Incremental value of global longitudinal strain in the long-term prediction of heart failure among patients with coronary artery disease

Heart failure (HF) remains a common complication for patients with coronary artery disease (CAD), especially after acute myocardial infarction. Although left ventricular ejection fraction (LVEF) is conventionally used to assess cardiac function for risk stratification, it has been shown in other settings to underestimate the risk of HF, compared to global longitudinal strain (GLS). Moreover, most evidence pertains to early onset HF. We sought the clinical and myocardial predictors for late-onset HF in patients with CAD. We analyzed echocardiograms (including GLS) in 334 patients with CAD (age 65±11 years, 77% male), enrolled in NIL-CHF, a prospective, randomized controlled trial that compared standard care with nurse-led intervention to prevent HF in individuals at risk of incident HF. Long-term (9 years) follow-up was obtained via data linkage. Analysis was performed using a competing risks model. Baseline LVEF values were normal or mildly impaired (EF≥40%) in all subjects. After a median of 9 years of follow-up, 50 (15%) of the 334 patients had new HF admissions, and 68 (20%) died. In a competing risk model, HF was associated with GLS (HR=1.15 [1.05-1.25], p=0.001), independent of eGFR (HR=0.98 [0.97-0.99], p=0.045), Charlson co-morbidity score (HR=1.64 [1.25-2.15], p<0.001), E/e' (HR=1.08 [1.02-1.14], p=0.01). GLS - but not conventional echocardiographic measures - added incremental value to a clinical model based on age, gender, and Charlson score (AUC 0.78 to 0.83, p=0.01). GLS was still associated with HF development in patients taking baseline ACEi (HR=1.21 [1.11-1.31], p<0.01) and baseline BB (1.17 (1.09, 1.26) p<0.01). Mortality was associated with older men, risk factors (hypertension or diabetes) and co-morbidities (AF and chronic kidney disease). GLS is independently associated with risk of incident HF in patients admitted with coronary artery disease and provides incremental prognostic value to standard markers. Identifying an at-risk subgroup using GLS may be the focus of future randomized controlled trails to enable targeted therapeutic intervention. [Abstract copyright: Copyright © 2021. Published by Elsevier Inc.