Patients with refractory metastatic renal cell carcinoma (RCC) were enrolled in a phase II study with teniposide (VM26) and cyclosporin A (CSA) to investigate (1) the effect of CSA on the response rate to VM26; and (2) the effect of CSA on the pharmacokinetics and pharmacodynamics of VM26. Sixteen patients initially received VM26 alone (200 mg m(-2) day(-1) i.v.). No objective responses were observed and all patients crossed over to receive at least an additional two courses (range 2-5) of VM26 plus CSA (5 mg kg(-1) 2h(-1) followed by 30 mg kg(-1) 48h(-1) i.v.). At the end of the 2-h loading dose of CSA, whole-blood CSA levels ranged from 2250 to 3830 ng ml(-1), whereas at the end of the 48-h CSA infusion, CSA ranged from 1830 to 4501 ng ml(-1). CSA significantly (P<0.01) increased the area under the curve (AUC) of VM26. The variation in the paired AUC of VM26 was 50%. Terminal half-life of VM26 was significantly (P<0.01) increased (1.72-fold) after CSA administration, whereas the systemic clearance of VM26 was decreased by 1.4-fold (P<0.01). The nadir neutrophil count after VM26 plus CSA (median 700 microl(-1), range <100 to 2860 microl(-1)) was lower than after VM26 alone (median 1900 microl(-1), range 200 to 6000 microl(-1)). Increased haematological toxicity after CSA could be explained by the increase in the VM26 AUC and by inhibition of P-glycoprotein (P-gp) activity in haematopoietic precursor cells. Bilirubin concentrations in the serum were increased after VM26 plus CSA compared with VM26 alone (P<0.01). Among the 15 patients evaluable for response, one had a minor response, eight had stable disease, and six had progressive disease. In conclusion, the dose of CSA we used achieved plasma concentrations within the effective range for P-gp inhibition. CSA affected both the pharmacokinetics and pharmacodynamics of VM26 in the patients, principally by increasing the plasma concentrations of the antineoplastic drug and VM26 haemopoietic toxicity.