Figure 1

Figure 1. Prevalence of Ebola virus (EBOV) and Plasmodium spp. RNA in patient samples submitted to the Centers for Disease Control and Prevention–National Institutes of Health diagnostic laboratory at the Eternal Love Winning...

During October 12, 2014–March 28, 2015, samples collected from 1,058 persons in Liberia with suspected EBOV infection were submitted to the CDC–NIH ELWA laboratory. The samples used for this research were collected for public health surveillance and not human subjects research, so institutional review board review and approval were not required. Early during the study period (October–November), most of the patients from whom samples were collected received a diagnosis of EBOV infection (Figure 1, panel A); toward the end, however, few cases of EBOV infection were diagnosed in Liberia. The number of patients who received a diagnosis of Plasmodium spp. parasitemia remained stable over time; thus, despite the lack of positive EBOV test results in the final months of the study period, overall, 40%–60% of patients each week received a diagnosis of EBOV infection, Plasmodium spp. parasitemia, or both (Figure 1, panel A). Of 1,058 samples tested, 259 (24.5%) were positive for EBOV alone, 243 (23%) were positive for Plasmodium spp. alone, and 47 (4.4%) were positive for both (Figure 1, panel B). Of 311 Plasmodium-positive samples that were further analyzed, 296 (95%) were positive for P. falciparum (Figure 1, panel C), confirming that P. falciparum was the main Plasmodium species causing parasitemia in our cohort.

Figure 2

Figure 2. Inverse parasite load in patients with Plasmodium spp. parasitemia over time by month of sample submission, for samples submitted to the Centers for Disease Control and Prevention–National Institutes of Health diagnostic...

The cycle threshold (C_t) values observed in the Plasmodium spp. qRT-PCR results are a proxy for parasite load (Figure 2). A high C_t value corresponds to a low-level parasitemia; the lower the C_t value, the higher the number of Plasmodium parasites detected. All patients in the cohort were triaged as having suspected EBOV infection, and thus all had clinical symptoms that might have been caused by EBOV, Plasmodium spp., or a different pathogen. Of note, significantly fewer Plasmodium parasites were detected in patients with EBOV infection than in patients who were not infected with EBOV (average Plasmodium C_t 24.7 vs. 20.37; p<0.01 by unpaired Student t-test), likely because clinical symptoms in these patients were caused by the EBOV infection rather than malaria. Some of the patients in our cohort with Plasmodium spp. parasitemia might have been asymptomatic carriers, especially those with a low-level parasitemia.

One could argue that because all patients with febrile illness who were seen at many of the ETUs were given antimalarial treatment when they first sought medical care, providing laboratory testing for Plasmodium spp. parasitemia is not useful. Presumptive artemisinin-based combination treatment is recommended for all patients seen at ETUs, followed by prompt malaria diagnostic testing so that appropriate measures can be taken if oral treatment cannot be sustained because of clinical symptoms (9,10). Moreover, mathematical models predict a large increase in malaria in the epidemic region because of lapses in malaria control (7). Therefore, Plasmodium spp. parasitemia testing of all febrile patients seen at healthcare facilities could be used to determine whether this increase is indeed occurring so that countermeasures can be scaled up accordingly.

Differential diagnostic testing should be expanded to identify the cause of disease in the patients in our cohort whose samples tested negative for EBOV and Plasmodium spp., as was attempted in 1 laboratory in Sierra Leone (15). However, this kind of testing was difficult for several reasons. First, during the height of the EVD epidemic in Liberia, most laboratories were already working close to capacity while testing for EBOV alone. Testing for additional pathogens would have required the allocation of additional resources, including equipment and personnel. Second, it might not be possible to perform diagnostics using a whole blood sample for all the pathogens of differential diagnostic importance that can cause signs and symptoms similar to those of EVD (e.g., typhoid, bacterial sepsis, shigellosis, cholera, leptospirosis, dengue fever, rickettsioses, relapsing fever, meningitis, viral hepatitis, influenza, Lassa fever). Third, each of these pathogens likely is the cause of disease in only a small subset of febrile patients. Last, the prevalence of these pathogens likely differs in each outbreak region. Because broad-spectrum antimicrobial drugs and antimalarial drugs are routinely administered to all patients seen at an ETU, consideration for additional testing should focus on endemic pathogens that would not respond to these treatments.

By using PCR-based detection for Plasmodium spp. parasitemia, the need for additional handling of clinical specimens possibly infected with EBOV was eliminated. Thus, the addition of PCR-based diagnostic tests to detect Plasmodium spp. does not pose an additional safety risk to laboratory staff. Also, it is less time-consuming to add additional PCR reactions to a PCR run than to separately perform microscopy or an RDT on each sample, and these additions would add only ≈15 minutes to the overall time needed from sample submission to reporting of results. Moreover, if a multiplex approach is used, the additions would not require extra time or reagents. Taken together, our findings suggest that PCR-based testing for Plasmodium spp. parasitemia can be implemented easily and safely in laboratories performing EBOV diagnostics to assist with case-patient management during EVD outbreaks in malaria-endemic areas.