Cerebrospinal fluid analysis, predictors of bacte

　　　　 作者：Seyed Mohammad Alavi， Naser Moshiri　　

【摘要】 Objective: Patients with cerebrospinal fluid (CSF) pleocytosis are routinely admitted to the hospital and treated with parenteral antibiotics， although few have bacterial meningitis (BM).The aim of this study was to evaluate predictors to differentiate BM from aseptic meningitis (ASM).Methods:The study was conducted in Razi hospital， a training center affiliated to Ahvaz Joundishapoor University of Medical Sciences in Iran. and all patients were 18 years old or above and were treated in the hospital between 2003 and 2007. Data of those who had meningitis， tested as CSF pleocytosis but had not received antibiotic treatment before lumbar puncture were retrospectively analyzed.Results: Among 312 patients with CSF pleocytosis， two hundred fifteen (68.9%) had BM and ninety seven (31.1%) had ASM. The mean age for patients with BM was (34.7±17.7) years (P=0.22， NS). Sixty percent of the BM cases and 61.2% of the ASM cases occurred in men (P=0.70， NS). We identified the following predictors of BM: CSFWBC count&>100 per micro liter， CSFglucose level&<40 mg/dL， CSFprotein level&>80 mg/dL. Sensitivity， specificity， PPV， NPV of these predictors， and LR for BM are 86.5%，52.6%，80.2%， 63.7% and 104.1 for CSFWBC count and 72.1%， 83.5%， 90.6%，57.4% and 164.2% for CSF glucose， and 49.7%， 91.8%， 93.4%，45.2% and 104.5% for CSF protein.Conclusion:The CSF WBC count should not be used alone to rule out bacterial meningitis. When it is combined with other factors such as CSF glucose and protein improved decision making in patients with suspected BM may occur.

【关键词】 Bacterial meningitis; Predictor; Cerebrospinal fluid

　　INTRODUCTION

　　Although， in developed countries immunization by conjugate vaccine against H. influenza has decreased the incidence of bacterial meningitis (BM)[16]， but， in developing countries BM continues as a major public health problem[2，7，8]. Unfortunately， meningitis still occurs and， if not treated promptly， brings with it severe morbidity and mortality[47，9，10]. A great majority of meningeal infections are caused by nonbacterial organisms such as viruses[11] and unnecessary use of broad spectrum antibacterial drugs lead to emerge multidrug resistant bacteria[7，1214]. The declining incidence of culture positive cerebrospinal fluid (CSF) samples and increased incidences of culturenegative CSF pleocytosis; aseptic meningitis (ASM) are a major confounding factor in approach to patients with suspected central nervous system (CNS) infection[15，16].Receiving antibacterial drugs before lumbar puncture (LP)， result in an increase in the proportion of patients with negative culture results in CSF examination[7，17]. Physicians often are concerned by any degree of CSF pleocytosis in patients who are being assessed for bacterial meningitis. These patients， when no other abnormal clinical or biochemical factors consistent with bacterial meningitis are identified， often are hospitalized and treated empirically with antibiotics[1821]. Previous studies have suggested formulas to estimate a patient's risk for BM[7，17，19，22，23]. It has been shown that the mean CSF white blood cell (WBC) count and CSF glucose are significantly different between patients with BM and those with ASM. Those studies have also shown that mean CSFWBC count， glucose and protein concentration varies in different countries and different pathogenic bacteria.Spanos et al reported that a CSFglucose concentration less than 34 mg/dL，a CSFprotein concentration more than 220 mg/dL， CSFWBC&>2 000/mm3 and CSFPMN&>1 180/mm3 were predictors of BM rather than ASM with 99% certainly[24].The foremost textbook of infectious disease[7，8] indicate that CSFWBC count greater than 5 per micro liter and CSF glucose lower than 60 mg/dL are abnormal.

　　Little information is available in Iran， on positive or negative predictive value (PPV or NPV)， likelihood ratio (LR)， or specific treatment guideline for given CSFWBC and glucose level. The aim of this study was to evaluate the PV and LR of CSFWBC，CSFglucose and CSF protein for BM in hospitalized patients in a teaching hospital in Ahvaz a city located in south west Iran.

　　MATERIAL AND METHODS

　　This study was a retrospective review of the CSF examination for the evaluation of BM. The study population included all patients aged 18 years or above who underwent a LP in the infectious disease ward of Razi hospital to diagnose BM between 2003 and 2007.Exclusion criteria were:(1) clotted samples， (2) systemic infection such as tuberculosis， brucellosis and HIV/AIDS， (3) any underlying medical condition that altered CSF findings such as neoplasms， metabolic diseases and collagen vascular disease， (4) receiving antibiotic before LP.

