Erythropoietin was administered to five anemic azotemic diabetic subjects for 1 year to assess the effect of increasing red cell mass on clinical well-being and the course of renal functional decline. None of the subjects manifested worsened hypertension or cerebrovascular or cardiovascular complications despite an increase in mean hematocrit from a baseline mean of 29.6% to a mean of 39.5%. The serum creatinine concentration after 1 year of treatment with erythropoietin was 3.7 mg/dL, which was unchanged from the baseline value of 3.5 mg/dL. Plasma viscosity remained constant as red cell mass increased. Although the viscosity of whole blood rose as the hematocrit increased, it was within the range of normal blood viscosity for an equivalent hematocrit. The favorable impact of erythropoietin treatment on three diabetic subjects who had macular edema and anemia is described. One hypothesis to explain the benefit of a raised hematocrit on both diabetic nephropathy and retinopathy is that the metabolic, hormonal, and hemodynamic components of the diabetic syndrome, in concert, produce tissue and cellular hypoxia that is ameliorated in part by the greater oxygen-transporting capacity of a raised red cell mass. The pseudohypoxia of diabetes may be implicated in the pathogenesis of diabetic neuropathy, retinopathy, muscular dysfunction, and nephropathy.