The venomous bites no one counts total at least 2.1 million a year
How many snakebites actually inject venom—and how many people are harmed or killed as a result? The researchers behind today’s paper set out to answer that with numbers big enough, and honest enough about uncertainty, to guide real-world decisions.
Their starting problem wasn’t biology but bookkeeping: in much of the world most at risk, bites that cause serious illness never show up in the ledger at all.
Who did what, and where it appears
The study’s estimates and maps come from researchers with experience in snakebite and tropical-disease epidemiology, published in PLOS Medicine in 2026 as “Estimates of global burden of snakebite: A literature review and geostatistical modelling study.” The authors write that their analysis updates their earlier global burden work from 2008.
Throughout this piece we refer to “the study” when citing those findings.
Why bites vanish from the record
The researchers state that over half of the countries where venomous snakes pose a threat don’t keep count of how many people are bitten or envenomed. That is the core data gap the work tries to bridge. In regions where formal reporting is thin, official statistics understate both the scale of illness and where it concentrates.
That absence matters for public health: plans that rely on administrative data alone will undershoot the need where the risk is highest.
What they measured, and how
The team reviewed published evidence up to March 31, 2025, drawing on studies from 79 countries. Because many countries report little or nothing, they turned to a model-based geostatistics framework to fill the blanks. In plain terms, that framework uses the pattern in places with data to infer a probability of envenoming and death in nearby places without data.
The approach is explicit about uncertainty: it does not replace missing counts with a single guess but yields a lower-bound estimate and a plausible upper bound, depending on how sparse the inputs are.
What the numbers say
Globally, the study estimates at least 2.1 million people experience snake envenoming each year. Accounting for underreporting, the true number could be as high as 7 million. On deaths, the model’s range is 274,000 to 513,000 annually.
The burden is not evenly spread. Sub‑Saharan Africa alone accounts for 45% of all global envenomings and fatalities. People there experience envenoming and death at nearly three times the rate seen in South Asia. Beyond survival, the study estimates over 800,000 survivors each year are left with long‑term physical disabilities.
Where non‑venomous bites fit into the picture
The analysis focuses on envenoming—the subset of bites in which venom enters the body—because that is what drives severe illness, death and long‑term disability. That choice matters when counting the burden: surveillance that records “snakebites” without distinguishing envenoming can both overstate danger in some places and, paradoxically, hide it where severe cases never reach a clinic.
By centering envenoming, the study puts policy attention on the bites that translate directly into mortality and disability in vulnerable communities.
The deeper cut
Inside the geostatistics: why a map beats a single number
Model‑based geostatistics treats risk as a spatial process: the log incidence (say, envenomings per 100,000) is modeled as a function of covariates plus a spatially correlated random field. Where surveillance is strong, the posterior is anchored by observed counts; where surveillance is weak, the spatial prior and covariates carry more weight. The payoff is twofold. First, the method yields pixel‑ or district‑level risk surfaces rather than a single global average, which is what planners need to allocate scarce interventions. Second, it returns a full posterior distribution, allowing the authors to report credible lower bounds (e.g., ≥2.1 million envenomings) alongside upper bounds (up to 7 million). That explicit uncertainty is not a hedge—it is information. If a country sits under a wide credible interval, that flags a surveillance problem as much as a disease problem.
The 2008 global burden estimates were largely meta‑analytic, aggregating patchy studies into regional totals. A geostatistical update in 2026 allows partial pooling across space: high‑incidence foci pull up estimates for adjacent under‑counted zones, but not uniformly—correlation decays with distance and differs by region. Practically, that is how the model can reconcile “nearly three times the rate” in sub‑Saharan Africa relative to South Asia while still producing overlapping uncertainty bands at subnational scales.
Limits the authors acknowledge—and what would change the answer
The central limitation is the missing data itself: the study notes that more than half of at‑risk countries do not keep counts of bites or envenomings. Even with model‑based geostatistics, the sparsest areas drive the widest uncertainty, and any systematic bias in the few available inputs could tilt the spatial field.
In practical terms, the true burden would be lower than the study’s lower bounds only if current reports systematically overcount envenomings in key anchor regions, or if areas assumed to be similar are, in fact, protected by unmeasured factors. Conversely, the true burden could exceed the upper bounds if underreporting is worse than represented in the input studies.
What’s known, and what isn’t—once, plainly
Known: the study’s model suggests ≥2.1 million envenomings and 274,000–513,000 deaths each year, with 45% of the global burden in sub‑Saharan Africa and nearly three‑fold higher rates there than in South Asia. Unknown: the exact country‑by‑country counts in places without surveillance; those require better routine reporting to pin down.
Why this matters for 2030 goals
The World Health Organization has set a target to halve snakebite mortality and disability by 2030. Updated, spatially explicit burden estimates are the baseline for measuring progress toward that goal and for focusing efforts where the modeled risk—and the uncertainty—are highest.
| Metric | Estimate / Detail |
|---|---|
| Countries with data reviewed | 79 |
| Evidence cutoff date | March 31, 2025 |
| Annual envenomings (lower bound) | 2.1 million |
| Annual envenomings (upper bound) | 7 million |
| Annual deaths (range) | 274,000–513,000 |
| Share in sub‑Saharan Africa | 45% of global envenomings and fatalities |
| Rate comparison | ~3× higher envenoming and death rates in sub‑Saharan Africa than South Asia |
| Annual survivors with long‑term physical disabilities | >800,000 |
| Method to fill data gaps | Model‑based geostatistics |
| Program target (WHO) | Halve snakebite mortality and disability by 2030 |
The paper: Estimates of global burden of snakebite: A literature review and geostatistical modelling study (PLOS Medicine, 2026)
Related research: The Global Burden of Snakebite: A Literature Analysis and Modelling Based on Regional Estimates of Envenoming and Deaths (PLoS Medicine, 2008); Public health aspects of snakebite care in West Africa: perspectives from Nigeria (Journal of Venomous Animals and Toxins including Tropical Diseases, 2013); Health seeking behavior following snakebites in Sri Lanka: Results of an island wide community based survey (PLOS Neglected Tropical Diseases, 2017)
Sources: Snakebites are a far bigger problem than we knew: new study maps the hidden dangers (phys.org)
Images: Cover: Everyman Science (AI illustration)
How this article was made: Everyman Science uses AI tools to structure, format and optimise its articles, and occasionally to produce illustrations where no free photograph exists. The reporting these articles are based on is human-produced and cited above. Spotted an error? Write to [email protected] and we will correct it. — The editors How we work.
Science desk team of Everyman Science, curating and reporting on the day’s most significant developments in research, space exploration, and technology.
