The pyogenic central nervous system infections of children include bacterial meningitis, focal abscesses including intracranial, extracerebral, and intracerebral abscess, ventricular shunt infections or ventriculitis, focal extracranial infections of the central nervous system (e.g., spinal, epidural, subdural or paraspinal abscesses), and postoperative or traumatic wound injury. Of these central nervous system infections, only bacterial meningitis is regularly associated with sepsis or bacteremia at the time of diagnosis, with rates of positive blood cultures ranging from <5% to >25%. Despite cryptic bacteremia as a cause of shunt infections, intracranial abscesses, and spinal infections, these infections are only infrequently accompanied by positive blood cultures at the time of diagnosis, with the exception of certain pathogens (e.g., Staphylococcus aureus).
Viral infection and its diagnosis will be discussed because it represents the infection causing the most frequent diagnostic confusion. Although it is associated with cerebrospinal fluid (CSF) pleocytosis and overlaps with some clinical and neurologic symptoms, it is defined by the absence of a recovery of a bacterial organism or specific bacterial antigen.
Methods of Review
Standard pediatric textbook references and studies of the treatment and epidemiology of bacterial meningitis from the English literature, published from 1960 to 2002, were examined. Representative studies that provided a complete definition used in the studies were included in the final definitions. A standardized definition of bacterial meningitis formulated by the Infectious Disease Society of America (IDSA) was adopted and compared with those of the treatment studies and the standard definitions used by the Centers for Disease Control, Active Bacterial Core, and National Nosocomial Infection Surveillance programs.
Definitions of Bacterial Meningitis Before 1992
Before 1992, definitions of bacterial meningitis were nonstandardized and defined for use in clinical antibiotic trials of the treatment of bacterial meningitis, which were conducted after the introduction of effective antibiotics in 1950. The definitions of bacterial meningitis in these studies are represented by studies conducted between 1977 and 1990 (1–8). Bacterial meningitis often was defined as a “compatible clinical syndrome” of meningitis plus a positive CSF culture, a positive bacterial antigen test, or positive blood culture. However, because cultures of blood or CSF were negative in cases of presumed bacterial meningitis in approximately 10%–30% of cases, a concept of “purulent unknown,” “pyogenic,” or “presumed” bacterial meningitis was used in some studies (9, 10). This was often defined as meningitis with negative cultures of blood and CSF in a patient with a compatible clinical syndrome and CSF changes consistent with bacterial meningitis with or without confirmation by a nonculture method such as bacterial antigen testing. In a review of studies of bacterial meningitis conducted from 1977 to 1990, the criteria for a diagnosis of presumed bacterial or purulent unknown meningitis included a compatible clinical syndrome consistent with bacterial meningitis plus one or two of the following: an elevated CSF leukocyte count, defined as varying from 5 to >500 white blood cells/mm3; a low CSF glucose of either ≤40 or 50 mg/dL or a CSF/serum ratio of <0.5; or an elevated CSF protein varying from ≥50 to 150 mg/dL; or autopsy evidence for bacterial meningitis.
Other positive blood cultures in association with a compatible clinical syndrome and CSF changes have been used variously as criteria for definite bacterial or probable bacterial meningitis. Adherence to strict conventional diagnostic CSF changes have often been less rigorous in patients with positive blood cultures, requiring as few as a total of only 5 leukocytes/mL. The prevalence of negative CSF cultures with positive bacterial cultures in patients given the diagnosis of bacterial meningitis has varied from <5% to >25%. Many tests for detection of bacterial antigens have been used, such as latex agglutination, enzyme immunoassay or enzyme linked immunoassay, and more recently, polymerization chain reaction (PCR) (11). Only latex agglutination is widely and commercially available. Unfortunately, these tests often have been associated with a significant lack of sensitivity and specificity for pneumococcal and meningococcal infections. Other nonspecific tests to differentiate bacterial from nonbacterial disease such as viral infections, including CSF lactate, C-reactive protein, procalcitonin, and some cytokines, or specialized stains for bacteria such as ethidium bromide, have remained largely experimental and never widely incorporated into the diagnosis of bacterial meningitis.
