Cm

LNa

b

c

ARRAA

KBCNP

1

t1rrrtttsv

F

T

0h

Clinical Neurology and Neurosurgery 115 (2013) 1820– 1825

Contents lists available at SciVerse ScienceDirect

Clinical Neurology and Neurosurgery

j o ur nal hom epage: www.elsev ier .com/ locate /c l ineuro

erebrospinal fluid lactate in post-neurosurgical bacterialeningitis diagnosis�

uis Patricio Maskina,∗, Federico Capparelli a, Andrea Morab, Alejandro Hlavnickaa,ora Orellanac, Maria Fernanda Díaza, Néstor Wainszteina, Marcelo Del Castillob

Intensive Care Unit, Raúl Carrea Neurological Research Institute, FLENI, Buenos Aires, ArgentinaInfectious Disease, Raúl Carrea Neurological Research Institute, FLENI, Buenos Aires, ArgentinaMicrobiology Department, Raúl Carrea Neurological Research Institute, FLENI, Buenos Aires, Argentina

a r t i c l e i n f o

rticle history:eceived 6 November 2012eceived in revised form 5 May 2013ccepted 17 May 2013vailable online 27 June 2013

eywords:acterial meningitisSF lactateeurosurgeryostoperative infection

a b s t r a c t

Background: Differential diagnosis between post-neurosurgical bacterial meningitis (PNBM) and asepticmeningitis is difficult. Inflammatory and biochemical cerebrospinal fluid (CSF) changes mimic thoseclassically observed after CNS surgery. CSF lactate assay has therefore been proposed as a useful PNBMmarker.Objective: To evaluate the diagnostic accuracy of CSF lactate as a PNBM marker in patients hospitalizedafter a neurosurgical procedure.Methods: Between July 2005 and June 2009, a prospective clinical study, in which all patients with clin-ical suspicion of PNBM were enrolled, was conducted at our neurosurgical Intensive Care Unit. PNBMdiagnosis was categorized as proven, probable or negative before the analysis.Results: Seventy-nine patients, 51 males with a mean age of 50 years (range 32–68 years) were included.Surgery was elective in 76% patients, mostly for brain tumors (57%); thirty PNBM episodes were identi-fied. CSF parameters were significantly different in glucose concentration (27 mg% vs. 73 mg%, p < 0.001),

–3 –3

lactate (8 mmol/L vs. 2.8 mmol/L, p < 0.001), CSF neutrophil pleocytosis (850 mm vs. 10 mm , p < 0.001),and protein levels (449 mg% vs. 98 mg%) between the PNBM and non-PNBM groups. The ROC curve thatbest fits PNBM diagnosis is lactate.Conclusion: Increased CSF lactate is a useful PNBM marker, with better predictive value than CSF hypo-glycorrhachia or pleocytosis. Lactate levels ≥4 mmol/L showed 97% sensitivity and 78% specificity, with

valu

a 97% negative predictive

. Introduction

Bacterial meningitis after neurosurgical procedures is rela-ively uncommon, with an estimated incidence between 0.3% and.5% following different intradural interventions [1]. However, itemains a significant problem associated with high morbidity, andeported mortality rates range between 20% and 50% [2]. Accu-ate and rapid diagnosis of bacterial meningitis is essential, sincehe infection outcome will depend on early and specific antibioticherapy [3,4]. Unfortunately, post-neurosurgical bacterial meningi-

is (PNBM) diagnosis is difficult. Bacterial and chemical meningitishare clinical signs and symptoms, such as neck stiffness, fever,omiting and headache [5]. Furthermore, standard cerebrospinal

� This study was conducted at the Raúl Carrea Neurological Research Institute,LENI in Buenos Aires, Argentina.∗ Corresponding author at: Montaneses 2325, C1428AQK Buenos Aires, Argentina.

el.: +54 11 5777 3200; fax: +54 11 5777 3209.E-mail address: p maskin@yahoo.com (L.P. Maskin).

303-8467/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.clineuro.2013.05.034

e.© 2013 Elsevier B.V. All rights reserved.

fluid (CSF) parameters are also unreliable for prompt PNBM diag-nosis [2,6]. Blood, surgical manipulation, surgical materials andbone dust appear to trigger inflammatory processes mimicking bac-terial meningitis CSF changes. Neutrophil CSF pleocytosis can bepresent in various infectious and non-infectious forms of meningi-tis as a consequence of subarachnoid-space inflammation [7–10].CSF testing, Gram staining and bacterial culture results are evenmore difficult to analyze, since up to 50% of patients with clini-cal signs of meningitis have already received steroids or antibioticsby the time a lumbar puncture is performed. Consequently, pro-longed empiric antibiotic therapy is frequently prescribed withoutPNBM confirmation cultures in patients presenting typical bacterialmeningitis CSF findings.

