Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Role of Peripheral Inflammatory Markers in Postoperative Cognitive Dysfunction (POCD): A Meta-Analysis

  • Linying Peng ,

    Contributed equally to this work with: Linying Peng, Liwei Xu

    Affiliation Department of Anesthesiology, The Third Xiangya Hospital of Central South University, Changsha, China

  • Liwei Xu ,

    Contributed equally to this work with: Linying Peng, Liwei Xu

    Affiliation Department of Gastrointestinal Surgery, The First Xiangya Hospital of Central South University, Changsha, China

  • Wen Ouyang 

    Affiliation Department of Anesthesiology, The Third Xiangya Hospital of Central South University, Changsha, China

Role of Peripheral Inflammatory Markers in Postoperative Cognitive Dysfunction (POCD): A Meta-Analysis

  • Linying Peng, 
  • Liwei Xu, 
  • Wen Ouyang



Postoperative cognitive dysfunction (POCD) is common following cardiac and non-cardiac surgery, but the pathogenic mechanisms remain unknown. Many studies suggest that an inflammatory response is a key contributor to POCD. The current meta-analysis shows that the levels of peripheral inflammatory markers are associated with POCD.


An online search was performed to identify peer-reviewed studies without language restriction that measured peripheral inflammatory markers of patients with and without POCD, using PubMed, ScienceDirect, SinoMed and the National Knowledge Infrastructure database. Extracted data were analyzed with STATA (version 12).The standardized mean difference (SMD) and the 95% confidence interval (95%CI) were calculated for each outcome using a random effect model. Tests of heterogeneity assessment of bias, and meta-regression were performed in the meta-analysis.


A total of 13 studies that measured the concentrations of peripheral inflammatory markers were included. The current meta-analysis found significantly higher concentrations of S-100β(SMD[95%CI]) (1.377 [0.423, 2.331], p-value < 0.001, N [POCD/non-POCD] =178/391, 7 studies), and interleukin(IL)-6 (SMD[95%CI]) (1.614 [0.603,2.624], p-value < 0.001, N[POCD/non-POCD] = 91/99, 5 studies), but not of neuron specific enolase, interleukin-1β, or tumor necrosis factor-α , in POCD compared with patients without POCD. In meta-regression analyses, a significant positive association was found between the SMD and the preoperative interleukin-6 peripheral blood concentration in patients with POCD (Coef.= 0.0587, p-value=0.038, 5 studies).


This study shows that POCD is indeed correlated with the concentrations of peripheral inflammatory markers, particularly interleukin-6 and S-100β.


Postoperative cognitive dysfunction (POCD) usually manifests as an alteration in orientation, memory, thinking, attention, insight or other aspects of central nervous function. Initially, it was thought to be associated with cardiac surgery. However, later studies showed that it has been associated with non-cardiac surgery and even with non-invasive procedures such as coronary angiography [1]. POCD can last for a few days to a few years. It decreases the patient’s quality of life and increases the cost of hospitalization and out-of-hospital care. It also increases surgical morbidity and mortality [2,3]. In patients over the age of 60, POCD was observed in 25.8% of patients at 1 week post-surgery and in 9.9% of patients at 3 months post-surgery [4].

Thus, POCD is an important concern for the anesthesiologist.

Considerable evidence suggests that an inflammatory response may be involved in the occurrence of POCD [5,6]. In the clinic, anesthesiologists regularly study the correlation between POCD and factors such as S-100β protein(S-100β), neuron specific enolase(NSE), interleukin-1β(IL-1β), interleukin-6(IL-6), interleukin-8(IL-8), interleukin-10(IL-10), tumor necrosis factor(TNF)-α, and C-reactive protein(CRP). However, direct evidence showing a relationship between POCD and inflammatory markers is lamentably absent from the literature. Furthermore, opposing results are observed at same experimental conditions and experimental designs. Therefore, we conducted a meta-analysis to pool and analyze the data and to determine the relationship between POCD and specific inflammatory markers.

Materials and Methods

Data Sources and Search Strategy

All analyses were performed according to PRISMA guidelines [7] and the Cochrane handbook for systematic reviews of interventions. PRISMA guidelines focus on randomized trials, but the PRISMA statement specifies that “PRISMA can also be used as a basis for reporting systematic reviews of other types of research”. Literature was searched using PubMed, ScienceDirect, SinoMed and the National Knowledge Infrastructure database. Searches were performed using the key words POCD and S-100β, NSE, IL-1β, IL-6, IL-8, TNF-α up to April 2013. All articles selected were included only human studies. The reference lists of relevant studies were searched for additional reports. No standardized review protocol has been published.

Study Selection

Original studies measuring inflammatory marker concentrations in living subjects with POCD were included. Inclusion criteria were as follows: 1.) case-control study including non-POCD subjects as controls; 2.) human subjects; 3.) explicit diagnostic criteria such as the Confusion Assessment Method [8], the Digit-Symbol-Substitution Test of the Wechsler Adult Intelligence Scale (as a measure of attention) [2,9], and a diagnosis based on DSM-III criteria, examined using the Mini-mental State Examination [10,11] and so on.

Studies were excluded if: 1.) the article was a case report; 2.) the trials measured inflammatory marker concentrations from peripheral blood cells following additional stimulation; 3.) the trials did not group the results according to POCD status; 4.) the date were in an unavailable format.

