Proof of concept of mass dog vaccination for the control and elimination of canine rabies

Summary

For more than 100 years, canine rabies vaccination has been available as a tool for rabies control and elimination. However, domestic dogs still remain a major reservoir for rabies, and although canine rabies has been eliminated through mass dog vaccination in some parts of the world, the disease continues to kill tens of thousands of people every year in Africa and Asia. This review focuses on the situation on those two continents, presenting evidence to show that canine rabies elimination is both epidemiologically and operationally feasible, and could be achieved across a wide range of settings in Africa and Asia. The challenges of achieving the large-scale, comprehensive dog vaccination coverage that is required are discussed, and opportunities for developing new strategies that generate multiple benefits for human and animal health and welfare are highlighted. Finally, the substantial progress that has been made in developing the tools, partnerships and frameworks needed to move towards global canine rabies elimination is outlined.

Introduction

Canine rabies is one of the world’s deadliest and most terrifying diseases and still poses a substantial public health problem in many parts of the world, particularly Asia and Africa, where tens of thousands of people die from dog-mediated rabies every year (1). Yet rabies is also a disease that is highly amenable to prevention, control and elimination, with scientific interest and policy successes in rabies control and elimination going back many decades.

Human rabies is 100% preventable through two complementary measures: firstly, post-exposure prophylaxis (PEP), which involves the administration of rabies immunoglobulin and a multi-dose course of rabies vaccination to people bitten by suspected rabid animals; and secondly, mass vaccination of animal reservoirs (primarily domestic dogs, the reservoir in the vast majority of human cases), which reduces the risk of human exposure and can ultimately result in rabies virus (RABV) elimination.

While PEP is highly effective, the intervention is expensive, with direct expenditure on PEP estimated at US$ 1.7 billion per year and indirect costs estimated at US$ 1.3 billion (1). Furthermore, many challenges remain for poorer people in remote, rural communities in accessing and completing PEP regimens (2, 3). As a result, human deaths still occur in countries with endemic canine rabies, despite substantial investments in PEP made in many countries, particularly in Asia (1). A complementary approach is needed to control infection at its source, targeting the animal reservoir. Controlling and eliminating rabies in the domestic dog population would not only reduce or remove the source of infection for more than 99% of human rabies deaths, but would also provide the most equitable approach to human rabies prevention, protecting people in the poorest communities (4).

While the effective alignment of PEP and dog vaccination is a central element of contemporary ‘One Health’ frameworks for human rabies prevention (5), these dual approaches have long been recognised. In pioneering the first human rabies vaccines, Louis Pasteur and Emile Roux initially demonstrated the effectiveness of an attenuated strain of RABV as a vaccine for protecting dogs (6). Despite recognising the need for interventions in the dog population, the onerous procedures for preparing vaccines at that time (drying rabbit spinal cords over potash) ruled out large-scale canine vaccination (6). With the development of rabies vaccines for dogs in 1915 (7), canine vaccination became a reality and was implemented on a large scale from the 1920s. This resulted in the elimination of canine rabies in several parts of the world; for example, in island and peninsula states in Asia (e.g. Japan, Chinese Taipei and Malaysia), in the United States of America (USA), in Western Europe and across parts of Latin America (7, 8, 9).

The ‘proof of principle’ for mass dog vaccination as part of the effective control and elimination of canine rabies has therefore been well demonstrated in several high- or high-to-middle income settings. The focus of this review is on canine rabies in Africa and Asia; first reviewing the evidence for the feasibility of canine rabies elimination through mass dog vaccination in these settings, and, second, by highlighting some of the challenges that low- and middle-income countries (LMICs) still face.

The feasibility of canine rabies elimination

Three key themes dominate the debate around canine rabies elimination in Asia and Africa:

• –the feasibility of reaching and sustaining adequate vaccination coverage in the domestic dog population to interrupt transmission and achieve elimination
• –the potential contribution of dog population management to rabies control and elimination
• –the role of wildlife as reservoirs of canine RABV.

What level of dog vaccination coverage is sufficient to eliminate canine rabies?

