Neurocysticercosis, caused by the larvae of the pork tapeworm Taenia solium (TsM), is the most common helminth infection of the CNS in humans worldwide^[1]. Neurocysticercosis (NCC) is also the leading cause of epilepsy in developing world.

Immune mechanisms are involved in the pathogenesis of neurocysticercosis. Cysticercal lesions develop in brain depending upon a combination of host immune-inflammatory response, mainly mediated by cytokines produced by cysticercal antigens^[2]. The invasion of the central nervous system by the TsM induces a blood brain barrier (BBB) disruption. As the parasite dies, the bioactive molecules of Taenia solium metacestode are released, and a strong inflammatory response are induced, that may contribute to the development of clinical symptoms and signs^[3].

Thus, disorder of immune system may involve in pathogenesis of NCC. We conducted a meta-analysis of CSF, sera and non TsM antigen stimulated peripheral blood mononuclear cell (PBMC) cytokines, cytokine receptors and soluble intercellular adhesion molecule (sICAM) levels in neurocysticercosis so as to investigate the influence of TsM infection to host immune responses in the neurocysticercosis patient in comparison with the control subject. Additionally, we discussed the effect of different body fluid components, different assay methods and the influence of anti-parasite treatment on cytokine- and cytokine receptor levels.

Methods and materials

A protocol for the review was prepared before the data collection.

Eligibility criteria

Inclusion and exclusion criteria were determined before the search was conducted. We included human studies comparing patients with neurocysticercosis with the controls, in which one or more of the following inflammatory markers was measured in the CSF, sera and/or no TsM antigen stimulated PBMC: IFN-γ , IL-2, IL-2 receptor (sIL-2R) , IL-4 , TNF-a, IL-6, IL-8, IL-10, IL-23, IL-17 and sICAM-1 . The exclusion criteria were: (i) case reports or very small case series; (ii) same patient data (such as duplicate publication); (iii) review articles, editorial, letters, author reply, comments, erratum, conference proceedings; (iv) insufficient data to reassess sensitivity or specificity from individual studies; (v) articles not within the field of our study.

Search strategy

The systematic search was performed by Chen XH and Wang F in PubMed, CNKI, Wan Fang and VIP Web of Science from 1990 to December 2014. The search strategy employed the following keywords: inflammation, cytokine, interleukin, inflammatory markers, tumor necrosis factor, immune and neurocysticercosis. Reference lists of retrieved articles were also searched by hand.

Study selection and data extraction

Data were collected using a standard data extraction form (Table 1). Information from articles in a language of English or Chinese was extracted.

Tab.1

Tab.1

Tab.1 Characteristics of the included studies Ref Author, year Subjects N (male/female) Age Cytokines measured 1 YanXiaobo 2004 NCC 41(21/20) 30.6(13-45) IFN-γ , TNF-α Control 20(12/8) 33(18-50) BBB disorder BBB normal 2 XuHongxiu2001 NCC 30(18/12) 10-62 IFN-γ , IL-4, 5, 6, 10 Control 31 before treatment after treatment 3 WangYing 2002 NCC 30(18/12) 43.14± 10.12(15-60) IFN-γ , TNF-α , sIL-2R, Control 30(18/12) 30.5± 4.2 IL-8 and sICAM-1, IL-2, 4 WenTingyuan2002 NCC 32(17/15) 14.8± 6.94(7-21) IL- 2, IL-6 Control 30(15/15) 16.4± 5.48(10-21) 5 Yehong2005 NCC 20(11/9) 14-54 IFN-γ , TNF-α , IL-2, Control 20/ / IL-10, IL-4 before treatment after treatment 6 Zhang Weizhe1997 NCC 31(17/14) 22-45 IFN-γ , TNF-α , IL-2, Control 30/ / 7 Wangfeng2003 NCC 40(23/17) 38.3± 6.5 TNF-α , IL-1β Control 25 epilepsy 23 non-epilepsy 17 8 Sunxiaohong2006 NCC 28(16/12) 38.5 TNF-α Controls 26(13/11) 28.6 9 liujunnian2000 NCC 30/ / TNF-α control 15 / 10 Liuruichun2006 NCC 45(28/17) 12-69(32.7) IL-6, SIL-2R Control 43 epilepsy 23 12-69 non-epilepsy 22 16-53 11 Wangminghui2001 NCC 17 5-11 TNF-α Controls 11 before-treatment after-treatment 12 Chenchuan2000 NCC 30(19/11) 14-67 TNF-α Controls 40 13 Zhoufen2002 NCC 76(49/27) 6-13 IL-2, SIL-2R, IL-6, IL-8 Controls 30(17/13) 7-12 14 Avantika Verma 2011 symptomatic 32(18/14) 24.66± 11.03 IFN-γ , TNF-α , IL-4, IL-10, asymptomatic 36(21/15) 27.40± 14.94 IL-17, IL-23, sICAM-1

