Prognostic superiority of global longitudinal strain beyond four-tiered ventricular hypertrophy in asymptomatic individuals

BACKGROUND: This study aims to explore the clinical correlates of myocardial deformations using speckle-tracking algorithm and to determine the prognostic utility of such measures in asymptomatic ethnic Chinese population. METHODS: Global longitudinal (GLS), circumferential strain (GCS), and torsion were analyzed using featured tissue-tracking algorithm among 4049 symptom-free ethnic Chinese population. Hypertrophy (LVH) was classified into 4 tiers: indeterminate, dilated, thick and thick/dilated, by gender-stratified partition of end-diastolic volume index (EDVi) and LV mass/EDV0.67. RESULTS: LVH (7.3%) showed substantially lower GLS (-20.3 ± 1.82% vs. -18.9 ± 2.08%) yet higher torsion (2.20 ± 0.90 vs. 2.39 ± 1.01, p < 0.001) than non-LVH participants. Those with thick LVH (n = 123) were more obese, had higher blood pressure and increased high-sensitivity C-reactive protein (hs-CRP); with dilated/thick LVH (n = 26) group demonstrating highest pro-brain natriuretic peptide (NT-proBNP) and worse GLS compared to indeterminate-/non-LVH groups. There were independent associations among larger EDVi, higher NT-proBNP and decreased torsion, and among greater LV mass/EDV0.67, worse GLS, greater GCS/torsion and hs-CRP. Over a median of 2.3 years (IQR: 1.2-4.8), the dilated, thick, and dilated/thick LVH categorizations were associated with higher risk of composite all-cause death and heart failure (HF) compared to non-LVH (adjusted hazard ratio [HR]: 3.65, 3.72, 6.01, respectively, all p < 0.05). Per 1% GLS reduction was independently associated with higher risk (adjusted HR: 1.31, p < 0.001) and improved risk prediction (p ≤ 0.001 by integrated discrimination improvement [IDI]: 3.5%, 95% CI: 1.5%-5.6%, and continuous net reclassification improvement [NRI]: 42.3%, 95% CI: 24.0%-60.6%) over LVH. CONCLUSION: GLS improved risk stratification of four-tiered classification of LVH in asymptomatic ethnic Chinese.