Primary exposure to Zika virus is linked with increased risk of symptomatic dengue virus infection with serotypes 2, 3, and 4, but not 1

Abstract

Infection with any of the four dengue virus serotypes (DENV1-4) can protect against or enhance subsequent dengue depending on preexisting antibodies and infecting serotype. Additionally, primary infection with the related flavivirus Zika virus (ZIKV) is associated with increased risk of DENV2 disease. Here, we measured how prior DENV and ZIKV immunity influenced risk of disease caused by DENV1-4 in a pediatric Nicaraguan cohort. Of 3412 participants in 2022, 10.6% experienced dengue cases caused by DENV1 (n = 139), DENV4 (n = 133), DENV3 (n = 54), DENV2 (n = 9), or an undetermined serotype (n = 39). Longitudinal clinical and serological data were used to define infection histories, and generalized linear and additive models adjusted for age, sex, time since last infection, and year, and repeat measurements were used to predict disease risk. Compared with flavivirus-naïve participants, primary ZIKV infection was associated with increased risk of disease caused by DENV4 (relative risk = 2.62, 95% confidence interval: 1.48 to 4.63) and DENV3 (2.90, 1.34 to 6.27), but not DENV1 infection. Primary DENV infection or DENV followed by ZIKV infection was also associated with increased risk of DENV4 disease. We reanalyzed 19 years of cohort data and demonstrated that prior flavivirus immunity and antibody titer had distinct associations with disease risk depending on incoming serotype. We thus find that prior ZIKV infection, like prior DENV infection, is associated with increased risk of disease with certain DENV serotypes. Cross-reactivity among flaviviruses should be considered when assessing vaccine safety and efficacy.

Fig. 1.. Nicaragua experienced a large dengue epidemic in 2022–2023 with co-circulation of all four serotypes.

(A to C) Monthly dengue and Zika cases (a), the proportion of individuals with various infection histories (B), and denV ieliSa titer distributions (c) by epidemic year in the Pediatric dengue cohort Study (2004 to 2022).

Fig. 2.. Prior infection history confers differential risk of symptomatic DENV1, DENV3, and DENV4 as well as differential risk of overall disease and disease severity.

Log-binomial and log-Poisson GLMs were used to estimate the risk of any dengue case, a case caused by DENV1, DENV2, DENV3, or DENV4, and DwWS/SD by DENV and ZIKV infection history. (A and B) Risk ratios for the 2022–2023 epidemic (A) or all years of the cohort study, 2004 to 2023 (B). Data are presented as risk ratios with 95% CIs. Symbols indicate the degree of significance. An open circle is nonsignificant, closed circle represents P < 0.05, solid triangle indicates P < 0.01, and solid square indicates P < 0.001. *P < 0.05, **P < 0.01, ***P < 0.001. DENV2 was not evaluated for the 2022–2023 epidemic because of small sample size.

Fig. 3.. History of prior flavivirus infection exacerbates risk of symptomatic DENV2 and DENV4 as well as overall disease and disease severity.

(A) Linear mixed-effects models were used to estimate the risks of a symptomatic DENV infection (case) by any serotype or from DENV1, DENV2, DENV3, or DENV4 infection. (B) Linear mixed-effects models were used to estimate the risks of DwWS/SD and DHF/DSS. Both (A) and (B) are derived from the Pediatric Dengue Cohort Study, 2004 to 2023 by immune status [naïve or flavivirus-immune (flavi-immune)]. Data are presented as risk ratios with 95% CIs. Symbols indicate the degree of significance. An open circle is nonsignificant, closed circle represents P < 0.05, and solid triangle indicates P < 0.01. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 4.. Risks of dengue differ by preexisting anti-DENV antibody titers and infecting serotype.

(A and B) Linear mixed-effects models were used to estimate the risk of any symptomatic DENV infection as well as a symptomatic infection caused by DENV1–4 by preexisting DENV iELISA titer during the 2022–2023 epidemic (A) or all years of the cohort study (2004 to 2023) (B). Data are presented as risk ratios with 95% CIs. Symbols indicate the degree of significance. An open circle is nonsignificant, closed circle represents P < 0.05, solid triangle represents P < 0.01, and solid square equates to P < 0.001. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 5.. DENV iELISA titer is a mediator that helps explain the relationship between infection history and disease outcome.

(A) Mediation path diagram (left) and effects and contribution of DENV iELISA titers (between 10 and 320) on disease risk attributed to prior flavivirus infection (right). (B) Mediation path diagram (left) and effects and contribution of DENV iELISA titers (>320) on disease risk attributed to prior flavivirus infection (right). Data are presented as risk ratios with 95% CIs. *P < 0.05, **P < 0.01.

Fig. 6.. Primary DENV and ZIKV infections, as well as the order of DENV or ZIKV infections, result in distinct antibody dynamics.

Antibody kinetics in the Pediatric Dengue Cohort Study, as measured using the DENV iELISA, are shown. Data are shown for 1 to up to 10 years for the following infection histories: DENV-immune, primary DENV, primary ZIKV, DENV-DENV, DENV-ZIKV, ZIKV-DENV, 2+DENV, and 2+DENV-ZIKV infections. Individual trajectories and group generalized additive model fits (black line; 95% CIs, gray shading) are shown.