Journal of Critical Care xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis☆ Waka Takahashi, Taka-aki Nakada, MD, PhD ⁎, Ryuzo Abe, Kumiko Tanaka, Yosuke Matsumura, Shigeto Oda Chiba University Graduate School of Medicine, Department of Emergency and Critical Care Medicine, Chiba, Japan
a r t i c l e
i n f o
a b s t r a c t
Keywords: Interleukin 6 Cerebrospinal fluid Bacterial meningitis Diagnosis Neurologic outcome
Purpose: Interleukin 6 (IL-6) is a proinflammatory cytokine produced during infections. We hypothesized that IL-6 levels in the cerebrospinal fluid (CSF) would be elevated in bacterial meningitis and useful for diagnosing and predicting neurologic outcomes. Materials and methods: For the differentiation of bacterial meningitis, serum and CSF samples were obtained from patients with an altered level of consciousness. Patients were classified into 3 groups: bacterial meningitis, nonbacterial central nervous system disease, and other site sepsis. Results: Of the 70 patients included in this study, there were 13 in the bacterial meningitis group, 21 in the nonbacterial central nervous system disease group, and 36 in the other site sepsis group. The CSF IL-6 level was significantly higher in the bacterial meningitis group than in the other 2 groups (P b .0001). Of the 5 CSF parameters assessed, CSF IL-6 level exhibited the largest area under the receiver operating characteristic curve (0.962), with a cut-off value of 644 pg/mL (sensitivity, 92.3%; specificity, 89.5%). To examine a potential association between a high CSF level and neurologic outcome, CSF IL-6 levels were divided into 4 quartiles, and each level was compared with the frequency of a good neurologic outcome. The frequency of a good neurologic outcome was significantly lower in the highest CSF IL-6 quartile than in the other 3 quartiles (odds ratio, 0.18; 95% confidence interval, 0.05-0.69; P = .013). Conclusions: Measurement of the CSF IL-6 level is useful for diagnosing bacterial meningitis. © 2014 Elsevier Inc. All rights reserved.
1. Introduction
and severity of sepsis, and it serves as a useful marker for predicting clinical outcomes in such conditions [6]. Furthermore, bacterial infections are reported to be associated with significantly higher blood IL-6 levels than viral pneumonia or urinary tract infections [7]. We thus hypothesized that CSF IL-6 levels would be elevated in bacterial meningitis and served as a useful marker for diagnosis. To test this hypothesis, CSF was sampled by lumbar puncture in critically ill patients needing a differential diagnosis of bacterial meningitis; the sample was assayed for IL-6 to evaluate the sensitivity and specificity of CSF IL-6 level as a diagnostic marker for bacterial meningitis. Furthermore, the association between CSF IL-6 level and neurologic outcomes was examined, as excessive inflammation in the central nervous system (CNS) can cause and seriously impair its function [8].
Although the incidence of bacterial meningitis has declined over the last 10 years due to widespread vaccination, the mortality rate and neurologic outcomes of bacterial meningitis have not yet improved [1]. To improve these, an early and accurate diagnosis is needed [1,2]. Although cerebrospinal fluid (CSF) examination is essential to diagnose bacterial meningitis, it is often difficult to interpret test results, including white blood cell (WBC) count, glucose levels, and CSF protein levels [2]. As these conventional markers are not sufficiently accurate for a diagnosis of meningitis, novel markers have been tested. These novel markers are useful for not only the diagnosis but also the prediction of neurologic outcomes have been tested [2-4]. Interleukin 6 (IL-6) is a proinflammatory cytokine produced during the acute phase response to stimuli such as trauma, surgical insult, or infection [5]. It has been reported that blood IL-6 level is elevated in a variety of diseases related to the inflammatory response [5]. The blood IL-6 level is also positively correlated with the degree of inflammation
☆ Conflict of interest: Waka Takahashi, Taka-aki Nakada, Ryuzo Abe, Kumiko Tanaka, Yosuke Matsumura, and Shigeto Oda declare that they have no conflict of interest. ⁎ Corresponding author. Chiba University Graduate School of Medicine, Department of Emergency and Critical Care Medicine, 1-8-1 Inohana, Chuo, Chiba 260-8677, Japan. E-mail address: [email protected] (T. Nakada).
2. Materials and methods 2.1. Patients The present study retrospectively analyzed 70 critically ill patients who were admitted to the intensive care unit of Chiba University Hospital (Chiba, Japan) between April 2008, and August 2012 and were subjected to CSF analysis for the diagnosis or differentiation of bacterial meningitis; the CSF analysis included an IL-6 assay. Cerebrospinal fluid samples were collected immediately upon suspicion of bacterial
0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2014.02.020
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
2
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
those mentioned above was established by critical care physicians and neurologists based on neurologic findings, hematology, CSF analysis, bacterial culture test, diagnostic imaging including computed tomography and magnetic resonance imaging, brain physiology including electroencephalography, and clinical course. Patients were classified into 3 groups based on the final diagnosis: bacterial meningitis, nonbacterial CNS disease, or other site sepsis [11-13]. The Institutional Research Ethics Board at Chiba University approved this retrospective observational study.
Table 1 Type of CNS diseases Type
n
Bacterial meningitis Nonbacterial CNS diseases Viral meningitis Autoimmune encephalitis Trauma Epilepsy Acute encephalitis Other site sepsis
13 21 6 2 6 5 2 36
2.2. Measurements meningitis owing to clinical symptoms, such as fever and/or acute disturbance of consciousness, neck stiffness, and/or seizures. As previously reported [9,10], diagnosis of bacterial meningitis was based on compatible clinical features and one of the following: (a) positive CSF culture, (b) negative CSF culture but either a positive CSF antigen test or identification of bacteria on Gram staining of the CSF or a positive blood culture, or (c) CSF white cell count of 500/mm 3 or more and rapid improvement after antibacterial therapy despite negative CSF and blood culture results. Diagnosis of viral meningitis or encephalitis was based on positive polymerase chain reaction results or antiviral antibodies in the CSF and blood. The final diagnosis of various CNS diseases other than
After CSF samples were collected by lumbar puncture, neutrophil counts (cells per cubic millimeter), monocyte counts (cells per cubic millimeter), protein level (milligrams per deciliter), glucose level (milligrams per deciliter), and IL-6 level (picograms per milliliter) in the CSF were immediately determined. We also determined WBC counts (cells per cubic millimeter), C-reactive protein (CRP) levels (milligrams per deciliter), and IL-6 levels (picograms per milliliter) in the blood samples collected at the same time or within the same day of collection of CSF samples. Interleukin 6 levels in the CSF and blood samples were measured using a chemiluminescent enzyme immunoassay system (Lumipulse f, Human IL-6 CLEIA; Fujirebio, Tokyo, Japan,
Table 2 Baseline characteristics of patients
Age (y) Male, n (%) APACHE score SOFA score Body temperature Mean arterial pressure Heart rate Glasgow Coma Scale score SIRS, n (%) Sepsis, n (%) Severe sepsis, n (%) Septic shock, n (%) 28-day mortality
Bacterial meningitis (n = 13)
Nonbacterial CNS diseases (n = 21)
Other site sepsis (n = 36)
P
60 (44-72) 8 (61.5) 24 (21-29) 7 (5-11) 38.8 (38.0-40.0) 75 (64-98) 120 (108-140) 9 (8-10) 12 (92.3) 3 (23.1) 8 (61.5) 1 (7.7) 0 (0)
32 (18-63) 11(52.3) 18 (17-21) 8 (4-9) 38.7 (38.0-39.7) 71 (65-86) 132 (118-150) 8 (6-9) 16 (76.2) 7 (33.3) 8 (38.1) 0 (0) 0 (0)
67 17 23 8 38.5 63 119 6 34 11 8 13 2
.0086 .67 .06 .85 .87 .0642 .38 .23 .0366 .30 .0241 .0021 .37
(49-77) (47.2) (19-31) (6-10) (37.7-39.9) (54-83) (105-140) (4-11) (94.4) (30.6) (22.2) (36.1) (5.6)
APACHE indicates Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; SIRS, systemic inflammatory response syndrome. Data are median (interquartile range) for continuous variables. P values were calculated using χ2 or Kruskal-Wallis tests.
