Marburg virus persistence and clinical complications among recovered persons in Rwanda: a longitudinal study protocol

Abstract

Introduction

Marburg virus (MARV) causes a potentially highly lethal haemorrhagic fever disease that threatens global health security, and knowledge of viral persistence, long-term clinical outcomes among recovered persons and protective immunity remain limited. Existing protocols and technical guidelines provide a framework for safe sample collection, clinical monitoring and outbreak response, yet systematic approaches to investigate these aspects in African settings are scarce. This protocol describes the Partnership Research on MARV (PREMAV) following the 2024 outbreak in Rwanda. PREMAV is a longitudinal study that was initiated to investigate the persistence of MARV in body fluids of recovered persons, characterises associated clinical complications, identifies factors associated with viral persistence, classifies the protective immune response and discovers genomic variants during viral persistence.

Methods and analysis

This is a protocol for a prospective cohort study that enrolled individuals who survived Marburg virus disease (MVD), individuals with natural antibodies for MARV, close contacts and low risk controls in Rwanda. Participants will undergo serial sampling of blood, urine, semen and other relevant specimens over 24 months post-recovery. This will be combined with continuous clinical assessments, and laboratory analyses will include molecular detection, serological profiling, immunological assays and genomic viral characterisation. Statistical analysis will integrate the various datasets to determine associations.

Ethics and dissemination

This study will adhere to the Declaration of Helsinki and obtain approval from the Rwanda National Ethics Committee. Written informed consent will be obtained from all participants with additional safeguards for vulnerable groups. Confidentiality will be ensured through de-identification and secure data storage. All biological samples will be handled in compliance with national and international biosafety standards. Psychosocial support will be provided to address stigma and mental health needs among survivors. The study will generate critical evidence on MARV persistence, clinical sequelae and risk factors, informing survivor care guidelines, transmission mitigation strategies and future outbreak preparedness. Findings will be disseminated through peer-reviewed publications, conferences and policy briefs targeting national and global stakeholders. De-identified data and genomic sequences will be curated and shared through secure platforms and public repositories in line with ethical and regulatory requirements.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Longitudinal cohort design with repeated sampling will enable detailed assessment of Marburg virus (MARV) persistence and immune dynamics over time.

• Comparative participant groups including known survivors, their close contacts with natural MARV antibodies and seronegative controls strengthen the analyses of seroprevalence and subclinical infections.

• Integrated immunological and genomic approaches provide a comprehensive framework to investigate viral persistence, immune correlates and transmission.

• Implementation of Partnership Research on MARV within Rwanda’s national health system enhances feasibility, supports local capacity building and ensures public health relevance.

• Observational study limitations such as loss to follow-up, restricted sample size and potential selection bias may affect generalisability and causal inference.

Introduction

Marburg virus disease (MVD) is a growing severe acute illness, primarily emerging in sporadic outbreaks in sub-Saharan Africa.¹ To date, there have been over 20 documented outbreaks, resulting in approximately 734 human infections and around 558 deaths, with case fatality rates ranging from 24% to 90%, depending on the outbreak.^{2 3} The outbreak in Rwanda was the third largest recorded globally, with 66 MVD PCR-confirmed cases and 15 deaths.⁴ The largest outbreak occurred in Angola and the Democratic Republic of the Congo, both with a high case fatality.⁵ The last reported MVD outbreak was documented in Ethiopia December 2025.⁶

The MVD is caused by two distinct negative-strand RNA viruses belonging to the species Orthomarburgvirus marburgense: Marburg virus (MARV) and Ravn virus (RAVV).⁷ Only a few cases have been attributed to RAVV, with most reports indicating that only three confirmed cases have been linked to RAVV.² The sporadic nature of MARV has made identifying its natural reservoirs challenging. Key potential sources for initial transmission mechanisms were suggested.⁸ However, studies have confirmed that Rousettus aegyptiacus, fruit bats are the primary natural reservoir for MARV.^{8 9} The transmission of MARV to humans is not yet fully understood, but it is likely linked to incidental exposure to infected bats and their contaminated excretions in caves.¹⁰ MVD then spreads through human-to-human transmission, with infection resulting from direct contact with blood, secretions, organs or other bodily fluids of infected people and indirect contact with environments contaminated by such fluids.¹

