Ebola Virus Glycoprotein IgG Seroprevalence in Community Previously Affected by Ebola, Sierra Leone

We explored the association of Ebola virus antibody seropositivity and concentration with potential risk factors for infection. Among 1,282 adults and children from a community affected by the 2014–2016 Ebola outbreak in Sierra Leone, 8% were seropositive for virus antibodies but never experienced disease symptoms. Antibody concentration increased with age.

). The EBOV antibody geometric mean concetration (GMC) was higher in participants ≥5 years of age than in younger children (Appendix Table 1). After adjusting for age and sex, only pig ownership remained associated with antibody concentration (adjusted GMC ratio 3.0, 95% CI 1.5-5.9; p<0.01) ( Table 2).
The 8.4% seroprevalence in our study is within the range of estimates (0%-24%) from prior studies; however, this range is large because of the use of different assays, different seroprevalence thresholds, different levels of exposure to EVD cases, and studies undertaken in different geographic areas and at different timepoints relative to reported outbreaks (1). Our estimate is similar to the baseline EBOV antibody seroprevalence (4.0%) measured in another Ebola vaccine trial conducted in Liberia during the 2014-2016 EVD outbreak that used the same assay and cutoff (5).
Similar to results from previous studies, our findings showed a statistically significant increase in EBOV antibody concentration with participants' age, possibly because of increased exposure of older age groups to EBOV or to other infections that could induce cross-reactive antibodies to the EBOV GP (6,7). Potential exposures to EVD, such as healthcare work, contact with EVD cases, and funeral attendance, which were associated with EBOV transmission in other studies (8), were not associated with EBOV antibody seropositivity or concentration in our study. However, few participants reporting those risk fac-tors, and our study might have lacked the power to detect such associations.
We found an independent association of both EBOV antibody seropositivity and concentration with residence in a household compound that owned >1 pigs during the Ebola outbreak. Pigs can be experimentally infected with EBOV and can transmit the virus to nonhuman primates (9). EBOV-specific antibodies have been found in pigs in Sierra Leone and Guinea, suggesting that pigs can be naturally infected by EBOV (10,11). Pigs in the Philippines have been found to be naturally infected with Reston virus, an EBOV strain that is not known to cause disease in humans. Reston virus-specific antibodies were found in healthy farmers in contact with the infected pigs, suggesting potential transmission from pigs to humans (12). However, we found no association of EBOV antibody with having other domestic animals, in particular dogs, which also could be infected with EBOV (13,14).
One strength of our study is that we conducted our study in an area with prolonged EBOV transmission during the 2014-2016 EVD outbreak. Further, we explored a wide range of potential risk factors for EBOV acquisition, and we used the FANG ELISA, which has been proven to be more precise and accurate than a commercial alternative (4).
The first limitation of our study is that the parent RCT did not require random sampling of potential participants' households, which could have affected the generalizability of our results to the  general population. The RCT recruitment was agestaggered, and the youngest age cohort (1-3 years of age) was recruited >2 years after the EVD outbreak ended. However, a sensitivity analysis suggested that year of recruitment had a negligible confounding effect on the lower EBOV antibody concentrations observed in the youngest children (Appendix Table 2). Our study was conducted at the end of the 2014-2016 EVD outbreak in Sierra Leone, when public health measures to contain EBOV transmission had been in place for several months and the population had received messages about EVD prevention. This factor could have caused an underreporting of behaviors considered to put persons at risk for EVD. For example, hunting and consumption of bushmeat was rarely reported by our participants, in contrast with some reports that describe frequent hunting and bushmeat consumption in West Africa (15). The association of both antibody seropositivity and concentration with pig ownership is based on only 18 participants who reported keeping >1 pigs in their household compound at the time of the outbreak. This association could have occurred by chance, although the evidence of an association is quite strong. The observed association also could be confounded by unrecorded risk factors among participants who also kept pigs, such as EBOV transmission clustering in participants from a household that also owned pigs. However, that possibility seems unlikely because none of the seropositive participants who owned pigs reported contact with an EVD case, and these participants all came from different households. Finally, we are not able to determine whether EBOV antibody seropositivity in this setting reflects true asymptomatic infection because we cannot exclude underreporting of earlier EVD symptoms and we have not yet investigated crossreactivity with other viral infections. Whether EBOV seropositivity reflects acquired immunity that might provide some protection against future EBOV infections also is unclear.
Our findings suggest that the role of pigs as potential, occasional reservoirs of EBOV needs to be investigated further. The presence of antibodies binding the EBOV GP could also suggest circulation of other infectious agents, probably viruses, inducing cross-reactivity with the EBOV GP, but this possibility needs further investigation. §Types of wild animals hunted by participants who answered yes included monkeys, duiker antelopes, bats, and rodents.

Conclusions
The incidence of EBOV infection during the 2014-2016 EVD outbreak in Sierra Leone could have been higher than previously reported; 8.4% of adults and children from a community affected by the outbreak who never experienced symptoms of EVD had serologic responses to EBOV above a cutoff threshold. Our study suggests that EBOV might cause asymptomatic infection, but whether underreporting of symptoms, FANG assay specificity, or exposure to other viral infections that could generate cross-reactive antibodies also contributed to the results is unclear. These questions would benefit from further investigation to help define the extent of future EVD outbreaks. Countries at high risk for EVD outbreaks should be aware of the risk of asymptomatic or paucisyntomatic infections.