Skip to main content
  • Loading metrics

Zoonotic risk factors associated with seroprevalence of Ebola virus GP antibodies in the absence of diagnosed Ebola virus disease in the Democratic Republic of Congo


18 May 2022: Bratcher A, Hoff NA, Doshi RH, Gadoth A, Halbrook M, et al. (2022) Correction: Zoonotic risk factors associated with seroprevalence of Ebola virus GP antibodies in the absence of diagnosed Ebola virus disease in the Democratic Republic of Congo. PLOS Neglected Tropical Diseases 16(5): e0010465. View correction



Ebola virus (EBOV) is a zoonotic filovirus spread through exposure to infected bodily fluids of a human or animal. Though EBOV is capable of causing severe disease, referred to as Ebola Virus Disease (EVD), individuals who have never been diagnosed with confirmed, probable or suspected EVD can have detectable EBOV antigen-specific antibodies in their blood. This study aims to identify risk factors associated with detectable antibody levels in the absence of an EVD diagnosis.


Data was collected from September 2015 to August 2017 from 1,366 consenting individuals across four study sites in the DRC (Boende, Kabondo-Dianda, Kikwit, and Yambuku). Seroreactivity was determined to EBOV GP IgG using Zaire Ebola Virus Glycoprotein (EBOV GP antigen) ELISA kits (Alpha Diagnostic International, Inc.) in Kinshasa, DRC; any result above 4.7 units/mL was considered seroreactive. Among the respondents, 113 (8.3%) were considered seroreactive. Several zoonotic exposures were associated with EBOV seroreactivity after controlling for age, sex, healthcare worker status, location, and history of contact with an EVD case, namely: ever having contact with bats, ever having contact with rodents, and ever eating non-human primate meat. Contact with monkeys or non-human primates was not associated with seroreactivity.


This analysis suggests that some zoonotic exposures that have been linked to EVD outbreaks can also be associated with EBOV GP seroreactivity in the absence of diagnosed EVD. Future investigations should seek to clarify the relationships between zoonotic exposures, seroreactivity, asymptomatic infection, and EVD.

Author summary

Ebola virus (EBOV) is spread through exposure to infected bodily fluids of a human or animal. While EBOV can lead to a severe disease, Ebola Virus Disease (EVD), it is possible for individuals to have anti-EBOV antibodies without ever getting sick with EVD. Seroreactivity (the detection of antigen-specific antibodies) suggests that that person has been exposed to EBOV or a similar virus in the past. Our study looked for seroreactive individuals who have never received an EVD diagnosis in four sites across the Democratic Republic of the Congo. Then we checked if animal exposures previously linked to EVD were more common among seroreactive individuals than non-seroreactive individuals. Among respondents from all four sites, 113 (8.3%) were seroreactive to EBOV. Additionally, contact with bats, rodents, and eating non-human primate meat were associated with seroreactivity, indicating these factors may be predictors of undocumented EBOV exposure events. These findings show that some EVD risk factors may be associated with EBOV seroreactivity without EVD diagnosis. Future research is needed to clarify the relationships between zoonotic exposures, seroreactivity, asymptomatic infection, and EVD.


Ebola virus (EBOV) is a zoonotic filovirus spread through exposure to infected bodily fluids or contact with an infected human or animal [1,2]. While EBOV ecology remains largely unknown, it is thought that outbreaks are initiated by zoonotic spillover events, where virus is passed from an infected animal to a human [3,4]. In several cases, outbreaks have been traced back to a single human-animal interaction, most commonly with a bat or non-human primate [48]. When the source is not traceable to a single zoonotic spillover event, outbreaks often originate in forested areas or with individuals who frequent forests [5]. Once introduced into the population, disease may then be propagated through human-to-human transmission by those who exhibit EVD symptoms [9].

