Fifteen RCTs (n=906 patients) were included in the review. Duration of follow-up ranged from five to 52 weeks. There was adequate allocation concealment, loss to follow-up was substantial (mean=8.5%) and nine trials reported financial ties with pump manufacturers.
Glycaemic control: For patients with type 1 diabetes, there was statistically significant improved glycaemic control (measured by levels of glycosylated haemoglobin HbA1c) in those who received continuous subcutaneous insulin infusion compared to those who received multiple daily injection (WMD:-0.18, 95% CI -0.27 to -0.10; 13 RCTs). However, for patients with type 2 diabetes (two RCTs), there was no statistically significant difference in glycaemic control between patients who received continuous subcutaneous insulin infusion compared with those who received multiple daily injection.
Hypoglycaemia: For patients with type 1 diabetes and for patients with type 2 diabetes, there was no significant difference in the risk of severe hypoglycaemic events or nocturnal hypoglycaemic events for those who received continuous subcutaneous insulin infusion compared with those who received multiple daily injections.
Multiple daily injection was found to significantly improve minor hypoglycaemia in parallel studies (WMD 0.68, 95% CI 0.16 to 1.2; three RCTs; n=89 patients) but in cross-over trials minor hypoglycaemia was unaffected. Subgroup analysis did not find any significant results.