Fast-acting insulin analog (lispro insulin), as well as slow/fast-acting analog mixtures (NPL-lispro insulin) allow a better control of postprandial glycemia in type 1 and type 2 diabetic patients (35% mean reduction in area under curve for postprandial glycemia, 2 mmol/l reduction in postprandial glucose peak), as compared with regular insulin or conventional mixtures of NPH and regular insulin. Analog pharmacokinetics allow to procede to insulin injection immediately before meal, which is convenient for the patients. When used alone, lispro insulin has no impact on basal glycemia, whereas twice daily injections of NPL-lispro mixtures allow a 1.8 mmol/l mean reduction of basal glycemia, as compared with conventional mixtures, meaning a more specific effect of NPL intermediate insulin. Other premeal blood glucose levels (lunch and dinner) are not improved by lispro insulin. Most studies did not establish a clear reduction in HbA1c with insulin analogs. When this is the case, this reduction averages 1.5% and could be more frequently observed in studies dealing with type 2 diabetic patients. Finally, the reduced incidence of delayed hypoglycemic episodes, which is one of the most attractive effects of insulin analogs, was only reported in a minority of studies. Surprisingly, the reduction of hypoglycemia incidence with analogs was more frequently reported in type 2 diabetics than in type 1 patients. Thus fast-acting insulin analogs feature interesting characteristics, noteworthy immediate premeal injection and a better postprandial glucose control. However, it is established that the determination of diabetes complications (particularly microangiopathy) mostly relies on average glucose control and not solely on postprandial glycemia. Indeed, most studies suggest that insulin analogs are at least as efficient as conventional insulins on HbA1c, but possibly not more.