Clinical sequelae among individuals with pauci-symptomatic or asymptomatic Ebola virus infection and unrecognised Ebola virus disease in Liberia: a longitudinal cohort study

Abstract

Background: Whether or not individuals with pauci-symptomatic or asymptomatic Ebola virus infection and unrecognised Ebola virus disease develop clinical sequelae is unknown. We assessed current symptoms and physical examination findings among individuals with pauci-symptomatic or asymptomatic infection and unrecognised Ebola virus disease compared with Ebola virus disease survivors and uninfected contacts.

Methods: Between June 17, 2015, and June 30, 2017, we studied a cohort of Ebola virus disease survivors and their contacts in Liberia. Surveys, current symptoms and physical examination findings, and serology were used to characterise disease status of reported Ebola virus disease, unrecognised Ebola virus disease, pauci-symptomatic or asymptomatic Ebola virus infection, or no infection. We pre-specified findings known to be differentially prevalent among Ebola virus disease survivors versus their contacts (urinary frequency, headache, fatigue, muscle pain, memory loss, joint pain, neurological findings, chest findings, muscle findings, joint findings, abdominal findings, and uveitis). We estimated the prevalence and incidence of selected clinical findings by disease status.

Findings: Our analytical cohort included 991 reported Ebola virus disease survivors and 2688 close contacts. The median time from acute Ebola virus disease onset to baseline was 317 days (IQR 271-366). Of 222 seropositive contacts, 115 had pauci-symptomatic or asymptomatic Ebola virus infection and 107 had unrecognised Ebola virus disease. At baseline, prevalent findings of joint pain, memory loss, muscle pain, and fatigue were lowest among those with pauci-symptomatic or asymptomatic infection or no infection, higher among contacts with unrecognised Ebola virus disease, and highest in reported survivors of Ebola virus disease. Joint pain was the most prevalent finding, and was reported in 434 (18%) of 2466 individuals with no infection, 14 (12%) of 115 with pauci-symptomatic or asymptomatic infection, 31 (29%) of 107 with unrecognised Ebola virus disease, and 476 (48%) of 991 with reported Ebola virus disease. In adjusted analyses, this pattern remained for joint pain and memory loss. Survivors had an increased odds of joint pain compared with unrecognised Ebola virus disease contacts (adjusted odds ratio [OR] 2·13, 95% CI 1·34-3·39); unrecognised Ebola virus disease contacts had an increased odds of joint pain compared with those with pauci-symptomatic or asymptomatic infection and uninfected contacts (adjusted OR 1·89, 95% CI 1·21-2·97). The adjusted odds of memory loss was more than four-times higher among survivors than among unrecognised Ebola virus disease contacts (adjusted OR 4·47, 95% CI 2·41-8·30) and two-times higher among unrecognised Ebola virus disease contacts than in those with pauci-symptomatic or asymptomatic infection and uninfected contacts (adjusted OR 2·05, 95% CI 1·10-3·84). By 12 months, prevalent findings had decreased in the three infected groups.

Interpretation: Our findings provide evidence of post-Ebola virus disease clinical sequelae among contacts with unrecognised Ebola virus disease but not in people with pauci-symptomatic or asymptomatic Ebola virus infection.

Funding: National Cancer Institute and National Institute of Allergy and Infectious Diseases of the National Institutes of Health.

Figure 1:. Antibody concentrations by group and p values for tests comparing concentrations in the close contact groups with Ebola virus disease survivors

Concentrations were compared using generalised estimating equation linear regression models that adjusted for relationships between survivors and close contacts. The dashed vertical line indicates the cutoff for seropositivity.

Figure 2:. Prevalence of selected symptoms and findings on physical examination at baseline

Statistically significant comparisons (p<0·05) are indicated by horizontal brackets. Survivors were compared with the unrecognised Ebola virus disease and pauci-symptomatic or asymptomatic groups, and the unrecognised Ebola virus disease and pauci-symptomatic or asymptomatic groups were compared to the uninfected group. Comparisons were made using logistic regression models that adjusted for age, sex, and enrolment site (except the model for uveitis, which adjusted for age and sex only) and used generalised estimating equations to adjust for relationships between survivors and close contacts.