Tab.1 Characteristics of the included studies

Tab.1

Tab.1

Tab.1 (Continued) Ref Author, year Subjects N (male/female) Age Cytokines measured Control 37(20/17) 25.06± 13.51 15 Wangxiaojuan2004 NCC 49(27/22) 33± 4.5 TNF-α StageⅠ (alive) 11 stageⅡ (early stage of degenerated and died) 14 stageⅢ (late stage of degenerated and died) 10 stageⅣ (inactive) 14 Controls 20 unknown

Tab.1 (Continued)

Two reviewers (Zou Y. and Tian XJ) independently screened all titles, abstracts and full texts of selected articles and conducted the data extraction and the quality assessment. Disagreements were resolved by consensus.

Cytokine, cytokine receptor and soluble intercellular adhesion molecule levels (mean and standard deviation) were extracted for each group of study participants along with relevant clinical data (age, sex, anti-parasite treatment, epilepsy or not, ) and information on assay method and body fluid fractions was analyzed. For two studies evaluating anti-parasite treatment on TNF-α level of neurocysticercosis patients, the levels of pre-treatment were chosen for comparison with post-treatment.

Review Manager v.5.2 (Cochrane Collaboration, Oxford, United Kingdom) was used for the analysis. We chose to calculate a standardized mean difference (SMD) for the principal summary measure for every cytokine included because of the variable assay types used (i.e. Flow Cytometry and ELISA) in evaluating cytokines, cytokine receptors and soluble intercellular adhesion molecule concentrations, and because different body fluid components were analyzed in the included studies, yielding very varying concentrations between studies. The standardized mean difference was calculated using a random effects model, which is more appropriate than a fixed-effect model when the true effect likely varies between studies due to heterogeneity not readily explained.

Cytokine parameters evaluated in only one study were not included in the quantitative analysis. Heterogeneity was investigated for combined measures by the chi-square test and inconsistency, indicating the impact of heterogeneity on the meta analysis, was calculated using the I-square value. Possible sources of significant heterogeneity were assessed and sensitivity analysis and subgroup analysis were performed where data permitted this.

Results of the data analysis were presented overall for neurocysticercosis patients versus control subjects, with subgroup results according to different body fluid component (Serum, PBMC and CSF), different measure method (ELISA or Flow Cytometry) or whether received anti parasite treatment (before treatment and after treatment) where applicable.

Results

Study selection

The search yielded 87 articles, of which 15 were included in the meta-analysis^{[4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]}. Figure 1 depicts the study selection process.

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 Flowchart of literature search, study screening and selection process

In total 904 individuals (518 patients with neurocysticercosis) were included. Patients with neurocysticercosis and controls were comparable as to age and gender (Table 2). The neurocysticercosis was diagnosed by the physician according to clinical symptoms, CSF/serum diagnosis and image examination. The sample of control comes from really healthy population or other diseases of differential diagnosis from neurocysticercosis.

Tab.2

Tab.2

Tab.2 Cytokine expression in patients with neurocysticercosis compared with those of the control group Outcome N studies neurocysticer cosis patients controls SMD (95% CI) pValue Heterogeneity References χ 2 pValue I2 IFN-γ 6 261 225 -19.13 [-29.84, -8.41] 0.000 5 595.9 < 0.000 01 99% Avantika Verma 2011, WangYing 2002, XuHongxiu2001, YanXiaobo 2004, Yehong2005, Zhang Weizhe1997, IL-2 5 189 140 -4.60 [-5.93, -3.26] < 0.000 01 50.48 < 0.000 01 92% WangYing 2002, WenTingyuan2002, Yehong2005, Zhang Weizhe1997, Zhoufen2002 TNF-α 11 425 389 0.49 [0.22, 0.76] 0.000 4 994.60 < 0.000 01 99% Avantika Verma 2011, Chenchuan2000, l iujunnian2000, Sunxiaohong2006, Wangfeng2003, Wangminghui2001, WangYing 2002, YanXiaobo 2004, Yehong2005, Zhang Weizhe1997, Wangxiaojuan2004 IL-4 3 118 125 20.46 [12.11, 28.81] < 0.000 01 1 201.01 < 0.000 01 100% Avantika Verma 2011, XuHongxiu2001, Yehong2005 IL-10 3 118 125 69.78 [40.20, 99.36] < 0.000 01 5 283.93 < 0.000 01 100% Avantika Verma 2011, XuHongxiu2001, Yehong2005 sIL-2R 3 196 146 72.95 [51.84, 94.07] < 0.000 01 5 588.72 < 0.000 01 100% Liuruichun2006, WangYing 2002, Zhoufen2002 IL-6 4 228 177 10.96 [-40.77, 62.68] 0.68 12 661.39 < 0.000 01 100% Liuruichun2006, WenTingyuan2002, XuHongxiu2001, Zhoufen2002 IL-8 2 152 60 54.37 [-49.75, 158.48] 0.31 10 127.29 < 0.000 01 100% WangYing 2002, Zhoufen2002