MVD typically manifests abruptly after an incubation period of 7 to 11 days (with a range of 2 to 21 days).⁸ In phase 1, the disease presents as a nonspecific, influenza-like illness, while phase 2 involves more severe symptoms affecting the abdomen, central nervous system, respiratory and vascular systems and often includes haemorrhagic manifestations. Common clinical signs include abdominal pain, anorexia, diarrhoea, fatigue, malaise, nausea, myalgia, maculopapular rash, sore throat and vomiting. Laboratory findings frequently show thrombocytopenia, elevated transaminase levels and electrolyte imbalances.^{11 12} MVD induces multiorgan dysfunction syndrome, following hypovolemic shock, which is the typical disease’s prognosis leading to mortality.¹³ Survivors often face long, challenging recoveries in hospital, complicated by secondary infections, orchitis and neuropsychiatric conditions.¹ Although not well documented, MVD survivors appear to experience prolonged clinical sequelae, including arthralgia, neurocognitive dysfunction and uveitis.¹⁴ Disease relapse has also been reported, suggesting that MARV may persist in survivors under yet unknown conditions.^{14 15} It is believed that viral persistence in disease survivors might play a crucial role in maintaining transmission among the human population. One case of sexual transmission has been linked to persistent MARV infection after recovery,¹⁶ and studies in nonhuman primates suggest that MARV can remain in the testes for extended periods post-recovery.¹⁷

Due to the sporadic nature of past MVD outbreaks, their high case fatality rates and the remote locations in which they occurred, little is known about the long-term sequelae, protective immunity and its persistence in recovered persons compared with its counterpart Ebola virus disease (EVD).¹⁸ Cohort studies of EVD survivors have documented persistent symptoms and potential infectiousness during convalescence,^18–20 but no similar in-depth investigations exist for MVD. The 2024 Rwanda outbreak, which left an estimated 51 survivors, provides a unique opportunity to thoroughly assess the long-term sequelae in MVD recovered persons, protective immunity, as well as understanding within-host genomic variations and its association with delayed virus clearance, and the potential for relapse and/or secondary transmission of MARV, the presence of persistent inflammation, and the durability of MARV-specific immunity.

This manuscript describes the design and implementation of the Partnership Research on Marburg Virus (PREMAV), a longitudinal cohort study established following the 2024 outbreak in Rwanda. The study aims to systematically investigate MARV persistence in body fluids, long-term clinical sequelae, host immune responses associated with protection and within-host viral genomic evolution. By integrating clinical, immunological and genomic data within a single framework, PREMAV seeks to address critical knowledge gaps on post-infection outcomes and viral biology. The findings are intended to inform survivor management, transmission risk assessment and future outbreak preparedness in Rwanda and similar settings.

Research gaps and study rationale

Despite the significant threat of MARV that places it among the top candidates for pathogens that might cause the next pandemic,²¹ little progress was achieved toward evidence-based strengthening of prevention and control measures. Therefore, critical knowledge gaps remain regarding viral persistence, protective immune responses and long-term clinical outcomes among survivors. Most existing studies are limited by small sample sizes, cross-sectional designs or focus solely on acute infection, leaving the natural history of viral clearance and sequelae poorly characterised. In particular, the dynamics of MARV persistence in body fluids, the protective immunity and the incidence of asymptomatic or subclinical infections among close contacts remain largely unexplored. These knowledge gaps are particularly important because the contemporary clinical and public health environment differs from historical MARV outbreak settings, with improved supportive care, enhanced diagnostics, access to investigational vaccination, post-exposure prophylaxis and potential therapeutic interventions that may influence survival, viral clearance, immune responses and long-term outcomes. Furthermore, genomic investigations of within-host viral evolution and transmission chains are scarce, constraining our ability to detect cryptic transmission events or viral adaptations. This study is designed to address these gaps through longitudinal sampling of survivors and their contacts, integrated serological, immunological and genomic analyses, and systematic assessment of clinical complications. By generating comprehensive data on MARV persistence, immunity and sequelae, the study will inform survivor care, public health guidelines and the development of effective countermeasures.

