Identifying Recent Cholera Infections Using a Multiplex Bead Serological Assay

• Published in: mBio 2022-12, Vol.13 (6), p.e0190022-e0190022
• Main Authors: Jones, Forrest K, Bhuiyan, Taufiqur R, Muise, Rachel E, Khan, Ashraful I, Slater, Damien M, Hutt Vater, Kian Robert, Chowdhury, Fahima, Kelly, Meagan, Xu, Peng, Kováč, Pavol, Biswas, Rajib, Kamruzzaman, Mohammad, Ryan, Edward T, Calderwood, Stephen B, LaRocque, Regina C, Lessler, Justin, Charles, Richelle C, Leung, Daniel T, Qadri, Firdausi, Harris, Jason B, Azman, Andrew S
• Format: Article
• Language: English
• Subjects: Antibodies, Bacterial; Bangladesh - epidemiology; Cholera - epidemiology; Cross-Sectional Studies; Epidemiology; Humans; Immunoglobulin A; Immunoglobulin G; multiplex bead assay; Research Article; seroincidence; serosurveillance; Vibrio cholerae; Vibrio cholerae O1
• ISSN: 2150-7511; 2150-7511
• DOI: 10.1128/mbio.01900-22

Estimates of incidence based on medically attended cholera can be severely biased. Vibrio cholerae O1 leaves a lasting antibody signal and recent advances showed that these can be used to estimate infection incidence rates from cross-sectional serologic data. Current laboratory methods are resource intensive and challenging to standardize across laboratories. A multiplex bead assay (MBA) could efficiently expand the breadth of measured antibody responses and improve seroincidence accuracy. We tested 305 serum samples from confirmed cholera cases (4 to 1083 d postinfection) and uninfected contacts in Bangladesh using an MBA (IgG/IgA/IgM for 7 Vibrio cholerae O1-specific antigens) as well as traditional vibriocidal and enzyme-linked immunosorbent assays (2 antigens, IgG, and IgA). While postinfection vibriocidal responses were larger than other markers, several MBA-measured antibodies demonstrated robust responses with similar half-lives. Random forest models combining all MBA antibody measures allowed for accurate identification of recent cholera infections (e.g., past 200 days) including a cross-validated area under the curve (cvAUC ) of 92%, with simpler 3 IgG antibody models having similar accuracy. Across infection windows between 45 and 300 days, the accuracy of models trained on MBA measurements was non-inferior to models based on traditional assays. Our results illustrated a scalable cholera serosurveillance tool that can be incorporated into multipathogen serosurveillance platforms. Reliable estimates of cholera incidence are challenged by poor clinical surveillance and health-seeking behavior biases. We showed that cross-sectional serologic profiles measured with a high-throughput multiplex bead assay can lead to accurate identification of those infected with pandemic Vibrio cholerae O1, thus allowing for estimates of seroincidence. This provides a new avenue for understanding the epidemiology of cholera, identifying priority areas for cholera prevention/control investments, and tracking progress in the global fight against this ancient disease.