The American College of Physicians 2018 guidance statement recommended an A1c goal of 7%-8% for patients with type 2 diabetes because of a lack of statistically significant MACE risk reduction through intensive glycemic control in long-term clinical trials (e.g., ACCORDION trial and VADT). However, interpreting clinical trial results merely based on statical inference (p<.05) may prevent a large number of patients from receiving treatments that may be appropriate otherwise. Also, these trials’ point estimates of MACE risk reduction are all in favor of intensive glycemic control, regardless of its statistical significance. This study aimed to use Bayesian analysis to re-interpret the ACCORD and VADT trials. The data from the ACCORD and VADT trials were simulated by the Building, Relating, Assessing, and Validating Outcomes (BRAVO) diabetes model to generate the prior distribution of the MACE risk reductions as a result of intensive glycemic control. From the published distribution of the MACE risk reduction as data input, the Gibbs sampler was used to generate the posteriors distribution of the MACE risk reduction as result of intensive glycemic control. Intensive glycemic control appeared to be 99.9% likely to reduce MACE for T2DM patients in both trials. The Bayesian interpretation could be used to supplement statistical inferences of intensive glycemic control in the ACCORD and VADT trials.
H. Shao: None. S. Yang: None. V. Fonseca: Consultant; Self; Abbott, Asahi Kasei Corporation, AstraZeneca, Bayer Inc., Novo Nordisk Inc., Sanofi. Research Support; Self; Boehringer Ingelheim Pharmaceuticals, Inc. Stock/Shareholder; Self; Amgen, Bravo4health. J. Bian: None. L. Shi: None.
