Introduction

In 2017, the World Health Organization (WHO) designated snakebite envenoming a neglected tropical disease [1], which was followed two years later by the launch of the Snakebite Strategy with the aim of halving mortality and morbidity from snakebite by 2030 [2]. The WHO estimates that each year 5 million people are bitten by snakes, with up to 2.7 million people becoming critically ill, 130,000 suffering mortality, and a further 400,000 becoming physically disabled [3]. In sub-Saharan Africa, annual estimates are 1 million snakebites with only 360,000 presenting to health facilities resulting in approximately 20,000 deaths [4,5]. Though robust studies are rare, snakebite incidence estimates have been reported from some African countries including Cameroon, Kenya, Mozambique, Nigeria, Rwanda, Senegal and South Africa [6–10].

Globally, the incidence and severity of snakebite envenoming has been underestimated with the condition receiving little attention [4,6,10,11]. There are a number of challenges in gathering accurate data on incidence and severity, which include poor recording of snakebite cases that present to health facilities, patients’ willingness to come to health facilities and their confidence in the health system to treat snakebite when there is lack of timely access to affordable and effective antivenom treatment, and patient beliefs and/or health seeking behaviour [12–16]. Snakebite victims commonly report self-administering first aid measures, or consulting traditional healers instead of, or prior to, coming to a health facility [11–13,17]. Patients who die before reaching a healthcare facility or who choose to only seek treatment from a traditional healer will not present to the health system and therefore will not be counted in routinely collected national statistics.

Epidemiological studies from Africa and Asia show snakebite envenoming to disproportionally affect rural, farming and impoverished populations [5,18–20]. Children and young adults comprise the majority of the cases [5,7–9,21,22] followed, to a lesser extent, by the elderly [23,24]. In addition, susceptibility to snakebite is increased in situations where individuals live in houses whose construction does not prevent snake ingress and that have limited lighting and sanitation facilities [25] or where people need to collect firewood [8], factors that are underpinned by poverty. Understanding and modelling risk of snakebite may be one strategy to reduce the incidence and improve treatment outcomes of snakebite [26–29]. These studies have used snakebite incidence data, habitat suitability of common venomous snakes, land use land cover, geographic and socioeconomic status of snakebite patients, and/or venomous species distribution and occurrence data to geo-spatially model the risk of snakebite within a particular region of a country.

Eswatini, a landlocked country in southern Africa, has a population of 1.3 million, of which almost 80% live rurally and 30% have a household income of under US$1.90 a day [30]. Despite its small area and population, snakebite is an important and challenging public health issue [13,31,32]. The country is geographically diverse, with four distinct agroecological zones determined by elevation and rainfall [33], creating suitable habitat for over 60 snake species in 39 genera [34,35]. Eleven of these species are venomous and medically important, including seven species that bite frequently and are associated with serious or life-threatening envenoming (category 1). Four species bite frequently but rarely cause serious or life threatening envenoming (category 2), as classified by WHO [36]. The seven category 1 species are the puff adder (Bitis arietans), black mamba (Dendroaspis polylepis), boomslang (Dispholidus typus), rinkhals (Hemachatus haemachatus), snouted cobra (Naja annulifera), Mozambique spitting cobra (Naja mossambica), and vine snake (Thelotornis capensis) [31]. The four category 2 species are berg adder (Bitis atropos), Bibron’s stiletto snake (Atractaspsis bibronii), rhombic night adder (Causus rhombeatus), and snouted night adder (Causus defillippii).

The number of bites by these medically important snakes in Eswatini has not previously been reported. In conjunction with our other studies assessing antivenom provision for Eswatini [37], the objectives of this study were to: 1) determine the number of snakebites in Eswatini by compiling those reported to health facilities; 2) describe the epidemiology of snakebite in the country and associated factors; and 3) to develop two risk maps: one for the probability of getting a snakebite and another for the probability of a poor outcome for a snakebite victim.

