The Multiple Origins of Ebola Disease Outbreaks

Abstract

Background

The origins of Ebola disease outbreaks remain enigmatic. Historically outbreaks have been attributed to spillover events from wildlife. However, recent data suggest that some outbreaks may originate from human-to-human transmission of prior outbreak strains instead of spillover. Clarifying the origins of Ebola disease outbreaks could improve detection and mitigation of future outbreaks.

Methods

We reviewed the origins of all Ebola disease outbreaks from 1976 to 2022 to analyze the earliest cases and characteristics of each outbreak. The epidemiology and phylogenetic relationships of outbreak strains were used to further identify the likely source of each outbreak.

Results

From 1976 to 2022 there were 35 Ebola disease outbreaks with 48 primary/index cases. While the majority of outbreaks were associated with wildlife spillover, resurgence of human-to-human transmission could account for roughly a quarter of outbreaks caused by Ebola virus. Larger outbreaks were more likely to lead to possible resurgence, and nosocomial transmission was associated with the majority of outbreaks.

Conclusions

While spillover from wildlife has been a source for many Ebola disease outbreaks, multiple outbreaks may have originated from flare-ups of prior outbreak strains. Improving access to diagnostics as well as identifying groups at risk for resurgence of ebolaviruses will be crucial to preventing future outbreaks.

Sudan virus, Ebola virus, ebolavirus, epidemiology, filovirus, nosocomial, outbreak, phylogeny, spillover, transmission

Despite global advancements in vaccines, diagnostics, and therapeutics for ebolaviruses, devastating outbreaks of Ebola disease (EBOD) continue to occur across Africa. Understanding where and when ebolaviruses emerge in human populations could enhance preparedness for future outbreaks. Since ebolaviruses were first detected in humans in 1976, it has been widely thought that all EBOD outbreaks originate from spillover of ebolaviruses from wildlife into humans [1], but recent outbreaks have changed our understanding of the emergence of ebolaviruses [2]. Therefore, reassessing historic EBOD outbreaks through modern epidemiologic and phylogenetic lenses could reveal patterns for preventing further outbreaks.

Ebolaviruses are members of the family Filoviridae, and viruses from 4 Ebolavirus species are known to cause disease in humans (Table 1). We now generally refer to the disease caused by all ebolaviruses in humans as EBOD, ever since the International Classification of Disease for filoviruses (ICD-11) in 2019 [3]. These different ebolaviruses likely have distinct reservoir hosts and ecological niches [4]. Ebola virus (EBOV) RNA has been detected in 3 fruit bat species [5], whereas Sudan virus (SUDV), Bundibugyo virus (BDBV), and Taï Forest virus (TAFV) have not been detected in any potential wildlife reservoirs. While no ebolavirus that is known to be pathogenic in humans has ever been isolated from a putative reservoir host, multiple bat species have been considered possible reservoir hosts and other mammals have been deemed likely incidental hosts [4]. Spillover from these reservoir hosts or incidental hosts has been historically determined to be the likely source of EBOD outbreaks.

Table 1.

Ebolaviruses Known to Cause Disease in Humans

Ebolavirus	Ebolavirus Species	Ebola Disease (EBOD)
Ebola virus (EBOV)	Zaire ebolavirus	Ebola virus disease (EVD)
Sudan virus (SUDV)	Sudan ebolavirus	Sudan virus disease (SVD)
Bundibugyo virus (BDBV)	Bundibugyo ebolavirus	Bundibugyo virus disease (BVD)
Taï Forest virus (TAFV)	Taï Forest ebolavirus	Other specified Ebola disease

However, recent findings have changed how we think about the emergence of ebolaviruses in humans. During the West Africa Ebola virus disease (EVD) epidemic, a male survivor infected his female partner 179 days after being diagnosed with EVD, likely through sexual transmission [6]. Towards the end of the West Africa epidemic, a new cluster of EVD cases were found in Guinea and Liberia, which were linked to sexual transmission from a survivor that occurred about 470 days after he had symptoms [7]. Subsequent cohort studies have demonstrated prolonged EBOV persistence in the semen of EVD survivors, further indicating risk for delayed sexual transmission [8]. EBOV was also found to persist in aqueous humor as well as cerebrospinal fluid in survivors who developed uveitis and meningoencephalitis after their initial convalescence [9, 10]. Therefore, the testes, eyes, central nervous system, as well as placenta and potentially breast milk were deemed immune-privileged sites where EBOV can persist and evade the immune system [11].