The ideal biochemical PNBM marker should be highly sensitiveand present for early measurement in the course of infection. CSFlactic acid has been proposed as a potential marker distinguish-

ing between infection and inflammation [11]. When CSF infectionis present, leukocytes increase anaerobic glucose metabolism,thus producing lactic acid and lowering pH [11–15]. Bacterialmetabolism, neutrophil glycolysis and anaerobic metabolism of

and N

bmautcwdsesdnHp

as

2

lg

–

–

–

2iirtfrititttahprsaaDtR

w[

(

L.P. Maskin et al. / Clinical Neurology

rain tissue during bacterial meningitis lead to CSF lactate accu-ulation [16]. Lactate slowly crosses the blood–brain barrier

nd has decreased CSF clearance, consequently representing aseful indicator of brain metabolism [17]. CSF lactate is usedo distinguish bacterial from viral meningitis in children (dis-riminatory threshold between 3.5 and 4 mmol/L [11–19]), asell as to discriminate between types of inflammatory response,emonstrating good correlation with infection in several retro-pective series [7–10]. In a retrospective study, Leib et al. [2] havevaluated glucose CSF:blood ratio and CSF lactate levels in neuro-urgical patients for proven or probable PNBM vs. no meningitisiagnosis. Sensitivity and specificity were obtained to discrimi-ate between them, using a CSF lactate cut-off value > 4 mmol/L.owever, meningitis diagnosis was strongly based on CSFleocytosis.

In this study, CSF lactate accuracy for PNBM diagnosis (provennd probable) was investigated prospectively and using a moretrict definition of PNBM.

. Patients and methods

Study design: Prospective observational diagnostic study.Primary objective: To evaluate the diagnostic accuracy of CSF

actate as a PNBM marker in patients hospitalized after a neurosur-ical procedure.

Secondary objectives

To evaluate optimal CSF lactate cut-off values to discriminatePNBM.

To evaluate CSF lactate correlation with other CSF markers inPNBM.

To describe CSF parameters and bacterial cultures in this groupof patients.

Patients and setting: The study was conducted between July005 and June 2009 at the FLENI “Raul Carrea” Institute of Neurolog-

cal Research. Patients hospitalized up to 30 days after neurosurgeryn an 11-bed neurological and neurosurgical adult ICU and the Neu-osurgery Ward were screened. Initial CSF tests performed duringhe first episode of clinically suspected PNBM in patients presentingever (temperature > 38.3 ◦C) or unexplained neurological deterio-ation were reviewed. Patients under antibiotic therapy for othernfections were excluded (patients with perioperative prophylac-ic antibiotics were not excluded). Demographic data obtainedncluded age, gender, APACHE II score, Glasgow Coma Scale (GCS),ype of surgery, pathology, antibiotic prophylaxis, and day whenhe lumbar puncture was performed. CSF samples were obtainedhrough lumbar puncture. CSF biochemistry included: gram stainnd culture, white and red blood cell counts (XS-1000i Sysmexematology analyzer, Sysmex Corporation, Kobe, Japan), as well asrotein, glucose and lactate levels; blood biochemistry was alsooutinely performed by the central laboratory within 2 h of CSFampling. Chemistry assays were carried out using commerciallyvailable kits following the manufacturer’s specifications (glucosend proteins using ROCHE Hitachi 902 Chemistry Analyzer, Rocheiagnostics, F. Hoffmann-La Roche AG, Basel, Switzerland; and lac-

ate using Cobas b 221 systems, Roche Diagnostics, F. Hoffmann-Laoche AG, Basel, Switzerland).

Prior to analysis and blinded to lactate values, patient statusas categorized following predefined criteria (adapted from Leib

2] and Eross [11]) as one of the following:

1) Proven PNBM: positive bacterial CSF culture or Gram stain, plusCSF WBC count ≥100/mL and/or hypoglycorrhachia (<40 mg/dLin CSF or CSF:plasma glucose ratio <0.4).

eurosurgery 115 (2013) 1820– 1825 1821

(2) Presumed PNBM: Patients who received antibiotic 24 h beforeCSF sample, plus ≥250 WBC/mL and a CSF:plasma glucose ratio<0.5.

