The move to intensive insulin therapy following the Diabetes Control and Complications Trial has highlighted the major deficiencies in conventional insulin therapy. Patients are exposed to a high risk of hypoglycaemia due to the delay in absorption of conventional soluble insulin and peaked action of medium acting preparations. Two rapid acting insulin analogues, insulin lispro and aspart have been developed with a reduced tendency to self-association. These demonstrate faster absorption and reach higher concentrations after s.c. injection compared to conventional human insulin and this more physiological action reduces post-prandial glucose to a greater extent. However the benefits in clinical trials have been relatively modest with little or no improvement in HbA1c and minor reductions in hypoglycaemic risk. These rather disappointing results are partly due to the inability of regulatory trials to explore clinical benefit but also because it has taken time to learn how to use these preparations to their best advantage. In patient with tightly controlled diabetes the use of analogues leads to major reductions in severe noctumal hypoglycaemia and this is a robust and important indication. It is still uncertain whether either of the new long-acting insulin analogues, insulin glargine or detemir will have significant clinical benefit but it is possible that the combination of both quick and long-acting analogues will lead to tight glycaemic control without the risk of severe hypoglycaemia.