Impact of 2025 ASE Update Recommendations for the Evaluation of Left Ventricular Diastolic Function in Patients With...

Left ventricular diastolic dysfunction (LVDD) is highly prevalent in end-stage renal disease (ESRD) but remains difficult to assess accurately. The 2025 American Society of Echocardiography (ASE) update has introduced left atrial reservoir strain (LARS) as a novel parameter for estimating left ventricular filling pressure.

This retrospective study included 109 maintenance hemodialysis (HD) patients who underwent comprehensive echocardiography immediately before and after dialysis, yielding a total of 218 paired examinations for analysis. Left ventricular (LV) diastolic function was evaluated according to both the 2016 ASE/EACVI and 2025 ASE algorithms. Concordance and reclassification between the two algorithms were assessed using Cohen's κ coefficient. Hemodynamic and echocardiographic parameters were compared to characterize volume-related changes.

Hemodialysis significantly reduced LV volumes, cardiac output, ratio of early diastolic mitral inflow velocity to mitral annular velocity (E/e' ratio), and pulmonary artery systolic pressure. These findings indicate effective preload reduction after HD, whereas GLS and LARS also decreased in parallel with the reduction in volume load. Using the 2016 criteria, 32% of patients were classified as indeterminate in the pre-HD cohort. Application of the 2025 ASE algorithm eliminated all indeterminate cases: 29 cases previously labeled as "normal" were reclassified as Grade 1 LVDD, and 47 indeterminate cases were reassigned to normal (n = 13), Grade 1 (n = 20), or Grade 2 (n = 14) LVDD. Overall concordance between the two systems was good (κ = 0.682, p < 0.001). Together, these reclassifications effectively reduced the rate of indeterminate classifications.

The updated 2025 criteria demonstrated improved classification specificity and reduced diagnostic uncertainty compared with the 2016 guidelines, facilitating earlier categorization of diastolic dysfunction and addressing the major limitation of a high rate of indeterminate classifications.