Prediction of bacterial meningitis based on cerebrospinal fluid pleocytosis in children.
Despite the advances in diagnosis and treatment of infectious diseases, meningitis is still considered as an important cause of mortality and morbidity, specially in the pediatric population. (1,2) Bacterial meningitis (BM) can cause serious complications and its severity depends not only on the causal microorganism, but also on host immune factors, immunization status, and geographic region. (3) The most common etiological agents are Neisseria meningitidis and Streptococcus pneumoniae, the latter being associated with a higher rate of severe and permanent sequelae, and mortality. (4,5) The implementation of vaccination programs allowed a remarkable reduction in incidence and mortality of infectious diseases. The incidence of invasive disease by Haemophilus influenzae (Hib) decreased dramatically in populations with high immunization coverage rates. (6,7) More recently, meningococcal conjugate type C and pneumococcal vaccines have also contributed to change the epidemiological profile of this disease. (8,9)
When approaching a child with meningitis it is known that an early introduction of antibiotic treatment assures rapid treatment of children with BM. However, antibiotic therapy results in systematic hospitalization and unnecessary antibiotic administration for children with aseptic or viral meningitis (VM), with the associated morbidity and economic costs. Therefore, distinguishing BM from other types of meningitis in the emergency department could help to limit unnecessary antibiotic use and hospital admissions. Because the consequences of delayed diagnosis of BM can be severe, any proposed diagnostic tool must achieve near 100% sensitivity. (10) Some criteria such as Gram staining, bacterial antigen testing of cerebrospinal fluid (CSF) as well as the classic biological markers in the blood (CRP level, white blood cell [WBC] count, and neutrophil count) or CSF (protein level, glucose level, WBC count, and neutrophil count) can be used to help predicting BM. Some scores like the BM score and the Meningitest have a high sensitivity and are proven to be valid when evaluating a child with meningitis. (11-13) More recently, some isolated factors (14,15) also proved to be good parameters to differentiate bacterial from VM. However, in some institutions, these results can be time consuming, and in some cases are impossible to be obtained. Therefore, our aim was to verify the possibility of using the CSF WBC count in an initial evaluation of BM. The objective of the present study was to establish a cutoff point of CSF WBC count that distinguished bacterial from viral and aseptic meningitis.
Children aged 29 days to 17 years, admitted to Centro Hospital Sao Joao, Oporto, Portugal, with CSF pleocytosis (considered as a WBC count [greater than or equal to] 7 [micro][L.sup.-1]), were enrolled in this retrospective study from January 1st, 2005 to December 31th, 2009. Cases of traumatic lumbar puncture (LP) and those who had received antibiotic treatment before LP were excluded.
The diagnosis of meningitis was based on history, physical examination and CSF laboratory findings. Meningitis was defined as bacterial according to identification of bacterial agents in Gram staining and/or positive bacterial culture. It was defined as viral if the reverse transcriptase polymerase chain reactions were positive, and the bacterial culture was negative. The other cases were considered as aseptic meningitis.
Statistical analysis of data was carried out using the SPSS 18 software. Differences between groups in continuous variables were tested for significance with the student's t test. Differences in frequencies of findings between groups were analyzed by Fischer's exact test. The p-value was considered significant if <0.05. The CSF markers were analyzed to determine specificity and sensitivity of each marker and combinations between them. We used the receiver operating characteristic (ROC) curve to evaluate clinical usefulness of the WBC count. The ROC curve represents the probability of true results in a disease as a function of the probability of false positive results of a test. The area under the curve represents the validity of a test with 1.00 being the highest and 0 the lowest. A classification for accuracy of a diagnostic test considers 0.90-1.00 = excellent; 0.80-0.89 = good; 0.70-0.79 = fair; 0.60-0.69 = poor; 0.50-0.59 = failure (Fig. 1).
