The mitral valve apparatus is a highly sophisticated cardiac structure with complex functions. It comprises six components: atrial wall, annulus, two leaflets, chordae tendineae, papillary muscles, and the left ventricular wall. It functions dynamically, with the mitral annulus—having a three-dimensional morphology—shifting from a rounder, flatter shape during diastole to an enhanced saddle shape during systole, reducing its anteroposterior diameter. This adaptation minimizes hemodynamic stress and optimizes leaflet coaptation. Any failure in the interaction among the mitral annulus, leaflets, chordae, ventricular function, and geometry can lead to coaptation failure, resulting in mitral regurgitation (MR).

Primary mitral regurgitation (MR) is a prevalent valvular disease, with mitral valve prolapse being the most common form, affecting approximately 2-3% of the global population. In populations over 75 years in developed countries, moderate-to-severe mitral prolapse has a prevalence of about 8-10%. Primary MR has several causes, including myxomatous degeneration (Barlow’s syndrome and fibroelastic deficiency), rheumatic valve disease, infective endocarditis, annular calcification, hypertrophic cardiomyopathy, carcinoid syndrome, radiation effects, and medication-related conditions. Secondary forms of MR include dilated cardiomyopathy, either ischemic or non-ischemic, where the valve-ventricle interaction is emphasized. Additionally, secondary MR may arise from annular dilation due to atrial remodeling associated with atrial fibrillation, where annulus-atrial interaction becomes relevant. Other forms of annular disruption are common in Barlow’s disease, primarily affecting the posterior annulus, suggesting that mechanisms leading to MR may combine both primary and secondary causes.

The current study attempts to address the high heterogeneity of primary MR to identify distinct risk profiles. The study aims to classify severe primary MR patients into differentiated phenotypes and relate these to long-term post-surgical prognosis. Researchers reviewed patients who underwent mitral valve surgery between 2006 and 2020 in three South Korean tertiary hospitals. Part of the cohort (n = 1629) was used for model development. Latent class analysis identified subgroups based on 15 prognostic variables described in the literature. The remaining cohort (n = 692) served to validate the model. The primary outcome measured was all-cause mortality following mitral valve surgery. Patients under 18 years, those with any cardiac reintervention, double mitral lesions, concomitant other valvular lesions, infective endocarditis, and secondary MR were excluded.

With a median follow-up of 6 years, 149 patients (9.1%) died in the model-development cohort. Univariate Cox analysis identified age, female sex, atrial fibrillation, left ventricular end-diastolic volume, left ventricular ejection fraction, left atrial size, and tricuspid regurgitation peak velocity as mortality predictors after mitral surgery. Latent class analysis identified five phenotypes; three younger patient groups (groups 1-3) and two older groups (groups 4-5): group 1, low-comorbidity subgroup; group 2, men with left ventricular enlargement; group 3, women with rheumatic disease; group 4, elderly patients with few comorbidities; and group 5, high-risk elderly patients. The 5-year survival rates for groups 1 to 5 were 98.5%, 96%, 91.7%, 95.6%, and 83.5%, respectively (p < 0.001). Group 5 (high-risk elderly) showed the lowest survival, followed by group 3 (women with rheumatic disease). The phenotypes’ predictive value was confirmed using the validation cohort (n = 692) and showed similar predictive performance to the International Mitral Regurgitation Database risk score.

The authors concluded that five distinct phenotypes of patients with severe primary MR were identified. They advocate that grouping patients by phenotypes may improve risk stratification for planning mitral valve surgery.

COMMENTARY:

The study by Kwak et al. uses an innovative statistical approach in cardiac surgery: latent class analysis. More commonly used in social sciences, this analysis identifies groups or clusters by applying a probabilistic model to the data. Unlike traditional clustering, where groups are defined arbitrarily by setting a range or predefined value, latent class analysis first describes the data distribution and then evaluates the probability that cases belong to one latent subgroup or another. Conversely, traditional clustering first identifies groups and then defines the model.

As with any study, this one has limitations. One primary issue is the model’s development using a cohort undergoing surgery. This limits the external validity of these subgroups for patients evaluated in outpatient settings who are not yet surgical candidates, as their subgroup assignment could change due to clinical evolution while awaiting surgery. Additionally, the 14-year time span is long enough for significant changes in clinical practice, surgical techniques, and postoperative care to have occurred. Finally, the latent class analysis model was based on data from a single center (n = 1629), which might not produce identical subgroups if another region’s cohort were analyzed.

In conclusion, primary MR is a heterogeneous entity. Statistical models, such as that described in this article, can help determine risk groups. However, they might be more useful in patients with mitral pathology who are not yet surgical candidates to better define the timing and priority for surgery and study the clinical impact based on the type of MR they present. This approach would enable more personalized care compared to the generic recommendations published in clinical guidelines.

REFERENCE:

Kwak S, Lee SA, Lim J, Yang S, Choi HM, et al. Long-term outcomes in distinct phenogroups of patients with primary mitral regurgitation undergoing valve surgery. Heart. 2023 Jan 27;109(4):305-313. doi: 10.1136/heartjnl-2022-321305.