Study objectives and outcomes

Primary objectives

The main objectives of the study are to generate comprehensive data on MARV persistence, immunity and clinical sequelae in survivors, through longitudinal sampling and integrated serological, immunological and genomic analyses, in order to inform survivor care, public health guidelines and the development of effective countermeasures.

Secondary objectives

• Estimate the incidence and risk factors for post-MVD medical complications including vision problems, immune function changes (eg, inflammatory markers), joint pain, neuropsychiatric problems, diabetes, hypertension and pregnancy complications and others.

• Determine the persistence of MARV in body fluids (eg, semen, vaginal secretions, breast milk, sweat, tears, cerebral spinal fluid).

• Characterise the longitudinal humoral immune response among MVD survivors by developing and standardising serological and neutralisation assays to quantify MARV-specific antibody magnitude, durability, functionality and waning over time. This will address a major gap in understanding correlates of long-term immunity, protection against reinfection and implications for vaccine development and survivor management.

• Record birth information (Apgar scores, birth weight and gestational age), collect cord blood samples and placenta samples for virologic studies, assess development milestones including growth parameters (body weight, length, head circumference) and measure MVD antibody levels for children born to MVD survivors during the study.

• Estimate the seroprevalence of MARV infection among contacts (including the “high-risk” contacts) of confirmed and probable MVD cases.

• Study the genomic evolution of MARV within individual human hosts over time and across different human hosts, with a focus on understanding viral diversity, adaptation and factors driving its transmission and pathogenicity.

Methods and analysis

Study design

This is an observational cohort study of MVD recovered persons. The cohort study is intended to provide follow-up information on a clinically well-characterised group of individuals who recovered from MVD. Consented participants who recovered from the MVD will be enrolled and risk factors (eg, age, sex, length of stay in the hospitalisation and time since discharge from the hospitalisation) for long-term complications will be assessed at baseline assessment. Participants will be followed for up to 2 years to determine the incidence and predictors of post-MVD medical complications and to evaluate the evolution of medical complications over time. Clinical follow-up visits will occur every 3 months and blood samples for immunological studies will be collected every 6 months (figure 1).

Figure 1. Baseline examination and follow-up visits.

The PREMAV study will be conducted as a single-centre longitudinal study, with participant enrolment and follow-up taking place at the Rwanda Military Referral and Teaching Hospital (RMTH). Routine clinical laboratory investigations, including haematology, clinical chemistry and radiological assessments, will be conducted at RMTH. Peripheral blood mononuclear cell (PBMC) isolation and storage, as well as serological and molecular testing, will be centralised at the Rwanda National Reference Laboratory in Kigali, Rwanda.

To determine whether the prevalence of medical complications is greater among MVD recovered persons than those who did not have MVD, selected close contacts or other unexposed people with negative serology for MARV will be examined and used as controls. Additionally, laboratory data and selected medical conditions (eg, diabetes, hypertension) from MVD recovered people will be collected and documented.

Inclusion criteria and exclusion criteria

Inclusion criteria

Eligible participants include volunteers of any age with a confirmed diagnosis of MVD during the outbreak who are listed in the MVD registry of recovered people. Participants must be willing to attend examinations at RMTH and to provide informed consent or assent. In addition, close contacts of MVD patients who are recorded in the national contact-tracing registry may volunteer. Individuals who test positive for MARV antibodies will be enrolled in the cohort and followed alongside PCR-confirmed recovered MVD patients. Furthermore, samples and associated data from MVD-deceased individuals will be incorporated into the analyses.

Exclusion criteria

Any condition in the judgement of the lead study physician that would make the volunteer unable to participate in the study. Specific exclusion criteria include:

• Severe comorbid illness such as advanced cardiovascular, hepatic, renal or neurological disease that may interfere with study procedures or outcomes.

• Acute medical conditions requiring hospitalisation or intensive treatment at the time of screening.

• Immunosuppressive therapy including systemic corticosteroids or cytotoxic drugs, which may alter immune responses and confound study analyses.

Other investigator-determined risks where participation would be unsafe or data integrity could be compromised, including inability to comply with follow-up visits or repeated sampling.