Data Extraction

Three independent raters examined each retrieved article. The results were compared between raters and any disagreements regarding inclusion were settled by consensus. The methods and results sections of each relevant article were analyzed, and cytokine concentration metrics for POCD and non-POCD groups were extracted, in the formats: mean (standard deviation, SD), median (interquartile range, IQR) or median (range, R). If data were presented in the format from which means and SDs were not extractable, these measures were tried to request from the corresponding author of the publication. If data were presented in the format median(R), our simulations show that the formula “R/4” is the best estimator of the standard deviation (variance) for moderately sized samples (15<n≤70). For large samples (n > 70), the formula “R/6” gives the best estimator for the standard deviation [12]. If data were presented in the format median (IQR), then the formula “IQR/1.35” was used to calculate the standard deviation [13].

Quality score evaluation

Three authors independently assessed the qualities of included studies using the Newcastle-Ottawa scale (NOS) [14]. The NOS ranges between zero and nine stars. A quality score was calculated on the basis of 3 major components of case–control studies: selection of study groups (0–4 stars), comparability of study groups (0–2 stars), and ascertainment of the exposure of interest (0–3 stars). Studies with a score of seven stars or greater were considered to be of high quality. Disagreement was settled as described above. (Table S2)

Statistical Analyses

The standardized mean difference (SMD) and 95% confidence interval (95%CI) were calculated for each outcome using a random effect model (the DerSimonian and Laird method). Random effect models were preferable if significant heterogeneity was expected because they assumed and accounted for variable underlying effects in estimates of uncertainty, including both within-study and between-study variance. A Q statistic was calculated using a chi-squared test to quantify the heterogeneity among combined results. A significant Q statistic indicates diversity in the characteristics of the combined trials. Inconsistency was calculated using an I2 index to determine the impact of heterogeneity.

To identify potential sources of heterogeneity, planned subgroup analyses were conducted in studies where control subjects were matched for age and gender, in studies excluding subjects with inflammatory comorbidity, and in studies using comparable assay methodology. Planned study level meta-regression analyses were conducted comparing SMDs to pre-operative peripheral inflammatory markers, age, and gender. Sensitivity analysis was used to investigate the influence of each individual study on the overall meta-analysis summary estimate.

The risk of publication bias was assessed using funnel plots and Egger’s tests. Funnel plots were used to plot the effect size of the component studies against the sample size, and Egger’s tests were used to detect skewness in the funnel plots, and by extension, publication bias in the data. The risk of selective reporting bias was examined within the studies. When the difference between groups was not significant, Altman’s method for describing confidence intervals was used. Meta-regression analysis (using metareg) was used to test whether effect sizes were influenced by specific study design features. Analyses were performed using STATA, version 12.0 (Stata Corporation, College Station, Texas, 2012)


Literature Search Findings

We included 13 eligible studies [2,811,1522] in our meta-analysis (Table S1). In this meta-analysis, we included 7 studies examining S-100β, 3 studies of IL-1β, 5 studies of IL-6, 4 studies of TNF-α, and 6 studies of NSE. Figure 1 shows the study selection procedure.

Cytokine Concentrations

Significantly higher peripheral blood inflammatory marker concentrations (SMD[95%CI]) were detected in POCD subjects compared with non-POCD subjects when examining S-100β(1.377 [0.423 , 2.331], p-value < 0.001 N[POCD/non-POCD] = 178/391,7 studies, Figure 2), and IL-6(1.614 [0.603,2.624], p-value < 0.001, N[POCD/non-POCD ]= 91/99 ,5 studies, Figure 3), but not when examining NSE, IL-1β, or TNF-α. (Table 1)

Figure 2. Studies of peripheral blood S-100β.

Forest plot for S-100β displaying the results of the random effects meta-analysis for the association between peripheral blood S-100β and postoperative cognitive dysfunction (POCD). The standardized mean difference is presented on the horizontal axis; positive values denote higher serum levels in POCD patients; negative values denote higher serum levels in control subjects. The dashed line is at the overall standardized mean difference. CI, confidence interval; SMD, standardized mean difference.

Figure 3. Studies of peripheral blood IL-6.

Forest plot for IL-6 displaying the results of the random-effects meta-analysis for the association between peripheral blood IL-6 concentrations and POCD. The standardized mean difference is presented on the horizontal axis; positive values denote higher serum levels in patients; negative values denote higher serum levels in control subjects. The dashed line is at the overall weighted mean difference. CI, confidence interval; SMD, standardized mean difference.

Main EffectHeterogeneity
CytokineStudiesN (P/NP)SMD95% CIZpChi2dfpI2
S-100β7178/391 1.377 0.423, 2.331 2.830.005117.3560.00094.9%
IL-6591/991.614 0.603,2.624 3.130.00243.78 40.00090.9%
TNF-α486/1050.374 -0.341,1.0891.030.30516.3130.00181.6%

Table 1. Summary of comparative outcomes for each measured peripheral inflammatory marker.

S-100β, S-100β protein; IL-6, interleukin-6; NSE, neuron specific enolase; IL-1β, interleukin-1β; TNF-α, tumor necrosis factor-α;N(P/NP), number of patients with POCD/number of patients without POCD;SMD, standardized mean difference (SMD);95% CI, 95% confidence interval
Download CSV

Investigation of Heterogeneity

All inflammatory marker concentrations were compared in pictograms per milliliter or milligrams per milliliter; however, significant heterogeneity was found in all comparisons. For NSE, an I2 index of 76% indicated moderate heterogeneity, but we observed no significance relationship between elevated plasma NSE concentrations and POCD (p-value = 0.078 in 5 studies POCD/non-POCD=135/282). S-100β and IL-6 demonstrate significant relationships with POCD status, but the heterogeneity is high (for S-100β: χ2 = 117.35([d.f. = 6] ), p-value = 0.000, I2=94.9%, τ2=1.5417; for IL-6: χ2 = 43.78([d.f. = 4], p-value = 0.000, I2 = 90.9%, τ2= 1.190) (Table 1).

