Left Ventricular Diastolic Function: Pathophysiology, Diagnosis, and Prognosis with Echocardiography

Authors, Journal, Affiliations, Type, DOI

• Author: Sherif F. Nagueh, MD
• Journal: JACC: Cardiovascular Imaging 2020;13(1):228–244 (special issue on non-invasive assessment of LV diastolic function); state-of-the-art review
• Affiliation: Methodist DeBakey Heart and Vascular Center, Houston Methodist, Houston, Texas
• Type: State-of-the-art narrative review
• DOI: 10.1016/j.jcmg.2018.10.038

Overview

This review sets out the physiology of diastole — myocardial relaxation, elastic recoil and chamber stiffness — and maps each to the echocardiographic signals used to measure it. It explains why e′ is the most feasible and reproducible index of relaxation, why E/e′ estimates filling pressure only in the presence of myocardial disease, and how the 2016 ASE/EACVI algorithm combines e′, E/e′, TR velocity and LA volume index to diagnose diastolic dysfunction and elevated filling pressure. It also covers diastolic strain rate, LA strain, LV untwisting, special populations (AF, tachycardia, pacing/LBBB, HCM, restrictive cardiomyopathy, mitral stenosis/MR, MAC, LVAD), the evidence base for prognosis, and a detailed rebuttal of studies disputing E/e′. It predates the 2025 ASE guideline revision.

Keywords

Diastole; Doppler; Echocardiography; Filling pressure; Left atrium

Key Takeaways

Physiology of relaxation

• Diastolic function has two main aspects: myocardial relaxation and chamber stiffness, modulated by RV–LV interaction, LA function, pericardial constraint, systolic properties and dyssynchrony.
• At the cellular level, diastole begins with calcium uptake into the sarcoplasmic reticulum (regulated by SERCA2a and phospholamban) and detachment of actin–myosin cross-bridges. In heart failure, SERCA2a expression is reduced and relaxation slows; cAMP-mediated phospholamban phosphorylation removes inhibition and produces positive lusitropy.
• At the chamber level, relaxation is measured by the rate of LV pressure decay during isovolumic relaxation, described by the time constant τ (tau) — the invasive gold standard, but not practical. Increased afterload slows and delays relaxation; preload effects are more debatable.

Echocardiographic indices of relaxation

• e′ (mitral annulus early diastolic velocity) has the highest feasibility, reproducibility and outcome association of the three main signals (the others being global isovolumic and early diastolic strain rate).
• e′ relates inversely to τ and is affected by elastic recoil — smaller end-systolic volume gives faster recoil and higher e′.
• Preload dependence of e′ differs by relaxation state: with normal relaxation, transmitral gradient directly affects e′; with impaired relaxation, LA pressure has minimal effect because e′ is delayed until LV diastolic pressure equals or exceeds LA pressure.
• E/e′ estimates LA pressure because E rises with LA pressure, e′ falls with τ, and LA pressure minimally affects e′ in myocardial disease. Average E/e′ <8 suggests normal LVFP; >14 suggests elevated LVFP.
• Annulus reversus (preserved septal e′, reduced lateral e′) suggests constrictive pericarditis. Reduced e′ also occurs with mitral stenosis and moderate-to-severe MAC; significant MR with normal LVEF can raise e′.
• E/e′ should not be used when e′ is affected by non-relaxation factors: constrictive pericarditis, moderate/severe MAC, mitral stenosis, severe primary MR with normal LVEF, prosthetic mitral valve or surgical ring. In LBBB or paced rhythm, average E/e′ is less reliable — combine with LA volume and PASP.

Strain-rate signals

• SR_IVR (strain rate during isovolumic relaxation) and SR_E (early diastolic strain rate) correlate inversely with τ and −dP/dt; SR_IVR correlates more strongly with invasive relaxation than SR_E.
• E/SR_IVR and E/SR_E relate directly to mean wedge pressure; E/SR_IVR was more accurate than E/e′ in segmental dysfunction and normal EF.
• Limitations: image quality, no standardisation, training burden, lower accuracy with tachycardia (frame rate).