In the absence of a confirmatory culture or other reliable diagnostic test for bacterial organisms in patients pretreated for a presumed bacterial infection, the spinal fluid profile (e.g., cell count, differential, protein, and glucose) in so-called partially treated bacterial meningitis is remarkably consistent and little changed from that of patients with positive CSF cultures who have not been pretreated with antibiotics (Table 1). Only the frequency of positive cultures and Gram-negative stain is reduced in those children with bacterial meningitis who have received antibiotics before lumbar puncture (12). On occasion, the diagnosis of bacterial meningitis can be confounded by the accidental traumatic lumbar puncture, which has been estimated to occur in 15%–20% of children. Some authors have suggested that if determination of the ratio of white blood cells/mm3 to red blood cells/mm3 in CSF exceeds the ratio in the peripheral blood, the excess white blood cells could represent inflammatory pleocytosis (13). However, prospective studies have often found that this determination is inaccurate. Therefore, in those cases in which the CSF findings do not support a bacterial infection, a positive culture or other reliable diagnostic study for microorganisms has remained the standard for a diagnosis of definite bacterial meningitis, with probable or possible categories requiring increasingly less stringent supportive criteria.
Definitions of Bacterial Meningitis Since 1990
Since 1990, published studies have usually employed a definition of bacterial meningitis that required the isolation of an organism in culture, either directly from the CSF or the blood. The Centers for Disease Control bacterial core surveillance project, which includes surveillance for all the major pathogens of community-acquired meningitis, including group B streptococci, Haemophilus influenzae, Neisseria meningitidis, and Streptococcus pneumoniae, has used a case definition of “isolation (e.g., culture positive) of… bacterial pathogens from a normally sterile site” (14). The Centers for Disease Control and National Nosocomial Infection Surveillance have defined meningococcal, haemophilus, and pneumococcal infection with surveillance definitions requiring positive cultures or positive antigen tests for meningococcal infections in the cases of purpura, with negative bacterial cultures or haemophilus antigen positivity only with infection due to Haemophilus influenzae in bacterial meningitis. Because of poor sensitivity and specificity, pneumococcal positive antigen tests have not been included as criteria in Centers for Disease Control and National Nosocomial Infection Surveillance definitions.
In 1992, IDSA formulated definitive guidelines for the eligibility criteria for clinical trials of antimicrobial agents for bacterial meningitis (15), thereby providing the first standardized definition of bacterial meningitis. Since that time, trials of antibiotics for the treatment of bacterial meningitis have often cited these criteria (Table 2) as the standard for the diagnosis of bacterial meningitis (16, 17). Thus, the currently proposed definitions for bacterial meningitis for infants, neonates, and children proposed here are extrapolated from both the historical definitions and the definitions provided by the 1992 IDSA guidelines authored by McCracken et al (Table 3) (15). Therefore, as much as possible, the definitions of bacterial meningitis provided coincide with those of the IDSA and have been reformulated to provide the definite, probable, and possible categories.
The IDSA criteria provided a diagnosis of bacterial meningitis in four scenarios, combining the results of a consistent clinical syndrome plus either a positive bacterial culture, bacterial antigen tests, and abnormalities of the CSF consistent with bacterial meningitis. The IDSA criteria assumed a diagnosis (e.g., definite) of bacterial meningitis only when a bacterial culture of the CSF or blood is positive in association with an abnormality of CSF leukocytes, glucose, or protein consistent with bacterial meningitis. If there is a negative blood or CSF culture, the diagnosis is accepted when both the CSF antigen test and CSF abnormalities are positive and a compatible clinical syndrome. Because this category is less stringent but provides a highly probable pathogenic diagnosis of bacterial infection, it fulfills the diagnostic category of probable bacterial meningitis. Because the IDSA guidelines were to fulfill the requirements for eligibility for entry into phase III U.S. Food and Drug Administration studies of bacterial infections, demonstration of the organism by culture or antigen detection was required for eligibility. Therefore, possible or probable bacterial meningitis as listed in Table 3 does not fulfill any of the IDSA criteria because detection of bacteria is not required, but this definition does match those of previous studies of the treatment and epidemiology of bacterial meningitis that utilized definitions of pyogenic or purulent unknown meningitis. Depending on the clinical setting, frequency of previous antibiotics, and quality of microbiological technical services, the possible category has been estimated to represent 5%–20% of cases.
Neonatal bacterial infections have required the isolation of the organism because neither the clinical syndrome nor the CSF changes associated with bacterial infection in older subjects are consistently present during bacterial infection of the central nervous system in neonates (Table 2) (18, 19). Although coagulase-negative staphylococci are frequent causes of sepsis in neonates, they have been infrequently associated with infections of the central nervous system, unless the infant has a device implanted in the central nervous system, such as a ventriculostomy or ventricular shunt (20). Therefore, the isolation of these organisms in infants without such devices nearly always represents a contaminant in the CSF culture and not a true pathogen.