Methods

Results

A total of 932 snakebites were recorded between October 2019 and September 2021. Most patients were bitten during the summer months between November to March each year, peaking in January (Fig 1). The average rainfall was 126 mm (71–253 mm), and mean minimum and maximum temperatures were 17 ⁰C (14 ⁰C– 19 ⁰C) and 27 ⁰C (24 ⁰C– 32 ⁰C) respectively during the peak biting months (S1 Fig). Patients reported being bitten throughout the 24-hour period, with 30% of snakebite patients reporting being bitten between 18:00 and 22:00 (S2 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Frequency of snakebite patients presenting to health facilities in Eswatini between October 2019 to September 2021.

https://doi.org/10.1371/journal.pntd.0011732.g001

Over 35% of patients were asymptomatic when they presented to the health facility. Just over half (55%) of bites were cytotoxic causing either mild swelling, painful progressive swelling or venom ophthalmia. Neurotoxic bites accounted for less than 5% of all bites and only two patients (0.2%) were recorded to present with bleeding syndrome. Bleeding syndrome is identified using the 20-minute whole blood clotting test [79], which should be carried out on all patients presenting with snakebite as part of the Eswatini national snakebite management guidelines [80]. The result of this test was not available in the snakebite patient database.

Geocoding snakebite localities revealed bites occurred across all elevations, ranging between 71 m to 1499 m (Fig 2). Generalised linear modelling with negative binomial demonstrated elevation to significantly explain snakebite incidence (p = 0.0137, pseudo R² = 0.03), with 3.8% (SE = ±1.4%) decrease in snakebite incidence for every 100 m gained in elevation. None of the other variables (gender, age, and occupation of the snakebite victims) in the model were significant. There were no associations between the seasons (spring, summer and autumn) and elevation. However, during the dry winter months of June to August, twenty-two patients presented with snakebite, with fifteen of these patients bitten at low elevations (below 500 m) and only one patient reported to be bitten at an elevation above 1000 m.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Incidence map of snakebites in Eswatini between October 2019 to September 2021.

Made using shape files from the Eswatini Central Statistics Office (with permission).

https://doi.org/10.1371/journal.pntd.0011732.g002

Sociodemographic characteristics of snakebite patients

More male than female snakebite patients were reported at the 16 health facilities, with the relative risk for males to females calculated as 1.29 (95% CI = 1.13–1.46; Table 1). Age of snakebite patients ranged from 0–100 years, with 55% of bites occurring in those under the age of 30 years, which reflects the population [42]. The age group that presented with snakebite to health facilities the most was 10–19 years (27%; Table 1). The relative risk for children aged 0–9 years was 0.51 (95% CI 0.42–0.62) and for 10–19-year-olds was 1.46 (95% CI 1.26–1.69).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sociodemographic overview of snakebite patients in Eswatini.

https://doi.org/10.1371/journal.pntd.0011732.t001

Most snakebite patients (35%) were students (ages 6 years and above), followed by those unemployed (23%), in elementary low paid occupations such as unskilled labourers and domestic workers (7%) and retirees (6%) (Table 1).

Variables associated with snakebite susceptibility

Most bites (65%) occurred outdoors, 13% were indoors, and the location of over 20% of participants was not recorded. Over half (60%) of all bites were to the lower extremities of the foot, ankle, heel or toe, whilst 14% were to the hand, finger or wrist. The position of the bite was significantly and moderately strongly associated with the bite circumstances (χ² = 327.64, d.f. = 64, p-value < 0.001; Cramer’s V = 0.24). Analysis of the activities undertaken when the bite occurred revealed 27% of snakebite patients were walking, with bites to the lower extremities associated with walking, fetching water or being outdoors (Fig 3A). Age was significantly but weakly associated with the position of the bite on the body (χ² = 59.02, d.f. = 24, p-value < 0.001; Cramer’s V = 0.15). Bites to the leg or foot were associated with younger patients, whilst bites to the arm or hand were associated with older patients (Fig 3B). Gender was not significantly associated with anatomical site of snakebite.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

(A) Body part bitten by snake by activity of the person at the time of bite, in Eswatini; (B) Body part bitten by snake by age group of snakebite patient, in Eswatini.

https://doi.org/10.1371/journal.pntd.0011732.g003

Assessing snakebite risk

Hazard risk.