Elastic recoil, titin and untwisting

• Titin is the elastic spring compressed in systole; its isoform ratio determines passive tension. N2BA/N2B is increased in HFrEF (more compliant), whereas N2B predominates in HFpEF (stiffer). PKG phosphorylation makes titin more compliant; PKC phosphorylation and disulfide bonds stiffen it.
• Restoring forces and elastic recoil are normal in HFpEF but reduced in HFrEF; LV untwisting rate is normal in HFpEF at rest but reduced with exercise.
• Caution: peak untwisting rate reflects relaxation only in HFrEF; in normal EF it reflects restoring forces.

Chamber stiffness

• Myocyte stiffness (titin) plus fibrosis (collagen I, cross-links, advanced glycation end-products) determine chamber stiffness; myocyte stiffness is reduced in HFrEF and increased in HFpEF.
• Echocardiographic surrogates: deceleration time (DT) of E shortens with increased operating stiffness; A-wave transit time to the LV outflow tract shortens with increased late-diastolic stiffness and LVEDP.

Diagnosing diastolic dysfunction — the 2016 ASE/EACVI approach

• The most robust assessment combines clinical status, 2D findings and Doppler; diastolic function should not be judged from Doppler alone in patients with normal EF.
• Myocardial disease is inferred from hypertensive disease with LA enlargement or pathological LVH, systolic dysfunction (reduced EF, annular s′, MAPSE, global longitudinal strain), and other structural findings.
• In the absence of myocardial disease, the 2016 guidelines rely on four parameters: annular e′, average E/e′, peak TR velocity, and LA maximum volume index. When the majority (4/4, 3/4, 3/3, 2/3 or 2/2) are normal or abnormal, diastolic function is normal or abnormal; without a clear majority the result is indeterminate. The design deliberately favours high specificity over sensitivity.
• e′ alone is not sufficient — 10–20% of normal people have isolated low e′.
• Once myocardial disease is established, mitral inflow estimates LVFP: E/A ≥2 usually indicates elevated LVFP, E/A ≤0.8 with E ≤50 cm/s indicates normal LVFP; intermediate values require E/e′, LAVi, TR velocity and pulmonary venous S/D. Discordance yields an indeterminate result.
• Additional Doppler findings that independently support elevated LVFP: Ar-A ≥35 ms, mitral inflow change with Valsalva, L-wave ≥50 cm/s, and comparison of mitral with tricuspid inflow.

LA strain and diastolic stress testing

• LA volumes relate directly but weakly to LA pressure; LA reservoir strain correlates inversely with wedge pressure and LVEDP irrespective of EF and tracks diastolic grade. However, it depends on LV systolic function, so it can be normal despite elevated LVFP when longitudinal function is preserved — and it is not validated in AF, MAC or mitral valve disease.
• LA stiffness can be estimated as E/e′ ÷ LA strain; HFpEF patients have reduced LA strain and increased LA stiffness, which relates to PASP and pulmonary vascular resistance.
• A subset of HFpEF patients have normal LVFP at rest but elevated pressures with exercise; diastolic stress echocardiography is therefore needed in symptomatic patients with normal resting pressures.

Special populations (Table 3)

• Atrial fibrillation: high peak E acceleration rate (≥1,900 cm/s²), IVRT ≤65 ms, DT ≤160 ms (with depressed EF), pulmonary venous diastolic DT ≤220 ms, E/Vp ≥1.4, septal E/e′ ≥11.
• Sinus tachycardia: IVRT ≤70 ms is specific (79%); pulmonary venous systolic filling fraction ≤40% is specific (88%); average E/e′ >14 has highest specificity but low sensitivity.
• Paced rhythm/LBBB: average E/e′ >14 but lower accuracy; Ar-A ≥30 ms, TR >2.8 m/s, LAVi >34 mL/m².
• HCM: average E/e′ >14, Ar-A ≥30 ms, TR >2.8 m/s, LAVi >34 mL/m².
• Restrictive cardiomyopathy: DT <140 ms, E/A >2.5, IVRT <50 ms, average E/e′ >14.
• Non-cardiac pulmonary hypertension: lateral E/e′ <8; TR velocity and LAVi should not be used to infer LVFP.
• Mitral stenosis: IVRT <60 ms, mitral A >1.5 m/s, IVRT/TE-e′ <4.2.
• Mitral regurgitation: IVRT <60 ms with depressed EF; IVRT/TE-e′ <5.6; average E/e′ >14 only with depressed EF.
• Moderate-to-severe MAC: E/A <0.8 → normal; >1.8 → elevated; in between, IVRT <80 ms → elevated.
• LVAD: combinations of E/A ratio and RAP/sPAP, LAVi and E/e′ thresholds.

