SETTING: Multidrug-resistant tuberculosis (MDR-TB) is a growing global health threat that often requires presumptive treatment in the absence of drug susceptibility testing (DST) results. OBJECTIVE : To compare two approaches to the treatment of MDR-TB contacts with no DST results who develop TB disease. DESIGN: We conducted a retrospective cohort study of adults treated for TB disease who were contacts of patients living with MDR-TB. Subjects had been treated according to one of two presumptive treatment strategies: 1) regimens containing exclusively first-line drugs, and 2) regimens that included both first-and second-line drugs that were adjusted if and when DST results became available. The primary endpoint was a composite of death and treatment failure. RESULTS : Household contacts of MDR-TB patients who developed TB disease and were treated with firstline regimens were significantly more likely to experience unfavorable end-of-treatment outcomes than those treated with presumptive MDR-TB regimens (RR 2.88, 95%CI 1.24-6.68). CONCLUSION: Household contacts of MDR-TB patients who develop TB disease but have no DST results should receive regimens containing second-line drugs selected based on the infecting strain of the index patient. Regimens containing only first-line anti-tuberculosis drugs significantly increase the risk of unfavorable outcomes.