Sensitivity analysis

Sensitivity analyses were conducted to ascertain the primary origin of the heterogeneity. For all inflammatory markers, no single study qualitatively changed the SMD, no point estimate of its "omitted" analysis lies outside the confidence interval of the "combined" analysis. Its "omitted" meta-analytic estimate differs in significance relative to the "combined" analysis. Those suggested that the results of this meta-analysis was stable.


Meta-regression analyses were used to assess the associations between the preoperative and postoperative peripheral blood concentrations of the studied inflammatory markers. We conducted this analysis only if the preoperative and postoperative peripheral blood concentrations had at least five measurements from different studies and had either a significant SMD or significant heterogeneity in the absence of a significant SMD. Meta-regression analyses found a significant positive association between the SMD and the preoperative IL-6 peripheral blood concentration in patients with POCD (Coef. = 0.0587, p-value = 0.038, 5 studies, Figure 4 and Table 2). No associations were found between the SMD and the preoperative S-100β peripheral blood concentration (Coef.= 9.554, p-value = 0.290, 6 studies, Figure 5 and Table 2).

Figure 4. Association of preoperative IL-6 blood concentration in POCD patients with plasma IL-6 SMD using meta-regression.

Plot of the preoperative IL-6 peripheral blood concentration vs. the peripheral blood IL-6 SMD in the 5 included studies showing a fitted random effects meta-regression line.( 5 studies, Coef.=0 .0587 ,p-value = 0.038 ) Coef., coefficient.

CytokineNumber of studiestau2I2 resAdjusted R2Coef.Std. Err. t P>|t| [95% CI]
S100β# 6 2.106 94.27%9.22% 9.5547.839 1.22 0.290 -12.21,31.32
IL-6##5 0.2118 61.02%87.46% 0.05870 .0166 3.54 0.038 0.0059 0 .1114

Table 2. Meta-regression between the SMDs of postoperative and preoperative S-100β and IL-6 peripheral inflammatory concentrations.

# Preoperative S-100β peripheral blood concentration of all patients
## Preoperative IL-6 peripheral blood concentration of patients with POCD
tau2, estimate of between-study variance by residual maximum likelihood model;
I2 res, percentage of residual variation due to heterogeneity;
Adjusted R2, Proportion of between-study variance explained;
Coef., coefficient; Std. Err., standard error; 95% CI, 95% confidence interval.
Download CSV
Figure 5. Association of preoperative S-100β blood concentration in all subjects with plasma S-100β SMD using meta-regression.

Plot of the preoperative S-100β peripheral blood concentration vs. the peripheral blood S-100β SMD in the 6 included studies, showing a random effects meta-regression line. (6 studies, Coef.= 9.554,p-value = 0.290 ) Coef., coefficient.

Publication and Selective Reporting Biases

Significant risk of publication bias was not detected, as demonstrated by funnel plots, and no significant correlations were found between effect size and sample size among studies of peripheral blood S-100β, NSE and inflammatory cytokines, including IL-1β, IL-6 and TNF-α. However, tests for funnel plot asymmetry were only recommended for use when at least 10 studies were included in the meta-analysis. Therefore, Egger’s test was implemented to evaluate asymmetry and publication bias. The results showed no evidence of publication bias: simultaneous reporting of negative results suggested a lower risk of reporting bias; the majority of included studies examined multiple biomarkers, and the majority of studies reported at least one significant comparison also reported at least one non-significant comparison. (Table 3)

Egger’s test (bias)
CytokineCoef95% CItp
S-100β5.998 -6.389,18.391.240.268
IL-610.19 -1.384 ,21.762.800.068

Table 3. Egger’s test results for publication and selective reporting Bias.

S-100β, S-100β protein; IL-6, interleukin-6; NSE, neuron specific enolase; IL-1β, interleukin-1β; TNF-α, tumor necrosis factor-α;Coef., coefficient; 95% CI, 95% confidence interval.
Download CSV


A study by van Harten and colleagues showed that an immunological reaction was one of the most important causative factors of cardiac POCD [23]. Buvanendran also found that cardiac and non-cardiac POCD was associated with an inflammatory reaction [24]. In this meta-analysis, we aimed to clarify the association between the inflammatory makers and POCD. Our results show that IL-6 and S-100β are POCD- related pro-inflammatory markers. Although both positive and negative results have been reported in individual studies, this meta-analysis strengthens the clinical evidence that POCD is accompanied by a peripheral inflammatory reaction. Other pro-inflammatory cytokines, such as IL-10 and IL-8, were not involved in this analysis due to the limited number of studies available.

The concentrations of IL-1 and TNF-α did not differ significantly between the subjects with POCD and the control subjects. Following surgery, elderly patients often suffered from POCD, which could persist long after physical recovery. Surgery-induced tissue damage could activate the peripheral innate immune system, resulting the release of inflammatory mediators [25]. 

IL-1β was a pro-inflammatory cytokine that contributed to neuro-inflammation in many central nervous system (CNS) disorders [26] and the development of the inflammatory response within the brain, along with TNF-α and IL-6 [5,27]. In rats, IL-1β messenger RNA (mRNA) was significantly increased at postsurgery days 1 and 3, with an increase in protein expression detected on day 1. There was a significant increase in TNF-α mRNA on day 1 after surgery, although there was no increase in protein expression [28].

