Objectives: To evaluate the use of insulin glargine in its licensed basal-bolus indication in terms of both clinical and cost-effectiveness.

Data sources: Electronic databases.

Review methods: A systematic review of the literature, involving a range of databases, was performed to identify all papers relating to insulin glargine.

Results: Nineteen studies met the inclusion criteria but full reports were available for only six. For type 1 diabetes patients, insulin glargine appears to be more effective than neutral protamine Hagedorn (NPH) in reducing fasting blood glucose (FBG) but not in reducing glycosylated haemoglobin (HbA1c) and there is some evidence that both insulins are as effective as each other in both FBG and HbA1c control. For type 2 patients for whom oral antidiabetic agents provide inadequate glycaemic control, there is no evidence that insulin glargine is more effective than NPH in reducing either FBG or HbA1c and some evidence that both insulins are as effective as each other in both FBG and HbA1c control. Evidence for control of hypoglycaemia is equivocal. In studies where insulin glargine is demonstrated to be superior to NPH in controlling nocturnal hypoglycaemia, this may be only apparent when compared with once-daily NPH and not twice-daily NPH. Further, this superiority of glargine over NPH in the control of nocturnal hypoglycaemia may relate to one formulation of insulin glargine (HOE901[80]) and not another (HOE901[30]). There is no conclusive evidence that insulin glargine is superior to NPH in controlling symptomatic hypoglycaemia and severe hypoglycaemia. Insufficient data are available to conclude whether insulin glargine is different from each of the commonly used NPH dosing regimens: once daily and more than once daily. Given the lack of a published evidence base for the cost-effectiveness of insulin glargine, the economic review concentrates on a review of the industry submission and an amended model. Three economic models are provided in the submission, two relating to type 1 diabetes and one relating to type 2 diabetes. All three models compare the cost--utility of insulin glargine against NPH insulin. In general, the structures of the models are poor and in all three models, mistakes relating to assumptions and calculations have been made. The assessment team believe that the cost per QALY estimates generated by the Aventis model may be an underestimate for several reasons. The cost-effectiveness of insulin glargine in both type 1 and type 2 diabetes is highly sensitive to the amount of utility associated with reducing the fear of hypoglycaemia.

Conclusions: The evidence suggests that, compared with NPH insulin, insulin glargine is effective in reducing the number of nocturnal hypoglycaemic episodes, especially when compared with once-daily NPH. There appears to be no improvement in long-term glycaemic control and therefore insulin glargine is unlikely to reduce the incidence of the long-term microvascular and cardiovascular complications of diabetes. Further research into insulin glargine is needed that addresses the quality of life issues associated with fear of hypoglycaemia and also the economic impact of balance of HbA1c control and incidence of hypoglycaemia achieved in practice. Studies examining the economic evidence on insulin glargine should be published.