Considerations for use of Ebola vaccine during an emergency response

3.1. At-risk countries

Thirteen African countries have documented at least one human case of one or more Ebolavirus species. Four species within the Ebola-virus genus are known to cause human disease: Zaire ebolavirus (EBOV), Sudan ebolavirus (SUDV), Tai Forest ebolavirus (TAFV), and Bundibugyo ebolavirus (BDBV). Outbreaks of more than 300 cases have occurred in the Democratic Republic of Congo (DRC), Guinea, Liberia, Sierra Leone, and Uganda (Fig. 1). EBOV, SUDV, and BDBV have caused multiple outbreaks in Africa with substantial case fatality [11] while TAFV has been documented in only one non-lethal human case related to a chimpanzee outbreak in Ivory Coast [12].

3.2. Human transmission

Human-to-human Ebolavirus transmission occurs through percutaneous or mucous membrane contact with blood or other infected body fluids [13,14]. Skin swabs have yielded virus during periods of high viremia [15] and contaminated surfaces and aerosolized droplets may be infectious [16], necessitating strict infection prevention and control procedures. Household caregivers and HCWs face elevated risk of infection, especially when the infected individual is near death [3–5,7,14,17–26]. Corpses remain infectious after death, and participating in funeral rituals that include washing and handling corpses has been implicated in out-break transmission [13,22,24,27,28]. Additionally, EVD survivors have the potential to transmit virus due to viral persistence in select body fluids during convalescence [29] (see Contacts of survivors section below).

Ebolavirus transmission dynamics are important when considering control strategies and potential impact of vaccination. Although the basic reproductive number (R_o, the average number of cases generated by an index case), is relatively low for Ebolavirus transmission, significant variation in individual transmission exists. R_o estimates for the 2014 outbreak were 1.2–2.5, similar to prior EVD outbreaks [30]. Transmission chain analyses from Guinea showed that the majority of infected individuals transmitted to 0–1 secondary cases, while few transmitted to multiple secondary cases [31,32]. However, the phenomenon of superspreading, where a single case leads to many secondary cases, has been documented [7,27,33] and may be a key factor for sustained transmission. Analyses of community-level safe burial data in Sierra Leone and of surveillance data from all three affected West African countries (Epi Info Viral Hemorrhagic Fever application) suggest that a small number of superspreaders were responsible for 61% [34] and 73% [35] of all infections, respectively. Early vaccination of contacts and contacts of contacts may have the potential to suppress superspreading events and prevent epidemic spread.

3.3. Clinical manifestations & immune response

The incubation period for Ebola infection is 2–21 days. Infection results in a generalized viral illness with early signs and symptoms difficult to distinguish from other endemic febrile illnesses including malaria [7,11,36]; patients may develop “wet” symptoms including nausea/vomiting, diarrhea, and hemorragic signs [2,3,6,7,24, 25, 37]. The incubation and early infection periods are important in relation to vaccination, since Ebola-exposed vaccinees are monitored for EVD symptoms, but may experience similar symptoms (e.g., fever) related to vaccine reactogenicity, particularly for replicating vaccines. Participants in ring vaccination protocols were provided with paracetamol to prevent or treat fever related to reactogenicity. In addition, Ebolavirus PCR testing for the glycoprotein (GP) gene may be positive in vaccinees due to the replication of the vaccine itself.

Humoral and cellular immune correlates of disease and protection are essential to interpret vaccine immunogenicity. The evolution of EVD involves increasing viremia at the most severe and infectious stages of disease, when high titers of virus can be detected in blood and numerous body fluids. Ebolavirus-specific immunoglobulin G (IgG) antibody response has been associated with EVD survival both in primates receiving post-exposure anti-body prophylaxis [38] and in humans in outbreak settings [39]. High viral load is associated with poor outcomes [40]. While evidence shows protection to be primarily antibody based, the role of cellular immune protection is less well understood. CD4+ and CD8+ T-cell function may be important among EVD survivors [41,42], and vaccines that promote CD4+ and CD8+ T-cell responses may improve efficacy.

