Introduction: ABCD conducted two nationwide audits on GLP-1 agonist use in real life clinical practice - exenatide from 2007-2009 and liraglutide from 2009-2010. We compared treatment outcomes and differences in diabetes treatment in these two audits.

Methods: Patients in the liraglutide audit who were previously on exenatide were excluded from the analyses. Pairwise comparisons of baseline and 6 month (taken at ±6 weeks) HbA1c and weight were assessed using paired t-tests. Differences in HbA1c and weight changes, baseline characteristics, rates of gastrointestinal side effects, insulin dose changes, and oral diabetes treatment changes (reduced/unchanged/increased) were compared using Student’s t-test and Chi Square tests. Significant group differences were entered into stepwise multivariate analyses of HbA1c change and weight change with an alpha to remove of 0.15.

Results: There were 6717 patients from 126 centres, and 2303 patients from 65 centres in the exenatide and liraglutide audits, respectively. Combined baseline characteristics were 54.7% male, 91.5% Caucasian, mean (±SD) age 55.0±10.8 years, HbA1c 9.43±1.70%, BMI 39.6±7.8 kg/m² and total insulin dose 1.0±0.9 U/kg/day. Patients on liraglutide had statistically higher baseline age, lower baseline HbA1c and weight, were more likely Caucasian and on insulin, and less likely to develop gastrointestinal side effects. They also had lower insulin dose reduction, less sulphonylurea and thiazolidinedione reduction, but more dipeptidyl peptidase-IV inhibitor reduction. There were 2866 and 2790 data for HbA1c and weight changes respectively. All HbA1c and weight changes from baseline were significant (p<0.001) for both agents. HbA1c reduction (exenatide v liraglutide) were 0.75±0.04 v 0.93±0.07% (difference, p=0.040) and weight reduction 6.5±0.1 v 3.7±0.2 kg (difference, p<0.001). In stepwise regression analyses, the differences in HbA1c and weight reduction between the two agents remained significant after adjustment for the differences in baseline characteristics and treatment changes between the two groups.

Conclusions: Within the limitations of a multivariate analysis and differences in diabetes practice over a 1-2 year period, treatment with liraglutide in real life clinical practice appear to be associated with greater HbA1c reduction but lesser weight reduction than treatment with exenatide.