Abstract

INTRODUCTION: Buruli ulcer (BU) caused by Mycobacterium ulcerans is effectively treated with rifampicin and streptomycin for 8 weeks but some lesions take several months to heal. We have shown previously that some slowly healing lesions contain mycolactone suggesting continuing infection after antibiotic therapy. Now we have determined how rapidly combined M. ulcerans 16S rRNA reverse transcriptase / IS2404 qPCR assay (16S rRNA) became negative during antibiotic treatment and investigated its influence on healing. METHODS: Fine needle aspirates and swab samples were obtained for culture, acid fast bacilli (AFB) and detection of M. ulcerans 16S rRNA and IS2404 by qPCR (16S rRNA) from patients with IS2404 PCR confirmed BU at baseline, during antibiotic and after treatment. Patients were followed up at 2 weekly intervals to determine the rate of healing. The Kaplan-Meier survival analysis was used to analyse the time to clearance of M. ulcerans 16S rRNA and the influence of persistent M ulcerans 16S rRNA on time to healing. The Mann Whitney test was used to compare the bacillary load at baseline in patients with or without viable organisms at week 4, and to analyse rate of healing at week 4 in relation to detection of viable organisms. RESULTS: Out of 129 patients, 16S rRNA was detected in 65% of lesions at baseline. The M. ulcerans 16S rRNA remained positive in 78% of patients with unhealed lesions at 4 weeks, 52% at 8 weeks, 23% at 12 weeks and 10% at week 16. The median time to clearance of M. ulcerans 16S rRNA was 12 weeks. BU lesions with positive 16S rRNA after antibiotic treatment had significantly higher bacterial load at baseline, longer healing time and lower healing rate at week 4 compared with those in which 16S rRNA was not detected at baseline or had become undetectable by week 4. CONCLUSIONS: Current antibiotic therapy for BU is highly successful in most patients but it may be possible to abbreviate treatment to 4 weeks in patients with a low initial bacterial load. On the other hand persistent infection contributes to slow healing in patients with a high bacterial load at baseline, some of whom may need antibiotic treatment extended beyond 8 weeks. Bacterial load was estimated from a single sample taken at baseline. A better estimate could be made by taking multiple samples or biopsies but this was not ethically acceptable.