Long-term outcomes of papillary muscle relocation anteriorly for functional mitral regurgitation

Abstract

Open in new tabDownload slide

OBJECTIVES

This study aimed to evaluate the outcomes of the patients who underwent restrictive annuloplasty (RA) plus papillary muscle relocation anteriorly (PMR-A) with the risk factors in mitral valve repair for functional mitral regurgitation (FMR).

METHODS

Eighty-six patients underwent mitral valve repair with RA for FMR. Thirty-five of them received additional bilateral papillary muscle relocation for severe leaflet tethering. The papillary muscles were relocated posteriorly (PMR-P) early in the study. Then, in the later period, the technique was modified to PMR-A, in which the papillary muscles were relocated anteriorly for 24 cases. The survival of the patients undergoing RA + PMR-A was examined retrospectively, adjusting for differences in patient background.

RESULTS

Twenty-three deaths were observed during the follow-up period out of the 86 cases. Independent preoperative risk factors for survival were left ventricular ejection fraction, patient age and B-type natriuretic peptide (BNP) level. Among the patients with BNP <1000 pg/ml, 5-year survival after RA plus PMR-A was 84.7%, while RA alone was 78.6% and RA + PMR-P 57.1%. Cox proportional hazards regression adjusted for the preoperative risk factors showed a significantly higher hazard ratio of RA + PMR-P to RA + PMR-A (12.77, P = 0.011), while the hazard ratio of RA alone to RA + PMR-A was not significantly different. Furthermore, reverse remodelling of the left ventricle was observed for 3 years only in RA + PMR-A.

CONCLUSIONS

Long-term survival for patients who underwent RA plus bilateral PMR-A was promising. Patients with significantly higher BNP had lower survival after valve repair for FMR.

INTRODUCTION

Functional mitral regurgitation (FMR) often results in severe heart failure due to left ventricular systolic dysfunction [1]. Various techniques with mitral subvalvular intervention to add to restrictive annuloplasty (RA) have been reported for FMR [2–5]. However, long-term outcomes of the subvalvular intervention are still poorly understood. In addition, preoperative risk factors that predict the outcomes are not sufficiently investigated. The B-type natriuretic peptide (BNP) level is a well-established predictor of the prognosis in heart failure [6–9], while little is known about long-term survival for FMR and preoperative BNP levels. This study retrospectively evaluated the long-term outcome of patients who underwent papillary muscle relocation anteriorly (PMR-A) for FMR with severe mitral apex tethering. Specifically, to adequately assess survival in a potentially heterogeneous patient population, we performed a risk-adjusted analysis of preoperative risk factors in patients who underwent multiple techniques for mitral valve repair for FMR.

PATIENTS AND METHODS

Ethics statement

The Ethics Review Committee of Tokyo Medical and Dental University School of Medicine has approved this study's ethical appropriateness (2014/08/19 Research number M2000-1806). Formal written consent was obtained from all patients.

Study patients

Between 2005 and 2019, 104 cases of cardiac surgery were performed for FMR due to ventricular dysfunction at Tokyo Medical and Dental University Hospital. Optimal medical treatment was provided preoperatively in all patients. Of these 104, 16 patients received a ventricular assist device as they met criteria. Of the 88 cases excluding the above 16, preoperative BNP levels were measured in 86 cases, for which RA was performed. Of these 86 undergoing RA, 35 patients were diagnosed with severe leaflet tethering due to advanced ventricular dilatation and given an additional papillary muscle relocation (PMR). Papillary muscle relocation posteriorly (PMR-P) was performed in 11 during the early study period to relocate the papillary muscles posteriorly towards the direction of the mitral annulus of the posterior leaflet. Later, the technique was modified, and PMR-A, which relocates the papillary muscles anteriorly, was performed for 24 patients. A total of 625 mitral valve surgery were performed during the same period, and 44 ventricular assist devices were implanted.

Study methods

As this study examined patient outcomes retrospectively, the potential for heterogeneity in the patient background was anticipated. Therefore, we aimed to identify preoperative risk factors in patients who underwent mitral valve repair with multiple techniques for FMR and to evaluate the outcomes of 24 patients who underwent RA + PMR-A by risk-adjustment analysis.

