key: cord-0258044-t9fpzbas
authors: Blasdell, K. R.; McNamara, B.; O'Brien, D. P.; Tachedjian, M.; Boyd, V.; Dunn, M.; Mee, P. T.; Clayton, S.; Gaburro, J.; Smith, I.; Gibney, K. B.; Tay, E. L.; Hobbs, E. C.; Waidyatillake, N.; Lynch, S. E.; Stinear, T. P.; Athan, E.
title: Environmental risk factors associated with the presence of Mycobacterium ulcerans in Victoria, Australia
date: 2022-01-11
journal: nan
DOI: 10.1101/2022.01.10.22269030
sha: cce4ef7bedcb8df8333cc749f1864770306eced1
doc_id: 258044
cord_uid: t9fpzbas

In recent years reported cases of Buruli ulcer (BU), caused by Mycobacterium ulcerans (MU), have increased substantially in Victoria, Australia, with the epidemic also expanding geographically. To develop an understanding of how MU circulates in the environment and transmits to humans we analyzed environmental samples collected from 115 properties of recent BU cases and from 115 postcode-matched control properties, for the presence of MU. Environmental factors associated with increased odds of MU presence at a property included certain native plant species and native vegetation in general, more alkaline soil, lower altitude, the presence of common ringtail possums (Pseudocheirus peregrinus) and overhead powerlines. However, only powerlines and the absence of the native plant Melaleuca lanceolata were associated with BU case properties. Samples positive for MU were more likely to be found at case properties and were associated with detections of MU in ringtail possum feces, supporting the hypothesis that MU is zoonotic, with ringtail possums the strongest reservoir host candidate. However, the disparity in environmental risk factors associated with MU positive properties versus case properties indicates a strong human behavioral component or the influence of other environmental factors in disease acquisition that requires further study.

Buruli ulcer (BU) is a neglected tropical disease, caused by the environmental pathogen 53 Mycobacterium ulcerans (MU). Affecting all age groups, the disease causes severe 54 destructive lesions of skin and soft tissue and results in significant morbidity, sometimes 55 leading to long term disability and deformity (O'Brien et al., 2015) . Endemic to more than 30 56 countries, the highest disease burden is in sub-Saharan Africa (Johnson et al., 2005 (Victorian DH, 2021b; Tai et al., 2018) . 64 Previous studies have identified several risk factors and potential transmission routes. In 65 Africa, BU foci are often associated with natural water bodies (Aiga et al., 2004; Bratschi et 66 al., 2014) and in Victoria, an outbreak was linked to exposure to a contaminated water 67 irrigation system at a golf course (Veitch et al., 1997) . In a questionnaire-based case control 68 study in one Victorian hotspot, the risk of having BU was found to be increased in people 69 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 who did not wash minor skin wounds immediately, did not frequently wear insect repellent or 70 long trousers outdoors, and who received mosquito bites to the lower legs or arms (Quek et 71 al., 2007) . Molecular detection of MU in mosquitoes collected from several localities within 72 the Victorian endemic area Lavender et al., 2011) , and the 73 demonstration that Ae. notoscriptus can act as mechanical vectors for BU in a mouse model 74 (Wallace et al., 2017) suggests that mosquitoes may be involved with BU transmission in 75 Victoria. Several studies have also suggested that MU may be a zoonotic pathogen in 76 Victoria. Evidence of infection and disease has been reported in several native and non-native 77 mammals (Elsner et al., 2008; Mitchell et al., 1987; O'Brien et al., 2011; O'Brien et al., 2013; 78 O'Brien et al., 2014; Van Zyl et al., 2010) , but there is increasing evidence that two common 79 possum species may be acting as reservoirs hosts in south east Australia (Fyfe et al., 2010) . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 Cases who completed the study questionnaire were purposely selected by postcode to ensure 118 a representative spread of sampling across the affected area based on reported BU prevalence 119 (i.e. more properties were surveyed in postcodes with more cases). Control properties were 120 then purposely selected based on postcode and matched 1:1 to case properties. Five different sample types were collected as outlined in Table 1 , namely soil, water, plants 131 (Supplemental Figure 4) , feces and insects. Soil texture was determined as per standard 132 protocols (https://www.dpi.nsw.gov.au/__data/assets/pdf_file/0008/168866/texture-133 salinity.pdf). Up to 20 environmental samples were collected per property. In addition, a 134 mains water sample was also collected from each property as a negative control, to validate 135 sample collection techniques and detect potential contamination. The total number of 136 observable water sources on a property was recorded, although samples were not always 137 collected from all sources. The presence of mosquito larvae in any of the water sources was 138 also recorded. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint travel restrictions imposed during the COVID-19 pandemic. For all return environmental 143 field surveys at a property, the property outline from the initial visit was used to enable the 144 same sample types to be collected from the same locations. Any different samples collected 145 and significant environmental changes between the two field surveys were recorded. All 146 samples were returned to the laboratory and maintained at -70°C until processed. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022 Samples for nucleic acid extraction were thawed prior to processing and transferred to 2ml (Fyfe et al., 2007) . As this assay detects other mycolactone-producing 171 Mycobacteria in addition to MU, any sample that tested positive by this assay (CT<40, 172 threshold 0.02) was subjected to confirmatory testing using two MU specific assays (IS2606 173 and KR; Fyfe et al., 2007) as well as an RNA-based assay targeting the MU 16S rRNA to 174 assess viability (Beissner et al., 2012) . This viability assay can also detect some strains of M. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022 We investigated relationships between each of the property outcomes (IS2404 detected, 191 confirmed, viable and case status) with environmental variables including garden type, and is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint fox and RT possum feces in particular (20.0% and 16.7% respectively)) ( properties, the interval between case notification date and field collection date did not affect 237 the odds of a property testing IS2404 detected, confirmed or viable (data on request). At is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint properties when considering all samples and when restricted to fecal samples or RT possum 258 feces only (Table 3) . No significant relationships were observed for the viability assay. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint low sample size, the presence of rabbit feces was also not included in the multivariable 280 models, even though properties with rabbit feces were more likely to have viable MU at the 281 property. Two significant relationships that were observed in univariate analyses but not in 282 multivariable analyses were the positive association between Melaleuca lanceolata and 283 IS2404 detected, confirmed and viable properties, and the higher soil salinity associated with 284 IS2404 and confirmed properties. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint 

