World Congress on Clinical Trials in Diabetes 2018

World Congress on Clinical Trials in Diabetes 2018

The World Congress on Clinical Trials in Diabetes 2018 took place in Vienna, Austria, from December 3 to 4. There, Dr Tina Vilsbøll (Steno Diabetes Center, Gentofte, Denmark) emphasized heterogeneity within the glucagon-like peptide-1 (GLP-1) receptor agonists' class-wide trend towards cardioprotection. Informed by an understanding of the different side effect, clinical, and pharmacodynamic profiles of different GLP-1 agonist molecules, Dr Vilsbøll implored the audience, in this age of cardioprotection, not to forget about the impressive metabolic effects of GLP-1 receptor agonists: “We have to keep in mind that reducing HbA1c is important in reducing microvascular complications. Please do not forget it on the cardioprotective battlefield.” However, this does not imply equality of GLP-1 agonists. Dr Vilsbøll pointed towards a recent review article on HbA1c and weight loss data from head-to-head trials of GLP-1 agonists in espousing the fundamental principle that longer-acting molecules have better metabolic effects.1 Overall, semaglutide, liraglutide, dulaglutide, and exenatide-QW (all of which are given once weekly, except once-daily liraglutide) are associated with more substantial HbA1c reductions than lixisenatide and exenatide (both of which are at least once daily). Notably, a similar pattern is also evident for weight loss.

Dr Philip Home (Newcastle University, Newcastle, UK) cast serious doubt on the ability of current statistical methods to adjust for confounding in pharmacoepidemiological studies. He showed how observational studies are hindered by the fact that they compare inherently different populations. In diabetes specifically, he outlined that healthcare providers prescribe different medications at different stages of disease and in different situations based on comorbidities and care settings. In addition, different types of healthcare providers see different cross-sections of the population with diabetes, and different medications are marketed to personal care providers than to various specialists. For example, metformin is avoided in those with heart failure or chronic kidney disease and is very common in early stage diabetes, whereas insulin tends to be prescribed later in the course of disease and is used in sicker patients. These prescribing trends may well drive the observation that people given metformin have better cardiovascular disease and mortality outcomes, whereas those given insulin have worse outcomes.2 Indeed, differences in the way drugs are prescribed are very difficult to control for outside of a randomized clinical trial, and Dr Home argued that study protocol and analysis plans must consider detailed concomitant disease states as well as prescribing physician type and the circumstances. Even changing body mass index or glucose control, which may be associated with a terminal condition, need to be censored. Dr Home drew on a number of published studies to demonstrate that the potential for confusion exists in publications in prestigious journals, not to argue against publication of observational studies, but rather to call for greater acknowledgement in manuscripts that these studies are hypothesis generating rather than conclusive.

In addition, Dr Francesco Giorgino (University of Bari Aldo Moro, Bari, Italy) raised interesting questions about the role of intensive HbA1c reductions in the context of cardiovascular outcomes trials (CVOTs) and expressed concern over wide variations in the design of CVOTs. Dr Giorgino addressed contemporary issues in CVOT conduct and design, focusing on the widespread “shift” from intensive glucose control to the benefits of non-glycemic effects. The previous era of diabetes outcomes trials investigated intensive glucose lowering compared with standard of care, but failed to demonstrate any effect on major cardiovascular events or mortality. In contrast, modern CVOTs of diabetes drugs have demonstrated improvements on cardiovascular outcomes, despite shorter follow-up and more modest glucose reduction than was achieved in studies of intensive glucose lowering. To be sure, CVOTs are not designed to examine glycemic control; indeed, they were designed for glycemic equipoise, to be achieved according to local clinical practice guidelines. As a result, the benefits of new diabetes medications on cardiovascular outcomes at low (or target) HbA1c levels is unknown. Dr Giorgino used this point to describe an oddity in the recent American Diabetes Association (ADA)/European Association for the Study of Diabetes (EASD) consensus document on management of type 2 diabetes (T2D),3 which ties its recommendations on the use of GLP-1 agonists and sodium-glucose cotransporter 2 (SGLT-2) inhibitors to HbA1c levels: “if HbA1c is not at target, add one of these agents according to the algorithm on cardiovascular disease; if HbA1c is at target, consider switching to one of these agents or setting a lower HbA1c target.”