Tab.2 Cytokine expression in patients with neurocysticercosis compared with those of the control group

IFN-γ

IFN-γ was investigated in six studies comprising 261 neurocysticercosis patients and 225 controls (WangYing 2002, XuHongxiu2001, YanXiaobo 2004, Yehong2005, Zhang Weizhe 1997, Avantika Verma 2011). Concentrations of IFN-γ were significantly lower in neurocysticercosis patients compared with controls (SMD-19.13 [-29.84, -8.41], P=0.000 5) (Figure 2). Removing any component subgroup (CSF, Serum or PBMC) did not change the overall result.

Fig.2

	Figure Option ViewDownloadNew Window
	Fig.2 Forest graph shows the meta-analysis of the IFN-γ expression with the neurocysticercosis patients and controls CI: confidence interval.

IL-2

Five studies evaluated IL-2, involving 189 neurocysticercosis patients and 140 controls (WangYing2002, WenTingyuan2002, Yehong2005, Zhang Weizhe1997, and Zhoufen2002). neurocysticercosis patients had significantly lower concentrations of IL-2 compared with controls (SMD -4.60, CI: [-5.93, -3.26], P < 0.000 01) (Figure 3). Removing the different body fluid subgroup (Serum or PBMC) did not change the overall result.

Fig.3

	Figure Option ViewDownloadNew Window
	Fig.3 Forest graph shows the meta-analysis of the IL-2 expression with the neurocysticercosis patients and controls CI: confidence interval.

TNF-α

Levels of TNF-α were examined in eleven studies (YanXiaobo 2004, WangYing 2002, Yehong2005, Zhang Weizhe1997, Wangfeng2003, Sunxiaohong2006, liujunnian2000, Wangminghui2001, Chenchuan2000, Avantika Verma 2011, Wangxiaojuan2004) involving a total of 425 neurocysticercosis patients and 389 controls. Overall, concentrations of TNF-α were higher among neurocysticercosis patients compared with controls (SMD 0.49, CI: [0.22, 0.76], P=0.000 04) (Figure 4). Removing different body fluid component subgroup (CSF, Serum or PBMC) did not alter the overall result.

Fig.4

	Figure Option ViewDownloadNew Window
	Fig.4 Forest graph shows the meta-analysis of the TNF-α expression with the neurocysticercosis patients and controls CI: confidence interval.

IL-4

Concentrations of IL-4 were evaluated among 118 neurocysticercosis patients and 125 controls in three studies (XuHongxiu2001, Yehong2005, and Avantika Verma 2011). In all, levels of IL-4 were significantly higher in neurocysticercosis patients compared with controls (SMD20.46, CI: [12.11, 28.81], P< 0.000 01) (Figure 5). Removing the different body fluid subgroup (serum or PBMC) did not alter the overall result.

Fig.5

	Figure Option ViewDownloadNew Window
	Fig.5 Forest graph shows the meta-analysis of the IL-4 expression with the neurocysticercosis patients and controls interval. CI: confidence

IL-10

Concentrations of IL-10 were evaluated among 118 neurocysticercosis patients and 125 controls in three studies (Avantika Verma 2011, XuHongxiu2001, and Yehong 2005). Over all, levels of IL-10 were significantly higher in neurocysticercosis patients compared with controls (SMD69.78, CI: [40.20, 99.36], P< 0.000 01) (Figure 6). Removing any body fluid subgroup (serum or PBMC) did not alter the overall result.