Adjusted analyses for each property type included variables with an association observed (at p<0.1) for any property type. Variables were excluded where the total number of properties with that variable was less than 30.

A total of 27 properties were visited twice. 

Positive became negative 6 (22.2) 5 (18.5) 4 (14.8)

Negative became positive 2 (7.4) 0 (0.0) 0 (0.0)

Under the assumption that properties with positive status at both visits were positive for the entire period between the two sampling visits, we documented one property that remained IS2404 detected for at least 8.7 months, which was also the longest time between visits for any property. Of properties that remained IS2404 detected, half (7/14) had an interval between sampling of over six months. For nine confirmed properties, five properties . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint remained confirmed for over six months, with the longest sampling interval of 7.9 months.

The two properties that remained positive for the viability assay were positive for over six months. Only for the IS2404 assay did any property become positive, with two properties that were initially IS2404 negative becoming positive at the second visit (7.4%; Table 5 ); all other changes of property status were from positive at the first visit, to negative at the second visit (Table 5 ).

The findings from this study support the hypothesis that BU may be a zoonotic disease in Australia, with native mammals, specifically species of possum, acting as reservoir hosts (Carson et al., 2014; O'Brien et al., 2014) . Consistent with previous findings, fecal samples were the sample type most commonly positive for MU and had the highest bacterial loads (Fyfe et al. 2010) . This was also the sample type most likely to remain positive at return properties and the only sample type that appeared to contain viable bacteria. In Australia, numerous species of both native and introduced mammals including feces from RT possum,

BT possum and rodents have tested positive for MU in the past, (Fyfe et al., 2010; Roltgen et al., 2017) ) as also found in this study. In addition, in our study fecal samples from wild fox (Vulpes vulpes) and rabbit (Oryctolagus cuniculus) tested positive; a first report for both species, although laboratory rabbits have been infected experimentally (Norden & Linell, 1951) . Fox feces collected during this study had the highest proportion of positives in IS2404 and confirmatory assays. However, RT possum feces were the sample type with the second highest proportion of positives by both these assays, the primary sample type positive by the viability assay and the sample type most commonly collected in this study (>28% of all . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint samples collected). Only feces from RT possums, BT possums and a single fox were found to be viable, suggesting that all three species may be involved in the transmission of MU.