The demographic characteristics of patients are summarized in Table 1. When excluding the cases of traumatic LP and those with previous antibiotic treatment (a total of 98) we found 295 patients with CSF pleocytosis. There was a female predominance in all types of meningitis, with 60.3% females in total. The medium age was 5.0 [+ or -] 4.3 years distributed as: 12.2% 1-3 months; 10.5% 3-12 months; 29.8% 12 months to 5 years; and 47.5% >5 years. BM was evenly distributed in all age groups, while VM was much more frequent among children >12 months. This difference of age distribution between viral and BM was significant (p <0.05). Thirty one children (10.5%) had BM, 156 (52.9%) VM and 108 (36.6%) AM. The rate of BM was 14.9% in 2005, 26.4% in 2006, 24.7% in 2007, 20.3% in 2008 and 13.6% in 2009. BM was the prevailing type of meningitis in 2007, representing 29% of all bacterial cases.
The etiology of meningitis is summarized in Table 2. BM was caused by N. meningitidis (48.4%), S. pneumoniae (32.3%), other Streptococcus species (9.7%), Staphylococcus aureus (3.2%), H. influenzae (3.2%), and Escherichia coli (3.2%). VM was caused by Enterovirus (98.1%), herpes simplex virus type 1 (1.3%), and varicella zoster virus (0.6%).
When analyzing CSF characteristics (Table 3) WBC count was significantly higher in patients with BM (mean, 4839 cells/[micro]L) as compared to patients with VM (mean, 159 cells/[micro]L, p < 0.001), with those with AM (mean, 577 cells/[micro]L, p<0.001) and with both (p<0.001). CSF protein level was also higher in BM than in VM (p < 0.01). Since in our hospital the differential counting of the cells with polymorphonuclear leukocyte count is not always performed, this was not subject of analysis.
Table 4 shows the sensitivity and specificity of different values of WBC count for BM patients. The diagnostic cutoff level of 321 WBC/[micro]L in CSF maximized was found to have optimum sensitivity (80.6%) and specificity (81.4%), with an area under the ROC curve of 0.837.
Hospitalization and treatment with broad-spectrum antibiotics in a child with CSF pleocytosis not caused by bacterial agents is frequent and constitute a source of parental stress and increased health costs. On the other hand, failure to promptly diagnose and treat BM can have devastating consequences. The ultimate confirmation of this diagnosis is CSF bacterial culture. However, physicians must make treatment decisions before culture results are available, and they depend on CSF findings to help them do so. Furthermore, a clinical prediction parameter to accurately identify patients at risk of BM is desirable. The search for simple CSF parameter predictor has been a concern of several authors. (16,17) Our present study analyzed in a retrospective way the CSF WBC count aiming at establishing a cutoff WBC value to predictive of BM. Several studies showed that the CSF profile alone could not reliably differentiate bacterial from other types of meningitis. (18-20) However, we found in our study a very large area under the ROC curve when testing WBC count at the cutoff of 321/[micro]L, as well as high sensitivity and specificity for this parameter, when comparing with similar studies, (21,22) and even when compared with other CSF parameters, like protein or glucose levels. (23,24) Only 14.7% of our patients with VM had a WBC count in LCR >321 [micro][L.sup.-1]. The fact that this parameter was statistically significant to differentiate BM from both VM and AM came as a surprise to us. Our intention was not to replace scores already studied and well documented, but to try to prove that a single simple parameter could, in an emergency setting, guide a clinical decision. Also, we do not want to downplay the importance of the clinical presentation and the physical examination for diagnosing BM.
The current knowledge showed the existence of very sensitive and specific parameters, including some well-studied scores, used to identify BM. Because these scores differentiate bacterial from nonbacterial meningitis better than a single laboratory value, the current proposal is a multivariable approach. Despite that, the CSF WBC count was also found to be a useful and rapid diagnostic test to distinguish between bacterial and nonbacterial meningitis in children. It can be useful as an initial approach or in situations or places when the time is limited or the resources are scant. A cutoff value of 321 WBC/[micro]L has highly sensitive and specific for the diagnosis of BM. As a retrospective study, its limitations are obvious. This study concerns the cases of a tertiary center, where the number of meningitis is probably higher when compared to other hospitals. Thus, it is questionable if this cutoff can be extrapolated to other settings.
Received 22 September 2012
Accepted 10 December 2012
Available online 18 April 2013
Conflict of interest
All authors declare that they have no conflict of interest.
