Prognostic Indicators for Ebola Patient Survival

Abstract

To determine whether 2 readily available indicators predicted survival among patients with Ebola virus disease in Sierra Leone, we evaluated information for 216 of the 227 patients in Bo District during a 4-month period. The indicators were time from symptom onset to healthcare facility admission and quantitative real-time reverse transcription PCR cycle threshold (Ct), a surrogate for viral load, in first Ebola virus-positive blood sample tested. Of these patients, 151 were alive when detected and had reported healthcare facility admission dates and Ct values available. Time from symptom onset to healthcare facility admission was not associated with survival, but viral load in the first Ebola virus-positive blood sample was inversely associated with survival: 52 (87%) of 60 patients with a Ct of >24 survived and 20 (22%) of 91 with a Ct of <24 survived. Ct values may be useful for clinicians making treatment decisions or managing patient or family expectations.

Keywords: Ebola; Ebola virus; Ebola virus disease; Sierra Leone; hemorrhagic fever; prognosis; survival; viruses; zoonoses.

Figure 1

Classification of patients with Ebola virus disease into study groups, Bo District, Sierra Leone, September 2014–January 2015.

Figure 2

Percent survival among 151 patients in the Ebola virus disease (EVD) primary cohort by cycle threshold (C_t) rounded to nearest integer, Bo District, Sierra Leone, September 2014–January 2015. Locally weighted smoothing line and 95% uncertainty intervals added to illustrate trend. The area of each dot is scaled to represent the number of confirmed EVD cases, by C_t. The trend line suggests a sharp increase in survival for patients with C_t values in the mid-20s (dotted line).

Figure 3

Scatterplot of outcome by cycle threshold (C_t) at time of first Ebola virus–positive test result and time to admission at any healthcare facility (primary cohort, n = 151), Bo District, Sierra Leone, September 2014–January 2015. Each circle represents an infected person. The dashed line indicates the classification threshold of the C_t value of 24. Observations are slightly horizontally jittered to reduce overplotting.