How Many Humans Are Killed by Dogs Worldwide?

Dogs kill an estimated 59,000 people worldwide each year, with the vast majority of those deaths caused not by mauling but by rabies transmitted through bites and scratches.¹ That figure dwarfs the number of people killed by direct physical attacks from dogs, which accounts for a few hundred deaths per year across high-income countries. The gap between those two realities, and the very different populations they affect, is central to understanding the global picture.

Rabies Is the Overwhelming Cause

When most people in wealthy countries think about fatal dog encounters, they picture a mauling. But globally, rabies spread by dogs accounts for the vast majority of dog-related human deaths. A widely cited modeling study estimated roughly 59,000 annual deaths from canine rabies, along with over 3.7 million disability-adjusted life years lost and $8.6 billion in economic costs each year.¹ Other estimates from separate research groups have placed the toll at 50,000 to 55,000 deaths annually.²

These numbers are far higher than many people expect, largely because rabies has been virtually eliminated in North America and Europe through decades of dog vaccination campaigns. In places where canine rabies still circulates freely, though, a single bite from an infected dog is almost always fatal once symptoms develop. Rabies can present as “furious” rabies, with agitation and hydrophobia, or as a paralytic form, but both are nearly 100 percent lethal without prompt treatment.³ This makes rabies unusual among infectious diseases: it is entirely preventable with vaccination, yet it still kills tens of thousands each year because the vaccine does not reach the people and dogs who need it most.⁴

It is worth noting that a 2022 analysis using the Global Burden of Disease framework arrived at a much lower figure of roughly 13,700 rabies deaths worldwide in 2019.⁵ That study did find a decline from 1990 levels, which is encouraging. But the enormous gap between that estimate and the 59,000 figure illustrates how difficult it is to count rabies deaths at all, a problem explored further below.

Where Deaths Are Concentrated

Over 95 percent of rabies deaths occur in Asia and Africa, with a large proportion of victims being children.² India alone bears the single heaviest burden, with estimates placing the country’s toll at 25,000 to 30,000 rabies deaths per year.² India’s outsized share reflects a combination of factors: a massive population, enormous numbers of free-roaming dogs, limited access to post-exposure treatment in rural areas, and inconsistent enforcement of dog vaccination programs.⁶

Sub-Saharan Africa is the other major hotspot, where rabies surveillance is weakest and access to both dog vaccination and human post-exposure prophylaxis is patchy. Countries in Latin America, by contrast, have made dramatic progress over the past few decades through coordinated mass dog vaccination campaigns, reducing human rabies deaths to near zero in many areas. The key variable across all these regions is not the number of dogs per se but whether those dog populations are vaccinated. Where canine rabies has been eliminated through vaccination, the threat essentially vanishes.

Fatal Dog Attacks in High-Income Countries

Strip away rabies and the picture changes completely. In countries where canine rabies has been eliminated, fatal dog encounters are overwhelmingly physical attacks: maulings that cause fatal bleeding, crushing injuries, or, less commonly, infections that turn deadly afterward. These events are real and tragic, but they are rare by any statistical standard.

A European study tallied fatalities recorded under the ICD-10 code W54 (“bitten or struck by dog”), which covers deaths from direct attacks but excludes rabies and other infectious complications. In 2016, 45 people died from dog attacks across Europe, a rate of about 0.009 per 100,000 inhabitants. The United States had a comparable rate of 0.011 per 100,000, and Canada’s was 0.007.⁷ In absolute numbers, the U.S. averages roughly 30 to 50 fatal dog attacks per year, depending on the source and methodology used. An analysis of U.S. death certificates from 1999 through 2016 identified 553 dog-encounter fatalities across that entire 18-year span.⁸

To put that in perspective, you are far more likely to die from a fall, a car accident, or a lightning strike than from a dog attack in any of these countries. The European study did note that fatality numbers had been climbing over its study period from 1995 to 2016, which may reflect changes in dog ownership patterns, the popularity of larger breeds, or simply better reporting.

