Short, highly effective, and inexpensive standardized treatment of multidrug-resistant tuberculosis.

RATIONALE: Based on expert opinion, the global guidelines for management of multidrug-resistant tuberculosis impose lengthy and often poorly tolerated treatments. OBJECTIVES: This observational study evaluates the effectiveness of standardized regimens for patients with proven multidrug-resistant tuberculosis previously untreated with second-line drugs in low-income countries. METHODS: Consenting patients were sequentially assigned to one of six standardized treatment regimens. Subsequent cohorts were treated with regimens adapted according to results in prior cohorts. The study was designed to minimize failure and default while reducing total treatment duration without increasing relapse frequency. MEASUREMENTS AND MAIN RESULTS: We report the treatment outcome of all patients with laboratory-confirmed, multidrug-resistant tuberculosis enrolled from May 1997 to December 2007. The most effective treatment regimen required a minimum of 9 months of treatment with gatifloxacin, clofazimine, ethambutol, and pyrazinamide throughout the treatment period supplemented by prothionamide, kanamycin, and high-dose isoniazid during an intensive phase of a minimum of 4 months, giving a relapse-free cure of 87.9% (95% confidence interval, 82.7-91.6) among 206 patients. Major adverse drug reactions were infrequent and manageable. Compared with the 221 patients treated with regimens based on ofloxacin and commonly prothionamide throughout, the hazard ratio of any adverse outcome was 0.39 (95% confidence interval, 0.26-0.59). CONCLUSIONS: Serial regimen formulation guided by overall treatment effectiveness resulted in treatment outcomes comparable to those obtained with first-line treatment. Confirmatory formal trials in populations with high levels of human immunodeficiency virus coinfection and in populations with a higher initial prevalence of resistance to second-line drugs are required.

Endogenous reactivation and true treatment failure as causes of recurrent tuberculosis in a high incidence setting with a low HIV infection.

OBJECTIVE: To determine the relative frequencies of reinfection vs. reactivation or treatment failure in patients from a high tuberculosis incidence setting with a low prevalence of HIV infection. METHOD: We performed DNA fingerprinting on serial isolates from one and multiple TB episodes from 97 retreatment patients; 35 patients had been previously cured, whereas 62 had not. RESULTS: DNA fingerprinting patterns of recurrence Mycobacterium tuberculosis isolates of 5 of the 35 previously cured patients did not match with those of the corresponding initial isolates, indicating reinfection. We did not document reinfection during treatment. Isolates from each of the remaining 30 previously cured patients had identical DNA fingerprinting results, indicating reactivation. DNA fingerprinting patterns of isolates from the 62 patients with persistently positive sputum smears were identical, suggesting treatment failure. CONCLUSION: These findings suggest that reinfection is not a common cause of relapse and treatment failure in this rural predominantly HIV-free population despite the high incidence of TB.