It was possible that our findings could not be interpreted as truly negative due to small sample sizes and significant heterogeneity between studies. Additional evidence demonstrated inflammatory activity with POCD, which could be the mechanism through which peripheral pro-inflammatory markers induced POCD. In mouse models, Barrientos and colleagues demonstrated that a single intracisternal administration of interleukin-1 receptor antagonist (IL-1RA) at the time of surgery was sufficient to block both the behavioral deficits and the neuro-inflammatory response. Injecting the same dose of IL-1RA peripherally failed to have a protective effect. These data provided strong support for the specific role of central, but not peripheral, IL-1β in POCD [29].

Following surgery under general anesthesia, TNF-α was the first cytokine to be released, and its concentration peaked (p-value < 0.05) 30 min after surgery. Terrando explained that TNF-α was important in cognitive decline. TNF-α was upstream of IL-1 and stimulates its production in the brain. Peripheral blockade of TNF-α could limit the release of IL-1 and prevent neuro-inflammation and cognitive decline in a mouse model of surgery-induced cognitive decline [30]. Belarbi found that the TNF-α synthesis inhibitor DT(3,6’dithiothalidomide) could significantly reverse hippocampus-dependent cognitive deficits induced by chronic neuro-inflammation. These results suggested that TNF-α was a critical mediator of chronic neuro-inflammation-induced neuronal dysfunction and cognitive impairment, and targeting its synthesis could be an effective therapeutic approach for several human neuron degenerative diseases [31].

Although POCD was a common complication of major surgery, its pathogenic mechanisms remained unknown. However, the contribution of several risk factors, including advanced age, duration of anesthesia, a second operation, low education level, postoperative infection, and respiratory complications, have been identified [4]. The apolipoproteinE4( ApoE4) allele was a known risk factor for Alzheimer’s disease [32], and one large study found a strong association between the ApoE4 allele and POCD in elderly patients who were given inhaled anesthetics [33]. However, McDonagh and colleagues were unable to found an association between cognitive decline and ApoE4 genotype, despite a robust sample size. Similarly, they also did not found association between perioperative B-type natriuretic peptide, CRP, D-dimer, matrix metalloproteinase-9, NSE, or S-100β levels and either POCD or the ApoE4 genotype [34].

A clinically relevant concentration of β-amyloid generation and aggregation was a key factor in the pathogenesis of Alzheimer’s disease [35]. These findings suggested that such an accumulation in elderly patients could potentially increase the risk of POCD. Wan et al. suggested that surgery could provoke astrogliosis, β-amyloid accumulation, and τ phosphorylation in elderly patients, which was likely to be associated with the cognitive decline [36].

Lipopolysaccharide (LPS) came from gram-negative bacteria and it was an endotoxin that could cause a strong immune response. When LPS bound the CD14/Toll-like receptor 4 /MD2 complex, it could stimulate the inflammatory cells to secrete a variety of cytokines. Some studies have examined the relationship between LPS and POCD [6, 37].

Observational studies could not clearly infer a causal relationship. Thus, the possibility that increased concentrations of pro-inflammatory cytokines were a consequence of POCD pathology should also be considered. Cibelli and colleagues explored whether systemic inflammation in response to surgical trauma triggers hippocampal inflammation and subsequent memory impairment in a mouse model of orthopedic surgery. Their studies suggested that inflammation play a critical role in the pathogenesis of POCD, as evidenced by the protection afforded to surgical animals by minocycline which was a nonspecific inhibitor of inflammation [5].

Peripheral pro-inflammatory signals can be actively propagated across the blood brain barrier (BBB) via crosstalk between peripheral and central immune cells [37]. Cytokines produced peripherally could cross the BBB via saturable transporters [38,39] or by passive diffusion through spaces between vascular endothelial cells [40]. Microglia was important in brain inflammation and inflammatory neurodegeneration as the resident innate immune cells of the CNS. Increasingly evidence illustrated that microglia played a crucial role in regulating both pathogenic and subsequent repair processes [41,42]. Guy C. Brown found that activation was accompanied by partial rounding up and mobility of the cells, proliferation, and the expression and release of pro-inflammatory cytokines, including TNF-α, IL-1 β, and IL-6 in the event of brain inflammation. These cytokines activated other microglia and astrocytes [43]. Microglia cells might secrete a variety of inflammatory mediators and neurotoxic factors [44]. Peripheral inflammatory markers might directly or indirectly cause a CNS inflammatory response. When excessive CNS inflammatory reactivity occurred, peripherally originating inflammatory markers could influence cognitive function by interfering with neuronal activity, which affected the function of synaptic connections, or by causing neuron toxicity and neuron degeneration, which resulted in impaired cognitive function [2].

The S-100β protein has been used as a marker of neuronal damage in various diseases [45]. S-100β belongs to the family of calcium-binding proteins. It was expressed primarily by astrocytes and found both intra- and extra-cellularly in brain tissue [46]. S-100β was usually elevated in the blood and CSF following nervous system damage due to a functional disturbance of membrane integrity and/or increased permeability of the BBB. Therefore, the S-100β protein might be a potential marker for POCD.

