A computational phenomapping approach to personalize the cardiometabolic benefits of novel glucagon-like peptide-1 receptor and glucose-dependent insulinotropic polypeptide agonists in patients with type 2 diabetes mellitus

According to the World Health Organization (WHO) 422 million people worldwide have diabetes, which is a direct or indirect cause of 1.5 million deaths annually. The most common type of diabetes is type 2 diabetes mellitus, where the blood sugar (glucose) levels are too high due to the body not making enough of a hormone called insulin, or not using it properly. Agonists (drugs or substances that bind to a receptor inside a cell or on its surface and causes the same action as the substance that normally binds to the receptor) of glucagon-like peptide-1 (GLP-1) receptor and glucose-dependent insulinotropic polypeptide (GIP) are a new group of glucose-lowering medications used to treat diabetes, which have been shown to improve several aspects of cardiovascular (heart and blood system) health and prevent complications related to diabetes. A series of randomized clinical trials (RCTs) have demonstrated the efficacy and safety profile of tirzepatide, a novel dual GIP and GLP-1 agonist, across a range of type 2 diabetes mellitus patient profiles highlighting marked effects on several metrics of glucose control and weight. However, the cardiometabolic benefit (i.e., the benefit relating to the chemical processes affecting the cardiovascular system and metabolic health) may vary across different patient populations and for different parameters. For instance, in a clinical trial of patients with obesity, tirzepatide provided substantial and sustained reductions in body weight when compared to placebo (sugar pill). On the other hand, a recent analysis of 7,215 patient data from seven RCTs with at least 26 weeks of treatment demonstrated no significant difference between patients receiving tirzepatide and those receiving placebo when assessing events such as a non-fatal heart attack, non-fatal stroke, death due to heart failure or hospitalization for unstable angina (chest pain caused by reduced blood flow to the heart muscle). Given the known heterogeneity (difference) of patients with type 2 diabetes mellitus, understanding the links between unique sets of characteristics (signature phenotypes) and the predicted benefit of blood sugar levels and weight loss response is important to better inform treatment practices to ensure that patients receive individualized treatment programs that are safe and efficacious. We hereby propose a study with the aim to provide personalized estimates of treatment efficacy and safety for GLP-1/GIP agonists. Our approach is based on our machine learning method, called 'computational trial phenomapping', that has been previously employed across a range of cardiometabolic trials and has detected heterogeneous treatment effects, including among patients with type 2 diabetes treated with newer antihyperglycemic therapies (glucose-lowering medication), such as sodium-glucose cotransporter-2 (SGLT2) inhibitors, a class of drugs that lower blood sugar levels by preventing the kidneys from reabsorbing sugar that is created by the body.