A total of 446 case-patients with M. ulcerans disease were managed at Barwon Health during the study period. Twenty (4.5%) patients did not have WHO disease category recorded and were excluded. Thus, we included 426 case-patients in the study (Figure 1). We determined the number of cases included per calendar year stratified by disease severity (Figure 2).

The findings of our study suggest a serious change in the epidemiology of M. ulcerans disease in southeastern Australia. There appears to be an increase in the proportion of severe cases in recent years, with a near doubling compared with earlier time periods to 25% of all case-patients who came to our health service during 2011–2017. More severe disease has major implications: case-patients with severe disease frequently require surgical treatment with tissue reconstruction (2) and hospitalization, and often have long-term deformities (4). Severe disease also causes major increases in cost of treatment (3). This finding is in contrast to nonsevere disease, which can usually be managed with oral antimicrobial drugs; outpatients have shown good outcomes (12–14). We also report a large proportion of severe cases that affected 1 of 5 patients in our cohort. However, these rates are lower than those reported for cohorts in Africa, in which the proportion of cases with severe disease (WHO categories 2 and 3 combined) was >70% (15–17). It has been postulated that strains in Africa are more virulent because of production of increased quantities and more potent forms of mycolactone (18).

The reason for the increasing proportion of severe cases in recent years is not clear but might be related to evolution of a more pathogenic strain of M. ulcerans in the region. In the area of M. ulcerans research, utility of whole-genome sequencing has been studied from an ecologic and epidemiologic perspective. Studies indicate a strong relationship between the genotype of an isolate and the geographic origin of disease at a national and regional level (19–21). However, there is no information available on the relationship between the genotype of an isolate and the clinical severity of disease in affected patients. In the related field of M. tuberculosis research, there has been increasing interest in using genomic information of individual disease isolates to predict clinical severity of disease in humans (phenotype), especially for M. tuberculosis strains with increased drug resistance (22). We advocate that genomic studies be conducted to explore whether particular M. ulcerans strains are associated with more severe disease, and if strains in disease-endemic regions are evolving over time to become more pathogenic. These studies might provide useful information for public health policy.

Other possible explanations for the increased proportion of severe cases of M. ulcerans disease could include environmental or climatic changes, which lead to a higher inoculum of organisms. In addition, there might have been a change in population dynamics or characteristics that make humans more susceptible to severe disease (23). Although human populations in disease-endemic regions are steadily increasing, these increases are not sufficient to explain the rapid increase in reported cases, and additional studies of these factors are needed.

Our analysis suggested that there was no major difference in the proportion of severe cases of M. ulcerans disease reported between the 2 peninsulas. This finding is consistent with our understanding of restricted genetic diversity of M. ulcerans from the same geographic location (20,24). A recent study of isolates from 11 disease-endemic countries in Africa identified only 2 specific M. ulcerans lineages; these lineages were subdivided into 4 major clusters, and most of these isolates were in cluster 1 (25). The spread of disease across both peninsulas in our study region, given their close geographic proximity, probably resulted from clonal expansion of the same genotype. This hypothesis is consistent with findings of a study of isolates from Australia in which 1 isolate from Frankston (Mornington Peninsula) had genomic sequences similar to those for isolates from Point Lonsdale and St. Leonards (Bellarine Peninsula) (20).

Apart from time period, we have identified 2 risk factors for severe M. ulcerans disease: a lesion situated over a joint and age. The reason for more severe disease occurring over joints is not clear. Reduced skin and subcutaneous tissue and the absence of muscle compared with adjacent regions might lead to lower tissue temperatures and increased growth of organisms (26,27). In addition, these factors might facilitate spread across tissue planes, and a reduced blood flow in the absence of muscle might relatively reduce local immune function. In contrast, increased tissue movement and physical stress over joints might increase lymph flow in the region and enhance spread of the infection, such as that hypothesized for hand, foot and mouth disease (28,29).

Our finding of older (>65 years) age as a risk factor for severe M. ulcerans disease is consistent with our previous studies that demonstrated that these patients were more likely to have multiple lesions or be categorized into WHO category 3 (30). This finding might be related to reduced immunity in patients in older age groups. Reduced immunity inhibits control of the mycobacterium and results in larger, more severe forms of M. ulcerans disease, which is similar to immunosuppressive effects of HIV (31). Reduced immunity in elderly persons might also be related to an increased prevalence of other concurrent conditions. Likewise, our study findings suggest that children might be at increased risk for severe disease, which might be related to relative immune suppression in an immune system that is still developing.

We found that, similar to findings from Africa (32), more prolonged signs and symptoms before diagnosis were not associated with severe disease. Therefore, severe disease is not directly related to a delay in diagnosis, as is often believed.

With the increasing number and severity of cases of M. ulcerans disease in Victoria, public health measures to prevent the disease are needed, although these measured are hampered because the environmental reservoir and mode of transmission of M. ulcerans are unknown. Current recommendations involve mosquito and insect bite preventive strategies, given some evidence that mosquitoes or another vector might be responsible for transmission (33,34). There is evidence of reduced risk for disease for persons outdoors in disease-endemic areas who reported regular use of insect repellent (35). In addition, sleeping under bed nets in Africa has been associated with a reduced incidence of disease (36). Possums have been postulated as a potential environmental reservoir in Australia (8,9), and more research is needed on the role of possums in the epidemiology of human M. ulcerans disease.

There were some limitations to this study. First, because we conducted an observational study, other unmeasured confounders associated with severe disease could affect the results. Second, there might be referral bias for case-patients managed at Barwon Health and increased numbers of severe cases referred over time because of the increasing expertise of clinicians in managing M. ulcerans disease. However, Barwon Health is the only tertiary referral center on the Bellarine Peninsula, and case-patients who are residents of this peninsula are seen by this health service regardless of disease severity. Therefore we believe that referral bias accounting for increasing disease severity over time is unlikely.

The epidemiology and pathogenicity of M. ulcerans disease in southeastern Australia is changing rapidly; we observed increases in numbers of cases per year and disease severity. Risk factors for severe disease include age, time period, and lesions located over a joint. Reasons for the changing epidemiology and pathogenicity are unknown but these factors urgently need to be identified and addressed to halt spread of this increasingly devastating disease.