Viral infection is the most common cause of inflammation of the central nervous system and acute febrile encephalopathy of children and are discussed briefly because of the diagnostic confusion they engender in the differential of bacterial infection. Infections of viral origin are generally classified as meningitis, meningoencephalitis, or encephalitis based on the relative clinical contributions of meningeal inflammation or involvement of the brain parenchyma. Differentiation from bacterial infection of the central nervous system is made on the basis of signs and symptoms, including clinical epidemiology and CSF changes. Clinical symptoms consistent with meningeal involvement are milder but overlap with those of bacterial infection, whereas in meningoencephalitis and encephalitis, the cerebral symptoms and signs predominate with variable associated meningeal involvement. Clinical symptoms may vary from mild to severe nonfocal generalized encephalopathy to that of only focal findings. CSF pleocytosis and changes in CSF chemistry are often distinct, generally with a mononuclear pleocytosis of 100–300 white blood cells/mm3, but can range up to a few thousand, a normal or modestly elevated protein, and a normal glucose (21). However, early in the illness, CSF white cell counts may exceed 1000/mm3, with a predominance of polymorphonuclear cells (22, 23). Therefore, the demonstration of viral antigens or nucleic acids, or viral specific antibodies, in CSF is required to definitively establish a diagnosis of viral disease of the central nervous system (24–26). When compared with viral culture, PCR for enteroviruses in CSF samples is frequently positive, compared with only about 10% for viral culture, yielding a sensitivity and specificity of 85.7% and 93.9%, respectively (26). Other studies using enteroviral PCR in young infants have demonstrated sensitivities approaching 100%, with specificities exceeding 90%. Because enteroviruses are the predominant cause of the “aseptic meningitis” syndrome, this single test is likely to exclude a viral etiology in >90% of children. Coupled with CSF, clinical, and bacteriologic findings that exclude bacterial infection (e.g., culture or bacterial antigen testing), a diagnosis of a viral etiology can be established in most children. However, the category cited above of possible bacterial meningitis is likely to significantly overlap in clinical symptoms and CSF changes with that of viral meningitis.
For causes of viral central nervous system infection, other than enteroviral infections, many other PCR tests are in development, and some, such as those for herpes viruses, are widely available but not yet approved by the U.S. Food and Drug Administration (27). However, because encephalitis has even a greater range in clinical findings and epidemiologic settings, the diagnosis will depend on the clinical syndrome and CSF changes consistent with a viral pathogenesis, and when available, a specific CSF culture, viral antibody, or PCR result that is associated with high sensitivity and specificity, such as those for herpes and West Nile meningoencephalitis (28, 29). In addition, for most cases, neither the CSF or cultures of either blood or CSF will meet the criteria for diagnosis of bacterial infection. Therefore, in those bacterial and viral syndromes with overlapping CSF findings, the diagnosis of the likely specific pathogenesis would depend on either the isolation of bacteria from an appropriate CSF or blood culture or the demonstration of a virus from CSF by either polymerase chain reaction, specific antiviral antibodies, or culture of the virus.
CSF Shunt Infections
Infection of the ventricular shunt occurs in up to 27% of children with hydrocephalus, and although shunt infections may have a “typical” clinical syndrome and CSF findings (vomiting, fever, central nervous system dysfunction, and CSF leukocytosis), these may vary by the time of onset after shunt placement, the pathogen causing infections, and site of infection in either the proximal (ventricular) or distal (atrial or peritoneal) catheter (30, 31). Therefore, the diagnosis of bacterial shunt infection cannot be established by CSF and clinical findings alone. Positive cultures acquired from the shunt fluid confirms a diagnosis of shunt infection, whereas positive cultures of CSF acquired from a lumbar site in the presence of typical signs and symptoms and CSF pleocytosis (with the presence of a shunt), may be consistent with presumptive shunt infection. Blood cultures are very infrequently positive with shunt infections but may occur in up to 42% of ventriculoatrial and 10% of ventriculoperitoneal shunt infections (30). Other sites rarely positive on culture include urine or overlying wounds of the shunt incision. Clinical symptoms and CSF changes are inconsistent (31, 32). In previous studies, fever has been present in 64%–92% of children with shunt infection, whereas other symptoms, such as shunt dysfunction, changes in sensorium, vomiting, meningismus, irritability, cellulitis over the shunt, wound inflammation, and abdominal tenderness, varied from <15% to 94% (31, 32). Because shunt infections are essentially foreign-body infections, alternating periods of quiescent and active symptoms, with an initial response and subsequent relapse in response to antibiotic therapy, are characteristic. Also, the infections are usually caused by commensal flora (e.g., staphylococci), and there is often a difficulty in recovery the organisms in fluids or culture of tissues not directly contiguous with the shunt. Therefore, only culture of the pathogen from the shunt reservoir or from fluids from the distal or proximal shunt tubing is the gold standard for diagnosis of shunt infection, whereas positive cultures from other sites, such as lumbar CSF or blood, are considered evidence for probable shunt infection if combined with clinical and CSF changes consistent with bacterial infection.