Data collection procedure for MVD recovered persons

At the time of enrolment, participants will be provided with a participant information sheet detailing the study objectives, procedures, potential risks, and expected benefits (online supplemental file 1): Section 1. This document will be given prior to any study-related activity to allow sufficient time for review and questions. Following review of the information sheet and checking for understanding, participants will be asked to sign a written informed consent form confirming their voluntary agreement to participate and conduct the study visits. For minors, assent forms will be provided alongside parental or guardian consent forms.

Each scheduled visit is expected to last approximately 2–3 hours, depending on the number of procedures performed. This includes registration, consultation, sample collection and closing procedures. Participants will be informed in advance of the expected time commitment. At each scheduled visit, participants will receive a visit summary sheet outlining the specific procedures to be performed, including sample collection, clinical assessments and any associated instructions. In the event of clinically relevant findings, participants will be given clinical results notification and referred to appropriate care providers as per national guidelines. The baseline examination and follow-up visits will be conducted as shown in figure 2.

Figure 2. Participant flow and procedures during data collection. PBMC, peripheral blood mononuclear cell; RMTH, Rwanda Military Referral and Teaching Hospital.

On enrolment visit, the study procedures are classified as either mandatory or optional to ensure methodological transparency. Mandatory procedures include informed consent, eligibility screening, demographic data collection, medical history and physical examination, clinical data recording and blood specimen collection. Optional procedures, while voluntary, are strongly encouraged and include questionnaires on depression, post-traumatic stress and stigma; semen and vaginal secretion sampling; urine specimen collection, protein and creatinine; detailed ophthalmological examination and survivor-specific documentation of the MVD event.

During clinical evaluation, if potentially treatable conditions are identified by study evaluations, such as secondary bacterial infections, malaria, nutritional deficiencies or diarrhoeal diseases, the participant will be referred to affiliated clinic for further evaluation and treatment in accordance with standard medical practice in Rwanda. Study physicians may serve as consultants for provision of standard medical care to participants (table 1).

Table 1. Follow-up and events schedule for MVD recovered persons.

Events	Baseline (Visit one or during discharge)	Follow-up¥
Informed consent	X
Eligibility screening	X
Demographic data	X
Medical history, symptoms review and physical examination	X	X
Clinical data	X	X
Details of MVD event*	X
Questionnaires on depression and post-traumatic stress	X	X
Questionnaire on stigma	X	X
Semen sample for MVD testing*		X
Blood specimen for PBMC	X	X
Blood specimen for antibody testing	X	X
Blood specimen for chemistry and haematology*	X	X
Vaginal secretion sample*		X
Urine specimen for MARV PCR, protein and creatinine*		X
Samples’ storage	X	X
Detailed eye exam	X

^*For recovered individuals at clinician discretion. ¥ (Every 3 months±1 months) for clinical evaluation and 6 months for immunological sample collection. Participants might also be evaluated whenever necessary.

MARV, Marburg virus; MVD, Marburg virus disease; PBMC, peripheral blood mononuclear cell.

Enrolment of MVD recovered persons

The baseline examination will be divided into two parts: (1) One part will include medical history, physical examination and the blood draw and (2) The second part will include more extensive targeted evaluations when clinically indicated. Individuals presenting with complications such as eye inflammation, visual disturbances or neurological symptoms (eg, repeated headache, neuropathic pain, cognitive changes) will be examined respectively by an ophthalmologist or neurologist, or any other required specialists in case of other complications. List of data and samples to be collected for the first part of the baseline examination are described in online supplemental file 1), Section 2.

Note that except for blood samples collected for immunological assessments, all other body fluids will be obtained at the clinicians’ discretion for medical care purposes. Approximately 2 weeks after the baseline visit, participants will be invited to return to discuss laboratory findings. This interim visit is not considered a scheduled study visit within the protocol, as its occurrence depends on the outcome of clinical assessments and the presence of results requiring timely medical attention or specialist consultation. The purpose of this visit is to ensure that participants with abnormal findings (eg, haematological or biochemical abnormalities, evidence of persistent viral shedding or complications requiring referral) receive appropriate care prior to the next scheduled follow-up. For participants without clinically significant findings, results are communicated at the subsequent routine study visit.

