Abstract Introduction Cardiometabolic dysregulation, particularly type 2 diabetes mellitus (T2DM), is associated with subtle myocardial pathologies, even in individuals without known cardiovascular disease. These subclinical changes, often undetectable by standard electrocardiogram (ECG) analytics, can offer valuable insights into the relationship between metabolic health and cardiac function. Novel ECG markers, developed through machine learning and artificial intelligence (AI), may provide the sensitivity needed to detect such changes. Aim This study aims to evaluate novel composite cardiac parameters from an analytics platform in the context of heart failure (HF), seeking to understand their relationship to cardiometabolic dysregulation. Methods Data from the MyoVasc study, a prospective cohort on HF, were analysed. Participants underwent extensive clinical phenotyping, including a Holter ECG assessment. Holter ECG data was processed based on machine learning and AI models on a cloud-based analytics platform (Cardiolyse, Helsinki, Finland), generating both traditional and proprietary ECG, and heart rate variability cardiac health markers. LASSO-penalized logistic regression adjusted for sex and age was performed to select the marker most strongly associated with T2DM. The identified marker was then evaluated in relation to cardiac function and structure as well as HF-related outcome. Results A total of 953 subjects (mean±SD age 64.6±10.5 years; 35.7% women) were included in the analyses. Symptomatic HF stage C/D was present in 54.8% (n=522) of subjects and diabetes mellitus in 22.4% (n=213). The Condition of Myocardium Reserves Score (CMRS) was the novel composite cardiac parameter selected based on its relationship with T2DM. In a Poisson regression model with robust variance adjusted for sex and age, a lower score was associated with T2DM (Prevalence ratio per SD 1.24, 95% confidence interval [1.14;1.36], P=0.0002). In multivariable linear regressions adjusted for traditional cardiovascular risk factors, comorbidities, and medication, a higher CMRS was associated with better global longitudinal strain (β per SD –0.82 [–1.11;–0.53], P<0.0001), diastolic function (left ventricular E/e’; β per SD –0.04 [–0.06;–0.01], P=0.0059) and systolic function (LV ejection fraction; β per SD 2.06 [1.46;2.67], P<0.0001). A lower CMRS independently predicted all-cause death (Hazard ratio per SD 1.25 [1.09;1.45], P=0.0020) and worsening of HF (HR per SD 1.44 [1.21;1.72], P<0.0001) in Cox regression analysis adjusted for the same confounders. Conclusion The Condition of Myocardium Reserves Score exhibited a strong association with cardiometabolic dysregulation. A higher CMRS indicated preserved cardiac function, while a lower score independently predicted increased risk of all-cause death and worsening HF. This novel score offers potential as a non-invasive tool for assessing cardiometabolic health and stratifying risk in patients with HF.
Read full abstract