In 1884, Etienne Nocard, a colleague of Louis Pasteur, questioned the feasibility of canine rabies elimination by asking: ‘For the removal of rabies – would it be feasible to vaccinate all the dogs?’ (7). However, since the concept of herd immunity was introduced in 1923, it is well recognised that not all individuals have to be vaccinated to protect the population (10, 11). The required level of vaccination (P_crit) depends on the basic reproductive number, R ₀, which describes the number of new infections generated, on average, by one infected individual in a fully susceptible population, and is described by the expression, P_crit = 1 – 1/R ₀ (12) (Fig. 1). The higher the value of R ₀, the greater the vaccination coverage required to achieve herd immunity and break disease transmission. Once vaccination is introduced, transmission is interrupted, and if sufficient hosts are vaccinated to bring the effective reproductive number (R) below 1, the infection cannot be sustained and elimination should be feasible (Fig. 1).

Fig. 1. Scheme illustrating the concept of the basic reproduction number, R ₀, and the potential for controlling and eliminating rabies through mass dog vaccination.

The clear icons represent susceptible dogs, the red icons represent rabid (infectious) dogs and the black icons represent vaccinated dogs. When adequate levels of dog vaccination coverage are obtained, the effective reproduction number (R) falls below 1, transmission is interrupted, canine rabies is brought under control and will ultimately be eliminated

Different approaches have been used to measure R ₀ for rabies, but all demonstrate a value that consistently falls between one and two, even across populations that differ widely in dog density (13, 14, 15, 16, 17, 18). This suggests that controlling rabies should be feasible in most, if not all, domestic dog populations across the world.

In line with these predictions, theoretical and empirical research demonstrates that rabies can be eliminated when 70% coverage is achieved (13, 14). Furthermore, economic analyses of vaccination campaigns carried out in rural Tanzania indicate that 70% coverage also represents the optimal scenario in terms of cost effectiveness (19).

While the low value of R ₀ suggests that control of dog rabies could be achieved at vaccination levels lower than 70%, this threshold provides a useful target in LMICs, where dog vaccination is usually implemented during annual campaigns, and where dog populations are typically characterised by high birth and death rates. As a result, population immunity declines rapidly between campaigns, and a campaign target of 70% is required for population immunity to remain at all times above the critical level (25%–40%) (Fig. 2), determined by values of R ₀. If vaccines were more widely available throughout the year, these declines could be reduced; pups could be vaccinated soon after birth and susceptible dogs, brought in to replace those that die, could be vaccinated soon after being acquired (Fig. 2). Strategies that allow greater access to vaccines throughout the year, particularly in remote and under-served communities, may become viable through the use of high-quality, thermotolerant dog rabies vaccines (21).

Fig. 2.

Figure showing (a) the decline in vaccination coverage resulting from annual mass vaccination campaigns, due to births of susceptible pups and natural mortality of vaccinated dogs (blue line) and (b) coverage sustained through programmes that ensure that puppies and other unvaccinated dogs are vaccinated throughout the year (green line)

Birth rates and death rates are derived from rural Tanzanian dog populations (20). Coverage achieved is predicted to remain above the critical threshold (lower dashed red line) if the annual campaign target of at least 70% (upper dashed red line) is reached

Operational research also suggests that reaching the required level of dog vaccination coverage is feasible. In sub-Saharan Africa, the vast majority of dogs have owners and, where campaigns are well organised and vaccine is delivered free of charge, there has been sufficient access to these animals to achieve the target level of vaccination coverage (22, 23).

In South and South-East Asia, the situation may be more challenging, as the result of a larger population of less accessible neighbourhood or ‘street’ dogs, but target levels of vaccination coverage have also been achieved in these communities (15, 24, 25, 26, 27).

While there is no evidence of any underlying cultural barriers to vaccinating dogs against rabies, several challenges remain when organising and planning such campaigns to ensure sufficient coverage. Community engagement, awareness and participation are critical, and strategies need to be adaptable and optimise all available resources. Understanding local contexts is important to ensure effective implementation; for example, in relation to disseminating information about the campaigns, the location of vaccination stations, incentives to participate and the timing of campaigns (28, 29, 30, 31). Vaccination campaigns must reach all communities, as even small gaps in coverage can hinder the progress towards elimination (15). The importance of detecting coverage gaps means that practical and cost-effective approaches are needed to evaluate vaccination coverage at the community level (32).

