This open-label randomized trial compared isoniazid (9 months) to rifampin (4 months) on toxicity and completion in a jailed population with latent tuberculosis infection. Rifampin resulted in fewer elevated liver function tests (risk ratio [RR] 0.39, 95% confidence interval [CI] [0.18, 0.86]) and less toxicity requiring medication withdrawal (RR 0.51, 95% CI [0.13, 2.01]), although one participant receiving rifampin experienced an allergic reaction. Completion was achieved for 33% receiving rifampin compared to 26% receiving isoniazid (p = .10). With careful monitoring rifampin is a safe and less toxic regimen and appears to be a reasonable alternative because of its shorter duration, allowing more people to complete treatment behind bars. Therapy completion in released inmates is unacceptably low and ensuring follow-up after discharge must be part of a decision to treat.