Answers are grounded in this paper and cite their sources.
Insulin resistance drives cardiometabolic disease, yet its molecular signatures and tissue origins remain incompletely characterized. Here we define a circulating protein signature that enables scalable assessment of insulin sensitivity. We applied multiworkflow proteomics to plasma from 161 individuals spanning the insulin sensitivity spectrum defined by the hyperinsulinaemic–euglycaemic clamp and identified 488 proteins associated with insulin sensitivity, capturing contributions from liver, adipose tissue and immune cells. A subset of these proteins shifted towards a more insulin-sensitive profile following an exercise intervention in men. We further developed a model combining 13 proteins, including IGFBP1, LEP, GDF15, PON3 and LDLR, with clinical variables (sex, HbA1c and the triglyceride-to-high-density lipoprotein cholesterol ratio) to predict insulin sensitivity ( R 2 = 0.67). In ~20,000 UK Biobank participants, the model outperformed the triglyceride-to-high-density lipoprotein cholesterol ratio in predicting type 2 diabetes (concordance index, 0.86 versus 0.74) and, unlike HbA1c, predicted hepatic, renal, cardiovascular and respiratory disease. In an independent cohort of 1,700 individuals, the model also showed greater predictive utility than fasting insulin resistance indices, including the homeostatic model assessment of insulin resistance. Together, these findings establish a tissue-resolved plasma proteomic framework for characterizing insulin resistance and provide a scalable approach to precision cardiometabolic risk stratification.
No takes yet. Share an insight, caveat, or question.
Northcote* et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: