Analysis of glycaemic control and weight change in patients initiated with human or analog insulin in an US ambulatory care setting

Background: Insulin is a mainstay in the treatment of type 1 diabetes and is a recommended option in patients with type 2 diabetes who fail to maintain glycaemic control on other non‐insulin therapies. The purpose of this study was to describe patient characteristics and evaluate changes in glycaemic control and weight in patients treated with insulin in an ambulatory care setting. Methods: Patients with diabetes were identified from the General Electric electronic medical record (EMR) database (1 September 2004 to 30 April 2008). Patients were ≥18 years, insulin naï ve, newly treated with monoinsulin therapy (index date). Baseline characteristics were identified overall and stratified by insulin type (basal, mixed, and rapid). Basal insulins were described by human versus analog and for insulin detemir and insulin glargine. Change in haemoglobin A1C (HbA1C) and weight from baseline (45 days pre‐ to 15 days postindex date) to 6 months (±90 days) were compared. Regression analyses were used to evaluate HbA1C outcomes across insulins and for the likelihood of gaining 0.9 kg (2 lbs) for detemir versus glargine controlling for baseline characteristics. Results: A total of 12 136 patients were included. A majority were initiated on a basal insulin (64.7%) followed by mixed (20.8%) and rapid (14.4%). Basal users had significantly higher mean body weight and lower mean baseline HbA1C than mixed users (p < 0.001 for all), and were significantly older, had higher baseline HbA1C and higher baseline body mass index (BMI) than rapid insulin users (p < 0.001 for all). Glargine patients had a significantly higher mean baseline HbA1C (p = 0.003) than detemir patients. The adjusted reduction in HbA1C was greater for rapid insulin than for mixed or basal insulin (p ≤ 0.05). The adjusted differences in HbA1C between basal human and basal analog insulins and between detemir and glargine were not statistically significant (p > 0.05). Patients using detemir were 30% less likely to gain 0.9 kg or more than glargine users (p < 0.05). Conclusions: HbA1C outcomes in the ambulatory care setting were generally not different between insulin classes. The likelihood of weight gain was less with insulin detemir than with insulin glargine. Thus, real‐world weight outcomes for basal analog insulin may differ by specific product.