Table 3 Blood and CSF data Bacterial meningitis (n = 13) Blood WBCs, cells/mm3 CRP, mg/dL Glucose, mg/dL IL-6, pg/mL CSF Pressure, cm H2O Monocytes, cells/mm3 Neutrophils, cells/mm3 Protein, mg/dL Glucose, mg/dL IL-6, pg/mL CSF and blood CSF/blood glucose ratio CSF/blood IL-6 ratio CSF/blood IL-6 log-fold change
17.1 11.9 140 186
(11.6-24.7) (5.5-16.2) (112-195) (44-415)
19 (6-27) 44 (3-286) 1222 (67-2045) 142 (58-408) 54 (16-77) 6680 (2136-55708) 0.44 (0.097-0.63) 138.26 (5.25-376.31) 1.85 (0.72-2.57)
Nonbacterial CNS diseases (n = 21) 10.7 3.1 121 132
(7.7-14.8) (2.2-5.1) (104-179) (23-238)
15 10 14 45 73 185
(14-20) (5-28) (6-150) (28-83) (59-116) (28-579)
1.34 (0.25-4.35) 1.05 (0.25-21.75) 0.02 (−0.61-1.34)
Other site sepsis (n = 36) 13.4 6.2 132 389
(7.1-18.1) (1.9-12.7) (114-175) (81-1697)
16 (9-25) 3 (0-6) 3 (1-9) 31 (22-51) 80 (71-97) 29 (7-92) 0.19 (0.05-0.74) 0.05 (0.019-0.120) −1.29 (−1.73-0.92)
P .06 .06 .74 .0126 .88 .0009 b.0001 b.0001 .0051 b.0001 .0304 b.0001 b.0001
CSF/blood IL-6 log-fold change was calculated using formula of log10 CSF IL-6 minus log10 blood IL-6. Data are median (interquartile range) for continuous variables. P values were calculated using χ2 or Kruskal-Wallis tests.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
detection range, 0.2-1000 pg/mL; coefficient of variation, intramediate precision, 2.14%-2.66%; and intermediate precision, 4.37%-7.04%) [6]. At the time of CSF sample collection, we recorded body temperature, arterial pressure, heart rate, Glasgow Coma Scale score, Acute Physiology and Chronic Health Evaluation II score, and Sequential Organ Failure Assessment score. Diagnosis of systemic inflammatory response syndrome, sepsis, severe sepsis, or septic shock was determined at the same time, according to the American College of Chest Physicians/Society of Critical Care Medicine consensus criteria established in 1992 [14]. The Glasgow Outcome Scale (GOS) was used to evaluate the neurologic outcome, with a GOS score of 5 at the time of discharge defined as a “good neurologic outcome.”
3
monocyte count, neutrophil count, protein level, glucose level, and IL-6 level), the CSF IL-6 level exhibited the largest AUC of 0.962 (Fig. 1A), with a cut-off value of 644 pg/mL (sensitivity, 92.3%; specificity, 89.5%) (Table 4A). The AUC compared with nonbacterial CNS disease (Fig. 1B) was 0.912 for IL-6, with a cut-off value of 1339 pg/mL (sensitivity, 84.6%; specificity, 85.7%) (Table 4B). In addition, the AUC for IL-6 combined with all other CSF tests was significantly higher than the AUC for all other CSF tests without including IL-6, with an AUC of 0.901 (95% confidence
2.3. Statistical analysis The primary outcome variable was the diagnosis of bacterial meningitis. For the primary analysis, we used a receiver operating characteristic (ROC) curve analysis of CSF markers specific to the diagnosis of bacterial meningitis relative to other bacterial diseases and nonbacterial CNS diseases. Area under the ROC curve (AUC), cutoff value given by the maximum of the Youden index (sensitivity + specificity − 1), sensitivity, and specificity were calculated. The secondary outcome was a good neurologic outcome, defined as a GOS score of 5 at the time of hospital discharge. Patients were divided into quartiles according to the CSF IL-6 level (quartile 1, low; quartile 4, high), and the number of patients with favorable outcomes were compared among these quartiles. For the secondary analysis, we used logistic regression analysis to test for differences in good neurologic outcomes relative to the CSF IL-6 level with correction for potential confounding factors including age and sex. We analyzed the association between CSF IL-6 and GOS in each disease population. Because IL-6 values were skewed, they were subjected to log10 transformation before statistical analysis. We tested for differences in baseline characteristics using a χ 2 test for categorical data and a Kruskal-Wallis test for continuous data. Differences were considered significant using 2-tailed P b .05. Analyses were performed using SPSS version 20 (SPSS, Chicago, IL) statistical software. 3. Results Of the 3706 patients admitted to our intensive care unit during the study period, 70 patients (13 in the bacterial meningitis group, 21 in the nonbacterial CNS disease group, and 36 in the other site sepsis group) were included in the present study (Table 1). Of the 70 patients, 17 were administered corticosteroids such as dexamethasone or hydrocortisone for treatment, but none had been administered corticosteroids before obtaining CSF samples by lumbar puncture. Comparison of baseline characteristics among the 3 groups revealed that patients in the nonbacterial CNS disease group were younger, severe sepsis was more frequent in the bacterial meningitis group, and septic shock was most frequent in the other site sepsis group (Table 2). The pathogens identified in the bacterial meningitis group were as follows: methicillinsensitive Staphylococcus aureus (n = 2), Streptococcus pneumoniae (n = 2), methicillin-resistant S aureus (n = 1), methicillin-resistant Staphylococcus epidermidis (n = 1), group B streptococcus (n = 1), group G streptococcus (n = 1), Enterobacter aerogenes (n = 1), Streptococcus dysgalactiae (n = 1), and unknown (n = 3). Hematologic tests revealed that WBC counts, CRP levels, and IL-6 levels were lower in the nonbacterial CNS disease group than in the other groups (Table 3). Bacterial meningitis patients had higher CSF monocytes, CSF neutrophils, CSF protein levels, CSF IL-6 levels, ratio of CSF to blood IL-6 levels, and lower CSF glucose levels than the other 2 groups (Table 3). Receiver operating characteristic curve analysis was performed to assess the usefulness of these CSF parameters in diagnosing bacterial meningitis. Among the 5 CSF parameters assessed (CSF
Fig. 1. Receiver operating characteristic curves of the CSF test for predicting bacterial meningitis. We performed ROC curve analysis to determine whether CSF parameters could predict bacterial meningitis using the data of (A) all patients and (B) patients with bacterial meningitis and nonbacterial CNS diseases. A, The AUCs were 0.962 for the CSF IL-6 level, 0.917 for the CSF neutrophil count, 0.871 for the CSF protein level, 0.779 for the CSF glucose level, and 0.725 for the CSF monocyte count. B, The AUCs were 0.912 for the CSF IL-6 level, 0.874 for the CSF neutrophil count, 0.826 for the CSF protein level, 0.742 for the CSF glucose level, and 0.625 for the CSF monocyte count.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
4
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
Table 4 Area under receiver operating characteristic curves for predicting bacterial meningitis. AUC (95% CI) A. AUC for predicting bacterial meningitis CSF IL-6 levels 0.962 (0.922-1.003) CSF neutrophil counts 0.917 (0.811-1.023) CSF protein levels 0.871 (0.779-0.964) CSF glucose levels 0.779 (0.606-0.952) CSF monocyte count 0.725 (0.542-0.908)
P
Cut-off value
Sensitivity, specificity
1.0 × 10−4 1.0 × 10−4 1.0 × 10−4 .0018 .0120
644 pg/mL 20 cells/mm3 50.5 mg/dL 69.5 mg/dL 17.0 cells/mm3
0.923, 0.846, 0.846, 0.692, 0.692,
0.895 0.825 0.684 0.839 0.825
1339 pg/mL 497 cells/mm3 46.5 mg/dL 55.0 mg/dL 23.5 cells/mm3
0.846, 0.769, 1.000, 0.615, 0.615,
0.857 1.000 0.571 0.850 0.714
B. AUC for predicting bacterial meningitis as compared with nonbacterial CNS diseases CSF IL-6 levels 0.912 (0.818-1.000) 1.0 × 101.0 × 10−4 CSF neutrophil counts 0.874 (0.730-1.017) 1.0 × 10−4 CSF protein levels 0.826 (0.688-0.964) 1.0 × 10−3 CSF glucose levels 0.742 (0.557-0.926) .0203 CSF monocyte count 0.625 (0.416-0.834) .2283 Receiver operating characteristic curve for glucose was analyzed using inverse values. P values were calculated using a null hypothesis of having an area of 0.5. The Youden index was used to determine cut-off values.
interval [CI], 0.796-1.006) for IL-6 combined with all other CSF tests and 0.824 for all other CSF tests without IL-6 (95% CI, 0.647-1.001). We then examined the association between CSF IL-6 level and neurologic outcomes. Patients were divided into 4 quartiles of CSF IL-6 levels (quartile 1, lowest; quartile 4, highest). We compared the frequency of a good neurologic outcome, defined as a GOS score of 5 at hospital discharge, among the 4 groups (Fig. 2). The frequency of a good neurologic outcome was significantly lower in quartile 4, with the highest CSF IL-6 levels (median range, 2933 [969-12182] pg/mL [25th to 75th percentile]) compared with the remaining 3 quartiles (per quartile, odds ratio [OR], 0.64; 95% CI, 0.42-0.98; P = .041; quartile 4 vs other quartiles, OR, 0.18; 95% CI, 0.05-0.69; P = .013; Table 5). In addition, we repeated the logistic regression analysis including disease type as a covariate. A non-significant trend in the same direction was observed (quartile 4 vs other quartiles, adjusted OR, 0.24; 95% CI, 0.03-1.71; P = .15). Finally, we analyzed the associations with disease types. There was a significant correlation between log CSF IL-6 and GOS in the viral meningitis patients (P = .0027, R2, 0.92, log CSF IL-6 = −0.59 × GOS + 0.37) but not in the bacterial meningitis patients (P = .45), despite the smaller sample size of the subpopulations.
