteria, 71 studies (_((((xxx))))_)[8, 23–30, 121, 124–184] were considered for this study, of which 23 studies had sufficient quantitative data and were included in the meta-analysis and reported in this study (_((((xxx))))_)[25, 28–30, 125–127, 131, 134, 140–142, 144, 147, 149, 152, 156, 175, 178–180, 183, 184]. The remaining 48 studies were not included in the meta-analysis because they mostly reported qualitative data and did not contain sufficient quantitative data. Our study does not report qualitative data and only focuses on the meta-analysis of quantitative data. Although Chen et al. [8] study investigated the concentration of cytokines, the reported data overlapped Zou et al. [184] study, published more recently, so we considered the latter in the meta-analysis. The meta-analysis was performed for 12 cytokines and chemokines, each of which was assessed in at least three comparison studies.

Fig 1

Flow diagram of literature search and study selection.

The values demonstrate the number of documents in each category.

3.2. Characteristics and quality of the included studies

The included studies’ publication time were between 1996 and 2021. CSF was used in 11 studies for evaluating the concentration of the cytokines [29, 30, 125, 126, 134, 140, 141, 149, 156, 178, 184], plasma/serum was examined for cytokines concentration in 2 studies [28, 142], and both CSF and serum/plasma were investigated in 9 studies [25, 127, 131, 144, 147, 152, 175, 179, 180, 183]. Twelve cytokines and chemokines were included in the CSF meta-analysis, namely IL-2 [149, 156, 175], IL-4 [125, 126, 156, 183], IL-6 [25, 125, 126, 134, 141, 156, 178, 184], IL-8 [29, 30, 125, 126, 140, 156], IL-10 [25, 125, 126, 134, 149, 156, 184], IL-17 [25, 125, 141, 178], TNF-α [134, 144, 156, 175, 184], IFN-γ [125, 126, 149, 156, 183], CCL2 [125, 126, 156], CXCL9 [140, 152, 156], CXCL10 [147, 156, 179, 180], and CXCL13 [25, 156, 179]. Five cytokines, including CXCL10 [127, 147, 179, 180], CXCL13 [25, 179, 180], IL-6 [25, 131, 142], IL-10 [25, 28, 131], and TNF-α [28, 142, 144, 175] were encompassed in the serum/plasma meta-analysis. The concentration of the cytokines were measured using enzyme-linked immunosorbent assay (ELISA) (_((((xxx))))_)[25, 28–30, 131, 134, 141, 142, 147, 149, 152, 175, 178–180, 183, 184], flowcytometry immunoassay [28, 127, 140], and multiplex assay [125, 126, 156]. The total number of the included subjects in the serum meta-analysis is 421, comprised of 241 patients and 180 controls. Concerning CSF meta-analysis, a total of 1062 individuals are available, including 670 patients 392 controls. The summary of the characteristics of the studies is available in the Table 1.

Table 1

Summary of findings from studies included in the quantitative analysis of serum/plasma or CSF levels in encephalitis disease.

First Author	Year	Country	CSF/Serum	Measured Cytokine(s)	Encephalitis Type	Patients N	Controls N	Assay Type	Female %	Mean Age	Encephalitis Type
Tsai, M. L. [30]	1996	Taiwan	CSF	IL-8	Infectious	11	11	ELISA	-	-
Singh, A. [29]	2000	India	CSF	IL-8	Infectious	30	27	ELISA	-	-	JE
Lin, T. Y. [142]	2002	Taiwan	Serum	IL-6, TNF-α	Infectious	8	21	ELISA	14.19	19.59
Wang, S. M. [28]	2003	Taiwan	Serum	IL-2, IL-4, IL-10, TNF-α, IFN-γ	Infectious	34	15	Multiplex Assay	-	-	EV71-associated BE
Leake, J. A. [149]	2004	USA	CSF	IL-2, IL-10, IFN-γ	Auto-Immune	11	28	ELISA	-	-	ADEM
Babu, G. N. [175]	2006	India	CSF & Serum	IL-2, TNF-α	Infectious	18	20	ELISA	-	-	JE
Lepej, S. Z. [147]	2007	Croatia	CSF & Serum	CXCL10	Infectious	19	10	Immunoassay	50.34	34.90	TBE patient
Wang, S. M. [127]	2008	Taiwan	CSF & Serum	CCL2, IL-8, CXCL10	Infectious	21	13	Multiplex Assay	-	-	EV71-associated BE
Zajkowska, J. [179]	2011	Poland	CSF & Serum	CXCL10, CXCL13	Infectious	15	8	ELISA	33.30	43.00	TBE
Ygberg, S. [125]	2016	Sweden	CSF	IL-4, IL-6, IL-8, IL-10, IL-17, CCL2, IFN-γ	Both	17	13	Multiplex Assay	64.68	51.29	13 anti-NMDAR + 4 Infectious
Singh, S. K. [131]	2017	India	CSF & Serum	IL-6, IL-10	Both	87	64	ELISA	98.00	-	13 AES+JE, 74 AES
Koper, O. M. [152]	2018	Poland	CSF & Serum	CXCL9	Infectious	24	13	ELISA	64.24	47.11	TBE
Ai, P [178]	2018	China	CSF	IL-6, IL-17	Auto-Immune	33	38	ELISA	54.46	37.55	Anti-NMDAR
Maric, L. S. [180]	2018	Croatia	CSF & Serum	CXCL10, CXCL13	Auto-Immune	23	20	ELISA	-	-	ADEM
Liu, B. [141]	2018	China	CSF	IL-6	Auto-Immune	24	31	ELISA	47.28	36.98	NMDAR
Liu, J. [140]	2018	China	CSF	IL-8, IL-10, CXCL9	Infectious	99	22	Multiplex Assay	39.67	28.79	EV71-associated BE
Lin, Y. T. [25]	2019	China	CSF & Serum	IL-6, IL-10, IL-17, CXCL13	Auto-Immune	16	9	ELISA	15.92	51.65	Anti-LGI1 encephalitis
Peng, Y. [134]	2019	China	CSF	IL-10, TNF-α	Auto-Immune	33	17	ELISA	56.00	34.87	NMDAR
Jiang, X. Y. [156]	2020	China	CSF	IL-2, IL-4, IL-6, IL-8, IL-10, TNF-α, CCL2, CXCL9, CXCL10, CXCL13, IFN-γ	Auto-Immune	147	35	Multiplex Assay	-	-	NMDAR
Ygberg, S. [126]	2020	Sweden	CSF	IL-4, IL-6, IL-8, IL-10, CCL2, IFN-γ	Infectious	37	19	Multiplex Assay	41.04	95.93	TBE
Zou, C. [184]	2020	China	CSF	IL-6, TNF-α	Both	46	21	-	35.70	61.54	33 NMDAR + 13 Viral
Li, J. [144]	2020	China	CSF & Serum	TNF-α	Infectious	84	50	ELISA	44.06	6.37	TBE
Xie, J. [183]	2021	China	CSF & Serum	IL-4, IFN-γ	Infectious	80	40	ELISA	52.50	7.83	40 Viral + 40 Supporative

Abbreviations: CSF, cerebrospinal fluid; IL-8, interleukin-8; ELISA, enzyme-linked immunoassay; JE, Japanese encephalitis; TNF-a, tumor necrosis factor-a; EV-71, enterovirus-71; BE, brainstem encephalitis; ADEM, acute disseminated encephalomyelitis; TBE, tick-borne encephalitis; NMDAR, N-methyl-D-aspartate receptor; AES, acute encephalitis syndrome; LGI1, leucine-rich glioma inactivated1; IFN-γ, Interferon γ

Using the JBI quality assessment tool, most of the studies achieved a quality score of 5 (62.5) or more, two studies had a score of 4 (50%) [175, 183], and the score of one study was 3 (37.5) [30]. A summary of the quality assessment of included studies is provided in Table 2. The funnel, drapery, and sensitivity analysis forest plots of the analyses are available in S1 File.