For every spillover event that results in EVD, there are likely many EBOV exposure events that produce subclinical or asymptomatic EBOV infection and serologic response[10]; Studies conducted throughout sub-Saharan Africa have identified individuals with detectable levels of antibodies that recognize EBOV antigens, yet without a history of EVD diagnosis[1117]. Several different hypotheses could explain this phenomenon: surveillance failures, undiagnosed EVD due to subclinical or asymptomatic infection [18], infection by another ebolavirus species, or infection by a less virulent, unidentified virus, which is antigenically cross-reactive with known ebolavirus species [5]. The latter hypothesis is supported by the recent discovery of Bombali virus, an ebolavirus that is not known to cause disease in humans [19,20]. Additionally, these antibodies could represent exposure to, but not infection by, EBOV, which could occur if an individual were exposed to inactivated virus or virus particles [2123]. Thus, evidence of exposure to EBOV in those without a history of EVD may originate from several potential etiologies, but ultimately suggests contact between an individual and EBOV or an EBOV-like virus.

Though the existence of EBOV seroreactivity among individuals with no history of EVD is well-documented[1117], minimal research characterizes risk factors associated with such seroprevalence. A small number of previous studies have focused on demographics or ecological predictors of seroprevalence, such as an individual’s occupation being located in forest vs. non-forested areas [16,2427]. However, few studies consider the association between seroprevalence and contact with specific animals that may carry EBOV, such as bats, or activities that may lead to zoonotic exposure. These more precise predictors of spillover events may be identified using serosurveys in areas where EBOV is thought to exist in the environment, including regions at high risk for EVD outbreaks.

The Democratic Republic of the Congo (DRC) has experienced twelve EVD outbreaks in history, and therefore is an optimal region in which to study zoonotic exposures to EBOV[28]. This study of individuals throughout the DRC aims to identify associations between zoonotic exposures and seroreactivity to EBOV in the absence of diagnosed EVD. If EBOV seroreactivity in the absence of EVD is correlated with protective immunity, identifying its predictors could have significant contributions to understanding natural Ebola resistance in the DRC. Relationships detected in this analysis will be used to better understand zoonotic EBOV exposure and seroprevalence of EBOV GP antibodies in the DRC.


Ethics statement

Ethical approval was obtained at UCLA Fielding School of Public Health (IRB#15–000333 for those in Boende, Kabondo-Dianda, and Yambuku; IRB#16–001346 for those in Kikwit) and the Kinshasa School of Public Health (ESP/CE/038/2015 for those in Boende, Kabondo-Dianda, and Yambuku; IRB#ESP/CE/022/2017 for those in Kikwit). All participants provided written informed consent and had the right to refuse participation at any time.

Study design

As a part of a larger study examining EBOV serology across DRC, we conducted a cross-sectional substudy using data collected between September 2015 –August 2017. We enrolled 1,937 individuals across four sites around the country: 736 residents of Boende Health Zone, 41 residents of Kabondo-Dianda Health Zone, 424 residents of Yambuku Health Zone, and 736 residents of either North or South Kikwit Health Zones (Fig 1). Demographically, these regions are distinct in that Boende and Yambuku are in the north western part of the country, and considered very rural and heavily forested, Kikwit is semi-urban located in the central southwestern portion of the country and is considered forest-savannah mosaic and Kabando Dianda is semi-urban and located in the southeastern portion of the country also considered forest-savannah mosaic. Three of these areas have experienced previous EVD outbreaks (Boende, Kikwit, and Yambuku), while one has not had any known outbreak of EVD (Kabondo-Dianda). Participants were included regardless of self-reported presence at an EVD outbreak or self-reported exposure to an EVD case or if they lived in an area with a history of EVD.

Fig 1. Map of the Democratic Republic of the Congo with health zones of study sites highlighted in red.

Base layer from

Convenience samples were gathered at all four sites, targeting individuals who work in a healthcare setting, including traditional healers and pastors involved in healthcare activities, based on the WHO definition of healthcare worker [29]. However, eligibility was not restricted to individuals in those professions. Participants were eligible for the study if they were over 18 years of age and healthy, defined as no fever (<38°C) or other self-reported acute illness at the time of enrollment. At each site, a survey was administered and blood samples were collected from consenting participants at a single visit. Blood samples were obtained from all consenting participants by venipuncture using red-top Vacutainer tubes for serum collection (BD Biosciences).

Among the eligible 1,937 individuals, 1,407 (72.6%) had available serological results. Serological results were obtained according to assay availability at time of collection. Of these 1,407 participants, 41 had previously received an EVD diagnosis (2.9%) and were excluded from the analysis. None of the individuals in our study had received any Ebola vaccine at the time of data collection. Participants were included regardless of self-reported presence at an EVD outbreak site or self-reported exposure to an EVD case, or if they lived in an area with a history of EVD. Thus, 1,366 individuals were included in this analysis. Subsections of this data have been used in previous publications[30,31].