Fig.6

	Figure Option ViewDownloadNew Window
	Fig.6 Forest graph shows the meta-analysis of the IL-10 expression with the neurocysticercosis patients and

sICAM-1

Levels of sICAM-1 were examined in two studies (Avantika Verma 2011, WangYing 2002) involving a total of 98 neurocysticercosis patients and 104 controls. The concentrations of sICAM-1 were higher among neurocysticercosis patients compared with controls (SMD 0.62, CI: 0.36-0.88, P< 0.001) (Figure 7). As all data of sICAM-1 come from serum of subjects, no different body fluid subgroup could be compared in this group.

sIL-2R

sIL-2R was investigated in three studies comprising 196 neurocysticercosis patients and 146 controls (Liuruichun2006, WangYing 2002, and Zhoufen2002). Concentrations of sIL-2R were significantly higher in neurocysticercosis patients compared with controls (SMD 72.95, CI: [51.84, 94.07], P< 0.000 01). Whereas, removal of the serum subgroup resulted in significantly lower sIL-2R (SMD -2.37, CI: [-2.78, -1.96], P< 0.000 01), which was supported by only one study (Liuruichun2006) (Figure 8).

Fig.7

	Figure Option ViewDownloadNew Window
	Fig.7 Forest graph shows the meta-analysis of the sICAM-1 expression with the neurocysticercosis patients and controls CI: confidence interval.

Fig.8

	Figure Option ViewDownloadNew Window
	Fig.8 Forest graph shows the meta-analysis of the sIL-2R expression with the neurocysticercosis patients and controls

IL-6 and IL-8 CI: confidence interval.

There were no significant differences overall, for replicated evaluations of IL-6 and IL-8 (Table 2, Figure 9 & Figure 10). In the subgroup analysis, no overall result was changed when removing any body fluid component subgroup but for IL-6, where removal of the serum subgroup resulted in an overall significantly higher standardized mean difference, indicating higher levels among neurocysticercosis patients compared with control subjects in CSF. This, however, was based on one remaining study (Liuruichun2006).

Fig.9

	Figure Option ViewDownloadNew Window
	Fig.9 Forest graph shows the meta-analysis of the IL-6 expression with the neurocysticercosis patients and controls CI: confidence interval.

Fig.10

	Figure Option ViewDownloadNew Window
	Fig.10 Forest graph shows the meta-analysis of the IL-8 expression with the neurocysticercosis patients and controls CI: confidence interval.

IL-5, IL-1β , IL-23 and IL-17

Concentrations of IL-5, IL-1β , IL-23 and IL-17 were only evaluated in one study respectively, thus not yielding replicated findings.

Correlation with clinical features

The influence of anti parasite treatment used in individual studies was evaluated where study numbers allowed this. Correlation between prevention of anti parasite treatment and cytokine parameter levels were reported in two studies for TNF-a, with no replication of positive findings. (WangYing 2002, Zhoufen 2002). There is no significant findings in TNF-α level in before-treatment and after treatment (SMD: -3.25, CI: [-8.72, 2.21], P=0.24) (Figure 11). As no enough data, correlation between clinical features (BBB disorder vs. BBB normal, epilepsy vs. non-epilepsy, symptomatic vs. asymptomatic and different clinic stages) and cytokine parameter levels could not be compared by meta analysis.

Fig.11

	Figure Option ViewDownloadNew Window
	Fig.11 Forest graph shows the meta-analysis of the TNF-α expression with the neurocysticercosis patients before treatment and after treatment CI: confidence interval.

Heterogeneity and sensitivity analysis

Heterogeneity in the standardized mean difference was assessed for each cytokine for all replicated comparisons between neurocysticercosis patients and control subjects, showing signs of significant heterogeneity for all comparisons (Table 2).

In the sensitivity analysis, the influence of body fluid components used in individual studies (CSF, Serum or PBMC) was evaluated where study numbers allowed this. For the overall significant finding of IFN-γ , IL-2, TNF-α , IL-4, IL-10 and sICAM-1, removal of studies analyzing CSF, serum or PBMC samples had no impact on results. Of the studies with overall no significant findings in IL-8, results were not altered when removing serum studies analyzing. but for sIL-2R and IL-6, where the overall concentration of sIL-2R and IL-6 became significantly higher among neurocysticercosis patients compared with control subjects when removing the rest studies analyzing on serum.

In addition, sensitivity analysis was performed, where possible, for the impact of assay type used for cytokine analysis. For the comparisons with overall significant standardized mean differences between neurocysticercosis patients and control subjects, the result of IFN-γ level remained significant after removal of studies employing flow cytometric assays (Figure 12).

Fig.12

	Figure Option ViewDownloadNew Window
	Fig.12 Forest graph shows the meta-analysis of the IFN-γ expression measured by ELISA, flow cytometry and other methods CI: confidence.