3.3. IL-2

Three studies [149, 156, 175] with 176 encephalitis patients and 83 controls reported CSF concentration of IL-2. Although the overall concentration of the IL-2 was higher in encephalitis patients, this difference was not significant, with a P-value of 0.05 (SMD, 0.82; 95% CI, -0.02–1.66). The heterogeneity was considerable with Q = 11.58 (P < 0.01) and I = 83%. The meta-regression for age was not applicable because data for age were only available in one study. In addition, subgroup analysis demonstrated non-significant increase in the concentration of IL-2 in infectious encephalitis (SMD = 0.87; 95% CI, -0.89–2.63; P = 0.33) with high heterogeneity (Q = 11.57; P < 0.01; I2 = 91%). There was only one study available for the autoimmune subgroup; hence, the overall effect for this subgroup and comparison between subgroups are not applicable. The Egger’s test revealed no publication bias (P = 0.91). Sensitivity analysis revealed that omitting Leake, J. A. 2004 study leads the overall effect to become significant (P = 0.02) (Fig 2A).

Fig 2

Forest plots for results of meta-analysis for IL-2 CSF concentrations (A), IL-4 CSF concentrations (B), IL-6 CSF concentrations (C), and IL-6 serum/plasma concentrations (D). SD, standard deviation; SMD, standardized mean difference; CI, confidence interval.

3.4. IL-4

241 patients and 106 controls from 4 studies [125, 126, 156, 183] were included in the meta-analysis of CSF concentration of IL-4. The meta-analysis showed slightly elevated IL-4 CSF levels, but this was not statistically significant (SMD, 1.74; 95% CI, -0.09–3.58; P = 0.06). The heterogeneity was significant with Q = 123.24 (P < 0.01) and I2 = 97%. The meta-regression demonstrated no significant effect for age as a moderator of the meta-analysis (P = 0.33). The subgroup analysis did not show significant alteration in IL-4 levels in both infectious (SMD, 2.73; 95% CI, -1.35–6.80; P = 0.19) or autoimmune encephalitis (SMD, 0.50; 95% CI, -0.89–1.90; P = 0.48). The autoimmune subgroup had less heterogeneity (Q = 4.51; P < 0.05; I = 78%) than the infectious group (Q = 118.72; P < 0.01; I = 98%). The chi-square test for assessing the disparity between subgroups demonstrated no significant difference (P = 0.31). There was no publication bias based on the Egger’s linear regression test for funnel plot asymmetry (P = 0.56). Omitting Ygberg, S. 2016 from autoimmune subgroup in the sensitivity analysis resulted in a significant overall effect (P = 0.04) (Fig 2B).

3.5. IL-6

Out of 10 studies that investigated the concentration of the IL-6, 7 reported its concentration in CSF [125, 126, 134, 141, 156, 178, 184], 2 reported its concentration in serum/plasma [131, 142], and 1 study reported both serum/plasma and CSF levels of IL-6 [25].

3.5.1. CSF IL-6

A total of 340 patients and 182 controls explored CSF concentration of IL-6. The CSF levels of the IL-6 in encephalitis patients are significantly higher than controls (SMD, 1.72; 95% CI, 0.79–2.66; P < 0.001). The overall heterogeneity is high, with Q = 129.24 (P < 0.01) and I = 94%. Meta-regression for age as moderator was not significant (P = 0.34). Subgroup analysis results are consistent and show that in both autoimmune and infectious encephalitis, CSF concentration of IL-6 is higher than controls (SMD (autoimmune), 1.90; 95% CI, 0.71–3.90; P < 0.01; SMD (infectious), 1.28; 95% CI, 0.75–1.80; P < 0.01). There was no significant difference between the two subgroups (P = 0.35). The tests for heterogeneity showed a high heterogeneity in the autoimmune subgroup (Q = 127.54; P < 0.01; I = 95%) and a lack of observed heterogeneity in the infectious subgroup (Q = 0.00; P = 0.96; I = 0.0%). The publication bias was not significant based on the Egger’s test (P = 0.05), and the sensitivity analysis demonstrated that Ai, P. can influence the overall effect size, without changing the statistical significance (Fig 2C).

3.5.2. Serum IL-6

Pooled data demonstrated that the serum/plasma levels of the IL-6 are not significantly higher (SMD, 0.56; 95% CI, -0.41–1.52; P = 0.26) in encephalitis patients (n = 111) than controls (n = 94). The overall heterogeneity was high (Q = 11.86; P < 0.01; I = 83%). Meta-regression was not applicable due to the lack of enough data. Subgroup analysis for infectious encephalitis indicated similar results to the overall pooled effect with high heterogeneity (SMD, 0.92; P = 0.28; 95% CI, -0.76–2.61; Q = 11.1, P < 0.01; I = 91%). Since there is only one study in the autoimmune subgroup, the overall effect for this subgroup and comparison between subgroups are not applicable. The Egger’s test showed no significant publication bias (P = 0.58). The sensitivity analysis suggested that omitting Singh, S. 2017 or Lin, Y. T. 2019 could affect the overall effect to be statistically significant. This happens because of the small number of studies and high SMD of the Lin, T. Y. 2002 study, which could affect the pooled effect in the absence of the mentioned studies (Fig 2D).

3.6. IL-8

Random effect meta-analysis of 6 studies [29, 30, 125, 126, 140, 156] demonstrated that encephalitis patients (n = 341) had higher CSF IL-8 concentration compared to control group (n = 126) for IL-8 (SMD, 1.03; 95% CI, 0.17–1.90; P < 0.05). The overall heterogeneity calculated is Q = 70.63 (P < 0.01), I = 92%. Meta-regression demonstrated significant effect of age on the CSF concentration of IL-8 (P < 0.05). In subgroup analysis, the CSF concentration of IL-8 in the autoimmune subgroup was significantly higher than controls (SMD, 1.15; 95% CI, 0.77–1.52; P < 0.01). In infectious subgroup, IL-8 CSF levels were not statistically different from controls (SMD, 1.09; 95% CI, -0.19–2.38; P = 0.10). The most observed heterogeneity was attributable to the infectious subgroup (Q = 66; P < 0.01; I = 94%), and there was no considerable observed heterogeneity in autoimmune subgroup (Q = 0.79; P = 0.37; I = 0.0%). The Between-subgroup test did not show any significant difference (P = 0.94). The Egger’s linear regression test for meta bias reported that publication bias was not significant (P = 0.68). Sensitivity analysis indicates that removing Singh, A. 2000 and Jiang, X. Y. 2020 studies leads to a non-significant overall effect. This is possibly due to the high effect size reported in Singh, A. 2000 study and robust large sample size in Jiang, X. Y. 2020, which greatly influences the overall result (Fig 3A).