Species richness of the eleven venomous species considered in this study peaks in the eastern parts of South Africa, Eswatini and neighbouring parts of Mozambique (Fig 4A), with only minor variation in species richness across Eswatini. Based on the modelled distribution of these species, any location in Eswatini harbours between 8–11 species of venomous snakes (S4 Fig). The modelled snakebite hazard (Fig 4B) indicates that areas with the lowest hazard risk correspond to urban areas, protected and conservation areas where human populations are low, as well as areas bordering South Africa in the west and southwest of Eswatini, and those bordering Mozambique in the east. Considering the relatively uniform species richness across the country, the areas with elevated snakebite hazard correspond to populated areas in both rural and peri urban areas.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4.

Species richness of the eleven venomous snakes (considered in this study) in Southern Africa (A), and the hazard (B), exposure (C), susceptibility (D) and health facility access scarcity (E) risk maps associated with snakebite in Eswatini. For maps B-D, the darker the area the greater the risk. For map E, the darker the area the scarcer the access to health facilities. Map A made with Natural Earth. Free vector and raster map data @naturalearthdata.com. Maps B-E made with shape files from the Eswatini Central Statistics Office (with permission).

https://doi.org/10.1371/journal.pntd.0011732.g004

Exposure risk.

The snakebite exposure risk was not uniform throughout the country, with the highest risk of exposure (Fig 4C) associated with densely populated and built-up areas. High resolution images of these high-risk areas indicate these to be peri urban areas close to tar roads and towns, with a mixture of land uses ranging from homesteads on small plots that had livestock and some crops to industrial estates. Lowest risk areas are characterized by very low human population density such as protected and conservation areas as well as commercial farmlands.

Susceptibility risk.

Susceptibility shows marked contrasts particularly between rural and urban environments. Notably, the southern administrative region of Shiselweni had the highest susceptibility (Fig 4D). Similarly, most rural areas exhibit high susceptibility to snakebites. High resolution imagery of these areas indicates that these high-risk areas correspond to rural communities that are typically situated away from towns and accessible mainly by unpaved roads. Such areas are also characterized by subsistence farming, higher rates of unemployment and poverty, and most use firewood for cooking, rely on the river or a communal tap for water in addition to typically having poor sanitation facilities such as outside pit latrines.

Access to health care (Resource scarcity).

Health facilities with the capacity to treat snakebite are distributed across the country (Fig 4E), however their accessibility, and consequently resource scarcity was unevenly distributed. Of the 20 facilities that can treat snakebites, most are concentrated within or in close proximity to urban areas. Where two or more facilities were situated next to or close by each other (i.e., in south of Hhohho region and east of Manzini region), this was reflective of private health facilities being predominantly located close to government health facilities, which invariably were in urban areas. As such, areas adjacent to health facilities with the capacity to treat snakebite had the lowest resource scarcity ranking.

Health facilities in northeast of Hhohho region, east of Manzini region and east-central Shiselweni region had very low resource scarcity values because of increased access and smaller distances to the nearest health facility. Such facilities and their surrounding low risk areas predominantly are typically accessed via paved roads. Areas west of the Manzini region; central, east and south of the Lubombo region; and large areas of the Shiselweni region have high resource scarcity.

Snakebite cluster risk mapping.

Cluster analysis (Fig 5A) revealed four clusters highlighting varying degrees of snakebite hazard, exposure, and susceptibility. The risk maps produced in this analysis show areas that have low (0) to high similarity (1) with areas with the worst-case scenario (highest number of reported snakebites). For instance, areas with high snakebite susceptibility are very similar in relation to the factors influencing snakebite, i.e. they have high levels of poverty, high proportion of households that use firewood for cooking, and high proportion of households that use candles for lighting, among other factors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5.

