IntroductionMaking a differential diagnosis between bacterial meningitis and aseptic meningitis is a critical clinical problem. The utility of a cerebrospinal fluid (CSF) lactate assay for this purpose has been debated and is not yet routinely clinically performed. To adequately evaluate this assay, a systematic review and meta-analysis of studies of the CSF lactate concentration as a marker for both bacterial meningitis and aseptic meningitis was performed.MethodsElectronic searches in PubMed, Scopus, the MEDION database and the Cochrane Library were conducted to identify relevant articles published before March 2009. A manual search of reference lists from selected articles was also conducted. Two reviewers independently selected relevant articles and extracted data on study characteristics, quality and accuracy.ResultsTwenty-five articles were identified that met the eligibility criteria. Diagnostic odds ratios were considerably homogenous (Chi-square P = 0.1009, I2 = 27.6%), and the homogeneity was further confirmed by a Galbraith plot and meta-regression analysis using several covariates. The symmetrical summary receiver-operator characteristic curve (SROC), fitted using the Moses-Shapiro-Littenberg method, was positioned near the upper left corner of the SROC curve. The Q value and area under the curve were 0.9451 and 0.9840, respectively, indicating excellent accuracy. The diagnostic accuracy of the CSF lactate concentration was higher than those of other four conventional markers (CSF glucose, CSF/plasma glucose quotient, CSF protein, and CSF total number of leukocytes) using a head to head meta-analysis of the 25 included studies.ConclusionsTo distinguish bacterial meningitis from aseptic meningitis, CSF lactate is a good single indicator and a better marker compared to other conventional markers.