(3) No-PNBM: neither proven nor presumed criteria.

In this study, all patients received routine antibiotic prophylaxisbefore and after craniotomy following hospital guidelines, usuallyfirst generation cephalosporins. Emergency surgery was defined,as a life-threatening neurosurgical condition, which required anurgent neurosurgical procedure to prevent further deteriorationand/or death. These included: extra-axial hematoma evacua-tion, posterior fossa hematoma drainage, and ICP monitoring forintracranial hypertension.

CSF lactate cut-off value: Previously published data on dif-ferential diagnosis between bacterial and nonbacterial meningitisused different thresholds for CSF lactate, ranging between 3.5 and4.2 mmol/L. We defined positive, or abnormal, CSF lactate levels as≥4 mmol/L, as in previous studies [2].

Statistical analysis: Sample size was calculated based on at least35 negative CSF lactate test results to determine a true negativepredictive value with a 95% confidence interval and ±10% error,assuming CSF lactate performance had a 90% negative predictivevalue for PNBM diagnosis.

Descriptive statistics were used when appropriate, and provenand presumed meningitis combined as a single group (PNBM)for statistical analysis. Positive CSF lactate (≥4 mmol/L) test resultcapacity to diagnose PNBM was assessed using specificity, sensi-tivity, as well as negative and positive predictive values. Receiveroperating characteristic (ROC) curves were applied to evaluate dis-criminatory power of each CSF marker. Continuous variables wereanalyzed using t-test or Wilcoxon rank-sum test, and categoricalvariables using chi square or Fisher’s exact test. For correlationanalysis, Pearson or Spearman correlation coefficient was appliedas appropriate. Two-sided tests were used and p values < 0.05 wereconsidered statistically significant. Statistical analysis was repeatedexcluding patients who had received antibiotic less than 24 h beforeCSF sample collection (proven PNBM vs. non-PNBM). We usedIntercooled Stata 8.2 (StataCorp LP, TX, USA) for statistical analysis.

Institutional Review Board approval was obtained for the study.No diagnostic or therapeutic decisions were made based on theprotocol; all procedures were conducted in accordance with IRBethical standards as per the Declaration of Helsinki (1975).

3. Results

During the 4-year study period, 2718 neurosurgeries were per-formed and 79 CSF samples obtained from suspected PNBM cases.Following the above mentioned clinical criteria, 26 patients hadproven PNBM (33%), 7 presumed PNBM (9%) and 46 no-PNBM (58%).

Demographic data are shown in Table 1. Fifty-one patients weremales (65%) and overall mean age was 50 (SD ± 18 years). Mostcommon indications for the surgical procedures performed wereintracranial tumors (45 patients) or hemorrhages (24 patients).Emergency surgery was performed in 24 patients and was asso-ciated with increased PNBM incidence. No differences in motordeficit, GCS scores, CSF fistula rates or intracranial pressure moni-toring were observed between groups. Mortality rate was 18% forPNBM (both proven and presumed), and 7% for patients in the non-meningitis group (p = 0.15).

As shown in Table 2, the most frequently isolated microor-ganisms were coagulase-negative Staphylococci and Acinetobacter

species. Nine cultures were negative (27% of all meningitis), 2 withproven PNBM and positive Gram stain, and the remaining 7 consid-ered to be presumed PNBM. All these negative culture CSF sampleswere obtained while patients were under antibiotic treatment.

1822 L.P. Maskin et al. / Clinical Neurology and Neurosurgery 115 (2013) 1820– 1825

Table 1Demographic data.