Patients who underwent mitral leaflet resection and suture were not included in this study because of degenerative lesions. The BNP reagent used in these patients was Shionoria BNP (Shionogi & Co., Ltd., Osaka, Japan). All 86 patients' preoperative background characteristics, and the pre- and postoperative echocardiographic data were collected and analysed to elucidate the background of patients undergoing PMR-A and their echocardiographic characteristics. In addition, on the transthoracic echocardiogram, the 4 angles formed between the anterior mitral leaflet and the annular plane, and the posterior mitral leaflet and the annular plane, at the end-diastole and end-systole were defined as anterior leaflet opening angle (ALOA), anterior leaflet closing angle (ALCA), posterior leaflet opening angle (PLOA) and posterior leaflet closing angle (PLCA) (Fig. 1) [10, 11]. These angles were measured and examined to determine whether they indicate tethering relief and represent surgery outcomes.

Surgical technique

The surgical procedure of mitral valve repair for FMR was determined according to the following criteria. First, the policy was to perform RA on all patients. Then, as previously described [4], bilateral PMR was performed for patients with advanced left ventricular dilatation and reduced leaflet mobility with highly tethered valve leaflets. A polytetrafluoroethylene mattress suture with pledget (CV-4, W.L. Gore and Associates, Newark, Delaware) was placed and ligated on the fibrous portion of both the anterior and posterior papillary muscle tips. The ligated CV-4 sutures were passed through the mitral annulus, followed by a run through the corresponding portion of the semi-rigid annuloplasty ring (Carpentier-Edwards Physio II; Edwards Life sciences, Irvine, CA, USA). After the annuloplasty ring was placed on the annulus, the CV-4 strings were towed and pulled up to reposition the papillary muscles while performing the saline test. The anterior mitral leaflet was consequently lifted towards the annular plane, and the CV4 sutures were ligated at the height where the leaflets were coapted and not prolapsed. Early in the study, the CV-4 sutures were carried towards the annuls of the posterior leaflet (Fig. 2A), and the papillary muscles were relocated towards the posterior annulus. This procedure was named PMR-P. Later, the relocation technique was modified; the CV-4 was carried towards the annuls of the anterior leaflet, and the papillary muscles were relocated towards the anterior annulus (Fig. 2B). This method was named PMR-A. Coronary artery bypass grafting, maze procedure and tricuspid valve surgery were performed concomitantly for eligible patients for these procedures. No patients were diagnosed requiring cardiac resynchronization therapy for the left bundle branch block. All patients received optimal medical treatment throughout the preoperative and postoperative periods. Anticoagulation therapy with warfarin was continued for a minimum of 3 months after surgery.

Statistical analysis

The study first reviewed the preoperative characteristics and echocardiographic data of patients who underwent surgery using Fisher's exact test and analysis of variance. The Gray test examined recurrent mitral regurgitation (MR) with death as a competing risk. To explore preoperative risk factors for survival, Kaplan–Meier curves, log-rank tests and multivariable analysis based on Cox proportional hazards model were used to examine risk factors for survival. Regarding the candidate risk factors, all variables in the table described below were treated as continuous values except for BNP, regarding which log-transformed values were used. Then, preoperative risk factors were selected by backward–forward stepwise model selection using the Akaike's Information Criterion. Then, cut-off values were set using a specific risk factor, and overall and stratified survival after surgery was investigated. To view the adequacy of PMR-A, univariable analysis with survival time as the objective variable, surgical technique variables as explanatory variables and multivariable analysis with the selected risk factors as adjustment variables were employed. Furthermore, the Wald test examined the survival of RA + PMR-A adjusted for risk factors. Finally, mixed effect model analyses of echocardiographic values were performed in patients with confirmed survival to 3-year follow-up for each procedure to observe trends in PMR-A. The model included the factors of procedure, time and their interaction as the fixed effect and random intercepts for each subject. The Kolmogorov–Smirnov test confirmed the distribution of outcomes. If a continuous variable was not normally distributed, the variable was log-transformed. All statistical analyses were performed using R (The R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Preoperative characteristics and surgical profile

The preoperative characteristics and surgical profiles of the patients are shown in Table 1. The mean age of overall 86 patients was 68 years, 80% (69 cases) were male, 45% (39 cases) had diabetes and the mean preoperative BNP level was 830 pg/ml. The proportion of patients with ischaemic MR differed significantly between procedures. Echocardiographic data showed that the mean left ventricular diameter and volume were larger than normal. The mean left ventricular ejection fraction (LVEF) was 35.9%. Reflecting the surgery criteria, preoperative echocardiographic data of the patients showed significant differences among the 3 surgical techniques in left ventricular end-systolic diameter (LVDs), left ventricular end-diastolic volume, LVEF and the degree of MR. Concomitant surgeries are shown in Table 1.

