Iron-catalyzed oxygen-free radical-induced oxidative stress mediates the pathogenesis of diabetes and cardiovascular disease (CVD). Diabetics (n = 473) and non-diabetics (n = 804) with CVD entered into a randomized trial of iron (ferritin) reduction by calibrated phlebotomy (, Identifier NCT00032357) had comparable iron measures at entry but diabetics had a greater burden of CVD and comorbidities, lower hemoglobin and hematocrit levels, and higher glucose levels than non-diabetics. Entry iron measures were lower in diabetics on hypoglycemic therapy compared to previously untreated diabetics. Diabetics and non-diabetics had comparable iron measures during follow-up. The Loess analysis of paired ferritin and hemoglobin, and paired ferritin and glucose levels in diabetics randomized to phlebotomy showed higher ferritin levels associated with lower hemoglobin and higher glucose levels. Progressive ferritin reduction in diabetics correlated with increasing hemoglobin and decreasing glucose levels, neither of which reached levels observed in non-diabetics. We postulate that phlebotomy-triggered autophagy (ferritinophagy) released redox-active iron sequestered intracellularly, worsening anemia and glucose utilization that corrected partially with ferritin reduction. Intracellular redox-active iron levels contributory to disease, not reflected in peripheral iron measures, may persist because of glycation of iron transport proteins in diabetes. These findings suggest novel strategies for disease prevention and improving outcomes in diabetes and CVD.