Thoracic duct drainage (TDD) was established for 21-115 days in 40 kidney recipients with an average removal per patient day of 4.7 1 lymph and 1.88 billion cells. Cellular and humoral immunity were depressed. TDD and immunosuppressive drugs were started at transplantation in 35 recipients of cross-match negative grafts. Although the results were better than in precedent non-TDD controls, eight patients rejected their grafts before a full TDD effect, and three of the eight developed predominantly anti-B lymphocyte cytotoxic antibodies which were probably responsible for positive cross-matches with their next donors. With continuing TDD, all eight patients had good initial function after early retransplantation. In five more "nontransplantable" patients with performed cytotoxic antibodies, TDD was started 30-56 days before transplantation. In these five pretreated patients, antibodies persisted with positive antidonor cross-matches. Hyperacute rejection occurred repeatedly in two patients with high anti-T (and anti-B) titers, but was surmounted in three patients with lower titers. From the clinical and immunologic data, we have concluded that TDD should be used for pretreatment of all cases with or without prior antibodies, and have suggested an adjustable management plan that takes into account new developments in antibody monitoring.