Background
Brucellosis occurs globally with highly variable incidence in humans from very low in North America and Western Europe to high in the Middle East and Asia. There are few data in Sub-Saharan Africa. This study estimated the incidence of human brucellosis in a pastoralist community in Kenya.
Conclusion
We report a high incidence of brucellosis in humans among members of this pastoralist community. Brucellosis was the most common cause of febrile illness in this community.
Methods
Between February 2015 and January 2016, we enrolled persons living in randomly selected households in Kajiado County. Free health care was offered at three facilities in the study area. Those who met the study clinical case definition completed a standardized questionnaire on demographics, clinical history and presentation. A blood sample was collected and tested by Rose Bengal test (RBT), then later tested at the Kenya Medical Research Institute laboratory for Brucella IgG and IgM by ELISA. Those who tested positive by both RBT and ELISA (IgG or IgM antibodies) were classified as confirmed while those that only tested positive for IgG or IgM antibodies were classified as probable. Further, sera were tested by polymerase chain reaction using a TaqMan Array Card (TAC) for a panel of pathogens causing AFI including Brucella spp. Annual incidence of brucellosis was calculated as the number of confirmed cases in one year/total number in the study population.
Results
We enrolled a cohort of 4746 persons in 804 households. Over half (52.3%) were males and the median age was 18 years (Interquartile range (IQR) 9 months- 32 years). A total of 236 patients were enrolled at three health facilities; 64% were females and the median age was 40.5 years (IQR 28-53 years). Thirty-nine (16.5%) were positive for Brucella antibodies by IgG ELISA, 5/236 (2.1%) by IgM ELISA and 4/236 (1.7%) by RBT. Ten percent (22/217) were positive by TAC. We confirmed four (1.7%) brucellosis cases giving an annual incidence of 84/100,000 persons/year (95% CI 82, 87). The incidence did not significantly vary by gender, age and location of residence.