Fig 3

Forest plots for results of meta-analysis for IL-8 CSF concentrations (A), IL-10 CSF concentration (B), IL-10 serum/plasma concentrations (C), and IL-17 CSF concentrations (D). SD, standard deviation; SMD, standardized mean difference; CI, confidence interval.

3.7. IL-10

Nine studies [25, 28, 125, 126, 131, 134, 149, 156, 184] have investigated the concentration of IL-10: 6 studies in CSF [125, 126, 134, 149, 156, 184], 2 studies in serum/plasma [28, 131], and 1 study in both [25].

3.8. CSF IL-10

The higher CSF concentration of IL-10 in encephalitis patients (n = 294) compared to controls (n = 141) was confirmed by meta-analysis (SMD, 1.31; 95% CI, 0.36–2.26; P < 0.01). The heterogeneity among the studies is significant (Q = 95.08; P < 0.0001; I = 93%). The meta-regression demonstrated no significant effect of the age on the CSF concentration of the IL-10 (P = 0.24). In the autoimmune subgroup the effect was consistent with the overall effect (SMD, 2.18; 95% CI, 0.60–3.77; P < 0.01), but the infectious subgroup exhibited no significant alteration of the CSF IL-10 levels (SMD, 014; 95% CI, -0.47–0.74; P = 0.66). The test for the difference between subgroups was significant (P = 0.02). The existing heterogeneity can be mostly ascribed to the autoimmune subgroup with Q = 81.74 (P < 0.01; I = 95), and the heterogeneity in the infectious was moderate (Q = 4.13; P = 0.13; I = 52%). The difference between subgroups was significant with P-value of 0.02. The Egger’s test for the publication bias was not significant (P = 0.16). The sensitivity analysis revealed no single study with significant impact of the overall effect (Fig 3B).

3.9. Serum IL-10

Similarly, the concentration of the IL-10 in serum/plasma was higher in encephalitis patients (n = 137) than controls (n = 88; SMD, 0.51; 95% CI, 0.21–0.80; P < 0.001). The between studies heterogeneity was low (Q = 2.16; P = 0.34; I = 7%). The meta-regression for age was not applicable since the data for age was only available in one study. There was only one study in the autoimmune subgroup, and the pooled effect and evaluation of the heterogeneity were not applicable. Serum/plasma alteration of IL-10 levels in the infectious group was not significant (SMD, 0.43; 95% CI, -0.05–0.92; P = 0.08). The heterogeneity was moderate (Q = 2.11; P = 0.15; I = 53%). The publication bias was not significant (P-value = 0.66). Furthermore, the sensitivity analysis demonstrated that omitting either Singh, S. 2017, or Lin, Y. T. 2019 will make the overall effect non-significant, which may be a consequence of the low number of studies available for the meta-analysis (Fig 3C).

3.10. IL-17

Four studies reported the CSF concentration of IL-17 or IL-17A for encephalitis patients (n = 90) and control group (n = 90). The overall levels of IL-17 were not significantly higher in encephalitis patients compared with controls (SMD, 0.44; 95% CI, -0.32–1.20; P = 0.26). The heterogeneity was high with Q = 19.51 (P < 0.001), and I = 80%. The meta-regression demonstrated a non-significant effect of age on the CSF concentration of the IL-17 (P = 0.54). The subgroup analysis demonstrated that in autoimmune subgroup IL-17 levels were not significantly altered (SMD, 0.41; 95% CI, -0.52–1.34; P = 0.38), with a high heterogeneity among studies (Q = 19.42; P < 0.01; I = 85%). The pooled estimates were not performable for the infectious subgroup because only one study belongs to it. The Egger’s test for funnel plot asymmetry reported no significant publication bias (P = 0.40). The sensitivity analysis demonstrated that removing Lin, Y. T. 2019, because of its reverse reported effect, increases the overall estimate and leads to significantly higher IL-17 levels in patients than controls (P < 0001) (Fig 3D).

3.11. CCL2

Three studies with 201 encephalitis patients and 66 controls investigated the CSF concentration of CCL2. The overall effect tended to show decreased concentration of CCL2 (SMD, -0.53; 95% CI, -1.08–0.01; P = 0.06). The between-study heterogeneity was moderate (Q = 7.11; P = 0.07; I = 58%). Moreover, the meta-regression demonstrated no significant moderating effect for age in the meta-analysis (P = 0.10). In the autoimmune subgroup the overall CCL2 levels were significantly decreased (SMD, -0.54; 95% CI, -0.90–-0.18; P <0.01), but this reduction was not significant in the infectious subgroup (SMD, -0.40; 95% CI, -1.89–1.10; P = 0.60). There was no difference between subgroups (P = 0.86). There was no heterogeneity in the autoimmune subgroup (Q = 0; P = 1; I = 0%), and the infectious subgroup is mostly responsible for the overall heterogeneity (Q = 6.85; P < 0.01; I = 85%). Publication bias was not significant (P = 0.70). Based on the sensitivity analysis, omitting the Ygberg, S. 2016 from infectious subgroup will result in significant overall effect size (Fig 4A).

Fig 4

Forest plots for results of the meta-analysis for CCL2 CSF concentrations (A) and CXCL9 CSF concentrations (B). SD, standard deviation; SMD, standardized mean difference; CI, confidence interval.

3.12. CXCL9

Three studies with 270 encephalitis subjects and 70 control subjects reported the CSF concentration of the CXCL9. The overall effect shows the CSF concentration of the CXCL9 is higher in encephalitis than in control, whereas non-significant (SMD, 2.50; 95% CI, -0.15–5.16; P = 0.06). The heterogeneity is substantially high (Q = 99.75; P < 0.01; I = 98%). The meta-regression for age was not applicable since the data for age was only available in two studies. The subgroup analysis for the infectious encephalitis revealed significantly increased levels of CXCL9 in the CSF (SMD, 3.58; 95% CI, 3.02–4.14; P < 0.01). There was no observed heterogeneity in the infectious subgroup (Q = 0.56; P = 0.46; I = 0%). Egger’s linear regression for the funnel plot asymmetry exhibited no significant publication bias (P = 0.33). The estimates for the autoimmune subgroup were not applicable because there is only one study. The removal of the Jiang, X. Y. 2020 effect in the sensitivity study causes the overall effect to be significant (P < 0.0001) (Fig 4B).

3.13. CXCL10

Out of 5 studies that reported the concentration of the CXCL10, one measured CSF concentration of CXCL10 [156], one measured serum/plasma [127], and three measured both concentrations [147, 179, 180].

3.13.1. CSF CXCL10

Meta-analysis of 204 encephalitis patients and 73 controls showed that the CSF levels of the CXCL10 are higher in the encephalitis group than in the control group (SMD, 0.86; 95% CI, 0.16–1.56; P = 0.02). The overall heterogeneity among studies is high (Q = 13.25; P < 0.01; I = 77%). The meta-regression for age as a moderator was not significant (P = 0.39). The subgroup analysis in the infectious encephalitis, similar to the overall meta-analysis, demonstrated a higher CSF concentration of the CXCL10 in the encephalitis patients compared to the controls (SMD, 1.16; 95% CI, 0.22–2.10; P = 0.02). The heterogeneity in infectious subgroup was moderate (Q = 7.72; P = 0.02; I = 74%). The autoimmune subgroup only had one study, and overall estimates were not applicable. Publication bias based on Egger’s linear regression test was not significant (P = 0.08). Sensitivity analysis revealed that omitting the Maric, L. S. 2018 study will cause the overall effect to become non-significant (Fig 5A).