4. Discussion The present study investigated CSF IL-6 levels in critically ill patients with CNS disorders. The results demonstrate that the CSF IL-6 level in patients with bacterial meningitis is significantly higher than that in patients with other conditions. An ROC curve analysis was conducted to assess the usefulness of various CSF parameters in diagnosing bacterial meningitis; this analysis revealed that the CSF IL-6 level exhibited the largest AUC, suggesting that CSF IL-6 was a useful diagnostic marker. In addition, a high CSF IL-6 level was associated with a poor neurologic outcome for viral meningitis, suggesting that this parameter could serve as a predictive marker for neurologic outcome in patients with viral meningitis. Cerebrospinal fluid parameters previously reported to be useful in diagnosing bacterial meningitis include an increase in WBC count [15,16], increase in protein level [9,16], and decrease in glucose level [9,16]. In the present study, the monocyte count, neutrophil count, and protein levels in the CSF were significantly higher in the bacterial meningitis group than in the other 2 groups, whereas the glucose level was significantly lower. Higher CSF IL-6 levels in patients with bacterial meningitis than in those with aseptic meningitis have been previously reported in the emergency department [17], pediatric patients with meningitis [18,19], and patients with postoperative infection after subarachnoid hemorrhage [20]; however, no comparative assessment using ROC curve analysis was performed. In the present study, the clinical usefulness of the CSF IL-6 level in diagnosing bacterial meningitis was compared with other CSF parameters (ie, monocyte count, neutrophil count, protein level, and glucose level) by ROC curve analysis. The results demonstrate that CSF IL-6 level exhibited the largest AUC among all CSF parameters assessed, and ROC of IL-6 combined with all other CSF tests showed significant differences as compared with ROC of all other CSF tests without including IL-6, suggesting the clinical usefulness of CSF IL-6 as a diagnostic biomarker for bacterial meningitis.
Table 5 Logistic regression analysis of good neurologic outcomes (GOS score of 5)
Fig. 2. Frequencies of patients with good neurologic outcome by the quartile of CSF IL-6 levels. Patients with high CSF IL-6 levels had a significantly lower rate of good neurologic outcome (P = .032). Good neurologic outcome was defined as GOS score of 5 (“good recovery”). Median CSF IL-6 levels (25%-75% tile) were 4.5 (3.3-6.4) pg/mL (quartile 1), 26 (21-35) pg/mL (quartile 2), 166 (82-235) pg/mL (quartile 3), and 2933 (969-12182) pg/mL (quartile 4). Quartile 1 had the lowest CSF IL-6 levels, whereas quartile 4 had the highest CSF IL-6 levels. P values were calculated using the χ2 test for trend.
A. CSF IL-6 levels per quartile Age per year Male CSF IL-6 levels per quartile
OR (95% CI)
P
0.98 (0.97-1.01) 1.12 (0.44-2.91) 0.64 (0.42-0.98)
.23 .80 .041
B. CSF IL-6 levels in quartile 4 compared with that in other quartiles Age per year 0.98 (0.97-1.01) Male 1.19 (0.45-3.12) CSF IL-6 levels, quartile 4 0.18 (0.05-0.69)
.11 .73 .013
Quartile 4 has the highest CFS IL-6 levels.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
The cut-off level of IL-6 for distinguishing bacterial meningitis was 644 pg/mL. Vazquez et al [17] demonstrated that CSF IL-6 concentrations more than 90 pg/dL can be used for diagnosing bacterial meningitis with a sensitivity of 100% and a specificity of 95% in adults, and Hsieh et al [18] demonstrated that a concentration of 10 pg/mL or more can help in distinguishing bacterial meningitis from aseptic meningitis in children with a sensitivity of 96% and a specificity of 51%. In these 2 previous reports, CSF samples were collected at the time of clinical suspicion of meningitis, and IL-6 was detected through the enzyme-linked immunosorbent assay technique, as was done in the current study. The discrepancy in CSF IL-6 cut-off values may be due to using different measurement devices for each study. In addition to the CSF IL-6 level, many reports have suggested that the CSF lactate level might serve as another useful biomarker in diagnosing bacterial meningitis. In an ROC analysis evaluating the ability of CSF parameters to differentiate between bacterial and viral meningitis, Viallon et al [3] reported an AUC of 0.96 (diagnostic cut-off level, 3.8 mmol/L) for the CSF lactate level, a value comparable with that for the CSF IL-6 level obtained in the present study. In addition, a meta-analysis demonstrated that the CSF lactate level served as a better marker for bacterial meningitis than conventional markers such as CSF glucose, CSF protein, and CSF cell count [21]. However, limitations of CSF lactate level as a diagnostic marker have also been reported: it is not elevated in patients who received or underwent pretreatment antibiotics, and it is elevated in those with noninfectious conditions, including stroke and head trauma [4,21]. Candidate biomarkers of bacterial meningitis in the CSF other than lactate that exhibit a high AUC include heparin-binding protein (AUC, 0.996) [22] and soluble triggering receptor expressed on myeloid cells 1 (AUC, 0.82) [23,24]. These biomarkers are expected to be useful in the diagnosis of bacterial meningitis in the clinical setting. The immunologic profile, including the levels of a wide range of cytokines or chemokines in the CSF of patients with bacterial meningitis, has been described. Coutinho et al [25] measured the levels of several cytokines, including IL-6, obtained from patients with bacterial meningitis, and, with the exception of one cytokine, all values were more than 1000 pg/mL. Morichi et al [26] reported that brain-derived neurotrophic factor, which affects CNS function, was significantly elevated in those with bacterial meningitis and that the CSF brain-derived neurotrophic factor level was correlated with the CSF IL-6 level. Grandgirard et al [27] demonstrated that the causative pathogen determined the inflammatory profile in the CSF as well as the outcome in patients with bacterial meningitis. On the other hand, Misra et al [28] reported that cytokine levels did not correlate with the stage of meningitis, outcome, or radiologic deterioration or improvement in patients with tuberculous meningitis. Central nervous system disorders due to bacterial meningitis are driven by 2 different mechanisms: an overwhelming inflammatory reaction and the direct effect of the bacterial toxin itself [8]. Microglial cells and meningeal macrophages play an important role in neuronal injury caused by the inflammatory response. These cells are involved in the initial stage of the CNS immune response by producing various proinflammatory cytokines, including IL-6 as well as soluble factors that recruit WBCs to the subarachnoid space [29]. Proinflammatory cytokines induce up-regulation of several adhesion molecules on the vascular endothelium, thereby allowing an influx of neutrophils and lymphocytes from the bloodstream into the infection site. Leukocytes that have migrated into the CNS release a variety of factors, such as reactive oxygen species, that contribute to vasospasm and vasculitis, leading to cellular injury, cell death in neurons, and eventually longterm neurologic deficits [29]. In fact, association between the severity of the subarachnoid space inflammatory response and mortality has been reported in experimental animal models of bacterial meningitis [30-32]. Modulation of the subarachnoid space inflammatory response is proposed as a mechanism that improves neurologic outcome in patients with meningitis by adjunctive dexamethasone
5
