Clinical Outcomes After Valve Intervention in Rheumatic Mitral Valve Disease

Study Population

This retrospective study included patients with symptomatic rheumatic MV disease who were referred for valve intervention—either PMV or MVR—between January 2010 and September 2016 at Hospital das Clínicas, Universidade Federal de Minas Gerais, Brazil.

Inclusion and exclusion criteria

Patients were included if they met the following criteria:

• - Confirmed rheumatic MV disease based on echocardiographic morphological features, including leaflet thickening, restricted leaflet mobility, commissural fusion, chordal thickening, and shortening (21, 22).

• - Severe MV disease, as defined by current guideline recommendations for intervention (8, 9, 11). Severe MS was defined as a mitral valve area ≤1.5 cm² and significant MR was defined as moderate or severe MR, according to established echocardiographic grading systems (23),

• - Symptomatic status, requiring intervention based on dyspnea (NYHA class II-IV) or pulmonary hypertension.

Exclusion criteria included patients with prior valve prosthesis dysfunction undergoing cardiac surgery for prosthetic valve replacement.

Given the retrospective nature of the study, the sample size was determined by the number of patients meeting the inclusion criteria within the study period. While retrospective studies may introduce selection bias, our approach aimed to minimize this by including all consecutive patients referred for intervention during the study period. Additionally, adjustments for potential confounders were applied in the statistical analysis.

The study protocol was approved by the institutional review board of Universidade Federal de Minas Gerais, Brazil.

Echocardiography

A standard transthoracic echocardiogram was performed following the American Society of Echocardiography guidelines (23). Mitral valve morphology was evaluated according to recommendations, taking into account valve thickening, mobility, calcification, and subvalvular chordal fusion, as well as the assessment of commissural morphology (10, 24, 25). Peak and mean transmitral diastolic pressure gradients were measured from pulsed-wave Doppler profiles recorded in the apical four-chamber view. The presence and severity of MR were evaluated according to guidelines (23). The tricuspid regurgitant velocity was recorded with continuous-wave Doppler imaging and used to determine the systolic pulmonary artery pressure using the modified Bernoulli equation.

Mitral valve interventions: surgical and percutaneous mitral valve procedure

MVR was performed with complete preservation of the subvalvular apparatus. The choice of prosthetic valve and suture technique was left to the surgeon’s discretion.

PMV was performed using an anterograde trans-septal approach using the Inoue balloon technique, as previously described (26). Procedural success was defined as a post-procedure MV area was ≥1.5 cm² without a significant increase in MR (25).

Study endpoint definitions

The primary outcome was a composite of long-term adverse events, including all-cause mortality, cardiovascular-related death, mitral valve replacement, repeat PMV, or stroke. Cardiovascular mortality was defined as death due to stroke, heart failure, sudden death, or procedure-related mortality. Other deaths were considered non-cardiovascular.

Patients who underwent unsuccessful PMV requiring subsequent MVR were analyzed in the PMV group (intention-to-treat principle). Outcome data were collected through clinic follow-ups and telephone interviews with patients or their relatives.

Statistical analysis

Categorical data were presented as numbers and percentages and compared using the Chi-square test. Continuous variables were expressed as mean ± standard deviation (SD) or median and interquartile range (IQR), depending on the distribution, and compared using the unpaired Student’s t-test or Mann-Whitney test as appropriate.

Cox proportional hazards regression analyses were performed to identify independent predictors of mortality and composite outcomes. In the univariable Cox hazard model, we included all relevant variables, and those statistical significances were entered into a multivariable model. To address potential confounding, variables were assessed for collinearity, and redundant or highly interdependent covariates were excluded from simultaneous inclusion in the analyses. The final multivariable model was then adjusted for baseline clinical and echocardiographic characteristics that could act as potential confounders.

As missing values were <2% for key variables, the final model was based on complete data, and no imputation techniques were applied.

Kaplan-Meier curves were constructed for all-cause mortality and composite outcomes and compared using the log-rank test.

A P-value <0.05 was considered statistically significant. Statistical analyses were conducted using SPSS version 18 (SPSS Inc., Chicago, IL).

Baseline Characteristics of the Study Population

Of the 334 patients who underwent mitral valve intervention, 302 had rheumatic MV disease. Forty-two patients who had prosthetic MV were excluded, and 14 patients were lost to follow-up, leaving 246 with complete outcome data for analysis. Ninety patients underwent MVR, whereas 156 underwent PMV (Figure 1). The mean age was 43.8 ± 12.9 years, and 198 (80%) were women. The types of MV dysfunction according to the procedure are shown in Figure 2A and 2B.

Figure 1

Figure 2

Among patients referred to PMV, the most frequent valve dysfunction was combined MS and regurgitation with predominance of stenosis (119 patients, 76%), whereas isolated MS was found in 27 patients (17%). Patients with combined balanced lesions who underwent PMV had the symptoms related to MS, without indications for surgery to treat the regurgitation lesion. A single patient with predominance of MR, deemed unsuitable for surgery secondary to cancer, underwent PMV for relief of pulmonary congestion (Figure 2A).