Another potential source for EVD outbreaks was identified after a survivor had relapse of infection 149 days after leaving an Ebola treatment unit, which caused a new chain of transmission [12]. Although risk factors for relapse of infection remain unknown, it was noted that the survivor had received antibody-based therapies during their initial treatment [12]. Researchers also found EBOV persistence in the central nervous system and recrudescence of infection in nonhuman primates who were treated with monoclonal antibody-based therapies, leading to concerns such therapies may contribute to relapse of EVD in humans [13].

Recently, phylogenetic analyses of EBOV genomes from the 2021 outbreak in Guinea showed that the outbreak strain was genetically similar to the Makona strain from the West Africa epidemic from 2013 to 2016 [2]. It was postulated that this outbreak may have originated from delayed sexual transmission, reactivation of latent infection, or unrecognized chains of transmission among humans instead of spillover from wildlife. Hereon we refer to outbreaks originating from such sources as flare-ups or resurgence events of human-to-human transmission. Unpublished sequences from the 2020 Equateur, 2021 North Kivu, and 2022 North Kivu outbreaks were also shown to be similar to sequences from prior outbreaks in those regions, and it is thought that these outbreaks were resurgence events, while the 2022 Equateur outbreak has been recognized to be likely a separate spillover event [14–18].

Investigating the spatiotemporal and phylogenetic relationships of prior EBOD outbreaks could further elucidate how outbreaks originate. While reviewing the exposure history and contexts of initial cases could help identify the sources of EBOD outbreaks, additional insights can be gained through phylogenetic analyses. During initial EVD outbreaks researchers found little genetic variation within the same epidemic chain and that the glycoprotein (GP) gene was conserved [19–22]. Rapid sequencing in the field during the West Africa EVD epidemic also initially revealed that EBOV sequences from human-to-human transmission had few mutations that were mostly synonymous or in noncoding regions [23]. However, as the outbreak grew and became the largest and longest EVD epidemic in history, multiple EBOV lineages were subsequently found, and a higher amount of genetic diversity was found in the GP gene [24]. The mean EBOV evolutionary rate from human-to-human transmission during this outbreak was estimated at approximately 1.2 × 10⁻³ nucleotide substitutions/site/year [24]. The EBOV genome is approximately 19 kb so this is roughly 23 substitutions per year.

While less is known about the evolutionary rate during persistent EBOV infection, researchers found low genetic divergence in reemerged strains from persistently infected individuals [25]. In samples from acutely infected individuals, researchers found an evolutionary rate of 0.96 × 10⁻³ nucleotide substitutions/site/year and similar or reduced evolutionary rates in semen, aqueous humor, and urine during persistent infection [26]. In fact, the virus from the EVD survivor who infected his partner through sexual transmission 470 days after symptom onset had an evolutionary rate of 0.19 × 10⁻³ substitutions/site/year, or approximately 4 substitutions per year [7]. Therefore, these evolutionary rates provide rough estimates of the expected variation one might see from very prolonged human-to-human transmission versus relapse of persistent infection as the origin of a new outbreak.

The evolutionary rate in the natural reservoir host for EBOV remains unknown. However, there was found to be significant sequence variation between EBOV strains from separate spillover events from wildlife [20, 22]. Because spillover events from wildlife occur with viruses from different populations, host species, and locations, it is expected that there would be greater genetic variation between viruses from different spillover events. Therefore, considering both epidemiologic and phylogenetic patterns is necessary to characterize the origins of EBOD outbreaks.

