A 50-year-old geologist was referred in April 2005 for assessment of chronic cellulitis and ulceration of the right hand, present for 11 weeks (Box 1). An intraoperative swab showed acid-fast bacilli on Ziehl–Neelsen staining. Buruli ulcer (Mycobacterium ulcerans infection) was confirmed by polymerase chain reaction (PCR) testing. Over the next 3 months, the patient received a combination of antibiotic and surgical therapy. At follow-up 1 week after ceasing antibiotic therapy (4 months after initial presentation), the skin graft had healed, and the patient remained well.

The patient lived in rural New South Wales but had travelled to the Mornington Peninsula and East Gippsland regions of Victoria 5 years previously. Two weeks before the onset of illness, he had been fishing in the Snowy Mountains. He also made frequent work trips to Papua New Guinea, with seven trips in the 15 months before illness. His most recent trip there lasted 18 days, and he returned to Australia almost 12 weeks before the onset of infection.

Variable number tandem repeat (VNTR) typing (Box 2) revealed that the isolate had the same profile as a previously described M. ulcerans strain from PNG, which differed from Victorian strains at loci 8 and 19.1 Similarly, the isolate had a restriction fragment length polymorphism (RFLP) band profile identical to that of the PNG strain, but differing from that of Victorian strains.2 These results suggest that the infection was acquired in PNG, not Australia. The minimum incubation period in this case was therefore 3 months.

The second patient was a 42-year-old Australian citizen usually resident in Holland. He presented in Melbourne in January 2006 with a skin ulcer over the left ankle (Box 1) which had been present for 5 months. Buruli ulcer had been correctly diagnosed in the Netherlands, and he had been treated with appropriate antibiotics for 6 weeks with no apparent response. The lesion was excised, a skin graft applied, and further antibiotic therapy given for 6 weeks, resulting in complete resolution. PCR testing performed on resection specimens from the ulcer was positive for M. ulcerans.

The patient had travelled widely in Africa between 5 and 2 months before the appearance of the skin lesion; 7 months before its appearance, he had spent 2 weeks sea kayaking near Eden on the southern NSW coast. He had not visited Victoria for at least 10 years.

M. ulcerans was not able to be cultured, so typing was performed on DNA extracted from the resection specimens. The two VNTR loci that discriminate between African and Victorian strains were examined.1 The estimated PCR product sizes of 420 base pairs (locus 4) and 650 base pairs (locus 8) were consistent with the Victorian profile.1 A second typing method, multilocus sequence typing, revealed that the patient’s strain lacked the C-to-T substitution in the sodA gene characteristic of African isolates.3 These data suggest that the infection was acquired in southern NSW and not Africa. The incubation period in this case was therefore about 7 months.