Survey measurements

Our explanatory variables were zoonotic exposures, which included both animal exposures and activities that could lead to contact with an infected animal. Surveys were conducted by trained interviewers in the participant’s preferred local language (French, Lingala, Swahili, or Kikongo), and collected data on sociodemographic and epidemiologic characteristics, including potential zoonotic exposures to EBOV. Exposures were first assessed by determining if the participant had ever performed each activity or had contact with the specified animal. Those indicating a previous exposure were then asked if that exposure had occurred in the past month.

Laboratory measurements

EBOV GP IgG seroreactivity was the primary outcome, determined using the manufacturer’s protocol for Zaire Ebola Virus Glycoprotein (EBOV GP antigen) ELISA kits (Alpha Diagnostic International, Inc.) at the National Institute for Biomedical research in Kinshasa, DRC. Details on methodology have been described elsewhere [31,32]. Samples were run in duplicate, and the average of the two results was used to determine seroreactivity. The manufacturer’s classification places any sample above 1.0 units/mL as reactive. However, this analysis considers results greater than 4.7 to be seroreactive in order to be conservative. This more stringent reactivity criteria has been shown to increase sensitivity to 96.7% and specificity to 97.7% for this assay [18]. Increased cutoffs for all EBOV antibody assays are commonly applied to Congolese cohorts, which are thought to exhibit high levels of assay cross-reactivity due to high rates of infectious disease within these populations [18,31,32]. No confirmatory assay or virus neutralization was used to rule out cross-reactivity as a source of positive results in reactive samples.


Descriptive statistics on sample characteristics and demographics were obtained for the full sample. Crude odds ratios (OR) for EBOV seroreactivity were then obtained for select sample characteristics. Adjusted ORs describing the associations between EBOV seroreactivity and ever having a zoonotic exposure were produced using logistic regression. To further explore the role of recent or frequent exposures, adjusted ORs were then obtained for each exposure in the past month among individuals who had ever had that exposure. This addition was made to distinguish possible risk from any exposure over a lifetime versus recent or frequent exposures, as it is possible that there may be a different biological response in both situations. For all ORs, a 95% confidence interval is provided; A 95% CI that did not cross the null value of 1.00 was considered to be evidence of an association. No corrections were made for multiple comparisons.

Adjusted ORs were obtained through multivariable logistic regressions that considered age as a continuous variable, sex, healthcare worker status, contact with an EVD case, and study site as confounders based on a priori assumptions depicted in our hypothesized directed acyclic graph (DAG) (Fig 2). Logistic regression was determined to be appropriate as there was independence of observations, linearity in the logit for continuous independent variables, an absence of multicollinearity, and an absence of influential outliers.[33] Additional analyses examining effect modification by study location were performed through the addition of interaction terms for each exposure by site to the specified models. A sensitivity analysis excluding all those with self-reported exposure to an EVD case was conducted. All statistical analyses were carried out using SAS software, version 9.4 (SAS Institute, Cary, NC).

Fig 2. Directed Acyclic Graph (DAG) depicting hypothesized causal structure of zoonotic exposure and Ebola virus seroreactivity among individuals with no history of Ebola Virus Disease.

Age and location are assumed to be simple confounders, influencing both presence of zoonotic exposures and Ebola virus (EBOV) seroreactivity. Sex is assumed to influence both zoonotic exposure and seroreactivity, and to be associated with healthcare worker status, as women were more likely to be healthcare workers in our sample. Being a healthcare worker is assumed to influence EBOV seroreactivity, due to potential for unknown exposures to Ebola Virus Disease (EVD) cases, and zoonotic exposure, due to relatively low contact with animals and forest when compared to non-healthcare workers. Working in healthcare is assumed to raise an individual’s risk of having known contact with an EVD case, which in turn is assumed to increase an individual’s risk of EBOV seroreactivity. Our DAG includes a residual confounding path through an unknown ancestor of zoonotic exposure and contact with an EVD case. In the analysis we controlled for contact with an EVD case as a proxy for this unknown confounder, which led to a more conservative analysis.


