C. M. Steffen MBBS, FRACS; M. Smith MSc, OAM; W. J. H. McBride MBBS, PhD.
This paper has not previously been presented at a conference or society
Professor William J. H. McBride, James Cook University, School of Medicine and Dentistry, Clinical School, Cairns Base Hospital, PO Box 902, Cairns, QLD 4870, Australia. Email: firstname.lastname@example.org
Background: As the third most common mycobacterial infection in the world after tuberculosis and leprosy, Mycobacterium ulcerans is a major health and development problem that has become the focus of a World Health Organisation (WHO) initiative seeking to reduce the burden of this disease. The Daintree River catchment in north Queensland is an endemic focus for Mycobacterium ulcerans infection, known locally as the ‘Daintree Ulcer’. The aim of this study is to analyse the changing pattern of the disease over the last 44 years in the region.
Methods: The study is a descriptive review of all human cases where infection had been acquired in an endemic region of far north Queensland. Demographic information, lesion characteristics, management and outcomes were recorded in a database.
Results: Over the period there were 92 cases of M. ulcerans infection. The average age was 41.7 years and 56.7% were male. There was distinct clustering of cases in several defined locations in the area. Most cases (90%) presented with ulcers, and all but three patients were treated surgically. Excision and skin grafting was more commonly used in the period between 1964–1988. Excision alone was the most common treatment used thereafter. Earlier recognition and smaller lesion size is thought to have contributed to the change in surgical practice.
Conclusions: Ulcers caused by M. ulcerans infection are a distinct clinical entity in tropical North Queensland. Early recognition, diagnosis and prompt surgical intervention has minimised morbidity associated with this disease.
The term ‘Daintree Ulcer’ was coined in the mid-20th century and referred to an ulcerating, ‘flesh-eating’ infection of the skin that seemed to occur only in people frequenting the catchment of the Daintree River and adjacent swampy coastal lowlands in tropical far North Queensland. Starting as a small nodule in the skin, it would progress to an extensive but usually painless ulceration, eventually healing with intense scarring. It did not respond to antibiotics.
It was not until the atypical mycobacterium was isolated from two Daintree ulcer cases in the early 1960s that it became clear that this ulcer was caused by the same infecting organism as the ‘Bairnsdale’ or ‘Buruli’ ulcers – similarly named after places where infections were occurring in Australia and Uganda, respectively. It was first cultured from a Bairnsdale ulcer and reported by McCallum in 1948.1 The disease is officially named ‘Buruli Ulcer’ after the focus in Uganda.2
Mycobacterium ulcerans is now the third most common mycobacterial infection in the world after tuberculosis and leprosy. As a major health and development problem, especially in West Africa, Buruli ulcer has become the focus of a World Health Organization (WHO) initiative seeking to reduce the burden of this disease.3
In North Queensland, the Daintree River gathers its waters from the mountainous rainforested region north-west of the small town of Mossman and flows into the sea at Cape Tribulation. Settlement is sparse, composed of some smaller communities and farms. The population is composed of settlers of mainly European ethnic origins, their descendants and indigenous Aboriginal families.
Our experience with the ‘Daintree ulcer’ is through the surgical outreach service to Mossman Hospital from Cairns Base Hospital. This local datum has not been published previously. Ulcers caused by M. ulcerans are a distinct clinical entity, and there is considerable local expertise in diagnosis and treatment. This paper presents an audit of cases over a 45-year period from 1964 to 2008.
This is a descriptive study of the characteristics of disease due to M. ulcerans acquired in the endemic region of the disease in the vicinity of Mossman.
Since 1964, patients with the ‘Daintree Ulcer’ caused by M. ulcerans have had details of their condition recorded in a database. Patients presenting to Mossman and Cairns Base Hospitals make up the majority. Hospital discharge and pathology laboratory records were used to ensure completeness of the database. Three patients were managed privately. The database records age, gender, ethnicity, place of residence, site of lesion, type of lesion, month and year of presentation, possible aetiology, management and outcome.
All cases were confirmed by at least two diagnostic modalities for M. ulcerans (positive smear for acid fast bacilli, positive polymerase chain reaction (PCR) using primers specific for M. ulcerans, positive culture for M. ulcerans or histology demonstrating typical features of the infection) and a clinical presentation consistent with M. ulcerans infection.
