Nicotinamide can protect the NOD mouse from diabetes if given early enough and in sufficient dose. The effect partly wanes with time. There is reduced islet inflammation. Similar protective effects can be demonstrated in quasi-experimental interventions in humans--both diabetes related and unrelated deemed at risk of developing diabetes by reason of having islet cell antibodies. Nicotinamide protects isolated islets in vitro from the toxicity of a number of agents, but only in doses that produce significant PARP inhibition, and increased intracellular levels of NAD. It is unlikely that the protective effect demonstrated in humans is due to significant PARP inhibition, as the levels of nicotinamide achieved with the doses used are too low. Other effects of the vitamin are more likely, e.g., increase in NAD pool size by de novo synthesis, or inhibition of free radical generation. The drug appears to be safe in the doses employed in humans.