5.1. Ring vaccination implementation

Ring vaccination is the most likely vaccination strategy (Table 1) to be used in future community Ebolavirus outbreak responses given the experience in the 2014–2016 West African outbreak; this strategy, particularly in the pre-approval period, will require substantial effort and mobilization of resources. In October 2015, the WHO SAGE on Immunization provisionally recommended that EV continue to be used in outbreak response settings for those at high risk of infection [54]. Following publication of the interim Guinea ring vaccination trial results, ring vaccination was used in all three countries under amended clinical trial protocols for “compassionate use” or “expanded access” to give high-risk individuals access to vaccination, including HCW/FLWs. In April 2017, SAGE recommended pre-licensure use of rVSV-ZEBOV using the ring vaccination delivery strategy to include contacts, contacts of contacts, and local and international HCW/ FLWs in outbreak-affected areas and in areas at risk of spread [8]. SAGE also recommended that in case of an outbreak caused by a species other than EBOV, the use of a candidate vaccine designed to target the viral species identified should be considered. However, candidate vaccines for other Ebolavirus species are not yet available.

5.2. Regulatory requirements

Licensure for rVSV-ZEBOV is not expected until 2019 at the earliest. Before licensure, outbreak response ring vaccination could take place within the expanded access framework in the context of appropriate regulatory approvals and a protocol approved by the manufacturer, the country’s national regulatory authority, and appropriate ethical review board(s) (Table 2). As with any clinical trial protocol, an expanded access protocol must include procedures for vaccine handling (including compliance with GCP, individual informed consent, safety monitoring, and adverse events management (follow-up, reporting, and provision of medical care). Duration of safety follow-up may be shorter than required for clinical trials if supported by available safety data for the populations included. Building on experience from prior ring vaccination, a standard protocol for all at-risk countries will facilitate planning across countries. After licensure, vaccine safety will need to be monitored through an adverse events following immunization surveillance system [55].

5.3. Vaccine stock and procurement

For prelicensure supply needs for expanded access or emergency use, Gavi, the Vaccine Alliance, recently entered into an advance purchase agreement with Merck^®, committing $5 million USD towards a stockpile of 300,000 rVSV-ZEBOV doses [56]. After licensure, countries will make requests to the International Coordinating Group (ICG) for the provision of Ebola Vaccine, an international partnership to manage, deploy, finance, and monitor emergency vaccine stockpiles [57]. If, in the future, vaccine is available at the country-level, doses of vaccine should be monitored for quantity and expiration.

5.4. Cold chain capacity

The rVSV-ZEBOV vaccine must be stored at −60 °C, requiring specialized cold chain equipment not used for routine immunization programs in Africa. Specialized equipment for storage and transport of vaccine, including −80 °C freezers, modified Arktek passive storage devices and Creedo cold store boxes were used to support ring vaccination activities in Guinea, Liberia and Sierra Leone; these countries are transitioning these resources to government management for emergency planning. Maintenance of the −60 °C cold chain requires careful coordination to allow effective vaccination in remote areas. Resource needs include equipment (−80 °C freezers and passive storage devices for transport and maintenance of thermostability over several days to weeks) with adequate space, power, backup supply, and trained personnel for monitoring the EV. Vaccine manufacturers aim to provide data to support stability at higher temperatures. Until such data become available EV delivery will be challenging, and in the meantime countries should ensure the availability of ultra-cold freezer capacity either. This could be centrally within the country or at an established regional hub shared by multiple countries with shipping, import/export agreements, and plans in place to rapidly transport vaccine to remote locations.