Ninety-four records were identified. Two cases were not included as the infection was acquired at one of the known Victorian endemic sites, giving 92 cases to date.
Median age at diagnosis was 42 years. Average age at diagnosis was 41.7 years. Fifty-three patients were male, 39 were female. The majority were of Caucasian ethnic origin (89 of 92). The remaining three patients were indigenous Australians.
Eighty-eight of 92 (96%) were living in the Douglas Shire at the time of diagnosis. Thirty-seven patients (40%) were from the Daintree locality on the Daintree River itself, and 30 (33%) from the swampy coastal area of Miallo/Wonga/Cooya Beach. Twenty-one lived in other parts of the Douglas Shire but visited the Daintree area frequently (Fig. 1).
Of the four non-residents, three had visited the Daintree area prior to the diagnosis of infection. One patient was from the Atherton tableland approximately 80 km away and had not visited the Daintree catchment at all.
At presentation for treatment, five patients had nodules only, the remainder had ulcerating lesions (83/92, 90%). Three patients presented with an extensive oedematous form of M. ulcerans infection. One patient had a plaque of oedema surrounding a small ulcer. All oedematous forms involved the upper limb.
Twenty-one (23%) patients recalled an insect bite or minor trauma at the site as a precipitating event, usually in the weeks prior to lesion development.
The predominant sites for M. ulcerans infections were the extremities, (89/92, 97%), and most of these (73/89, 82%) were on the lower limb.
The majority of the patients were treated surgically. Forty-seven patients underwent excision alone, and 40 had excision with skin grafting (Fig. 2). Excision plus skin graft was the commoner procedure up to 1988, simple excision becoming more frequent after 1988 because of earlier presentation with smaller lesions (Fig. 3). There were two patients who required amputation, a hand and distal forearm amputation in 1966 and a below knee amputation in 1998 in a male with very extensive M. ulcerans infection and concurrent human immunodeficiency virus infection. The patients with oedematous type infections required extensive excision of skin and subcutaneous tissue and meshed skin grafts.
Three patients did not undergo surgery. One had disseminated carcinoma and was treated palliatively, ‘topical’ treatment in one resulted in healing and a third patient refused treatment and was lost to follow-up.
Two patients died during the period of treatment – one from a pulmonary embolus while recovering from surgery, the other from disseminated carcinoma.
Initial surgical treatment was not curative for 10 patients. The first five, in the period from 1964 to 1988 underwent multiple procedures and received adjuvant therapies in the form of local heat application and various combinations of antibiotics known to be active against M. tuberculosis, including amikacin and rifampicin, plus ethambutol or clofazamine. Healing was achieved in all the patients but with extensive scarring and long hospital stays.
After 1988, all healed after limited further excision and grafting only. Three were on the foot, one on the thigh and one was in a patient with the oedematous form of infection.
Over the entire 45-year period of the study, there was an average of just over two cases per year. However, over the 27-year period from 1964 to 1990, there were 27 cases and from 1991 to 2008, there were 65 cases (chi-squared P < 0.001) (Fig. 4).
The indigenous people of the Mossman–Daintree area have been aware of this disease entity long before European settlement, and there are numerous anecdotal reports. In the 1960s, indigenous people of the region moved from missions in areas where acquisition is known to occur, to an area from which no cases have been recorded. Overall, the incidence for indigenous and non-indigenous populations is the same.
The first documented but not microbiologically proven cases at Mossman Hospital were in the early 1950s, with a consistent record of cases dating from 1964.4
It is still an uncommon infection in this region, with one to four cases per year most years. Most cases present during the dry season and increased numbers presented in the dry seasons after flooding from cyclones or ‘big wets’ in 1974 and 1998/1999. This pattern has been observed in other series.5
Of interest are 10 asynchronous cases in genetically related family members, raising the possibility of a genetic predisposition. The single case of an individual who had not visited the area at all suggests that there are other local sites for infection, not surprising given the worldwide distribution of M. ulcerans and reports of cases from various other sites in Eastern Australia.6–9
An index of suspicion combined with knowledge of the appearance and behaviour of M. ulcerans infection is obviously important.6 While some infections are known to resolve without intervention, most progress.10,11
The Daintree ulcer usually starts as a nodule that ulcerates. The ulcer is characteristically painless unless secondarily infected. The edges are undermined, often extensively, and there is a rim of induration beyond the ulcer that is characteristic and indicative of active infection in this tissue.