All patients (n = 79) No meningitis group (n = 46) Meningitis group (n = 33) p value

Age 50 (32–68) 49 (32–66) 53 (33–71) 0.34Male (%) 51 (65) 34 (74) 17 (52) 0.04Lumbar puncturea (days after surgery) 6 (3–8) 5 (2–6) 8 (5–10) 0.001GCSa 13 (9–14) 12 (9–14) 13 (6–14) 0.97Motor deficita 12 (15) 5 (11) 7 (21) 0.21Antibiotic prophylaxis (<24 h) 57 (74) 37 (83) 22 (65) 0.65Indication for surgeryb

Emergency procedure 24 (30) 18 (39) 6 (18) 0.046Tumor 45 (57) 24 (52) 21 (64) 0.3Intracranial bleeding 24 (30) 18 (39) 6 (18) 0.046Posterior fossa 17 (22) 10 (22) 7 (21) 0.96

CSF fistulae 9 (11) 6 (13) 3 (9) 0.72Intracranial pressure monitoring 22 (28) 13 (28) 9 (27) 0.91Mortality 9 (11) 3 (7) 6 (18) 0.15

Data expressed as absolute number (percentage).Abbreviations: GCS: Glasgow Coma Scale; CSF: cerebral-spinal fluid.

a Obtained at time of lumbar puncture expressed as median (25–75 quartile).b Sum of surgery category may be more than 100%, because some patients were subjected to more than one procedure.

Table 2Cerebrospinal fluid bacteriological culture results.

Bacteria isolated Positive gram staining Positive cultures Percentage (%)

Coagulase-negative Staphylococci 4 6 18Acinetobacter 3 4 12Staphylococcus aureus 2 3 9Escherichia coli 2 2 6Enterococcus faecalis 0 2 6Klebsiella pneumoniae 2 2 6Achromobacter xylosoxidam 1 1 3Bacillus 1 1 3Corynebacteriun sp. 1 1 3Pseudomonas aeruginosa 0 1 3Streptococcus mitis 0 1 3Negative Culture 2 0 27a

Total 18 24 100

Data expressed as absolute numbers, except percentages.and w

n

pp(tp

wgnp

dP(9tc

d0

eCrW

a Nine cultures were negative (27% of meningitis cases), two were proven PSBM

egative culture CSF samples were taken under antibiotic treatment.

Mean lactate CSFs were significantly higher for patients withroven PNBM than for no-PNBM patients (8 mmol/L vs. 2.8 mmol/L;

< .001), and for presumed PNBM than for no-PNBM patients7.2 mmol/L vs. 2.8 mmol/L; p < .001) (Fig. 1). There were no sta-istically significant differences in CSF lactate between proven andresumed PNBM (p = 0.38).

CSF parameters are listed in Table 3. CSF WBC and protein levelsere higher in PNBM. Inversely, CSF glucose level and CSF:blood

lucose ratio were lower in these groups (p < .001). There wereo statistically significant differences in CSF parameters betweenroven and presumed PNBM cases.

Evaluation of CSF lactate ≥4 mmol/L performance as a markeristinguishing between PNBM (proven and presumed) and non-NBM showed 97% sensitivity (95%CI 84–100%), 78% specificity95%CI 64–89%), 76% positive predictive value (95%CI 60–88%) and7% negative predictive value (95%CI 86–100%). At this cut-off level,here was only one false negative detected in a presumed PNBMase under antibiotic therapy.

ROC analysis of CSF markers is shown in Fig. 2, with CSF lactateemonstrating excellent discriminatory power (AUC 0.96, 95%CI.93–1.00%).

Correlation between CSF lactate and different CSF mark-

rs was also assessed. CSF lactate inversely correlated withSF blood glucose ratio (r = −0.732; p < 0.001), but directly cor-elated with CSF protein levels (r = 0.67; p < 0.001) and CSF

BC count (r = 0.76; p < 0.001). No correlation was observed

ere Gram-positive, and the remaining seven were categorized presumed PSBM. All

between CSF lactate and CSF red blood cells count (r = 0.255;p = 0.06).

Average number of days between surgery and CSF sampling wasgreater in PNBM than in no-PNBM patients (8 vs. 5 days, p < 0.001).

Statistical analysis was repeated after excluding patients whohad received antibiotic 24 h before CSF sampling. Then, there wereonly two groups: proven PNBM (19 patients) and non-PNBM (37patients) cases. CSF lactate ≥4 mmol/L also demonstrated good dis-criminatory power (AUC 0.96; 95%CI 0.93–1.00%), distinguishingbetween proven PNBM and no-PNBM, with 100% sensitivity (95%CI82–100%), 78% specificity (95%CI 62–90%), 70% positive predictivevalue (95%CI 42–86%) and 100% negative predictive value (95%CI88–100%).

4. Discussion

Our study confirms that CSF lactate has high sensitivity andacceptable specificity in distinguishing between PNBM and asepticmeningitis, even in patients under antibiotic treatment. To the bestof our knowledge, this is the largest prospective study on the use ofCSF lactate to discriminate between PNBM and other inflammatoryCSF conditions.