　　Data were extracted from the medical files of patients with meningitis， which contained information on all blood and CSF samples， submitted for culture， biochemical and cytological examination. Additional information was extracted to ensure that inclusion and exclusion criteria were met. Selection criteria for prediction BM were : (1) positive CSF culture ， (2) diagnosis of BM including presumptive meningitis， (3) CSF WBC count， (4) CSF glucose and protein data， (5) serum glucose data， (6) completed blood cell count(CBC)， (7) clinical variables such as fever， headache， neck stiffness， level of consciousness(LOC)， seizures and focal deficits， and (8) receiving antibiotics before LP. Patients were placed in 2 groups; (1) BM consisted of patients whose CSF culture was positive for common CNS pathogen， CSF pleocytosis in association with decreased CSF glucose and increased CSF protein， and (2) ASM which included all other patients with abnormal CSF examination. Data were analyzed with SPSS (version 16， USA) by using descriptive， chi square， Fisher exact test and likelihood ratio (LR). Sensitivity， specificity， PPV， NPV and LR were determined. The NPV describes the probability that the patient does not have BM and is a valuable indicator to rule out BM. The LR is the likelihood that a given test result is expected in a patient with BM compared with the likelihood that the same result would be expected in a patient without BM. Because LR is less likely to change with the prevalence of the disorder， is preferred to PPV， NPV and specificity. Cutoff values for predictors were: (1) CSF WBC; less than 10，10 to 100， and greater than 100 cell per microlter， (2) CSF glucose &<40 mg/dL，(3) CSF protein &>45 mg/dL， 45 mg/dL～80 mg/dL and greater than 80 mg/dL，(4) CBCPMN&>10×103per micro liter[25，26].

　　RESULTS

　　We identified 457 patients who met the inclusion criteria and excluded 77 patients because of underlying diseases， 65 patients who had received antibiotic treatment before LP， and 15 patients who were suspected to have tuberculosis， brucellosis， or HIV/AIDS. Ptients could also be excluded for more than one reason. Among 312 remaining patients with CSF pleocytosis， 215(68.9%) had BM (definite and presumptive) and 97(31.1%) had ASM. The mean age for patients with BM was (34.7±17.7) years (P=0.22， NS). Sixty percent of the BM cases and 61.2% of the ASM cases occurred in men (P=0.70， NS). Demographic and clinical characteristics are summarized in Table 1. As shown in Table 1， there is no significant difference between 2 groups with the respect to clinical findings. We identified the following predictors of BM: CSFWBC count&>100 per microliter， CSFglucose level&<40 mg/dL， CSFprotein level&>80 mg/dL， CSFPMN&> 80%， and peripheral blood PMN of 10 000 or more. CSF characteristics are shown in Table 2.Mean WBC count， CSF glucose and CSF protein concentration in patients with aseptic meningitis were (356.4±1 631.8) per microliter， (68.7±16.1) mg/dL and (32.6±17.7) mg/dL seperately. On the other hand， the corresponding data in patients with bacterial meningitis were (4 021±10 352.1) per microliter，(48.5±17.3) mg/dLand (89.8±76.8) mg/dL. There were significant difference between them (P=0.001， 0.000 1 and 0.000 1).

　　Sensitivity，specificity， Positive predictive value， negative predictive value and LR of CSF WBC count&>100/μL were 86.5%， 52.6%， 80.2%， 63.7% and 104.1; for CSFglucose (mg/dL)， the correspongding results were 72.1%， 83.5%， 90.6%， 57.4% and 164.2; for CSFprotein (mg/dL)， were 49.7%， 91.8%， 93.4%， 45.2% and 104.5%. These data demonstrated significant differences between these predictors. Of 80 patients without these predictors， 4 cases were culture positive for common bacterial pathogen (3 N.meningitidis， 1 S. pneumoniae)， this means that NPV for these predictors is at least 95%.Table 1 Demographic and clinical variables in patients with meningitis admitted in Razi hospital，Table 2 Crebrospinal fluid analysis in aseptic and bacterial meningitis in patients with meningitis admittedin Razi hospital， Ahvaz， Iran， 2003～2007(n，%).

