Abstract. Objectives . To explore the possible pathogenetic role of erythropoietin (EPO) in the anaemia associated with cyclosporin (Cs) in newly diagnosed patients with insulin‐dependent diabetes mellitus (IDDM). Design . A multicentre randomized placebo controlled prospective trial of Cs immunosuppression for 12 months in IDDM patients. Setting . Patients were recruited from the out‐patient clinics of diabetes centres in Europe and Canada. Subjects . Patients 9–35 years old with a clinical diagnosis of ketonuric IDDM entering less than 6 weeks after diagnosis. 188 patients were originally recruited; 105 patients completed the investigation, 52 patients being treated with Cs, and 53 patients receiving placebo. Interventions . Random allocation to receive either Cs or placebo. The initial dose of Cs was 10 mg kg −1 BW day −1 . Therapy was maintained for 12 months. Main outcome measures . B‐Haemoglobin, s‐creatinine, and s‐EPO concentrations were monitored before, during and after therapy with either Cs or placebo. Results . Blood‐haemoglobin (Hgb) fell from 8.5 ± 0.8 to a nadir of 7.8 ± 0.9 mmol l −1 at 6 months ( P < 0.0001) in IDDM patients treated with Cs but not in the placebo patients (8.5 ± 0.8 to 8.8 ± 0.9 mmol l −1 , NS). The mean serum EPO levels remained unaltered throughout the 6‐month period of Cs and placebo therapy. No significant differences in serum EPO levels between Cs and placebo‐treated diabetic patients were found after 6 months of treatment. Conclusions . The light anaemia associated with Cs therapy in IDDM patients is not related to an insufficient production of EPO, but is caused by other, as yet unknown mechanisms, unrelated to the nephrotoxic action of this drug.