CSF acidosis due to bacterial infection has been known for along time [12–14], and CSF lactate has been available circa 1920.However, despite its use in the pediatric population to differenti-ate between bacterial and viral meningitis [11,20], the use of CSF

L.P. Maskin et al. / Clinical Neurology and Neurosurgery 115 (2013) 1820– 1825 1823

Fig. 1. Lactate CSF levels were significantly higher for patients with PNBM (both proven and presumed) compared to no-PNBM patients (p < 0.0001). Data is expressed asmedian (IQR 25–75%).

Table 3Cerebrospinal fluid parameters results.

No meningitis group (n = 46) Proven meningitis group (n = 26) Presumed meningitis group (n = 7) p value

Leucocytes (mm−3) 10 (3–120) 850 (460–3360) 2500 (840–7500) <0.001a

<0.001b

0.11c

Glucose (mg/dL) 73 (52–84) 27 (13–36) 32 (15–42) <0.001a

<0.001b

0.9c

CSF:blood glucose ratio 0.59 (0.43–0.65) 0.21 (0.1–0.3) 0.19 (0.13–0.26) <0.001a

<0.001b

0.51c

Proteins (mg/dL) 98 (53–176) 449 (175–812) 334 (88–428) <0.001a

<0.05b

0.14c

Lactate (mmol/L) 2.8 (2.2–3.7) 8 (6.3–11) 7.2 (5.7–10.4) <0.001a

<0.001b

0.38c

a Data are expressed as median (25–75 quartile). p values were between the following groups: no meningitis vs. proven meningitis.

laTase

aLitgdop

b No meningitis vs. presumed meningitis.c Proven meningitis vs. presumed meningitis.

actate in the adult population has been limited because of the avail-bility of the test and questions that arose regarding its utility [2].he CSF lactate threshold proposed was between 3.5 and 4 mmol/L,chieving an acceptable sensitivity and specificity [17,18]. Recently,everal meta-analyses studied the accuracy of CSF lactate to differ-ntiate bacterial from aseptic meningitis [21,22].

In the post-neurosurgical setting, only one retrospective studyssessing CSF lactate role in PNBM diagnosis has been published [2].eib et al. retrospectively collected data from 477 lumbar puncturesn PNBM suspected cases, and analyzing 77 of them, found a bet-er correlation with CSF lactate (cut-off: 4 mmol/L) than CSF/blood

lucose ratio (cut-off: 0.4), in positive PNBM cases. However, theirefinition of PNBM also included inflammatory changes, like thenes observed in meningitis cases. Our prospective findings sup-ort this, as we also found better correlation with CSF lactate than

with hypoglycorrhachia. Tavares et al. [23] evaluated the diagnos-tic usefulness of different CSF parameters for PNBM detection, andfound CSF glucose and lactate are good discriminatory markers forbacterial infection (ROC 0.85 and 0.85 respectively). We found sim-ilar results for lactate (0.96, range: 0.93–1.00). Our study showedthat CSF lactate was also effective in distinguishing between CSFsamples taken while on antibiotic therapy, when CSF cultures werenegative; a finding also published in the report by Cunha [19] andsuggested by other authors [24,25].

CSF lactate levels were commonly thought to be influenced byCSF red blood cell counts, due to their intrinsic metabolism. Also,

increased lactic levels secondary to red blood cell production havebeen documented in mock CSF solutions [26], as well as in caninemodels [27]. Therefore, interpretation of lactate concentrationsin blood-stained CSF may require estimating a rise in lactic acid

1824 L.P. Maskin et al. / Clinical Neurology and N

Fig. 2. Receiver operating curve (ROC) showing sensitivity and specificity ofdifferent markers in cerebrospinal fluid for post-surgical bacterial meningitis diag-n0

aCCow

bttpNCwmiflbettttnhhttc

ild

ieisb

cat[

t

[

[

[

[

osis. Lactate level showed an excellent discriminatory value (AUC 0.96, 95%CI.93–1.00%).

ttributable to red cells [28]. We reviewed our results for potentialSF blood contamination affecting lactate concentration and foundSF lactate values were not significantly affected by the presencef red cells, in agreement with results from Begovac’s study [29] asell as Leib’s [2].

