New echocardiography guideline cuts uncertainty but misses more elevated filling pressures

A large prospective study has put the newest British Society of Echocardiography (BSE) recommendations for assessing diastolic function to a demanding test: direct comparison against pressures measured inside the beating heart. The multicenter investigation, published in Echo Research & Practice, found that the 2024 guidelines nearly eliminate the frustrating “indeterminate” category that plagued the 2013 version, but they do so at a measurable cost in sensitivity, missing a substantially larger share of patients whose invasive pressures show elevated left ventricular filling. To close that gap, the researchers also proposed a complementary algorithm that, in both a derivation cohort and an independent validation cohort, achieved a more balanced diagnostic profile than either guideline alone.

Assessing left ventricular diastolic function and elevated left ventricular filling pressure (LVFP) is clinically important but remains challenging in practice. Echocardiography is the most widely used non-invasive technique for detecting elevated filling pressures, with invasive hemodynamic assessment serving as the reference standard. The BSE first published recommendations in 2013; although practical, they were criticized for classifying a substantial proportion of patients as indeterminate, leaving clinicians without an actionable answer. In 2024, the society introduced updated recommendations designed to improve feasibility and diagnostic clarity, incorporating global longitudinal strain, left atrial reservoir strain and age-specific tissue Doppler e′ cut-offs alongside conventional indices, with separate algorithms for heart failure patients in sinus rhythm and a dedicated pathway for atrial fibrillation.

The new study is among the first invasive comparative validations of the two guideline generations. Conducted across three university hospitals with five cardiac catheterization laboratories, it screened 809 consecutive patients referred for clinically indicated left heart catheterization between June 2024 and June 2025. Of these, 716 patients in sinus rhythm fulfilled the eligibility criteria and underwent comprehensive transthoracic echocardiography within 120 minutes before catheterization. The cohort was split into a derivation group of 401 patients and a temporally independent validation group of 315, recruited during a subsequent period with complete separation and no patient overlap. Patients with conditions that could interfere with the reliability of the algorithms were excluded, including atrial fibrillation and other significant arrhythmias, moderate-to-severe valvular disease, prosthetic valves, congenital heart disease, poor acoustic windows and hemodynamic instability.

The invasive reference standard deserves attention because LVFP is not a single number. The researchers measured left ventricular end-diastolic pressure (LVEDP), the most frequently used reference, which reflects ventricular pressure at the end of diastole and LV preload, and LV pre-A pressure, measured immediately before atrial contraction and therefore less affected by active atrial filling, making it a closer surrogate for mean left atrial and pulmonary capillary wedge pressures. Using a 6-Fr pigtail catheter and a standardized protocol with measurements taken at end-expiration and averaged over at least three cardiac cycles, the team defined elevated LVEDP as 16 mmHg or greater and elevated LV pre-A pressure as above 15 mmHg. Interventional cardiologists were blinded to the echocardiographic findings, and the sonographers were blinded to the invasive results. Reproducibility was excellent, with intraclass correlation coefficients of 0.90 and 0.91 for LVEDP and LV pre-A respectively, and minimal inter-center bias on Bland–Altman analysis.

The headline result concerns classification feasibility. BSE-2024 reduced the proportion of indeterminate diastolic function classifications from 60.6% under BSE-2013 to just 0.7%, redistributing patients into more balanced normal (49.1%) and impaired (50.1%) categories, compared with 14.2% and 25.2% under the older guideline. A similar pattern emerged for filling pressure classification: BSE-2024 assigned 81.5% of patients to normal and 17.7% to elevated LVFP, versus 26.7% and 12.7% for BSE-2013, primarily by minimizing the indeterminate category. In reclassification terms, 82.3% of patients deemed indeterminate by BSE-2013 were called normal by BSE-2024, and 60.8% of patients classified as having elevated filling pressure under the 2013 guideline were downgraded to normal, while the majority, 89.7%, remained normal under both.

Against the invasive reference standards, however, the picture was more sobering. After excluding indeterminate cases, both recommendations showed comparable but modest discrimination. For detecting elevated LVEDP, the area under the curve was 0.602 for BSE-2024 and 0.630 for BSE-2013, a difference that was not statistically significant (P = 0.459); AUCs for elevated LV pre-A pressure were likewise not significantly different (P = 0.921). BSE-2024 achieved higher specificity for LVEDP (93.5% versus 81.9%) but substantially lower sensitivity (27.8% versus 44.2%), with a similar trade-off for LV pre-A pressure (specificity 92.1% versus 79.4%; sensitivity 38.2% versus 50.8%). Net reclassification improvement and integrated discrimination improvement showed no significant overall gain (NRI −0.056, P = 0.41; IDI −0.056, P = 0.40), indicating that BSE-2024 mainly altered classification without materially changing overall discrimination.