(1.) Kim KS. Acute bacterial meningitis in infants and children. Lancet Infect Dis. 2010;10:32-42.
(2.) Levy C, de La Rocque F, Cohen R. Epidemiology of pediatric bacterial meningitis in France. Med Mal Infect. 2009;39: 419-31.
(3.) Agrawal S, Nadel S. Acute bacterial meningitis in infants and children: epidemiology and management. Paediatr Drugs. 2011;13:385-400.
(4.) Chandran A, Herbert H, Misurski D, Santosham M. Long-term sequelae of childhood bacterial meningitis: an under appreciated problem. Pediatr Infect Dis J. 2011;30: 3-6.
(5.) de Jonge RC, van Furth AM, Wassenaar M, Gemke RJ, Terwee CB. Predicting sequelae and death after bacterial meningitis in childhood: a systematic review of prognostic studies. BMC Infect Dis. 2010;10:232.
(6.) Giufre M, Cardines R, Caporali MG, Accogli M, D'Ancona F, Cerquetti M. Ten years of Hib vaccination in Italy: prevalence of non-encapsulated Haemophilus influenzae among invasive isolates and the possible impact on antibiotic resistance. Vaccine. 2011;29:3857-62.
(7.) Kalies H, Grote V, Siedler A, Grondahl B, Schmitt HJ, von Kries R. Effectiveness of hexavalent vaccines against invasive Haemophilus influenzae type b disease: Germany's experience after 5 years of licensure. Vaccine. 2008;26:2545-52.
(8.) Riordan A. The implications of vaccines for prevention of bacterial meningitis. Curr Opin Neurol. 2010:319-24.
(9.) Tsai CJ, Griffin MR, Nuorti JP, Grijalva CG. Changing epidemiology of pneumococcal meningitis after the introduction of pneumococcal conjugate vaccine in the United States. Clin Infect Dis. 2008;46:1664-72.
(10.) Dubos F, Korczowski B, Aygun DA, et al. Serum procalcitonin level and other biological markers to distinguish between bacterial and aseptic meningitis in children: a European multicenter case cohort study. Arch Pediatr Adolesc Med. 2008;162:1157-63.
(11.) Dubos F, Korczowski B, Aygun DA, et al. Distinguishing between bacterial and aseptic meningitis in children: European comparison of two clinical decision rules. Arch Dis Child. 2010;95:963-7.
(12.) Nigrovic LE, Kuppermann N, Macias CG, et al. Clinical prediction rule for identifying children with cerebrospinal fluid pleocytosis at very low risk of bacterial meningitis. JAMA. 2007;297:52-60.
(13.) Zimmerli W. How to differentiate bacterial from viral meningitis. Intensive Care Med. 2005;31:1608-10.
(14.) Alkholi UM, Abd Al-Monem N, Abd El-Azim AA, Sultan MH. Serum procalcitonin in viral and bacterial meningitis. J Glob Infect Dis. 2011;3:14-8.
(15.) 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.
(16.) Negrini B, Kelleher KJ, Wald ER. Cerebrospinal fluid findings in aseptic versus bacterial meningitis. Pediatrics. 2000;105:316-9.
(17.) Kanegaye JT, Nigrovic LE, Malley R, et al. Diagnostic value of immature neutrophils (bands) in the cerebrospinal fluid of children with cerebrospinal fluid pleocytosis. Pediatrics. 2009;123:967-71.
(18.) Lindquist L, Linne T, Hansson LO, Kalin M, Axelsson G. Value of cerebrospinal fluid analysis in the differential diagnosis of meningitis: a study in 710 patients with suspected central nervous system infection. Eur J Clin Microbiol Infect Dis. 1988;7:374-80.
(19.) Graham AK, Murdoch DR. Association between cerebrospinal fluid pleocytosis and enteroviral meningitis. J Clin Microbiol. 2005;43:1491.
(20.) Levy ML, Wong E, Fried D. Analysis of 650 lumbar punctures. Clin Pediatr. 1990;29:254-61.
(21.) Ray P, Badarou-Acossi G, Viallon A, et al. Accuracy of the cerebrospinal fluid results to differentiate bacterial from non bacterial meningitis, in case of negative gram-stained smear. Am J Emerg Med. 2007;25:179-84.