Fig 5

Forest plots for results of the meta-analysis for CXCL10 CSF concentrations (A), CXCL10 serum/plasma concentrations (B), CXCL13 CSF concentrations (C), and CXCL13 serum/plasma concentrations (D). SD, standard deviation; SMD, standardized mean difference; CI, confidence interval.

3.13.2. Serum CXCL10

In contrast to the CSF meta-analysis, Serum analysis reported no significant alteration in the CXCL10 concentration (npatients = 78; ncontrols = 51; SMD, 0.12; 95% CI, -1.27–1.51; P = 0.87). The observed heterogeneity was high with Q = 37.08 (P < 0.0001; I = 92%). The meta-regression for age as a moderator variable was not significant (P = 0.52). The subgroup analysis was not conductible since the included studies only belong to infectious encephalitis. The publication bias was not significant (P = 0.53). The sensitivity analysis revealed that omitting the studies could lead to a change in the direction of the overall effect. Except for Lepej, S. Z. 2007, whose omission leads to a slightly significant overall effect, removal of the other studies does not lead to a significant change in the overall effect (Fig 5B).

3.14. CXCL13

Among 4 studies that measured CXCL13 concentration, one investigated the CSF concentration [156], one assessed serum/plasma concentration [180], and two measured both CSF and serum/plasma concentration [25, 179].

3.14.1. CSF CXCL13

178 encephalitis patients and 52 controls entered the CSF meta-analysis for CXCL13. There was a non-statistically significant increase in the concentration of CXCL13 in patients (SMD, 091; 95% CI, -0.23–2.05; P = 0.12). The overall heterogeneity was high with Q = 13.57 (P < 0.01; I = 85%). The subgroup analysis in autoimmune encephalitis group demonstrated no significant alteration of the CXCL13 concentration with high between-study heterogeneity (SMD, 1.39; 95% CI, -0.66–3.44; P = 0.18; Q = 12.23; P < 0.01; I = 92%). The meta-regression for age was not applicable since the data for age was only available in two studies. The estimates for the infectious subgroup are not available since there was only one study. Publication bias based on Egger’s test was not significant (P = 0.59). The sensitivity analysis revealed that the removal of Lin, Y. T. 2019 could lead to a significant overall effect size with P-value = 0.04 (Fig 5C).

3.14.2. Serum CXCL13

54 cases and 37 controls were investigated for serum/plasma concentration of the CXCL13. The meta-analysis came up with no significant difference in the serum concentration of CXCL13 in encephalitis patients and controls (SMD, 0.04; 95% CI, -1.48–1.56; P = 0.96). The sample size was small, and CI was wide. The heterogeneity was high with Q = 21.38 (P < 0.0001; I = 91%). The meta-regression demonstrated a significant effect of age on the serum concentration of CXCL13 (P < 0.0001). The subgroup analysis results are only available for the infectious encephalitis group, which, similar to the overall effect, does not implicate any significant alteration in CXCL13 concentration (SMD, -0.52; 95% CI, -2.13–1.10; P = 0.53). The heterogeneity in infectious subgroup was also high (Q = 8.74; P < 0.01; I = 89%). There was no significant publication bias (P = 0.13). Sensitivity analysis reported that omitting Lin, Y. T. 2019 study leads to a slightly significant overall effect size (P = 0.04) (Fig 5D).

3.15. TNF-α

Seven studies investigated the concentration of TNF-α: two were related to serum/plasma concentration [28, 142], three to CSF concentration [134, 156, 184], and two measured both CSF and serum/plasma concentration [144, 175].

3.15.1. CSF TNF-α

Increased CSF concentration of the TNF-α was seen in the meta-analysis of 213 encephalitis patients and 73 controls (SMD, 2.61; 95% CI, 1.29–3.94; P < 0.001). The observed heterogeneity was high with Q = 90.82 (P < 0.0001; I = 96%). The meta-regression demonstrated a significant effect of age on the CSF concentration of CXCL13 (P < 0.0001). The effect in the autoimmune encephalitis sub-group was similar to the overall effect (SMD, 1.77; 95% CI, 0.45–3.09; P < 0.01). Subgroup-analysis for infectious encephalitis demonstrated no significant alteration in the concentration of the TNF-α (SMD, 10.91; 95% CI, -9.02–30.84; P = 0.28). The test for subgroup difference demonstrated no significant difference (P = 0.37). The heterogeneity was high in both infectious (Q = 61.13; P < 0.01; I = 98%) and autoimmune (Q = 29.21; P < 0.01; I = 93%) subgroups. The Egger’s linear regression test for funnel plot asymmetry demonstrated significant publication bias with P-value less than 0.001. The sensitivity analysis came up with no change in the significance of the effect when putting each of studies aside (Fig 6A).

Fig 6

Forest plots for results of the meta-analysis for TNF-α CSF concentrations (A), TNF-α serum/plasma concentrations (B), and IFN-γ CSF concentrations (C). SD, standard deviation; SMD, standardized mean difference; CI, confidence interval.

3.15.2. Serum TNF-α

The meta-analysis demonstrated that the concentration of the TNF-α is increased in the encephalitis patients compared to controls (npatients = 144; ncontrol = 106; SMD, 9.06; 95% CI, 2.85–15.27; P < 0.01). The observed heterogeneity of the reported studies was high (Q = 317.69; P < 0.0001; I = 99%). The meta-regression for age was not applicable since the data for was only available in two studies. The subgroup analysis was not available because all studies investigated infectious encephalitis patients. Publication bias was not significant with P-value = 0.14. The sensitivity analysis displayed that omitting each study except Li, J. 2020, which does not change the significance of the overall effect, will lead to a non-significant overall effect size (Fig 6B).

3.16. IFN-γ

252 encephalitis patients and 134 controls were included in the IFN-γ CSF concentration meta-analysis, which showed a tendency to increased concentration of IFN-γ in the former group (SMD, 1.40; 95% CI, -0.06–2.85; P = 0.06). The overall heterogeneity was high (Q = 123.06; P < 0.0001; I = 96%). The meta-regression demonstrated no significant effect of the age on the CSF concentration of the IFN-γ (P = 0.30). The effects in the infectious (SMD, 2.06; 95% CI, -0.50–4.62; P = 0.11) and autoimmune (SMD, 0.35; 95% CI, -0.01–0.70; P = 0.06) subgroups were consistent with the overall effect. The heterogeneity in infectious subgroups was high (Q = 117.52; P < 0.01; I = 97%), but the heterogeneity in autoimmune subgroup was low (Q = 0.4; P = 0.53; I = 0%). Publication bias was not significant with P-value = 0.34. Sensitivity analysis via leaving one study out at each time, showed that omitting Ygberg, S. 2016 and Leake, J. A. 2004 can lead to increased overall effect with a P-value of 0.046 and 0.050, respectively (Fig 6C).

4. Discussion

The present meta-analysis was conducted to explore cytokine levels in patients with encephalitis compared to controls without encephalitis, as well as to compare the cytokine concentrations in autoimmune encephalitis and infectious encephalitis. The cytokines IL-6, IL-8, IL-10, CXCL10, and TNF-α were significantly higher in the CSF of patients compared to controls. Also, there was a significant difference between infectious and autoimmune encephalitis regarding CSF levels of IL-10 and CXCL9. Increased serum levels of TNF-α were observed in patients with encephalitis. The Meta-analyses results have been summarized in Table 3, Figs 7 and 8, S1 and S2 Figs.

