Maternally-derived antibody titer dynamics and risk of hospitalised infant dengue disease

Dengue virus (DENV) immunity is complex. Maternally-derived DENV antibodies initially provide protection against infection, however, as antibodies decay they can enhance disease severity upon infection. Quantifying antibody titers that are associated with disease risk is complicated by their dynamic nature and imperfect measurement processes. It also remains unknown whether long-term trends in birth rates, population-level infection risks, and maternal ages have altered immune profiles in child-bearing women, leading to shifts in age-specific infant disease risks. Here, we analyse DENV antibody data from two infant cohorts (N=165 infants with 665 blood draws) and 40 years of infant dengue hospitalisation data from Thailand. We use mathematical models to reconstruct maternally-derived antibody dynamics and estimate hospitalisation risk by titer. We find the relative risk of dengue hospitalisation ranges from 0.13 (0.00-0.89) in 1 month olds to 3.52 (3.25-3.79) in 8 month olds, compared to the risk in 12 month olds. We estimate the highest risk of infant dengue hospitalisation occurs at PRNT50 titers of 6.0 (5.7-6.6). Our inferred titer-related risk estimates are consistent with previously identified titer-based correlates of severe disease among older individuals experiencing secondary DENV infections, suggesting a common mechanism of risk enhancement from pre-existing antibodies. Finally, we describe how the shifting epidemiology of dengue in Thailand, combined with declining birth rates, have decreased the absolute risk of infant dengue disease by 64% over a 40 year period while having minimal impact on the mean age of infant hospital-attended dengue disease.