Snakebite risk cluster risk map (A), and snakebite outcome risk cluster map (B). Made with shape files from the Eswatini Central Statistics Office (with [permission).

https://doi.org/10.1371/journal.pntd.0011732.g005

Cluster 1 (indicated in yellow, Fig 5A) is characterized by low exposure but high snakebite hazard and moderate susceptibility, resulting in a moderate risk of snakebite. EAs in this cluster is dominated by peri urban areas and other high population density areas. Cluster 2 (indicated in orange, Fig 5A) is comprised of EAs that have moderate snakebite exposure and susceptibility but very low snakebite hazard, resulting in a moderate to high risk of being bitten. Cluster 3 (indicated in green, Fig 5A) is comprised of EAs that have very high snakebite exposure and high hazard, but very low susceptibility, resulting in a very low risk of being bitten. Using high-resolution satellite imagery, these EAs were mostly associated with densely populated urban areas. Building density, building occupation density, and the presence of multiple land uses are the exposure variables that push the exposure risk score up. Snake sightings in these areas, as well as snakebite localities recorded as the nearest town would have also increased the reported hazard risk for this cluster. However, the lower levels of poverty, and better access to solid cement or brick buildings, indoor lavatories, high percentage of inhabitants using electricity for lighting and cooking, resulted in very low susceptibility of snakebite, and consequently the low risk of snakebite.

Cluster 4 (indicated in red, Fig 5A) contains EAs that have low snakebite exposure and hazard but very high susceptibility. Enumeration areas in this cluster were at the greatest risk of snakebite. Using high-resolution satellite imagery, these areas were typically associated with rural areas with poor socio-economic conditions and high rates of poverty.

Snakebite outcome cluster risk mapping.

In total, four (4) clusters emerged as before, with differences in access to healthcare. Availability and fast access to healthcare was moderately scarce in cluster 1, high in clusters 2 and 4, and plentiful in cluster 3.

Densely populated peri urban areas were identified to be where a person is most at risk of being bitten by a venomous snake and have a poor outcome (Cluster 4, indicated in red, Fig 5B). Approximately 50% of EAs in the southern Shiselweni region were identified to be at the highest risk of being bitten by a snake and encountering a poor outcome due to scarcity of fast access to health facilities with the capacity to treat snakebite, and high susceptibility to snakebite. High rates of poverty, outdoor pit latrines, and a high percentage of inhabitants using firewood or paraffin for cooking and lighting accounts for the high susceptibility score in this cluster. High-resolution satellite imagery shows EAs in clusters 2 and 4 in this region to be accessible by unpaved road resulting in delayed access to health facilities where life and limb saving treatment can be accessed. EAs in Lubombo region, in the east of Eswatini, were primarily in cluster 2. Whilst access to health resources is challenging in this region, the large areas of commercial agriculture such as sugarcane plantations, as well as protected and conservation areas, lower venomous species richness. Low population density drove down the cluster risk for snakebite outcomes in this region. Both Manzini and Hhohho regions had a diverse mix of clusters with about 25% of EAs in each region identified to be at low and moderate risk of poor snakebite outcomes (clusters 1 and 3, Fig 5B). These EAs were largely the same as those identified in the snakebite cluster risk map (Fig 5A). High resolution imagery confirmed a highway linking EAs in clusters 1 and 3 that traverse the regional boundary, indicating good access to health facilities associated with urban areas. Health facilities not located in urban areas, such as those in the southeast and southwest of Manzini region, and northeast of Manzini region bordering Hhohho region, whilst accessible by tar road, have populations typically living in surrounding area that are only accessible by unpaved road, thereby delaying access to snakebite treatment and increasing risk of poor snakebite outcomes.

Discussion

Understanding snakebite burden, epidemiology and risks enables governments and health agencies to make informed decisions when allocating resources as well as designing and implementing treatment and prevention measures to achieve the global goal of reducing morbidity and mortality from snakebite by 50% [2]. Despite the reported widespread incidence of snakebites in Eswatini, this is the first analysis through modelling which also identifies geographical locations that are at various levels of snakebite risk and at risk of poor outcomes from snakebite.