Table 3

Summary of the meta-analyses.

Marker	Source	Subgroup	No. Studies	No. Cases	No. Controls	Meta-analysis			Heterogeneity				P-value for subgroup difference
						SMD	95% CI	P-value	I 2	τ2	Q	P-value1
IL-2	CSF	Autoimmune	1	147	35	NA	NA	NA	NA	NA	NA	NA	NA
		Infectious	2	29	48	0.87	-0.89–2.63	0.33	91%	1.47	11.57	<0.01
		Overall	3	176	83	0.82	-0.02–1.66	0.05	83%	0.45	11.58	<0.01
IL-4	CSF	Autoimmune	2	151	41	0.50	-0.89–1.90	0.48	78%	0.82	4.51	<0.01	0.31
		Infectious	3	90	65	2.73	-1.35–6.80	0.19	98%	12.69	118.72	<0.01
		Overall	5	241	106	1.74	-0.09–3.58	0.06	97%	4.15	123.24	<0.01
IL-6	CSF	Autoimmune	7	290	157	1.90	0.71–3.09	<0.01	95%	2.38	127.54	<0.01	<0.01
		Infectious	2	50	25	1.28	0.75–1.80	<0.01	0%	0	0	0.96
		Overall	9	340	182	1.72	0.79–2.66	<0.01	94%	1.83	129.24	<0.01
	Serum	Autoimmune	1	16	9	NA	NA	NA	NA	NA	NA	NA	NA
		Infectious	2	95	85	0.92	-0.75–2.61	0.28	91%	1.35	11.1	<0.01
		Overall	3	111	94	0.56	-0.41–1.52	0.26	83%	0.59	11.86	<0.01
IL-8	CSF	Autoimmune	2	151	41	1.15	0.77–1.52	<0.01	0	0	0.79	0.37	0.94
		Infectious	5	190	85	1.09	-0.19–2.38	0.10	94%	1.99	66	<0.01
		Overall	7	341	126	1.03	0.17–1.90	0.02	92%	1.19	70.63	<0.01
IL-10	CSF	Autoimmune	5	233	88	2.18	0.60–3.77	<0.01	95%	2.96	81.74	<0.01	0.02
		Infectious	3	61	53	0.14	-0.47–0.74	0.66	52%	0.14	4.13	0.13
		Overall	8	294	141	1.31	0.36–2.26	<0.01	93%	1.64	95.08	<0.01
	Serum	Autoimmune	1	16	9	NA	NA	NA	NA	NA	NA	NA	NA
		Infectious	2	121	79	0.43	-0.05–0.92	0.08	53%	0.07	2.11	0.15
		Overall	3	137	88	0.51	0.21–0.80	<0.01	7%	0.01	2.16	0.34
IL-17	CSF	Autoimmune	4	77	84	0.41	-0.52–1.34	0.38	85%	0.73	19.42	<0.01	NA
		Infectious	1	13	6	NA	NA	NA	NA	NA	NA	NA
		Overall	5	90	90	0.44	-0.32–1.20	0.26	80%	0.47	19.51	<0.01
CCL2	CSF	Autoimmune	2	151	41	-0.54	-0.90–-0.18	<0.01	0	0	0	1	0.86
		Infectious	2	50	25	-0.40	-1.89–1.10	0.60	85%	1.00	6.85	<0.01
		Overall	4	201	66	-0.53	-1.08–0.01	0.06	58%	0.17	7.11	0.07
CXCL9	CSF	Autoimmune	1	147	35	NA	NA	NA	NA	NA	NA	NA	NA
		Infectious	2	123	35	3.58	3.02–4.14	<0.01	0	0	0.56	0.46
		Overall	3	270	70	2.50	-0.15–5.16	0.06	98%	5.34	99.75	<0.01
CXCL10	CSF	Autoimmune	1	147	35	NA	NA	NA	NA	NA	NA	NA	NA
		Infectious	3	57	38	1.16	0.22–2.10	0.02	74%	0.50	7.72	0.02
		Overall	4	204	73	0.86	0.16–1.56	0.02	77%	0.37	13.25	<0.01
CXCL10	Serum	Infectious	4	78	51	0.12	-1.27–1.51	0.87	92%	1.84	37.08	<0.01	NA
CXCL13	CSF	Autoimmune	2	163	44	1.39	-0.66–3.44	0.18	92%	2.01	12.23	<0.01	NA
		Infectious	1	15	8	NA	NA	NA	NA	NA	NA	NA
		Overall	3	178	52	0.91	-0.23–2.05	0.12	85%	0.84	13.57	<0.01
CXCL13	Serum	Autoimmune	1	16	9	NA	NA	NA	NA	NA	NA	NA	NA
		Infectious	2	38	28	-0.52	-2.13–1.10	0.53	89%	1.20	8.74	<0.01
		Overall	3	54	37	0.04	-1.48–1.56	0.96	91%	1.62	21.38	<0.01
TNF-α	CSF	Autoimmune	3	213	73	1.77	0.45–3.09	<0.01	93%	1.24	29.21	<0.01	0.37
		Infectious	2	102	70	10.91	-9.02–30.84	0.28	98%	203.47	61.13	<0.01
		Overall	5	315	143	2.61	1.21–3.94	<0.01	96%	1.83	90.82	<0.01
	Serum	Infectious	4	144	106	9.06	2.85–15.27	<0.01	99%	38.36	317.69	<0.01	NA
IFN-γ	CSF	Autoimmune	2	151	41	0.35	-0.01–0.70	0.06	0	0	0.4	0.53	0.19
		Infectious	4	101	93	2.06	-0.50–4.62	0.11	97%	6.59	117.52	<0.01
		Overall	6	252	134	1.40	-0.06–2.85	0.06	96%	3.07	123.06	<0.01

1 P-value for Cochrane Q test

Abbreviations: CSF, cerebrospinal fluid; NA, not applicable

Fig 7

Summary of the overall effects and heterogeneity of the concentration of the cytokines in the CSF.

SMD, standardized mean difference; CI, confidence interval.

Fig 8

Summary of the overall effects and heterogeneity of the concentrations of the cytokines in the serum/plasma.

SMD, standardized mean difference; CI, confidence interval.

Encephalitis is a condition that can be the result of various causes. It can be classified into infectious and autoimmune encephalitis. Infectious encephalitis is caused when an infectious agent enters the body, mainly viruses, but some cases of primary encephalitis are caused by bacteria, fungi, or parasites. The common viral causes of encephalitis include HSV, JEV, varicella zoster, measles, mumps, rabies, and rubella [185-187].

Herpes simplex encephalitis (HSE) is one of the most common types of infectious encephalitis. It is caused by the HSV-1 virus–a neurotropic virus that infects the trigeminal nerve of humans [188]. Despite its widespread prevalence, it is still unknown whether HSE is caused by viral reactivation inside the trigeminal ganglia, direct primary infection of the olfactory mucosa, or other infected central nervous system (CNS) neurons in the brain [189]. Prior investigations have demonstrated that cytokines, such as TNF-α, IFN-γ, and IL-1, play an important role in the pathogenesis of HSE [190, 191].

