|
Abstract:
|
Safety of antidiabetic drugs had become an important issue, and FDA recommends for CV outcome trials (CVOTs) to show that any new anti-diabetic drugs should not result in undesirable increase of CV risk. Glucagon-like peptide-1 (GLP-1) receptor agonist (RA) is a class of antidiabetic drug that lowers glucose by promoting insulin secretion. There are several drugs in this class are approved by the FDA and available on the market with phase 3 clinical trials: ELIXA, LEADER, and SUSTAIN-6. It is perceivable that leveraging strength from historical trial information in both treatment and control groups and apply to the current data analysis will increase the study power. Several Bayesian approaches exist that would be particularly suitable and will be further investigated. As we increase the power parameter value, we are borrowing more from the historical data, so the hazard ratio becomes closer to the historical ones with a smaller standard error (SE). Most historical borrowing methods give similar results. Overall, not much historical information is borrowed due to the heterogeneity between LEADER, SUSTAIN-6, and ELIXA.
|
