UC Berkeley

Background

Zika and chikungunya are arboviral diseases of growing concern in regard to human health. Chikungunya virus (CHIKV) and Zika virus (ZIKV) recently expanded their traditional niches from Africa and Asia into the Americas. Starting in late 2013 and extending through 2016, Latin American countries experienced massive chikungunya epidemics followed by even larger Zika epidemics. Established epidemiological knowledge about Zika and chikungunya mostly derive from a few studies, usually of small or modest sample sizes. The clinical aspects of Zika in children are not well understood, as the existing literature on the topic is based on small studies using strict criteria for laboratory testing and suboptimal statistical methods. In addition, because the two arboviral diseases share a similar set of signs and symptoms, making an accurate diagnosis can be difficult in the absence of laboratory methods. Spatial studies of chikungunya and Zika to date are based on a sample of cases (i.e., instances of symptomatic infection), which creates two problems. First, spatially analyzing only cases conflates the separate processes of infection and disease; second, spatial case-only analyses ignore inapparent infections, which constitute the majority of CHIKV and ZIKV infections. Together, the existing limitations in the Zika and chikungunya literature preclude a full understanding of the viruses’ epidemiological, clinical, and spatial impacts at the population level.

Methods

The Pediatric Dengue Cohort Study (PDCS) is an ongoing, prospective cohort. At the time of the present studies, it enrolled approximately 3,700 children aged 2-14 years in Managua, Nicaragua. It was established in 2004 to study dengue virus infections but expanded to cover CHIKV and ZIKV in 2014 and 2015, respectively. We analyzed demographic, clinical, serological, and geospatial data from participants of the PDCS during the 2004-2019 timeframe. The main parameters of interest were: the percentage of CHIKV infections that are inapparent; the prevalence of anti-ZIKV antibodies (ZIKV seroprevalence), the sensitivity of the Zika case definitions from the World Health Organization and the Pan American Health Organization, the prevalence of signs and symptoms caused by symptomatic ZIKV infections, and the spatiotemporal patterns of CHIKV and ZIKV infections and cases. We used generalized additive models, Monte Carlo resampling, Bayesian statistics, hierarchical agglomerative clustering, generalized estimating equations, geostatistics, and non-parametric approaches in the analysis of PDCS data. Systematic searches of the CHIKV and ZIKV literature were used to augment and verify analyses conducted on PDCS data.

Results

The percentage of CHIKV infections that were inapparent varied by CHIKV lineage. Epidemics of the Asian CHIKV lineage exhibited a median proportion of inapparent infection of 48% (95% credible interval: 20%, 77%) and the CHIKV East/Central/South African lineage exhibited a median proportion of inapparent infection of 18% (95% credible interval: 0%, 47%). An inverse association between CHIKV seroprevalence and the percentage of CHIKV infections that were inapparent was observed in the PDCS and for most studies in the CHIKV literature, verifying a decade-old hypothesis in the field. After a three-month Zika epidemic in Managua, Nicaragua, ZIKV seroprevalence was 36% among children (aged 2-14 years), 46% among the overall population (aged 2-80 years), and 56% among adults (aged 15-80 years), as measured by the biotinylated Zika NS1 blockade-of-binding assay. The clinical spectrum of Zika was observed to vary across age; older children exhibited a higher number and wider range of signs, symptoms, and complete blood count findings than younger children. Official case definitions for Zika exhibited low sensitivity (20-32%) in children because the case definitions were derived from clinical data in adults, and adults tend to exhibit different clinical manifestations of Zika than children. The 2014-2015 chikungunya epidemics and the 2016 Zika epidemic had different spatiotemporal patterns and exhibited distinct spatial clusters of infections. Overall, distance from participants’ households to a cemetery abutting the study area was a robust indicator of infection risk in both the large 2015 chikungunya epidemic and the large 2016 Zika epidemic.

Conclusions

Our results substantially update the epidemiological, clinical, and spatial knowledge regarding ZIKV and CHIKV. Our studies, often the largest of their kind, provide novel insights into the viruses and their respective diseases. In addition, our epidemiological results provide hypotheses that can be explored in future virological and immunological studies. Taken together, our studies provide practical knowledge that ministries of health and international health organizations can use to better prepare for future epidemics of Zika and chikungunya.