Early postoperative results and postoperative MR

Early postoperative results showed that 45.3% (39 patients) of the 86 patients required a respirator for >72 hours, and 12.8% (11 patients) required new postoperative haemodialysis. Thirty-day mortality was 1.2% (1 patient). The pre- and postoperative echocardiographic data are shown in Supplementary Material, Table S1. The mean left ventricular end-diastolic diameter (LVDd) and LVDs reduced postoperatively in all techniques, but no improvement in LVEF was observed. Tenting height and tenting area were decreased in all techniques. However, mixed-model analyses revealed that the degree of decrease varied depending on the surgical technique, with RA alone having slighter decreases and RA + PMR-P having greater decreases (tenting height P = 0.002, tenting area P = 0.004). As for the mitral leaflet angles, there was no significant change in the mean ALOA in the overall cases. However, focusing on each surgical technique, ALOA increased from 56.3° to 68.8° in RA + PMR-A and decreased from 55.7° to 41.0° in RA + PMR-P. A mixed model for the interaction between the time and surgical technique factors showed a significant difference (P < 0.001), showing that the change in ALOA was significantly different among the different surgical techniques. The 5-year MR recurrence rate, moderate MR or more, of all 86 patients was 21.2% (Supplementary Material, Fig. S1A); the recurrence rate of RA alone, RA + PMR-A and RA + PMR-P was 20.4%, 22.0% and 22.5%, respectively (Supplementary Material, Fig. S1B), and there were no significant differences in recurrence.

Survival, preoperative risk factors and adjusted survival by surgical technique

The median follow-up period was 2.3 years; of the 86 patients, 23 deaths were confirmed. Of these, 11 were in RA alone (cardiac cause 5, pulmonary 3, sepsis 1, cancer 1, unknown cause 1), 7 in RA + PMR-A (cardiac cause 5, pulmonary 1, unknown cause 1) and 5 in RA + PMR-P (cardiac cause 2, cerebral 1, unknown cause 2). There were 13 deaths out of 64 ischaemic aetiology and 10 out of 22 nonischaemic cases. One stroke was observed during the follow-up period in each RA + PMR-A and RA + PMR-P, but no other valve-related events such as thromboembolic events or endocarditis were observed.

A univariable and multivariable analysis by Cox regression was performed to identify independent preoperative risk factors for survival in overall patients (Table 2). Seven factors such as patient age, history of hypertension, history of diabetes, estimated glomerular filtration rate, BNP level, ischaemic aetiology and LVEF were listed as candidates with univariable analysis. The BNP values were found not to follow a normal distribution, so the values were log-transformed. The Cox model identified patient age, log BNP level and LVEF as independent preoperative risk factors. Other factors were excluded during the stepwise process.

The BNP level had the lowest P-value in the univariable analysis among these preoperative risk factors. Hence, all patients were divided into 2 groups with a 1000 pg/ml BNP level cut-off value. The survival of overall patients at 5 years was 66.2% (Fig. 3A); that of RA alone, RA + PMR-A and RA + PMR-P was 71.9%, 65.8% and 43.6%, respectively, showing no significant differences (log-rank test, P = 0.182). The survival of the patients at 5 years of <1000 pg/ml (66 cases) was 78.0%, while the survival of ≥1000 pg/ml (20 cases) was 28.4%, which was significantly poor (log-rank test P = 0.001, Fig. 3B). Next, the RA + PMR-A patients were divided into 2 groups by this cut-off value: those with <1000 pg/ml (17 patients) had significantly better survival at 5 years of 84.7%, while those with 1000 pg/ml and above (7 patients) had a survival of 28.6% (log-rank test P = 0.003, Fig. 3C). The distribution of the BNP level is shown in Supplementary Material, Fig. S2.

