Aim: Addition and titration of basal insulin is usually effective in improving glycaemic control in type 2 diabetes, but fear of hypoglycaemia remains a barrier. Ability to predict which patients might be at greatest risk of hypoglycaemia might facilitate individualization of treatment and improve safety. The aim of this study was to obtain information about clinical characteristics which might predict risk of hypoglycaemia during initiation of basal insulin.
Methods: Patient-level data from 2251 participants in 11 studies in which insulin glargine was started and titrated using similar treat-to-target methods was pooled and analysed with logistic regression models.
Results: Participants had mean age 58 years, diabetes duration 8.9 years, body mass index 31.0 and baseline A1c 8.8%. They attained mean A1c 7.1% during 6 months of treatment with final mean glargine dosage 0.44 units/kg. Symptomatic hypoglycaemia occurred in 52%, glucose-confirmed hypoglycaemia (blood glucose <50 mg/dl) in 17%, repeated glucose-confirmed events in 7% and severe hypoglycaemia in 1.5%. Independent predictors of glucose-confirmed hypoglycaemia were younger age, lower body mass index, use of a sulphonylurea in addition to metformin, lower attained A1c and lower dosage of glargine.
Conclusions: These findings confirm low rates of clinically important hypoglycaemia using this method, and suggest that higher risk of hypoglycaemia may be suspected when patients needing insulin are younger, less obese and taking metformin and a sulphonylurea, and especially when A1c levels ≤7.0% are attained with glargine dosage ≤0.4 units/kg.
© 2013 Blackwell Publishing Ltd.