Necrosis of underlying subcutaneous fat leads to the ulceration and is due to a cytopathic toxin, mycolactone, which has several variants produced by different strains of M. ulcerans.12 The Australian variants are considered less biologically active than the African variants.13 The toxin also dampens the host immune response in the vicinity of the infection producing a typically indolent ulcer.14
Later in the disease process, there may be healing of ulcers with scarring and extension of the disease to adjacent areas or the development of satellite lesions. Infrequently, the infection may involve bone. In the less common oedematous variant of the disease, there is extensive involvement of skin and subcutaneous tissue (e.g. whole forearm) and rapid extension of the infection.
In the endemic region, a clinically consistent lesion that is smear positive for acid-fast bacilli indicates a high likelihood of M. ulcerans infection. In the past, confirmation relied on culture of the organism, which could take many months. With the advent of PCR testing, confirmation can be achieved within days, enabling early definitive treatment.15
The mainstay of treatment remains wide local excision with or without skin grafting. In our series, the majority of cases up to 1988 were treated with excision and grafting as they were larger lesions. After 1988, the proportion of lesions treated by excision alone rose. Increasing awareness of M. ulcerans infection at the community and local health provider level combined with improved diagnosis by PCR testing are thought to be factors underlying a trend to smaller lesions at presentation.
The ‘recurrences’ in the early part of the series also highlight that during the time it took to establish the diagnosis, the infection could progress significantly. The difficulty then was to gain clear margins.
Four episodes of recurrence in the latter part of the series were on a much smaller scale. Three were at sites where wide margins were difficult to obtain, namely a heel, a toe and a malleolus. A fourth recurrence was in the thigh of a two-year-old boy following an attempt at simple excision. Local anaesthetic, infiltrated widely around the initial excision site, for post-operative pain relief may have contributed to the rapid and extensive recurrence. The fifth recurrence was in a case of oedematous M. ulcerans infection involving the forearm, where margins were, not unexpectedly, inadequate, and further excision was necessary.
Simultaneous excision and grafting resulted in several graft failures in the series. A revised recommended technique is to excise the lesions to beyond the indurated area and to delay grafting for several weeks.3 This has given better graft take rates and a reduction in the area requiring grafting.
Several patients early in our series were treated with antibiotics and some had received Bacillus Calmette-Guérin (BCG) vaccinations as recommended in recent publications and by the WHO Global Buruli Ulcer Inititative.16,17 The small number treated in this way does not enable us to draw any conclusions as to efficacy.
Though research shows there is a role for adjuvant antibiotics to reduce recurrence in larger lesions, in the context of our mostly small lesions amenable to minor surgery, toxicity and variable response to antibiotics have favoured surgery as an initial and only treatment. The combination of the oral agents rifampicin and either clarithromycin or ciprofloxacin may be more useful in our context than regimens that include aminoglycosides.18,19
A proportion of patients in this series gave very definite histories of spider or insect bites preceding development of their Daintree ulcer. That M. ulcerans is contracted from an environmental source has long been known, given that it is a skin infection and occurs in specific geographic localities. However, culture of M. ulcerans from the environment or potential vectors has, until recently, been an elusive goal. With positive cultures from various water insects now obtained, the concept of infection via insect bite is supported.20–22
Further evidence supporting the role of mosquitoes in transmission has also been recently published.23,24 The reasons for the unique distribution of this infection in North Queensland are currently unknown. Insect exposure is consistent with the finding that the commonest site of ulceration in our patients was the exposed extremities. The usual mode of dress for local people is shorts, short-sleeved shirt and minimal footwear. Preventive measures such as more protective clothing, insect repellent and avoidance of implicated rivers and streams have been suggested as means to reduce likelihood of infection.18
The ‘Daintree ulcer’ has been a long-standing health problem for the people of this well-circumscribed region of North Queensland from prior to European settlement. Its presence in our region rarely receives mention in publications on M. ulcerans.25 It is nevertheless important for the affected communities to be aware of, and for local health practitioners to be able to diagnose and manage it effectively in its early clinical stages. Left untreated or treated with ineffective conservative measures, it can result in serious disfigurement, loss of physical function and social and economic disadvantage. The region is a popular tourist destination, and recognition of this condition in returned travellers is particularly important.26