This study was designated as a retrospective analysis from the Heart Failure Network Rhineland registry, which is a multicenter, prospective, observational registry of symptomatic patients with MR who undergo TMVR at three high-volume centers in Germany (University Hospitals of Bonn, Cologne, Duesseldorf) [14, 15]. From this registry, we identified consecutive patients who underwent their first edge-to-edge TMVR with the MitraClip system (Abbott Vascular, Santa Clara, California) between October 2011 and October 2018. Patients with available pre-procedural serum creatinine and total bilirubin results were included in the analysis. We excluded patients who were lost to follow-up within one month after TMVR. Patients agreed to participate in our registry which was approved by the Ethical Committee of the individual center in accordance with the Declaration of Helsinki.

The indication for TMVR was moderate-to-severe or severe MR accompanied by symptomatic HF according to the New York Heart Association (NYHA) functional classification in patients considered as inoperable or at high surgical risk. After a standardized diagnostic workup including transesophageal echocardiography (TEE), the decision to perform the intervention was taken by the interdisciplinary heart team of each center. Procedures were performed under general anesthesia or sedation with three-dimensional TEE and fluoroscopic guidance. Details of the device system and procedure have previously been well described [16]. The discretion of whether a second or third device needed to be used was left up to the treating physicians. Device success was defined as post-procedural MR ≤ 2 and post-procedural mean transmitral valve pressure gradient < 5 mmHg after deployment of one or more clips to achieve leaflet approximation and retrieval of the delivery system.

Hepatorenal function was assessed using the MELD-XI score which was calculated as 5.11 × ln (serum total bilirubin in mg/dl) + 11.76 × ln (serum creatinine mg/dl) + 9.44 [4]. The creatinine and total bilirubin measurements were performed at each institutional laboratory with other parameters, including N-terminal pro-B-type natriuretic peptide (NT-proBNP) and hemoglobin. A high MELD-XI score was defined as > 11, according to a previous report [5, 7]. Post-procedural blood data, including creatinine and total bilirubin, were collected at 1- and 12-month follow-up after TMVR. In addition, anemia was defined according to the World Health Organization criteria (hemoglobin < 13 g/dl for male patients; hemoglobin < 12 g/dl for female patients).

We assessed echocardiographic parameters performed at baseline and discharge according to the current guidelines [17]. The severity of MR was graded as follows: grade 0, none; 1 + , mild; 2 + , moderate; 3 + , severe. All measurements were reviewed by an independent cardiologist dedicated to echocardiographic evaluation at each center.

The primary endpoint was a composite outcome, consisting of all-cause mortality and hospitalization due to worsening HF within 2 years after TMVR. As secondary endpoints, all-cause mortality and hospitalization due to worsening HF within 2 years after TMVR were examined separately. All suspected adverse events were independently adjudicated by the local heart team according to the criteria of the Mitral Valve Academic Research Consortium 2 [18]. The need for hospitalization due to worsening HF was determined based on the attending physicians’ discretion without any prespecified criteria. The occurrence of clinical events was recorded from admission and outpatient medical records. Telephone interviews were also performed with the patients’ general practitioners or family. In addition, HF medication, including beta blockers, renin–angiotensin system (RAS) inhibitors, and aldosterone antagonists, and dosage with a standardized furosemide equivalent were recorded at baseline [19].