Discussion

This meta-analysis included 15 studies, totally 518 neurocysticercosis patients in various stage of infection and 386 control subjects. Of the 13 different cytokines, cytokine receptor and soluble inter-cellular adhesion molecule investigated, 9 were examined in multiple studies with overall significant findings for six cytokine parameters. Specifically, the meta-analysis demonstrated significantly elevated levels of TNF-α , IL-4, IL-10 and sICAM-1 in neurocysticercosis patients compared with controls. Overall depressed levels of IFN-γ and IL-2 in neurocysticercosis patients compared with control subjects were significant. There were no significant differences in levels of sIL-2R, IL-6 and IL-8 between neurocysticercosis patients and control subjects, while there were no replicated comparisons for IL-5, IL-1β , IL-23 and IL-17. The sensitivity analysis revealed no clear effect of assay type used for the analysis and no significant findings in TNF-α level in before-treatment and after treatment, but sIL-2R and IL-6 may have different expression characters in CSF and serum (Liuruichun 2006).

Primarily, Th1 and Th2 immune response were involved in the brain granulomas nearby TsM lesions. The brain granulomas in neurocysticercosis patients^[19], naturally infected swine^[20] and murine^[21] were associated with a Th1 (IFN-γ and IL-18 )and Th2 profile(IL-4, IL-13, and IL-10). In CSF, Th2 cytokines IL-5, IL-6, IL-10 and proinflammatory cytokine TNF-α were elevated of inflammation and active nuerocysticercosis patients compared with non-inflammation or inactive of those^{[22, 23, 24]}. As for reasons such as not expressed by mean SD, no control group nor whole text available, those articles were not include in meta analysis. Whereas, those results were partly consistent with our meta analysis result which indicates that pro-inflammation and Th2 cytokines were predominant in CSF of NCC immune, especially in symptomatic and inflammatory patients. But for expression of Th1 cytokines, such as IFN-γ in CSF, serum and PBMC (TsM antigen stimulated), the result was paradoxical. It is supposed that IFN-γ level was no significant difference in CSF of inflammation and non-inflammation neurocysticercosis patients and a little lower in CSF of active NCC and inactive NCC patients than those of controls^{[24, 25]}. But in Taenia solium cyst fluid antigens stimulated PBMC, IFN-γ level was significantly higher in symptomatic compared with asymptomatic neurocysticercosis patients^[26]. In references included in our meta analysis, only one result support IFN-γ in elevated in serum of NCC patients(Avantika Verma 2011), but most studies showed that which was lower in central and peripheral(serum and no TsM antigen stimulated PBMC) of NCC group than those of controls(YanXiaobo 2004, XuHongxiu2001, WangYing 2002, Yehong2005, Zhang Weizhe1997). The above results imply that NCC has different immune characters in central and peripheral immune system. Th2 cytokines were predominant in NCC immune response, whereas the trend of Th1 cytokines still needs more researches.

While the interplay between the peripheral immune system and the central immune system is not clearly understood, there is some evidences shown that neurocysticercosis is a systemic condition. There is no influence of removing the PBMC subgroup on overall significant finding regarding IFN-γ , IL-2, TNF, IL-4 and IL-10, which implies that certainly cytokines' character is identical in CNS and peripheral system of neurocysticercosis patients. However, the overall no significant finding regarding sIL-2R and IL-6 became significant after removing the PBMC subgroup. Although which was based on two comparisons in a single study (Liuruichun2006), the result still implies that certainly cytokines' character is different in CNS and peripheral system of neurocysticercosis patients.

Several limitations in the present study need attention. First, the collection criteria of control were different which could influence cytokine parameter levels. Second, the collection criteria of neurocysticercosis patients was obscure, the inclusion of multi-stages patients could pose significant challenges. Third, the inclusion of studies employing various assay procedures and analyzing different body fluid components, could potentially introduce heterogeneity and bias results.

Our sensitivity analysis, demonstrated no significant impact of assay type on overall significant results. The present results should be interpreted with caution because of significant heterogeneity between studies and low study numbers for some cytokine measures. There are some measures to minimize heterogeneity within studies: strictly criterion of enrolled controls, maximize excluded out the disease impacted cytokine levels; matched pair samples in central and peripheral system; chosen sensitive and consistent laboratory methods, and assessing intra-individual alteration of immune responses.

Conclusion

In conclusion, this meta-analysis demonstrated decreased level of IFN-γ and IL-2, but elevated levels of TNF-α , IL-4, IL-10 and sICAM-1 in neurocysticercosis patients compared with controls, which implies disorder of Th1/Th2 and predominant of pro-inflammation cytokines were involved in the pathogenesis of the disease. Further researches on the role of cytokine played in central and peripheral immune system in neurocysticercosis are warranted.