Comparison of dengue, chikungunya, and Zika among children in Nicaragua across 18 years: a single-centre, prospective cohort study.

BACKGROUND

Dengue, chikungunya, and Zika are diseases of major human concern. Differential diagnosis of these three diseases is complicated in children and adolescents due to overlapping clinical features (signs, symptoms, and complete blood count results). Few studies have directly compared these three diseases. We aimed to use 18 years of primary care observations from a paediatric cohort to characterise the distinguishing features of dengue, chikungunya, and Zika.

METHODS

This single-centre prospective cohort study was based on the ongoing Pediatric Dengue Cohort Study (PDCS), which started on Aug 30, 2004, in District II of Managua, Nicaragua. The PDCS was initiated to study dengue virus infections in children who attended the Health Center Sócrates Flores Vivas (HCSFV) for their medical needs; over the years, the PDCS expanded the age range (2 to <10 years expanded to 2 to <18 years). The PDCS also expanded eligibility criteria to include chikungunya virus and Zika virus before they entered the geographical study area in August, 2014 and January, 2016, respectively. For this study, we included laboratory confirmed cases of dengue, chikungunya, and Zika enrolled in the PDCS between Jan 19, 2006, and Dec 31, 2023, and evaluated at the HCSFV. We assessed clinical features (clinical records and laboratory results) during the first 10 days of illness using generalised additive models, day-specific and disease-specific prevalence estimates, and machine learning models.

FINDINGS

We characterised 1405 dengue, 517 chikungunya, and 522 Zika cases. The median age was 10·0 years (IQR 7·0-12·7); 1165 (47·7%) cases were male and 1279 (52·3%) were female. The prevalence of many clinical features shown by dengue, chikungunya, and Zika cases differed substantially overall, by age, and by day of illness. The presence of basophilia (prevalence difference 42·3% [95% CI 37·4 to 47·0]), monocytopenia (13·0% [10·0 to 16·4]), abdominal pain (19·1% [15·7 to 22·9]), and leukopenia (41·1% [36·2 to 45·6]) best distinguished dengue; the presence of arthralgia (60·5% [56·3 to 64·2]) and absence of papular rash (-14·9% [-17·2 to -12·7]), leukopenia (-32·0% [-36·7 to -27·1]), and conjunctival injection (-4·9% [-6·6 to -3·3]) best distinguished chikungunya; and the presence of generalised rash (35·0% [30·1 to 39·7]) and absence of fever (-37·3% [-41·7 to -33·0]), headache (-36·2% [-41·1 to -31·2]), myalgia (-30·1% [-33·9 to -26·2]), and lymphocytopenia (-41·9% [-46·6 to -37·1]) best distinguished Zika. Dengue and chikungunya cases showed similar temperature dynamics during acute illness, and their mean temperatures were higher than Zika cases. 62 laboratory confirmed afebrile dengue cases, which would not be captured by any widely used international case definition, presented most similarly to afebrile Zika cases, but five (8·1%) had warning signs of dengue disease severity. Based on boosted regression tree models, the presence of arthralgia and absence of basophilia and leukopenia most distinguished chikungunya, the presence of basophilia and leukopenia most distinguished dengue, and the absence of fever most distinguished Zika.

INTERPRETATIONS

These findings substantially update the understanding of dengue, chikungunya, and Zika in a paediatric population and identify various clinical features that could improve differential diagnoses. The occurrence of afebrile dengue warrants reconsideration of current guidance.

FUNDING

US National Institutes of Health.

TRANSLATION

For the Spanish translation of the abstract see Supplementary Materials section.