Model-based Meta-analysis of the HbA1c Lowering Effect of PF-04971729, a Sodium Glucose Co-transporter-2 Inhibitor (SGLT2i), In Comparison to Other SGLT2i and Anti-diabetic Agents (ADA)

Author(s): Jaap Mandema, Kevin Sweeney, Steven Terra, Vaishali Sahasrabudhe

PF-04971729 is a potent, selective SGLT2i in development for treatment of type 2 diabetes mellitus (T2DM). Since there is growing recognition of the need for comparative effectiveness of various ADA, a model was developed to quantify time course of dose vs HbA1c response of PF-04971729 relative to other ADA including SGLT2i, DPP4 inhibitors (DPP4i), GLP-1 agonists (GLP1), sulfonylureas (SU), thiazolidinediones (TZD), and metformin. A systematic literature review yielded 153 randomized controlled trials representing >67000 T2DM patients and 21 drugs. PF-04971729 data were obtained from a 12-week, randomized, placebo-controlled study in T2DM patients on metformin background. The model indicated that SGLT2i have the fastest onset time for HbA1c lowering followed by DPP4i, metformin, SU, TZD and GLP1. A significant loss of effect over time was predicted for all drug classes except SGLT2i and TZD. There was no significant difference in maximal effect (Emax) across ADA within a class; however, Emax was dependent on baseline HbA1c and time (Emax = -0.70% [95% CI -0.62 to -0.78] for SGLT2i at 12 weeks at baseline HbA1c of 8%). Figure 2 illustrates model-estimated and observed dose response for various SGLT2i and other classes of ADA. Estimated differences in HbA1c lowering between PF-04971729 25 mg and top doses of other SGLT2i ranged from -0.01 to -0.13%. This analysis offers a quantitative framework to leverage external data and thus enables an indirect comparison of novel ADA with existing treatments.