Cerebral Abscess and Other Focal Central Nervous System Infections
Other central nervous system infections such as cerebral abscess, subdural empyema, intracerebral abscess, subdural abscess, paraspinal and spinal abscess, and other central nervous system anatomic infections are usually suspected on the basis of localizing, peripheral, or central neurologic symptoms and confirmed by imaging techniques such as magnetic resonance imaging or computerized tomography (33, 34). A full review of each of these clinical entities is beyond the scope of this work but are available in standard pediatric, pediatric infectious disease, neurology, and neurosurgical textbooks. In almost all cases of focal neurologic abscesses, there is known or suspected predisposing cause. In children, the most frequent are pyogenic disease of the sinuses and chronic otitis media. Unless associated with a known hematogenous source for bacteremia (e.g., congenital heart disease with endocarditis, lung, or abdominal abscess or previous septic episode), these infections are infrequently associated with positive blood cultures. The diagnosis is confirmed with the site-specific appropriate imaging technique (e.g., computerized tomography or magnetic resonance imaging, or both), but confirmation of the bacterial pathogenesis requires demonstration of the organism in culture of spontaneously drained or surgically acquired fluids, culture of tissue, or culture of appropriate autopsy materials. Paraspinal or spinal abscesses in children are rare clinical infections and most often occur as a result of associated infections of the vertebrae, the intervertebral disk, or from trauma or infection of retroperitoneal organs or soft tissues (35).
The presence of a clinically consistent syndrome and CSF changes in association with appropriate culture findings from relevant materials is necessary to confirm a definite clinical infection of the central nervous system, including bacterial meningitis. The expanding array of laboratory techniques that can demonstrate infection of the central nervous system with specific bacterial antigens, antiviral antibodies, viral antigens, and PCR for specific bacterial or viral nucleic acids have augmented the diagnostic armamentarium, but these may not be readily available in all clinical settings. Thus, this necessitates the procurement of CSF or other fluids or tissue for bacterial culture from presumptively infected sites. In some cases, positive blood cultures may provide a definitive etiologic diagnosis of a bacterial infection. Bacterial PCR is in the process of development and, in the future, may supplant culture as the diagnostic modality of choice.
The IDSA definition of bacterial meningitis (Table 2) was based on principals that were learned in nearly three decades of clinical studies of the treatment and epidemiology of bacterial meningitis. However, they were provided to meet the requirements of enrollment into phase III U.S. Food and Drug Administration studies and fall somewhat short of the clinical definition of bacterial meningitis that may not always be supported by isolation of the bacterial agent. The definitions cited in Table 3 are an attempt to combine and expand the 1992 IDSA criteria to those that are compatible with studies of sepsis. Thereby, categories that include definite, probable, and possible criteria for definition are provided.
1. Overturf GD, Steinberg EA, Underman AE, et al: Comparative trial of carbenicillin and ampicillin for purulent meningitis. Antimicrob Agents Chemother
2. Baken JS, Bruun JN, Gaustad P, et al: Penetration of amoxicillin and potassium clavulanate into the cerebrospinal fluid of patient with inflamed meninges. Antimicrob Agents Chemother
3. Modai J, Vittecoq D, Decazes JM, et al: Penetration of aztreonam I the cerebrospinal fluid of patients with bacterial meningitis. Antimicrob Agents Chemother
4. Peltola H, Anttila M, Renkonen OV: Randomised comparison of chloramphenicol, ampicillin, cefotaxime, and ceftriaxone for childhood bacterial meningitis. Finnish Study Group. Lancet
5. Lecour H, Seara A, Miranda M, et al: Treatment of 160 cases of acute bacterial meningitis with cefotaxime. J Antimicrob Chemother
1984; 14(Suppl B):195–202