How Fatal Attacks Happen

Fatal maulings follow certain recurring patterns. A Kentucky medical examiner review found that in 10 out of 11 fatal cases, the victim suffered multiple bite marks and blunt force injuries to the head, neck, trunk, and limbs.⁹ The sheer volume of injury is what distinguishes fatal attacks from the millions of dog bites that happen each year and heal without lasting harm. In one case from that same series, the victim’s immediate cause of death was bacterial sepsis from the bite rather than the trauma itself; that person had no spleen, which left them unable to fight off infection.⁹

Infection as a delayed cause of death is uncommon but not unheard of. A case report described a patient with advanced liver disease who died from a respiratory infection that started as a soft tissue infection from an untreated dog bite. The combination of an underlying immune-compromised state and a lack of treatment allowed the infection to spread systemically.¹⁰ These cases are extreme outliers, but they highlight that dog bites can become dangerous even when the bite itself is not severe, particularly for people with weakened immune systems.

A Canadian study of fatal attacks from 1990 to 2007 found that most involved owned, known dogs in residential settings. Children who had unsupervised access to areas with dogs were heavily represented. Rural and remote areas had higher rates, and free-roaming dog packs in rural communities were responsible for the majority of fatalities on Indigenous reserves.¹¹ The involvement of multiple dogs rather than a single dog was another common thread. Attacks by packs tend to be more sustained and harder to interrupt.

Children Are Disproportionately at Risk

Across every setting studied, young children are overrepresented among both fatal and nonfatal dog bite victims. Among U.S. dog-encounter fatalities from 1999 to 2016, about 30 percent of victims were children aged four or younger.⁸ In studies of nonfatal pediatric dog bites, children under six account for roughly half of all injuries in the 0-to-19 age group, and they suffer bites to the head, neck, and face at much higher rates than older children or adults. Those bites are more often severe and more frequently need surgery.¹²

The reasons are straightforward. Young children are small enough that their heads and faces are at the dog’s mouth level. They have less ability to recognize threat signals from a dog, less strength to defend themselves, and they are more likely to provoke a dog unintentionally through sudden movements or face-to-face contact. A study of pediatric facial dog bites found that the average victim was about four and a half years old, boys were slightly more often bitten than girls, and a familiar dog, not a stray, was usually responsible. Children five and under were also more likely to need hospitalization.¹³ The consistent finding that known dogs in home settings are behind most attacks on young children underscores that supervision is the most practical prevention tool for this age group.

Why the True Death Toll Is So Uncertain

The wide range in global estimates, from roughly 14,000 to 59,000, is not just a matter of different counting years or rounding. It reflects a genuine surveillance gap. In the countries where canine rabies kills the most people, the infrastructure to count those deaths accurately is weakest. A key review of surveillance limitations pointed out that passive surveillance data for canine rabies in the highest-burden regions are inadequate for decision making. Existing legislation is poorly enforced, international guidance on reporting is inconsistently implemented, and a potentially large proportion of rabies victims die outside the health system entirely.¹⁴

Rabies disproportionately affects poor, rural communities, places where a person bitten by a dog may not have access to a clinic, let alone a diagnostic laboratory. If someone dies at home in a remote village and no autopsy or lab confirmation is performed, that death likely never enters any national database. Researchers have tried to compensate for this with statistical models that extrapolate from bite incidence, dog rabies rates, and the probability of exposure leading to death when treatment is unavailable. Those models are useful but carry wide confidence intervals. The 59,000 figure, for instance, comes with 95 percent confidence bounds stretching from 25,000 to 159,000.¹ So the real number could be substantially lower or several times higher than the central estimate, and nobody can say for certain which end is closer to the truth.

