Statistics from Altmetric.com
Editor,—Nitric oxide (NO) is implicated in the pathogenesis of bacterial sepsis.1 The potential sources of NO include endothelial, smooth muscle, and inflammatory cells. Serum concentrations of nitrogen oxides are increased in patients with bacterial sepsis, compared with non-septic controls.2 3NO has also been shown to be a neurotransmitter, and has been implicated in acute and chronic brain pathology.1 We hypothesised that NO production is increased in bacterial meningitis; that this would be reflected by an increased concentration of nitrogen oxides in the cerebrospinal fluid (CSF); and that the CSF concentrations of nitrogen oxides would discriminate bacterial meningitis from other causes of fever and childhood encephalopathies. Reagent strips have recently been used for the rapid diagnosis of meningitis.4 We thought that if CSF nitrogen oxides identified children with bacterial meningitis it may be a further step towards improving the current diagnostic accuracy of reagent strips.
Children undergoing lumbar puncture as part of routine investigation were studied prospectively. The study was approved by the institutional ethics committee. A 1 ml aliquot of CSF was centrifuged at 10 000 revolutions per minute for 15 minutes at 2ºC, and the supernatant aspirated and stored at −20ºC. CSF nitrate was converted to nitrite by incubation of 300 μl of CSF with nitrate reductase and NADPH for 100 minutes at 37ºC. The reaction was terminated by addition of zinc sulphate (1.5% w/v, final concentration) to precipitate protein. The sample was centrifuged at 2000 g for five minutes at 4ºC and the nitrite was determined in the supernatant. Nitrite levels were measured using the reaction of the Griess reagent with NO2 −forming a chromophore. An aliquot of 100 μl of sample was added to 100 μl of freshly prepared Griess reagent in a microtitre plate. After a five minute period to allow colour development, the absorbance was determined in a Behring plate reader at 570 nm. The concentration of NO2 − was quantified by comparison with a standard curve constructed using known concentrations of NO2 − (0.1-100 μmolar).
Other data recorded were: CSF total and differential white cell count; CSF protein and glucose concentration; viral and bacterial culture results; and the final diagnosis, based on CSF results, other investigations, and clinical findings. The CSF nitrogen oxide assay was performed by one investigator (AS) blinded to the clinical and laboratory details.
Forty six children were studied. The number of subjects (in parentheses) and the median values of nitrogen oxides according to diagnostic group were: bacterial meningitis (12), 10.5 μmol/l; viral meningitis (7), 5.8 μmol/l; encephalopathy of unknown aetiology (7), 6.9 μmol/l; fever with no central nervous system infection (19), 6.6 μmol/l (fig 1). There were no differences in CSF concentrations of nitrogen oxides between the four groups (Kruskal-Wallis test, p = 0.50). One child with encephalitis due to mycoplasma, who did not fit clearly into any of the diagnostic groups, had a CSF nitrogen oxide concentration of 482 μmol/l; the assay was repeated with the same result. Overall there were no relationships between CSF nitrogen oxide concentrations and CSF white cell count (p=0.60). Such a relationship would not be surprising if the major source of nitrogen oxides in acute central nervous system disease were inflammatory cells. We conclude that measuring CSF nitrogen oxides will not reliably distinguish bacterial meningitis from other causes of fever requiring lumbar puncture. Use of the nitrite patch on reagent strips is unlikely to add to the diagnostic yield that can be achieved by testing for protein, glucose, and leucocytes.4