　　Bacterial meningitis is an important and life threatening infection[7]. Reducing morbidity and mortality is critically dependent on rapid diagnosis[27，28]and， perhaps more importantly， on the timely initiation of appropriate antimicrobial therapy[17]. Differentiating BM from ASM is the first step in making decision for treatment. Bonsu et al reported that among children with cerebrospinal fluid (CSF) pleocytosis， the task of separating aseptic from bacterial meningitis is hampered when the CSF Gram stain result is unavailable， delayed， or negative[29]. Previous studies suggested laboratory predictors with controversial results[7，17，1926].The present study showed that CSFWBC count more than 100 per micro liter is a useful tool for prediction of BM. There is a significant relation between increased CSF WBC count and increased risk of BM. When the CSF WBC is100/μL or more， the PPV is high and more likelihood for BM. When the CSFWBC is less than 100/μL NPV is high and likelihood for BM is low and is similar to CSFWBC count less than 10/ μL. This finding is consistent with previous studies[2526].In suspected case of meningitis with low CSF WBC count; other laboratory tests should be employed to differentiate BM from ASM. In this study 4 cases of confirmed BM( CSF culture positive for common pathogenic agent) had low WBC count in their CSF samples， but had at least one of the following criteria: CSF glucose&<40 mg/dL，CSF protein&> 80 mg/dL，or CSFPMN count&>10 000 cell/mL. The presence of these criteria independently indicate that the patient is at an increased risk for BM. This finding is in agreement with previous studies[26，30，31]. Presence of 4 case of BM in the absence of these variables (3 predictors) means that NPV for these predictors is not 100%， hence， low CSFWBC， normal CSF glucose and protein concentration does not rule out BM. Previous researchers; Freedman et al[26]， Wong et al[30] and Polk et al[31] reported a number of definite BM in patients without CSF pleocytosis. The present study also explained that other laboratory tests such as CSF glucose concentration and CSF protein concentration in combination with CSF pleocytosis may be useful to differentiate BM from ASM. When the glucose level of CSF is lower than 40 mg/dL， PPV is high (90.6%) and likelihood for BM is high， but when CSF glucose is normal or is higher than 40 mg/dL， NPV is less than 60% and does not rule out BM. When CSFprotein is higher than 80 mg/dL， PPV is 93.4% and more likelihood for BM， but when CSF protein is normal or lower than 80 mg/dL， NPV is decreased and BM can not be ruled out. Our study showed that， although these predictors as well as previous studies are useful in separating BM from ASM， but， CSF WBC cutoffs and protein level are similar or different of those studies[17，19-26，29，3233] . This finding suggests that cutoffs these predictors may be varied in different countries. As are described in literature[7，26，29]， this study showed that clinical variables such as fever， head ache， neck stiffness， nausea， vomiting， Kernic， Bruzensky， seizure， and LOC are not predictors of BM.

参考文献

1 Dery M， Hasbun R. Changing epidemiology of bacterial meningitis. Curr Infect Dis Rep 2007; 9(4):301307.

　　2 FayeKetté H， Doukou ES， Boni C， AkouaKoffi C， DialloTouré K， KacouN'Douba A， et al. Agents of community acquired purulent meningitis in the child: epidemiologic trends in Abidjan， Cte d'Ivoire， from the year 1995 to 2000. Bull Soc Pathol Exot 2003; 96(4):313316.

　　3 Kyaw MH， Christie P， Jones IG， Campbell H. The changing epidemiology of bacterial meningitis and invasive nonmeningitic bacterial disease in scotland during the period 198399. Scand J Infect Dis 2002; 34(4):289298.

　　4 Neuman HB， Wald ER. Bacterial meningitis in childhood at the Children's Hospital of Pittsburgh: 1988～1998. Clin Pediatr (Phila) 2001; 40(11):595600.

　　5 Theodoridou MN， Vasilopoulou VA， Atsali EE， Pangalis AM， Mostrou GJ， Syriopoulou VP， et al. Meningitis registry of hospitalized cases in children: epidemiological patterns of acute bacterial meningitis throughout a 32year period. BMC Infect Dis 2007; 30; 7:101.

　　6 Peltola H， Salo E， Saxen H. Incidence of Haemophilus influenzae type b meningitis during 18 years of vaccine use: observational study using routine hospital data. BMJ 2005; 330:1819.

　　7 Tunkel AR， ScheldWM. Acute meningitis. In: Mandell GL， Bennett JE， Dolin R. (eds). Principle and Practice of infectious diseases. 6th ed. USA， Philadelphia:Churchill Livingstone， 2005.

　　8 Behrman RE， Kliegman RM， Jenson HB. Nelson Textbook of Pediatrics. 16th ed. Philadelphia， Pa: WB Saunders Co; 2000.

　　9 Husain EH， AlShawaf F， Bahbahani E， ElNabi MH， AlFotooh KA， Shafiq MH， et al. Epidemiology of childhood meningitis in Kuwait. Med Sci Monit 2007; 13(5):CR220223.

　　10Fleisher GR. Central nervous system infections. In: Fleisher GR， Ludwig SL， Henretig FM， eds. Textbook of Pediatric Emergency Medicine. 5th ed. Philadelphia， Pa: Lippincott Williams && Williams; 2006.788795.

　　11 Centers for Disease Control and Prevention.Outbreaks of aseptic meningitis associated with echoviruses 9 and 30 and preliminary surveillance reports on enterovirus activityUnited States， 2003. MMWR Morb Mortal Wkly Rep 2003; 52:761764.