therapy [8]. In IL-6 knockout mice, decreased cerebral edema, bloodbrain barrier disruption, and intracranial pressure have been reported [33], and dexamethasone therapy in patients with bacterial meningitis induced a significant decrease in CSF IL-6 level compared with a placebo in a randomized trial [34]. A significant correlation between the CSF IL-6 level and neurologic outcome in viral meningitis observed in the present study appears to be consistent with these previously reported findings. Wang et al [35] reported that, in children with echovirus 71 meningitis, the CSF IL-6 level was significantly higher for pulmonary edema and autonomic nervous system dysregulation than for isolated brainstem encephalitis. The present study has the following limitations. First, this was a retrospective, observational study. It is unknown whether the measurement of CSF IL-6 would improve the outcome of patients with bacterial meningitis. Second, the sample size was relatively small, and the causative bacteria were variable. In particular, we performed a logistic regression analysis of neurologic outcomes and CSF IL-6 levels including 3 disease types as covariates. However, as the neurologic outcome is associated with the disease type, testing for the association between CSF IL-6 level and neurologic outcomes in one category is optimal. Thus, our results of neurologic outcome and CSF IL-6 are limited due to the small sample size. In addition, the baseline characteristics of nonbacterial CNS diseases describe a younger and less severely ill group of patients, a potential confounder for the prognostic value of CSF IL-6. 5. Conclusions In conclusion, measurement of the CSF IL-6 level was useful in diagnosing bacterial meningitis. References [1] Thigpen MC, Whitney CG, Messonnier NE, Zell ER, Lynfield R, Hadler JL, et al. Bacterial meningitis in the United States, 1998-2007. N Engl J Med 2011;364:2016–25. [2] Brouwer MC, Thwaites GE, Tunkel AR, van de Beek D. Dilemmas in the diagnosis of acute community-acquired bacterial meningitis. Lancet 2012;380:1684–92. [3] Viallon A, Desseigne N, Marjollet O, Birynczyk A, Belin M, Guyomarch S, et al. Meningitis in adult patients with a negative direct cerebrospinal fluid examination: value of cytochemical markers for differential diagnosis. Crit Care 2011;15:R136. [4] Sakushima K, Hayashino Y, Kawaguchi T, Jackson JL, Fukuhara S. Diagnostic accuracy of cerebrospinal fluid lactate for differentiating bacterial meningitis from aseptic meningitis: a meta-analysis. J Infect 2011;62:255–62. [5] Rincon M. Interleukin-6: from an inflammatory marker to a target for inflammatory diseases. Trends Immunol 2012;33:571–7. [6] Oda S, Hirasawa H, Shiga H, Nakanishi K, Matsuda K, Nakamua M. Sequential measurement of IL-6 blood levels in patients with systemic inflammatory response syndrome (SIRS)/sepsis. Cytokine 2005;29:169–75. [7] Chalupa P, Beran O, Herwald H, Kasprikova N, Holub M. Evaluation of potential biomarkers for the discrimination of bacterial and viral infections. Infection 2011;39:411–7. [8] Mook-Kanamori BB, Geldhoff M, van der Poll T, van de Beek D. Pathogenesis and pathophysiology of pneumococcal meningitis. Clin Microbiol Rev 2011;24:557–91. [9] Durand ML, Calderwood SB, Weber DJ, Miller SI, Southwick FS, Caviness VS, et al. Acute bacterial meningitis in adults. A review of 493 episodes. N Engl J Med 1993;328:21–8. [10] Schwarz S, Bertram M, Schwab S, Andrassy K, Hacke W. Serum procalcitonin levels in bacterial and abacterial meningitis. Crit Care Med 2000;28:1828–32. [11] Hirohata S, Kanai Y, Mitsuo A, Tokano Y, Hashimoto H, Subcommittee NR. Accuracy of cerebrospinal fluid IL-6 testing for diagnosis of lupus psychosis. A multicenter retrospective study. Clin Rheumatol 2009;28:1319–23. [12] Helmy A, De Simoni MG, Guilfoyle MR, Carpenter KL, Hutchinson PJ. Cytokines and innate inflammation in the pathogenesis of human traumatic brain injury. Prog Neurobiol 2011;95:352–72. [13] Vezzani A, French J, Bartfai T, Baram TZ. The role of inflammation in epilepsy. Nat Rev Neurol 2011;7:31–40. [14] Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. Chest. American College of Chest Physicians/Society of Critical Care Medicine; 1992. p. 1644–55. [15] van de Beek D, de Gans J, Spanjaard L, Weisfelt M, Reitsma JB, Vermeulen M. Clinical features and prognostic factors in adults with bacterial meningitis. N Engl J Med 2004;351:1849–59. [16] Spanos A, Harrell Jr FE, Durack DT. Differential diagnosis of acute meningitis. An analysis of the predictive value of initial observations. JAMA 1989;262:2700–7.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
6
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
[17] Vazquez JA, Adducci Mdel C, Coll C, Monzon DG, Iserson KV. Acute meningitis prognosis using cerebrospinal fluid interleukin-6 levels. J Emerg Med 2012;43:322–7. [18] Hsieh CC, Lu JH, Chen SJ, Lan CC, Chow WC, Tang RB. Cerebrospinal fluid levels of interleukin-6 and interleukin-12 in children with meningitis. Childs Nerv Syst 2009;25:461–5. [19] Mukai AO, Krebs VL, Bertoli CJ, Okay TS. TNF-alpha and IL-6 in the diagnosis of bacterial and aseptic meningitis in children. Pediatr Neurol 2006;34:25–9. [20] Hopkins SJ, McMahon CJ, Singh N, Galea J, Hoadley M, Scarth S, et al. Cerebrospinal fluid and plasma cytokines after subarachnoid haemorrhage: CSF interleukin-6 may be an early marker of infection. J Neuroinflammation 2012;9:255. [21] Huy NT, Thao NT, Diep DT, Kikuchi M, Zamora J, Hirayama K. Cerebrospinal fluid lactate concentration to distinguish bacterial from aseptic meningitis: a systemic review and meta-analysis. Crit Care 2010;14:R240. [22] Linder A, Akesson P, Brink M, Studahl M, Bjorck L, Christensson B. Heparinbinding protein: a diagnostic marker of acute bacterial meningitis. Crit Care Med 2011;39:812–7. [23] Bishara J, Hadari N, Shalita-Chesner M, Samra Z, Ofir O, Paul M, et al. Soluble triggering receptor expressed on myeloid cells-1 for distinguishing bacterial from aseptic meningitis in adults. Eur J Clin Microbiol Infect Dis 2007;26:647–50. [24] Determann RM, Weisfelt M, de Gans J, van der Ende A, Schultz MJ, van de Beek D. Soluble triggering receptor expressed on myeloid cells 1: a biomarker for bacterial meningitis. Intensive Care Med 2006;32:1243–7. [25] Coutinho LG, Grandgirard D, Leib SL, Agnez-Lima LF. Cerebrospinal-fluid cytokine and chemokine profile in patients with pneumococcal and meningococcal meningitis. BMC Infect Dis 2013;13:326.
[26] Morichi S, Kashiwagi Y, Takekuma K, Hoshika A, Kawashima H. Expressions of brainderived neurotrophic factor (BDNF) in cerebrospinal fluid and plasma of children with meningitis and encephalitis/encephalopathy. Int J Neurosci 2013;123:17–23. [27] Grandgirard D, Gaumann R, Coulibaly B, Dangy JP, Sie A, Junghanss T, et al. The causative pathogen determines the inflammatory profile in cerebrospinal fluid and outcome in patients with bacterial meningitis. Mediators Inflamm 2013;2013:312476. [28] Misra UK, Kalita J, Srivastava R, Nair PP, Mishra MK, Basu A. A study of cytokines in tuberculous meningitis: clinical and MRI correlation. Neurosci Lett 2010;483:6–10. [29] Gerber J, Nau R. Mechanisms of injury in bacterial meningitis. Curr Opin Neurol 2010;23:312–8. [30] Tunkel AR, Wispelwey B, Scheld WM. Bacterial meningitis: recent advances in pathophysiology and treatment. Ann Intern Med 1990;112:610–23. [31] Quagliarello V, Scheld WM. Bacterial meningitis: pathogenesis, pathophysiology, and progress. N Engl J Med 1992;327:864–72. [32] Nudelman Y, Tunkel AR. Bacterial meningitis: epidemiology, pathogenesis and management update. Drugs 2009;69:2577–96. [33] Paul R, Koedel U, Winkler F, Kieseier BC, Fontana A, Kopf M, et al. Lack of IL-6 augments inflammatory response but decreases vascular permeability in bacterial meningitis. Brain 2003;126:1873–82. [34] Mai NT, Tuan TV, Wolbers M, Hoang DM, Nga TV, Chau TT, et al. Immunological and biochemical correlates of adjunctive dexamethasone in Vietnamese adults with bacterial meningitis. Clin Infect Dis 2009;49:1387–92. [35] Wang SM, Lei HY, Su LY, Wu JM, Yu CK, Wang JR, et al. Cerebrospinal fluid cytokines in enterovirus 71 brain stem encephalitis and echovirus meningitis infections of varying severity. Clin Microbiol Infect 2007;13:677–82.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis☆ Waka Takahashi, Taka-aki Nakada, MD, PhD ⁎, Ryuzo Abe, Kumiko Tanaka, Yosuke Matsumura, Shigeto Oda Chiba University Graduate School of Medicine, Department of Emergency and Critical Care Medicine, Chiba, Japan
a r t i c l e
i n f o
a b s t r a c t
Keywords: Interleukin 6 Cerebrospinal fluid Bacterial meningitis Diagnosis Neurologic outcome
Purpose: Interleukin 6 (IL-6) is a proinflammatory cytokine produced during infections. We hypothesized that IL-6 levels in the cerebrospinal fluid (CSF) would be elevated in bacterial meningitis and useful for diagnosing and predicting neurologic outcomes. Materials and methods: For the differentiation of bacterial meningitis, serum and CSF samples were obtained from patients with an altered level of consciousness. Patients were classified into 3 groups: bacterial meningitis, nonbacterial central nervous system disease, and other site sepsis. Results: Of the 70 patients included in this study, there were 13 in the bacterial meningitis group, 21 in the nonbacterial central nervous system disease group, and 36 in the other site sepsis group. The CSF IL-6 level was significantly higher in the bacterial meningitis group than in the other 2 groups (P b .0001). Of the 5 CSF parameters assessed, CSF IL-6 level exhibited the largest area under the receiver operating characteristic curve (0.962), with a cut-off value of 644 pg/mL (sensitivity, 92.3%; specificity, 89.5%). To examine a potential association between a high CSF level and neurologic outcome, CSF IL-6 levels were divided into 4 quartiles, and each level was compared with the frequency of a good neurologic outcome. The frequency of a good neurologic outcome was significantly lower in the highest CSF IL-6 quartile than in the other 3 quartiles (odds ratio, 0.18; 95% confidence interval, 0.05-0.69; P = .013). Conclusions: Measurement of the CSF IL-6 level is useful for diagnosing bacterial meningitis. © 2014 Elsevier Inc. All rights reserved.