The second part of the baseline examination will include an eye examination (eg, visual acuity, intraocular pressure) for MVD recovered persons who completed the baseline visit. We anticipate that some recovered individuals may test positive for MARV in certain body fluids, such as semen, although we do not expect viremia. If a semen test returns positive, the participant will be informed and provided with counselling, emphasising the risks of transmission through unprotected sex.^{22 23} However, if a blood test is MARV positive, this will prompt the initiation of all necessary precautions and measures for managing positive MVD patients.²⁴ The patient will be isolated and receive treatment and supportive care, while close contacts will be monitored and tested as needed.

Follow-up data collection

The MVD recovered individuals enrolled will be followed longitudinally with yearly detailed ophthalmological examinations. Ophthalmological examinations may occur more frequently to evaluate new or changing ocular pathology.²⁵ Also, except for blood samples collected for immunological assessments, all other body fluids will be obtained at the clinicians’ discretion for medical care purposes (online supplemental file 1), Section 2.

Bodily fluids

Participants will be asked to provide bodily fluids (eg, sweat, tears, for adult males’ semen or vaginal and breast milk in lactating women) to determine the persistence of MARV. Body secretions in MVD survivors may potentially contain live MARV.²⁶ Thus, appropriate protective precautions will be taken by all personnel involved in obtaining, handling and shipping of these specimens, as currently recommended.²⁷ All procedures involving the collection and handling of bodily fluids will be conducted under strict biosafety protocols to minimise the risk of viral transmission to healthcare workers and study personnel. Staff engaged in specimen collection will be trained in infection prevention and control practices and will consistently use personal protective equipment, including gloves, gowns, masks or respirators and eye protection, throughout the process. Specimen collection will be performed in controlled clinical environments with measures in place to reduce exposure risk, and all materials used will be appropriately decontaminated and safely discarded after use.

Following collection, samples will be transported under sterile conditions ensuring triple-layer containment and clear labelling to prevent accidental exposure during transit within 6 hours of collection to the NRL which is equipped with biosafety level-3, where engineering controls such as biosafety cabinets are employed. Standard decontamination procedures will be applied to all work surfaces and equipment.

Participants with evidence of MARV in their bodily fluids will be referred to a case manager for counselling. In addition to the foreseen study visit, they will be seen approximately every 2–3 weeks for repeated testing until there is no evidence of MARV in their bodily fluids, that is, at least two consecutive negative PCR results.

Aliquots of all collected samples will be frozen at −80°C and stored under standardised biobanking conditions. These aliquots will be shipped to a biosafety level-4 laboratory for viral culture, in accordance with international biosafety and transport regulations, to allow direct investigation of the infectivity of samples. Beyond immediate infectivity testing, these stored specimens might provide a valuable resource for future collaborative studies, including live virus neutralisation assays, evaluation of antibody-virus complex characteristics and development of animal model experiments.

Data collection procedure for close contacts of MVD patients

MVD recovered individuals will be asked if they would like to identify close contacts up to five household members, friends or neighbours who were near at the time of diagnosis and/or after recovery from MVD. In addition, individuals who were sexual partners after recovery will be invited to participate in study visits. Close contacts of any age will be eligible to participate if they do not have a history of MVD and are able to provide appropriate informed consent (either personally or through a legal guardian) (online supplemental file 1), Section 3.

Those identified will be invited to attend a clinic examination at which demographic information will be collected, a medical history will be obtained, a physical examination, including a vision screen, will be performed and a blood sample will be obtained; additional data collected will be similar to those collected from MVD recovered individuals (table 1). A brief questionnaire on contact with the MVD patients will also be administered (online supplemental file 1), (online supplemental table S1). Selected close contact participants with and without eye problems detected at screening may be referred for the detailed eye examination as controls. Other controls with negative serology for MARV may also be identified among the close contacts. These controls will be invited to attend baseline and follow-up examinations that include data collected like that of close contacts (table 1).

Data collection procedure for children of MVD survivors born during the study

For enrolment of children born to MVD survivors during this study, a consenting procedure will be followed to ensure the parents voluntarily agree that the child participates (table 2). List of data to be collected at the time of enrolment are described in online supplemental file 1), Section 4.

Table 2. Follow-up and events schedule for children born to Marburg virus disease survivors during the study.

