The glycaemic and insulinaemic responses to different carbohydrates vary and these have been suggested to affect performance. The purpose of the present study was to determine the effects of pre-exercise ingestion of glucose (GLU), galactose (GAL) and trehalose (TRE) on metabolic responses at rest and during exercise and on subsequent time-trial (TT) performance. Eight well-trained male cyclists completed three exercise trials separated by at least 3 days. At 45 min before the start of exercise subjects consumed 500 ml of a beverage containing 75 g of either glucose, galactose or trehalose. The exercise trials consisted of 20 min of submaximal steady-state exercise (SS) at 65% of maximal power output immediately followed by a [mean (SEM)] 702 (25) kJ TT. Plasma glucose concentration 15 min postprandial was significantly higher in GLU compared to GAL and TRE ( P<0.05). This was accompanied by a more than twofold greater rise in plasma insulin concentration in GLU compared to GAL and TRE (118% and 145%, respectively). During SS exercise four subjects in GLU and one subject in TRE developed a rebound hypoglycaemia (plasma glucose concentration less than 3.5 mmol.l(-1)). No differences were observed in TT performance between the three trials. Pre-exercise ingestion of trehalose and galactose resulted in lower plasma glucose and insulin responses prior to exercise and reduced the prevalence of rebound hypoglycaemia. Despite the attenuated insulin and glucose responses at rest and during exercise following pre-exercise ingestion of galactose and trehalose, there was no difference in TT performance compared with pre-exercise ingestion of glucose.