Reactive oxygen species (ROS) have been mainly viewed as unwanted byproducts of cellular metabolism, oxidative stress, a sign of a cellular redox imbalance, and potential disease mechanisms, such as in diabetes mellitus (DM). Antioxidant therapies, however, have failed to provide clinical benefit. This paradox can be explained by recent discoveries that ROS have mainly essential signaling and metabolic functions and evolutionally conserved physiological enzymatic sources. Disease can occur when ROS accumulate in nonphysiological concentrations, locations, or forms. By focusing on disease-relevant sources and targets of ROS, and leaving ROS physiology intact, precise therapeutic interventions are now possible and are entering clinical trials. Their outcomes are likely to profoundly change our concepts of ROS in DM and in medicine in general. A New Approach to Diabetes Mellitus and Reactive Oxygen Species Diabetes mellitus (DM) and its related end-organ damage, such as diabetic nephropathy, neuropathy, retinopathy, and cardiomyopathy, are major causes of death and long-term disability. Their underlying mechanisms are incompletely understood, which is why none of the current antidiabetic therapies target the underlying causes or are curative, but focus instead on normalizing surrogate parameters or risk factors such as blood glucose or hypertension [1]. Hence, our lack of mechanistic understanding of lifestyle change-resistant diabetic end-organ damage, together with the increasing prevalence of DM, represent a significant major unmet medical need. One mechanism that has been suggested for decades to cause pancreatic b cell dysfunction and diabetic end-organ damage is 'oxidative stress' (see Glossary), originally defined as an overproduction of reactive oxygen species (ROS). Antioxidants were considered the obvious therapeutic countermeasure but, clinically, have consistently disappointed [2]. Even worse, meta-analyses of clinical trials show that antioxidants may not only be ineffective, but harmful, and even increase mortality [2]. Recently, however, important conceptual breakthroughs in our understanding of ROS in general and DM in particular explain the failure of antioxidants and point towards entirely different mechanism-based and possibly curative therapeutic approaches. Our new understanding of ROS requires that many long-held misconceptions, such as the 'redox balance hypothesis' and the view that ROS are primarily stressors, disease triggers, and metabolic waste products, must be overcome. Instead, the many physiological roles of ROS and the existence of at least seven evolutionarily conserved ROS-producing enzymes (NOX1-5,