Clinical and programmatic mismanagement rather than community outbreak as the cause of chronic, drug-resistant tuberculosis in Buenaventura, Colombia, 1998

SETTING: Buenaventura, Colombia. OBJECTIVE: To assess whether antituberculosis drug resistance was generated by poor management or community transmission. DESIGN: Treatment-failure and new tuberculosis (TB) patients identified bet ween May 1997 and June 1998 were interviewed and their treatment histories reviewed. Bacteriologic testing, including drug susceptibility profiles (DS P) and DNA fingerprinting by restriction fragment length polymorphism (RFLP ), was performed and human immunodeficiency virus (HIV) testing was offered . RESULTS: DSP and RFLP fingerprints were obtained for isolates from 34 of 64 treatment-failure patients; 25 (74%) were resistant to greater than or equ al to one drug. Fifteen of the 25 patients consented to HIV testing; none w ere positive. An average of 2.8 major treatment errors per patient was iden tified. RFLP from the treatment-failure patients revealed 20 unique isolate s and six clusters (isolates with identical RFLP); 4/6 clusters contained i solates with different DSP. Analysis of the RFLP from both treatment-failur e and new patients revealed that 44/111 (40%) isolates formed 18 clusters. Four of 47 (9%) new patients had multidrug-resistant TB (MDR-TB). Eleven is olates belonged to the Beijing family, related to the MDR strain W. CONCLUSION: Drug resistance in Buenaventura results from both poor manageme nt and community transmission. Dependence on DSP to identify TB transmissio n is inadequate when programmatic mismanagement is common.