Marburg Virus Disease (MVD) is a highly pathogenic zoonotic disease affecting both humans and non-human primates (1). It is caused by the Marburg virus (MARV) of the Filoviridae family under the genus Orthomarburgvirus (2). The virus is classified as a Category A and Risk Group 4 pathogen by the World Health Organization (WHO) and the Centers for Disease Control and Prevention (CDC) (3), and it is considered to have pandemic potential, posing a significant threat to global health security (4). MVD manifests as a hemorrhagic fever syndrome and multiple organ failure in humans (5, 6). The disease has an average fatality rate of approximately 50% (ranging from 24 to 88%), depending on the outbreak’s severity and the quality of healthcare provided during the outbreak (7). In the absence of a known cure, patient management is restricted to symptomatic and supportive care, along with the treatment of concomitant complications (8).

The Egyptian fruit bat, Rousettus aegyptiacus, is identified as the natural reservoir of the MARV, as evidenced by serological and molecular findings in bats (9, 10), and its association with at least 80% of all human MVD outbreaks (11). Although the epidemiology of MVD is not fully understood, the distribution of MARV appears to align with the geographical range of its reservoir bats (6, 11, 12). Furthermore, MVD outbreaks in human populations are significantly influenced by anthropogenic activities, such as contact with infected animals in laboratory settings or after extended periods in caves (7). With the exception of the United States, Germany, the Netherlands, and Russia (Koltsovo, Soviet Union), all reported MVD outbreaks have occurred in Africa (13), suggesting that the disease is predominantly an African issue. Most of these outbreaks lack epidemiological links and have been sustained either through continuous human-to-human transmission following an initial spillover event, as observed in Angola between 2004 and 2005 (3, 14), or through repeated introductions of the virus, as seen in the Democratic Republic of Congo (5). In this review, we examined the prevalence of MARV in bats, rodents, non-human primates (NHP) and domestic animals from the first recorded outbreak to May 2025. We also reviewed the duration of antibodies, and the routes of viral shedding by the selected identified species.

We followed the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 expanded checklist (Supplementary file 1) (15). The review questions were formulated using the CoCoPop framework (Condition, Context, Population) (16). The conditions of interest were MARV infection, the duration of MARV antibody persistence in exposed animals, and potential transmission routes. The population comprised bats, rodents, NHP, and domestic animals, and the context encompassed both wild and domestic settings. Population, Intervention, Comparator, and Outcome (PICO)–based framework was not applied, as this has been associated with reduced recall in non-interventional reviews (17). As this review did not evaluate an intervention or comparator, application of the PICO framework was considered methodologically inappropriate. Accordingly, the CoCoPop framework was adopted as the most appropriate and rigorous structure to guide the review questions, eligibility criteria, and search strategy. This review was not registered in PROSPERO because, at the time of protocol development, the review type “incidence or prevalence of a disease or condition” was not available for registration for systematic reviews involving animals as the primary study population.

A total of 16,735 publications were obtained from the initial search. After applying the predefined filters (publication year and article type) and removing 69 duplicate records, 518 publications were screened; of which 488 were excluded based on the eligibility criteria (Figure 1). Papers were excluded based on three major reasons; reports of published studies by other sources, e.g., CDC (0.2%; 1/488), preliminary reports of studies prior to the publication of the study (0.2%; 1/488), and studies that did not meet the inclusion criteria (99.6%; 486/488). The full texts of the remaining 30 articles were retrieved and evaluated. None of the retrieved articles were excluded after evaluation (Figure 1). Prevalence of MARV was discussed in 26 articles; thus, bats (22), domestic animals (3), NHP (1) and rodents (2). Five publications provided insights into the duration of MARV antibodies and isolation and shedding of MARV in bats.

Only 30 records met the inclusion criteria after successive screening processes. A record was included in the study if it was in English, a full article, investigated the prevalence of MARV RNA and antibodies, provided insights into the duration of MARV antibodies, or isolation or shedding routes of MARV RNA in bats and other animals. The PRISMA diagram was generated using the PRISMA Flow Diagram tool (18).

