Predictive value of cerebrospinal fluid lactate for the diagnosis of bacterial meningitis in patients post-neurosurgical operation with blood-contaminated cerebrospinal fluid

Abstract

To evaluate the diagnostic value of cerebrospinal lactate for the diagnosis of bacterial meningitis in patients post-neurosurgical operation (PNBM) with blood-contaminated cerebrospinal fluid (CSF). A prospective observational study was conducted. 101 patients underwent neurosurgical operation and clinically suspected PNBM admitted to neurosurgical intensive care unit (NSICU) of the First Affiliated Hospital of Sun Yat-sen University from October 2015 to December 2016 were enrolled. Based on red blood cell quantitative test in CSF, the patients were divided into blood-contaminated and non blood-contaminated CSF groups. According to the PNBM diagnostic criteria of 2008 Centers for Disease Control and Prevention/National Healthcare Safety Network (CDC/NHSN), all patients were divided into PNBM group and non-PNBM group. The biochemical indexes levels in CSF were compared among the groups. Receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic power of CSF lactate for PNBM in blood-contaminated patients. A total of 101 suspected PNBM patients were enrolled. In 77 blood-contaminated CSF patients, 39 patients were diagnosed as PNBM (account for 50.6%); in 24 non-blood-contaminated patients, 12 patients were diagnosed as PNBM (account for 50.0%). (1) In non-PNBM patients, the lactate level in blood-contaminated CSF was significantly higher than that of non-blood-contaminated CSF (mmol/L: 3.5±1.3 vs. 2.3±1.1, P < 0.01). In PNBM patients, there was no significant difference in lactate level between blood-contaminated CSF and non blood-contaminated CSF (mmol/L: 6.8±2.1 vs. 6.9±2.5, P > 0.05). (2) In both blood-contaminated and non blood-contaminated CSF, white blood cell (WBC), protein and lactate levels in PNBM group were significantly higher than those in non-PNBM group [WBC (×106/L): 660.0 (67.5, 1 105.0) vs. 41.0 (15.0, 142.5) in blood-contaminated CSF, 168.0 (86.5, 269.5) vs. 34.5 (7.0, 83.5) in non-blood-contaminated CSF; protein (mg/L): 4 757.8 (2 995.2, 10 219.8) vs. 1 292.8 (924.2, 1 936.2) in blood-contaminated CSF, 39 247.3 (14 900.6, 62 552.2) vs. 1 441.6 (977.3, 2 963.9) in non blood-contaminated CSF; lactate (mmol/L): 6.8±2.1 vs. 3.5±1.3 in blood-contaminated CSF, 6.9±2.5 vs. 2.3±1.1 in non blood-contaminated CSF, all P < 0.05], and glucose and CSF glucose/blood glucose ratio in PNBM group were significantly lower than those in non-PNBM group [glucose (mmol/L): 2.5±1.2 vs. 4.4±1.6 in blood-contaminated CSF, 1.9±1.4 vs. 3.4±0.9 in non blood-contaminated CSF; CSF glucose/blood glucose ratio: 0.28±0.15 vs. 0.46±0.16 in blood-contaminated CSF, 0.24±0.16 vs. 0.45±0.11 in non blood-contaminated CSF, all P < 0.01]. (3) It was shown by ROC curve analysis that CSF lactate level was a good diagnostic parameter for PNBM both in blood-contaminated and non blood-contaminated CSF, and the area under ROC curve (AUC) was 0.91 and 0.97, respectively. When the cutoff value of lactate in non blood-contaminated CSF was 3.35 mmol/L, the sensitivity was 100%, and the specificity was 91.7%. When the cutoff value of lactate in blood-contaminated CSF was 4.15 mmol/L, the sensitivity was 92.3%, and the specificity was 71.1%, and the combination of CSF lactate and glucose achieved better diagnostic specificity (AUC = 0.96, sensitivity was 97.4%, specificity was 84.2%). Blood in CSF led to the elevation of CSF lactate as compared with that in non-blood-contaminated CSF of patients with PNBM. CSF lactate was still a good diagnostic parameter for PNBM both in blood-contaminated patients, and the combination of CSF lactate and glucose achieved better diagnostic specificity.