1. Introduction
and severity of sepsis, and it serves as a useful marker for predicting clinical outcomes in such conditions [6]. Furthermore, bacterial infections are reported to be associated with significantly higher blood IL-6 levels than viral pneumonia or urinary tract infections [7]. We thus hypothesized that CSF IL-6 levels would be elevated in bacterial meningitis and served as a useful marker for diagnosis. To test this hypothesis, CSF was sampled by lumbar puncture in critically ill patients needing a differential diagnosis of bacterial meningitis; the sample was assayed for IL-6 to evaluate the sensitivity and specificity of CSF IL-6 level as a diagnostic marker for bacterial meningitis. Furthermore, the association between CSF IL-6 level and neurologic outcomes was examined, as excessive inflammation in the central nervous system (CNS) can cause and seriously impair its function [8].
Although the incidence of bacterial meningitis has declined over the last 10 years due to widespread vaccination, the mortality rate and neurologic outcomes of bacterial meningitis have not yet improved [1]. To improve these, an early and accurate diagnosis is needed [1,2]. Although cerebrospinal fluid (CSF) examination is essential to diagnose bacterial meningitis, it is often difficult to interpret test results, including white blood cell (WBC) count, glucose levels, and CSF protein levels [2]. As these conventional markers are not sufficiently accurate for a diagnosis of meningitis, novel markers have been tested. These novel markers are useful for not only the diagnosis but also the prediction of neurologic outcomes have been tested [2-4]. Interleukin 6 (IL-6) is a proinflammatory cytokine produced during the acute phase response to stimuli such as trauma, surgical insult, or infection [5]. It has been reported that blood IL-6 level is elevated in a variety of diseases related to the inflammatory response [5]. The blood IL-6 level is also positively correlated with the degree of inflammation
☆ Conflict of interest: Waka Takahashi, Taka-aki Nakada, Ryuzo Abe, Kumiko Tanaka, Yosuke Matsumura, and Shigeto Oda declare that they have no conflict of interest. ⁎ Corresponding author. Chiba University Graduate School of Medicine, Department of Emergency and Critical Care Medicine, 1-8-1 Inohana, Chuo, Chiba 260-8677, Japan. E-mail address: [email protected] (T. Nakada).
2. Materials and methods 2.1. Patients The present study retrospectively analyzed 70 critically ill patients who were admitted to the intensive care unit of Chiba University Hospital (Chiba, Japan) between April 2008, and August 2012 and were subjected to CSF analysis for the diagnosis or differentiation of bacterial meningitis; the CSF analysis included an IL-6 assay. Cerebrospinal fluid samples were collected immediately upon suspicion of bacterial
0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2014.02.020
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
2
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
those mentioned above was established by critical care physicians and neurologists based on neurologic findings, hematology, CSF analysis, bacterial culture test, diagnostic imaging including computed tomography and magnetic resonance imaging, brain physiology including electroencephalography, and clinical course. Patients were classified into 3 groups based on the final diagnosis: bacterial meningitis, nonbacterial CNS disease, or other site sepsis [11-13]. The Institutional Research Ethics Board at Chiba University approved this retrospective observational study.
Table 1 Type of CNS diseases Type
n
Bacterial meningitis Nonbacterial CNS diseases Viral meningitis Autoimmune encephalitis Trauma Epilepsy Acute encephalitis Other site sepsis
13 21 6 2 6 5 2 36
2.2. Measurements meningitis owing to clinical symptoms, such as fever and/or acute disturbance of consciousness, neck stiffness, and/or seizures. As previously reported [9,10], diagnosis of bacterial meningitis was based on compatible clinical features and one of the following: (a) positive CSF culture, (b) negative CSF culture but either a positive CSF antigen test or identification of bacteria on Gram staining of the CSF or a positive blood culture, or (c) CSF white cell count of 500/mm 3 or more and rapid improvement after antibacterial therapy despite negative CSF and blood culture results. Diagnosis of viral meningitis or encephalitis was based on positive polymerase chain reaction results or antiviral antibodies in the CSF and blood. The final diagnosis of various CNS diseases other than
After CSF samples were collected by lumbar puncture, neutrophil counts (cells per cubic millimeter), monocyte counts (cells per cubic millimeter), protein level (milligrams per deciliter), glucose level (milligrams per deciliter), and IL-6 level (picograms per milliliter) in the CSF were immediately determined. We also determined WBC counts (cells per cubic millimeter), C-reactive protein (CRP) levels (milligrams per deciliter), and IL-6 levels (picograms per milliliter) in the blood samples collected at the same time or within the same day of collection of CSF samples. Interleukin 6 levels in the CSF and blood samples were measured using a chemiluminescent enzyme immunoassay system (Lumipulse f, Human IL-6 CLEIA; Fujirebio, Tokyo, Japan,
Table 2 Baseline characteristics of patients
Age (y) Male, n (%) APACHE score SOFA score Body temperature Mean arterial pressure Heart rate Glasgow Coma Scale score SIRS, n (%) Sepsis, n (%) Severe sepsis, n (%) Septic shock, n (%) 28-day mortality
Bacterial meningitis (n = 13)
Nonbacterial CNS diseases (n = 21)
Other site sepsis (n = 36)
P
60 (44-72) 8 (61.5) 24 (21-29) 7 (5-11) 38.8 (38.0-40.0) 75 (64-98) 120 (108-140) 9 (8-10) 12 (92.3) 3 (23.1) 8 (61.5) 1 (7.7) 0 (0)
32 (18-63) 11(52.3) 18 (17-21) 8 (4-9) 38.7 (38.0-39.7) 71 (65-86) 132 (118-150) 8 (6-9) 16 (76.2) 7 (33.3) 8 (38.1) 0 (0) 0 (0)
67 17 23 8 38.5 63 119 6 34 11 8 13 2
.0086 .67 .06 .85 .87 .0642 .38 .23 .0366 .30 .0241 .0021 .37
(49-77) (47.2) (19-31) (6-10) (37.7-39.9) (54-83) (105-140) (4-11) (94.4) (30.6) (22.2) (36.1) (5.6)
APACHE indicates Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; SIRS, systemic inflammatory response syndrome. Data are median (interquartile range) for continuous variables. P values were calculated using χ2 or Kruskal-Wallis tests.