This systematic review evaluated the prevalence of MARV in bats, domestic animals, and rodents, the duration of antibodies, and potential routes of viral shedding. This is particular important because African countries continue to face numerous public health challenges, including zoonotic diseases of pandemic potential (39). Yet, research output from the continent remains among the lowest globally (40). Although collaborative research efforts are gradually increasing, they remain insufficient, particularly given the urgent need for funding and effective prevention and control strategies (39, 41). Despite the high case fatality rate of MVD and the fact that most outbreaks occur in Africa (42), much of the research on the virus originates from developed countries (43). These nations, motivated partly by the fear of cross-border transmission due to increased global travel and trade, tend to provide the bulk of available research funding (43).

The COVID-19 pandemic served as a stark reminder that infectious diseases are a global threat, and often emerging from regions with limited surveillance and research capacity. In this context, systematic reviews are especially valuable for African researchers. They synthesize available evidence, identify knowledge gaps, and can inspire new lines of inquiry that attract donor interest. This systematic review contributes to bridging part of the existing research gap, providing a foundation for future investigations and interventions aimed at controlling zoonotic diseases such as MARV infection, which disproportionately affect African populations (42).

This review has identified 11 African countries where outbreaks of Marburg disease in humans have occurred (10, 44, 45). Following the outbreaks, surveys have been conducted in bats, rodents, dogs, and livestock in South Africa, Uganda, Kenya, Zimbabwe, Sierra Leone, the Democratic Republic of the Congo, and Ghana to detect the presence of MARV antibodies and viral particles. These surveillance activities were conducted in high-risk areas, including caves with significant bat and rodent populations, livestock in outbreak zones, and areas where bats and livestock intersect. There is a pressing need for comprehensive surveillance across all affected countries to elucidate the epidemiology of the disease and origins of the outbreaks.

The elevated levels of MARV antibodies and the increased presence of MARV in Rousettus aegyptiacus support the hypothesis that this bat species is a primary reservoir of MARV. However, the isolation of MARV and the presence of MARV antibodies in other apparently healthy bat species, including the relatively common Eidolon helvum, necessitate consideration of their role in maintaining MARV. The wide range of reported seroprevalence (0.93–53.04%) is biologically plausible and expected given variances in bat ecology, signalment and demography and methodology across the identified studies. Additionally, temporal variation across studies spanning several decades is expected to contribute to fluctuations in seroprevalence due to outbreak cycles and undulating immunity in sampled bats even in the same location.

The duration of MARV antibodies in R. aegyptiacus is estimated to be at least 4 months, with potential loss within 3 months of natural horizontal infection or 5 months following vertical transmission (21). This may explain the lower success rates in identifying MARV-specific antibodies in bats. MARV has been isolated from or demonstrated in R. aegyptiacus tissues including the kidney and liver (14, 24, 29, 31). However, there is limited evidence conclusively supporting the shedding of the virus through the oral route as the virus has been detected in oral fluid (26). However, the absence of MARV in saliva and its presence in other tissues, as reported by Towner et al. (24), led to the hypothesis that transmission through partially eaten fruits, as observed for Ebola and Nipah virus, is unlikely for MARV. Recent experiments show the persistence of the virus for up to 6 h on partially eaten fruits by infected R. aegyptiacus (46), which necessitates the need for further investigation into the routes of transmission. Given the elusive duration of MARV antibodies, the brief viremic stage of the virus in bats, and the demonstration of the virus in non-R. aegyptiacus bat species, further revealed that other bat species may play a significant role in the maintenance and circulation of MARV.