RT possums were more likely to be found at confirmed and viable properties and their feces were more likely to be positive for MU at case properties. Previous studies have also found a correlation between the geographic location of cases and the presence of positive possum feces (Carson et al., 2014; Fyfe et al 2010) , supporting the reservoir host hypothesis. The findings from this study suggest that the presence of RT possums per se at a property does not increase the risk of the residents contracting BU, but the presence of RT possums positive for MU does. This intrinsically makes sense but requires effective communication to local residents to discourage the indiscriminate removal or translocation of possums, which are a protected species in Victoria. In the UK, removal of badgers as part of bovine tuberculosis (TB) control measures led to increased bovine TB prevalence in some regions. This was hypothesized to be because culling disrupted badger social organization, leading to longdistance movement and dispersal of individual badgers, resulting in increased TB transmission among badgers (Donnelly et al., 2003; Donnelly et al., 2006; Woodroffe et al., 2005) . Increases in Leptospira carriage in rat populations subjected to indiscriminate lethal control methods in Vancouver, Canada have also been attributed to altered social structure and subsequent increases in aggressive interactions (Lee et al., 2018) . As possums are territorial, removal or disturbance of individual resident animals impacts both social interactions and movement patterns (Matthews et al., 2004) , which may in part help explain the shifting dynamics of this disease and the expansion of the Victorian endemic area.

Movement of MU into previously unaffected areas may also be facilitated by infected foxes, which demonstrate considerably larger home ranges than RT possums: individual foxes in a similar coastal habitat in New South Wales were found to have a mean home range of 135 ha . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101 /2022 doi: medRxiv preprint (Meek & Saunders, 2000) , compared to <1ha for RT possums (Lindenmayer et al., 2008) .

However, further research into the role of foxes in BU transmission is needed.

It is thought that MU can also persist outside of a vertebrate host, although the duration and the environmental conditions required have not been well defined (Bratschi et al., 2014; Garchitorena et al., 2015) . Overall, soil was the second most commonly positive sample type, and particular soil characteristics were associated with positive properties. It is possible that the higher conductivity, salinity and alkalinity detected at these properties may enhance environmental survival of MU and/or aid transmission between hosts. The link between MU and slightly alkaline soil was unexpected as these bacteria have been associated with mildly acidic pH conditions in two aquatic communities in Cameroon (Garchitorena et al., 2015) .

Under laboratory conditions MU also appears to grow better at mildly acidic pH, although growth can also occur under mildly alkaline conditions (Portaels & Pattyn, 1982) . It is possible that in soil (in contrast to water), other biotic factors may interact with pH to make alkaline conditions more favorable to MU. However, as only a minority of IS2404 positive soil samples were confirmed positive and none were considered viable, the detection of MU in soil may represent the presence of DNA from non-viable, degrading MU. If this is the case, then these environmental conditions may favor the preservation of MU DNA rather than bacterial survival. There seemed to be little association between MU and water at the scale analyzed in this study. Very few water sources returned IS2404 positive results and only a single water source was confirmed positive (from a bucket). This is consistent with previous environmental surveys in the region that also found low rates of MU positivity in soil and water (Fyfe et al 2010) . There was also no association between property status and the number of water sources present or the presence of bore water. This suggests that in Victoria, water plays a limited role in determining the fine-scale distribution of MU. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022 Infection through puncturing injuries received from plants, biting insects and other objects contaminated with MU has also been hypothesized as a transmission pathway to humans (Street et al., 1991; Wallace et al., 2017) . Relatively few insects were screened in this study making it difficult to assess associations with cases. However, one confirmed mosquito (Aedes notoscriptus)was detected in a case property, consistent with MU mosquito positivity reported in previous field surveys in this region Lavender et al., 2011) . Bellarine Peninsulas (Yugovic, 2002) , and are utilized heavily by possum species for denning and as food sources (K. Blasdell, personal observation; Pahl, 1987) . Interestingly, Moonahs were less likely to be found at case properties, although this may be due to the tendency of local residents (particularly those personally affected by MU) to discourage possums from visiting their properties through environmental modification due to the perception that possums are carriers of these bacteria. Certainly, the gardens of some properties visited by the researchers had been re-landscaped or modified by their owners post BU diagnosis (K Blasdell, personal observation). As gardens containing native or mixed vegetation were more likely to be positive for MU than those containing mainly non-native vegetation, this suggests that native environments may promote better survival of the bacteria, potentially because they appear to support denser populations of native mammalian hosts, such as possums.

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022 To persist, possums require a suitable area of habitat containing sufficient resources. Habitat patches below a certain size (such as most urban and suburban gardens) are unlikely to provide these requirements unless they are well connected to other similar patches (Goddard et al., 2010) . For example, individual BT possums in Melbourne, Australia regularly foraged in several residential gardens despite denning in urban forest fragments (Harper et al., 2005) , whilst in New Zealand, BT possum occupancy of urban gardens decreased with increasing housing density and decreasing green cover (Adams et al., 2013) . Assuming that RT possums respond in a similar way, this may explain the association between larger properties and positive status. However, this could also be a geographic effect, as properties surveyed in the Mornington Peninsula (the current epicenter of BU in Victoria/Australia) were larger than those surveyed closer to Melbourne (Bayside area). Powerlines were more likely to be found at IS2404 detected, confirmed and case properties. As possums regularly use power lines to travel around urban areas (K Blasdell, personal observation), this feature might promote connectivity between properties and facilitate the presence of these potential hosts. Similar to our study, BU prevalence was found to increase with decreasing elevation in Benin, with the authors proposing that MU survival might be promoted by the wetter conditions often found at lower altitudes (Sopoh et al., 2011) .

