Common for all older studies is the use of less potent anti-tuberculosis chemotherapy as compared with the present. The results of these studies cannot without reservation be used in the present setting. The newer, prospective, randomized placebo-controlled trials include rather few patients and their number is limited. The results must be interpreted carefully. There is no reason to give prophylactic isoniazide treatment to Mantoux-positive patients or patients with earlier tuberculosis who start treatment with corticosteroids. If allergic reactions to one of the antituberculous drugs emerge during therapy, the treatment can, if necessary, be continued if corticosteroids are added. Pericardial tuberculosis and atelectasis in children with endobronchial tuberculosis should be treated with corticosteroids, as can pleural disease with prolonged fever and exudation. Cases of severe pulmonary tuberculosis may be treated with supplementary steroids. The effect seems modest. Patients with tuberculous meningitis, stages II and III seem to benefit from corticosteroid-treatment.