Aims The use of an erythropoesis-stimulating agent, darbepoetin alfa (DA), to treat anaemia in patients with diabetes mellitus and chronic kidney disease was associated with a heightened risk of stroke and neutral efficacy in the Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT), despite epidemiological data suggesting the contrary. However, this association has not been evaluated in another randomized, placebo-controlled trial. Methods and results Reduction of Events by Darbepoetin Alfa in Heart Failure (RED-HF) was a randomized placebo-controlled trial of DA in 2278 patients with systolic heart failure and anaemia, enrolled from 2006 to 2012 and followed for a median of 28 months. Within RED-HF, 816 patients had diabetes mellitus and chronic kidney disease [estimated glomerular filtration rate (eGFR) 20-60 mL/min/1.73 m2] and met inclusion criteria for TREAT. TREAT-like RED-HF patient data were analysed alone and combined at the patient level with the 4038 TREAT patients. In RED-HF, the annualized event rate of stroke was 2.3 in patients on DA and 1.1 in patients randomized to placebo (P = 0.051). Analysis of the combined group (n = 4854) confirmed a nearly two-fold increase in stroke risk [hazard ratio (HR) 1.94, 95 confidence interval (CI) 1.43-2.63] and an overall neutral effect on mortality (HR 1.00, 95 CI 0.89-1.12) of raising haemoglobin with DA. Conclusion The placebo-controlled cohort of heart failure patients with anaemia, diabetes mellitus, and chronic kidney disease from RED-HF provides confirmation of the increased stroke risk associated with DA use identified in TREAT.