　　12 Welinder Olsson C， Dotevall L， Hogevik H， Jungnelius R， Trollfors B， Wahl M， et al. Infection with PantonValentine leukocidinpositive methicillinresistant Staphylococcus aureus t034.Emerg Infect Dis 2008;14(8):12711272.

　　13 Kaplan S. Streptococcus pneumonia: impact of antibiotic resistance in pediatrics. Curr Probl Pediatr 1997; 27:187195.

　　14 Kyaw MH， Lynfield R， Schaffner W， Craig AS， Hadler J， Reingold A， et al. Effect of introduction of the pneumococcal conjugate vaccine on drugresistant Streptococcus pneumoniae. N Engl J Med 2006;354:14551463.

　　15 Duan QJ， Shang SQ， Wu YD. Rapid diagnosis of bacterial meningitis in children with fluorescence quantitative polymerase chain reaction amplification in the bacterial 16S rRNA gene. Eur J Pediatr 2008 Jun 12.[Epub ahead of print]

　　16 Husain EH， Bahbahani E， AlShawaf F， ElNabi MH， Shafiq MH， AlAteeqi N， et al. Aseptic meningitis among children in Kuwait. Med Princ Pract 2008; 17(2):122125.

　　17 Ziai WC， Lewin JJ. Update in the diagnosis and management of central nervous system infections. Neurol Clin 2008; 26(2):427468

　　18 Stille CJ， RifasShiman SL， Kleinman K， Kotch JB， Finkelstein JA. Physician responses to a communitylevel trial promoting judicious antibiotic use. Ann Fam Med 2008; 6(3):206212.

　　19 Hoen B， Viel JF， Paquot C， Gerard A， Canton P. Multivariate approach to differential diagnosis of acute meningitis. Eur J Clin Microbiol Infect Dis 1995; 14:267274.

　　20 Oostenbrink R， Moons KG， Twijnstra MJ， Grobbee DE， Moll HA. Children with meningeal signs: predicting who needs empiric antibiotic treatment. Arch Pediatr Adolesc Med 2002; 156:1.

　　21Bonsu BK， Harper MB. Differentiating acute bacterial meningitis from acute viral meningitis among children with cerebrospinal fluid pleocytosis: a multivariable regression model. Pediatr Infect Dis J 2004; 23:511517.

　　22 Reilly BM， Evans AT. Translating clinical research into clinical practice: impact of using prediction rules to make decisions. Ann Intern Med 2006; 144:201209.

　　23 Nigrovic LE， Kuppermann N， Malley R. Development and validation of a multivariable predictive model to distinguish bacterial from aseptic meningitis in children in the postHaemophilus influenzae era. Pediatrics 2002; 110:712719.

　　24 Spanos A， Harrel E， Durack DT.Differential diagnosis of acute meningitis. An analysis of the predictive value of initial observation.JAMA 1989; 262:27002707.

　　25 Nigrovic LE， Kuppermann N， Macias CG， Cannavino CR， MoroSutherland DM， Schremmer RD，et al. Clinical prediction rule for identifying children with cerebrospinal fluid pleocytosis at very low risk of bacterial meningitis. JAMA 2007; 3; 297(1):5260.

　　26 Freedman SB， Marrocco A， Pirie J， Dick PT. Predictors of Bacterial Meningitis in the Era After Haemophilus influenzae. Arch Pediatr Adolesc Med 2001; 155:13011306.

　　27 Jansen GJ， Mooibroek M， Idema J， Harmsen HJ， Welling GW， Degener JE. Rapid identification of bacteria in blood cultures by using fluorescently labeled oligonucleotide probes. J Clin Microbiol 2000; 38:814817.

　　28 Poppert S， Essig A， Stoehr B， et al. Rapid diagnosis of bacterial meningitis by realtime PCR and fluorescence in situ hybridization. J Clin Microbiol 2005; 43:33903397.

　　29 Bonsu BK， Ortega HW， Marcon MJ， Harper MB. A decision rule for predicting bacterial meningitis in children with cerebrospinal fluid pleocytosis when gram stain is negative or unavailable. Acad Emerg Med 2008; 15(5):437444.

　　30 Wong VK， Hitchcock W， Mason WH. Meningococcal infections in children: a review of 100 cases. Pediatr Infect Dis J 1989; 8:224227.

　　31 Polk DB， Steele RW.Bacterial meningitis presenting with normal cerebrospinal fluid. Pediatr Infect Dis J 1987; 6(11):10401042.

　　32 Bonsu BK， Harper MB. Accuracy and test characteristics of ancillary tests of cerebrospinal fluid for predicting acute bacterial meningitis in children with low white blood cell counts in cerebrospinal fluid. Acad Emerg Med 2005; 12(4):303309.

　　33 Negrini B， Kelleher KJ， Wald ER. Cerebrospinal fluid findings in aseptic versus bacterial meningitis. Pediatrics 2000; 105(2):316319.