Cardiovascular magnetic resonance imaging markers of ageing: a multi-centre, cross-sectional cohort study

TL;DR

This study uses automated cardiac MRI to track age-related heart changes and predict functional heart age in healthy and unhealthy aging.

Contribution

A fully automated CMR model to predict functional heart age, validated across healthy and diseased populations.

Findings

CMR-derived functional heart age matched chronological age in healthy individuals.

Functional heart age was significantly higher than chronological age in patients with hypertension, diabetes, and atrial fibrillation.

The model showed higher functional heart age in obesity classes II and III.

Abstract

Cardiac ageing involves a series of anatomical and physiological changes contributing to a decline in overall performance. Cardiac magnetic resonance (CMR) provides comprehensive structural and functional assessment for detecting age-related cardiovascular remodelling. We aimed to develop a fully automated CMR model to predict functional heart age.

This international, multi-centre, retrospective observational study enrolled 191 healthy individuals with normal body mass index (BMI), free of metabolic, cardiovascular, and respiratory disease as the derivation cohort. Left atrial (LA) end-systolic volume and LA ejection fraction were selected for the final model. The model was validated on 366 patients with BMI >25 kg/m2 and one or more comorbidities [hypertension, diabetes mellitus (DM), atrial fibrillation (AF), and obesity]. In healthy individuals [median age: 34 years, 105 (55%) female], CMR-derived functional heart age was similar to the chronological age [bias: 0.05%, 95% confidence interval (CI): 9.56–9.67%, P = 0.993]. In the validation cohort [median age: 53 years, 157 (43%) female], CMR-derived functional heart age was 4.6 years higher than chronological age (95% CI: 1.6–7.6 years, P = 0.003). Cardiac magnetic resonance-derived functional heart age was significantly higher in patients with hypertension (P < 0.001), DM (P < 0.001), and AF (P < 0.001) than age-matched healthy controls. Moreover, CMR-derived functional heart age was higher than the chronological age in obesity Class I (P = 0.07), obesity Class II (P = 0.11), and obesity Class III (P < 0.001).

This study highlights the time course of structural and physiological changes in the heart during healthy and unhealthy ageing. We propose simple equations that should help communicate subtle changes in heart assessment with ageing.

ClinicalTrials.gov: NCT05114785

Graphical Abstract

Figures