The 5-year survival of RA + PMR-A was then examined, excluding patients with significantly higher BNP levels. The survival of RA + PMR-A in patients with BNP <1000 pg/ml was 84.7%, while that of RA alone was 78.6% and RA + PMR-P 57.1% (Fig. 3D). To view the adequacy of PMR-A, univariable analysis and multivariable analysis with the selected risk factors as adjustment variables were performed. The hazard ratio of RA + PMR-P to RA + PMR-A was 12.77 (P = 0.039), adjusting for preoperative risk factors (Table 3). Wald test showed a P-value of 0.032 for the 3 techniques.

Changes in echocardiographic measurements

The Kolmogorov–Smirnov results of echocardiographic measurements were not significant, and the values were treated as a normal distribution. Supplementary Material, Table S2 compares the mitral valve echocardiographic data of patients with <1000 pg/ml preoperative BNP levels. Tenting height and tenting area decreased postoperatively in each surgical technique, especially tenting height, a significant difference in the time and surgical technique interaction (Fig. 4A). There was no significant difference in the change of mean ALOA. However, focusing on the surgical techniques, ALOA increased from 54.9° to 68.1° in patients with RA + PMR-A and decreased from 60.0° to 37.6° in RA + PMR-P (Fig. 4B). A mixed model showed a significant difference in the interaction between the time and surgical technique (P < 0.001), indicating that PMR-A significantly increased ALOA and PMR-P decreased it. There were no significant interactions between the time and surgical procedure for ALCA, PLOA and PLCA.

Next, the ventricular diameters and LVEF of the patients with BNP <1000 pg/ml who survived for 3 years after surgery were followed yearly to review the transitions. Supplementary Material, Table S3 shows the results. The mean LVDd in RA + PMR-A and RA alone decreased for 3 years, while no significant change was in RA + PMR-P (Fig. 4C). A mixed effect model exhibited a P-value of <0.001 in the interaction between the time and surgical procedure. The LVDs also decreased in PMR-A (P < 0.001) (Fig. 4D); however, no significant change was observed in RA alone or RA + PMR-P. The mixed effect model showed in the interaction a significant difference (P < 0.001). Similarly, the LVEF increased from 33% to 60% for 3 years in PMR-A (Fig. 4E), but no change in LVEF in RA alone or RA + PMR-P. The P-value of the interaction was 0.002.

DISCUSSION

This study showed that survival after PMR-A for patients was promising despite severe leaflet tethering. The study also demonstrated that notably high preoperative BNP levels strongly predicted poor survival. In addition, reverse remodelling of the left ventricle was observed in RA + PMR-A. Therefore, we consider that the study results facilitate our surgical policy. The latest guidelines recommend transcatheter edge-to-edge repair over surgery for FMR [6]. Although long-term outcomes of surgery with subvalvular intervention had not been well demonstrated, this study showed a preferable survival. Thus, PMR-A could pose a substitute therapy in the future. We acknowledge that the present study did not examine whether the remodelling directly affected the long-term outcomes, but this matter needs to be clarified in the future.

Preoperative BNP level

In several guidelines, BNP is a biomarker of heart failure and is considered the most critical factor in predicting prognosis [6–9]. Preoperative BNP levels have been reported to predict prognosis after cardiac surgery [12, 13]. This study revealed that the preoperative BNP level was a powerful predictor for postoperative survival in FMR, so the BNP should be paid close attention to predicting the prognosis after surgery for FMR. High risk is predicted when surgical mitral valve repair is performed for FMR patients with a markedly high BNP level. A left ventricular assist device or heart transplantation might be an option in such cases.

The direction of PMR

We have carefully observed changes in mitral valve morphology on postoperative echocardiography in the study patients, and we suspected that PMR-P might not be efficient in directing blood flow to the left ventricle. PMR-P is a technique that pulls both papillary muscles towards the posterior mitral annulus, but simultaneously, this technique pushes the posterior mitral annulus down towards the apex through the papillary muscles. This displacement of the posterior annulus may induce a tilt of the mitral annular plane and limit the opening of the anterior leaflet. We considered this issue a disadvantage of PMR-P and developed PMR-A as a surgical technique to overcome it [4].

