Daily administration of isoniazid for one year to persons infected with Mycobacterium tuberculosis is effective in considerably reducing the risk of disease. From a practical viewpoint, this approach to prevention is less than ideal because it results in considerable costs as health care providers monitor for possible hepatotoxic effects and because it is difficult to maintain compliance for 12 months. The efficacy and toxicity of isoniazid preventive therapy regimens of 12, 24, and 52 weeks' duration were recently assessed in a study conducted in Eastern Europe. We used data from this study to conduct a cost-effectiveness analysis of the three alternative regimens. The results indicate that over a wide range of assumptions, a regimen of 24 weeks' duration is more cost-effective than either the 12- or 52-week regimen. Using base case estimates, the cost per case prevented for the 24-week regimen was $7,112, compared with $16,024 for the 52-week regimen. Among a cohort of 1,000 persons treated, each additional case prevented by the 52-week regimen would cost $80,807. Thus, a shorter course of isoniazid preventive therapy is relatively cost-effective compared with current policy.