Immuno-epidemiological analysis of dengue to enhance surveillance.

As dengue continues to emerge globally, it is vital surveillance systems in endemic countries optimise routine case report data to accurately monitor dengue burden and target limited control interventions. Typical dengue surveillance practices, that often rely on case counts, are heavily distorted by underreporting. The WHO therefore promotes integrating additional surveillance practices to better describe dengue transmission. Across the Philippines, recently established laboratory surveillance routinely collects molecular and serological metrics from cross-sectional surveys of suspected dengue case reports. Research in this thesis aimed to investigate how analysis of laboratory surveillance data could be enhanced to better characterise dengue transmission dynamics across the country. The variable clinical manifestations associated with dengue are influenced by successive serotype (DENV1-4) infections individuals experience and contribute to disease underreporting. Severe dengue disease is associated with a second DENV infection. However, distinguishing primary and secondary immune status remains challenging as molecular and serological kinetics change rapidly during disease and existing methods rely on paired sera collected from patients. Here, mixture modelling approaches were adopted to characterise DENV antibody dynamics and develop a dengue immune status algorithm that could determine primary and post-primary (secondary, tertiary or quaternary) status among acute-stage dengue case reports using single serum samples. This framework achieved 90.5% agreement with the WHO gold standard method using paired sera. Surveillance metrics from this algorithm were then investigated as potential surrogate indicators of the dengue force of infection (FOI) estimated using catalytic models of age-seroprevalence and compared using Pearson’s R correlation coefficient. Across cities, the mean annual age of reporting primary infections strongly correlated (ρ: -0.85, p-value<0.001) with the FOI and highlighted prominent spatio-temporal heterogeneity in dengue burden. Notably, results also revealed reported dengue incidence was higher in cities with lower dengue FOI (ρ: -0.69, p-value:0.009) suggesting case reports represent inferior indicators of dengue burden. Common dengue serological diagnostics detect flavivirus cross-reactive antibodies and growing evidence suggests prior Zika virus exposure exacerbates subsequent dengue disease. Therefore, serological evidence of Zika was explored among dengue case reports. Findings revealed historical Zika exposure was widespread across the Philippines and an estimated 5.7% (95%CI: 3.0–10.4%) of the population became infected annually. To enhance dengue surveillance practices in low resource settings where laboratory testing is unfeasible, logistic regression models were utilised to determine dengue immune status using point-of-care rapid diagnostic tests. On specific days of disease, certain combinations of rapid test outcomes gave rise to clear immune status classifications. Together, findings in this report demonstrate how characterising dengue immune status can enhance laboratory surveillance to accurately monitor dengue transmission intensity.