The predominant valvular lesion in the patients who underwent surgery was combined MS and MR with predominance of stenosis (33 patients, 37%), followed by combined balanced lesions (32 patients, 36%). Isolated MR was detected in 21 patients (23%) and isolated MS without any degree of regurgitation in only 4 patients (4%) (Figure 2B).

The baseline clinical characteristics of the study population stratified by the type of MV intervention are summarized in Table 1. Patients were similar regarding age and NYHA functional class. The PMV group had a predominance of women, whereas atrial fibrillation was more frequent in the surgical group.

Similarly, use of diuretics and angiotensin-converting-enzyme inhibitors was more frequent in the surgical group. Mitral valvuloplasty had previously been performed in 72 patients (29%), including either percutaneous or surgical intervention. Patients who were referred for PMV had more frequent previous valvuloplasty than those referred to MVR.

Echocardiographic features

Echocardiographic parameters in the two groups are presented in Table 2. The left ventricular ejection fraction was slightly higher in the MVR group. The largest difference between the groups was in left ventricular diameters. The patients in the surgical group presented with larger end-diastolic and end-systolic dimensions compared to the PMV group. Severe MS defined by valve area less than 1 cm² was found in 98 patients (63%) and moderate (valve area between 1–1.5 cm²) in 58 patients (37%) of the PMV group. In the MVR group, severe MS was found in 47 patients (52%), moderate in 19 (21%), mild (area more than 1.5 cm²) in 17 patients (19%), and no MS in 7 cases (8%). Regarding the degree of MR, moderate and severe MR was observed in 6 patients (4%) in the PMV group compared with 67 patients (74%) in the MVR group.

As expected, echocardiographic Wilkins’ score was lower in the patients who underwent PMV. Systolic pulmonary artery pressure and mean transmitral gradients were similar between the groups.

Functional tricuspid regurgitation associated with MV lesions was found in 60 patients (24%). Rheumatic aortic valve involvement was frequent in this population, with 124 patients (50%) presenting some degree of aortic valve disease, usually mild aortic regurgitation (39%). Fifty-two patients had associated aortic stenosis (21%) with a mean valve area of 1.22 cm² (range 0.8–1.6 cm²).

The surgical procedure for MVR was combined with aortic valve replacement in 14 patients (16%), correction of functional tricuspid regurgitation with annuloplasty in 7 (8%), and coronary artery bypass surgery in 2 (2%).

Long-term outcomes

The overall mean follow-up time was 2.8 years (range 1 day–7.8 years); the median was 3.2 years for the surgical group and 2.7 years for the valvuloplasty group (p = 0.08). During the follow-up period, 45 adverse clinical events were observed. Twelve patients died (2 non-cardiac deaths), 22 patients underwent MVR, and 7 repeated PMV. Additionally, 4 patients had a stroke. The outcomes according to the type of valve intervention are shown in Table 4. The mortality was higher in the MVR group, whereas the need for valve intervention (repeated PMV and valve replacement) was higher in the PMV group.

Outcomes in PMV group

PMV resulted in adequate valve opening without an increase in MR or other complications in 131 patients (84%). The procedure was considered unsuccessful in 25 patients (16%), due to insufficient valve opening in 8 patients (5%) or a greater than one-grade increase in MR grade in 17 patients (11%). Additionally, 6 patients (4%) required emergency MVR surgery due to severe MR following PMV. During a median follow-up of 2.7 years, 31 adverse clinical events were observed, including 3 deaths, 20 mitral valve replacements, 7 repeat PMVs, and 1 stroke (Table 3). Long-term outcome after PMV was strongly determined by the quality of immediate procedural results (hazard ratio (HR) 3.352; 95% CI 2.396–4.690, p < 0.001).

Outcomes in surgical group

In the MVR group, there were 3 operative deaths. Eleven patients required reoperation due to cardiac tamponade (n = 8), major bleeding (n = 2), and mediastinitis (n = 1). Additionally, nine patients presented with postoperative infections. During a mean follow-up of 3.2 years, 14 adverse events were observed, including 9 deaths (2 non-cardiac deaths), 2 needed further valve replacement, and 3 strokes (Table 3).

Predictors of long-term outcome

Univariable Cox proportional-hazards analysis for predicting all-cause mortality and the composite outcome are shown in Table 4. Clinical and echocardiographic parameters that express the severity of MV disease or its hemodynamic consequences were tested in the univariable analysis.

Unadjusted survival curves are shown in Figure 3. PMV outcomes were comparable to MVR. At 1-year follow-up, event-free survival rates were 91% for PMV patients and 92% for MVR patients (P = 0.470). After two years, there was a non-significant trend for a more favorable outcome for MVR versus PMV; 5-year event-free survival rates were 85% for MVR versus 71% for PMV (P = 0.07). After six years of intervention, event-free survival was similar between the two types of intervention, with survival rates of 75% for MVR versus 65% for PMV (p = 0.233).