Events	Baseline (visit 1)	Follow-up (every 6 months±3 months)
Parental/guardian informed consent	X
Demographic data	X
Medical history, symptoms review and physical examination	X	X
Weight, length, mid-arm circumference and head circumference	X	X
Neurodevelopment milestones	X	X
Birth history	X
Cord blood sample (if available)	X
Placenta sample (if available) for MARV RT-PCR	X
Blood specimen for antibody testing	X	X
Samples’ storage	X	X

MARV, Marburg virus.

Data monitoring and quality assurance

Data monitoring

The observational study will be conducted under the direction of a protocol team. On a regular basis, the protocol team will review summary statistics for participants enrolled. The analyses will be shared with key representatives of the Ministry of Health for swift decision if needed. Data collection plans may be modified, and new protocols or sub-studies may be planned based on accumulating data.

Data quality assurance

To minimise data recording errors, a digital system (REDCap) will be used and will incorporate fields with validations limiting entries to plausible data. Cleaning and quality assessment of the anonymised raw data, reaching the server at a national data centre will be conducted initially by the study data manager. Automated high-frequency checks, including assessments of completeness and indicators of irregularities, will be included in the quality checks. Reports on data quality will be shared with the field implementation team. Any systematically missing or inconsistent data will be addressed in consultation with the field team.

Sample size

The study consecutively enrolled all consenting MVD-recovered patients (40 of 51 individuals who recovered), along with their close contacts and a control group matched to the recovered patients and contacts on demographic characteristics. Additionally, 44 individuals with serological evidence of natural MARV infection were identified and enrolled. This resulted in a total of 84 individuals (59 males and 25 females) under active follow-up at RMTH. With the inclusion of at least three close contacts per cohort participant (3:1), the total number of participants is expected to exceed 300 participants. If the number of affected individuals remains low, the analysis will focus on describing complications and the potential infectiousness due to the persistence of MARV in body fluids.

Immunoassay platforms for characterisation of humoral and cellular immune responses to Marburg virus

Three key immunoassays were identified based on their development maturity and performance characteristics and will be used to characterise immunological biomarkers and assess their association with protection against MARV infection. The first two are antibody-based assays: a pseudovirion neutralisation assay to measure functional neutralising antibody responses and a MARV glycoprotein (GP)-specific ELISA to quantify binding antibody levels.²⁸ The technology transfer will be supported to enable local implementation of the assays in Rwanda, including establishment at a central sample storage and testing site to strengthen in-country capacity. The third assay is an enzyme-linked ImmunoSpot (ELISPOT) assay, which will be used to detect and quantify antigen-specific cellular immune responses by measuring cytokine secretion following stimulation with MARV antigens. Together, the pseudovirion neutralisation assay and the MARV GP-specific ELISA provide quantitative and functional measures of antibody responses, while the ELISPOT assay enables assessment of antigen-specific T-cell activity.

The integration of these orthogonal readouts allows for a more comprehensive immunological profile of infected individuals. This multi-dimensional approach increases the likelihood of identifying robust immune correlates of protection by enabling correlation analyses between specific immune markers and clinical outcomes, including susceptibility to infection, disease severity and durability of immunity over time.

Genomic characterisation of Marburg virus in human hosts

Genomic analyses will be conducted to characterise the genetic diversity of MARV within and across infected human populations, with the aim of identifying viral variants potentially associated with transmissibility, persistence, disease severity and zoonotic potential. These analyses will enable a deeper understanding of viral evolution in human hosts and may support the identification of genomic signatures linked to clinically relevant outcomes, thereby contributing to improved interpretation of transmission dynamics and pathogenesis. RNA extraction will be performed at the NRL under appropriate biosafety conditions, as previously described.²⁹ Next-generation sequencing (NGS) will be performed using platforms like Illumina to achieve deep sequencing with coverage of 500x - 1000x. This depth will allow for the detection of low-frequency variants.

Data analysis

Analyses will be carried out for MVD recovered persons who are confirmed to be on the MVD Treatment Centre registry. Simple descriptive statistics will be used to describe the MVD recovered persons enrolled. Characteristics according to age, sex and time since diagnosis of MVD will be summarised. Associations will be summarised with χ² statistics. Logistic regression will be used to study the association of multiple factors with selected conditions at study entry, for example, major vision problems. Logistic regression will also be used to determine whether the prevalence of medical complications at baseline for MVD survivors differs from close contacts or other controls who are antibody negative to MARV after adjusting for factors such as age and sex.