Although return surveys where only conducted at a small proportion of properties, the findings suggest that MU bacteria can remain at a specific location for a considerable period of time (>6 months). This has also been found in Cameroon, where a village water source remained positive for over two years (Bratschi et al., 2014) . However, as each property was only sampled at two time points, it is possible that undetected changes may have occurred at properties during that interval, and additionally that a property might remain positive for MU for longer than the maximum 8.7 months observed here. Although it is unknown what factors changed between sampling points for those properties where MU status did alter, some . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint environmental changes were observed at some of these properties that may have impacted the presence and survivability of MU. For example, at one property that became IS2404 positive at the return visit, de-vegetation and construction of a new house on the neighboring plot, which had previously been vacant and covered in native flora, may have resulted in the movement of infected wildlife onto the sampled plot. Most properties changed from positive to negative, which may suggest that the environmental disease risk in this region decreased slightly over the study period. At individual properties this may be because the resident infected possum (or other host) dies and is replaced by a non-infected individual, although this requires further exploration. However, two properties did become positive by IS2404, demonstrating this is a dynamic situation.

This study's findings cast some light on (1) what the 'ideal property' for MU presence looks like and (2) how this differs from what the 'ideal' case property looks like. The results suggest that the average MU property is a larger property located at lower altitudes with soil that is slightly alkaline. It has overhead powerlines and contains native vegetation, particularly coastal tea trees, which in turn support a healthy population of RT possums. In contrast, the 'ideal' case property has overhead powerlines present, is less likely to contain Moonahs, but more likely to contain MU-positive wild mammals, especially RT possums.

Further longitudinal investigation is required to understand these differences in environmental associations between properties where MU was detected and properties where human cases occurred. It should be noted that although there was a clear association between the presence of infected RT possums and human cases, MU-positive RT possum feces and other samples were also found at many control properties. Human behavioral impacts on BU disease risk will be assessed through the analysis of the questionnaires collected as part of this case-control study (results to be published separately). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022 Our findings provide additional evidence to support the hypothesis that MU is a zoonotic pathogen, at least in the Victorian endemic area. Although RT possums are clearly the strongest reservoir host candidate, the high proportion of fox feces that were IS2404 and confirmed positive, along with evidence of viability in one sample, indicate the foxes may also contribute to the circulation of this pathogen. Further research is needed to explore whether the more mobile fox could introduce MU to new areas.

Predicting summer site occupancy for an invasive species, the common brushtail possum (Trichosurus vulpecula), in an urban environment

Assessing waterrelated risk factors for Buruli ulcer: a case-control study in Ghana. The American journal of tropical medicine and hygiene

Detection of Viable Mycobacterium ulcerans in Clinical Samples by a Novel Combined 16S rRNA Reverse Transcriptase/IS 2404 Real-Time qPCR Assay. PLoS neglected tropical diseases

Buruli ulcer: a new case definition for Victoria

Mycobacterium ulcerans persistence at a village water source of Buruli ulcer patients. PLoS neglected tropical diseases

Potential wildlife sentinels for monitoring the endemic spread of human Buruli ulcer in South-East Australia. PLoS neglected tropical diseases

Impact of localized badger culling on tuberculosis incidence in British cattle

Positive and negative effects of widespread badger culling on tuberculosis in cattle

Localised Mycobacterium ulcerans infection in a cat in Australia

Development and application of two multiplex real-time PCR assays for the detection of Mycobacterium ulcerans in clinical and environmental samples

A major role for mammals in the ecology of Mycobacterium ulcerans. PLoS neglected tropical diseases

Mycobacterium ulcerans dynamics in aquatic ecosystems are driven by a complex interplay of abiotic and biotic factors

Scaling up from gardens: biodiversity conservation in urban environments. Trends in ecology & evolution

Home range and den use of common brushtail possums (Trichosurus vulpecula) in urban forest remnants

Buruli ulcer (M. ulcerans infection): new insights, new hope for disease control

Mycobacterium ulcerans in mosquitoes captured during outbreak of Buruli ulcer, southeastern Australia. Emerging infectious diseases

Risk of Buruli ulcer and detection of Mycobacterium ulcerans in mosquitoes in southeastern Australia. PLoS neglected tropical diseases

