Each row is a cut-off tested in a specific study, against a specific reference standard. Where a study does not report sensitivity and specificity, the row says so and gives the figure it does report instead. The numbers come from PubMed abstracts; where I calculated them myself from the counts, the "Source" column says so.
| Parameter | Cut-off | What it predicts, and against what | Sens. | Spec. | Other figure | Study | Source | Trust | Main limitation | DOI |
|---|
The atrium does not measure the regurgitation: it measures how long, and at what pressure, the atrium has received it. That is why it works poorly as a test of severity and well as a marker of consequences and timing.
Indexed LA ≤ 34 mL/m²
In chronic MR it makes severe MR unlikely: it is an exclusion argument in the ASE 2017 recommendations. If the other signs say severe with a normal atrium, think of acute MR (recent chordal rupture, endocarditis) or recheck the measurements.
Source: ASE 2017 recommendations. No sensitivity or specificity found for this use.
40–59 mL/m²
Compatible with significant chronic MR, but also with AF, diastolic dysfunction, age, athletes. Hall 1997 found no independent correlation between atrial size and EROA or regurgitant volume.
Source: Hall 1997, PubMed abstract.
≥ 60 mL/m²
In 492 patients with organic MR in sinus rhythm, 5-year survival was 53% against 76% expected; mortality 2.8 times that of patients with an atrium < 40 mL/m² on medical therapy. After surgery the excess risk disappeared. Since 2025 it is one of the ESC criteria for class I in asymptomatic patients.
Source: Le Tourneau 2010, PubMed abstract; summary of the ESC/EACTS 2025 guidelines.
Sensitivity and specificity depend on the reference standard. Almost all studies use angiography, expert echocardiographic grading or quantitative Doppler. Uretsky 2015 showed that integrated echo correlates poorly with remodelling after surgery, while CMR correlates well: every row that uses echo as the reference inherits that error.
The samples are small: with 28 or 37 patients a 95% specificity has a very wide confidence interval, which the abstracts do not report.
Many series are mixed, primary and secondary. In secondary MR the orifice is elliptical and the PISA underestimates; the numbers do not transfer automatically to primary MR.
The grading thresholds come from ASE 2017 and EACVI 2022, which I did not re-read in full text for this sheet.