Estimating the seroincidence of scrub typhus using antibody dynamics following infection

Introduction Scrub typhus is an acute febrile illness caused by the bacterium Orientia tsutsugamushi . Characterizing the population-level burden of scrub typhus is challenging due to the lack of accessible and accurate diagnostics. In this study, we present a novel approach that utilizes information on antibody dynamics after infection to generate population-level scrub typhus seroincidence estimates from cross-sectional serosurveys. Methods We use data from three cohorts of scrub typhus patients enrolled in Chiang Rai, Thailand, and Vellore, India, and representative population data from two serosurveys in and around the Kathmandu valley, Nepal, and Vellore, India. The samples were tested for IgM and IgG responses to Orientia tsutsugamushi -derived recombinant 56-kDa antigen using commercial ELISA kits. We used Bayesian hierarchical models to fit two-phase models to the antibody responses from scrub typhus cases and used the joint distributions of the peak antibody titers and decay rates to estimate population-level incidence rates in the cross-sectional serosurveys. We compared this new method to a traditional cut-off-based approach for estimating seroincidence. Results Median IgG antibodies persisted above OD 1.7 for 22 months, while IgM displayed longer persistence than expected, with 50% of participants having an OD >1 for 5 months. We estimated an overall seroincidence of 18 per 1000 person-years (95% CI: 16-21) in India and 4 per 1000 person-years (95% CI: 3-6) in Nepal. Among 18 to 29-year-olds, the seroincidence was 8 per 1000 person-years (95% CI 4 -16) in India and 9 per 1000 person-years (95% CI: 6-14) in Nepal. In both India and Nepal, seroincidence was higher in urban and periurban settings compared to rural areas. Compared to our method, seroincidence estimates derived from age-dependent IgG-seroprevalence without accounting for antibody decay underestimated the disease burden by 50%. By incorporating antibody dynamics, the approach described here provides more accurate age-specific infection risk estimates, emphasizing the importance of considering both IgG and IgM decay patterns in scrub typhus seroepidemiology. Conclusion The sero-surveillance approach developed in this study efficiently generates population-level scrub typhus seroincidence estimates from cross-sectional serosurveys. This methodology offers a valuable new tool for informing targeted prevention and control strategies, ultimately contributing to a more effective response to scrub typhus in endemic regions worldwide.