One dose containing 3,100 PFU of a live attenuated Oka-strain varicella vaccine (Varilrix, Smith Kline-RIT, Belgium) was administered subcutaneously to 45 children, 26 of whom were suffering from acute leukaemia and 19 from solid malignant tumours. Their immunological status had been severely compromised by chemotherapy as evidenced by markedly low values for all immunological parameters. Of the 31 children seronegative for varicella at the time of vaccination, 70%, 85%, 75%, and 40% had varicella serum antibodies at 1, 2, 3, and 12 months after vaccination, respectively. Evaluation of the immunological parameters indicated that they were of no predictive value regarding the antibody response of the patients to the vaccine. Eight children (18%) developed varicella after vaccination. In one case, the disease was caused by the vaccine virus while in another, the vaccine virus was probably also responsible. The remaining six cases were caused by wild virus. The antibody response and accompanying protection against disease induced by the vaccine in severely immunodepressed patients is acceptable, but clearly lower and of shorter duration than in normal subjects. Thus, in immunocompromised patients, booster doses of the vaccine may be required at relatively short intervals.