Main

Type 2 diabetes is an escalating health problem of enormous proportions1. International guidelines highlight the need for more personalized treatment1, but the concept has not yet been systematically examined in randomized trials specifically designed to evaluate treatment responses in patients with different characteristics. This is important to reduce the risk of biases compared with observational studies, meta-analyses or post hoc analyses of previously conducted trials1,2.

The choice of antihyperglycaemic treatment is usually based on comorbidities, baseline cardiovascular risk, side effects, cost and clinical assumptions, but rarely on measurements of pathophysiological features driving the deteriorating metabolic state that ultimately leads to complications2. Furthermore, evaluations of glucose-lowering drugs have mainly been based on average efficacy data, and there is a major gap in our understanding of treatment response heterogeneity1. To address the current knowledge gaps and facilitate the cost-effective use of drugs, the latest international guidelines emphasize the need to investigate treatment efficacy in different subgroups of patients1.

Interestingly, a recent analysis of 9,000 patients with diabetes highlighted five clusters, each with different characteristics and risk of complications3. Two of these clusters are particularly aggressive. One cluster has been coined ‘severe insulin-deficient diabetes’ (SIDD), which features young age at onset, low body mass index (BMI) and poor insulin secretion. The second cluster, termed ‘severe insulin-resistant diabetes’ (SIRD), presents at older age and is associated with high BMI and high insulin resistance. Similar clusters have been reproduced in several multiethnic cohorts4,5,6,7,8,9,10.

This could potentially provide a tool to distinguish individuals with different pathophysiology. However, the clinical relevance of such stratification for predicting treatment response has also been questioned, as it assumes homogeneity within each cluster8,11,12,13,14. An alternative option to stratifying patients into subgroups would be using continuous variables that reflect individual pathophysiology8,15,16,17,18,19. Evaluating the most feasible approaches to predict the individual response to common drugs is critical to guiding future clinical and scientific work in precision medicine.

In this trial, patients with SIDD or SIRD characteristics were randomly assigned to receive semaglutide, a glucagon-like peptide 1 receptor agonist (GLP1ra), or dapagliflozin, a sodium–glucose cotransporter 2 inhibitor (SGLT2i). GLP1ra and SGLT2i drugs are increasingly used and have shown cardiovascular benefits in patients with established cardiovascular or renal disease. However, for most patients with type 2 diabetes, it is currently unclear who benefit most from these drugs. In particular, it is unknown to what extent the glucose-lowering efficacy depends on the pathophysiological characteristics of the patient. The trial represents the first randomized comparison of a GLP1ra and an SGLT2i in stratified subgroups, allowing for side-to-side comparisons of the efficacy of these two drug classes in patients with different pathophysiology. We aimed to address two main questions of clinical and scientific importance: (1) whether knowledge of the SIDD or SIRD subgroup could help inform the decision of adding semaglutide or dapagliflozin to metformin in terms of metabolic benefits and (2) whether continuous pathophysiological measures could be used to identify which patients are likely to benefit most from these drugs in terms of metabolic improvement.

Läs abstrakt och hela artikeln som pdf free
https://www.nature.com/articles/s42255-023-00943-3

Nyhetsinfo
www red DiabetologNytt