Ten RCTs (nine parallel and one cross over design) were included in the review (3,825 patients; 3,048 adults and 777 children). In four trials allocation concealment was unclear, and in one trial randomisation was unclear. There was a lack of blinding in all of the studies.
HbA1c values (10 RCTs): The meta-analysis showed a significant reduction of HbA1c levels in detemir insulin group compared with NPH insulin (WMD -0.073, 95% CI -0.135 to -0.011, p= 0.021; Ι² = 0%). Statistically significant effects that favoured detemir insulin were also found in adults, the population of patients with HbA1c above 8% and duration of intervention six months or less.
Fasting Plasma Glucose (10 RCTs): The meta-analysis showed a significant reduction in the detemir group compared with the NPH group; with significant heterogeneity (WMD -0.977mmol/l, 95% CI -1.395 to -0.558, p<0.001; Ι²=66.5%).
Hypoglycaemic episodes (eight RCTs): The meta-analysis showed that detemir group had a significant reduction of the number of patients with all-day hypoglycaemic episodes (RR 0.978, 95% CI 0.961 to 0.996,p=0.016), nocturnal hypoglycaemic episodes (RR 0.877, 95% CI 0.816 to 0.942, p<0,001) and severe hypoglycaemic episodes (RR 0.665, 95% CI 0.547 to 0.810, p<0.001). There was no statistically significant difference in severe nocturnal hypoglycaemic episodes between the two groups.
Body weight (six RCTs): Pooled results showed significant reduction in body weight gain in the detemir group compared with the NPH group (WMD -0.77kg, 95% CI -0.992 to -0.567, p<0.001).