Knowing how EBOD outbreaks originate and whether they are more likely to occur from spillover or resurgence of human-to-human transmission could change the way that we prepare for future outbreaks. Therefore, the goal of this study is to identify the origins of all EBOD outbreaks from 1976 to 2022. Through comparing epidemiologic and phylogenetic relationships, we aim to determine the origins and characteristics of each EBOD outbreak. Additionally, we aim to analyze if the detection of ebolaviruses and EBOD outbreaks has changed, and what gaps remain in detecting and preventing future outbreaks.

METHODS

We investigated the epidemiology of all known EBOD outbreaks from 1976 to 2022. The US Centers for Disease Control and Prevention's chronology of EBOD outbreaks was used as a guide to determine outbreaks (excluding laboratory-acquired and exported cases) [27]. We then reviewed the primary literature as well as World Health Organization (WHO) and international outbreak reports. For each outbreak, we sought to determine the earliest infected individual, known as the primary case. If this was not possible, we identified the first case recognized by health care authorities, the index case. We labeled some individuals as both primary and index cases if they fulfilled both of these criteria. In certain circumstances, we identified multiple primary/index cases from the same outbreak (if multiple individuals were exposed to the same animal or if there was an epizootic event). We categorized each primary/index case as confirmed, probable, or suspected according to WHO Integrated Disease Surveillance and Response EVD case definitions [28]. We created a dataset of all primary/index cases from 1976 to 2022 along with their associated locations, dates of onset, demographics, suspected sources, and diagnostics [29]. We also made a dataset of all EBOD outbreaks from 1976 to 2022, including locations, most likely origins, cases, deaths, case fatality rates, and factors associated with transmission [29].

To determine whether EBOD outbreaks originated from spillover or resurgence of human-to-human transmission, we conducted a phylogenetic analysis with representative sequences from each outbreak (Supplementary Table 1). We then compared epidemiological contexts with phylogenetic relationships to determine the most likely origin of each outbreak. We aligned 26 ebolavirus whole-genome sequences from 1976 to 2021 using MAFFT [30] with the default parameters in Geneious Prime version 2023.0 (https://www.geneious.com) and constructed a maximum likelihood tree from PhyML [31] using the GTR nucleotide substitution model with 100 bootstraps in Geneious Prime. Given representative sequence similarities, we also performed separate alignments of multiple EBOV whole-genome sequences from the 1976 and 1977 outbreaks, as well as the 2007 and 2008 outbreaks in the Democratic Republic of the Congo (Supplementary Table 2). We used a distance matrix to compare the number of nucleotide differences between aligned sequences. We considered the previously described EBOV evolutionary rates in humans when comparing nucleotide differences, with 23 nucleotide substitutions/year as the upper boundary in variation one might expect from an outbreak originating from sustained human-to-human transmission. Therefore, genome sequences from consecutive years that had alignment differences of less than 23 nucleotide substitutions and no epidemiologic evidence of spillover were considered to potentially originate from resurgence of human transmission.

We created a map of the initial locations of each outbreak using QGIS (http://www.qgis.org). We used the point coordinates from prior analyses [4, 32], as well as where initial cases were treated for the 2017–2022 outbreaks. For outbreaks that were possibly from resurgence events, we calculated the number of days from the first case of a possible flare-up to the date that the previous zoonotic outbreak was declared to be over. We also performed a logistic regression analysis to determine the association between the number of cases in an outbreak and whether that outbreak strain was associated with a subsequent flare-up. We conducted statistical analyses using GraphPad Prism 9 (https://www.graphpad.com/).

