Buruli ulcer: differences in treatment outcome between two centres in Ghana.

OBJECTIVES

Assess treatment effects by following up patients treated for Buruli ulcer in two hospitals with different treatment aspects, including widely differing surgical practices.

PATIENTS/

METHODS

Treated patients were retrospectively identified from hospital records. Between 1994 and July 2000, 136 patients had been admitted for Buruli ulcer in both hospitals, and lived in areas covered in the research period. 78 (57%) Patients were included in the study. Treatment and status of the patient were analysed.

RESULTS

27 (35%) Patients were not healed. Of the 33 patients treated in hospital A, six (18%) were not healed at follow-up, whereas of the 45 patients treated in hospital B, 21 (47%) were not healed. The length of stay in hospital A was significantly longer (P=0.002), and more operations on average were done per patient (P=0.002). In a univariate analysis, treatment in hospital A; the use of rifampicin (P=0.013); and BCG vaccination status (P=0.04) were all significantly associated with ulcer healing. Using a logistic regression model for multivariate analysis, only treatment as given in hospital A, with standard practice of wide surgical excision, appeared to predict ulcer healing independently (P=0.02).

CONCLUSIONS

This study shows large differences in treatment outcome between the two hospitals; the results support the hypothesis that extent of surgical treatment influences the chance of healing of Buruli ulcer.