To describe the molecular epidemiology of tuberculosis (TB)-related deaths in a well-managed program in a low-HIV area, we analyzed data from a cohort of 454 pulmonary TB patients recruited between March 1995 and October 2000 in southern Mexico. Patients who were sputum acid-fast bacillus smear positive underwent clinical and mycobacteriologic evaluation (isolation, identification, drug-susceptibility testing, and IS6110-based genotyping and spoligotyping) and received treatment from the local directly observed treatment strategy (DOTS) program. After an average of 2.3 years of follow-up, death was higher for clustered cases (28.6 vs. 7%, p=0.01). Cox analysis revealed that TB-related mortality hazard ratios included treatment default (8.9), multidrug resistance (5.7), recently transmitted TB (4.1), weight loss (3.9), and having less than 6 years of formal education (2). In this community, TB is associated with high mortality rates.