Aetiologies of acute undifferentiated febrile illness at the emergency ward of the University of Gondar Hospital, Ethiopia.

OBJECTIVE: Causes of acute febrile illness (AFI) often remain undetermined in developing countries, due to overlap of symptoms and limited available diagnostics. We aimed to assess the aetiology of AFI in adults in a referral hospital in northwest Ethiopia. METHODS: While all participants were tested for malaria by rapid diagnostic test (RDT), microscopy was only done on physician's request. Dengue virus (DENV) infections were detected using an RDT and ELISAs and dengue, yellow fever and chikungunya cases were identified by PCR. Bacterial aetiologies were investigated using blood culture and PCR. RESULTS: The aetiology of acute infection was identified for 20.5% of 200 patients enrolled. Eleven percent tested positive for Plasmodium, while microscopy was only requested for half of the identified malaria cases. For 4.0% of the Plasmodium-infected patients, an acute or past DENV (co-)infection was detected. We found 7.5% acute and 13.0% past DENV - all serotype 3 - infections. Bacterial infections were observed in 4.5% of the patients. CONCLUSION: Malaria is still a considerable aetiology of AFI and dengue is underrecognised. There are areas where both diseases occur concomitantly, and the DENV-3 serotype presumably spreads from Sudan to northern Ethiopia. As only 20.5% of the aetiologies were identified, a broader testing platform is required.

Evaluation of C-reactive protein and myxovirus resistance protein A to guide the rational use of antibiotics among acute febrile adult patients in Northwest Ethiopia.

OBJECTIVES: In low-resource settings, treatment is often given empirically without knowledge of the aetiology due to a lack of diagnostics. In the search for reliable rapid tests to guide treatment work-up, this study was performed to determine whether two biomarkers could differentiate bacterial from non-bacterial infections in acute febrile patients. METHODS: Adults with acute fever were recruited at a referral hospital in Ethiopia. The QuikRead Go test was used to quantify C-reactive protein (qCRP) and the FebriDx test was used for combined qualitative detection of the bacterial CRP marker with myxovirus resistance protein A (MxA), a viral biomarker. RESULTS: Of the 200 patients included in this study, most presented with 2-3 days of fever, headache, and joint pain. Antibiotics were prescribed for 83.5% and antimalarials for 36.5%, while a bacterial infection was only confirmed in 5% and malaria in 11%. The median qCRP level for confirmed bacterial infections was 128 mg/l. The FebriDx and QuikRead Go test had an overall agreement of 72.0%. CONCLUSIONS: An over-prescription of antibiotics for febrile patients was observed, even for those with low CRP levels and without a confirmed bacterial infection. The added value of the FebriDx was limited, while the combined use of rapid tests for qCRP and malaria should be considered for the management of acute febrile illness and antibiotic stewardship.