What Actually Reduces Dog-Related Deaths

The most effective intervention against the global toll is mass dog vaccination. Canine rabies has been eliminated through mass dog vaccination in much of the Americas, Europe, Japan, and parts of Southeast Asia. A review of the evidence concluded that eliminating canine rabies is both epidemiologically and operationally feasible across a wide range of settings in Africa and Asia, the two continents where most deaths still occur.¹⁵ The roadblock is not science but logistics and money: vaccinating enough dogs in vast, underserved regions to break the chain of transmission.

On the human side, post-exposure prophylaxis (PEP), a series of vaccine doses given promptly after a bite, is highly effective at preventing rabies if administered before symptoms appear. The problem is that PEP is expensive relative to incomes in the countries that need it most, and vaccine supplies are often unavailable in rural clinics.¹⁶ Researchers have investigated switching from the standard intramuscular delivery method to intradermal injection, which uses smaller doses and stretches vaccine supplies further. One modeling study concluded that a universal switch to intradermal delivery would improve both affordability and accessibility, potentially reducing deaths.¹⁷

Managing free-roaming dog populations is another approach. Fertility control programs, when they reach a high enough proportion of female dogs, have been linked to reductions in human bite rates.¹⁸ Culling programs, by contrast, tend to backfire. Modeling work has shown that lethal removal programs skew dog populations toward younger animals, which are more likely to be aggressive and come into conflict with people.¹⁹ Removing adult dogs also opens ecological space that gets refilled quickly through breeding and immigration from surrounding areas, meaning the population bounces back with a less stable age structure.

Breed-Specific Legislation and Its Track Record

In high-income countries where rabies is not the issue, debate over fatal dog attacks often centers on whether certain breeds should be restricted or banned. Several countries and jurisdictions have enacted breed-specific legislation targeting breeds perceived as more dangerous. The evidence on whether this approach works is not encouraging.

A ten-year retrospective study in Spain found that the implementation of breed-specific legislation in 1999 and 2002 did not produce a reduction in dog bite-related fatalities.²⁰ A Danish study using time-series analysis reached a similar conclusion, finding that banning specific breeds had a highly limited effect on overall dog bite injury rates, and that requiring muzzles and leashes in public for those breeds also had little measurable impact.²¹ The reasons likely include the difficulty of reliably identifying breeds by appearance, the substitution of other large breeds by owners who want imposing dogs, and the fact that most fatal attacks involve failures of supervision and housing rather than breed alone.

One U.S. trauma center’s data did find that among patients whose attacking dog’s breed was recorded, pit bull attacks were associated with higher injury severity scores, higher hospital charges, and a higher risk of death compared with other breeds.²² The sample was small, and breed identification was available for only about a third of patients, but it illustrates why the debate persists: some breeds do appear capable of inflicting more severe injuries when an attack occurs. Whether legislation targeting those breeds reduces total injuries at a population level, however, is a different question, and the available evidence suggests it does not.

The Cost of Nonfatal Bites

Fatal encounters represent only the tip of a much larger problem. Tens of millions of dog bites occur worldwide each year, and even in countries where deaths are rare, the medical and economic burden is substantial. A study of dog and cat bite injuries to the hand found that patients requiring long-term intravenous antibiotics or multiple surgeries accumulated medical expenses exceeding $77,000 per case.²³ Hand injuries from dog bites are particularly consequential because of the complex anatomy involved: tendons, joints, and nerves packed into a small space, all vulnerable to both the initial trauma and subsequent infection.

Beyond direct medical costs, dog bites carry psychological consequences that are harder to quantify. Lasting fear of dogs, post-traumatic stress symptoms, and the impact of disfiguring facial scars in children can affect quality of life for years. These burdens never make it into the mortality statistics, but they are part of the full picture of how dog-human conflict plays out. For the estimated 59,000 families per year who lose someone to canine rabies, and the hundreds more who lose a child or loved one to a mauling, the numbers are not abstract. They represent a problem that is, frustratingly, almost entirely preventable with tools we already have.