RESULTS

We identified a total of 35 EBOD outbreaks (Table 2); 24 of 35 (68.5%) of EBOD outbreaks were due to EVD. If a primary case seemed more likely due to spillover and there was an epidemiologic association with particular wildlife exposure, we included these wildlife species as suspected sources (Supplementary Table 3). Only outbreaks of EVD were recognized as potentially originating from resurgence of prior outbreak strains. TAFV, BDBV, and SUDV have not been associated with transmission via relapse or epizootic events, so outbreaks from these viruses were characterized as separate spillover events, likely from different wildlife hosts. The potential origins of EBOD outbreaks included a single spillover event from a wildlife host, multiple spillover events from a wildlife host, epizootic events with spillover from multiple different wildlife hosts, and resurgence of human-to-human transmission/flare-ups from previous outbreaks. At least one EVD outbreak may have originated from both a spillover event and a resurgence event. Seven of 24 (25.9%) of EVD outbreaks were found to possibly originate from resurgence events, 5 of which occurred after the 2013–2016 epidemic. The initial locations and origins for each outbreak are shown in Figure 1.

Figure 1.

Locations of Ebolavirus emergence 1976–2022. The locations and potential origins of each Ebola disease outbreak are shown. Potential resurgence of human-to-human transmission generally occurred in close proximity to the locations of prior outbreaks.

Table 2.

Ebola Disease Outbreak Origins 1976–2022

Year	Country	Initial Location	Virus	Origin, Suspected Source	Cases (CFR %)
1976	South Sudan	Nzara	SUDV	Spillover	284 (53)
1976	DRC	Yambuku	EBOV	Spillover	318 (88)
1977	DRC	Bonduni	EBOV	1976 Yambuku outbreaka	1 (100)
1979	South Sudan	Nzara	SUDV	Spillover	34 (64)
1994	Gabon	Mékouka	EBOV	Spillover	52 (60)
1994	Cote d’Ivoire	Tai National Park	TAFV	Spillover	1 (0)
1995	DRC	Kikwit	EBOV	Spillover	315 (81)
1996	Gabon	Mayibout	EBOV	Spillover	31 (67)
1996	Gabon	Booué	EBOV	Spillover	60 (75)
2000–2001	Uganda	Rwot-Obillo	SUDV	Spillover	425 (53)
2001–2002	Gabon	Mendemba	EBOV	Spillover	65 (82)
2001	ROC	Olloba	EBOV	Spillover	59 (75)
2003	ROC	Yembelangoye; Mvoula	EBOV	Spillover	143 (90)
2003	ROC	Mbandza	EBOV	Spillover	35 (83)
2004	South Sudan	Yambio	SUDV	Spillover	17 (41)
2005	ROC	Parc d’Odzala	EBOV	Spillover	12 (83)
2007	DRC	Luebo	EBOV	Spillover	264 (71)
2007	Uganda	Kabango	BDBV	Spillover	149 (25)
2008–2009	DRC	Luebo	EBOV	2007 Luebo outbreaka	32 (47)
2011	Uganda	Nakisamata	SUDV	Spillover	1 (1)
2012	DRC	Isiro	BDBV	Spillover	62 (55)
2012	Uganda	Nyanswiga	SUDV	Spillover	24 (71)
2012	Uganda	Luwero	SUDV	Spillover	7 (57)
2013–2016	Guinea	Meliandou	EBOV	Spillover	28 656 (40)
2014	DRC	Ikanamongo	EBOV	Spillover	69 (71)
2017	DRC	Likati	EBOV	Spillover	8 (50)
2018	DRC	Ikoko-Impenge	EBOV	Spillover	54 (61)
2018–2020	DRC	Mangina	EBOV	Spillover	3470 (66)
2020	DRC	Mbandaka	EBOV	Spillover and 2018 Equateur outbreaka	130 (42)
2021	DRC	Butembo	EBOV	2018–2020 North Kivu outbreaka	12 (50)
2021	Guinea	Gouéké	EBOV	2013–2016 epidemica	23 (52)
2021	DRC	Butsili	EBOV	2018–2020 North Kivu outbreaka	11 (55)
2022	DRC	Mbandaka	EBOV	Spillover	5 (5)
2022	DRC	Beni Health Zone	EBOV	2018–2020 North Kivu outbreaka	1 (1)
2022	Uganda	Mubende	SUDV	Spillover	164 (47)

Abbreviations: CFR, case fatality rate; DRC, Republic of the Congo; EBOV, Ebola virus; ROC, Republic of the Congo; SUDV, Sudan virus; TAFV, Taï Forest virus.

