AIMS: We investigated whether differences in duration of first insulin use in type 2 diabetes remain after adjustment for potential confounders, and what factors are associated with longer use. METHODS: People prescribed a first insulin (2000-2007) after 2-3 non-insulin glucose lowering treatments (OGLD) were identified from the THIN UK primary care database and grouped by insulin, detemir (n=165), glargine (n=1011) or NPH (n=420). Time from beginning insulin to the prescription of another insulin type or a glucagon-like peptide was compared between insulins in a Cox model adjusting for: demographics, HbA1c, history of vascular complications and cardiovascular risk factors. The strength of association between duration of use and these variables was investigated. RESULTS: The adjusted hazard ratios compared to glargine for treatment change were 1.58 (95% CI 1.25, 2.00) for detemir and 1.49 (1.25, 1.78) for NPH. Lower mean treatment HbA(1c) correlated with longer time to a different insulin regimen (Spearman rank correlation -0.30, p<0.01) as were continuing OGLDs, older age, longer time from diagnosis, lower body mass index, lower HbA(1c), and no heart failure at baseline. CONCLUSIONS: People who began treatment with glargine and those with better on-treatment HbA(1c) remained on their first insulin for longer than those who began detemir or NPH.
AIMS: We investigated whether differences in duration of first insulin use in type 2 diabetes remain after adjustment for potential confounders, and what factors are associated with longer use. METHODS:People prescribed a first insulin (2000-2007) after 2-3 non-insulinglucose lowering treatments (OGLD) were identified from the THIN UK primary care database and grouped by insulin, detemir (n=165), glargine (n=1011) or NPH (n=420). Time from beginning insulin to the prescription of another insulin type or a glucagon-like peptide was compared between insulins in a Cox model adjusting for: demographics, HbA1c, history of vascular complications and cardiovascular risk factors. The strength of association between duration of use and these variables was investigated. RESULTS: The adjusted hazard ratios compared to glargine for treatment change were 1.58 (95% CI 1.25, 2.00) for detemir and 1.49 (1.25, 1.78) for NPH. Lower mean treatment HbA(1c) correlated with longer time to a different insulin regimen (Spearman rank correlation -0.30, p<0.01) as were continuing OGLDs, older age, longer time from diagnosis, lower body mass index, lower HbA(1c), and no heart failure at baseline. CONCLUSIONS:People who began treatment with glargine and those with better on-treatment HbA(1c) remained on their first insulin for longer than those who began detemir or NPH.
Authors: James H Flory; Dylan S Small; Patricia A Cassano; David J Brillon; Alvin I Mushlin; Sean Hennessy Journal: J Comp Eff Res Date: 2014-01 Impact factor: 1.744