Aims: We compared the effects of continuous subcutaneous insulin infusion (CSII) with those of multiple daily insulin (MDI) injections on glycaemic control, risk of hypoglycaemic episodes, insulin requirements and adverse events in type 1 and type 2 diabetes mellitus.
Methods: The electronic databases MEDLINE, EMBASE and CENTRAL were systematically searched for randomised controlled trials up to March 2007. A systematic review and meta-analysis were performed.
Results: Overall, 22 studies were included (17 on type 1 diabetes mellitus, two on type 2 diabetes mellitus, three on children). With regard to adults with type 1 diabetes mellitus, our meta-analysis found a between-treatment difference of -0.4% HbA(1c) (six studies) in favour of CSII therapy. Available median rates of mild or overall hypoglycaemic events were comparable between the different interventions (1.9 [0.9-3.1] [CSII] vs 1.7 [1.1-3.3] [MDI] events per patient per week). Total daily insulin requirements were lower with CSII than with MDI therapy. In patients with type 2 diabetes mellitus, CSII and MDI treatment showed no statistically significant difference for HbA(1c). The incidence of mild hypoglycaemic events was comparable between the treatment groups. In adolescents with type 1 diabetes mellitus, glycated haemoglobin and insulin requirements were significantly lower in the CSII groups; no data were available on hypoglycaemic events. The only study performed in younger children did not provide enough data for conclusive inferences. No overall conclusions were possible for severe hypoglycaemia and adverse events for any of the different patient groups due to rareness of such events, different definitions and insufficient reporting.
Conclusions/interpretation: CSII therapy in adults and adolescents with type 1 diabetes mellitus resulted in a greater reduction of glycated haemoglobin, in adult patients without a higher rate of hypoglycaemia. No beneficial effect of CSII therapy could be detected for patients with type 2 diabetes mellitus.