^aPotential resurgence.

Forty-eight primary/index cases were identified for the 35 EBOD outbreaks. The mean age of these cases was 32 years (range, 2–65 years). Thirty of 48 (62.5%) primary/index cases had available demographic information: of these thirty, 9 (30%) were hunters, 6 (20%) were children, 2 (6.67%) were health care workers, and 3 (10%) were pregnant. One primary/index case had a husband who was an EVD survivor. Full case details are available in our Index and Primary Cases dataset [29] and demographic details are included in Supplementary Table 4. The initial diagnoses for index cases included: malaria, unknown viral hemorrhagic fever, yellow fever, dysentery/gastroenteritis, and salmonellosis. Thirty-six of 48 (75%) were diagnosed only via clinical review of signs/symptoms and epidemiological history, whereas the other cases were diagnosed via reverse transcription polymerase chain reaction (RT-PCR), serology, and/or virus isolation. From 2021 to 2022 all 6 index cases were confirmed via RT-PCR.

Thirty-one of 35 (88.6%) EBOD outbreaks involved multiple cases, for which there were documented chains of transmission. Of these 31 outbreaks, 25 (80.6%) had reported nosocomial transmission, 14 (45.1%) had transmission from burial practices, and 11 (35.5%) were associated with bushmeat contact. An average of 637 days (range, 222–1690 days) occurred between the end of an original zoonotic outbreak and a potential subsequent flare-up. A logistic regression analysis showed that the odds of an EVD outbreak being associated with a future flare-up increased as the total cases increased (odds ratio, 1.013; 95% confidence interval, 1.002–1.028).

The maximum likelihood phylogenetic tree was constructed from 26 ebolavirus whole-genome sequences (16 EBOV, 7 SUDV, 2 BDBV sequences, and 1 TAFV) (Figure 2). There were multiple outbreaks in which phylogenetic and epidemiologic evidence supported the presumption of spillover as the etiology. The 1994, 1995, 1996, 2002, 2003, 2013, 2014, 2017, and 2018 EBOV genomes all had greater sequence variation and/or suspected wildlife sources, so these outbreaks likely started from different spillover events. Only sequences from the EBOV GP or nucleoprotein were available for some of the 2001–2003 outbreaks in the Republic of the Congo (ROC) and Gabon, so these were not included in our phylogenetic tree. Reviewing previous analyses of these sequences and their contexts revealed that there were 4 separate outbreaks during this period due to multiple epizootic and spillover events [33–35]. For each of these outbreaks there were multiple suspected primary/index cases in the context of different spillover sources, chains of transmission, and sequences. We were also unable to find a whole-genome sequence for the 2005 ROC outbreak, where the primary/index cases were 2 hunters with bushmeat contact [36]. We also found that the representative sequences from the 2013–2016 epidemic and the 2021 outbreak in Guinea were closely related (28 nucleotide differences), which was consistent with prior analyses [2].

Figure 2.