The hypothesis that livestock, rodents, and dogs contribute to the maintenance of MARV in at-risk areas has recently been explored by four research groups. Although none of these groups detected the virus in these animals, two study groups (34, 35) identified the presence of MARV antibodies in selected livestock and dogs. This hypothesis arises from the unique interactions between dogs, livestock, and at-risk humans in outbreak regions, such as Gabon (35), and wildlife, as observed in Ghana (34). The observed lower MARV antibody detections in livestock and dogs in Gabon, compared to higher levels reported in similar species in Ghana, warrants further investigation. Such variations may reflect differences in ecological conditions, viral circulation, host exposure (47, 48), and sampling strategies and methodologies (34, 35, 49) between the two regions. The detection of MARV antibodies in domestic animals, particularly those in close contact with humans, raises important public health considerations. While the presence of antibodies indicates previous exposure rather than active infection, it nonetheless suggests that these animals may serve as sentinels of viral spillover events or potential intermediate hosts in areas where fruit bats, the natural reservoirs of MARV, are abundant.

None of the included studies performed neutralization assays such as the plaque reduction neutralization test (PRNT), which remains the most specific method for confirming virus-specific antibody responses. However, such assays are rarely applied in MARV research because they require biosafety level-4 containment. Consequently, the observed seropositivity in some studies (27, 34, 35) could, in part, reflect cross-reactivity or low-affinity binding rather than true infection due to shared antigenic regions among filoviruses (50). Furthermore, differences in antigen targets, assay platforms, signal detection thresholds, and cut-off definitions across these serologic tests can lead to variability in sensitivity and specificity between studies. The use of Western Blot (20) and paired sample testing are supplementary approaches that improve confidence in the specificity of the serological tests for MARV antibodies. Experimental work has shown that sera from fruit bats experimentally infected with one filovirus can cross-react with antigens from other members of the Filoviridae family (51). Similar challenges have been reported in other serosurveys of filoviruses (26, 52). Therefore, while antibody detection in wildlife offers valuable clues about potential viral exposure, the absence of confirmatory neutralization testing limits the specificity of these findings and calls for follow-up studies employing assays with higher confirmatory value.

The possibility of domestic animals acting as mechanical or biological bridges for zoonotic transmission cannot outrightly be ruled out, especially in rural communities where livestock, dogs, and humans share habitats and resources. This highlights the importance of adopting a One Health approach to surveillance, integrating animal, human, and environmental health data to understand MARV dynamics. Future research should aim to clarify the epidemiological significance of MARV antibodies in domestic species, whether they indicate incidental exposure or potential roles in viral maintenance; especially when MARV was not demonstrated in any of the exposed species. Strengthening serological surveillance and diagnostic capacity in Africa will be critical to early detection and prevention of future outbreaks. The observed lower MARV antibody detections in livestock and dogs in Gabon detected using Luminex (specificity of 97 across species and target proteins), compared to the higher levels in similar species in Ghana detected using an in-house ELISA (no specificity estimate provided), may primarily be due to differences in the specificity of the tests conducted. However, while the MARV antibodies detected in Gabon may reflect potential exposure to infected individuals, those identified in Ghana are more likely due to repeated exposure to potentially infected bats, as the sampled livestock and hunting dogs frequently interacted with bats in the wild or in forests bordering human settlements. The presence of antibodies in these animals necessitates further investigation into their precise role in the maintenance and transmission of MARV in at-risk and endemic areas.

Only four studies successfully isolated MARV in Rousettus aegyptiacus while none used any confirmatory assays to verify MARV antibody specificity, thereby reducing the confidence in the reported serological findings. Both limitations are likely driven by restricted access to BSL-4 laboratory facilities. In addition, substantial heterogeneity in study design, sampled populations, diagnostic approaches, and outcome definitions limits comparability between studies.

This review highlights the widespread detection of MARV across multiple African countries and species. Evidence supports Rousettus aegyptiacus as a primary reservoir, but other bat species may also play critical roles. The recent identification of MARV antibodies in some livestock and dogs warrants further investigation into the role of these animals in the maintenance of MARV. Therefore, there is the need for expanded and targeted multispecies surveillance of MARV in high-risk countries to clarify reservoirs and host distribution as well as the transmission dynamics of MARV.