The seasonal distribution of snakebites followed weather patterns in Eswatini, with October to March being the ‘wet’ season characterised by rainfall and warmer temperatures, similar to other countries [6,7,22,26,81–84]. As shown in other studies [82], patients were mostly bitten at low geographical elevations where, relative to the rest of the country, temperatures are warmer. Historical trends have shown annual average temperatures have increased by 3⁰C between 1961–2020, and climate change is expected to see mean annual temperatures rise by a further 2⁰C by 2069 [85]. Consequently, this is expected to affect the distribution of biodiversity and ecosystems in Eswatini, including species movements leading to an increase in human-wildlife interactions, including snakebite [86].

Patients aged 10–19 years, and males had a greater relative risk of being bitten than other patients, supporting snakebite data from other countries [7,9,22,26,81–84,87,88]. Targeted public health prevention interventions to these populations through schools, community structures, peer educators should be prioritised, using traditional and social media [31,89]. Most bites were to the foot and 25% of patients were bitten whilst walking, as also reported in other countries [6–9,21,24,87,88]. The peak time of day for snakebites to occur was in the evening and early night between 6-10pm, when most people would be fetching firewood and/or water, preparing evening meals, or securing the homestead before sleep, reflecting findings from other countries [8,22,83,87]. Rural health clinics are normally closed during this time, further challenging timely snakebite treatment in high risk rural and peri urban areas poorly accessible by unpaved roads. Until November 2022, snakebite data was not collected by the national health programme from rural clinics because these clinics do not have the human resources nor the capacity to treat snakebite. Health centres and hospitals typically function with a smaller sized night shift team. Adapting medical team compositions to ensure snakebite expertise is available during the evening and night could be one strategy to reduce the morbidity and mortality associated with snakebite. Incorrect and/or inappropriate application of a tourniquet leads to swelling, confusion of snakebite clinical presentation, and for cytotoxic envenoming cause local sequestering of the venom and accelerated tissue necrosis [90,91]. Application of herbal remedies particularly to broken skin increases the risk of superficial and systemic infections [91]. This study found two thirds of patients had administered some sort of first aid measure, with tourniquets being the most common. Therefore, public programmes promoting more effective snakebite first aid management need to be prioritised particularly in high risk and poor outcome areas. To date, snake safety and snakebite prevention and first aid in Eswatini has been spearheaded by the Eswatini Antivenom Foundation. Interventions employed have included volunteer snake rescuers who remove snakes found in homes and workplaces, public snake-handling trainings, freely available snake identification information, and community and school outreach teaching snake safety and first aid [92]. In collaboration with the Ministry of Health and other national organisations, further interventions are planned including measuring their impact.

Snakebite envenoming is frequently a medical emergency requiring immediate critical care by a medical team. Eswatini has developed national snakebite treatment guidelines [80], and antivenom is listed as a vital medicine to be distributed to all health centres and hospitals through vertical national programmes in its essential medicines list. This enables available antivenom to be prioritised for procurement by the Ministry of Health [93]. Intravenous administration of antivenom therapy neutralizes venom toxins circulating in the patient’s system and is the cornerstone of medical management of snakebite envenoming [94]. However due to critical shortages of SAIMR polyvalent antivenom affecting the southern African region, the procurement of antivenom in Eswatini has been very challenging in recent years. This study found a quarter of all patients presenting with snakebite received SAIMR polyvalent antivenom. Further training, at all levels of the health system, in the clinical uptake of the national snakebite guidelines [80] and support for rational antivenom use is indicated as antivenom was reported to be administered inappropriately to patients presenting with no symptoms of snakebite and to one patient who presented with venom ophthalmia, whilst not given to patients that did need it. Short term adverse reactions to antivenom were reported in 35% of patients, and this percentage has been observed in other studies [22,83,84,95,96]. Subcutaneous adrenaline is used prior to administration of antivenom to mitigate severe adverse reactions [97,98]. However, this study showed 39% of those who received both adrenaline pretreatment and antivenom experienced an adverse reaction, with two patients experiencing severe reactions necessitating discontinuation of antivenom treatment. Further investigations are warranted to elucidate the severity, and the short- and long-term effects of antivenom use. The data also calls into question the rationale used by clinicians for performing fasciotomy, a procedure which is only indicated for the rare event of vascular compromise due to compartment syndrome.