However, low lactic acid CSF values do not always rule outacterial meningitis. Other factors may reduce isolate concentra-ions, and interpretation of findings should always be carried inhe context of each clinical case [30]. CSF lactate is another tool forost-surgical meningitis diagnosis, and should be used with care.ew biomarkers, such as sCD163 [31], procalcitonin (PCT) [31,32],-reactive protein (CRP) [32], or the heparin-binding protein [33],ere proposed to try to distinguish between bacterial and asepticeningitis. The PCT diagnostic accuracy was higher for distinguish-

ng between bacterial and viral infection in patients with spinaluid pleocytosis. The heparin-binding protein also distinguishedetween bacterial meningitis and other meningeal infections. How-ver, none of the markers was useful as an independent tool forhe clinical diagnosis of patients with purulent meningitis, andhey were seldom used to distinguish between PNBM and asep-ic meningitis. In the past decade, imaging techniques have beenested to characterize biochemical profiles. Proton nuclear mag-etic resonance-based metabonomics [34] and spectroscopy [35]ave been tested for rapid meningitis diagnosis. These pilot studiesave suggested that meningitis can be diagnosed through imagingechniques, analyzing the metabolic profile. However, because ofheir costs and complexity, these techniques are not available tolinicians, and are used only for research purposes.

The exact incidence of bacterial meningitis after neurosurgerys unknown, probably due to the difficulty in its diagnosis and theack of a gold standard. Reports range from 0.0007% to 14.8% [36,37],epending on several factors.

Staphylococcal group species were the most common pathogensn PNBM in this series [38–40], as it is usually described in the lit-rature. Acinetobacter and Klebsiella were also identified. Althoughn the past Gram negative bacteria seldom caused central nervousystem infections after neurosurgery in adults, they are currentlyecoming an increasingly common cause of PNBM [41,42].

Time between surgery and meningitis symptoms was signifi-antly longer in the PNBM group than in the non-PNBM one. Aslready mentioned, we believe this time difference may be due

o earlier development of chemical meningitis compared to PNBM6,7].

Nevertheless, the study has several limitations. First, informa-ion was collected from a single institution with standardized

[

[

eurosurgery 115 (2013) 1820– 1825

clinical management, which may not reflect worldwide standardof care. Second, since there was not an appropriate definition forPNBM, we chose clinical criteria to discriminate between provenPNBM, probable PNBM and absence of meningitis, but these are notuniversally accepted criteria. Zarrouk et al. published a local guideto manage aseptic meningitis, analyzing data on before and afterresults, in which PNBM was defined as a positive CSF culture andwith WBC counts > 100/mL, and there was no increase in mortalityor neurological morbidity [30]. Our criteria were similar to thoseused in Zarrouk’s study, and it may allow confident use of thesedefinitions. Third, the study was not powered to have CSF lactatedifferentiate between those with proven and presumed PNBM, andthis could be an area for future research. Finally, arterial lactate wasnot always measured in plasma, and its influence in CSF lactate isunknown. But lactate seldom passes through the brain–blood bar-rier, so the CSF lactate measurement should not be influenced byplasma levels [43].

In conclusion, our study shows that CSF lactate is reliable to dis-tinguish between PNBM and aseptic meningitis. High sensitivityand negative predictive values are achieved with a cut-off valueof 4 mmol/L, and this value is specially useful to rule out post-neurosurgical bacterial meningitis in the adequate clinical setting.

Funding

None.

Conflict of interest statement

All authors declare no conflict of interest in relation to the sub-ject of this study.

References

[1] Blomstedt GC. Infections in neurosurgery: a retrospective study of 1143patients and 1517 surgeries. Acta Neurochirurgica 1985;78:81–90.

[2] Leib S, Boscacci R, Gratzl O, Zimmerli W. Predictive value of cerebrospinalfluid (CSF) lactate levels versus CSF/blood glucose ratio for the diagnosisof bacterial meningitis following neurosurgery. Clinical Infectious Diseases1999;29:69–74.

[3] Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Durationof hypotension before initiation of effective antimicrobial therapy is the crit-ical determinant of survival in human septic shock. Critical Care Medicine2006;34:1589–96.

[4] Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. SurvivingSepsis Campaign: international guidelines for management of severe sepsis andseptic shock. Intensive Care Medicine 2013;39:165–228.

[5] Sanchez GB, Kaylie DM, O‘Malley MR, Labadie RF, Jackson G, Haynes DS. Chemi-cal meningitis following cerebellopontine angle tumor sugery. Otolaryngology– Head and Neck Surgery 2008;138:368–73.

