Algorithms for the Identification of Anthrax Meningitis During a Mass Casualty Event Based on a Systematic Review of Systemic Anthrax From 1880 Through 2018.

Background: During an anthrax mass casualty event, prompt identification of patients with anthrax meningitis is important. Previous research has suggested use of a screening tool based on neurological symptoms and signs.
Methods: Using historical anthrax patient data from 1880 through 2018, we analyzed risk factors for meningitis. We developed lists of symptoms and signs (ie, algorithms) for predicting meningitis with high sensitivity and specificity. We evaluated both single and paired algorithms as screening tools.
Results: A single algorithm with 1 or more neurological symptoms or signs identifying patients with likely meningitis achieved high sensitivity (86%; 95% confidence interval [CI], 71%-100%) and specificity (90%; 95% CI, 82%-98%). Pairing algorithms with the same symptoms and signs (severe headache, altered mental status, meningeal signs, and "other neurological deficits") improved specificity (99%; 95% CI, 97%-100%) but left 17.3% of patients in a middle "indeterminate" meningitis category and in need of additional diagnostic testing to determine likely meningitis status. Pairing algorithms with differing symptoms and signs also improved specificity over the single algorithm (92%; 95% CI, 85%-99%) but categorized just 2.5% of patients as indeterminate.
Conclusions: Our study confirms prior research suggesting quick and reliable assessment of patients for anthrax meningitis is possible based on the presence or absence of certain symptoms and signs. A single algorithm was adequate; however, if we assumed low-resource diagnostic testing was feasible for some patients, pairing algorithms improved specificity. Pairing algorithms with differing symptoms and signs minimized the proportion of patients requiring additional diagnostics.
(Published by Oxford University Press on behalf of Infectious Diseases Society of America 2022.)