What is the contribution of dog population management to rabies control?

An important implication of the finding that R ₀ varies little with dog population density is that control measures that aim to reduce dog density, such as culling, are likely to be ineffective (33). For example, rabies transmission in the Ngorongoro District, Tanzania, with a density of 1.4 dogs/km² is very similar to that seen in Bali, Indonesia, with a density of 250 dogs/km² (14, 15). Thus, dog numbers in Bali would have to be reduced to unfeasibly low levels (e.g. reducing a population of >200 dogs/km² to <1 dog/km²) to have any impact on RABV transmission. Culling campaigns are not only ineffective but, worse, can accelerate the spread of disease as people move dogs away from culling areas (15). Indiscriminate culling can also lead to further problems, including:

• –a decline in population immunity, with vaccinated dogs being killed and susceptible dogs brought in as replacements
• –increased community hostility and antagonism
• –welfare concerns.

Although authorities often feel under pressure to respond to rabies outbreaks by reducing dog populations through culling, efforts must focus on mass dog vaccination.

While there is broad consensus that indiscriminate culling has no role in rabies control, considerable debate remains about the role of other aspects of population management (34). In theory, reducing population turnover (e.g. through interventions that improve life expectancy and/or reduce fecundity) could reduce the decline in vaccination coverage between campaigns (Fig. 2), and thereby sustain population immunity. Current interventions focus largely on surgical sterilisation, and contributions to rabies control are proposed on the basis of reduced birth rates (and hence population turnover), fewer dogs (leading to reduced dog vaccination costs) and a decrease in aggressive behaviour (thus potentially reducing human dog-bite injuries and PEP costs).

However, there is still only limited empirical evidence that dog population management tools have improved the cost-effectiveness of human rabies prevention, over and above mass dog vaccination alone (34). Existing strategies that rely on surgical dog sterilisation are unlikely to be cost effective. Even in populations in India, with a large proportion of unowned ‘street dogs’, dog vaccination is more cost effective, in terms of achieving human health outcomes, than combined strategies involving vaccination and surgical sterilisation (18). The development of cheaper methods of sterilisation could enhance opportunities for more cost-effective linked strategies. Nonetheless, control options must allow for rabies vaccination to be delivered independently of sterilisation, to avoid any perception of rabies vaccination being associated with enforced sterilisation.

Strategies that focus on improving dog life expectancy have generally received less attention than reproductive control. Surgical sterilisation has been reported to increase life expectancy (35), but other simple interventions, such as de-worming, also have the potential to enhance longevity, as well as addressing other disease problems. These include cystic echinococcosis, an important public health concern in several countries of Asia and Africa (36), and coenurosis, a major cause of small ruminant mortality in some parts of Africa (37).

Where free-roaming dogs cause a nuisance, a further benefit of dog sterilisation has been improved attitudes and tolerance towards dogs (35, 38), resulting in greater ‘buy-in’ from authorities and communities for rabies control efforts (34). Indeed, several rabies control programmes have been initiated primarily in response to demands to manage the problems associated with poorly supervised neighbourhood dogs (38, 39, 40, 41). In general, increasing dog-owner access to animal health services that address a wider suite of diseases and/or nuisance problems is likely to result in more responsible dog ownership, which is widely advocated as part of all successful rabies control strategies (5).

One major limitation is the lack of data to formally quantify the impact of dog sterilisation or other management tools on rabies control outcomes. With substantial investments being directed towards dog population management, there is an urgent need to collect such indicator data, and clear opportunities to do so exist within ongoing humane dog population control programmes (42).

Does wildlife rabies affect the feasibility of canine rabies elimination?

Rabies can infect all mammalian species, and the abundance of susceptible wildlife hosts, particularly in Africa, has often been considered a major obstacle for dog rabies control that would render elimination efforts futile (43). However, although rabies can infect all mammals, not all species can maintain infection and act as reservoirs. Various features of host demography, ecology and virus-host interactions mean that only a few hosts are capable of maintaining infection as reservoirs (44, 45).

