Context: American Thoracic Society guidelines recommend a 9-month course of therapy with isoniazid for treatment of persons with latent tuberculosis infection who are at high risk for reactivation of disease. Major liver injury has been reported with the alternative regimen, a 2-month course of pyrazinamide and rifampin. Objective: To evaluate the rate of completion and incidence of hepatotoxicity of a short regimen of pyrazinamide and rifampin for latent tuberculosis as compared with standard isoniazid therapy before and after instituting an intensive monitoring program. Design, setting, and participants: Prospective cohort study of 224 patients in a community setting between 1999 and 2001. Interventions: Daily pyrazinamide and rifampin for 2 months or daily isoniazid for 6 months. Main outcome measures: Treatment completion, hepatotoxicity (fourfold increase of alanine transaminase [ALT]), severe hepatotoxicity (40-fold increase in ALT). Results: Treatment was completed by 71% (78 of 110 patients) in the pyrazinamide/rifampin group and by 59% (67 of 114 patients) in the isoniazid group (p = 0.07). Hepatotoxicity (ALT > 160 U/L) was documented in 13% (14 of 110 patients) in the pyrazinamide/rifampin group and in 4% (5 of 114 patients) in the isoniazid group (p = 0.03). Severe hepatotoxicity (ALT > 1,600 U/L) occurred in 2 of 43 patients (5%) receiving pyrazinamide/rifampin prior to instituting intensive monitoring. Once more intensive monitoring of liver enzymes was implemented, severe hepatotoxicity occurred in none of 67 patients. Conclusion: The risk of hepatitis in patients receiving pyrazinamide/rifampin for prevention of latent tuberculosis is increased threefold as compared to patients receiving isoniazid. When patients were monitored more intensively, severe hepatotoxicity did not develop, but the difference did not reach statistical significance (p = 0.15).