Oral hypoglycaemic drugs have widely differing pharmacokinetic properties. Possible pharmacodynamic benefits include greater efficacy and fewer adverse effects. In general, it has not been possible to demonstrate unequivocal differences in clinical efficacy between the sulphonylureas during long term use, although there are clear differences in potency. These differences have been emphasised to the extent that the term 'second-generation' has been used for the most potent sulphonylureas, but there is little to suggest that potency is of any therapeutic significance. Trials to study differences in efficacy have rarely been of acceptable design. They have often used fixed doses of drugs, begging the question of whether true potency ratios have been established for chronic treatment. They have rarely involved substantial numbers of patients in double-blind crossover studies with a suitable washout period. Trials which show that there is a clear relationship between drug concentrations in blood and drug effects (whether therapeutic effects or adverse effects such as severe hypoglycaemia) are generally lacking. Qualitative and semiquantitative analysis of adverse effects supports the concept that drugs with a long half-life (e.g. chlorpropamide), renally excreted active metabolites (e.g. acetohexamide) or unusual properties (e.g. glibenclamide, which accumulates progressively in islet tissue) are more likely to cause prolonged hypoglycaemia, which may be fatal. The major adverse effect of treatment with biguanides is lactic acidosis, and this probably occurs more commonly in patients treated with phenformin than those treated with metformin because of pharmacogenetic variation in phenformin metabolism. The available evidence therefore favours the use of drugs with a short elimination half-life which are extensively metabolised and which have no active metabolites.