Effects of culling on Leptospira interrogans carriage by rats. Emerging infectious diseases

The use of hollows and dreys by the common ringtail possum (Pseudocheirus peregrinus) in different vegetation types. Australian journal of zoology

Champion of the suburbs" or "Our tormentors". Urban Wildlife: more than meets the eye

Handbook of Biological Statistics

Home range and movement of foxes (Vulpes vulpes) in coastal New South Wales

Epidemiology of Mycobacterium ulcerans infection in koalas (Phascolarctos cinereus) on Raymond Island, southeastern Australia

A new type of pathogenic Mycobacterium

Localised Mycobacterium ulcerans infection in four dogs

Mycobacterium ulcerans infection in two alpacas

Clinical, microbiological and pathological findings of Mycobacterium ulcerans infection in three Australian Possum species. PLoS neglected tropical diseases

Mycobacterium ulcerans in the elderly: more severe disease and suboptimal outcomes

Tackling the worsening epidemic of Buruli ulcer in Australia in an information void: time for an urgent scientific response

The changing epidemiology worldwide of Mycobacterium ulcerans

Feeding-Behavior and Diet of the Common Ringtail Possum, Pseudocheirus-Peregrinus, in Eucalyptus Woodlands and Leptospermum Thickets in Southern Victoria

Growth of mycobacteria in relation to the pH of the medium

Risk factors for Mycobacterium ulcerans infection, southeastern Australia. Emerging infectious diseases

Mycobacterium ulcerans DNA in Bandicoot Excreta in Buruli Ulcer-Endemic Area

Buruli ulcer prevalence and altitude, Benin. Emerging infectious diseases

Nontuberculous mycobacterial infections of the skin: report of fourteen cases and review of the literature

Increased severity and spread of Mycobacterium ulcerans, Southeastern Australia. Emerging infectious diseases

Mycobacterium ulcerans infections in two horses in south-eastern Australia. Australian veterinary journal

A large localized outbreak of Mycobacterium ulcerans infection on a temperate southern Australian island

Mycobacterium ulcerans low infectious dose and mechanical transmission support insect bites and puncturing injuries in the spread of Buruli ulcer. PLoS neglected tropical diseases

Effects of culling on badger Meles meles spatial organization: implications for the control of bovine tuberculosis

Mornington Peninsula: Ecological vegetation class profiles

Supplemental Figure 1: Key indigenous (panels A-D) and non-indigenous (panel E) plants recorded for each property. A -Melaleuca lanceolata

D -Allocasuarina verticillata/littoralis (Drooping and black sheoaks)

Supplemental Figure 4: Examples of plant samples collected; Panel A -selection of 'spiky' plants sampled; Panel B -selection of fruits with evidence of mammalian gnaw marks

We gratefully acknowledge Loretta Vaughen for help with accessing the VPHS database. We also thank all who assisted with field collections and questionnaire deployment, and all of the project participants, with particular thanks to those who allowed us to conduct field surveys at their properties.

This study was funded by NHMRC Partnership Project Grant APP1152807, "Controlling Buruli ulcer in Victoria", with co-funding provided by the Victorian Department of Health.

KRB created the study design, collected, analyzed and interpreted the data and drafted the manuscript. BM analyzed and interpreted the data and drafted the manuscript. DPO'B assisted with the study design, interpreted the data and drafted the manuscript. MT, SC and JG assisted with the study design and collected the data. VB, MD and PTM assisted with the study design, collected, analyzed and interpreted the data. IS, KBG, ET and SEL assisted with the study design, interpreted the data and drafted the manuscript. ECH and NW . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022 interpreted the data and drafted the manuscript. TPS and EA assisted with the study design, interpreted the data and drafted the manuscript. All authors reviewed and edited the manuscript. It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10.1101/2022.01.10.22269030 doi: medRxiv preprint Supplemental Figure 5 : Directed acyclic graph (DAG) used for assessing the potential for confounding by covariates and for identifying the appropriate confounders to be included in each adjusted model. Supplemental Figure 6 : Map of affected area, illustrating the MU status by suburb. Suburbs containing at least one 'viable' property were classified as viable. Suburbs without 'viable'properties but with at least one 'confirmed' property were classified as confirmed. Suburbs without 'viable' or 'confirmed' properties but with at least one 'IS2404 detected' property were classified as 'IS2404 detected'. Suburbs without any 'IS2404 detected' properties were classified as negative. N.B. Geographical boundaries are not available by postcode and some postcodes contain more than one suburb.. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted January 11, 2022. ; https://doi.org/10. 1101 /2022