We have previously demonstrated that experimental hypoglycaemia in adults with type 1 diabetes causes an abnormal electrocardiogram (ECG), with increases in QT interval and dispersion. These abnormalities in cardiac repolarisation indicate a risk of ventricular tachycardia and sudden death in other conditions, including ischemic heart disease and congenital long QT syndrome. We have hypothesised that they could contribute to the dead in bed syndrome--the recently described sudden unexpected death in young people with type 1 diabetes--which occurs around three times more frequently than in those without diabetes. It is clearly impossible to explore the causes of a rare and fatal complication by direct observation. We have therefore explored the pathophysiology in a series of experimental studies involving non-diabetic subjects and surrogate endpoints. These have demonstrated that abnormal cardiac repolarisation occurs consistently during insulin-induced hypoglycaemia and that either potassium infusion or beta-blockade prevents increased QT dispersion but only partially prevents QT lengthening. The sympathoadrenal discharge induced by hypoglycaemia alters cardiac repolarisation by both direct and indirect (by reducing extracellular potassium) mechanisms. Other factors that might contribute to the clinical risk of cardiac arrhythmias during nocturnal hypoglycaemia include autonomic neuropathy. This is associated with prolonged QT interval in the non-hypoglycaemic state and has been proposed as a cause of sudden death in those affected. We have examined cardiac repolarisation during clamped hypoglycaemia in patients with type 1 diabetes, with and without autonomic neuropathy. Our data demonstrate lengthening of QTc (QT interval corrected for heart rate) during hypoglycaemia in all groups with no significant differences between the groups, suggesting that autonomic dysfunction does not contribute to hypoglycaemia-induced QTc lengthening in type 1 diabetes. Our hypothesis would be strengthened by demonstrating similar changes during clinical hypoglycaemia. We have recently completed studies in prepubescent children and adults that show modest but significant changes in QTc during nocturnal hypoglycaemia in both populations. We have also demonstrated that pre-treatment with beta-blocking agents prevents abnormal cardiac repolarisation during experimental hypoglycaemia. This has identified a possible treatment if we can identify patients at high risk. Further work is necessary to determine whether we can reliably identify patients who could be at special risk during hypoglycaemia and who might benefit from protection with agents such as beta-blockers. Sudden death in young people with diabetes is, thankfully, rare. However its consequences are so devastating that an excess risk of 3 to 4 times the non-diabetic population seems sufficient to warrant further investigation of the mechanisms that may cause it.