5.5. Social mobilization

Experience with EV has shown that successful vaccination depends on community vaccine knowledge, acceptance, and cooperation. Just after publication of the interim ring trial results, about 86% of community members in Guinea responded positively about the need for an EV and 84% reported that their family would accept a safe and effective vaccine if offered. Acceptability and interest were highest among those who were more educated, had children who had received immunizations in the routine childhood immunization schedule, understood Ebola virus transmission modes, had witnessed response teams, or knew Ebola-affected persons [58]. Studies in Sierra Leone [59] and Nigeria [60] during the epidemic period also showed good community acceptance of a hypothetical EV. Qualitative research among HCW/FLWs in two regions of Nigeria showed that willingness to receive an EV was significantly higher among individuals who had interacted with EVD cases; fear and hesitancy related to misconceptions about EV safety were mitigated through an education program [61]. In Liberia, ring vaccination activities during an Ebola virus cluster response were difficult to implement due to mistrust and misconceptions about EVD and EV. In one vaccination effort, none of the community contacts, half of HCW contacts, and only a third of the targeted population accepted vaccination, despite community engagement activities (J. Mann, personal communication). Because no community contacts accepted vaccination, the ring vaccination strategy was modified to include a geographically bounded ring to define contacts of contacts.

Broadly targeted education and advocacy may help to avoid vaccine hesitancy in both the pre- and post-licensure phases. Social mobilization activities for vaccines require collaboration among the national ministry of health, the United Nations Children’s Fund and other non-governmental organizations and partners. Coordination with the national immunization program is recommended to avoid confusion between sensitization messages for EV and other childhood immunizations [16]. Targeting at risk populations is particularly important before EV is licensed, since reactive ring vaccination, not mass vaccination, has been recommended by SAGE [8].

5.6. Alternate emergency vaccination strategies

While most EV experience is in the context of ring vaccination, certain scenarios may warrant alternate emergency vaccination strategies including geographic or mass vaccination and reactive HCW/FLW vaccination (Table 1). Both geographically-targeted and mass vaccination strategies might miss individuals who are not present at the time of vaccination, whereas careful listing of contacts as part of ring vaccination might locate those individuals. In instances of widespread transmission, limited vaccine and supplies may be a barrier to mass vaccination. Geographically-targeted vaccination may be most feasible if areas of transmission are well-defined, but high-quality contact tracing is not possible. In a geographically-targeted strategy, a post-vaccination household coverage survey is recommended. In some situations, a combination of strategies may be justified.

5.7. Implementation research

To inform policy decisions, vaccine implementation research is needed to determine the feasibility of introduction in various target populations and settings and sources of vaccine acceptance or hesitancy. Implementation research has been valuable for the introduction for other new vaccines [62].

6.1. Healthcare and frontline workers

Risk of EVD transmission to HCW/FLWs is known to be greater than risk to the general population [23], especially early in an outbreak if infection prevention and control practices are not strong. Preemptive vaccination for HCWs in high-risk countries may avert another large-scale outbreak. In the 13 countries in Africa with history of Ebola outbreaks, there are more than 236,000 HCWs [63], and HCW vaccination coverage against other endemic diseases is limited. Considering resource limitations in these countries, HCW/FLW EV implementation would require external funding sources. A strategy for phased vaccination in HCW/FLW subgroups or regions may be needed. Skrip and Galvani suggest prioritizing subgroups using indicators such as recency of transmission, country capacity for vaccination and monitoring of safety and effectiveness in the population, followed by phased vaccination [63]. International responders, medevac teams, and personnel from high containment laboratories could also benefit from preemptive vaccination.

Data to support duration of immunogenicity of single dose vaccines have not yet been published, but results are expected from clinical trials in Guinea, Liberia, and Sierra Leone up to 12 months post-vaccination, and in the PREVAIL study up to 5 years post-vaccination. Multi-dose boosted regimens may also be promising for long-term protection (see Continued vaccine development: duration of effectiveness). Results will inform whether prophylactic strategies for HCWs will be useful in preventing future outbreaks and the needed frequency of revaccination.