The authors are careful to flag a critical methodological caveat: direct performance comparisons between the two guidelines should be interpreted cautiously. BSE-2013 excluded roughly 60% of patients as indeterminate before its diagnostic metrics were calculated, whereas BSE-2024 retained nearly all patients. This asymmetry may bias the comparison and could overestimate the older guideline’s performance, and it partly explains how the newer recommendations can maintain comparable AUCs while classifying a far larger share of the cohort. The low agreement between the two algorithms also led the researchers to conclude that they should not be considered directly interchangeable despite their similar discrimination values.

To improve on BSE-2024, the team developed an exploratory complementary two-step escalation layer, triggered only when BSE-2024 classified left atrial pressure as normal. Step one applies abnormal cut-offs for tricuspid regurgitation velocity above 2.8 m/s, averaged E/e′ above 14, left atrial volume index above 34 mL/m², and left atrial reservoir strain thresholds that depend on ejection fraction (below 30% if LVEF is at least 50%, below 24% if it is lower). If any parameter is abnormal or missing, step two uses an LASr-anchored adjudication with pulmonary venous tie-breakers: LASr above the threshold confirms normal filling pressure, LASr below 18% indicates elevated pressure, and grey-zone values are adjudicated using the pulmonary venous systolic-to-diastolic ratio and the Ar–A duration difference. In the derivation cohort, this algorithm raised the LVEDP AUC from 0.602 to 0.793, lifted sensitivity from 27.8% to 78.8% with specificity of 79.8%, and for LV pre-A pressure improved sensitivity from 38.2% to 85.5% at 67.3% specificity, while leaving a modest 13.7% of studies indeterminate.

Crucially, these gains held up in the independent validation cohort, where the unmodified algorithm achieved 88.3% sensitivity and 87.9% specificity for elevated LVEDP and 95.7% sensitivity with 64.3% specificity for LV pre-A pressure. Feasibility of the required pulmonary venous and strain measurements was reasonable: in the derivation cohort, the S/D ratio was obtained in 83.3% of patients and Ar–A duration in 80.6%, allowing definitive classification in 86.7%, with comparable figures in the validation cohort. Tricuspid regurgitation velocity was measurable in about 80% of patients. The authors note that pulmonary venous feasibility varies widely in previous reports, from roughly 65% to 95%, which may limit real-world applicability in centers that do not routinely acquire these views.

The study has limitations the authors themselves emphasize. All patients were referred for clinically indicated catheterization, introducing spectrum bias, so the findings may not generalize to unselected outpatient populations. Echocardiography was performed shortly before rather than during catheterization, although hemodynamic variables were comparable between the two time points. Patients with atrial fibrillation were excluded to maintain comparability with BSE-2013, limiting generalizability to that group, and the exploratory composite endpoint combining either elevated LVEDP or LV pre-A pressure requires further investigation. The authors conclude that BSE-2024 substantially improves classification feasibility by nearly eliminating indeterminate cases but at the cost of reduced sensitivity, that the complementary algorithm offers a more balanced profile warranting further external validation before routine clinical implementation, and that future work should invasively validate both approaches in larger independent cohorts.

Subject of Research: Invasive validation of the BSE-2024 and BSE-2013 echocardiography guidelines for diastolic function and left ventricular filling pressure assessment

Article Title: Invasive validation of BSE-2024 vs. BSE-2013 using LVEDP and LV pre-A pressure in patients undergoing cardiac catheterization: a multicenter study with a complementary algorithm

Article References: Barakat, A., Khaddam, A., Alkhatib, M. A., Alsaadi, R., & Alsaadi, A. R. (2026). Invasive validation of BSE-2024 vs. BSE-2013 using LVEDP and LV pre-A pressure in patients undergoing cardiac catheterization: a multicenter study with a complementary algorithm. Echo Research & Practice, 13(1), Article 27. https://doi.org/10.1186/s44156-026-00132-4

Image Credits: AI Generated

DOI: 10.1186/s44156-026-00132-4

Keywords: echocardiography, diastolic function, left ventricular filling pressure, BSE-2024 guidelines, BSE-2013 guidelines, LVEDP, LV pre-A pressure, cardiac catheterization, left atrial reservoir strain, diagnostic accuracy, validation study, complementary algorithm