Li and colleagues found that the serum levels of the pro-inflammatory markers IL-6 and S-100β increased after general anesthesia in total hip-replacement surgery, and these increases may be associated with the occurrence of POCD [2].In our meta-analysis, we provide statistical evidence to support this association. Studies have shown that enhanced S-100β upregulated the cyclooxygenase-2 expression by way of RAGE in microglia [47], and in those with Alzheimer disease cyclooxygenase-2 inhibitor impeded neuro-inflammation and induced amelioration of cognitive function [48]. Rui-Lin Li indicated that RAGE signal transduction pathways mediated S-100β-induced neuro-inflammation, which might play a critical role in driving the pathogenesis of cognitive decline after surgery. And it could also activate the nuclear factor κB (NF-kB) signal pathways and gave rise to upregulation of pro-inflammatory mediators, such as IL-1β [49]. It is difficult to distinguish cause and effect in the relationship between inflammation and cerebral tissue damage. In the rat brain, both experimental subarachnoid hemorrhage [50] and global forebrain ischemia [51] induced upregulation of inflammatory gene pathways, as represented by the transcription factor NF-κB. Moreover, focal cerebral ischemia caused not only local inflammation but also widespread cerebral inflammation [52].

POCD is a growing and largely underestimated problem without a defined etiology. Globally, experts were focused on discovering risk factors for the occurrence of POCD and developing precautions to reduce its incidence rate [5355], but results have thus far been inconsistent. Meta-regression analyses utilizing the subgroup of studies reporting preoperative mean IL-6 levels suggested that the postoperative level of some inflammatory markers might have contribution to the occurrence of POCD. In Cibelli model, IL-6 plasma levels and transcription in the hippocampus appeared to increase after surgery. IL-6 had facilitating effects on IL-1β in mediating inflammation and causing hippocampal-dependent memory impairment [5]. Hudetz indicated that elevated postoperative IL-6 and CRP concentrations were associated with the subsequent development of short-term and medium-term impairment of cognitive functions after coronary artery surgery [56]. The relationship between inflammation and cognitive deficit was studied in a healthy young population after infusion of IL-6 to simulate the physiological acute-phase response, and the results indicated an explicit decrease in self-reported concentration and cognitive abilities [57]. The meta-regression analyses also expressed the same tendency. Assessment of hippocampal expression of the pro-inflammatory cytokines IL-6 confirmed that aged mice showed compared to adults embraced higher basal expression of these inflammatory genes in agreement with prior reports that showed aging primes microglial cells [58].

As the global aging progresses, the search for new preventative measures and treatment strategies to maintain higher brain functions throughout life is of major economic and medical importance [59]. In recent years, numerous studies have suggested that multiple strategies, including aerobic exercise, art therapy, and caloric restriction could enhance cognitive fitness [6062]. It would improved perioperative brain functional recovery that searching for biomarkers of brain functional reserve and the methods which could enhance it. IL-6 might be a biomarker that could guide the prevention and treatment of POCD. There was growing interest in inflammation-targeted therapeutics.

The present meta-analysis examining the role of peripheral inflammatory marks in POCD has several methodological limits. The literature search was limited to unstimulated cytokines (IL-1β, IL-6, TNF-α), NSE, and S-100β; however, other inflammatory markers (e.g. IL-10, high mobility group box 1(HMGB-1), NF-κB) have also been informative [63,64]. This meta-analysis was also limited by the substantial inconsistency in most comparisons, necessitating the use of random effects models that produced wider confidence intervals for most inflammatory markers(e.g. S-100β, IL-6, IL-1β, and TNF-α) .Most studies reported large standard deviations, suggesting substantial unexplained inter-individual variation in inflammatory marker concentrations. Some potential sources of heterogeneity, including assay methodology, age, gender, and medical comorbidity, were investigated.

The present study was also limited by the use of a categorical diagnosis of POCD; thus, POCD severity effects could not be systematically investigated. Meta-regression analyses in subgroups of studies reporting preoperative mean IL-6 and S-100β protein levels suggested that the preoperative level of some inflammatory markers might have contributed to the heterogeneity.

Furthermore, in cross-sectional studies, continuous and stochastic elevations could not be distinguished, and differences in the secretion patterns and half-lives of different cytokines might have a considerable impact on their measured concentrations. Most inflammatory markers immediately reached the peak postoperative value [30], but the peak serum concentration was not always the most valuable measure in the prediction of outcome [10]. One study found that the early and pronounced release of inflammatory markers was clinically unimportant but that the increase after 4-48h was significantly higher in patients with POCD than in patients without POCD [9].

Differences in enzyme linked immunosorbent assay (ELISA) techniques were investigated as a potential source of heterogeneity; however, many other assay parameters and methodological differences in the collection and handling of the samples that could not be systematically addressed might have contributed to heterogeneity.

Funnel plots and Egger's tests did not suggest the presence of publication bias. However, many observational studies were not registered with clinical trials data-bases, so the scope of the unpublished literature could not be ascertained and effects of bias at the study and outcome levels could not be ruled out. In addition, this article was mainly based on studies published in the English and Chinese language, and bias might have exist in the literature published in other languages.

In conclusion, our meta-analysis provides evidence that POCD is indeed correlated with the concentrations of IL-6 and S-100β. Furthermore, the meta-regression shows that IL-6 might serve an indicator to guide the prevention and treatment of POCD. Because POCD is a common postoperative complication and is related multiple factors, further investigation of the inflammatory response regulatory pathways, and the specific contributions of IL-6 and gene polymorphisms is necessary and meaningful.

Supporting Information

Checklist S1.

Reporting items for the meta-analyses.


Table S1.

Characteristics of the association studies that were included in the systematic review and that examined the S-100β and peripheral inflammatory markers.


Table S2.

Newcastle-Ottawa quality assessment scale for evaluation to the quality of each study.



This work was supported by JianBin Tong, Yi Wang and Xuyuan Kuang.

Author Contributions

Conceived and designed the experiments: LP LX WOY. Performed the experiments: LP LX. Analyzed the data: LP LX. Contributed reagents/materials/analysis tools: LX. Wrote the manuscript: LP LX. Final approval of the submitted version: WOY.


