Molecular screening reveals non-uniform malaria transmission in western Kenya and absence of Rickettsia africae and selected arboviruses in hospital patients

Background In sub-Saharan Africa, malaria is the common diagnosis for febrile illness and related clinical features, resulting in the under-diagnosis of other aetiologies, such as arboviruses and Rickettsia. While these may not be significant causes of mortality in malaria-endemic areas, they affect the daily life and performance of affected individuals. It is, therefore, important to have a clear picture of these other aetiologies to institute correct diagnoses at hospitals and improve patient outcomes. Methods Blood samples were collected from patients with fever and other clinical features associated with febrile illness at selected hospitals in the malaria-endemic counties of Busia, Bungoma, and Kakamega, and screened for Crimean-Congo haemorrhagic fever, Sindbis, dengue and chikungunya viruses, Rickettsia africae, and Plasmodium spp. using high-throughput real-time PCR techniques. A logistic regression was performed on the results to explore the effect of demographic and socio-economic independent variables on malaria infection. Results A total of 336 blood samples collected from hospital patients between January 2018 and February 2019 were screened, of which 17.6% (59/336) were positive for Plasmodium falciparum and 1.5% (5/336) for Plasmodium malariae. Two patients had dual P. falciparum/P. malariae infections. The most common clinical features reported by the patients who tested positive for malaria were fever and headache. None of the patients were positive for the arboviruses of interest or R. africae. Patients living in Busia (OR 5.2; 95% CI 2.46–11.79; p < 0.001) and Bungoma counties (OR 2.7; 95% CI 1.27–6.16; p = 0.013) had higher odds of being infected with malaria, compared to those living in Kakamega County. Conclusions The reported malaria prevalence is in line with previous studies. The absence of arboviral and R. africae cases in this study may have been due to the limited number of samples screened, low-level circulation of arboviruses during inter-epidemic periods, and/or the use of PCR alone as a detection method. Other sero-surveys confirming their circulation in the area indicate that further investigations are warranted.