The sulphonylurea drugs have been the mainstay of oral treatment for patients with diabetes mellitus since they were introduced. In general, they are well tolerated, with a low incidence of adverse effects, although there are some differences between the drugs in the incidence of hypoglycaemia. Over the years, the drugs causing the most problems with hypoglycaemia have been chlorpropamide and glibenclamide (glyburide), although this is a potential problem with all sulphonylureas because of their action on the pancreatic beta cell, stimulating insulin release. Other specific problems have been reported with chlorpropamide that occur only rarely, if at all, with other sulphonylureas. Hyponatraemia secondary to inappropriate antidiuretic hormone activity, and increased flushing following the ingestion of alcohol, have been well described. The progressive beta cell failure with time results in eventual loss of efficacy, as these agents depend on a functioning beta cell and are ineffective in the absence of insulin-producing capacity. Differences in this secondary failure rate have been reported, with chlorpropamide and gliclazide having lower failure rates than glibenclamide or glipizide. The reasons for this are unclear, but the more abnormal pattern of insulin release produced by glibenclamide may be partly responsible and, indeed, may explain the increased risk of hypoglycaemia with this agent. Previously reported increased mortality associated with tolbutamide therapy has not been substantiated, and more recent data have shown no increased mortality from sulphonylurea treatment. Indeed, benefit from glycaemic control, regardless of the agent used--insulin or sulphonylurea--was reported by the United Kingdom Prospective Diabetes Study. Nevertheless, there is still ongoing controversy in view of the experimental evidence, mainly from animal studies, of potential adverse effects on the heart from sulphonylureas, but these are difficult to extrapolate into clinical situations. Most of these studies have been carried out with glibenclamide, which makes comparison of possible risk difficult. Other cardiovascular risk factors may be modified by gliclazide, which seems unique among the sulphonylureas in this respect. Its reported haemobiological and free radical scavenging activity probably resides in the azabicyclo-octyl ring structure in the side chain. Reduced progression or improvement in retinopathy has been reported in comparative trials with other sulphonylureas, and the effect is unrelated to improvements in glycaemia. There are differences between the sulphonylureas in some adverse effects, risk of hypoglycaemia, failure rates and actions on vascular risk factors. As a group of drugs, they are very well tolerated, but differences in overall tolerability can be identified.