(22.) Carbonnelle E. Laboratory diagnosis of bacterial meningitis: usefulness of various tests for the determination of the etiological agent. Med Mal Infect. 2009;39:581-605.
(23.) Lussiana C, Loa Clemente SV, Pulido Tarquino IA, Paulo I. Predictors of bacterial meningitis in resource-limited contexts: an Angolan case. PLoS One. 2011;6:25706.
(24.) Celik N. Differential diagnosis of bacterial and viral meningitis in childhood acute meningitis: a statistical model. Mikrobiyol Bul. 2007;41:63-9.
Sofia Agueda *, Teresa Campos, Ana Maia
Department of Pediatrics, Centro Hospitalar Sao Joao, EPE, Porto, Portugal
* Corresponding author at: Hospital Sao Joao, Department of Pediatrics (Servico de Pediatria), Alameda Professor Hernani Monteiro, 4200-319, Porto, Portugal.
E-mail address: email@example.com (S. Agueda).
Table 1--Demographic features of all groups. Viral meningitis Bacterial meningitis Number of patients 156(52.9%) 31 (10.5%) Gender Male 60 12 Female 96 19 Age Medium age [+ or -] SD 5.4 [+ or -] 3.9 3.6 [+ or -] 5.0 (years) <3 months 13 5 [greater than or equal to] 8 9 3 months and <12 months [greater than or equal to] 48 10 12 months and <5 years [greater than or equal to] 87 7 5 years Year 2005 15 8 2006 59 7 2007 32 9 2008 24 5 2009 26 2 Aseptic CSF meningitis pleocytosis (total) Number of patients 108(36.6%) 295 Gender Male 45 117 (39.7%) Female 63 178 (60.3%) Age Medium age [+ or -] SD 4.9 [+ or -] 4.9 5.0 [+ or -] 4.3 (years) <3 months 18 36 (12.2%) [greater than or equal to] 14 31 (10.5%) 3 months and <12 months [greater than or equal to] 13 71 (29.8%) 12 months and <5 years [greater than or equal to] 46 140 (47.5%) 5 years Year 2005 21 44 (14.9%) 2006 12 78 (26.4%) 2007 32 73 (24.7%) 2008 31 60 (20.3%) 2009 12 40 (13.6%) Table 2--Agents involved in bacterial and viral meningitis. Bacterial meningitis Viral meningitis Agents Neisseria meningitidis Enterovirus (98.1%), identified (48.4%), Streptococcus herpes simplex type 1 pneumoniae (32.3%), virus (1.3%), varicella other Streptococcus zoster virus (0.6%) species (9.7%) and other agents (9.7%) Table 3--Laboratory findings in all groups. Bacterial meningitis Viral meningitis CSF WBC count 4839 [+ or -] 5235.7 159 [+ or -] 246.8 (cells/[micro]L) CSF protein (mg/dL) 2.1 [+ or -] 1.9 0.7 [+ or -] 1.0 Aseptic meningitis p -Value CSF WBC count 577 [+ or -] 1690.2 <0.001 (cells/[micro]L) CSF protein (mg/dL) 1.3 [+ or -] 2.1 <0.01 Table 4--Sensitivity and specificity (%) of white blood cell count values for bacterial meningitis patients. WBC count values 69-248 255-321 324-875 (cells/[micro]L) Sensibility 83.9 80.6 77.4 Specificity 50.4-75.8 76.5-81.4 81.4-93.9 WBC count values 905-940 980-1096 (cells/[micro]L) Sensibility 71 66.7 Specificity 93.9-94.3 94.3-96.2
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Original article|
|Author:||Agueda, Sofia; Campos, Teresa; Maia, Ana|
|Publication:||The Brazilian Journal of Infectious Diseases|
|Date:||Jul 1, 2013|
|Previous Article:||Chloroquine sensitizes biofilms of Candida albicans to antifungal azoles.|
|Next Article:||Gestational toxoplasmosis in Parana State, Brazil: prevalence of IgG antibodies and associated risk factors.|