When the body is infected with HSV, toll-like receptors (TLRs) such as TLR2, TLR3, TLR7, and TLR9 help identify the virus and then, through signaling cascades, trigger intracellular responses. Consequently, these responses stimulate the activation and migration of immune cells [191, 192]. Moreover, the activation of TLRs via HSV-1 proteins or nucleic acids can result in the expression of chemokines and pro-inflammatory cytokines that accelerate the activation of innate immune reactions [93, 193, 194]. The immune reactions stimulate numerous types of immune cells to produce a wide range of cytokines and chemokines (CXCL9), including interleukins, TNF-α, and interferon regulatory factors (IRF3, IRF7, and type I IFNs) [190, 191]. Type I interferons (IFN-α and IFN-β) are major contributors to combat viral encephalitis, including HSE. Non-immune cells also produce interferons in response to pathogens to regulate the immune system’s activity. Studies have proven that overexpression of these cytokines or chemokines could exacerbate brain damage over time. In summary, when HSV-1 infects neurons, innate immunity through IFN signaling prevents the virus from spreading [195].

Adaptive immune responses to HSV-1 and HSV-2 are complicated, involving a variety of factors, including nuclear factor kappa B (NFκB) and Janus kinase (JAK)/signal transducer and activator of transcription (STAT) [196]. Interferons could activate IFN stimulated genes (ISGs) that produce antiviral proteins in infected and surrounding cells [196]. The binding of interferons to the IFNα/β receptor activates the JAK-STAT pathway in adjacent cells. This event inhibits viral replication [197]. The innate immune response against HSV is facilitated by plasmacytoid dendritic cells (pDCs) and Natural Killer (NK) cells [196, 198]. Besides interferon, it has been demonstrated that TLR signaling could trigger the expression of proinflammatory cytokines, including pro–IL-1 β leading to neural death [199].

Wang et al. [200] reported that when comparing the expression of cytokines/chemokines in brain tissues from the experimental and control groups, 13 out of 62 components in the array were found to be differently expressed in the experimental group. Notably, all of the differentially expressed proteins (DEPs) were upregulated in the experimental group and belonged to a variety of cytokine families, including the interleukin family (IL-1, IL-2, IL-12f, IL-4, and IL-6), the CCL series (CCL-5, Macrophage inflammatory protein-1 (MIP-1 or CCL-3), and Monocyte chemoattractant protein-1 (MCP-1 or CCL-2)), the CXCL family (keratinocyte-derived cytokine), and the tumor necrosis factor family (TNF-α and soluble tumor necrosis factor receptor-Ι (sTNF-RΙ)). They mentioned that the majority of DEPs were inflammatory proteins that performed important roles in the recruitment and activation of inflammatory cells. It had not been previously documented that the cytokines IL-1, MIP-1, and sTNF-RΙ were upregulated in HSE [200]. It is generally established that IL-1, regulated upon activation, normal T cells expressed and secreted (RANTES or CCL-5), and KC stimulate the accumulation of inflammatory cells in the affected tissues. MIP-1 and MCP-1 are also known to stimulate the formation of monocyte/macrophage cell lines, whereas IL-2, TNF-α, and IL-12 are known to promote the activation and differentiation of T cells. Furthermore, IL-4 and IL-6 increase B cell activation and differentiation [201, 202]. These elements could also trigger the expression of additional cytokines and probably have a critical role in increasing inflammation and brain tissue damage in acute viral meningitis [203, 204].

The Japanese Encephalitis Virus (JEV) is the most prevalent cause of viral encephalitis in Southeast Asia, transmitted by mosquitoes. After JEV has entered the bloodstream, the viral replication peaks without any cell death or TNF-α release; then, monocytes become activated and differentiate into monocyte-derived dendritic cells (MDDCs) and monocyte-derived macrophages (MDMs) [205]. Pro-inflammatory cytokines such as TNF-α, IL-12, and IL-6 are produced by DCs infected with JEV; however, IL-10 is not produced by macrophages infected with JEV [206]. The microglial cells, which are CNS resident macrophages, could also be infected by JEV. Microglial cells play a critical role in the CNS during the JEV infection because they appear as a virus reservoir [207]. Activated microglia produce pro-inflammatory cytokines, namely TNF-α and IL-6, resulting in neuronal cell death [208]. Moreover, the study of Winter et al. [100], which was a large investigation into cytokines and chemokines for any acute viral encephalitis both in terms of the number of patients and number of investigated parameters, has demonstrated that an unregulated pro-inflammatory response in flavivirus encephalitis can be detrimental. TNF-α, IL-8, and IFN-α have been found to be related to poor outcomes in prior investigations of Japanese encephalitis (JE) patients [29, 54, 209]. In Winter et al.’s study [100], it has been established that TNF-α, IL-8, and IFN-α are critical in the development of JE, as well as the importance of IL-6 and RANTES. In addition, the West Nile virus and other flaviviruses have been shown to harm the CNS in a variety of animal models [210, 211].

Non-infectious, immune-mediated inflammation of the brain parenchyma is called autoimmune encephalitis. It can affect the cortex or deep gray matter, as well as white matter, the meninges, and the spinal cord [212-215]. Cytokines and chemokines in autoimmune encephalitis may give insight into the pathophysiology of autoimmune encephalitis. Studies in patients with NMDAR antibody-associated autoimmune encephalitis have shown an increase in serum IL-2 [172] and CSF IL-6, IL-17, CXCL-10, and IL-1β. Both innate and adaptive immunity, specifically Th1 and Th17, are involved in the secretion of these cytokines [13, 42, 151, 172, 216].

Monocytes and microglia have been found to produce CXCL13 [146]. it is a B cell chemoattractant that was found in the CSF of NMDAR antibody-associated autoimmune encephalitis patients. A decrease in CXCL13 levels was associated with a positive outcome after therapy [146]. The levels of BAFF and APRIL in the CSF of autoimmune encephalitis patients were linked to functional results in a study. According to the opposite findings, no increase in CSF levels of BAFF or APRIL was seen in research comparing NMDAR antibody-associated autoimmune encephalitis with viral encephalitis [217, 218].

Interferon-γ, IL-17, IL-12, and IL-23 levels in the CSF of autoimmune encephalitis patients with antibodies to cell surface proteins are greater than those of autoimmune encephalitis patients with antibodies to intracellular antigens [129]. According to recently published research, patients with NMDAR antibody-associated autoimmune encephalitis had increased IL-6, pentraxin-3, CD40L, and IL-17A in their CSF [141]. Patients with new-onset refractory status epilepticus who had autoimmune epilepsy were studied in a study. It was discovered that the CSF had higher IL-6, TNF-α, IL-2, and IL-4 and elevated levels of IL-6 and TNF-α in the peripheral tissues were observed [219]. Eighty-six percent of patients had improvement in seizure activity and a return to normal cytokine levels after receiving therapy with a monoclonal antibody targeting the IL-6 receptor (Tocilizumab) [219].