Whilst it was not possible to definitively identify the snakes responsible for the bites, over half of bites presented with syndromes suggesting cytotoxic envenoming (mild swelling or painful progressive swelling), less than 5% of bites presented with progressive weakness associated with neurotoxic envenoming, and 37% presented with no symptoms suggestive of a bite from a non-venomous snake or a dry bite from a venomous snake. Antivenom usage reflects this pattern, with less than 25% of bites receiving antivenom and less than 30% presenting with clinical symptoms necessitating need for antivenom. Furthermore, the frequency of cytotoxic envenoming reflects the venomous snake species in Eswatini with six out of the eleven venomous snake species inducing swelling. These six species are Bibron’s stiletto snake (Atractaspsis bibronii), puff adder (Bitis arietans), snouted night adder (Causus defillippii), rhombic night adder (Causus rhombeatus), rinkhals (Hemachatus haemachatus), and Mozambique spitting cobra (Naja mossambica). Additionally, mildly venomous snakes that do not require medical attention such as the herald snake (Crotaphopeltis hotamboeia) and the olive grass snake (Psammophis mossambicus) that are common across Eswatini and can also cause mild swelling. Other snakebite studies from South Africa and Sri Lanka have found cytotoxic envenoming to dominate clinical presentation [22,82]. The frequency of cytotoxic envenoming could be explained by potential biases in the patient population presenting to health facilities for treatment. Given the rapid onset of neurotoxic envenoming, such patients may die en route to a health centre / hospital, or at home because relatives take the patient home rather than to the health facility, as reported elsewhere [10]. Consequently, the victim is not included in the health system records. Community surveys will facilitate understanding the extent and reasons for underreporting of snakebite, and the common species involved [10,99,100].

For a snakebite incident to occur, there needs to be a snake-human interaction [101]. Previous studies that have attempted to uncover the incidence of snakebite using various approaches have generally used snake species richness or occurrence data as a key determinant of snakebite risk, with varying degrees of accuracy [26–28]. In contrast, our data, shows that in the small geographical area of Eswatini, species richness alone does not explain snakebite incidence. Snake species richness only indicates the number of snake species that occur or potentially occur in a given area. However, if such species are in low abundance, then the risk of being bitten diminishes. Unfortunately, abundance data for snakes are not available in Eswatini, as is the case for most of the world, which makes it impossible to adequately characterize the true probability of encountering a snake in Eswatini. Nevertheless, the findings in this study show that the exposure, vulnerability and susceptibility to snakebites varies geographically. As it is not yet possible to differentiate between dry bites from venomous species and non-venomous snakebites, the risk modelling of snakebite used distribution of venomous snakes and all snakebite data.

While most of our findings on the geography of bites are in agreement with the current literature [20,26,28,82], subtle patterns and/or clusters emerged. The cluster analysis revealed areas with homogenous levels of exposure, susceptibility, and health care resource scarcity. The cluster of areas with the highest rates of snakebite risk were seen in the rural and agricultural outskirts of the country. As similarly observed in other parts of the world, areas with the worst snakebite outcome are those with the underserved populations in terms of access to healthcare and infrastructure [19,20]. This points to the need for ensuring equitable distribution of healthcare resources to address the problem of snakebite, particularly in the most susceptible, and exposed populations in the rural areas of Eswatini. Other socio-economic factors also need to be considered to better understand these variations. Poverty and its subsequent consequences are one of the main drivers of these sub-national variations in susceptibility and healthcare resource availability [19]. For example, the lack of sanitation facilities, use of firewood for cooking and heating as well as the type of dwellings that people live in, all of which are factors that are a function of income, determine the risk of snakebite. This is in addition to what has been observed in previous studies which show that the rural population is at high risk of snakebite due to high exposure levels associated with their lifestyle and occupation [8,18,25]. It is, therefore, recommended that interventions to reduce snakebite in Eswatini focus on those geographic areas or clusters that have been shown to have high susceptibility and high healthcare resource scarcity.