Controversies addressed

• E/e′ in acute decompensated HFrEF: the review argues that a study reporting poor accuracy had methodological problems — wedge pressure measured supine but imaging in lateral decubitus, no wedge-position confirmation, implausibly low PCWP in NYHA III–IV patients, and poor Doppler alignment. Other studies in similar populations support E/e′.
• E/e′ in normal EF/HFpEF: in normal hearts, e′ is directly related to transmitral gradient and no longer corrects for relaxation, so E/e′ does not track LVFP — a real limitation. But studies claiming poor accuracy in HFpEF are small, often include patients with normal wedge pressure, and are contradicted by others.
• The meta-analysis (Sharifov 2016) is criticised for mixing simultaneous and non-simultaneous studies, using different reference standards (LVEDP vs mean wedge vs pre-A pressure), not excluding conditions in which E/e′ is invalid, and including normal subjects in whom there is no range of filling pressure to correlate.
• Doppler vs natriuretic peptides: Doppler-assessed LVFP correlated more strongly with mean wedge pressure than BNP, and E/e′ tracked changes in wedge pressure whereas BNP did not.

Prognosis

• Mitral inflow velocities and DT predict mortality and HF hospitalization in HFrEF; pulmonary venous velocities, Vp, e′, E/e′ and LA volumes also predict events.
• Diastolic strain rate and LA strain predict outcomes after MI and in AF; global longitudinal strain predicts outcome in HFpEF.
• A small RCT showed fewer hospitalizations when HFrEF therapy was guided by echocardiography rather than clinical assessment alone.

Limitations of the document

• Narrative review by a single author who chaired the 2016 ASE/EACVI guideline — the review is partly a defence of that guideline, and the rebuttal of contrary studies (notably the E/e′ meta-analysis) is advocacy as well as appraisal.
• No systematic search or formal evidence grading; studies are selected to illustrate mechanisms and the author's position.
• Predates the 2025 ASE guideline update, which demoted LAVi to a supplemental variable, introduced LA reservoir strain into the primary algorithm, and refined special-population recommendations. Its "four-parameter" algorithm and cut-offs are therefore historical.
• Normal-value and cut-off tables are drawn from heterogeneous sources, and many thresholds are not externally validated across sexes, ages and ethnicities.
• Limited discussion of three-dimensional echocardiography, automated strain and AI, and no practical workflow guidance for busy laboratories.
• The section on diastolic stress testing is brief, with little detail on protocol, contraindications or interpretation thresholds compared with dedicated reviews.
• Several claims rely on small mechanistic studies (e.g. E/SR_IVR accuracy, LA stiffness), which have not been replicated at scale.

Key Concepts Mentioned

• concepts/LV-Diastolic-Function — the concept this review defines and operationalises
• entities/HFpEF — the phenotype in which diastolic assessment matters most
• entities/Heart-Failure — prognosis and therapy guidance
• entities/HCM — special-population filling-pressure assessment
• concepts/Left-Atrial-Strain — LA reservoir, conduit and pump function
• concepts/Left-Atrial-Strain — LA reservoir, conduit and pump strain
• entities/ATTR-Amyloidosis — restrictive physiology

Key Entities Mentioned

• American Society of Echocardiography (ASE) and European Association of Cardiovascular Imaging (EACVI) — 2016 guideline
• Nagueh et al. — 2009 and 2016 guideline documents
• Houston Methodist — author affiliation
• Sharifov et al. — E/e′ meta-analysis critiqued in the review

Wiki Pages Updated

• Created: wiki/sources/lvdiastolic-echo-jaccimaging-2020.md
• Updated: wiki/concepts/LV-Diastolic-Function.md
• Updated: wiki/entities/HFpEF.md
• Updated: wiki/sourceindex.md
• Updated: log.md