The growing interest in tricuspid regurgitation (TR) treatment from a percutaneous approach represents the natural progression of edge-to-edge therapy as implemented in mitral clipping. Unlike the mitral valve, the tricuspid valve pathology predominantly presents as functional insufficiency, primarily classified under Carpentier’s type I and IIIB mechanisms. These correspond to isolated and primary forms, respectively, in the recently published reclassification, which we reviewed in a prior blog post. The prevalence of TR is considerable in the general population, contradicting the long-held assumption in clinical cardiology that TR is a benign condition suitable for medical treatment only, with invasive intervention reserved for cases of refractory and advanced-stage heart failure (HF), as discussed in previous blog entries.

The surgical community was prompted by Dreyfus’s work to move tricuspid valve treatment forward to scenarios with annular dilation (septo-lateral diameter >40 mm echographically, and antero-posterior >70 mm intraoperatively), independent of TR grade, preferably with semi-rigid prosthetic rings. These findings have reshaped the perspective that tricuspid valve improvement could follow left-sided valve repair, positioning early TR intervention as an opportunity to enhance the prognosis in surgically treated HF patients. However, a significant subset of patients with isolated tricuspid disease, whether primary or secondary to left-sided valvular disease repair (rheumatic disease, irreversible pulmonary hypertension, atrial fibrillation), remains under conservative care with pharmacological treatment, primarily diuretics, lacking alternative treatment options.

In this context, the TRI-SCORE scoring system emerged from an initiative to analyze data from the TRIGISTRY registry, which collected clinical and echographic variables in patients with significant isolated TR. This was an altruistic, non-sponsored project involving 33 centers across 10 countries, including those from Europe (Spain included), the US, Canada, and Israel. Data from a total of 2,413 patients were analyzed over a two-year survival period, of which 1,217 received conservative management, 551 isolated tricuspid valve surgery (excluding any concomitant procedures, and incorporating both repair and replacement techniques), and 645 percutaneous treatment (without concomitant procedures and exclusively using edge-to-edge techniques, excluding other device implants).

Without delving into methodological aspects (which we will discuss below), invasive therapy, both surgical and percutaneous, demonstrated a survival benefit over conservative treatment in patients with low TRI-SCORE scores (≤3 points): 93% surgery, 87% percutaneous, and 79% conservative; p = .0002. There were no significant differences among groups for intermediate (4-5 points) or high (≥6) scores. Only 3% of the surgical group failed to achieve successful TR correction (≥2+ residual grade) and were thus excluded from a subanalysis on successful repair. However, in the percutaneous group, there was a benefit over conservative treatment for intermediate scores when adequate repair was achieved (two-year survival 81% vs. 71%, p = .009). The high-score category continued to show no therapeutic benefit from invasive treatment over conservative, even with successful repair.

The authors conclude that the TRI-SCORE system shows good alignment with isolated TR survival rates and that, compared to conservative management, early correction with surgical or percutaneous treatment improves two-year survival in patients with low and intermediate scores when residual TR is ≤ grade 2+.

COMMENTARY:

In developing commonly used surgical risk assessment systems (EuroSCORE II and STS-PROM score), TR patients were not adequately represented. As a result, the STS system would not be applicable, and EuroSCORE would only equate risk to other isolated procedures, excluding aortic surgery or coronary revascularization. This has led to a poor fit for these systems in categorizing TR patients by risk and a lack of predictive power regarding actual surgical risk. Thus, a new tool to manage these patients effectively has long been awaited. The TRIGISTRY registry was a first approach, gathering various data, including mortality rates of approximately 2%, 9%, and 16% for low, medium, and high TRI-SCORE scores, far from the 20-40% average mortality of classic series when patients were operated in advanced HF stages.

Given this background and other risk stratification systems, TRI-SCORE shows good calibration, i.e., patient scoring matches clinical outcomes and, therefore, established risk categories. Nevertheless, despite its described merits, several critiques can be made, which we can discuss based on findings that will soon populate cardiology and cardiac surgery literature:

1. The sample size in the score system’s creation is limited. Only around 2,400 patients were included, in contrast with over 22,000 for EuroSCORE II or the more than 7.5 million cases in the STS database. This reduces statistical power, limiting the ability to identify statistically significant differences. This limitation is exacerbated by the fact that half the patients received conservative treatment, leaving each invasive alternative with just a quarter of the sample. While recruitment may have been challenging due to the limited casuistry of isolated TR treatment, this restricts the tool’s utility, developed to provide alternatives to conservative management. Also, the two-year maximum follow-up limits its utility in long-term therapeutic decision-making.
2. The apparent goodness of fit results from an overly simplistic design. This is due to the limited sample volume used. Typically, score systems are developed in two phases: an initial development phase using half the sample, followed by a validation phase using the other half to test its fit. TRI-SCORE was directly developed on the TRIGISTRY registry, which could present significant validity issues. The authors do not report patient characteristics across treatment groups and note lower morbidity in the surgical group than in conservative and, especially, percutaneous options. It is logical to expect a good fit on the sample used for development; however, this may not be representative (high-level centers, selected patients, etc.), potentially limiting the tool’s generalizability.
3. The eight variables include: age (cut-off at 70 years), NYHA functional class (III-IV threshold), right HF signs, daily furosemide dose (125 mg threshold, equivalent to 40 mg torasemide), glomerular filtration rate (cut-off at 30 ml/min), total bilirubin (threshold at the high normal limit), left ventricular ejection fraction (60% threshold), and right ventricular function. Each variable contributes cumulatively with a score of 0 to 12 points, where variables such as right HF signs, furosemide dose >125 mg, GFR <30 ml/min, and elevated bilirubin score 2 points, with the remaining variables weighted at one point each. In an excess of statistical confidence, they assign each score a hospital mortality risk based on TRIGISTRY data, which is inadequate due to previously noted validity and sample size limitations. Thus, it becomes an additive score without factoring in exponential clinical effects.
4. Variable selection also warrants a note. Risk scoring systems typically derive from a multivariate analysis, as the basis for the predictive model. Variables with a significant independent relationship with the outcome (two-year mortality in this case) are included, each weighted according to relative significance. However, TRI-SCORE development lacks transparency on method selection, implying that the model was built using the eight TRIGISTRY variables. Furthermore, biases associated with data collection (variable definition, interobserver variability in echocardiograms, etc.) produce paradoxical examples, such as right ventricular function having less weight than other clinical variables; hypertension not included, nor atrial fibrillation or therapy techniques; or redundant variables like right HF signs and NYHA functional class, among other crucial considerations for TR treatment.

Thus, once again, haste and pressure to generate evidence have negatively impacted a well-intentioned initiative. Rather than aiding decision-making for TR patients, TRI-SCORE likely serves as a call to consider invasive treatment early in the disease, before organ damage from right HF raises procedural risks and influences prognosis, even post-valve repair. Thus, two-year survival results suggest early intervention, improving reparability and, in surgical cases, achieving outcomes superior to percutaneous (93% two-year survival vs. 87%; p = .0002). Further research is imminent.

REFERENCE

Dreyfus J, Galloo X, Taramasso M, Heitzinger G, Benfari G, Kresoja KP, et al.; TRIGISTRY investigators. TRI-SCORE and benefit of intervention in patients with severe tricuspid regurgitation. Eur Heart J. 2023 Aug 25. doi: 10.1093/eurheartj/ehad585.