6. Schaad UB, Suter S, Gianella-Borradori A, et al: A comparison of ceftriaxone and cefuroxime for the treatment of bacterial meningitis in children. N Engl J Med
7. Overturf GD, Cable DC, Forthal DN, et al: Treatment of bacterial meningitis with ceftizoxime. Antimicrob Agents Chemother
8. Valmari P, Peltola H, Ruuskanen O, et al: Childhood bacterial meningitis: Initial symptoms and signs related to age and reasons for consulting a physician. Eur J Pediatr
9. Underman AE, Overturf GD, Leedom JM: Bacterial meningitis: 1978. Dis Mon
10. Carter PE, Barclay SM, Galloway WH: Changes in bacterial meningitis. Arch Dis Child
11. Gray LD, Fedorko DP: Laboratory diagnosis of bacterial meningitis. Clin Microbiol Rev
12. Saez-Llorens X, McCracken GH: Acute bacterial meningitis beyond the neonatal period. In:
Principles and Practices of Pediatric Infectious Diseases. Second Edition. Long SS, Pickering LK, Prober CG (Eds). New York, Churchill Livingstone, 2003, pp 264–271
13. Yogev R: Meningitis. In:
Pediatric Infectious Diseases: Principles and Practices. Second Edition. Jenson HB, Baltimore RS (Eds). New York, WB Saunders Company, 2002, pp 630–650
14. Active Bacterial Core Surveillance. August 29, 2003, Centers for Disease Control, National Center for Infectious Diseases, Division of Bacterial and Mycotic Diseases. Available at: http://www.cdc.gov/ncidod/dbmd/abcs/default.htm
. Accessed March 12, 2005
15. McCracken GH, Sande MA, Lentnek A, et al: Evaluation of new anti-infective drugs for the treatment of acute bacterial meningitis. Clin Infect Dis
1992; 15(Suppl 1):S182–S188
16. Saez-Llorens X: Quinolone treatment for pediatric bacterial meningitis: A comparative study of trovafloxacin and ceftriaxone with or without vancomycin. Pediatr Infect Dis J
17. Odio CM, Puig JR, Feris JM, et al: Prospective, randomized, investigator blinded study of the efficacy and safety of meropenem vs cefotaxime therapy in bacterial meningitis in children. Pediatr Infect Dis J
18. Unhanand M, Mustafa MM, McCracken GH, et al: Gram-negative enteric bacillary meningitis: A twenty-one year experience. J Pediatr
19. Klinger G, Chin CN, Beyene J, et al: Predicting the outcome of neonatal bacterial meningitis. Pediatrics
20. Gladstone IM, Ehrenkranz RA, Edberg SC, et al: A ten year review of neonatal sepsis and comparison with the previous fifty-year experience. Pediatr Infect Dis J
21. Sawyer MH: Enteroviral infections: Diagnosis and treatment. Pediatr Infect Dis J
22. Rorabaugh ML, Berlin LE, Heldrich F, et al: Aseptic meningitis in infants younger than 2 years of age: Acute illness and neurologic complications. Pediatrics
23. Singer JL, Maur PR, Riley JP, et al: Management of central nervous system infections during and epidemic of enteroviral aseptic meningitis. J Pediatr
24. Robinson CC, Willis M, Meagher A, et al: Impact of polymerase chain reaction results on management of pediatric patients with enteroviral meningitis. Pediatr Infect Dis J
25. Ahmed A, Brito F, Goto C, et al: Clinical utility of the polymerase chain reaction for diagnosis of enteroviral meningitis in infancy. J Pediatr
26. van Vliet KE, Glimaker M, Lebon P, et al: Multicenter evaluation of Amplicor enterovirus
PCR test with cerebrospinal fluid from patients with aseptic meningitis. J Clin Microbiol
27. Read SJ, Jeffery KJM, Bangham CRM: Aseptic meningitis and encephalitis: The role of PCR in the diagnostic laboratory. J Clin Microbiol
28. Smalling TW, Sefers SE, Li H, et al: Molecular approaches to detecting herpes simplex virus and enteroviruses in the central nervous system. J Clin Microbiol
29. Marfin AA, Gubler DJ: West Nile encephalitis: An emerging disease in the United States. Clin Infect Dis
30. Schoenbaum SC, Gardner P, Shillito J: Infections of the cerebrospinal fluid shunts: Epidemiology, clinical manifestations and therapy. J Infect Dis
31. Odio C, McCracken GH, Nelson JD: CSF shunt infections in pediatrics. Am J Dis Child
32. Ronan A, Hogg GG, Klug GL: Cerebrospinal fluid shunt infections in children. Pediatr Infect Dis J
33. Saez-Llorens X: Brain abscess in children. Semin Pediatr Infect Dis
34. Mathisen GE, Johnson JP: Brain abscess. Clin Infect Dis
35. Auletta JJ, John CC: Spinal epidural abscesses in children: A 15 year experience and review of the literature. Clin Infect Dis