Our meta-analysis has several limitations. Some of the included cytokines lacked enough original studies to perform the meta-analysis. Despite performing subgroup analysis for the types of encephalitis (infectious or autoimmune), The etiologies of the disease, which may affect the concentration of cytokines in CSF or serum, were not included in the meta-analysis due to insufficient data. Moreover, due to inadequate data, the meta-regression for age, as a possible moderator variable, was not applicable in all cytokines and chemokines. Also, the sample size of most studies was small. Regarding the measurement methods, the studies used different methods and kits, which resulted in heterogeneity. Additionally, the heterogeneity was high in most of the meta-analyses. Sensitivity analysis was conducted to investigate the probable sources of heterogeneity, which remained high in some instances. To the best of our knowledge, this is the first meta-analysis investigating the serum and CSF concentrations of the cytokines in encephalitis. Also, the levels of the cytokines have been compared between autoimmune and infectious encephalitis in subgroup analysis.

5. Conclusions

This meta-analysis serves as evidence that encephalitis patients had greater CSF concentrations of IL-6, IL-8, IL-10, CXCL10, and TNF-α when compared to controls. TNF-α also showed increased concentration in serum. Moreover, it was observed that IL-10 had higher levels in autoimmune encephalitis compared to autoimmune encephalitis. Contrastingly, CXCL9 had a higher concentration in infectious encephalitis. Accordingly, the interleukin antagonists could be investigated as a potential adjunctive treatment in encephalitis. The diagnostic and prognostic sensitivity and specificity of these cytokines may also be evaluated in future studies. Furthermore, the prospect of antiviral medication and other viable treatment methods needs much more extensive exploration to avoid disease progression and the frequently severe sequalae after encephalitis infection.

Completed PRISMA checklist.

(PDF)

Click here for additional data file.

Full search strategy for each database.

(PDF)

Click here for additional data file.

Meta-analysis protocol registered in PROSPERO.

(PDF)

Click here for additional data file.

Funnel plots, drapery plot, and leave-one-out analyses for all cytokines.

(PDF)

Click here for additional data file.

Summary of serum/plasma meta-analyses.

(TIF)

Click here for additional data file.

Summary of CSF meta-analyses.

(TIF)

Click here for additional data file.

15 Jul 2022

30 Jul 2022

Reviewers' comments:

Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Comments From the Reviewer #1:

• Comment:

- Dear Authors,

This is a well-written study describing the alterations in the levels of different cytokines in patients with encephalitis. The authors reviewed dozens of articles and selected 23 of them for meta-analysis.

�  Response:

- Thank you very much for the time you put into reviewing our manuscript and thank you for your comments. We tried our best to address all your comments, and the detailed answers are in the following sections.

• Major comment:

- My major concern is that the authors write that 71 studies were eligible for study inclusion, but only 23 were included in the meta-analysis. However, the criteria for selecting 23 out of 71 were not described and should be explained.

�  Response:

- Thank you for your comment and for pointing this out. 71 studies reported data about changes in the concentration of cytokines in encephalitis patients, but only 23 had enough reliable quantitative data to be included in the meta-analysis. Most of the data available in the other studies were not quantitative. We did not have qualitative synthesis, and we only performed meta-analyses in our study. Based on your comment, the section was edited, and more explanation was added to clarify this item.

“Based on the inclusion and exclusion criteria, 71 studies [8, 23-30, 121, 125-185] were considered for this study, of which 23 studies had sufficient quantitative data and were included in the meta-analysis and reported in this study [25, 28-30, 126-128, 132, 135, 141-143, 145, 148, 150, 153, 157, 176, 179-181, 184, 185]. The remaining 48 studies were not included in the meta-analysis because they mostly reported qualitative data or did not contain sufficient quantitative data. Our study does not report qualitative data and only focuses on the meta-analysis of quantitative data.”

• Minor Comment:

- In the abstract, the authors write: “Also, these cytokines and chemokine might potentially prevent severe sequelae related to encephalitis infection.”

However, this sentence seems loosely related the manuscript, as the manuscript does not deal with the role of cytokines in the course of encephalitis and its compilations.

�  Response:

- Thank you very much for your comment and for pointing this out. We removed this sentence from the abstract.

Comments From the Reviewer #2:

• General Comments:

• Comment

- The authors undertake a meta-analysis of existing literature concerning cytokine/chemokine concentrations in serum/plasma and CSF samples from patients diagnosed with encephalitis compared to in-study controls. In principle, this is a study of interest that could inform diagnostic and therapeutic strategies to deal with encephalitis. However, as is common with meta-analysis studies, the high degree of variability between studies and the generalizations employed (ie. Infectious v. autoimmune encephalitis) lessens confidence in the findings. Despite these drawbacks, with additional selection criteria, subgroup analyses, and a clear and inclusive discussion of the limitations of this study, this work could be publishable.

�  Response:

- Thank you very much for your insightful and constructive comments. We carefully checked them and tried to address all of them. The detailed answers are in the following sections.

• Comment:

- While useful as a starting point, the assumption/generalization that the cytokine/chemokine profile for any given autoimmune encephalitis or encephalitis of an infectious origin is going to be similar is flawed and may be diluting important findings. Multiple studies within this meta-analysis describe the etiology of encephalitis. For example, Japanese encephalitis and NMDA-autoimmune patients are described in multiple studies with overlapping cytokine/chemokines analyzed. The authors should provide additional subgroup analyses that characterize confirmed cases to provide a more accurate cytokine/chemokine profile in those conditions. Conversely, analyzing diagnostically unconfirmed cases as a group may reveal a unique cytokine/chemokine signature and should be untaken.

�  Response:

- Thank you very much for your comment and for pointing this out. The number of studies that met our inclusion criteria to be considered for meta-analysis was limited. Due to this limitation, we only were able to conduct subgroup analysis for autoimmune encephalitis or encephalitis of infectious origin. Although, as you mentioned, it is of high importance to consider the etiology of the encephalitis in the meta-analysis, the data were not enough to perform a robust meta-analysis. Moreover, based on the extracted data, different origins of the encephalitides did not have too much difference in concentrations of cytokines and chemokines in most cases. However, based on your comments, this analysis was of high interest if it was possible, and we have added this problem to the limitations of the study, and we will consider it for further studies. The limitations are updated in the discussion part, which is as follows:

“Despite performing subgroup analysis for the types of encephalitis (infectious or autoimmune), The etiologies of the disease, which may affect the concentration of cytokines in CSF or serum, were not included in the meta-analysis due to insufficient data. Moreover, the meta-regression for age, as a possible moderator variable, was not applicable in all cytokines and chemokines due to inadequate data.”

• Comment:

- The authors state that (line 163) “Regarding the studies that had investigated more than one type of encephalitis, we pooled the mean and SD of the concentration of these groups”. In line with comment 1), this is inappropriate as written. Were the data pooled regardless of being encephalitis of infectious or autoimmune origin as this would imply? This needs to be clarified. Infectious vs. autoimmune data should be subgrouped and if the etiologic agent/condition is known, the data should be subgrouped again.

�  Response:

- Thank you for your delicate comment. We agree with you that this part is not well-stated. When there is more than one group (effect) in a study, the meta-analysis will lead to a unit-of-analysis issue (effects from one study may affect the overall effect more than they should, Cochrane Handbook, 2022, section 6.2.1). The “Cochrane Handbook for Systematic Reviews of Interventions” suggest multiple methods to deal with this problem and has stated that combining similar groups is the recommended method (v5, 2011, section 9.3.1). Other methods are also available such as multi-level meta-analysis or separating groups. Because a small number of the studies reported enough data, we used the combining method for groups with similar types of encephalitis in each study to overcome the unit-of-analysis issue. The combination of effects was performed whenever the types of encephalitis were similar. Moreover, in most cases, the etiology of the encephalitis was also similar. However, as you mentioned, the sentence was a little vague, and we changed and restated it as follows:

“When studies have reported concentrations of cytokines in more than one encephalitis group, if the types of encephalitis were similar (all groups had infectious encephalitis or all groups had autoimmune encephalitis), we pooled the mean and SD of the concentrations in those groups.”

