In controlled clinical trials, several regimens have proven effective in treating tuberculosis. The drugs must be given in combination to avoid the emergence of resistant organisms. The simplest regimen is based on the administration of isoniazid and rifampin for 9 months. Ethambutol or streptomycin generally is added for an initial 2-8 weeks. Good results also have been reported using twice weekly administration of the isoniazid and rifampin after an initial 1 month of daily treatment. Relapse rates of less than 3% have been reported with these regimens. A table shows the recommended doses for the drugs. If neither isoniazid nor rifampin can be used, the patient needs to be given at least 2 and preferably 3 drugs to which his organisms are known to be susceptible. Additionally, these drugs must be administered for at least 18 months. Several circumstances are associated with a higher rate of drug resistant organisms and call for susceptibility studies: tuberculosis among individuals known to have a higher prevalence of drug resistance such as Asians or Hispanics; previous treatment; persistence of culture-positive sputum after 3 months of therapy; and exposure to drug resistant tuberculosis. The recommendations for the treatment regimen need to be changed in several special situations, including: infection with mycobacteria other than "M. tuberculosis;" tuberculosis in children; disease in organ systems other than the lungs; and the presence of certain patient characteristics, concurrent diseases, or the development of adverse drug reactions. Much scientific data supports the value of isoniazid (INH) in the prevention of tuberculosis. Isoniazid preventive therapy is recommended for several groups, listed in order of priority: household members and other close associates of potentially infectious tuberculosis cases; newly infected persons; persons with significant reactions to tuberculin skin test and abnormal chest roentgenograms; persons with significant reactions to tuberculin skin tests who are in special clinical situations; and tuberculin skin test reactors under age 35 with none of the other risk factors. Prior to administration of INH preventive therapy, it is important to: exclude bacteriologically positive or progressive tuberculous disease; question for a history of prior INH administration to exclude those who have had a adequate course of the drug; check for contraindications to the administration of INH for preventive therapy; and identify patients for whom preventive therapy is not contraindicated but in whom special precautions are indicated. Other drugs might be effective for preventive therapy, but as yet there are currently no data available documenting the efficacy of any drug other than INH.