Overall, 8.3% of our sample (113/1366) was seroreactive to EBOV GP. Most of our sample was male (62.6%) and between 26 to 59 years of age (78.7%). Nearly a third of our sample had finished primary school, and another third had completed secondary school as their highest level of education. Three quarters of participants (76.6%) were married or cohabitating with a partner, with only 15.4% reporting being single. Over two-thirds of the sample reported working in a healthcare setting, including traditional healers and pastors. A large majority (77.8%) reported that they had never had contact with a confirmed, suspected, or probable EVD case (Table 1).

Table 1. Sample Characteristics among 1,366 individuals from Boende, Kabondo-Dianda, Yambuku, and Kikwit in the Democratic Republic of the Congo, 2015–2017.

In our sample, males were more commonly seroreactive. Meanwhile, a lower proportion of healthcare workers were seroreactive than non-healthcare workers. Proportionally more participants from Yambuku and Boende were seroreactive when compared to participants from Kabondo-Dianda, though not significantly. Both age and self-reported contact with a confirmed, suspected, or probably EVD case did not show meaningful associations with seroreactivity (Table 2).

Table 2. Seroreactivity to Ebola virus GP by select characteristics among 1,366 individuals from Boende, Kabondo-Dianda, Yambuku, and Kikwit in the Democratic Republic of the Congo, 2015–2017.

Among the animal exposures, those who had ever come into contact with a bat had 1.64 times (95% CI 1.06, 2.54) the odds of EBOV seroreactivity as those who had never had contact with a bat, holding confounders constant. Any contact with rodents was also associated with an increased odds of EBOV seroreactivity, as was eating non-human primate meat (Table 3).

Table 3. Adjusted odds ratios of seroreactivity to Ebola virus GP for various zoonotic exposures among individuals in the Democratic Republic of the Congo, 2015–2017.

Among those who indicated they had ever had contact with a rodent, having contact with rodents in the past month was associated with a decrease in EBOV GP reactivity (Table 4).

Table 4. Adjusted odds ratios of seroreactivity to Ebola virus GP for recent zoonotic exposures among individuals who have ever experienced each respective exposure in the Democratic Republic of the Congo, 2015–2017.

There was no evidence of interaction in our models assessing effect modification of exposure effects by study site; all interaction terms were non-significant. Additionally, there were no significant or meaningful changes from the presented results when the analysis was restricted to those with no known EVD exposure.


In our study, we found associations between EBOV seroreactivity without diagnosed EVD and two zoonotic exposures previously tied to EVD outbreaks: contact with bats and eating non-human primate meat. These findings expand our understanding of zoonotic exposure, EBOV seroreactivity, and EVD risk within the DRC. Namely, it is possible that some exposures to bats or eating non-human primate meat in this region might result in EBOV seroreactivity without diagnosed EVD. If so, there may be contributing factors, such as environmental influences, chiropteran physiological conditions [3436], or reduced virus viability in meat cooked for consumption[37], among these less consequential zoonotic interactions which result in low-consequence EBOV exposure. Future research should work to identify and evaluate potential variations in virus viability and viral load transfer associated with different types of human-bat interactions and non-human primate consumption to further describe EBOV zoonotic exposure and resulting risk of seroreactivity or EVD diagnosis.

Additionally, ever having rodent contact was associated with increased odds of seroreactivity in our sample, though rodent contact has never been implicated as a risk factor diagnosed EVD. The association between rodents and EBOV seroreactivity is plausible, given that the EBOV genome has been sequenced in samples from rodents in central Africa [8,38]. In contrast, those who had contact with rodents in the past month had significantly lower odds of seroreactivity than individuals who had contact with rodents that did not report contact in the last month. It is unclear why recent contact with rodents would have a protective effect. Future research should attempt to further understand the role of rodents in EBOV epidemiology.