Continuous variables were tested for normal distribution using the Kolmogorov–Smirnov test. Normally distributed variables are presented as the mean ± standard deviation and compared using t tests. In contrast, non-normally distributed variables were expressed as medians with an interquartile range (IQR) and compared between groups using the Mann–Whitney U test. Categorical data were presented as numbers and percentages, and the differences between groups were evaluated using the chi-square test. Paired t test or Wilcoxon signed rank test compared change in hepatorenal function after TMVR within the same patient. Logistic regression analysis was performed to detect parameters that were related to the MELD-XI score. The variables with p < 0.05 on univariate analysis were incorporated into a multivariable regression model. The receiver-operating characteristic (ROC) curve was used to assess the validity of the cut-off value of the MELD-XI score for the 2-year composite outcome in the present study. In addition, Harrell’s C-statistic was used to compare the predictive ability of hepatorenal functional markers for the composite outcome by area under the curve analysis, accounting for censoring. Kaplan–Meier cumulative event curves for the composite outcome, HF hospitalization, and all-cause mortality were generated using two groups according to the high or low MELD-XI score. Differences between the groups were compared using the log-rank test. Univariate and multivariable Cox-proportional hazard models were used to calculate the hazard ratios (HRs) with 95% confidence intervals (CIs) for MELD-XI score for the composite outcome within 2 years after TMVR. In univariate analysis, we analyzed the HRs of conventional covariables that were determined according to previous reports. In multivariable analyses, covariates were included that showed significance (p < 0.05) in the univariate analyses, considering multicollinearity. A stratified analysis was performed to assess effects of interactions between high MELD-XI score (> 11) and clinical parameters on the composite outcome. The stratified analysis consisted of the following parameters: sex (male vs female), age (≥ 75 years vs < 75 years), etiology of MR (functional MR vs degenerative MR), MR severity at baseline (moderate-to-severe vs. severe), left ventricular ejection fraction (LVEF) (> 30% vs ≤ 30%), tricuspid annular plane systolic excursion (TAPSE) (≥ 15 mm vs < 15 mm), tricuspid regurgitation (TR) severity at baseline (≥ severe vs < severe), device success (yes vs no), and residual MR (≥ Grade II vs < Grade II), and additionally the p value for interaction between subgroups and high MELD-XI score (> 11) was examined. In addition, we performed a restricted cubic spline with 4 knots at the MELD-XI score to model a relationship between the MELD-XI score and the composite outcome. Statistical significance was set at p < 0.05. All analyses were conducted using Stata 15.1 (StataCorp, College Station, TX, USA) or JMP version 14.0 for Mac (SAS Institute Inc., Cary, NC, USA).

Of 881 patients, the mean age was 77.0 ± 9.0 years and 58.0% were of male sex (Table 1). Eighty-three percent of the patients were classified in NYHA functional class III or IV. The expected surgical mortality rate was elevated, as evidenced by the median logistic EuroSCORE of 16.4% (IQR 9.2–28.9%). Pre-procedural MR severity was moderate-to-severe or severe in 13% and 87% of the patients, respectively. The mean left ventricular ejection fraction (LVEF) was 44.6 ± 15.1%, and functional MR was the etiology of MR in 61% of the patients.

Clip implantation failed in 12 patients (1.4%), and device success was achieved in 93.0% of the patients. The mean number of implanted clips was 1.5 ± 0.6 (Supplementary Table 1). Echocardiography at discharge showed that residual MR of less than 2 + was prevalent in 56.0% of the patients.

The mean MELD-XI score was 11.0 ± 5.9 (Table 1), and the distribution of the MELD-XI score is presented in Fig. 1. In addition, the median estimated glomerular filtration rate (eGFR) was 48.0 ml/min/1.73m² (IQR 34.4–63.1 ml/min/1.73m²), and the median total bilirubin was 0.63 mg/dl (IQR 0.44–0.97 mg/dl). Fifteen patients received hemodialysis therapies. MELD-XI score was correlated to sex, MR effective regurgitation orifice area (EROA), left ventricular ejection fraction (LVEF), LV end-diastolic volume, TAPSE, right atrial area, left atrial volume, severe or more TR (Table 2). However, on multivariable analysis, male sex had the strongest correlation (standardized β: 0.27; 95% CI 0.16–0.37; p = 0.003), followed by LVEF (standardized β: − 0.19; 95% CI − 0.32 to − 0.26; p = 0.004), MR EROA (standardized β: 0.15; 95% CI 0.05–0.24; p = 0.003), and severe or more TR (standardized β: 0.13; 95% CI 0.02–0.22; p = 0.02).

