| Objective To develop and validate simple multivariable models to distinguish viral meningitis from other common infectious meningitis in adults using objective parameters on the basis of clinical, laboratory features and radiographic image.Method Retrospective cohort study of all adults with meningitis admitted to Xiangya Hospital, Second Xiangya Hospital, Hunan People's Hospital and Changsha Central Hospital from January 2004 through December 2010, who satisfied diagnostic criteria for viral, bacterial, tuberculosis or cryptococcal meningitis. We compared the clinical, laboratory and photography features of the adults of viral meningitis with bacterial meningitis, tuberculosis meningitis, cryptococcal meningitis and non-viral meningitis separately. Features independently predictive of viral meningitis among the different meningitis were modelled by multivariate logistic regression to create distinct diagnostic rules. The performance of each diagnostic aids was assessed by resubstitution and test data methods.Results Three modes were developed to distinguish viral meningitis from tuberculosis meningitis, bacterial meningitis and non-viral meningitis. A differential diagnostic rule of viral meningitis and tuberculosis meningitis was 87.4% sensitive and 94.3% specific of a score of 7 by resubstitution. Meanwhile,96.4% sensitive and 79.2% specific of a score of 8 was developed in the differential diagnostic mode of viral meningitis and bacterial meningitis.90.2% sensitive and 84.4% specific of a score of 10 was developed in the differential diagnostic mode of viral meningitis and non-viral meningitis.Conclusion This study suggests that simple clinical and laboratory data can help in the diagnosis of adults with viral meningitis. Without further validation in other patients, the practical value of the modes need further elevation. The models developed by this research may allow physicians to differentiate between viral and other infectious meningitis in feature. |
PDF Full Text Request