Most Cerebrospinal Fluid Cultures in Children With Bacterial Meningitis Are Positive Within Two Days

Infants and children with suspected bacterial meningitis receive empiric antibiotic therapy for at least 3 days, while attempts are made to establish a definitive diagnosis.1 This "3-day rule" derives from the microbiologic convention that cerebrospinal fluid (CSF) culture plates can be discarded if there is no growth at 72 h.2 Similar conventions regarding the management of children hospitalized for suspected bacteremia have been challenged by data suggesting that blood cultures yield pathogens within 48 h.3-5 Modern techniques may permit earlier identification of most CSF pathogens, affording opportunities for early discontinuation of antibiotics and hospital discharge. Methods. All children discharged from Kosair Children's Hospital with diagnoses of central nervous system infection between January 1, 1992, and September 30, 1997, were identified. The following were excluded from analysis: patients in the neonatal intensive care unit; patients with indwelling intracranial devices; posttraumatic or postsurgical patients; and patients with congenital or acquired immunodeficiencies. Clinical and microbiologic data were abstracted from medical records onto standardized forms. Microbiology laboratory records were available for patients with CSF cultures obtained after January 1, 1995, and laboratory records were requested from referral hospitals when appropriate. CSF cultures were performed by standard techniques that did not change during the study period. Specimens were first Gram-stained (centrifuged sediment was used if >1 ml was received), then streaked onto blood and chocolate plates and inoculated into a thioglycolate broth tube. Positive Gram stains were immediately reported to the ward. Plates were incubated in 5% carbon dioxide at 37°C and capped broth tubes were incubated in room air at 37°C. Cultures were inspected for growth between 7 and 10 a.m. every day (specimens received between 10 p.m. and 10 a.m. were first inspected on the next day). As soon as colonies were observed on the plates they were Gram-stained, and the results were called to the ward and entered into the hospital computer system. The time to growth was calculated as the difference between the date of inoculation and the date a positive result was recorded in the patient's chart. (It was recognized that this method would overestimate the time to positive culture, but microbiology laboratory records were not available for the majority of patients). Results. Of 369 patients identified with central nervous system infection, 40 met exclusion criteria, 3 had incomplete records and 221 had negative CSF cultures. Of the remaining 105, 99 had culture-positive bacterial meningitis. According to the medical record 62 cultures were positive within 1 day and 96 cultures within 2 days (Table 1). The most common pathogens were Neisseria meningitidis (n = 34), Streptococcus pneumoniae (n = 33), group B Streptococcus (GBS; n = 18), Escherichia coli (n = 6) and Haemophilus influenzae type b (n = 4; no cases identified after 1992).TABLE 1: Culture-positive cases of bacterial meningitis No systematic attempt was made to examine all CSF culture results obtained by the microbiology laboratory during the study period. However, 6 infants with contaminated cultures were identified in this review because they were given discharge diagnoses of meningitis. In 4 of these cases the organism was isolated after the second day. These cases are outlined below: (1) coagulase-negative Staphylococcus and Enterococcus grew after 3 days from the CSF of a 2-week-old with group B streptococcal bacteremia. CSF polymorphonuclear cell (PMN) count was 35 000 cells/mm3 and the infant was treated for GBS meningitis even though the culture grew contaminants. The case was not included as a culture-positive case of meningitis because the CSF did not grow GBS; (2) an Acinetobacter species was identified after 3 days in the CSF of a 3-month-old with 11 PMN cells/mm3. The child was subsequently proved to have enteroviral meningitis; (3) alpha-hemolytic Streptococcus was isolated after 3 days from the culture of a 1-month-old whose CSF had 323 PMN cells/mm3 but no organisms on Gram stain; (4) coagulase negative Staphylococcus grew after 5 days from the CSF of a 16-year-old treated with antibiotics before lumbar puncture. Gram stain revealed no bacteria and PMN count in the CSF was 8 cells/mm3. Microbiologic records were available for 35 of 99 patients with culture-positive bacterial meningitis. The days to positive culture noted in the medical records of 28 cases was in agreement with the microbiology laboratory record. In 7 cases the recording of the positive culture in the medical record was delayed by 1 day. Three patients had CSF cultures that grew after the second day. An 18-month-old female child admitted with fever and an abnormal neurologic examination had a CSF PMN count of 7 cells/mm3, and no organisms were seen on Gram-stained smear. Both medical and microbiology laboratory records confirmed that the CSF grew S. pneumoniae 3 days after inoculation. Because S. pneumoniae was isolated from blood cultures within 24 h, antibiotics would not have been stopped despite the negative CSF culture on the second day. A 2-month-old male infant was admitted for suspected sepsis and meningitis. CSF PMN count was 95 cells/mm3. CSF glucose concentration was 3 mg/dl and protein concentration was 211 mg/dl. Gram-positive cocci were seen in the CSF, and although CSF cultures were not positive until the third day, GBS grew from the blood within 24 h. Antibiotics would not have been stopped based on negative CSF cultures because of documented bacteremia and a positive CSF Gram-stained smear. A 2-week-old male infant had watery diarrhea, fever of 101.2°F, CSF with 2 mononuclear cells/mm3 and normal glucose and protein values. According to the medical record Escherichia coli grew from the CSF broth culture on the third day. Repeat CSF culture and Gram-stained smear obtained while receiving antibiotics were negative. Cell count, chemistries and viral diagnostics were not performed on this specimen. This infant probably did not have bacterial meningitis. However, because he was presumptively treated, the isolate was not dismissed as a contaminant for purposes of analysis. Discussion. A minority of children admitted to the hospital with suspected meningitis are proved to have bacterial disease. Nevertheless empiric antibiotics are usually continued in patients with CSF pleocytosis awaiting confirmation of negative cultures. Empiric therapy carries both financial costs and iatrogenic risks. In one study febrile infants subsequently found to have aseptic meningitis were hospitalized for a mean of 6 days, and nearly one-fourth experienced a complication related to therapy.6 Discontinuation of antibiotics should occur when bacterial disease is reliably excluded (and not highly suspected based on CSF findings), either by negative cultures (in the absence of antibiotic pretreatment) or by definitive establishment of an alternative diagnosis. The current study shows that in 97% of cases, pathogens are recovered from CSF within 2 days. Of the three children with cultures positive after the second day, 2 had already been identified as having bacterial disease on the basis of Gram stain results or positive blood cultures. Therefore antibiotics would not have been stopped based on a negative CSF culture on the second day. The third infant probably had a contaminated CSF culture. There are several limitations of this study. Cases inappropriately coded at discharge could have been missed. In addition medical records were used to determine the interval to positive culture. Although this method is not as accurate as microbiology laboratory records, it probably resulted in overestimation rather than underestimation. The objective of this study was to determine the time interval required to isolate a pathogen from the CSF of children with culture-proved bacterial meningitis. Not all children with bacterial meningitis have positive CSF cultures, however. These results cannot be extrapolated to the child clinically suspected to have bacterial meningitis despite a negative CSF culture, a situation that might be seen with antibiotic pretreatment. However, in routine cases failure to recover a pathogen from CSF within 2 days makes it very unlikely that bacterial meningitis will be definitively diagnosed. Acknowledgment. K.B. is supported by the Kosair Charities Fellowship in Pediatric Infectious Diseases. Kristina Bryant, M.D. Gary S. Marshall, M.D.