Capillary blood sampling as a means of monitoring blood cyclosporine levels has replaced venipuncture in some medical centers. As the validity of capillary venipuncture for analysis of cyclosporine has not been documented, we sought to validate the capillary blood collection technique by comparing it with serum samples collected simultaneously by venous phlebotomy. Forty paired capillary- and venous samples were collected from 36 cardiac transplantation patients and analyzed, using a polyclonal immunoassay. The values obtained were compared using regression correlation. The correlation coefficient for all 40 samples was 0.859. However, we discovered that the first 7 capillary specimens were processed incorrectly. The correlation coefficient for the other 33 samples was 0.995 (99% confidence interval 0.987-0.998). The excellent correlation between serum samples obtained from capillary sampling and from venous sampling, together with the ease of obtaining capillary blood specimens, make 'fingerstick' sampling the method of choice for monitoring cyclosporine levels in infants and children.