Figure 3

In the Cox multivariable proportional hazards analysis, adjusted for baseline differences between groups, age and NYHA functional class emerged as independent predictors of the composite outcomes (Table 5). In terms of mortality, baseline variables reflecting the impact of MV disease, including atrial fibrillation, left ventricular ejection fraction, and age, were associated with an increased risk of death. The PMV group had a lower mortality rate compared to the surgical group (3.2% vs. 12.2%; P = 0.005) (Figure 4). However, in the multivariable proportional hazards analysis, only age and left ventricular ejection fraction remained independent predictors of mortality risk (Table 5). Notably, left ventricular ejection fraction ≤60% was associated with a greater mortality risk (hazard ratio [HR] of 5.3, 95% confidence interval [CI] of 1.4–19.4; P = 0.013) compared with the value >60%.

Figure 4

Intervention strategies in rheumatic MV disease

Definitive treatment of symptomatic rheumatic MV disease is based on either surgical MVR or PMV. While MV repair is superior to MVR in select patients due to lower mortality (20), it is associated with higher rates of reoperation (16, 20).

The choice of intervention depends on the pattern of rheumatic valve involvement, including the predominant valve lesion, symptoms, and hemodynamic consequences (8, 9). However, mixed valve disease presents a challenge, as current noninvasive and invasive assessments have limitations in accurately defining severity. Furthermore, the optimal timing of intervention for patients with moderate mixed mitral disease remains unclear due to limited evidence.

Recent studies suggest that PMV indications have expanded to include patients with unfavorable valve anatomy, reflecting both epidemiological shifts in patient populations and advancements in interventional techniques (27, 28, 29, 30). Our findings confirm this trend, as 63% of patients underwent percutaneous intervention. However, the expansion of PMV indications should only be considered in centers with experienced operators and immediate surgical backup to manage complications (29).

Previous studies comparing surgical commissurotomy and PMV in isolated MS have demonstrated similar outcomes (14, 17, 31). Reyes et al. (14) demonstrated that balloon valvuloplasty and open surgical commissurotomy yielded comparable initial results and low restenosis rates, maintaining good functional status for at least three years. Further studies reinforced these findings, demonstrating equivalent early hemodynamic improvements in patients with isolated MS (15). Nevertheless, in cases of concomitant MR or suboptimal valve anatomy, MVR is often preferred (8, 9). For patients who failed PMV in particular, the surgical approach should be considered. In our study, 59 patients (38%) referred for PMV had undergone prior valvuloplasty, emphasizing that MV restenosis is a frequent indication for repeat intervention. For patients failing PMV, surgical intervention remains the preferred approach.

The long-term durability of valvuloplasty, whether surgical or percutaneous, in patients with extensive valve deformity remains uncertain. Late symptomatic deterioration often occurs, and reoperation rates remain high (17), even in experienced centers. Studies reported freedom from reoperation at 20 years in only 50%–60% of cases (32). A large series from Korea showed that MV repair was accomplished in 22% of patients undergoing surgery for RHD, and among them, one-third developed significant MS or MR within ten years. (19). Similarly, Bouleti et al. (33) reported that even after a successful PMV, two-thirds of patients required repeat intervention over a 20-year follow-up, with the majority undergoing MVR.

In ischemic MR, MVR has demonstrated superior durability in correcting regurgitation, potentially influencing long-term outcomes (34). Song et al. (35) found that patients with an echocardiographic score >8 and atrial fibrillation had better long-term outcomes with MVR, suggesting that early valve replacement should be considered in these cases. Another study by Kim et al. (36) compared MVR with repeat PMV in patients with mitral restenosis after previous PMV, showing that MVR provided superior long-term outcomes and was more effective than repeat PMV for up to 15 years. Kim et al. (36) also demonstrated that outcome was initially similar between groups for 40 months, but MVR had a clear advantage after a long follow-up period. In contrast, our study found no outcome differences between the two interventions. This may be due to our shorter follow-up period compared to Kim’s 15-year study.

A key consideration is whether patients should undergo MVR earlier, given that many will ultimately require a prosthetic valve in the long term (17). There are no comparative studies to answer this question definitively. The issue of repair versus replacement in rheumatic MV disease has been debated for years, with some studies reporting better survival with repair, despite its greater reoperation rate (16, 20, 37, 38).

Although MVR improves survival in symptomatic patients, it carries substantial perioperative risks and long-term complications. Operative mortality for isolated MVR ranges from 3% to 8% in most centers (39). In the present study, the overall mortality rate was 4.9% (12 out of 246), even considering that 16% of the surgeries involved both mitral and aortic replacement. In addition, MVR frequently leads to lifelong anticoagulation and poses long-term risk of prosthesis-related complications.

Study limitations

Our study has several limitations inherent to its retrospective design, including potential selection bias and missing data. The heterogeneous patient population included a spectrum of rheumatic MV disease, ranging from isolated pliable MS to severe predominant MR and mixed mitral disease, making direct comparisons a challenge in disease management.

Although selection bias may exist due to differences in intervention strategies, which were influenced by clinical characteristics, valve anatomy, institutional management protocols, and operator expertise. Additionally, while missing data were limited (<2% for key variables), no imputation techniques were applied. Despite these limitations, our study provides valuable insights into the outcomes of MV interventions in a real-world setting and highlights the factors associated with long-term prognosis in patients with rheumatic MV disease.