Maximum likelihood phylogenetic tree of ebolavirus whole genomes. Representative whole-genome sequences of EBOV, SUDV, TAFV, and BDBV from 1976 to 2021 were used to construct the phylogenetic tree. Numbers above the main nodes are bootstrap values (100 replicates). Each tip is labeled by the ebolavirus, year of onset, and initial country or province/district of each outbreak. Abbreviations: BDBV, Bundibugyo virus; DRC, Republic of the Congo; EBOV, Ebola virus; ROC, Republic of the Congo; SUDV, Sudan virus; TAFV, Taï Forest virus

There were a few historic outbreaks that may have originated from resurgence of prior outbreak strains based on phylogenetic and epidemiologic evidence. The 1977 sequence from Tandala/Bonduni was 99.97%–99.99% similar (2–5 nucleotide differences) to the 3 sequences from the 1976 outbreak in Yambuku (isolates named Mayinga, deRoover, and Ecran). In fact, the Bonduni sequence was more similar to the deRoover isolate (2 nucleotide differences) than all the 1976 isolates were to each other (5 nucleotide differences). Given this sequence similarity, it is possible that the 1977 outbreak may have originated from a resurgence event instead of spillover. The locations of the index cases are roughly 390 kilometers apart and there was no clear epidemiologic chain of transmission. However, during the 1976 outbreak cases were detected within 120 kilometers of Yambuku, and an additional cluster occurred via even further travel by a nurse to Kinshasa [37]. Therefore, a human origin of the 1977 outbreak could not be ruled out.

Given close similarities between the 2007 and 2008 Luebo strains, an additional alignment was done for the 2008 sequence with 5 sequences from the 2007 outbreak in the same location. The sequence from the 2008 Luebo outbreak had 99.9%–99.94% sequence similarity (11–19 nucleotide differences) to the 2007 viruses that occurred in the same location. A prior analysis of the nucleotide substitutions between these outbreak strains postulated that the 2008 virus could be a direct descendent of the 2007 strain [38]. Given the sequence similarities and epidemiologic relationships between these outbreaks, it is possible that the 2008 outbreak originated from resurgence of human-to-human transmission instead of spillover.

The 1979 SUDV sequence from Nzara was 99.75% similar (46 nucleotide differences) to the 1976 outbreak in the same region. The 2004 SUDV strain from Yambio was 99.6% similar (70 nucleotide differences) to the nearby 1976 Nzara sequence. The 2000, 2011, and 2012 SUDV sequences were more divergent, and the preliminary 2022 sequence was not included in our phylogenetic analysis [39]. The evolutionary rate of SUDV from sustained human-to-human transmission or persistent infection has not been determined, and it is unknown whether SUDV can cause flare-ups, so it was assumed that these outbreaks were all separate spillover events. The 2012 Isiro BDBV sequence was 98.6% similar (252 nucleotide differences) to the 2007 Uganda BDBV sequence; there were no identified sources for these outbreaks, and given the sequence differences it was assumed that they occurred from separate spillover events.

DISCUSSION

We identified multiple origins for EBOD outbreaks. While wildlife spillover seemed the likely etiology for the majority of outbreaks, there are both recent and historical outbreaks that may have emerged from human-to-human transmission of prior outbreak strains instead of spillover from wildlife. We identified 2 historic outbreaks (1977 Tandala/Bonduni and 2008 Luebo) that were previously deemed due to spillover events, but it now seems difficult to exclude resurgence of human-to-human transmission as the potential origin for these outbreaks. Since the 2013–2016 West Africa epidemic, 5 EBOD outbreaks have been considered to originate from possible flare-ups of previous outbreak strains. Given these findings we must reconsider the ways that we prepare for future outbreaks.

Previous research on ebolaviruses has focused on mitigating spillover through reducing contact with wildlife and bushmeat. While roughly a third of all EBOD outbreaks and primary/index cases involved suspected bushmeat contact, this is not the only source that we must consider. Identifying which human populations are at risk for reemergence of outbreak strains and increasing surveillance among these populations will also be key to outbreak prevention. Additionally, understanding risk factors for resurgence, such as whether certain antibody-based therapies increase the likelihood of viral persistence, will also be important to determining populations at risk for future outbreaks.