6.2. Contacts of survivors

Prophylactic vaccination of contacts of EVD survivors may prevent new cases. Some EVD survivors may continue to transmit virus in the convalescent phase due to viral persistence over long periods in areas of the body considered “immune privileged”, where the virus is sequestered by a barrier that allows it to evade removal by the immune system [64]. Most evidence on viral persistence comes from semen testing in survivors following recovery after acute infection [36,65–70]. Epidemiologic and molecular investigations suggest rare but important sexual transmission from male survivors during the 2014–2016 epidemic. The virus was documented to persist more than 500 days in the semen of a survivor in Guinea, resulting in sexual transmission and resurgence of disease [29]. Viral persistence and risk of transmission from the eye [71], CNS [72], placenta/amniotic fluid [73], and breast milk [15,74,75] continue to be studied but data are currently limited [76].

In the 2014–2016 epidemic resources to support survivors were directed towards programs for semen testing, sexual transmission prevention education (abstinence, condom use), and clinical care. In early 2016, WHO initiated vaccination of contacts of male survivors in Guinea as part of an investigational protocol. Implementation challenges for vaccinating this population include: (1) risk of transmission may be greater with informal sexual partners (e.g., short-term partners, male partners of men, sex workers) who are not aware of the survivor’s past EVD history; (2) identification and sensitization of sexual partners without causing stigmatization may be difficult; and, (3) the at-risk population likely includes minors who require parental consent for receipt of counseling and vaccine administration. Despite these challenges, this population may be considered for vaccination in a future outbreak if a licensed EV becomes available.

8.1. Multi-species coverage

The ability for a vaccine to provide protective immunity against multiple Ebolavirus species is particularly important, given that three Ebola species have caused human outbreaks (EBOV, SUDV, and BDBV) and are likely to cause future outbreaks. A broadly protective EV would allow at-risk countries to plan around the characteristics of a single product for rapid distribution from global or regional stockpiles. This would be desirable in regions where outbreaks of more than one type of filovirus or Ebolavirus outbreak has occurred, for instance in the DRC and Uganda, which have documented previous outbreaks from multiple filovirus species.

While some evidence of Ebolavirus species serologic cross-reactivity in humans [81] and cross-protection in animal models [82–85] suggests potential for a cross-protective monovalent vaccine, more promising for broad protective immunity are vaccine candidates that are multivalent or contain blended antigens from multiple Ebolavirus species [86]. The MVA booster dose in the Ad26-ZEBOV and MVA-BN-Filo vaccine regimen in clinical trials includes pooled proteins from EBOV, SUDV, TAFV, and Marburg virus. This approach has the potential to provide broad protection, but additional evidence of safety and effectiveness is needed. Additional blended and multivalent vaccines are in early clinical trials.

8.2. Duration of protection

Rapid and durable protection is highly desirable for vaccination during an outbreak, especially to provide prophylactic protection for HCW/FLWs. Two-dose boosted regimens are promising for prolonged protection. Prime boost clinical trials of Ad26-EBOV and MVA-BN-Filo have found European participants to have sustained immune responses 12 months after vaccination at varied prime-boost intervals (15, 29, or 57 days) [87]. Additional immunogenicity results are expected from African trials of these regimens. Longer intervals between the first and second dose are expected to elicit a more robust and prolonged response but possibly at the cost of early protective immunity and the chance of loss to follow-up among vaccinees.

8.3. Safety in special populations

With a replication-competent vaccine such as rVSV-ZEBOV, vaccinees have a period of transient viremia that could be more profound and prolonged in immunocompromised individuals. HIV-infected individuals are being evaluated in rVSV-ZEBOV and Ad26-EBOV/MVA-BN-Filo prime-boost trials. VSV infection in a pregnant vaccinee is also a theoretical risk. Platforms such as non-replicating recombinant vectors, nucleic acid, or VLP vaccines, would be preferable for use in populations with a high prevalence of immunodeficiency (e.g., HIV infection), or where ruling out pregnancy is not feasible before vaccination. HIV-infected individuals are being evaluated in rVSV-ZEBOV and Ad26-EBOV/MVA-BN-Filo prime-boost trials.