Table 3 Blood and CSF data Bacterial meningitis (n = 13) Blood WBCs, cells/mm3 CRP, mg/dL Glucose, mg/dL IL-6, pg/mL CSF Pressure, cm H2O Monocytes, cells/mm3 Neutrophils, cells/mm3 Protein, mg/dL Glucose, mg/dL IL-6, pg/mL CSF and blood CSF/blood glucose ratio CSF/blood IL-6 ratio CSF/blood IL-6 log-fold change
17.1 11.9 140 186
(11.6-24.7) (5.5-16.2) (112-195) (44-415)
19 (6-27) 44 (3-286) 1222 (67-2045) 142 (58-408) 54 (16-77) 6680 (2136-55708) 0.44 (0.097-0.63) 138.26 (5.25-376.31) 1.85 (0.72-2.57)
Nonbacterial CNS diseases (n = 21) 10.7 3.1 121 132
(7.7-14.8) (2.2-5.1) (104-179) (23-238)
15 10 14 45 73 185
(14-20) (5-28) (6-150) (28-83) (59-116) (28-579)
1.34 (0.25-4.35) 1.05 (0.25-21.75) 0.02 (−0.61-1.34)
Other site sepsis (n = 36) 13.4 6.2 132 389
(7.1-18.1) (1.9-12.7) (114-175) (81-1697)
16 (9-25) 3 (0-6) 3 (1-9) 31 (22-51) 80 (71-97) 29 (7-92) 0.19 (0.05-0.74) 0.05 (0.019-0.120) −1.29 (−1.73-0.92)
P .06 .06 .74 .0126 .88 .0009 b.0001 b.0001 .0051 b.0001 .0304 b.0001 b.0001
CSF/blood IL-6 log-fold change was calculated using formula of log10 CSF IL-6 minus log10 blood IL-6. Data are median (interquartile range) for continuous variables. P values were calculated using χ2 or Kruskal-Wallis tests.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
detection range, 0.2-1000 pg/mL; coefficient of variation, intramediate precision, 2.14%-2.66%; and intermediate precision, 4.37%-7.04%) [6]. At the time of CSF sample collection, we recorded body temperature, arterial pressure, heart rate, Glasgow Coma Scale score, Acute Physiology and Chronic Health Evaluation II score, and Sequential Organ Failure Assessment score. Diagnosis of systemic inflammatory response syndrome, sepsis, severe sepsis, or septic shock was determined at the same time, according to the American College of Chest Physicians/Society of Critical Care Medicine consensus criteria established in 1992 [14]. The Glasgow Outcome Scale (GOS) was used to evaluate the neurologic outcome, with a GOS score of 5 at the time of discharge defined as a “good neurologic outcome.”
3
monocyte count, neutrophil count, protein level, glucose level, and IL-6 level), the CSF IL-6 level exhibited the largest AUC of 0.962 (Fig. 1A), with a cut-off value of 644 pg/mL (sensitivity, 92.3%; specificity, 89.5%) (Table 4A). The AUC compared with nonbacterial CNS disease (Fig. 1B) was 0.912 for IL-6, with a cut-off value of 1339 pg/mL (sensitivity, 84.6%; specificity, 85.7%) (Table 4B). In addition, the AUC for IL-6 combined with all other CSF tests was significantly higher than the AUC for all other CSF tests without including IL-6, with an AUC of 0.901 (95% confidence
2.3. Statistical analysis The primary outcome variable was the diagnosis of bacterial meningitis. For the primary analysis, we used a receiver operating characteristic (ROC) curve analysis of CSF markers specific to the diagnosis of bacterial meningitis relative to other bacterial diseases and nonbacterial CNS diseases. Area under the ROC curve (AUC), cutoff value given by the maximum of the Youden index (sensitivity + specificity − 1), sensitivity, and specificity were calculated. The secondary outcome was a good neurologic outcome, defined as a GOS score of 5 at the time of hospital discharge. Patients were divided into quartiles according to the CSF IL-6 level (quartile 1, low; quartile 4, high), and the number of patients with favorable outcomes were compared among these quartiles. For the secondary analysis, we used logistic regression analysis to test for differences in good neurologic outcomes relative to the CSF IL-6 level with correction for potential confounding factors including age and sex. We analyzed the association between CSF IL-6 and GOS in each disease population. Because IL-6 values were skewed, they were subjected to log10 transformation before statistical analysis. We tested for differences in baseline characteristics using a χ 2 test for categorical data and a Kruskal-Wallis test for continuous data. Differences were considered significant using 2-tailed P b .05. Analyses were performed using SPSS version 20 (SPSS, Chicago, IL) statistical software. 3. Results Of the 3706 patients admitted to our intensive care unit during the study period, 70 patients (13 in the bacterial meningitis group, 21 in the nonbacterial CNS disease group, and 36 in the other site sepsis group) were included in the present study (Table 1). Of the 70 patients, 17 were administered corticosteroids such as dexamethasone or hydrocortisone for treatment, but none had been administered corticosteroids before obtaining CSF samples by lumbar puncture. Comparison of baseline characteristics among the 3 groups revealed that patients in the nonbacterial CNS disease group were younger, severe sepsis was more frequent in the bacterial meningitis group, and septic shock was most frequent in the other site sepsis group (Table 2). The pathogens identified in the bacterial meningitis group were as follows: methicillinsensitive Staphylococcus aureus (n = 2), Streptococcus pneumoniae (n = 2), methicillin-resistant S aureus (n = 1), methicillin-resistant Staphylococcus epidermidis (n = 1), group B streptococcus (n = 1), group G streptococcus (n = 1), Enterobacter aerogenes (n = 1), Streptococcus dysgalactiae (n = 1), and unknown (n = 3). Hematologic tests revealed that WBC counts, CRP levels, and IL-6 levels were lower in the nonbacterial CNS disease group than in the other groups (Table 3). Bacterial meningitis patients had higher CSF monocytes, CSF neutrophils, CSF protein levels, CSF IL-6 levels, ratio of CSF to blood IL-6 levels, and lower CSF glucose levels than the other 2 groups (Table 3). Receiver operating characteristic curve analysis was performed to assess the usefulness of these CSF parameters in diagnosing bacterial meningitis. Among the 5 CSF parameters assessed (CSF
Fig. 1. Receiver operating characteristic curves of the CSF test for predicting bacterial meningitis. We performed ROC curve analysis to determine whether CSF parameters could predict bacterial meningitis using the data of (A) all patients and (B) patients with bacterial meningitis and nonbacterial CNS diseases. A, The AUCs were 0.962 for the CSF IL-6 level, 0.917 for the CSF neutrophil count, 0.871 for the CSF protein level, 0.779 for the CSF glucose level, and 0.725 for the CSF monocyte count. B, The AUCs were 0.912 for the CSF IL-6 level, 0.874 for the CSF neutrophil count, 0.826 for the CSF protein level, 0.742 for the CSF glucose level, and 0.625 for the CSF monocyte count.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
4
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
Table 4 Area under receiver operating characteristic curves for predicting bacterial meningitis. AUC (95% CI) A. AUC for predicting bacterial meningitis CSF IL-6 levels 0.962 (0.922-1.003) CSF neutrophil counts 0.917 (0.811-1.023) CSF protein levels 0.871 (0.779-0.964) CSF glucose levels 0.779 (0.606-0.952) CSF monocyte count 0.725 (0.542-0.908)
P
Cut-off value
Sensitivity, specificity
1.0 × 10−4 1.0 × 10−4 1.0 × 10−4 .0018 .0120
644 pg/mL 20 cells/mm3 50.5 mg/dL 69.5 mg/dL 17.0 cells/mm3
0.923, 0.846, 0.846, 0.692, 0.692,
0.895 0.825 0.684 0.839 0.825
1339 pg/mL 497 cells/mm3 46.5 mg/dL 55.0 mg/dL 23.5 cells/mm3
0.846, 0.769, 1.000, 0.615, 0.615,
0.857 1.000 0.571 0.850 0.714
B. AUC for predicting bacterial meningitis as compared with nonbacterial CNS diseases CSF IL-6 levels 0.912 (0.818-1.000) 1.0 × 101.0 × 10−4 CSF neutrophil counts 0.874 (0.730-1.017) 1.0 × 10−4 CSF protein levels 0.826 (0.688-0.964) 1.0 × 10−3 CSF glucose levels 0.742 (0.557-0.926) .0203 CSF monocyte count 0.625 (0.416-0.834) .2283 Receiver operating characteristic curve for glucose was analyzed using inverse values. P values were calculated using a null hypothesis of having an area of 0.5. The Youden index was used to determine cut-off values.
interval [CI], 0.796-1.006) for IL-6 combined with all other CSF tests and 0.824 for all other CSF tests without IL-6 (95% CI, 0.647-1.001). We then examined the association between CSF IL-6 level and neurologic outcomes. Patients were divided into 4 quartiles of CSF IL-6 levels (quartile 1, lowest; quartile 4, highest). We compared the frequency of a good neurologic outcome, defined as a GOS score of 5 at hospital discharge, among the 4 groups (Fig. 2). The frequency of a good neurologic outcome was significantly lower in quartile 4, with the highest CSF IL-6 levels (median range, 2933 [969-12182] pg/mL [25th to 75th percentile]) compared with the remaining 3 quartiles (per quartile, odds ratio [OR], 0.64; 95% CI, 0.42-0.98; P = .041; quartile 4 vs other quartiles, OR, 0.18; 95% CI, 0.05-0.69; P = .013; Table 5). In addition, we repeated the logistic regression analysis including disease type as a covariate. A non-significant trend in the same direction was observed (quartile 4 vs other quartiles, adjusted OR, 0.24; 95% CI, 0.03-1.71; P = .15). Finally, we analyzed the associations with disease types. There was a significant correlation between log CSF IL-6 and GOS in the viral meningitis patients (P = .0027, R2, 0.92, log CSF IL-6 = −0.59 × GOS + 0.37) but not in the bacterial meningitis patients (P = .45), despite the smaller sample size of the subpopulations.
