Background
The 2014–15 Ebola virus disease (EVD) epidemic strained health systems in west Africa already overburdened with other diseases, including malaria. Because EVD and malaria can be difficult to distinguish clinically, and rapid testing was not available in many Ebola Treatment Units (ETUs), guidelines recommended empirical malaria treatment. Little is known, however, about the prevalence and characteristics of patients entering an ETU who were infected with malaria parasites, either alone or concurrently with Ebola virus.

Methods
Data for sociodemographics, disease characteristics, and mortality were analysed for patients with suspected EVD admitted to three ETUs in Sierra Leone using a retrospective cohort design. Testing for Ebola virus was done by real-time PCR and for malaria by a rapid diagnostic test. Characteristics of patients were compared and survival analyses were done to evaluate the effect of infection status on mortality.

Findings
Between Dec 1, 2014, and Oct 15, 2015, 1524 cases were treated at the three ETUs for suspected EVD, of whom 1368 (90%) had diagnostic data for malaria and EVD. Median age of patients was 29 years (IQR 20–44) and 715 (52%) were men. 1114 patients were EVD negative, of whom 365 (33%) tested positive for malaria. Of 254 EVD positive patients, 53 (21%) also tested positive for malaria. Mortality risk was highest in patients diagnosed with both EVD and malaria (35 [66%] of 53 died) and patients diagnosed with EVD alone (105 [52%] of 201 died). Compared with patients presenting to ETUs without malaria or EVD, mortality was increased in the malaria positive and EVD positive group (adjusted hazard ratio 9·36, 95% CI 6·18–14·18, p<0·0001), and the malaria negative and EVD positive group (5·97, 4·44–8·02, p<0·0001), but reduced in the malaria positive and EVD negative group (0·37, 0·20–1·23, p=0·0010).

Interpretation
Malaria parasite co-infection was common in patients presenting to ETUs and conferred an increased mortality risk in patients infected with Ebola virus, supporting empirical malaria treatment in ETUs. The high mortality among patients without EVD or malaria suggests expanded testing and treatment might improve care in future EVD epidemics.