Abstract
Background The emergence of dengue in malaria-endemic countries with limited diagnostic resources, such as Yemen, can be problematic because presumptive treatment of febrile cases as being malaria is a common practice. Co-infections with dengue and malaria are often overlooked and misdiagnosed as being a mono-infection because of clinical similarities. In Hodeidah city, Yemen, the capacity to conduct the diagnosis can be aggravated by the war context. To assess the magnitude of the problem, we determined the proportions of malaria, dengue and co-infection in relation to clinical characteristics among febrile outpatients. Methods This cross-sectional study included 355 febrile outpatients from Hodeidah city during the malaria transmission season (September 2018 -February 2019). Sociodemographic and clinical characteristics were collected using a pre-designed, structured questionnaire. Malaria was confirmed using microscopy and rapid diagnostic tests (RDTs), while dengue was confirmed using RDTs. Results Mono-infection proportions of 32.4% for falciparum malaria and 35.2% for dengue were found, where about two-thirds of dengue patients had a recent probable infection. However, co-infection with falciparum malaria and dengue was detected among 4.8% of cases. There was no statistically significant difference between having co-infection and mono-infection with malaria or dengue in relation to the sociodemographic characteristics. On the other hand, the odds of co-infection were significantly lower than the odds of malaria among patients presenting with sweating (OR = 0.1, 95% CI: 0.05-0.45; p <0.001), while the odds of co-infection were 3.5 times significantly higher than the odds of dengue among patients presenting with vomiting (OR = 3.5, 95% CI: 1.20-10.04; p <0.021). However, there were no statistically significant differences between having co-infection and mono-infection (malaria or dengue) in relation to other clinical characteristics. Conclusions Mono-infection with malaria or dengue can be detected among about one-third of febrile outpatients in Hodeidah, while almost 5.0% of cases can be co-infected. Sociodemographic and clinical characteristics cannot easily distinguish malaria patients from dengue-infected or co-infected ones, reinforcing the necessity of laboratory confirmation and avoidance of treating febrile patients as being presumed malaria cases.