In addition to zoonotic exposure findings, this analysis also describes EBOV seroreactivity across the DRC. A substantial portion of our sample, 8.3%, was seroreactive for antibodies to EBOV GP, with site specific seroreactivity in our sample ranging from 3.4% to 13.7%. Published EBOV seroprevalence estimates are greatly heterogeneous, varying widely depending on definition of reactivity, sample selection methods, and study area qualities such as ecology or rural versus urban settings [11]. Nevertheless, the estimates found in this study fall in line with previously reported seroprevalence estimates for comparable populations in sub-Saharan Africa [1117,39]. Demographic predictors of EBOV seroreactivity in our sample were male gender and not being a healthcare worker. These findings contribute to our understanding of EBOV epidemiology, as there has been limited research into the demographic predictors of seroprevalence in the absence of EVD diagnosis for comparison. Other studies have identified higher seroprevalence in adults and women in similar cohorts [13,16,24].

While these findings are informative, this study has limitations. Our explanatory variables are self-reported and therefore may be subject to misclassification. Limitations of recall were addressed in part through the collection of reported exposure over the past month in addition to ever experiencing an exposure, with the assumption that recall over the previous month would be more accurate. Additionally, classification of serological results into a binary variable may not have accurately captured zoonotic exposure history. Our assay cut-off value is based on the manufacturer’s classification and previous research [31,32], and is reported to have 96.7% sensitivity and 97.7% specificity [18]. Given the cross-sectional study design, our findings are strictly associational. Further research, including collection and analysis of longitudinal data, will be vital to establishing temporality and uncovering any causal relationships behind the associations observed here.

This analysis provides new insights into the epidemiology of EBOV infection and disease in an endemic region of Central Africa. Our study is the first to evaluate associations between various zoonotic exposures and seroreactivity to EBOV GP in a large, geographically diverse sample of Congolese individuals who have never been diagnosed with EVD. This analysis demonstrates that some zoonotic exposures that have previously been implicated in EVD outbreaks can also be linked to EBOV GP seroreactivity in the absence of EVD diagnoses. Future investigations should seek to clarify the relationships between zoonotic exposures, seroreactivity, asymptomatic infection, and EVD.

Supporting information

S1 Codebook. Variable names, corresponding survey questions, and values for the dataset used in this analysis.



