BackgroundLeft ventricular hypertrophy (LVH) is highly prevalent in haemodialysis (HD) patients and is associated with an increased risk of death. Roxadustat and recombinant human erythropoietin (rHuEPO, abbreviated as EPO) are the main treatment strategies for renal anaemia in HD patients, but it has not been clear whether there is a difference in their effect on LVH.MethodsIn this multi‐centre, prospective, randomized trial of 12‐month duration, study participants were randomized in a 1:1 ratio to the roxadustat group or the EPO group. The doses of both treatment regimens were adjusted so that the patients had a haemoglobin level of 10.0–12.0 g per dL. The primary study endpoint was the change from baseline to 12 months in the left ventricular mass index (LVMI, g/m2) measured by echocardiography.ResultsIn total, 114 patients were enrolled. The mean age was 50 years, and the median dialysis duration was 33 months. Sixty‐one patients were men, and 24 were diabetic. LVMI decreased from 116.18 ± 27.84 to 110.70 ± 25.74 g/m2 in the roxadustat group. However, it increased from 109.35 ± 23.41 to 114.99 ± 28.46 g/m2 in the EPO group, with a significant difference in the change in LVMI between the two groups [−5.48 (−11.60 to 0.65) vs. 5.65 (0.74 to 10.55), p < 0.05]. Changes in left ventricular mass, end‐diastolic volume and 6‐min walk test seemed superior in the roxadustat group. There were no significant differences in other cardiac geometry, biochemical parameters and major adverse cardiovascular events between the two groups.ConclusionsCompared to EPO, roxadustat is more helpful in the regression of LVH in HD patients.