  1. 1. Silbert B, Evered L, Scott DA (2011) Cognitive decline in the elderly: is anaesthesia implicated? Best Pract Res Clin Anaesthesiol 25: 379-393. doi: PubMed: 21925403.
  2. 2. Li YC, Xi CH, An YF, Dong WH, Zhou M (2012) Perioperative inflammatory response and protein S-100beta concentrations - relationship with post-operative cognitive dysfunction in elderly patients. Acta Anaesthesiol Scand 56: 595-600. doi: PubMed: 22224444.
  3. 3. Steinmetz J, Christensen KB, Lund T, Lohse N, Rasmussen LS (2009) Long-term consequences of postoperative cognitive dysfunction. Anesthesiology 110: 548-555. doi: PubMed: 19225398.
  4. 4. Moller JT, Cluitmans P, Rasmussen LS, Houx P, Rasmussen H et al. (1998) Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. ISPOCD investigators. International Study of Post-Operative Cognitive Dysfunction. Lancet 351: 857-861. doi: PubMed: 9525362.
  5. 5. Cibelli M, Fidalgo AR, Terrando N, Ma D, Monaco C et al. (2010) Role of interleukin-1beta in postoperative cognitive dysfunction. Ann Neurol 68: 360-368. doi: PubMed: 20818791.
  6. 6. Fidalgo AR, Cibelli M, White JP, Nagy I, Maze M et al. (2011) Systemic inflammation enhances surgery-induced cognitive dysfunction in mice. Neurosci Lett 498: 63-66. doi: PubMed: 21575676.
  7. 7. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC et al. (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLOS Med 6: e1000100. PubMed: 196315071962251219621070.
  8. 8. van Munster BC, Korse CM, de Rooij SE, Bonfrer JM, Zwinderman AH et al. (2009) Markers of cerebral damage during delirium in elderly patients with hip fracture. BMC Neurol 9: 21. doi: PubMed: 19473521.
  9. 9. Linstedt U, Meyer O, Kropp P, Berkau A, Tapp E et al. (2002) Serum concentration of S-100 protein in assessment of cognitive dysfunction after general anesthesia in different types of surgery. Acta Anaesthesiol Scand 46: 384-389. doi: PubMed: 11952437.
  10. 10. Rasmussen LS, Christiansen M, Rasmussen H, Kristensen PA, Moller JT (2000) Do blood concentrations of neurone specific enolase and S-100 beta protein reflect cognitive dysfunction after abdominal surgery?ISPOCD Group. Br J Anaesth 84: 242-244. doi: PubMed: 10743460.
  11. 11. Beloosesky Y, Hendel D, Weiss A, Hershkovitz A, Grinblat J et al. (2007) Cytokines and C-reactive protein production in hip-fracture-operated elderly patients. J Gerontol A Biol Sci Med Sci 62: 420-426. doi: PubMed: 17452737.
  12. 12. Hozo SP, Djulbegovic B, Hozo I (2005) Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol 5: 13. doi: PubMed: 15840177.
  13. 13. Higgins JP, White IR, Anzures-Cabrera J (2008) Meta-analysis of skewed data: combining results reported on log-transformed or raw scales. Stat Med 27: 6072-6092. doi: PubMed: 18800342.
  14. 14. Wells GA, Shea B, O’Connell D, Peterson J, Welch V et al. (2011) The Newcastle-Ottawa scale (NOS) for assessing the quality of nonrandomized studies in meta-analysis. Available: Accessed 20 August 2013.
  15. 15. Tian HZ (2012) Abdominal postoperative cognitive impairment in patients with perioperative inflammatory cytokine expression analysis. China Pract Med7: 134-135.
  16. 16. Yang ZY, Xu YH, Fei FY (2010) Postoperative cognitive dysfunction and expressions of IL-1β,IL-6 and TNF-α in the elderly. J Clinical Anesthesiol 26: 764-766.
  17. 17. Wu WC, Dong M, Cao YJ (2012) Relationship of NSE and S-100β with POCD in Elderly Patients Undergoing Elective Limb Surgery. Clinical Med Eng 19: 1699-1702.
  18. 18. Peng Y, Xiong H, Dong LP, Li C (2010) Relationship between Postoperative Cognitive Dysfunction and Time Course of Serum Neuron-specific Enolase and S-100β Protein in Elderly Patients. J Nanchang Univ (Medical Science) 50: 21-23.
  19. 19. Xu YH, Liang Q, Yang ZY (2010) Elderly patients with early postoperative cognitive dysfunction and neuron-specific enolase expression. Chinese Journal of Gerontology 30: 3158-3159.
  20. 20. Chen MH (2009) The anesthesia and S100beta protein elderly patients after abdominal surgery cognitive dysfunction relationship, Modern. Med Health 25: 867-868.
  21. 21. Hong T,Wen DX, Hang YN (2006) Relationship of postoperative serum S100ββ protein concentration and cognitive dysfunction in elderly patients underwent elective abdominal surgery. J Clinical Anesthesiol22: 571-573.
  22. 22. Shi LY, Xu L, Wan YJ (2012) Relationship between postoperative cognitive dysfunction and the expression of inflammatory cytokines of plasma. Shanghai Medical J35: 115-117.
  23. 23. van Harten AE, Scheeren TW, Absalom AR (2012) A review of postoperative cognitive dysfunction and neuroinflammation associated with cardiac surgery and anaesthesia. Anaesthesia 67: 280-293. doi: PubMed: 22321085.