We found that as cases increase during a particular EVD outbreak, there appear to be greater odds of subsequent resurgence. Therefore, early detection and mitigation of EBOD outbreaks is not only critical for stopping current outbreaks but also for preventing future outbreaks. Identifying regions that have had high numbers of survivors from prior outbreaks could help target surveillance and improve early detection of future outbreaks. While we found that there were fewer historical outbreaks that seemed to originate from delayed sexual transmission or relapse of infection, a potential confounding factor is that detection of flare-ups may have increased recently due to enhanced diagnostic methods.

We also found that the majority of EBOD outbreaks were associated with nosocomial transmission. Equipping and training responders with personal protective equipment and the appropriate resources to quickly diagnose and isolate patients will be crucial to stopping health care-acquired transmission. Additionally, multiple transmission events occurred among family groups during burials and by sexual transmission from survivors. Community-based approaches are therefore essential for educating about ebolavirus transmission while also mitigating stigma.

Our epidemiologic and phylogenetic analyses provide a rough estimation of what outbreaks could plausibly occur from resurgence of prior strains, but there are multiple limitations. Given limited sequence data from wildlife, the genetic diversity in wildlife reservoirs remains unknown. Although it has been observed that there are greater differences between strains from separate spillover events compared to human-to-human transmission, further wildlife and human sequences are needed to clarify these relationships and the degree of genetic variation. We also used representative whole-genome sequences from each outbreak to estimate phylogenetic relationships because of limited availability of multiple sequences from different outbreaks. We used previously published estimates of EBOV evolutionary rates as a guide, but estimates may vary depending on the outbreak strain and size of the outbreak. Additionally, ebolaviruses may exhibit heterogeneous evolutionary rates during persistent infection [26], and further studies among survivors are needed to determine these phylogenetic relationships.

Another limitation in our phylogenetic analysis is that methods for sequencing and virus isolation have changed over time, which could influence comparisons of evolutionary relationships. The ability to detect genetic variation within and between outbreaks has improved, making it difficult to compare sequences determined multiple decades ago to those from recent outbreaks. Virus isolates used to go through plaque purification prior to sequencing, which could introduce cell-adapted mutations and clonal selection, which could minimize nucleotide variations. Resequencing archived samples using newer methods and as well as more in-depth phylogenetic analyses including evolutionary rate estimations could provide further information about genetic relationships between ebolaviruses.

The evolutionary rates of SUDV, BDBV, and TAFV are less characterized, and it is also unknown whether these viruses can cause resurgent outbreaks through sexual transmission or relapse of infection. Additional phylogenetic analyses and cohort studies of survivors from infections from these other ebolaviruses are needed to determine genetic variation and the risk for resurgence. A recent genomic analysis of BDBV sequences from the 2012 outbreak supports the hypothesis that multiple spillover events could have contributed to this outbreak, but further epidemiologic and phylogenetic data are needed [40].

Ebolaviruses can emerge in human populations in multiple ways. Recently, there have been large EBOD outbreaks, which create the potential for future resurgence. Therefore, we need additional research and resources dedicated to improving risk assessment and early detection in human populations. Improving the rapid detection and isolation of primary and index cases will be essential to preventing further EBOD outbreaks.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Author contributions. S. D. J. contributed conceptualization, analysis, and figures, and wrote the original draft. S. D. J. and V. J. M. reviewed and edited the manuscript.

Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Data availability. All relevant data are within the manuscript and its online files. Datasets and analyses are available online (https://github.com/sethdjudson/ebola-origins) and can be cited at https://doi.org/10.6084/m9.figshare.22257433.v4.

Financial support. This work was supported by the National Institute of Health (grant number T32 AI007291–27 to S. D. J.). V. J. M. is supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

Supplement sponsorship. This article appears as part of the supplement “10^th International Symposium on Filoviruses.”

Author notes

Presented in part: IDWeek, 18–21 September 2022, San Diego, CA, USA, poster presentation by S. D. J.

Potential conflicts of interest. Both authors: No reported conflicts. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.