4. Discussion The present study investigated CSF IL-6 levels in critically ill patients with CNS disorders. The results demonstrate that the CSF IL-6 level in patients with bacterial meningitis is significantly higher than that in patients with other conditions. An ROC curve analysis was conducted to assess the usefulness of various CSF parameters in diagnosing bacterial meningitis; this analysis revealed that the CSF IL-6 level exhibited the largest AUC, suggesting that CSF IL-6 was a useful diagnostic marker. In addition, a high CSF IL-6 level was associated with a poor neurologic outcome for viral meningitis, suggesting that this parameter could serve as a predictive marker for neurologic outcome in patients with viral meningitis. Cerebrospinal fluid parameters previously reported to be useful in diagnosing bacterial meningitis include an increase in WBC count [15,16], increase in protein level [9,16], and decrease in glucose level [9,16]. In the present study, the monocyte count, neutrophil count, and protein levels in the CSF were significantly higher in the bacterial meningitis group than in the other 2 groups, whereas the glucose level was significantly lower. Higher CSF IL-6 levels in patients with bacterial meningitis than in those with aseptic meningitis have been previously reported in the emergency department [17], pediatric patients with meningitis [18,19], and patients with postoperative infection after subarachnoid hemorrhage [20]; however, no comparative assessment using ROC curve analysis was performed. In the present study, the clinical usefulness of the CSF IL-6 level in diagnosing bacterial meningitis was compared with other CSF parameters (ie, monocyte count, neutrophil count, protein level, and glucose level) by ROC curve analysis. The results demonstrate that CSF IL-6 level exhibited the largest AUC among all CSF parameters assessed, and ROC of IL-6 combined with all other CSF tests showed significant differences as compared with ROC of all other CSF tests without including IL-6, suggesting the clinical usefulness of CSF IL-6 as a diagnostic biomarker for bacterial meningitis.
Table 5 Logistic regression analysis of good neurologic outcomes (GOS score of 5)
Fig. 2. Frequencies of patients with good neurologic outcome by the quartile of CSF IL-6 levels. Patients with high CSF IL-6 levels had a significantly lower rate of good neurologic outcome (P = .032). Good neurologic outcome was defined as GOS score of 5 (“good recovery”). Median CSF IL-6 levels (25%-75% tile) were 4.5 (3.3-6.4) pg/mL (quartile 1), 26 (21-35) pg/mL (quartile 2), 166 (82-235) pg/mL (quartile 3), and 2933 (969-12182) pg/mL (quartile 4). Quartile 1 had the lowest CSF IL-6 levels, whereas quartile 4 had the highest CSF IL-6 levels. P values were calculated using the χ2 test for trend.
A. CSF IL-6 levels per quartile Age per year Male CSF IL-6 levels per quartile
OR (95% CI)
P
0.98 (0.97-1.01) 1.12 (0.44-2.91) 0.64 (0.42-0.98)
.23 .80 .041
B. CSF IL-6 levels in quartile 4 compared with that in other quartiles Age per year 0.98 (0.97-1.01) Male 1.19 (0.45-3.12) CSF IL-6 levels, quartile 4 0.18 (0.05-0.69)
.11 .73 .013
Quartile 4 has the highest CFS IL-6 levels.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
The cut-off level of IL-6 for distinguishing bacterial meningitis was 644 pg/mL. Vazquez et al [17] demonstrated that CSF IL-6 concentrations more than 90 pg/dL can be used for diagnosing bacterial meningitis with a sensitivity of 100% and a specificity of 95% in adults, and Hsieh et al [18] demonstrated that a concentration of 10 pg/mL or more can help in distinguishing bacterial meningitis from aseptic meningitis in children with a sensitivity of 96% and a specificity of 51%. In these 2 previous reports, CSF samples were collected at the time of clinical suspicion of meningitis, and IL-6 was detected through the enzyme-linked immunosorbent assay technique, as was done in the current study. The discrepancy in CSF IL-6 cut-off values may be due to using different measurement devices for each study. In addition to the CSF IL-6 level, many reports have suggested that the CSF lactate level might serve as another useful biomarker in diagnosing bacterial meningitis. In an ROC analysis evaluating the ability of CSF parameters to differentiate between bacterial and viral meningitis, Viallon et al [3] reported an AUC of 0.96 (diagnostic cut-off level, 3.8 mmol/L) for the CSF lactate level, a value comparable with that for the CSF IL-6 level obtained in the present study. In addition, a meta-analysis demonstrated that the CSF lactate level served as a better marker for bacterial meningitis than conventional markers such as CSF glucose, CSF protein, and CSF cell count [21]. However, limitations of CSF lactate level as a diagnostic marker have also been reported: it is not elevated in patients who received or underwent pretreatment antibiotics, and it is elevated in those with noninfectious conditions, including stroke and head trauma [4,21]. Candidate biomarkers of bacterial meningitis in the CSF other than lactate that exhibit a high AUC include heparin-binding protein (AUC, 0.996) [22] and soluble triggering receptor expressed on myeloid cells 1 (AUC, 0.82) [23,24]. These biomarkers are expected to be useful in the diagnosis of bacterial meningitis in the clinical setting. The immunologic profile, including the levels of a wide range of cytokines or chemokines in the CSF of patients with bacterial meningitis, has been described. Coutinho et al [25] measured the levels of several cytokines, including IL-6, obtained from patients with bacterial meningitis, and, with the exception of one cytokine, all values were more than 1000 pg/mL. Morichi et al [26] reported that brain-derived neurotrophic factor, which affects CNS function, was significantly elevated in those with bacterial meningitis and that the CSF brain-derived neurotrophic factor level was correlated with the CSF IL-6 level. Grandgirard et al [27] demonstrated that the causative pathogen determined the inflammatory profile in the CSF as well as the outcome in patients with bacterial meningitis. On the other hand, Misra et al [28] reported that cytokine levels did not correlate with the stage of meningitis, outcome, or radiologic deterioration or improvement in patients with tuberculous meningitis. Central nervous system disorders due to bacterial meningitis are driven by 2 different mechanisms: an overwhelming inflammatory reaction and the direct effect of the bacterial toxin itself [8]. Microglial cells and meningeal macrophages play an important role in neuronal injury caused by the inflammatory response. These cells are involved in the initial stage of the CNS immune response by producing various proinflammatory cytokines, including IL-6 as well as soluble factors that recruit WBCs to the subarachnoid space [29]. Proinflammatory cytokines induce up-regulation of several adhesion molecules on the vascular endothelium, thereby allowing an influx of neutrophils and lymphocytes from the bloodstream into the infection site. Leukocytes that have migrated into the CNS release a variety of factors, such as reactive oxygen species, that contribute to vasospasm and vasculitis, leading to cellular injury, cell death in neurons, and eventually longterm neurologic deficits [29]. In fact, association between the severity of the subarachnoid space inflammatory response and mortality has been reported in experimental animal models of bacterial meningitis [30-32]. Modulation of the subarachnoid space inflammatory response is proposed as a mechanism that improves neurologic outcome in patients with meningitis by adjunctive dexamethasone
5
