Preexisting Japanese encephalitis virus neutralizing antibodies and increased symptomatic dengue illness in a school-based cohort in Thailand

Abstract

Background: Dengue viruses (DENVs) and Japanese encephalitis virus (JEV) have significant cross-reactivity in serological assays; the clinical implications of this remain undefined. An improved understanding of whether and how JEV immunity modulates the clinical outcome of DENV infection is important as large-scale DENV vaccine trials will commence in areas where JEV is co-endemic and/or JEV immunization is routine.

Methods and findings: The association between preexisting JEV neutralizing antibodies (NAbs) and the clinical severity of DENV infection was evaluated in a prospective school-based cohort in Thailand that captured asymptomatic, non-hospitalized, and hospitalized DENV infections. Covariates considered included age, baseline DENV antibody status, school of attendance, epidemic year, and infecting DENV serotype. 942 children experienced at least one DENV infection between 1998 and 2002, out of 3,687 children who were enrolled for at least one full year. In crude analysis, the presence of JEV NAbs was associated with an increased occurrence of symptomatic versus asymptomatic infection (odds ratio [OR]= 1.55, 95% CI: 1.08-2.23) but not hospitalized illness or dengue hemorrhagic fever (DHF). The association was strongest in children with negative DENV serology (DENV-naive) (OR=2.75, 95% CI: 1.12-6.72), for whom the presence of JEV NAbs was also associated with a symptomatic illness of longer duration (5.4 days for JEV NAb+ versus 2.6 days for JEV NAb-, p=0.048). JEV NAbs were associated with increased DHF in younger children with multitypic DENV NAb profiles (OR=4.05, 95% CI: 1.18 to 13.87). Among those with JEV NAbs, the association with symptomatic illness did not vary by antibody titer.

Interpretation: The prior existence of JEV NAbs was associated with an increased probability of symptomatic as compared to asymptomatic DENV illness. These findings are in contrast to previous studies suggesting an attenuating effect of heterologous flavivirus immunity on DENV disease severity.

Figure 1. Clinical severity of dengue infections by strata of preexisting DENV and JEV immunity.

The proportions of dengue (DENV) infections resulting in symptomatic illness (1a), hospitalized illness (1b), and dengue hemorrhagic fever (1c) are shown. Data are stratified by preexisting DENV immunity (naïve [DENV -], monotypic [DENV 1+], multitypic older and younger than 10 years of age [DENV >1+]) and preexisting Japanese encephalitis virus (JEV) neutralizing antibodies (NAbs) (positive [+] or negative [-]). The JEV NAb positive groups are indicated by hash lines for each stratum of DENV immunity. Odds ratios (ORs) estimate the odds of being experiencing the disease severity of interest (dengue hemorrhagic fever [DHF], hospitalized illness [Hosp], or symptomatic illness [Sx]) in the presence of JEV NAbs over the odds of experiencing the disease severity of interest in the absence of JEV NAbs. Values in parentheses indicate the 95% confidence intervals for the ORs. Error bars indicate the 95% confidence intervals for proportions.

Figure 2. Distribution of JEV neutralizing antibody titers among those with detectable JEV antibodies.

Log10 titers are shown for asymptomatic (dark gray) and symptomatic (striped) infections. JEV: Japanese encephalitis.

Figure 3. Associations between JEV neutralizing antibodies and clinical severity of DENV infections, by DENV serotype.

The total number of cases of non-hospitalized dengue fever (DF), hospitalized DF, and dengue hemorrhagic fever (DHF) are shown by dengue (DENV) serotype (3a). The proportions of infections developing symptomatic illness (3b) and hospitalized illness (3c) are shown by Japanese encephalitis (JEV) neutralizing antibody (NAb) status and infecting serotype. JEV NAb positive groups are represented by hash lines; JEV NAb negatives by white bars. For figure 3b, the infecting DENV serotype was imputed for asymptomatic and RT-PCR-negative symptomatic infections, using information on the predominate DENV serotype in circulation at a child's school for the time interval during which they were infected, and restricting to those schools that had only one serotype in circulation that year (as detected by RT-PCR). Analysis was restricted to RT-PCR-positive, and therefore symptomatic, infections in figure 3c. Error bars indicate the 95% confidence interval for the proportion of children developing symptomatic illness (3b) and hospitalized illness (3c).