  1. 1. Feldmann H, Sprecher A, Geisbert TW. Ebola. Campion EW, editor. N Engl J Med. Massachusetts Medical Society; 2020;382: 1832–1842. pmid:32441897
  2. 2. Genus: Ebolavirus—Filoviridae—Negative-sense RNA Viruses—ICTV [Internet]. [cited 21 Feb 2021]. Available:
  3. 3. Feldmann H, Geisbert TW. Ebola haemorrhagic fever. Lancet. Elsevier; 2011;377: 849–862. pmid:21084112
  4. 4. Leroy EM,Epelboin A, Mondonge V, Pourrut X, Gonzalez JP, Muyembe-Tamfum JJ, et al. Human ebola outbreak resulting from direct exposure to fruit bats in Luebo, Democratic Republic of Congo, 2007. Vector-Borne Zoonotic Dis. Mary Ann Liebert, Inc. 140 Huguenot Street, 3rd Floor New Rochelle, NY 10801 USA; 2009;9: 723–728.
  5. 5. Groseth A, Feldmann H, Strong JE. The ecology of Ebola virus. Trends in Microbiology. Elsevier Current Trends; 2007. pp. 408–416. pmid:17698361
  6. 6. Marí Saéz A, Weiss S, Nowak K, Lapeyre V, Zimmermann F, Düx A, et al. Investigating the zoonotic origin of the West African Ebola epidemic. EMBO Mol Med. EMBO; 2015;7: 17–23. pmid:25550396
  7. 7. Transmission | Ebola Hemorrhagic Fever | CDC [Internet]. [cited 6 Jul 2020]. Available:
  8. 8. Pourrut X, Kumulungui B, Wittmann T, Moussavou G, Délicat A, Yaba P, et al. The natural history of Ebola virus in Africa. Microbes and Infection. Elsevier Masson SAS; 2005. pp. 1005–1014. pmid:16002313
  9. 9. Vetter P, Fischer WA, Schibler M, Jacobs M, Bausch DG, Kaiser L. Ebola Virus Shedding and Transmission: Review of Current Evidence. J Infect Dis. Oxford University Press; 2016;214: S177–S184. pmid:27443613
  10. 10. Glennon EE, Jephcott FL, Restif O, Wood JLN. Estimating undetected ebola spillovers. PLoS Negl Trop Dis. Public Library of Science; 2019;13: e0007428. pmid:31194734
  11. 11. Bower H, Glynn JR. A systematic review and meta-analysis of seroprevalence surveys of ebolavirus infection. Scientific Data. Nature Publishing Groups; 2017. pp. 1–9. pmid:28140390
  12. 12. Formella M, Gatherer D. The serology of Ebolavirus—A wider geographical range, a wider genus of viruses or a wider range of virulence? Journal of General Virology. Microbiology Society; 2016. pp. 3120–3130. pmid:27902321
  13. 13. Johnson ED, Gonzalez JP, Georges A. Haemorrhagic fever virus activity in equatorial Africa: distribution and prevalence of filovirus reactive antibody in the Central African Republic. Trans R Soc Trop Med Hyg. Oxford Academic; 1993;87: 530–535. pmid:8266402
  14. 14. Bouree P, Bergmann JF. Ebola virus infection in man: A serological and epidemiological survey in the Cameroons. Am J Trop Med Hyg. The American Society of Tropical Medicine and Hygiene; 1983;32: 1465–1466. pmid:6650749
  15. 15. Johnson ED, Gonzalez JP, Georges A. Filovirus activity among selected ethnic groups inhabiting the tropical forest of equatorial Africa. Trans R Soc Trop Med Hyg. Oxford Academic; 1993;87: 536–538. pmid:8266403
  16. 16. Nkoghe D, Padilla C, Becquart P, Wauquier N, Moussavou G, Akué JP, et al. Risk factors for zaire ebolavirus-specific IgG in rural gabonese populations. J Infect Dis. J Infect Dis; 2011;204. pmid:21987749
  17. 17. Mulangu S, Alfonso VH, Hoff NA, Doshi RH, Mulembakani P, Kisalu NK, et al. Serologic Evidence of Ebolavirus Infection in a Population With No History of Outbreaks in the Democratic Republic of the Congo. Ebolavirus Seroprevalence • JID. 2018: 217. pmid:29329455
  18. 18. Richardson ET, Kelly JD, Barrie MB, Mesman AW, Karku S, Quiwa K, et al. Minimally Symptomatic Infection in an Ebola ‘Hotspot’: A Cross-Sectional Serosurvey. Bausch DG, editor. Public Library of Science; 2016;10: e0005087. pmid:27846221
  19. 19. Forbes KM, Webala PW, Jääskeläinen AJ, Abdurahman S, Ogola J, Masika MM, et al. Bombali virus in mops condylurus bat, kenya. Emerg Infect Dis. Centers for Disease Control and Prevention (CDC); 2019;25: 955–957. pmid:31002301
  20. 20. Goldstein T, Anthony SJ, Gbakima A, Bird BH, Bangura J, Tremeau-Bravard A, et al. The discovery of Bombali virus adds further support for bats as hosts of ebolaviruses. Nat Microbiol. Nature Publishing Group; 2018;3: 1084–1089. pmid:30150734
