The integration of patient preferences into quantitative risk-benefit assessment of sodium glucose co-transporter-2 inhibitors for the management of type 2 diabetes

Background: The Sodium Glucose Co-Transporter-2 (SLGT2) Inhibitors are the newest class of antihyperglycemic medications available on the market. These agents have gained quick popularity due to demonstrated cardiovascular benefits among patients with pre-existing cardiovascular disease. While we have estimates for the probabilities of benefits and harms for SGLT2 inhibitors, the overall balance of risk and benefits that reflects the values of patients is unknown. Objectives and Methods: 1. To conduct a systematic review and meta-analysis of the current state of knowledge surrounding post-market safety concerns of the SGLT2 inhibitors, including acute kidney injury (AKI), diabetic ketoacidosis (DKA), urinary tract infection (UTI), bone fracture and amputation, in patients with type 2 diabetes. 2. To estimate the strength of preferences, relative importance, and trade-offs that Canadians with type 2 diabetes make between characteristics of glucose-lowering medications using a discrete choice experiment (DCE). 3. To bring together Canadian patient preferences for attributes of diabetes therapies with probabilities of efficacy and safety retrieved from the literature, to compare the SGLT2 inhibitors and glucagon-like peptide-1 (GLP1) receptor agonists, using quantitative benefit-risk assessment (BRA) following the incremental net benefit (INB) framework. Results: The analysis of the adverse outcomes of the SGLT2 inhibitors, suggested no significant increase in the risk of AKI, DKA, UTI (exception: high dose dapagliflozin) or bone fractures. Amputation was poorly reported, however CANVAS trials do show an increased risk. The DCE showed that all eight examined attributes for diabetes medications, including cost, risk of macrovascular and microvascular events, risk of minor side effects, severe hypoglycemia, serious long term consequences, and life expectancy were each shown to significantly influence choice. Life expectancy and cost were more important to patients. Finally, the BRA demonstrated that there was a minimal difference in INB between the SGLT2 inhibitors and the GLP1 receptor agonists, but favored the SGLT2 inhibitors (INB = 0.2) and results were consistent in sensitivity analysis. Conclusion: This program of research used emerging methods, including a network metaanalysis, a DCE, and preference-weighted BRA to examine the balance between risks and benefits of the SGLT2 inhibitors and GLP1 agonists. These studies resulted in a final INB that favored SGLT2 inhibitors, though magnitude was small. More importantly, this research identifies several challenges and limitations, including gaps in methodological guidance that still exist to successfully integrate patient preferences into BRA.