The utility of cerebrospinal fluid parameters in the early microbiological assessment of meningitis

Abstract

The measurement of cerebrospinal fluid (CSF) protein, glucose, and white cell count (WCC) is an essential part of the initial examination of CSF. The aim of this study was to determine the utility of CSF parameters in assessing the likely aetiological agent. A total of 2290 CSF samples from a 13-year period were retrospectively reviewed. The initial parameters were compared between bacterial, viral, and cryptococcal meningitis and cases where no pathogen was found. A protein concentration of <600 mg/L and a WCC <90 × 10⁶/L were found to be optimal cut-offs for excluding bacterial meningitis. A WCC of <25 × 10⁶/L was found to be optimal for excluding cryptococcal meningitis and a WCC of <10 × 10⁶ for excluding viral meningitis. Decreased glucose concentration was found to be a poor indicator of the aetiological agent.

Introduction

A rapid and accurate diagnosis of bacterial meningitis is imperative due to the serious nature, aggressive course, and high morbidity and mortality of the condition (Hoffman and Weber, 2009). Nonbacterial meningitis is responsible for 82–94% cases of meningitis (Nigrovic et al., 2002), and clinical differentiation is challenging. Accurate and rapid diagnosis would prevent patients with nonbacterial meningitis from unnecessary investigations, hospitalization, and use of antibiotics (Parasuraman et al., 2001, Raymond, 2000, Swingler et al., 1994). Currently, the gold standard for the diagnosis of bacterial meningitis is cerebrospinal fluid (CSF) culture. This is diagnostic in 70–85% of cases if prior antibiotics are not given (Bohr et al., 1984). Frequently, antibiotics are given before CSF is collected, reducing the diagnostic value of culture. Routine culture takes approximately 2–5 days, further limiting its clinical usefulness in the acute setting. In the last decade, nucleic acid amplification tests have been utilised clinically for the diagnosis of a variety of common causes of meningitis, but due to the technical expertise required for these tests they are not frequently available in the initial evaluation of suspected meningitis.

The decision to commence antibiotics is often based on clinical symptoms as well as on the preliminary CSF results, which are readily available within an hour. Bacterial meningitis scores utilising these parameters have been developed in children to distinguish between aseptic and bacterial meningitis (Bonsu and Harper, 2004).

CSF protein is composed of albumin, immunoglobulin, and transferrin, along with other enzymes and globulins (Profeta, 1996). An increased CSF protein level is an index of neurological disease, reflecting a pathological increase in permeability of endothelial cells (Lindquist et al., 1988). In adults, the normal CSF protein levels range from 230 to 380 mg/L (Scheld et al., 1997). Bacterial, cryptococcal, and tuberculous meningitis have been shown to reliably increase CSF protein levels in individual studies (Negrini et al., 2000, Sabetta and Andriole, 1985, Stockstill and Kauffman, 1983). Other processes also cause an increase in CSF protein including viral encephalitis (Koskiniemi et al., 1984), subarachnoid haemorrhage, central nervous system (CNS) vasculitis, CNS neoplasm (Jerrard et al., 2001), Guillain–Barré syndrome (Segurado et al., 1986), and increasing age (Eeg-Olofsson et al., 1981, Statz and Felgenhauer, 1983). CSF protein has been previously assessed as a diagnostic parameter, and a concentration of greater than 1500 mg/L is specific (99%) but not sensitive (55%) for culture-proven bacterial meningitis (Lindquist et al., 1988). Additionally, a CSF protein level of greater than 500 mg/L has been assessed as a cut-off in a paediatric population (Dubos et al., 2006).

Other parameters routinely examined in a CSF sample include the Gram's stain, glucose, white blood cell (WCC), and red blood cell (RBC) counts.

Gram's stain of the CSF has a high specificity for bacterial meningitis; however, its low sensitivity limits it usefulness (Feigin et al., 1992). In patients with clinical symptoms of bacterial meningitis, a diagnostic CSF Gram's stain allows empirical antibacterial treatment to start (Tunkel et al., 2004); however, in 30–40% of patients with bacterial meningitis, the Gram's stain will be unhelpful (Spanos et al., 1989).

Classically, in bacterial meningitis, CSF glucose concentration is low and CSF WCC is high particularly when compared to viral meningitis. However, in recent years, these parameters have been shown not to be reliably predictive (Dubos et al., 2006, Ray et al., 2007).

Blood can be introduced to the CSF sample because of a traumatic tap or intracerebral event and elevate the RBC. The presence of an elevated RBC complicates the interpretation of other CSF parameters when assessing the likelihood of meningitis because it is known to elevate the protein concentration (Mehl, 1986), but the magnitude of this effect is only known in the paediatric population (Nigrovicv et al., 2011). This is of particular relevance because up to 30% are traumatic taps (Nigrovic et al., 2007).

Tuberculosis (TB) and cryptococcal infection further complicate the diagnosis of meningitis, as clinical symptoms can be similar to that of bacterial meningitis. The diagnosis of TB meningitis relies on CSF culture as the gold standard. Additional factors that have been found to be predictive of TB meningitis include a CSF protein content of >1000 mg/L (Youssef et al., 2006). The diagnosis of cryptococcal meningitis is based on the India ink stain, cryptococcal antigen test, and culture (Berlin and Pincus, 1989). The CSF white cell count, differential, protein and glucose concentration are indistinguishable from that of TB meningitis (Bicanic and Harrison, 2004, Satishchandra et al., 2007).

The aim of this study was to determine whether the WCC, protein, and glucose concentrations in CSF samples obtained over a 13-year period from a tertiary-level teaching hospital in North Queensland, Australia could be used to predict the final microbiological diagnosis as defined by modern definitive methods for detecting bacterial, viral, cryptococcal, and tuberculous meningitis.

In the current era of rational pathology testing and evidence-based clinical practice, this study may provide further proof that CSF WCC, protein, and glucose concentration can be used to aid in treatment decisions and direct further testing of CSF in cases of suspected meningitis.