The aim of this study was to quantify and where possible objectively confirm the magnitude of non-compliance (NC) in our paediatric renal transplant recipients. A total of 94 paediatric transplants were performed between 1984 and 1989; 17 were excluded due to graft loss (2), death (3), oxalosis (2) and transfer to the adult unit (10). NC was assessed as missed clinic visits plus medication shortages or actual admission of NC. NC was found in 22% (17/77) of transplanted patients. NC showed no correlation with parental marital status, sex, distance lived from the hospital, pre-emptive transplant status or total lymphoid irradiation. Most NC was peripubertal with a smaller NC in the late teenager group. Social class correlated positively with NC; 82.3% of NC was from social classes III and IV, who formed 52.4% of the patients. NC in social class II (3/26) was significantly different from social class IV (12/24) (P = 0.01); 91% of black patients with NC were from social class IV. Race, corrected for social class, failed to reach significance (P > 0.05). Confirmation of compliance was sought from retrospective cyclosporin A (CsA) trough levels (twice daily dosage). Concomitant phenytoin therapy and CsA given as a daily dosage were excluded as significant confounding variables. The CsA dosage was not significantly different between the compliant (C) and patients with NC. Patients with NC were 8 times more likely to have a CsA level < 10 ng/ml (P = 0.0026) than C patients.(ABSTRACT TRUNCATED AT 250 WORDS)