Cutaneous anthrax cases in this region of Uganda during 2015–2017 appear to have been caused by skinning, butchering, and carrying carcasses of infected animals that died suddenly. The 68 patients represent the largest collection of cutaneous anthrax cases ever reported in Uganda. Although a large number of suspect cases were unconfirmed, the laboratory confirmation of anthrax from a dead bull and a human in the same community as the other suspect case-patients strongly suggests that suspect cases were indeed anthrax. Clinical characteristics of case-patients were also characteristic of cutaneous anthrax.
Handling carcasses of animals with suspected and confirmed anthrax has previously been established as a risk factor for cutaneous anthrax  . The location of the black eschars for most case-patients – on the hands and back - has also been observed in other studies demonstrating livestock-human transmission of anthrax , and is likely related to animal carcasses being carried over the shoulder. The community practice of smoking meat before consumption may have prevented cases of gastrointestinal anthrax. Whether or not this was protective is difficult to confirm; vegetative forms of B. anthracis are easily killed during normal cooking, but spores are much more resistant to adverse conditions and require heating to at least 100° C for 15 minutes for inactivation . Although we were unable to identify any gastrointestinal anthrax cases, illnesses might have been missed at health facilities or otherwise unreported.
Adult males were most affected in this cluster of infections. This is unsurprising, as this group is most likely to be engaged in exposures to animal carcasses. However, two case-patients were under five years of age. For these case-patients, the only two not involved in direct contact with livestock carcasses, infection may have occurred when spores were brought into the house by a parent or other contact involved in skinning, carrying, butchering, or preparing the meat. This type of exposure has been described previously with pediatric anthrax cases , and may explain the illnesses among the youngest case-patients.
Although suspected cases had been occurring throughout the study period, the Ministry of Health was only alerted to these cases in mid-2017. Anthrax testing is possible in Uganda, but it is infrequent because of incomplete reporting in veterinary surveillance systems, knowledge gaps among healthcare workers about anthrax, and limited access to laboratory services. Because of the infrequent testing, both human and veterinary anthrax are almost certainly underreported in Uganda. Investigation of suspected cases over a longer period (i.e., before 2015) or a prospective study could shed some light on the degree of underreporting.
The epidemic curve showed noticeable peaks in human cases for two years during March-May, which corresponds with the dry season in Uganda. During the dry season, livestock may be at increased risk of contracting anthrax from soil due to increased exposure to spores in soil/dust through close grazing and looser soil . In addition, stressors, such as poor nutrition or overheating, lower immunity, and reduced resource availability, may lead to increased animal density in localized areas such as along river banks and increase possible opportunities for exposures (15). For humans, food scarcity during the dry season may also force them to seek alternate sources of food, which could lead to their eating meat that they might otherwise avoid.
Cutaneous anthrax cases have been reported at the animal-human interface in several other regions globally including Bhutan, the Republic of Georgia, Kenya, and Tanzania    . In some cases, the outbreaks have been related to chronic food insecurity . In a household survey in Zambia conducted after an anthrax outbreak linked to eating infected hippo meat, 23% of respondents mentioned that they would eat meat from dead animals again, despite the risk, due to food shortages . Although we did not systematically collect data on this, several respondents reported to us anecdotally that they could not afford butchery-sourced meat.
Cutaneous anthrax is usually curable with prompt antibiotic therapy; approximately 10–20% of untreated patients die . Although this therapy is readily available in Uganda, anecdotal evidence suggests that not all physicians are familiar with the presentation, and misdiagnosis can lead to delays in appropriate treatment. The late care-seeking behavior and subsequent difficulty with diagnosis of Mr. A likely contributed to his eventual death. Although most patients in this outbreak reported that they sought and received treatment, apart from Mr. A, we were unable to assess whether or not they were diagnosed and treated correctly.
The number of human cases and the time period over which they occurred also suggest more widespread livestock anthrax in this region, something that could potentially be prevented by animal vaccination. Effective herd immunity against anthrax in cattle requires at least 80% of the animals in an area to be vaccinated . During an outbreak, all livestock in areas around the affected area should be vaccinated . Despite anthrax being considered a priority zoonotic disease in Uganda, mass animal vaccinations are not currently done, primarily due to the low number of reported cases, nor are reactive vaccinations done during an outbreak. While farmers can choose to vaccinate their animals privately, many do not. No national guidelines are currently in place requiring livestock vaccination. The development of such guidelines would be dependent upon knowledge about its cost-effectiveness, which would depend on the true incidence of the disease. In Zimbabwe, the government previously conducted national annual anthrax vaccination campaigns for livestock, but currently vaccinations are restricted to high-risk areas only. In spite of this, livestock producers in Zimbabwe are advised to ensure that farm animals are vaccinated against anthrax annually .
Because we conducted the investigation retrospectively, there is a possibility of recall bias including symptoms and date of onset, hence masking the actual anthrax prevalence in the district. In addition, it is likely that some cases were missed, which would have led to an underestimation of the magnitude of the outbreak. It is also possible that some were not true anthrax, leading to an overestimation. We had sparse data about the livestock deaths, preventing certainty that they were due to anthrax. In addition, we could not confirm the vast majority of cases, nor did we identify any cases of gastrointestinal anthrax. Although some patients had nonspecific gastrointestinal illnesses at the time of investigation, many were already taking antibiotic therapy, rendering sample testing ineffective. It is possible that gastrointestinal anthrax occurred and was missed.
Conclusions and recommendations
The cutaneous anthrax cases occurring over a 3-year period in Uganda were associated with handling carcasses of animals that died suddenly. The district set up a One Health structure (a collaboration between human and veterinary health departments, including a single laboratory that processes both human and veterinary samples) to facilitate epidemic preparedness and response as a result of this outbreak. We recommended education of the public and of healthcare workers about signs and symptoms of anthrax, prompt treatment of illness following exposure to animals that died suddenly, safe disposal of animals that died suddenly, consumption of meat only from slaughtered healthy animals inspected by a trained veterinarian, and investigations to establish the magnitude and risk factors for anthrax transmission in animals. Vaccination of livestock in the district may be warranted; however, a burden study would inform cost-effectiveness.