Epidemiological Investigation Of Hospital-acquired Meningitis And Establish Mathematical Models For Predicting Meningitis

Part One A retrospective analysis of postoperative intracranial neurosurgery bacterial meningitisObjective Postoperative intracranial neurosurgery bacterial meningitis is a common complication. This study examined the incidence rate of bacterial meningitis in patients who underwent intracranial neurosurgery, and also investigated the risk factors. Distribution of pathogens also was evaluated. For the control of postoperative intracranial neurosurgery bacterial meningitis.Methods Sampling the patients who underwent one intracranial neurosurgery in Huashan hospital, Fudan University in shanghai,2008.Results The study included1165patients. A total of79bacterial meningitis patients were found, including8pathogens. Incidence of postoperative intracranial neurosurgery bacterial meningitis rate was6.78%. The most frequent etiological agents are A. baumanii. Logistic regression analysis identified male sex, implant, parenteral nutrition use, external ventricular drainage and negative pressure drainage as risk factors for bacterial meningitis (P<.05).Conclusion Postoperative intracranial neurosurgery bacterial meningitis is a common and serious complication. A number of invasive operation increased the risk of postoperative bacterial meningitis. Surgery broke the normal physiological environment and cerebrospinal fluid showed atypical. The low rate of pathogen detection brings the diagnosis and treatment difficult. This report highlights those areas requiring attention to prevent bacterial meningitis. Part Two Cerebrospinal fluid levels of cytokines in patients with meningitisObjective Cytokines participate in inflammation immunity reaction in person. The origins of cytokines of cerebrospinal fluid and serum are different and are two relatively independent systems. To analyze the significance of27kinds of cytokines level in patients with meningitis or central nervous system tumor.Methods The levels of cytokines were determined by multiplex fluorescent immunoassay.Results The levels of IL-1β, IL-4, IL-8, IL-9, IL-10, IL-12(p70), IL-17, Eotaxin, INF-γ, IP-10, PDGF, MIP-1β, RANTS, TNF-a and VEGF of CSF in six group patients had a significant difference by one-way analysis of variance. The concentration of IL-4of CSF in patients with aseptic meningitis was higher than those of control group. The concentration of IL-4, IL-8, IP-10, Eotaxin and MIP-1β of CSF in patients with bacterial meningitis was higher than those of control group. The concentration of IL-4, IFN-γ, IP-10and TNF-α of CSF in patients with tuberculous meningitis was higher than those of control group. The concentration of IL-1β and IL-17of CSF in patients with bacterial meningitis was higher than those of tuberculous meningitis. The concentration of IL-1β,IL-8, IL-9, IL-12(p70), PDGF and MIP-1β of CSF in patients with bacterial meningitis was higher than those of fungal meningitis. The concentration of IP-10and RANTES of CSF in patients with tuberculous meningitis was higher than those of fungal meningitis.Conclusion The detection of the levels of cytokines in CSF probably plays a role in diagnosing central nervous system infection. Part Three Establish mathematical models for predicting meningitisObjective To establish mathematical models for predicting meningitis by using conventional laboratory indicators and multiple cytokines, and to evaluate its clinical value of predicting different kinds of meningitis. To search new pathway for diagnoise of meningitis by validating the mathematical models for exploring research.Methods Collecting cerebrospinal fluid of124patients admitted to Huashan hospital. The levels of cytokines were measured by multiplex fluorescent immunoassay. Screening for diagnostic value of the index according to the single factor and Spearman correlation analysis. Using logistic regression to analyse evidently relavant indicators. The models predicting for different kinds of meningitis were analyzed by receiver operating characteristic curve(ROC). Statistical analysis was performed by State software version7.0.Results We establish4models according to different kinds of meningitis. Logistic regression analysis showed that age and CSF glucose/serum glucose was indepent predictor of model1. CSF protein, IL-lra, IL-4and IL-17were indepent predictors of model2. IFN-γ was indepent predictor of model3. PDGF-bb and TNF-α were indepent predictors of model4. We finally built the predicting models and got index. ROC curve analysis revealed that the area under the curve (AUC) was0.794in model1,0.759in model2,0.736in model3,0.822in model4. A cutoff point of0.106had76.92%sensitivity,63.06%specificity,20.00%positive predictive value and95.95%negative predictive value in model1. A cutoff point of0.996had76.19%sensitivity,66.29%specificity,34.78%positive predictive value and92.19%negative predictive value in model2. A cutoff point of0.189had71.88%sensitivity,63.04%specificity,40.74%positive predictive value and85.71%negative predictive value in model3. A cutoff point of0.645had73.96%sensitivity,78.75%specificity,82.88%positive predictive value and69.23%negative predictive value in model4.Conclusion The mathematical models using conventional laboratory indicators and multiple cytokines have value for predicting meningitis.