Of the 881 patients, 415 patients (47.1%) had high MELD-XI score (> 11), and patients characteristics and echocardiographic findings of each group are shown in Table 1. Patients with the high MELD-XI score (> 11) had a lower LVEF and larger LV end-diastolic volume compared to those with low MELD-XI score (≤ 11) (41.4 ± 15.0% vs. 47.3 ± 14.7%; p < 0.0001, and 149 ml (IQR 109–192 ml) vs. 118 ml (IQR 91–160 ml); p < 0.0001, respectively). In contrast, there were no significant differences in the rate of device success and length of hospitalization.

Post-procedural changes in the MELD-XI score and other parameters of hepatorenal function were evaluated after TMVR (Supplementary Table 2). Compared to the value at baseline, the median MELD-XI score did not change significantly at 1- and 12-month follow-up after TMVR (11.8 ± 6.3; p = 0.86 and 11.2 ± 5.8; p = 0.79, respectively).

Within 2 years after TMVR, 184 patients (20.9%) died, 203 patients (23.0%) were rehospitalized due to worsening HF, and 337 patients (38.2%) experienced the composite outcome (Supplementary Table 1). Patients with the high MELD-XI score (> 11) had a higher incidence of the 2-year composite outcome compared to those with the low MELD-XI score (≤ 11) (54.7% vs. 36.9%; log-rank p < 0.0001). The Kaplan–Meier curves for each outcome are shown in Fig. 2. In addition, the high MELD-XI score (> 11) was associated with a higher incidence of the 2-year composite outcome in spite of MR etiology, such as functional MR and degenerative MR (Supplementary Fig. 1).

From the ROC curve for the 2-year composite outcome, the optimal cut-off value of the MELD-XI score was also 11 [area under the curve (AUC): 0.62; p < 0.0001] (Supplementary Fig. 2). The Harrel’s C-statistic of MELD-XI score for the 2-year composite outcome was 0.60 (95% CI 0.57–0.63), which was higher than those of eGFR or total bilirubin (C-statistic: 0.57, 95% CI 0.54–0.61; p = 0.049, and C-statistic: 0.55; 95% CI 0.52–0.59; p = 0.004, respectively) (Supplementary Table 3).

In the multivariable Cox-proportional hazard analysis, covariates were included which showed significance (p < 0.05) in univariate analysis (Supplementary Table 4). Patients with high MELD-XI score (> 11) had a higher incidence of the composite outcome within 2 years after TMVR compared to those with low MELD-XI score (≤ 11) (adjusted HR 1.34; 95% CI 1.02–1.77; p = 0.04) (Table 3). Moreover, MELD-XI was an independent predictor of the incidence of 2-year composite outcome (adjusted HR 1.02; 95% CI 1.00–1.05; p = 0.048).

The association between high MELD-XI score (> 11) and the composite outcome in subgroups is shown in Fig. 3. In the stratified analysis for the composite outcome, there were no significant interactions across the subgroups except for LVEF ≤ 30% (p for interaction = 0.02).

Our restricted cubic spline demonstrated a consistent increasing hazard of the composite outcome in the MELD-XI score of greater than 11. However, in the MELD-XI score of less than 11, a significant interaction between the MELD-XI score and the incidence of composite outcome was not observed (Fig. 4).

To the best of our knowledge, this is the first multicentric study assessing the prognostic impact of the MELD-XI score in patients undergoing TMVR; however, several limitations must be acknowledged. First, this was a retrospective study. Therefore, a certain patient selection bias might have impacted our results. Second, we could not distinguish acute hepatorenal injury from chronic dysfunction because we evaluated the MELD-XI score from only the latest laboratory data before TMVR. Finally, we could not assess post-procedural changes in the MELD-XI score after TMVR and could not investigate the prognostic impact of post-procedural changes in the MELD-XI score.