[6] Ross D, Rosegay H, Pons V. Differentiation of aseptic and bacterial menin-gitis in postoperative neurosurgical patients. Journal of Neurosurgery1988;69:669–74.

[7] Forgacs P, Geyer CA, Freiberg SR. Characterization of chemical meningitis afterneurological surgery. Clinical Infectious Diseases 2001;32:179–85.

[8] Maskin LP, Mora A, Orellana N, et al. Evaluations of cerebrospinal fluid lac-tate in the diagnosis of post-neurosurgical bacterial meningitis. Intensive CareMedicine 2007;33:S194.

[9] Salord F, Boussaid O, Eynard N, Perret C, Grando J, Chacornac R. Value ofD(−) lactate determination for the fast diagnosis of meningitis after cran-iotomy: an initial study. Annales Francaises D Anesthesie Et De Reanimation1994;13:647–53.

10] Genton B, Berger JP. Cerebrospinal fluid lactates in 78 cases of adult meningitis.Intensive Care Medicine 1990;16:196–200.

11] Eross J, Silink M, Dorman D. Cerebrospinal fluid lactic acidosis in bacterialmeningitis. Archives of Disease in Childhood 1981;56:692–8.

12] Levinson A. The hydrogen-ion concentration of cerebrospinal fluid: studies inmeningitis. The Journal of Infectious Diseases 1917;21:556–70.

13] Killian JA. Lactic acid of normal and pathological spinal fluids. Proceedings of

the Society for Experimental Biology and Medicine 1925;23:255–7.

14] De Sanctis AG, Killian JA, Garcia T. Lactic acid of spinal fluid in meningitis.American Journal of Diseases of Children 1933;46:239–49.

15] Fierro H, Chacón A, Gutierrez G, Munoz O. pH and lactic acid values inthe cerebrospinal fluid as a diagnostic aid for differentiating viral and

and N

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

L.P. Maskin et al. / Clinical Neurology

purulent meningoencephalitis. Archivos de Investigacion Médica (México)1977;8(3):185–97.

16] Guerra-Romero L, Guerra-Romero L, Täuber MG, Fournier MA, Tureen JH. Lac-tate and glucose concentrations in brain interstitial fluid, cerebrospinal fluid,and serum during experimental pneumococcal meningitis. The Journal of Infec-tious Diseases 1992;166:546–50.

17] Cameron PD, Boyce JM, Ansari BM. Cerebrospinal fluid lactate in meningitis andmeningococcaemia. Journal of Infection 1993;26:245–52.

18] Pavese P, Francois P, Lafond JL, Kayemba Kay’S S, Bosson JL. Assay of lactic acidin the cerebrospinal fluid for the diagnosis of bacterial meningitis. Strategies forthe choice of discriminatory threshold. La Presse Médicale 1997;26(12):551–4.

19] Cunha BA. Distinguishing bacterial from viral meningitis: the critical impor-tance of the CSF lactic acid levels. Intensive Care Medicine 2006;32:1272–3.

20] Berg B, Gärdsell P, Skånsberg P. Cerebrospinal fluid lactate in the diagnosisof meningitis. Diagnostic value compared to standard biochemical methods.Scandinavian Journal of Infectious Diseases 1982;14:111–5.

21] Huy NT, Thao NT, Diep DT, Kikuchi M, Zamora J, Hirayama K. Cerebrospinalfluid lactate concentration to distinguish bacterial from aseptic meningitis: asystemic review and meta-analysis. Critical Care 2010;14:R240.

22] Sakushima K, Hayashino Y, Kawaguchi T, Jackson JL, Fukuhara S. Diagnosticaccuracy of cerebrospinal fluid lactate for differentiating bacterial meningitisfrom aseptic meningitis: a meta-analysis. Journal of Infection 2011;62:255–62.

23] Tavares WM, Machado AG, Matushita H, Plese JPP. CSF markers for diagnosisof bacterial meningitis in neurosurgical postoperative patients. Arq Neurop-siquiatr 2006;64(3-A):592–5.

24] Komorowski RA, Farmer SG, Knox KK. Comparison of cerebrospinal fluid C-reactive protein and lactate for diagnosis of meningitis. Journal of ClinicalMicrobiology 1986;24:982–5.