To assess the representativeness of the MVD survivors who are enrolled in this study, data from the treatment centre registry on demographics and time of MVD hospitalisation for MVD recovered persons will be used. The rate of medical complications during follow-up will be summarised with life-table methods. Cox regression analysis will be used to study predictors of major outcomes during follow-up. Similar analyses will be carried out for comparisons of MVD recovered persons with controls. Seroprevalence rates for antibodies to MARV in enrolled contacts will be estimated. Trajectories over follow-up of antibody levels of MVD survivors, and of household members and sexual partners with antibody levels at entry, will be summarised. Predictors of antibody changes will be studied with longitudinal regression models. To estimate the longevity of MARV in body fluids, we will conduct survival analysis, as previously performed in the EVD study.³⁰

Patient and public involvement

None.

Risks related to participation

The risks of having blood taken include transient pain, bleeding, bruising, light-headedness, fainting and rarely, infection or a blood clot where the needle enters the body. The amount of blood drawn for research purposes shall not exceed 10.5 mL/kg to a maximum of 50 mL for adults over a 6-month period. For children, no more than 5 mL/kg will be drawn for research on a single day and no more than 9.5 mL/kg will be drawn over any 8 week period.

Examination of the eye may cause discomfort that could include temporary stinging from eye drops, temporary blurring of vision or light sensitivity from dilating drops or temporary irritation after collecting conjunctival swab specimen.

Participants may possibly experience mental distress and feelings of stigmatisation if MARV is detected by PCR in body fluid such as semen. Appropriate counselling will be done for all participants who volunteer to have body fluids tested. In addition, results of PCR testing will be kept confidential and only shared with the participant.

Study withdrawal

Every effort will be made to follow participants for the duration of the study. Following enrolment, participants may withdraw from follow-up in the study at any time at their request. They may resume participation at any time. However, participants who choose to withdraw from the study while still testing positive for MARV in one or more body fluids will be permitted to do so. In such cases, study-specific procedures will cease, but the participant will continue to be followed-up post-recovery in accordance with WHO and RBC guidelines.²⁷ This approach ensures continued clinical monitoring, appropriate case management and public health surveillance to mitigate the risk of onward transmission while fully respecting the participant’s right to withdraw from research participation.

Ethics and dissemination

Ethical considerations

This study will be conducted in full compliance with the Declaration of Helsinki,³¹ Good Clinical Practice guidelines and relevant Rwandan regulations, with oversight by the Rwanda National Ethics Committee (RNEC).³² Prior to initiation, the protocol, informed consent forms and participant materials received the RNEC approval in October 2024 (approval number, RNEC 132/2024), and any subsequent amendments will be submitted for review and approval. All participants provided written or marked informed consent, with procedures in place to ensure comprehension for illiterate volunteers. Participant confidentiality was strictly maintained, and compensation for travel, time and study-related procedures, including blood collection and lumbar puncture was provided according to RNEC guidelines.

Adverse events (AEs) and serious AEs related to study procedures (blood draws, non-invasive body fluid collection and lumbar puncture) were recorded and reported to the RNEC following expedited and annual reporting requirements. Protocol deviations, both major and minor, were documented and assessed for potential impacts on participant safety and scientific integrity.

Dissemination

The findings from this study will be disseminated through multiple complementary channels to ensure timely translation into policy, practice and further research. Results will be published in peer-reviewed international journals and presented at regional and global scientific conferences focusing on emerging infectious diseases and outbreak preparedness. At the national level, interim and final analyses will be shared with the Rwanda MoH and the RBC through policy briefs, technical reports and stakeholder workshops to support evidence-based decision-making for survivor care and outbreak response strategies.

In addition, study findings will be communicated to participating communities and MVD survivors through structured feedback sessions, community engagement meetings and simplified information materials to promote awareness and reduce stigma associated with MVD. Data generated from immunological and genomic analyses will be shared, in accordance with ethical approvals and data governance frameworks, with international collaborators and global research consortia to advance the understanding of MARV epidemiology and inform vaccine and therapeutic development where appropriate.

