Anthrax Outbreaks among Domestic Ruminants Associated with Butchering Infected Livestock and Improper Carcass Disposal in Three Districts of Uganda, 2016-2018

Background During January 2017-December 2018, multiple human anthrax outbreaks in Arua, Kween, and Kiruhura districts, Uganda were caused by exposure to domestic ruminants that died of anthrax. We investigated to determine the scope of anthrax outbreaks in domestic ruminants, identify possible exposures, and recommend evidence-based control measures. Methods We defined a suspected case-animal as sudden death of a domestic ruminant with unclotted blood oozing from body orifices during January 2016–December 2018 in Arua, Kween and Kiruhura districts. A probable case-animal was a suspected case-animal with a positive rapid diagnostic test using the Active Anthrax Detect Rapid Diagnostic Test, and/or the microscopic identification of Gram-positive rods in biological samples. A case-kraal was defined as an enclosure of cattle/sheep or mixed species with ≥1 suspected case-animal. We reviewed district veterinary records and actively searched for case-livestock. We conducted separate case-control studies in the affected districts to compare exposures between case-kraals and control-kraals (i.e., kraals with no suspected case-livestock during the same time period as the case-kraal), frequency-matched by village, with ratios of 1:1 in Arua, 1:4 in Kiruhura, and 1:2 in Kween. We estimated overall associations in all three districts using pooled analysis. Results We identified 1,971 suspected case-livestock (attack rate [AR] =1.4/1000) in 229 (7.3/1,000) of 31,500 kraals. Cattle (AR=2.1/1,000), goats (AR=0.48/1,000), and sheep (AR=0.10/1,000) were all affected. Of the three districts, Arua was the most affected (AR=3.1/1,000), followed by Kween (AR=1.8/1,000) and Kiruhura (AR=0.065/1,000). The epidemic curve indicated continuous outbreaks in Arua and Kween districts. Human outbreaks were reported during or after the onset of livestock outbreaks in all three (OR=8.0; 95%CI=5.2-12) and (OR=1.7, 95%CI=1.1-2.4) before the outbreak were significant risk factors for being a case-kraal. Conclusions Ugandan districts with human anthrax outbreaks had concurrent livestock anthrax outbreaks associated with nearby butchering and improper carcass disposal of livestock with suspected anthrax. We recommended anthrax vaccination for domestic ruminants, proper carcass disposal, increased surveillance for sudden livestock deaths, increased capacity for laboratory confirmation, and sensitization to livestock-keepers about anthrax control. AR: Attack Rate; CDC: US Centers for Disease Control and Prevention; AAD-RDT: Active Anthrax Detect Rapid Diagnostic Test; CI: Confidence Interval; DVO: District Veterinary Officer; MAAIF: Ministry of Agriculture, Animal Industry and Fisheries; MoH: Ministry of Health; NADDEC: National Animal Disease Diagnostic Epidemiology Centre; OR: Odds Ratio; PEPFAR: President’s Emergency Plan for AIDS Relief.


Introduction
Anthrax is a zoonotic disease caused by the Gram-positive spore-forming bacteria Bacillus anthracis (1). It is primarily a disease of domestic and wild herbivores, and presents as a peracute or acute condition, usually resulting in death (2). Livestock become infected when they either inhale or ingest B. anthracis spores from contaminated soil, contaminated water, or from carcasses of livestock that have died of anthrax (3). Upon entering the host, the spores germinate, multiply, and cause disease. The most susceptible species are cattle, sheep, and goats, but anthrax can also infect other mammals such as pigs, horses, dogs, cats, and primates (4,5). The incubation period of anthrax in susceptible livestock is 1-14 days after oral exposure (3). Infected livestock may have high fevers, blood-stained urine, bleeding from orifices, or no signs or symptoms at all (3). At death, livestock may demonstrate oozing blood from body orifices and rapid bloating of the carcass (6). It is postulated that the massive releases of capsulated bacilli from the lymphoid tissues, especially the spleen, leads to terminal septicemia and death within a day or two (7). B. anthracis occurs in a vegetative form within the host, and a spore form in the environment (8). The spore form is extremely resistant to adverse conditions, and can survive for years in soil, or in the wool and hair of infected livestock. After an animal dies of anthrax, vegetative B. anthracis cells within the carcass sporulate and further contaminate the surrounding land, subsequently posing a risk to grazing livestock, such as cattle, goats, and sheep (9,10). B. anthracis spores can remain viable on pastureland for decades. As a result, livestock outbreaks can be protracted, as exposures continue over many months or even years (9,10). Human infections are most often associated with exposure to infected livestock or their products, such as meat, hides, bones, and other materials (11).
Anthrax has been documented in Uganda since at least 1918 (12), and livestock outbreaks have been reportable since 1959 (9). In recent years, outbreaks have occurred sporadically -including livestock and wildlife-and humans living in and around Queen Elizabeth National Park (13)(14)(15). Control of anthrax in livestock and, as a result, in humans, requires strong surveillance, annual vaccination of livestock, and rapid outbreak response involving immediate safe carcass disposal, ring vaccination of cattle, antibiotic treatment of sick livestock, and awareness campaigns (9,(16)(17)(18). However, in Uganda, no policy currently exists regarding annual routine vaccination against anthrax among livestock, although some farmers privately buy prophylactic antibiotics and vaccines for their livestock during suspected anthrax outbreaks.
During January 2017 to December 2018, human anthrax outbreaks were reported in three geographically distant districts of Uganda. Investigations of these human anthrax outbreaks revealed that they were caused by eating and handling meat from livestock that died of suspected anthrax (unpublished data). Although these findings suggested that outbreaks had also occurred recently or were occurring among domestic ruminants, no livestock cases were reported during this time. We investigated to identify corresponding The districts of Kween and Kiruhura are located within the "cattle corridor region" of Uganda (23), which is a broad zone stretching from southwestern to northeastern Uganda, dominated by pastoral rangelands. In Kiruhura and Kween Districts, approximately 60% of the population is engaged in livestock-rearing, including imported and indigenous cattle, goats, poultry, and pigs, and approximately 30% are involved in growing crops. This corridor has high rainfall variability, including periodic late seasonal onset of rains leading to drought (24). In Arua District, most of the population -approximately 90% -work in the crop sector, primarily growing maize, coffee, and beans, while fewer than 1% work in livestock-rearing.

