Early diagnosis model for meningitis supports public health decision making

• Published in: The Journal of infection 2011-07, Vol.63 (1), p.32-38
• Main Authors: Close, Rebecca M, Ejidokun, Oluwatoyin O, Verlander, Neville Q, Fraser, Graham, Meltzer, Margie, Rehman, Yasmin, Muir, Peter, Ninis, Nelly, Stuart, James M
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Aseptic meningitis; Biological and medical sciences; C-reactive protein; C-Reactive Protein - analysis; Cerebrospinal fluid; Chemoprevention; Child; Child, Preschool; Decision Making; Diagnosis, Differential; Differential diagnosis; Early Diagnosis; England; Exanthema; Female; General aspects; Hospitals; Humans; Infant; Infant, Newborn; Infectious Disease; Logistic Models; Male; Medical sciences; Meningitis; Meningitis - blood; Meningitis - microbiology; Meningitis - prevention & control; Meningococcal infection; Meningococcal Infections - blood; Meningococcal Infections - microbiology; Middle Aged; Neisseria meningitidis; Peripheral blood; Predictive model; Prospective Studies; Public health; Public Health Practice; Regression analysis; Sentinel Surveillance; Sepsis - blood; Sepsis - microbiology; Young Adult
• ISSN: 0163-4453; 1532-2742
• DOI: 10.1016/j.jinf.2011.03.013

Summary Objective To develop a predictive model for rapid differential diagnosis of meningitis and meningococcal septicaemia to support public health decisions on chemoprophylaxis for contacts. Methods Prospective study of suspected cases of acute meningitis and meningococcal septicaemia admitted to hospitals in the South West, West Midlands and London Regions of England from July 2008 to June 2009. Epidemiological, clinical and laboratory variables on admission were recorded. Logistic regression was used to derive a predictive model. Results Of the 719 suspect cases reported, 385 confirmed cases were included in analysis. Peripheral blood polymorphonuclear count of >16 × 109 /l, serum C-reactive protein of >100 mg/l and haemorrhagic rash were strongly and independently associated with diagnosis of bacterial meningitis and meningococcal septicaemia. Using a simple scoring system, the presence of any one of these factors gave a probability of >95% in predicting the final diagnosis. Conclusion We have developed a model using laboratory and clinical factors, but not dependent on availability of CSF, for differentiating acute bacterial from viral meningitis within a few hours of admission to hospital. This scoring system is recommended in public health management of suspected cases of meningitis and meningococcal septicaemia to inform decisions on chemoprophylaxis.