We describe the snakebite data collected from the sixteen hospitals and health centres across Eswatini serving as referral snakebite centres. Data from rural and community clinics is excluded as these clinics refer patients with a systemic envenoming to one of the sixteen included hospitals and health centres and these referrals cannot be identified based on the programmatically collected and anonymized data. The number of snakebites without signs of envenoming in our study is therefore an underestimate. We preferred this over distorted numbers of patients with systemic envenoming being counted twice after being referred from these rural and community clinics. In November 2022, snakebite became a notifiable condition in Eswatini, requiring every type of health facility, irrespective of ability to treat snakebite, to report snakebite cases. This move will provide Eswatini with a more accurate picture of the national snakebite burden presenting to the health system in the future. Additionally, community surveys should be considered to better elucidate the snakebite burden in Eswatini, as has been studied in other countries [6,10].

The 2020–2021 snakebite season overlapped with the COVID-19 pandemic. Increased demand for hospital beds and concerns of acquiring COVID-19 may have led to fewer presentations of snakebite to hospitals and health centres]. Finally, to comprehensively assess the snakebite burden in Eswatini, future studies should also include collecting data from animals affected by snakebites.

Conclusions

Snakebite is a challenging medical and societal issue, affecting all areas and populations in Eswatini. Two cluster risk maps were developed based on risk of snakebite and risk of poor outcomes from a snakebite. Whilst Eswatini is geographically relatively well covered with rural clinics, only hospitals and health centres have the equipment and staffing to safely treat a venomous snakebite with antivenom. Consequently, large areas of the country are at the highest risk of snakebite with poor outcomes. If Eswatini is to meet the global goal of reducing snakebite morbidity and mortality by 50% by 2030, especially in the backdrop of severe shortages of antivenom in the region, the high-risk areas identified in these risk maps are recommended priority areas for resource allocation and targeted snakebite prevention programmes. These risk maps also highlight the urgent need for better, simpler snakebite prevention and treatment tools that can be employed by less skilled health care workers stationed at lower-level community rural clinics.

Supporting information

Acknowledgments

We are grateful to the following healthcare workers from the 16 health facilities for reporting snakebite cases and the Eswatini Ministry of Health for supporting this study: Lindokuhle Dladla, Augustine Dlamini, Bongekile Dlamini, Dr Bongie Dlamini, Doris Dlamini, Eunice Dlamini, Lindiwe Dlamini, Menzi Dlamini, Mzweleni Dlamini, Dr Nomonde Dlamini, Quinton Dlamini, Dr S Dlamini, Sr Sabelo Dlamini, Sabusiswa Dlamini, Thendeka Dlamini, Victoria Dlamini, Zakhele Dlamini, Lydia Gwebu, Mandla Kunene, Dr Lubi, Bongiwe Magagula, Dr Makengo, Mbongiseni Mathobela, Ayanda Maziya, Sr Mhlanga, Dr Zabenguni Mkhathswa, Phumzile Mkhonta, Dr Muchada, Sr Ncobile, Mancoba Ndzinisa, Dr Mande Nelson, Dr Tandzile Nhlenetfwa, Valvini Phakathi, Dr Samiceliso Shabngu, Mgcineni Sibandze, Nomalungelo Simelane, Phindile Tembe, Echo Van der Wal, Kate Zwane and Nqobile Zwane. The authors would like to thank Dr Nicholas Riches and Dr Anna Trelfa who were instrumental in converging the snakebite research interests of the Eswatini Antivenom Foundation and the Liverpool of School of Tropical Medicine.

We are also grateful to the Eswatini Antivenom Foundation community snake rescue volunteers and the public for reporting snake sightings.