• Comment:

- It does not appear that patient age is considered in these analyses. Several of these studies were performed in pediatric populations and some patient populations contained very old individuals. Cytokine/chemokine levels are differentially regulated at younger and older ages. Statistical tests should be run to ensure that age is not significantly influencing the analysis.

�  Response:

- Thank you so much for your comment and for pointing this out. Based on your comment, we conducted a meta-regression for “age” when it was possible. The mean or median age of participants of each study has been added to the meta-analysis as a moderator (meta-regression). Meta-regression was performed for cytokines/chemokines with more than two studies that reported mean or median age. The meta-regression process has been added to the method section, and the results of meta-regression for all cytokines/chemokines were added to the results section (age meta-regression for serum TNF-α, CSF CXCL13, CSF CXCL9, serum IL-10, serum IL-6, CSF IL-2 concentrations was not applicable). Moreover, the meta-regression demonstrated that IL-8 CSF concentration, CXCL13 serum concentration, and TNF-α CSF concentration are affected by the age of the studies as a moderator factor.

Meta-regression for age has been explained in the method section:

“Meta-regression was also performed based on the mean age of the participants of each study if it was possible (more than two studies with reported age were present in the meta-analysis; studies that did not report the mean or median age of the participants were omitted from meta-regression).”

• Comment:

- The discussion section needs to be largely rewritten. The second and the majority of third paragraphs (line 526-556) read as a textbook synopsis of adaptive and innate immune responses to HSV in the brain and largely do not relate directly to work presented.

Furthermore, little attention is paid to the limitations and caveats of this study, of which there are many. Discussion should include, patient and control populations age, gender, condition (ie. healthy v. nonspecific health conditions), location, immune experience and assay sensitivity.

�  Response:

- We appreciate your comment. We have thoroughly checked and edited the discussion section based on your comments, and the limitations of the study also have been revised and added. The discussion is hugely revised and rewritten, and the limitations have been added. The edited discussion is added to the revised manuscript files.

• Minor Comments:

• Comment:

- Line 73-74- “More recently, there has been a growing interest towards autoimmune encephalitis, such as Anti-N-Methyl-D-Aspartate Receptor (NMDAR)”. This is an incomplete sentence.

�  Response:

- Thank you for your comment. The revised sentence is as follows: “More recently, a growing number of studies have investigated biomarkers for diagnosis and therapeutic targeting of autoimmune encephalitides, such as Anti-N-Methyl-D-Aspartate Receptor (NMDAR) [2-5].”

• Comment:

- Line 83- “causative” agents.

�  Response:

- Done.

• Comment:

- Line 95-96- Cytokines are also produced by cells in the tissue. References should be provided for these statements.

�  Response:

- Thank you for your comment. References number 15-18 have been added.

15. Xie WR, Deng H, Li H, Bowen TL, Strong JA, Zhang JM. Robust increase of cutaneous sensitivity, cytokine production and sympathetic sprouting in rats with localized inflammatory irritation of the spinal ganglia. Neuroscience. 2006;142(3):809-22.

16. DeLeo JA, Colburn RW, Nichols M, Malhotra A. Interleukin-6-mediated hyperalgesia/allodynia and increased spinal IL-6 expression in a rat mononeuropathy model. J Interferon Cytokine Res. 1996;16(9):695-700.

17. Schäfers M, Svensson CI, Sommer C, Sorkin LS. Tumor necrosis factor-alpha induces mechanical allodynia after spinal nerve ligation by activation of p38 MAPK in primary sensory neurons. J Neurosci. 2003;23(7):2517-21.

18. Heijmans-Antonissen C, Wesseldijk F, Munnikes RJ, Huygen FJ, van der Meijden P, Hop WC, et al. Multiplex bead array assay for detection of 25 soluble cytokines in blister fluid of patients with complex regional pain syndrome type 1. Mediators Inflamm. 2006;2006(1):28398.

• Comment:

- Line 103-104- Chemokine are also produced by cells in tissue.

�  Response:

- Thank you for your comment. This part was edited in the manuscript.

“Chemokines are a group of cytokines that are released mainly by leukocytes to induce chemotaxis in damaged tissues and attract white blood cells. In addition, chemokines could be secreted by tissue-resident cells, such as neurons and glial cells in the brain [20-22].”

20. Mélik-Parsadaniantz S, Rostène W. Chemokines and neuromodulation. J Neuroimmunol. 2008;198(1):62-8.

21. Ubogu EE, Cossoy MB, Ransohoff RM. The expression and function of chemokines involved in CNS inflammation. Trends Pharmacol Sci. 2006;27(1):48-55.

22. Jaerve A, Müller HW. Chemokines in CNS injury and repair. Cell Tissue Res. 2012;349(1):229-48.

• Comment:

- Line 108- Lower case “chemokines”.

�  Response:

- Done.

• Comment:

- Line 123- Wrong protocol is linked.

�  Response:

- Thank you for your comment. The URL has been edited in the manuscript.

https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021289298

• Comment:

- Line 135- All patients with cancer should be excluded, correct? Tumors for example can change basal cytokine/chemokine levels themselves or induce changes in secretion by other cells.

�  Response:

- Thank you for your comment. We agree with you. Cancer can affect the concentration of the cytokine/chemokine in CSF and serum. Included studies did not contain cancer patients, and paraneoplastic encephalitis patients were the only cancer patients. Based on your comment, we added cancer as an exclusion criterion, but we also keep paraneoplastic encephalitis as an exclusion to also emphasize this group. So, the item has been changed as follows:

“6) All cancer patients and patients with paraneoplastic encephalitis, because of the existence of cancerous tissue or remote neoplasia that might dysregulate cytokines profile.”

• Comment:

- Line 206- add “in”

�  Response:

- Done.

• Comment:

- Can not see all of tables 1 and 3. Transform to landscape.

�  Response:

- Done.

• Comment:

- Table 2- Should 2: be 0:?

�  Response:

- Yes, done.

• Comment:

- Line 370- add “the”

�  Response:

- Done.

• Comment:

- Line 627-629- Heterogeneity was high in “most” of the meta-analyses.

�  Response:

- Done.

• Comment:

- Line 640- “Accordingly, the interleukin antagonists should be investigated as adjunctive treatment in encephalitis.”- This may be flawed logic as the induced immune response if often critical to resolution of infectious encephalitis. This language should be tempered.

�  Response:

- Thank you for your comment. The sentence has been tempered as follows:

“Accordingly, the interleukin antagonists could be investigated as a potential adjunctive treatment in encephalitis.”

18 Aug 2022

Cytokines and chemokines profile in encephalitis patients: a meta-analysis

PONE-D-22-04861R1

Dear Dr. Rezaei,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Karin E. Peterson, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Dear Authors,

The manuscript has been improved substantially. All comments have been addressed.

I have no additional remarks.

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

**********

23 Aug 2022

PONE-D-22-04861R1

Cytokines and chemokines profile in encephalitis patients: a meta-analysis

Dear Dr. Rezaei:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Karin E. Peterson

Academic Editor

PLOS ONE