Twenty-eight RCTs (number of participants unclear) were included in the review. Eighteen trials were parallel group and 10 used a cross-over design. Only four trials were double blinded and only four had adequate concealment of treatment allocation.
Type 1 diabetes (18 RCTs)
Insulin aspart resulted in a significant decrease in glycated haemoglobin levels (WMD -0.11%, 95% CI -0.16 to -0.06; Ι²=17.5%; 13 RCTs) compared with regular human insulin. When the results were stratified according to insulin regimen, a significant difference remained for insulin aspart given as continuous subcutaneous insulin infusion (WMD -0.31%, 95% CI -0.55 to -0.08) and as a basal bolus (WMD -0.12%, 95% CI -0.17 to -0.06) but not between mixtures.
Insulin aspart was associated with improvements in post-breakfast glucose (WMD -1.43mmol/L, 95% CI -1.75, -1.11; five RCTs), post-lunch glucose (WMD -1.11mmol/L, -1.61 to -0.61; five RCTs) and post-dinner glucose (WMD -0.97mmol/L, -1.25 to -0.69; six RCTs), but there was no significant difference in fasting glucose levels compared with regular human insulin.
There was an increased risk of hypoglycaemic episode with insulin aspart (RR 1.06, 95% CI 1.01 to 1.10; six RCTs) but no difference in the risk of nocturnal or severe hypoglycaemic episodes.
Patients who received insulin aspart were significantly more satisfied with treatment based on Diabetes Treatment Satisfaction Questionnaire scores (SMD 0.30, 95% CI 0.20 to 0.40; three RCTs) and with treatment flexibility based on the same questionnaire (WMD 0.31, 95% CI 0.15 to 0.47; two RCTs).
Type 2 diabetes (11 RCTs)
There were no significant differences in glycated haemoglobin levels between insulin aspart and regular human insulin groups, although after exclusion of one trial (in which all participants used a continuous glucose monitoring system) a significant improvement was found with insulin aspart (WMD -0.10%, 95% CI -0.19 to -0.02; eight RCTs).
Insulin aspart was associated with improvements in daily mean postprandial glucose (WMD -1.18mmol/L, 95% CI -1.88, -0.47; three RCTs), post-breakfast glucose (WMD -0.83mmol/L, -1.45 to -0.21; 3 RCTs) and post-lunch glucose (WMD -1.32mmol/L, -2.16 to -0.49; one RCTs), but there was no significant difference in fasting glucose levels compared with regular human insulin.
There was no difference in the risk of any type of hypoglycaemic episode.