  24. 24. Buvanendran A, Kroin JS, Berger RA, Hallab NJ, Saha C et al. (2006) Upregulation of prostaglandin E2 and interleukins in the central nervous system and peripheral tissue during and after surgery in humans. Anesthesiology 104: 403-410. doi: PubMed: 16508385.
  25. 25. Rosczyk HA, Sparkman NL, Johnson RW (2008) Neuroinflammation and cognitive function in aged mice following minor surgery. Exp Gerontol 43: 840-846. doi: PubMed: 18602982.
  26. 26. Wu MD, Montgomery SL, Rivera-Escalera F, Olschowka JA, O'Banion MK (2013) Sustained IL-1beta expression impairs adult hippocampal neurogenesis independent of IL-1 signaling in nestin neural precursor cells. Brain Behav Immun.
  27. 27. Barrientos RM, Frank MG, Hein AM, Higgins EA, Watkins LR et al. (2009) Time course of hippocampal IL-1 beta and memory consolidation impairments in aging rats following peripheral infection. Brain Behav Immun 23: 46-54. doi: PubMed: 18664380.
  28. 28. Wan Y, Xu J, Ma D, Zeng Y, Cibelli M et al. (2007) Postoperative impairment of cognitive function in rats: a possible role for cytokine-mediated inflammation in the hippocampus. Anesthesiology 106: 436-443. doi: PubMed: 17325501.
  29. 29. Barrientos RM, Hein AM, Frank MG, Watkins LR, Maier SF (2012) Intracisternal interleukin-1 receptor antagonist prevents postoperative cognitive decline and neuroinflammatory response in aged rats. J Neurosci 32: 14641-14648. doi: PubMed: 23077050.
  30. 30. Terrando N, Monaco C, Ma D, Foxwell BM, Feldmann M et al. (2010) Tumor necrosis factor-alpha triggers a cytokine cascade yielding postoperative cognitive decline. Proc Natl Acad Sci U S A 107: 20518-20522. doi: PubMed: 21041647.
  31. 31. Belarbi K, Jopson T, Tweedie D, Arellano C, Luo W et al. (2012) TNF-alpha protein synthesis inhibitor restores neuronal function and reverses cognitive deficits induced by chronic neuroinflammation. J Neuroinflammation 9: 23. doi: PubMed: 22277195.
  32. 32. Horsburgh K, McCarron MO, White F, Nicoll JA (2000) The role of apolipoprotein E in Alzheimer's disease, acute brain injury and cerebrovascular disease: evidence of common mechanisms and utility of animal models. Neurobiol Aging 21: 245-255. doi: PubMed: 10867209.
  33. 33. Cai Y, Hu H, Liu P, Feng G, Dong W et al. (2012) Association between the apolipoprotein E4 and postoperative cognitive dysfunction in elderly patients undergoing intravenous anesthesia and inhalation anesthesia. Anesthesiology 116: 84-93. doi: PubMed: 22108393.
  34. 34. McDonagh DL, Mathew JP, White WD, Phillips-Bute B, Laskowitz DT et al. (2010) Cognitive function after major noncardiac surgery, apolipoprotein E4 genotype, and biomarkers of brain injury. Anesthesiology 112: 852-859. doi: PubMed: 20216394.
  35. 35. Tanzi RE, Bertram L (2005) Twenty years of the Alzheimer's disease amyloid hypothesis: a genetic perspective. Cell 120: 545-555. doi: PubMed: 15734686.
  36. 36. Wan Y, Xu J, Meng F, Bao Y, Ge Y et al. (2010) Cognitive decline following major surgery is associated with gliosis, beta-amyloid accumulation, and tau phosphorylation in old mice. Crit Care Med 38: 2190-2198. doi: PubMed: 20711073.
  37. 37. Maier SF (2003) Bi-directional immune-brain communication: Implications for understanding stress, pain, and cognition. Brain Behav Immun 17: 69-85. doi: PubMed: 12676570.
  38. 38. Banks WA, Farr SA, Morley JE (2002) Entry of blood-borne cytokines into the central nervous system: effects on cognitive processes. Neuroimmunomodulation 10: 319-327. doi: PubMed: 12907838.
  39. 39. Skinner RA, Gibson RM, Rothwell NJ, Pinteaux E, Penny JI (2009) Transport of interleukin-1 across cerebromicrovascular endothelial cells. Br J Pharmacol 156: 1115-1123. doi: PubMed: 19298391.
  40. 40. Rivest S, Lacroix S, Vallières L, Nadeau S, Zhang J et al. (2000) How the blood talks to the brain parenchyma and the paraventricular nucleus of the hypothalamus during systemic inflammatory and infectious stimuli. Proc Soc Exp Biol Med 223: 22-38. doi: PubMed: 10632958.
  41. 41. Block ML, Zecca L, Hong JS (2007) Microglia-mediated neurotoxicity: uncovering the molecular mechanisms. Nat Rev Neurosci 8: 57-69. doi: PubMed: 17180163.
  42. 42. Ransohoff RM, Perry VH (2009) Microglial physiology: unique stimuli, specialized responses. Annu Rev Immunol 27: 119-145. doi: PubMed: 19302036.
  43. 43. Brown GC, Neher JJ (2010) Inflammatory neurodegeneration and mechanisms of microglial killing of neurons. Mol Neurobiol 41: 242-247. doi: PubMed: 20195798.