Case definition and case-finding
We defined a suspected livestock case as sudden death of a domestic ruminant with unclotted blood emerging from body orifices from January 2016 to December 2018 in the districts of Arua, Kiruhura, and Kween. A probable case was a suspected case that tested positive for anthrax by the Active Anthrax Detect Rapid Diagnostic Test (AAD-RDT, described later) or identification of Gram-positive rods via microscopy of livestock tissue samples. To identify cases, we reviewed district veterinary anthrax records and conducted active community case-finding in the affected villages. We recorded the cases in a line list, which included signs and symptoms, livestock location, identification information, species, breed, date and site of death, method of carcass disposal, and vaccination status (ever vaccinated, and vaccinated during the past year).

Descriptive epidemiology
We described the characteristics of domestic ruminants (livestock) affected by anthrax by calculating the attack rates (AR) by species, sex, and breed. Denominator data for livestock populations were provided by the district veterinary offices. We evaluated the time of death using district-specific epidemic curves, and place of death using maps.

Hypothesis generation interviews
We conducted individual hypothesis-generating interviews with 229 leaders or owners of case-kraals (140 case-kraals in Arua, 22 case-kraals in Kiruhura, and 67 case-kraals in Kween). We collected information on potential exposures for livestock anthrax, including method and location of grazing (specifically, grazing near the riverbank during a drought), purchase of livestock from areas known to have anthrax outbreaks at the time of purchase, presence of a wildlife migration route on or near the pastureland, method of livestock carcass disposal, and vaccination status of livestock before an anthrax outbreak.
Proper carcass disposal was defined as burial of a suspected anthrax-infected carcass at least six feet deep followed by disinfection of the land, while improper carcass disposal was defined as failure to bury a suspected anthrax-infected carcass at least six feet deep and/or failure to disinfect the site and/or leaving the carcass on the pastureland to rot or throwing the carcass in the river.

Case-control studies
We conducted three separate case-control studies in Arua, Kiruhura, and Kween districts.
The units of study were case-and control-kraals. In all three districts, a case-kraal was defined as one with at least one case-animal during January 2016-December 2018. A control-kraal was matched to a case-kraal by village and had not had any livestock that died suddenly with unclotted blood coming from body orifices during the same time period. For all three case-control studies, the same questionnaire tool was used, and questions were asked in the same way. We interviewed the heads or owners of the case-and control-kraals on the same exposures as used during hypothesis generation. We asked about these exposures with specific reference to the time period before the human outbreaks occurred in their respective districts.

Environmental and laboratory investigations
We assessed the environment to identify carcass disposal sites or remains, migration routes near the pastureland, and presence of natural features such as a rivers, swamps, or game reserve/game parks.
At the time of our investigation, livestock cases were still occurring. We were able to test samples from livestock carcasses from Arua, Kween and Kiruhura districts. We obtained nasal swabs, ear clips, and dried skin for testing using Gram stain and AAD-RDT (a lateral flow immunoassay) (25). These tests were used to identify probable livestock cases.

