OBJECTIVE To study the prevalence, risk factors, and genotypes of drug-resistant Mycobacterium tuberculosis in Karachi. METHODS Pulmonary tuberculosis (TB) patients were recruited in a cross-sectional study (2006-2009). Drug susceptibility testing was performed for culture-positive cases (n=1004). Factors associated with drug resistance were evaluated using logistic regression analysis. Strains were typed using spoligotyping and mycobacterial interspersed repetitive units-variable number tandem repeat (MIRU-VNTR). The associations of genotype and drug resistance were explored using the Chi-square test. RESULTS Resistance rates - new and previously treated - were as follows: multidrug-resistant (MDR)-TB, 2.4% and 13.9%, respectively; rifampin (RIF) monoresistance, 0.1% and 0.6%, respectively; any isoniazid (INH) resistance, 8.9% and 28.5%, respectively; and INH monoresistance, 3.0% and 6.3%, respectively. Prior TB treatment was a risk factor for MDR-TB (adjusted odds ratio (AOR) 6.8, 95% confidence interval (CI) 3.5-13.1) and INH monoresistance (AOR 2.4, 95% CI 1.1-5.2). Additional risk factors included low socioeconomic status for INH monoresistance (AOR 3.3, 95% CI 1.7-6.5), and belonging to Balouchi (AOR 9.2, 95% CI 2.5-33.4), Sindhi (AOR 4.1, 95% CI 1.2-13.5), or Pakhtun (AOR 3.4, 95% CI 1.0-11.2) ethnicity for MDR-TB. Although Central Asian strain (55.6%) was the most prevalent genotype, MDR-TB was significantly associated with Haarlem (H) genogroup (crude OR 9.2, 95% CI 3.6-23.8). CONCLUSIONS An MDR-TB rate of 2.4% is reported in new patients. Low RIF monoresistance supports the use of RIF as a marker for MDR-TB in this population. The need to strengthen TB care in the identified at-risk groups is emphasized. Based on INH resistance rates, a review of national treatment/prevention regimens relying on INH is suggested.