Diabetes mellitus ranks among the top ten causes of mortality throughout the world [1,2]. With the rapid advancement of medicine, treatments without side effect for the long-term control of this disorder becomes important. Alternative therapies have recently received attention. Physical exercise in the control of diabetes has been criticised because an increase in glucagon and adrenaline secretion could raise the plasma glucose [3]. Acupuncture has been used in traditional Chinese medicine to relieve symptoms of diabetes mellitus for many years [4].The Zhongwan acupoint is located on the abdominal wall associated with the pancreas [5]. This area contains muscle and adipose tissue, both of which are insulin sensitive tissues. Electroacupuncture (EA) is widely used in traditional Chinese medicine because it combines the effects achieved with traditional needle acupuncture with those of massage [4]. It causes a prolonged contraction of the abdominal muscle fibres and ATP expenditure in addition to an increase in blood circulation resulting in faster glu- Diabetologia (1999) Summary Acupuncture at the Zhongwan acupoint has been widely used in traditional Chinese medicine to relieve symptoms of diabetes mellitus. Our study investigated the effect on plasma glucose of electroacupuncture applied at the Zhongwan acupoint in rat diabetic models. Plasma concentrations of insulin, glucagon and b-endorphin were also determined using radioimmunoassay. A decrease in plasma glucose was observed in rats after electroacupuncture (15 Hz, 10 mA) for 30 min at the Zhongwan acupoint. This was observed in normal rats and rat models with Type II (non-insulin-dependent) diabetes mellitus. No significant effect on plasma glucose was observed in rat models with Type I (insulin-dependent) diabetes mellitus; neither the streptozotocin (STZ)-induced diabetic rats nor the genetic (BB/W) rats. Further, the hypoglycaemic action of electroacupuncture stimulation disappeared in rats with insulin-resistance induced by an injection of human long-acting insulin repeated daily to cause the loss of tolbutamide-induced hypoglycaemia. An insulin-related action can thus be hypothesised. This hypothesis is supported by an increase in plasma insulin-like immunoreactivity after electroacupuncture stimulation in normal rats. Participation of glucagon was ruled out because there was no change in plasma glucagonlike immunoreactivity resulting from electroacupuncture stimulation. In addition to an increase in plasma b-endorphin-like immunoreactivity, the plasma glucose lowering action of electroacupuncture stimulation at Zhongwan acupoint was abolished by naloxone in a sufficient dose to block opioid receptors. Thus we suggest that electroacupuncture stimulation at the Zhongwan acupoint induces secretion of endogenous b-endorphin which reduces plasma glucose concentration in an insulin-dependent manner. [Diabetologia (1999) 42: 250±255]