Data analysis
We conducted descriptive analysis overall and by district, using frequency distributions, percentages, and rates. We analyzed the case-control data for each district independently using Epi-info version 7.2.2.6 to identify the exposures of interest. Finally, we conducted a pooled analysis to estimate the summary odds ratios (OR) and their associated 95% confidence intervals (CI) associated with the exposures of interest for all three districts, using Stata version 13.1.
Of the three districts, Arua had the highest AR (3.1/1000), followed by Kween (1.8/1000), and Kiruhura (0.065/1000) (Figure 1a). Although kraals are located throughout Arua district (not shown) all the case-kraals in Arua were located along the Nile River ( Figure   1b). In Kiruhura District, two clusters of case-kraals were identified, one in Engari Sub-County and the other in Nyakashashara Sub-county; a few other sub-counties were also sporadically affected (Figure 1c). In Kween District, case-kraals were located in one subcounty close to the Pian Upe Game Reserve (Figure 1d). Hypothesis-generating interviews with leaders of case-kraals from each district suggested that different districts might have different risk factors (Table 2). Leaders also mentioned that they sometimes provided antibiotics (in particular, oxytetracylines) to their livestock when they heard about outbreaks in neighboring kraals, with the hope of protecting their own kraal.

Findings of case-control studies
The district-specific case-control studies showed that, in Arua District, butchering the carcasses of livestock suspected to have died of anthrax on/near the pastureland was associated with 7.5 times higher odds of illness. In Kiruhura District, case-kraals had a 76 times higher odds of having suspected case-livestock butchered on/near the pastureland than control kraals. In Kween District, having suspected case-livestock butchered on/near the pastureland was associated with a 3.8-fold higher odds of being a case-kraal (Table 3).
When we asked the kraal leaders about livestock anthrax vaccination, they reported that none of the livestock in either the case-or control-kraals had ever been vaccinated against anthrax.
In the pooled analysis, having suspected anthrax-infected dead livestock butchered on or near the pastureland before the outbreak was associated with 8-fold higher odds of being a case-kraal (Table 4).

Environmental and laboratory investigation findings
In all three districts affected by the outbreak, we observed carcasses of livestock reported In our study, cattle were more affected than goats and sheep. This is consistent with studies in both Kenya and China that identified cattle as being more affected by anthrax than goats and sheep (26,28). Cattle ingest a large amount of soil from the ground when grazing, while goats typically browse on grass only; this may lead to higher levels of exposure among cattle, compared with goats. This characteristic is thought to be a major factor associated with different levels of infection among different species (9).
Interestingly, the attack rates among male livestock were more than twice as high as among female livestock, and more than six times higher among domestic than imported livestock. The reasons for this are unknown, but may relate to the ways in which livestock ingest the spores during grazing or browsing, or variation in lethal doses among livestock (9). livestock died of suspected anthrax in Arua. The reason for the lack of reported human cases during this time is unclear. It is possible that human cases -and, possibly, deathssimply went unrecognized or unreported. It is also possible that at least some of the livestock deaths were due to other causes, or were due to anthrax but were not followed by human anthrax cases. Improved laboratory capacity to diagnose anthrax would facilitate improved understanding of the dynamics between livestock and human anthrax in Uganda.
Vaccination is generally considered a cost-effective prevention method for anthrax among livestock, and, as a result, among humans (4,16,17,30,31). Although farmers reported using antibiotic prophylaxis during outbreaks (with a protective outcome for their livestock), none of the case-kraals were reportedly vaccinated against anthrax. This finding suggests a possible opportunity to promote routine annual vaccination in Uganda.

Studies evaluating the cost-effectiveness of vaccination may help guide policy decisions
In 2017, anthrax was categorized as the most important of seven priority zoonoses in Uganda (33), and in June 2019, a national symposium was held to enhance One Health efforts to combat anthrax in Uganda. The symposium included: i) sharing of up-to-date spatial and temporal data on anthrax outbreaks in both livestock and humans in Uganda; ii) validating the national anthrax control strategy; iii) forming a National Anthrax Uganda should focus on these issues as the country improves its approach to combating anthrax.

Limitations
Our investigation had several limitations. Given that anthrax-infected livestock die suddenly, it is possible that livestock that may have died of anthrax in more rural areas went unnoticed. Additionally, some farmers may have failed to notice the absence of some members of their herd due to their large herd size. Failure to account for all anthraxrelated livestock deaths in the area may have resulted in an underestimation of the scope of the outbreak. In addition, only a small subset of reported livestock deaths were tested, due to lack of available specimens and testing materials. Therefore, some of the suspected anthrax deaths might have been deaths due to other etiologies, and some anthrax deaths may have been unidentified.

Conclusions And Recommendations
In conclusion, there was evidence of livestock anthrax outbreaks in Arua, Kiruhura, and guidelines and are free to publish the work. We obtained verbal informed consent from leaders or owners of case-and control-kraals before interviews began. Participants were told that their participation in the study was voluntary and refusal to participate would not result in any negative consequences. During data analysis, we used unique identifiers to ensure confidentiality of the information provided by livestock owners.

Consent for Publication
Not applicable content and approved it for publication. All the authors reviewed the manuscript to ensure scientific integrity and intellectual content.      Stratified epidemiological curves of livestock anthrax outbreaks per Subcounty in Kiruhura District, Uganda: 2016