25] Cabec a HL, Gomes HR, Machado LR, Livramento JA. Dosage of lactate in thecerebrospinal fluid in infectious diseases of the central nervous system. ArqNeuropsiquiatr 2001;59:843–8.

26] Froman C, Smith AC. Metabolic acidosis of the cerebrospinal fluid associatedwith subarachnoid haemorrhage. Lancet 1967;1:965–7.

27] Shannon DC, Shore N, Kazemi H. Acid–base balance in haemorrhagic cere-brospinal fluid. Neurology 1972;22:585–9.

28] Venkatesh B, Morgan TJ, Boots RJ, Hall J, Siebert D. Interpreting CSF lactic

acidosis: effect of erythrocytes and air exposure. Critical Care & Resuscitation2003;5:177–81.

29] Begovac J, Bace A, Soldo I, Lehpamer B. Lactate and glucose in cerebrospinalfluid heavily contaminates with blood. Acta Medica Croatica 1991;45(4–5):341–5.

[

[

eurosurgery 115 (2013) 1820– 1825 1825

30] Zarrouk V, Vassor I, Bert F, Bouccara D, Kalamarides M, Bendersky N, et al.Evaluation of the management of postoperative aseptic meningitis. ClinicalInfectious Diseases 2007;44:1555–9.

31] Knudsen TB, Larsen K, Kristiansen TB, Møller HJ, Tvede M, Eugen-Olsen J,et al. Diagnostic value of soluble CD163 serum levels in patients suspectedof meningitis: comparison with CRP and procalcitonin. Scandinavian Journalof Infectious Diseases 2007;39:542–53.

32] Shimetani N, Shimetani K, Mori M. Levels of three inflammation markers, C-reactive protein, serum amyloid A protein and procalcitonin, in the serum andcerebrospinal fluid of patients with meningitis. Scandinavian Journal of Clinical& Laboratory Investigation 2001;61:567–74.

33] Linder A, Akesson P, Brink M, Studahl M, Björck L, Christensson B. Heparin-binding protein: a diagnostic marker of acute bacterial meningitis. Critical CareMedicine 2011;39:812–7.

34] Coen M, O’Sullivan M, Bubb WA, Kuchel PW, Sorrell T. Proton nuclear mag-netic resonance-based metabonomics for rapid diagnosis of meningitis andventriculitis. Clinical Infectious Diseases 2005;41:1582–90.

35] Himmelreich U, Malik R, Kuhn T, Daniel HM, Somorjai RL, Dolenko B, et al.Rapid etiological classification of meningitis by NMR spectroscopy based onmetabolite profiles and host response. PLoS ONE 2009;4:e5328.

36] Baltas I, Tsoulfa S, Sakellariou P, Vogas V, Fylaktaki M, Kondodimou A. Posttrau-matic meningitis: bacteriology, hydrocephalus, and outcome. Neurosurgery1994;35:422–7.

37] Buckwold FJ, Hand R, Hansebout RR. Hospital-acquired bacterial meningitis inneurosurgical patients. Journal of Neurosurgery 1977;46:494–500.

38] Korinek AM. Risk factors for neurosurgical site infections after craniotomy: aprospective multicenter study of 2944 patients. The French Study Group ofNeurosurgical Infections, the SEHP, and the C-CLIN Paris-Nord. Service Epi-demiologie Hygiene et Prevention. Neurosurgery 1997;41:1073–81.

39] Blomstedt GC. Infections in neurosurgery: a retrospective study of 1143patients and 1517 operations. Acta Neurochirurgica (Wien) 1985;78:81–90.

40] McClelland 3rd S, Hall WA. Postoperative central nervous system infection:incidence and associated factors in 2111 neurosurgical procedures. ClinicalInfectious Diseases 2007;45:55–9.

41] Lu C-H, Chang W-N, Chuang Y-C, Chang H-W. Gram-negative bacillary meningi-tis in adult’s post-neurosurgical patients. Surgical Neurology 1999;52:438–44.

42] Briggs S, Ellis-Pegler R, Raymond N, Thomas M, Wilkinson L. Gram-negativebacillary meningitis after cranial surgery or trauma in adults. ScandinavianJournal of Infectious Diseases 2004;36:165–73.

43] Cameron PD, Boyce JMH, Ansari BM. Cerebrospinal fluid lactate in meningitisand meningococcaemia. Journal of Infection 1993;26:245–52.