Usefulness of estimated glucose disposal rate in detecting heart failure: results from national health and nutrition examination survey 1999-2018.

Abstract

Background: Estimated glucose disposal rate (eGDR) is a novel, clinically available, and cost-effective surrogate of insulin resistance. The current study aimed to assess the association between eGDR and prevalent heart failure (HF), and further evaluate the value of eGDR in detecting prevalent HF in a general population.

Methods: 25,450 subjects from the National Health and Nutrition Examination Survey 1999-2018 were included. HF was recorded according to the subjects' reports. Logistic regression was employed to analyze the association between eGDR and HF, the results were summarized as Per standard deviation (SD) change. Then, subgroup analysis tested whether the main result from logistic regression was robust in several conventional subpopulations. Finally, receiver-operating characteristic curve (ROC) and reclassification analysis were utilized to evaluate the potential value of eGDR in improving the detection of prevalent HF.

Results: The prevalence of reported HF was 2.96% (753 subjects). After adjusting demographic, laboratory, anthropometric, and medical history data, each SD increment of eGDR could result in a 43.3% (P < 0.001) risk reduction for prevalent HF. In the quartile analysis, the top quartile had a 31.1% (P < 0.001) risk of prevalent HF compared to the bottom quartile in the full model. Smooth curve fitting demonstrated that the association was linear in the whole range of eGDR (P for non-linearity = 0.313). Subgroup analysis revealed that the association was robust in age, sex, race, diabetes, and hypertension subgroups (All P for interaction > 0.05). Additionally, ROC analysis displayed a significant improvement in the detection of prevalent HF (0.869 vs. 0.873, P = 0.008); reclassification analysis also confirmed the improvement from eGDR (All P < 0.001).

Conclusion: Our study indicates that eGDR, a costless surrogate of insulin resistance, may have a linear and robust association with the prevalent HF. Furthermore, our findings implicate the potential value of eGDR in refining the detection of prevalent HF in the general population.