Current Infectious Disease Reports

, Volume 13, Issue 4, pp 305–307

Do Delays in Performing Lumbar Puncture After Administration of Antibiotics Alter the Results of CSF Cultures?

Authors

    • Vanderbilt University Medical Center
Clinical Trial Report

DOI: 10.1007/s11908-011-0188-6

Cite this article as:
Bloch, K.C. Curr Infect Dis Rep (2011) 13: 305. doi:10.1007/s11908-011-0188-6

Michael B, Menezes BF, Cunniffe J, et al.: Effect of delayed lumbar punctures on the diagnosis of acute bacterial meningitis in adults. Emerg Med J 2010,27:433–438.

Rating: •Of importance.

Keywords Lumbar puncture · Acute bacterial meningitis ·Antibiotics · Cerebrospinal fluid cultures

Introduction: Bacterial meningitis is associated with significant morbidity and mortality despite the availability of antibiotics active against all common community-acquired pathogens. One of the major determinants of an adverse outcome is a delay in antibiotic administration, prompting the publication of national guidelines in the United States [1] and Britain [2] advocating the rapid administration of empiric antibiotics shortly after presentation. However, the need to institute treatment quickly must be balanced against the risk of performing a lumbar puncture in a patient with a contraindication to this procedure. When lumbar puncture is delayed in patients who first undergo neuroimaging, antibiotics may potentially sterilize the cerebrospinal fluid (CSF), decreasing the likelihood of making a microbiologic diagnosis.

Aims: To determine whether national guidelines on the management of bacterial meningitis are being followed, and whether delays in performing lumbar puncture after administration of antibiotics alters the results of CSF cultures.

Methods: Patients ≥15 years of age hospitalized at a district general hospital in Northwest England with a discharge diagnosis of meningitis between January 2004 and December 2006 were identified. Patients who did not have lumbar puncture performed or whose CSF had less than 5 white blood cells (WBCs)/μL were excluded. Chart and electronic information system review were performed to identify baseline clinical and microbiologic data. Contraindications to an immediate lumbar puncture were determined according to the British Infection Society meningitis guidelines [2].

Patients were stratified by Meningitest criteria as to the probability of bacterial or viral meningitis [3]. Patients with a positive CSF bacterial culture or CSF viral polymerase chain reaction (PCR) were considered to have confirmed bacterial or viral meningitis, respectively. Patients without a microbiologic diagnosis were classified as likely bacterial meningitis if the peripheral WBC was greater than 15,000 cells/μL and the CSF had all of the following characteristics: WBC greater than 1,700 cells/μL, proportion of neutrophils greater than 80%, protein greater than 2.3 g/L, and CSF/serum glucose ratio less than 0.35. Patients who had negative microbiologic evaluation and did not meet the above criteria for probable bacterial meningitis were classified as likely viral meningitis.

Results: A total of 92 patients met inclusion criteria for the study. Bacterial meningitis was diagnosed in 24 (26.1%) cases, with confirmatory cultures in 16 (66.7%) cases. Neisseria meningitidis was cultured in 75% of the confirmed cases, followed by Streptococcus pneumoniae (18.8%) and Listeria monocytogenes (6.3%). Eight (33.3%) cases lacked microbiologic confirmation but met Meningitest criteria for likely bacterial meningitis. The remaining 68 (73.9%) cases were classified as viral meningitis, with a single virologically confirmed case of herpes simplex virus type 2 meningitis. The remainder met criteria for likely viral meningitis with no pathogen identified.

Among the 84 patients for whom sufficient data existed to determine timing of antibiotic administration and lumbar puncture, no patient had CSF obtained prior to antibiotics, and only three patients (4%) had lumbar puncture performed within 2 h of antibiotic administration. CSF bacterial cultures were positive in 73% of the 11 patients who had lumbar puncture performed within 4 h of antibiotic administration compared with 11% of the 71 patients who received antibiotics more than 4 h earlier. Bacteria were isolated from CSF samples up to 8 h after initiation of antibiotics. None of the 34 patients with lumbar puncture performed more than 8 h after antibiotic dosing had a positive culture result.

Delay in lumbar puncture was attributed to performance of a head CT in 62 (67%) cases. Among patients with CT scan prior to CSF sampling, 67.5% did not met British Infection Society guidelines for neuroimaging, and none of the scans in this subgroup substantiated a radiographic contraindication to lumbar puncture.

Discussion: The results confirm that consensus guidelines for the management of suspected bacterial meningitis published in 2003 were not being followed in the ensuing 3 years in this district general hospital. Although the guidelines provide criteria as to when a lumbar puncture can be safely performed at presentation, in this study two thirds of the patients who meet these criteria had a CT scan performed regardless of the lack of an indication. Even among cases that did not have neuroimaging performed, there was a significant delay in obtaining CSF, and in no case was CSF obtained prior to antibiotics. The failure to adhere to consensus guidelines with respect to timing of lumbar puncture resulted in an inability to identify a pathogen in a third of the cases classified as likely bacterial meningitis.

Comments

Bacterial meningitis represents a true infectious disease emergency, as delays in considering the diagnosis and instituting empiric antimicrobial treatment may have fatal consequences [4]. However, balanced against the need for rapid treatment is the importance of identifying the causative organism. Microbiologic diagnosis allows the clinician to tailor antibiotic treatment to the specific pathogen based on in vitro susceptibility pattern, determines the optimal duration of treatment, and, in the case of N. meningitidis infection, alerts the public health system to administer antimicrobial prophylaxis to prevent secondary infection. Inability to culture a pathogen, particularly if attributed to sterilization of CSF by prior antibiotics, necessitates costly and prolonged empiric parenteral antibiotic therapy. Recognition of the importance of the initial management of suspected bacterial meningitis has led to the publication of national guidelines in the United States [1], Britain [2], and Europe [5]. The study by Michael et al. represents a hospital-based cross-sectional study of adherence to the British guidelines, and emphasizes the consequences when these practices are not followed.

The most commonly identified reason for a delay in lumbar puncture is the perception that neuroimaging is mandatory prior to obtaining CSF. This has been refuted in a study of patients with bacterial meningitis, in whom 97% of patients lacking a predisposition for a central nervous system mass lesion, new-onset seizures, or focal neurologic deficits had normal head CT scans [6]. Yet despite evidence-based recommendations to reserve CT scanning for the subset of cases with increased probability of a mass lesion, a survey of clinicians in England found that 78% of responders would perform imaging prior to lumbar puncture in a low-risk patient [7]. The study by Michael et al. confirms that CT scan prior to lumbar puncture, even in the absence of a documented indication, continues to be commonly performed.

The consequences of prior antibiotic therapy have been most thoroughly studied in children with bacterial meningitis, in whom antibiotics have been shown to significantly decrease the yield of CSF culture and alter both the total CSF WBC and the proportion of neutrophils [8, 9]. The rapidity of sterilization of CSF differs by organism, with CSF cultures among children with N. meningitidis uniformly negative 2 h after antibiotic administration compared with S. pneumoniae, which was consistently cultured up to 4 h after antibiotic administration [10]. In the present study, one third of cases with presumed bacterial meningitis were culture negative following antibiotic treatment, but it is likely given the stringent laboratory criteria of the Meningitest that additional cases of culture-negative bacterial meningitis were misclassified as viral. Clinically, it is often challenging to differentiate partially treated bacterial meningitis from viral or other causes of aseptic meningitest. Because the consequences of inadequately treating bacterial meningitis are so dire, empiric antibiotics are frequently continued for a full course of treatment when these two diagnoses cannot be reliably distinguished, with associated unnecessary costs and morbidity.

Delays in performing lumbar puncture are often appropriate or unavoidable, and when this is the case, the paramount consideration is rapid antibiotic treatment. Obtaining two sets of blood cultures prior to the first dose of antibiotics is crucial in this setting, because up to 74% of untreated patients with bacterial meningitis will have an associated bacteremia [10]. Additional diagnostic testing that may be useful in pretreated patients includes CSF PCR for common bacterial pathogens, although this test is rarely available outside of research settings [1].

However, as the study by Michaels et al. demonstrates, all too frequently lumbar puncture is unnecessarily delayed despite guidelines outlining initial management of suspected bacterial meningitis. Whether similar deviations are occurring on a widespread basis is not known, and further studies documenting adherence to these recommendations are warranted. In an era in which patient safety and measures of quality of care are increasingly emphasized, health care systems would benefit from the implementation of guideline-driven algorithms to balance the need for early empiric antibiotic therapy with rapid diagnostic evaluation in the management of bacterial meningitis.

Disclosure

No potential conflict of interest relevant to this article was reported.

Copyright information

© Springer Science+Business Media, LLC 2011