CSF lactate alone is not a reliable indicator of bacterial ventriculitis in patients with ventriculostomies

Introduction In a febrile patient with a ventriculostomy, diagnosing or excluding bacterial or microbial ventriculitis is difficult, as conventional markers in analysis of cerebrospinal fluid (CSF) are not applicable due to presence of blood and inflammation. CSF lactate has been shown to be a useful indicator of bacterial meningitis in CSF obtained via lumbar puncture, but little and heterogenous data exist on patients with ventriculostomies. Methods We reviewed all CSF analyses obtained via ventriculostomy in patients admitted to our tertiary medical center between 2008 and 2013, and constructed receiver operating characteristic (ROC) curves to evaluate the accuracy of CSF lactate concentration in discriminating a positive CSF culture from a negative one in setting of ventriculostomy and prophylactic antibiosis. Results Among 467 CSF lactate values, there were 22 corresponding CSF cultures with bacterial growth. Sensitivities and specificities for CSF lactate at threshold values 3, 4, 5 and 6 mmol/L showed sensitivity and specificity greater than 70% for CSF lactate threshold 4 mmol/L. The lowest threshold value of 3 mmol/L resulted in higher sensitivity of 81.8%, and the highest chosen threshold value resulted in high specificity of 94.2%, but these values had poor corresponding specificity and sensitivity, respectively. The area under the curve was 0.82 (95% CI 0.72, 0.91). Conclusion Our data from a large sample of CSF studies in patients with ventriculostomy indicate that no single value of CSF lactate provided both sensitivity and specificity high enough to be regarded as reliable test.