Endemic chikungunya fever in Kenyan children: a prospective cohort study

Abstract

Background: Chikungunya fever (CHIKF) was first described in Tanzania in 1952. Several epidemics including East Africa have occurred, but there are no descriptions of longitudinal surveillance of endemic disease. Here, we estimate the incidence of CHIKF in coastal Kenya and describe the associated viral phylogeny.

Methods: We monitored acute febrile illnesses among 3500 children visiting two primary healthcare facilities in coastal Kenya over a 5-year period (2014-2018). Episodes were linked to a demographic surveillance system and blood samples obtained. Cross-sectional sampling in a community survey of a different group of 435 asymptomatic children in the same study location was done in 2016. Reverse-transcriptase PCR was used for chikungunya virus (CHIKV) screening, and viral genomes sequenced for phylogenetic analyses.

Results: We found CHIKF to be endemic in this setting, associated with 12.7% (95% CI 11.60, 13.80) of all febrile presentations to primary healthcare. The prevalence of CHIKV infections among asymptomatic children in the community survey was 0.7% (95% CI 0.22, 2.12). CHIKF incidence among children < 1 year of age was 1190 cases/100,000-person years and 63 cases/100,000-person years among children aged ≥10 years. Recurrent CHIKF episodes, associated with fever and viraemia, were observed among 19 of 170 children with multiple febrile episodes during the study period. All sequenced viral genomes mapped to the ECSA genotype albeit distinct from CHIKV strains associated with the 2004 East African epidemic.

Conclusions: CHIKF may be a substantial public health burden in primary healthcare on the East African coast outside epidemic years, and recurrent infections are common.

Fig. 1

The Kilifi Health and Demographic Surveillance System. The location of Ngerenya and Pingilikani dispensaries within the KHDSS

Fig. 2

Study population. The flow of study participants, reasons for exclusion and results from CHIKV RT-PCR screening are shown

Fig. 3

Recurrent CHIKF episodes. The distribution of febrile episodes for each of 19 children who had at least two CHIKV RT-PCR positive febrile episodes is shown. The timing of the CHIKV RT-PCR positive results is shown, with the index episode shown as the first data point in the follow up timeline. Children with prefix ‘NGE’ were resident in Ngerenya and those with ‘PIN’ from Pingilikani, respectively

Fig. 4

Phylogenetic analysis of CHIKV genomes from Kenya. BEAST MCC tree of 123 ECSA genomes collected across the world, including 24 from Kenya is shown in the left panel. On the right panel is an extraction highlighting the CHIKV sequences generated in this study. The tips are coloured according to geographic location or child identity (as in Fig. 3, for those with paired index-recurrent episode sequence pairs), with posterior probability support shown for select nodes with probability > 0.8. Sample MT526800 was from a child with a single CHIKF episode whose RT-PCR cycle threshold (Ct) value was low (Ct = 19). The same and more detailed MCC tree can be found in Figure S3 in Supplementary Material