Partnership and roles

PREMAV project is led by RBC in collaboration with local partners, including Rinda Ubuzima, Rwanda Military Teaching Hospital (RMTH), Centre for Family Health Research, and the University of Rwanda, along international partners. CEPI supports the establishment of standardized sample collection procedures and systematic bio-archiving for immunological assessments aimed at identifying immune correlates of protection (CoP) against Marburg virus. The U.S. CDC supports the development of the national MVD recovery program, hosted at the RMTH, within which the PREMAV platform is embedded, and provides technical guidance for the implementation and analysis of the MVD serosurvey. The Pandemic Sciences Institute (PSI) at the University of Oxford, funded by the UK's Foreign, Commonwealth & Development Office (FCDO) supports components of the national recovery program and will contribute to the analyses of clinical sequelae and viral genomics

Discussion

EVD survivor protocols have been critical in defining post-acute filovirus outcomes, demonstrating that virus persistence in immune-privileged sites (eg, semen, ocular fluid and CNS) can occur for prolonged periods and is associated with relapse risk and sexual transmission. Longitudinal EVD cohort studies have also characterised chronic clinical sequelae, including musculoskeletal, ocular and neurocognitive complications, and have informed structured survivor follow-up, clinical care and psychosocial support frameworks.^33–37 However, comparable systematic survivor protocols and longitudinal data remain largely lacking for MVD, representing a key gap in understanding viral persistence, long-term outcomes and survivor care needs for this infection.

This protocol describes a comprehensive longitudinal cohort study designed to address critical knowledge gaps in the epidemiology, clinical sequelae, immunology and transmission dynamics of MVD. Despite the high case fatality rates and recurrent outbreaks of MVD in sub-Saharan Africa, evidence on long-term outcomes among survivors remains extremely limited.³⁸ Leveraging the recent Rwanda outbreak, this study will provide a timely and unique opportunity to generate high-resolution data on post-MVD complications, viral persistence and immune responses in a well-characterised cohort.

A key strength of this protocol lies in its integrated, multidisciplinary design, combining clinical follow-up, immunological profiling and viral genomic analysis. While prior research on filoviruses, particularly EVD, has demonstrated the persistence of viral RNA in immune-privileged sites and the occurrence of post-viral syndromes, comparable data for MVD are scarce.^39–42 By systematically investigating viral persistence across multiple body fluids, this study will contribute to a more nuanced understanding of potential reservoirs of infection in survivors and the risk of delayed transmission, including sexual transmission. This is particularly relevant given historical reports suggesting prolonged viral persistence and rare cases of secondary transmission.^{22 23 26}

The inclusion of close contacts and seronegative controls further strengthens the study by enabling comparative analyses of clinical outcomes and serological responses as seen in similar seroprevalence investigations of MARV and other filoviruses.^43–46 This approach will allow for the estimation of excess morbidity attributable to MVD and improve understanding of asymptomatic or subclinical infections in human populations.⁴⁴ Additionally, the assessment of seroprevalence among contacts will provide insights into the extent of undetected transmission, which remains a critical gap in current outbreak investigations.

Incorporation of neutralisation assays, ELISA, ELISPOT and NGS for genomic analyses will enable exploration of within-host viral evolution, transmission chains and enable characterisation of both humoral and cellular immunity over time,⁴⁶ similar to approaches taken in recent outbreak genomic studies.^{7 29 47} By applying NGS to longitudinal samples, this study will explore within-host viral evolution, transmission chains and the emergence of variants associated with persistence or pathogenicity.²⁹ These data will enhance molecular epidemiological frameworks and may inform outbreak response strategies, particularly in identifying cryptic transmission events or viral adaptation patterns.

From a public health perspective, the study has important implications for post-outbreak surveillance and survivor care. The identification of long-term complications, including ocular, neurological and metabolic sequelae will support the development of evidence-based clinical guidelines for survivors’ management. Characterisation of long-term clinical outcomes in survivors with appropriate control groups has been shown to reveal persistent sequelae years post-infection.³⁸ Furthermore, findings related to viral persistence and transmission risk will directly inform infection prevention policies, including recommendations for safe sexual practices and monitoring of survivors.

The study also emphasises capacity building and local ownership, with laboratory assays and genomic analyses being implemented in Rwanda through international collaborations. This approach not only ensures sustainability but also strengthens national preparedness for future outbreaks of emerging infectious diseases.