Unravelling the infection dynamics of multi-host pathogens is challenging and requires the integration of multiple lines of evidence. In the wildlife-rich Serengeti ecosystem in Tanzania, observational studies, epidemiological modelling and genetic analyses suggest that domestic dogs are the only population essential for rabies maintenance, with occasional short-lived chains of infection in wildlife resulting from spillover from domestic dogs (46, 47, 48).

This is supported by data generated through detailed contacttracing studies in wildlife-rich communities adjacent to the Serengeti National Park (14). Where dog rabies has been well controlled through mass dog vaccination, wildlife cases have also disappeared. They have only re-occurred in the wake of well-documented incursions of dog rabies from areas beyond the vaccination zone (Fig. 3). Despite regular reports of wildlife rabies in the Serengeti National Park throughout the late 1980s and 1990s (47), only one wildlife case has been documented inside the park since 2004, the period during which mass dog rabies vaccination has been implemented (Fig. 2), with no evidence of onward or sustained transmission in wildlife. A further case in a civet (Civettictis civetta) was caused by a novel lyssavirus, highly divergent from canine RABV (49) and unconnected to any other cases.

Fig. 3. Rabies cases in domestic dogs (grey) and wildlife (red), identified through detailed contact tracing studies in Ngorongoro District, Tanzania (49).

Mass dog vaccination was initiated in 2003 (shown by a black arrow). The timing of wildlife cases occurring in neighbouring Serengeti National Park is shown by the star symbols. A case in a civet was caused by the highly divergent Ikoma lyssavirus in 2009 (⋆), and one in an aardwolf (Proteles cristata) was caused by canine rabies in 2013 (⋆), with no evidence of onward transmission

While wildlife in the Serengeti ecosystem appear not to sustain independent cycles of infection, the extent to which this applies more globally is unclear. In Africa, considerable debate has revolved around the role of jackals (Canis mesomelas, Canis adjustus) as potential reservoirs. Although there is genetic evidence for the independent circulation of a canine RABV variant in jackals in northern South Africa (50), it remains to be seen whether this cycle will be sustained in the absence of dog rabies, which has now been well controlled in the area through large-scale, mass vaccination campaigns (51). If so, vaccination of jackals will likely be needed to achieve elimination in a few localities. However, this should not represent an insurmountable barrier, given earlier studies that have demonstrated the safety, efficacy and feasibility of oral vaccination in jackals (52, 53, 54).

What are some of the key challenges to achieving global canine rabies elimination?

Despite the optimism generated by epidemiological and operational research from pilot studies, awareness is growing about the challenges faced in reaching and sustaining the vaccination coverage needed to achieve rabies elimination, as well as the timescales required to move from control to elimination. These have been exemplified in Latin America, where successes have been underpinned by long-term investments to sustain dog vaccination over large geographical areas and by coordination across national boundaries (55). In Africa and Asia, progress has been made in the development of operational tools to support national rabies plans and in the establishment of regional rabies networks (56, 57). However, only a few countries have started to implement large-scale control programmes and these will require a level of investment in mass dog vaccination over and above anything seen to date (58). The recent launch of a strategic plan to end human deaths from dog-mediated rabies by 2030 (‘Zero by Thirty’) by the ‘United Against Rabies’ collaboration – the World Health Organization (WHO), the Food and Agriculture Organization of the United Nations (FAO), the World Organisation for Animal Health (OIE), and the Global Alliance for Rabies Control (GARC) – should act as an important catalyst to progress (59).

In summary, a large body of historical and contemporary evidence suggests that canine rabies elimination is feasible through interventions based on mass dog vaccination, and that reaching an adequate level of coverage to control and eliminate rabies is feasible in dog populations across Africa and Asia, where canine rabies remains endemic. The veterinary profession was pivotal in achieving the elimination of another multi-host pathogen, rinderpest, through mass vaccination campaigns. There is no reason why this success could not be repeated for canine rabies, so that dog-mediated rabies becomes the first zoonotic disease to be eliminated worldwide.