therapy [8]. In IL-6 knockout mice, decreased cerebral edema, bloodbrain barrier disruption, and intracranial pressure have been reported [33], and dexamethasone therapy in patients with bacterial meningitis induced a significant decrease in CSF IL-6 level compared with a placebo in a randomized trial [34]. A significant correlation between the CSF IL-6 level and neurologic outcome in viral meningitis observed in the present study appears to be consistent with these previously reported findings. Wang et al [35] reported that, in children with echovirus 71 meningitis, the CSF IL-6 level was significantly higher for pulmonary edema and autonomic nervous system dysregulation than for isolated brainstem encephalitis. The present study has the following limitations. First, this was a retrospective, observational study. It is unknown whether the measurement of CSF IL-6 would improve the outcome of patients with bacterial meningitis. Second, the sample size was relatively small, and the causative bacteria were variable. In particular, we performed a logistic regression analysis of neurologic outcomes and CSF IL-6 levels including 3 disease types as covariates. However, as the neurologic outcome is associated with the disease type, testing for the association between CSF IL-6 level and neurologic outcomes in one category is optimal. Thus, our results of neurologic outcome and CSF IL-6 are limited due to the small sample size. In addition, the baseline characteristics of nonbacterial CNS diseases describe a younger and less severely ill group of patients, a potential confounder for the prognostic value of CSF IL-6. 5. Conclusions In conclusion, measurement of the CSF IL-6 level was useful in diagnosing bacterial meningitis. References [1] Thigpen MC, Whitney CG, Messonnier NE, Zell ER, Lynfield R, Hadler JL, et al. Bacterial meningitis in the United States, 1998-2007. N Engl J Med 2011;364:2016–25. [2] Brouwer MC, Thwaites GE, Tunkel AR, van de Beek D. Dilemmas in the diagnosis of acute community-acquired bacterial meningitis. Lancet 2012;380:1684–92. [3] Viallon A, Desseigne N, Marjollet O, Birynczyk A, Belin M, Guyomarch S, et al. Meningitis in adult patients with a negative direct cerebrospinal fluid examination: value of cytochemical markers for differential diagnosis. Crit Care 2011;15:R136. [4] Sakushima K, Hayashino Y, Kawaguchi T, Jackson JL, Fukuhara S. Diagnostic accuracy of cerebrospinal fluid lactate for differentiating bacterial meningitis from aseptic meningitis: a meta-analysis. J Infect 2011;62:255–62. [5] Rincon M. Interleukin-6: from an inflammatory marker to a target for inflammatory diseases. Trends Immunol 2012;33:571–7. [6] Oda S, Hirasawa H, Shiga H, Nakanishi K, Matsuda K, Nakamua M. Sequential measurement of IL-6 blood levels in patients with systemic inflammatory response syndrome (SIRS)/sepsis. Cytokine 2005;29:169–75. [7] Chalupa P, Beran O, Herwald H, Kasprikova N, Holub M. Evaluation of potential biomarkers for the discrimination of bacterial and viral infections. Infection 2011;39:411–7. [8] Mook-Kanamori BB, Geldhoff M, van der Poll T, van de Beek D. Pathogenesis and pathophysiology of pneumococcal meningitis. Clin Microbiol Rev 2011;24:557–91. [9] Durand ML, Calderwood SB, Weber DJ, Miller SI, Southwick FS, Caviness VS, et al. Acute bacterial meningitis in adults. A review of 493 episodes. N Engl J Med 1993;328:21–8. [10] Schwarz S, Bertram M, Schwab S, Andrassy K, Hacke W. Serum procalcitonin levels in bacterial and abacterial meningitis. Crit Care Med 2000;28:1828–32. [11] Hirohata S, Kanai Y, Mitsuo A, Tokano Y, Hashimoto H, Subcommittee NR. Accuracy of cerebrospinal fluid IL-6 testing for diagnosis of lupus psychosis. A multicenter retrospective study. Clin Rheumatol 2009;28:1319–23. [12] Helmy A, De Simoni MG, Guilfoyle MR, Carpenter KL, Hutchinson PJ. Cytokines and innate inflammation in the pathogenesis of human traumatic brain injury. Prog Neurobiol 2011;95:352–72. [13] Vezzani A, French J, Bartfai T, Baram TZ. The role of inflammation in epilepsy. Nat Rev Neurol 2011;7:31–40. [14] Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. Chest. American College of Chest Physicians/Society of Critical Care Medicine; 1992. p. 1644–55. [15] van de Beek D, de Gans J, Spanjaard L, Weisfelt M, Reitsma JB, Vermeulen M. Clinical features and prognostic factors in adults with bacterial meningitis. N Engl J Med 2004;351:1849–59. [16] Spanos A, Harrell Jr FE, Durack DT. Differential diagnosis of acute meningitis. An analysis of the predictive value of initial observations. JAMA 1989;262:2700–7.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
6
W. Takahashi et al. / Journal of Critical Care xxx (2014) xxx–xxx
[17] Vazquez JA, Adducci Mdel C, Coll C, Monzon DG, Iserson KV. Acute meningitis prognosis using cerebrospinal fluid interleukin-6 levels. J Emerg Med 2012;43:322–7. [18] Hsieh CC, Lu JH, Chen SJ, Lan CC, Chow WC, Tang RB. Cerebrospinal fluid levels of interleukin-6 and interleukin-12 in children with meningitis. Childs Nerv Syst 2009;25:461–5. [19] Mukai AO, Krebs VL, Bertoli CJ, Okay TS. TNF-alpha and IL-6 in the diagnosis of bacterial and aseptic meningitis in children. Pediatr Neurol 2006;34:25–9. [20] Hopkins SJ, McMahon CJ, Singh N, Galea J, Hoadley M, Scarth S, et al. Cerebrospinal fluid and plasma cytokines after subarachnoid haemorrhage: CSF interleukin-6 may be an early marker of infection. J Neuroinflammation 2012;9:255. [21] Huy NT, Thao NT, Diep DT, Kikuchi M, Zamora J, Hirayama K. Cerebrospinal fluid lactate concentration to distinguish bacterial from aseptic meningitis: a systemic review and meta-analysis. Crit Care 2010;14:R240. [22] Linder A, Akesson P, Brink M, Studahl M, Bjorck L, Christensson B. Heparinbinding protein: a diagnostic marker of acute bacterial meningitis. Crit Care Med 2011;39:812–7. [23] Bishara J, Hadari N, Shalita-Chesner M, Samra Z, Ofir O, Paul M, et al. Soluble triggering receptor expressed on myeloid cells-1 for distinguishing bacterial from aseptic meningitis in adults. Eur J Clin Microbiol Infect Dis 2007;26:647–50. [24] Determann RM, Weisfelt M, de Gans J, van der Ende A, Schultz MJ, van de Beek D. Soluble triggering receptor expressed on myeloid cells 1: a biomarker for bacterial meningitis. Intensive Care Med 2006;32:1243–7. [25] Coutinho LG, Grandgirard D, Leib SL, Agnez-Lima LF. Cerebrospinal-fluid cytokine and chemokine profile in patients with pneumococcal and meningococcal meningitis. BMC Infect Dis 2013;13:326.
[26] Morichi S, Kashiwagi Y, Takekuma K, Hoshika A, Kawashima H. Expressions of brainderived neurotrophic factor (BDNF) in cerebrospinal fluid and plasma of children with meningitis and encephalitis/encephalopathy. Int J Neurosci 2013;123:17–23. [27] Grandgirard D, Gaumann R, Coulibaly B, Dangy JP, Sie A, Junghanss T, et al. The causative pathogen determines the inflammatory profile in cerebrospinal fluid and outcome in patients with bacterial meningitis. Mediators Inflamm 2013;2013:312476. [28] Misra UK, Kalita J, Srivastava R, Nair PP, Mishra MK, Basu A. A study of cytokines in tuberculous meningitis: clinical and MRI correlation. Neurosci Lett 2010;483:6–10. [29] Gerber J, Nau R. Mechanisms of injury in bacterial meningitis. Curr Opin Neurol 2010;23:312–8. [30] Tunkel AR, Wispelwey B, Scheld WM. Bacterial meningitis: recent advances in pathophysiology and treatment. Ann Intern Med 1990;112:610–23. [31] Quagliarello V, Scheld WM. Bacterial meningitis: pathogenesis, pathophysiology, and progress. N Engl J Med 1992;327:864–72. [32] Nudelman Y, Tunkel AR. Bacterial meningitis: epidemiology, pathogenesis and management update. Drugs 2009;69:2577–96. [33] Paul R, Koedel U, Winkler F, Kieseier BC, Fontana A, Kopf M, et al. Lack of IL-6 augments inflammatory response but decreases vascular permeability in bacterial meningitis. Brain 2003;126:1873–82. [34] Mai NT, Tuan TV, Wolbers M, Hoang DM, Nga TV, Chau TT, et al. Immunological and biochemical correlates of adjunctive dexamethasone in Vietnamese adults with bacterial meningitis. Clin Infect Dis 2009;49:1387–92. [35] Wang SM, Lei HY, Su LY, Wu JM, Yu CK, Wang JR, et al. Cerebrospinal fluid cytokines in enterovirus 71 brain stem encephalitis and echovirus meningitis infections of varying severity. Clin Microbiol Infect 2007;13:677–82.
Please cite this article as: Takahashi W, et al, Usefulness of interleukin 6 levels in the cerebrospinal fluid for the diagnosis of bacterial meningitis, J Crit Care (2014), http://dx.doi.org/10.1016/j.jcrc.2014.02.020