  21. 21. Warfield KL, Bosio CM, Welcher BC, Deal EM, Mohamadzadeh M, Schmaljohn A, et al. Ebola virus-like particles protect from lethal Ebola virus infection. Proc Natl Acad Sci U S A. National Academy of Sciences; 2003;100: 15889–15894. pmid:14673108
  22. 22. Henao-Restrepo AM, Camacho A, Longini IM, Watson CH, Edmunds WJ, Egger M, et al. Efficacy and effectiveness of an rVSV-vectored vaccine in preventing Ebola virus disease: final results from the Guinea ring vaccination, open-label, cluster-randomised trial (Ebola Ça Suffit!). Lancet. Lancet Publishing Group; 2017;389: 505–518. pmid:28017403
  23. 23. Kennedy SB, Bolay F, Kieh M, Grandits G, Badio M, Ballou R, et al. Phase 2 Placebo-Controlled Trial of Two Vaccines to Prevent Ebola in Liberia. N Engl J Med. Massachussetts Medical Society; 2017;377: 1438–1447. pmid:29020589
  24. 24. Mbala P, Baguelin M, Ngay I, Rosello A, Mulembakani P, Demiris N, et al. Evaluating the frequency of asymptomatic Ebola virus infection. Philos Trans R Soc B Biol Sci. Royal Society Publishing; 2017;372: 20160303. pmid:28396474
  25. 25. Busico KM, Marshall KL, Ksiazek TG, Roels TH, Fleerackers Y, Feldmann H, et al. Prevalence of IgG Antibodies to Ebola Virus in Individuals during an Ebola Outbreak, Democratic Republic of the Congo, 1995. J Infect Dis. University of Chicago Press; 1999;179: S102–S107. pmid:9988172
  26. 26. Moyen N, Thirion L, Emmerich P, Dzia-Lepfoundzou A, Richet H, Boehmann Y, et al. Risk Factors Associated with Ebola and Marburg Viruses Seroprevalence in Blood Donors in the Republic of Congo. Kasper M, editor. PLoS Negl Trop Dis. Public Library of Science; 2015;9: e0003833. pmid:26047124
  27. 27. Becquart P, Wauquier N, Mahlakõiv T, Nkoghe D, Padilla C, Souris M, et al. High Prevalence of Both Humoral and Cellular Immunity to Zaire ebolavirus among Rural Populations in Gabon. Montgomery JM, editor. PLoS One. Public Library of Science (PLoS); 2010;5: e9126. pmid:20161740
  28. 28. Centers for Disease Control and Prevention (CDC). Years of Ebola Virus Disease Outbreaks [Internet]. [cited 28 Jan 2020]. Available:
  29. 29. Walker D, Mosqueira NR, Penny ME, Lanata CF, Clark AD, Sanderson CFB, et al. Cost effectiveness of measles eradication Final Report. Bull World Health Organ. 2015;528: 1–36. pmid:18419391
  30. 30. Doshi RH, Hoff NA, Bratcher A, Mukadi P, Gadoth A, Nicholson BP, et al. Risk Factors for Ebola Exposure in Health Care Workers in Boende, Tshuapa Province, Democratic Republic of the Congo. J Infect Dis. Oxford University Press (OUP); 2020; pmid:33269402
  31. 31. Hoff NA, Mukadi P, Doshi RH, Bramble MS, Lu K, Gadoth A, et al. Serologic Markers for Ebolavirus Among Healthcare Workers in the Democratic Republic of the Congo. J Infect Dis EBOV Serol Markers Among HCWs DRC • JID. 2019: 517. pmid:30239838
  32. 32. Rimoin AW, Lu K, Bramble MS, Steffen I, Doshi RH, Hoff NA, et al. Ebola Virus Neutralizing Antibodies Detectable in Survivors of the Yambuku, Zaire Outbreak 40 Years after Infection. pmid:29253164
  33. 33. Stoltzfus JC. Logistic Regression: A Brief Primer. Acad Emerg Med. John Wiley & Sons, Ltd; 2011;18: 1099–1104. pmid:21996075
  34. 34. O’Shea TJ, Cryan PM, Cunningham AA, Fooks AR, Hayman DTS, Luis AD, et al. Bat flight and zoonotic viruses. Emerg Infect Dis. Centers for Disease Control and Prevention (CDC); 2014;20: 741–745. pmid:24750692
  35. 35. Brook CE, Dobson AP. Bats as “special” reservoirs for emerging zoonotic pathogens. Trends in Microbiology. Elsevier Ltd; 2015. pp. 172–180. pmid:25572882
  36. 36. Banerjee A, Baker ML, Kulcsar K, Misra V, Plowright R, Mossman K. Novel Insights Into Immune Systems of Bats [Internet]. Frontiers in Immunology. Frontiers Media S.A.; 2020. p. 26. pmid:32117225
  37. 37. Haddock E, Feldmann F, Feldmann H. Effective chemical inactivation of Ebola virus. Emerg Infect Dis. Centers for Disease Control and Prevention (CDC); 2016;22: 1292–1294. pmid:27070504
  38. 38. Morvan JM, Nakouné E, Deubel V, Colyn M. Forest Ecosystems and Ebola. Bull la Soc Pathol Exot 93. 2000; 172–175. pmid:11030051
  39. 39. Dean NE, Halloran ME, Yang Y, Longini IM. Transmissibility and Pathogenicity of Ebola Virus: A Systematic Review and Meta-analysis of Household Secondary Attack Rate and Asymptomatic Infection. 2016; pmid:26932131