  44. 44. Choi DK, Koppula S, Suk K (2011) Inhibitors of microglial neurotoxicity: focus on natural products. Molecules 16: 1021-1043. doi: PubMed: 21350391.
  45. 45. Gonçalves CA, Leite MC, Nardin P (2008) Biological and methodological features of the measurement of S100B, a putative marker of brain injury. Clin Biochem 41: 755-763. doi: PubMed: 18454941.
  46. 46. Basile AM, Fusi C, Conti AA, Paniccia R, Trefoloni G et al. (2001) S-100 protein and neuron-specific enolase as markers of subclinical cerebral damage after cardiac surgery: preliminary observation of a 6-month follow-up study. Eur Neurol 45: 151-159. doi: PubMed: 11306858.
  47. 47. Bianchi R, Adami C, Giambanco I, Donato R (2007) S100B binding to RAGE in microglia stimulates COX-2 expression. J Leukoc Biol 81: 108-118. PubMed: 17023559.
  48. 48. Peng M, Wang YL, Wang FF, Chen C, Wang CY (2012) The cyclooxygenase-2 inhibitor parecoxib inhibits surgery-induced proinflammatory cytokine expression in the hippocampus in aged rats. J Surg Res 178: e1-e8. doi: PubMed: 22959208.
  49. 49. Li RL, Zhang ZZ, Peng M, Wu Y, Zhang JJ et al. (2013) Postoperative impairment of cognitive function in old mice: a possible role for neuroinflammation mediated by HMGB1, S100B, and RAGE. J Surg Res. PubMed: 23899512.
  50. 50. Ma CX, Yin WN, Cai BW, Wu J, Wang JY et al. (2009) Toll-like receptor 4/nuclear factor-kappa B signaling detected in brain after early subarachnoid hemorrhage. Chin Med J (Engl) 122: 1575-1581. PubMed: 19719951.
  51. 51. Clemens JA, Stephenson DT, Dixon EP, Smalstig EB, Mincy RE et al. (1997) Global cerebral ischemia activates nuclear factor-kappa B prior to evidence of DNA fragmentation. Brain Res. Mol Brain Res 48: 187-196. doi: PubMed: 9332715.
  52. 52. Dhawan J, Benveniste H, Nawrocky M, Smith SD, Biegon A (2010) Transient focal ischemia results in persistent and widespread neuroinflammation and loss of glutamate NMDA receptors. NeuroImage 51: 599-605. doi: PubMed: 20206701.
  53. 53. Tilvis RS, Kähönen-Väre MH, Jolkkonen J, Valvanne J, Pitkala KH et al. (2004) Predictors of cognitive decline and mortality of aged people over a 10-year period. J Gerontol A Biol Sci Med Sci 59: 268-274. doi: PubMed: 15031312.
  54. 54. Monk TG, Weldon BC, Garvan CW, Dede DE, van der Aa MT et al. (2008) Predictors of cognitive dysfunction after major noncardiac surgery. Anesthesiology 108: 18-30. doi: PubMed: 18156878.
  55. 55. Slater JP, Guarino T, Stack J, Vinod K, Bustami RT et al. (2009) Cerebral oxygen desaturation predicts cognitive decline and longer hospital stay after cardiac surgery. Ann Thorac Surg 87: 35-44; discussion: 19101265.
  56. 56. Hudetz JA, Gandhi SD, Iqbal Z, Patterson KM, Pagel PS (2011) Elevated postoperative inflammatory biomarkers are associated with short- and medium-term cognitive dysfunction after coronary artery surgery. J Anesth 25: 1-9. doi: PubMed: 21061037.
  57. 57. Späth-Schwalbe E, Hansen K, Schmidt F, Schrezenmeier H, Marshall L et al. (1998) Acute effects of recombinant human interleukin-6 on endocrine and central nervous sleep functions in healthy men. J Clin Endocrinol Metab 83: 1573-1579. doi: PubMed: 9589658.
  58. 58. Ye SM, Johnson RW (1999) Increased interleukin-6 expression by microglia from brain of aged mice. J Neuroimmunol 93: 139-148. doi: PubMed: 10378877.
  59. 59. Pressley JC, Trott C, Tang M, Durkin M, Stern Y (2003) Dementia in community-dwelling elderly patients: A comparison of survey data, medicare claims, cognitive screening, reported symptoms, and activity limitations. J Clin Epidemiol 56: 896-905. doi: PubMed: 14505776.
  60. 60. Baker LD, Frank LL, Foster-Schubert K, Green PS, Wilkinson CW et al. (2010) Effects of aerobic exercise on mild cognitive impairment: a controlled trial. Arch Neurol 67: 71-79. doi: PubMed: 20065132.
  61. 61. Hattori H, Hattori C, Hokao C, Mizushima K, Mase T (2011) Controlled study on the cognitive and psychological effect of coloring and drawing in mild Alzheimer's disease patients. Geriatr Gerontol Int 11: 431-437. doi: PubMed: 21518170.
  62. 62. Witte AV, Fobker M, Gellner R, Knecht S, Flöel A (2009) Caloric restriction improves memory in elderly humans. Proc Natl Acad Sci U S A 106: 1255-1260. doi: PubMed: 19171901.
  63. 63. He HJ, Wang Y, Le Y, Duan KM, Yan XB et al. (2012) Surgery upregulates high mobility group box-1 and disrupts the blood-brain barrier causing cognitive dysfunction in aged rats. CNS Neurosci Ther 18: 994-1002. doi: PubMed: 23078219.
  64. 64. Terrando N, Eriksson LI, Ryu JK, Yang T, Monaco C et al. (2011) Resolving postoperative neuroinflammation and cognitive decline. Ann Neurol 70: 986-995. doi: PubMed: 22190370.