It is noteworthy how PMR-A and PMR-P differently affect the left ventricle. Both procedures bring the papillary muscles closer to the mitral annulus and presumably produce similar vectors in the long axis of the left ventricle. On the other hand, they are considered to generate different vectors in the short axis. We conducted a literature search to see if previous studies have described this difference in the short-axis vector. As a result, it was reported that in experimental ovine models, the posterolateral papillary muscles were displaced posterolaterally in the ischaemic MR of the left circumflex artery. Also, the anterior papillary muscles were displaced laterally in the nonischaemic FMR and the posterolateral papillary muscles posterolaterally [14]. Among the echo video recordings kept at our hospital, we present a video that could make it easy to understand the changes in the short-axis dimension of the papillary muscles (Video 1). We consider that PMR-A, rather than PMR-P, is reasonable to compensate for the papillary muscle displacement. However, the referred previous study was based on an experimental animal model, and a more rigorous examination of the surgical mechanism would require a quantitative echocardiographic quantitative echocardiographic/MRI imaging would be needed to examine more rigorous surgical mechanisms. Future research is needed in this area, and we hope that research in this area will become more active in the future.

Mitral leaflet angles on mitral valve tethering

This study defined the angles between the mitral valve leaflets and the annular plane, ALOA, ALCA, PLOA and PLCA and examined their potential as new tethering index values. PMR-A increased ALOA and PMR-P decreased it. We consider that decreased ALOA may mean restricted anterior leaflet opening in diastole and may have caused mitral stenosis. RA is known to have the potential to cause postoperative functional mitral stenosis [11]. In addition, an increased mitral pressure gradient after the transcatheter edge-to-edge repair is known to worsen the long-term prognosis of MR [15]. Our supplemental analysis also suggests that the pressure gradient may be increased in PMR-P in Supplementary Material, Table S4 and Supplementary Material, Fig. S3. The survival difference among the procedures in this study may indicate that ALOA is an appropriate predictor of postoperative outcome. Measuring ALOA is effortless and requires no particular expertise. This study did not use a cut-off value for the ALOA and had determined the indication for surgery on case-by-case basis based on the motion of the echocardiogram. Based on the results of this study, we now believe that PMR-A should be performed in cases where the ALOA is less than 60 degrees.

Ischaemic and nonischaemic aetiology

The aetiology of MR differed significantly among the groups, with a higher proportion of nonischaemic MR in the PMR-A group. In addition, the univariable analysis showed that survival was worse for nonischaemic MR. In patients with ischaemic aetiology, concomitant CABG is thought to increase LVEF and induce reverse remodelling by revascularization. At the same time, there is no evidence that coronary revascularization is effective in nonischaemic MR, and little is known about promising factors related to postoperative left ventricular function and survival. Thus, it does not seem easy to assess postoperative survival in the same way for nonischaemic procedures. However, in this study, postoperative survival in nonischaemic MR was good in PMR-A cases with BNP less than 1000 pg/ml (Supplementary Material, Fig. S4). In addition, LVDd also was decreased in the PMR-A group, even for nonischaemic aetiologies. Therefore, we believe that PMR-A is effective in nonischaemic patients.

The annular shape

The saddle shape of the mitral annulus has been reported to be associated with valve durability and haemodynamics after valve repair for MR [16, 17]. This study used Physio II rings, which have the effect of correcting annular dilatation and maintaining saddle shape to some extent. We have no data in this study on how PMR-A impacted the saddle shape of the annulus. However, based on previous reports, we believe that the association with saddle shape is an issue to be investigated. Further studies will be needed to clarify this topic.

Limitations

There are several important limitations to this study. The study is a single-centre, nonrandomized, retrospective study with a small patient cohort. Furthermore, PMR-A is a procedure developed based on our experience and therefore was performed later than PMR-P. We are also aware that due to several circumstances, echocardiographic data were unavailable, and some indices could not be used or analysed. Thus, we cannot preclude the possibility of bias in the study outcomes and estimates. The background of the patients for the procedures performed also differed. The present study is not an exact comparison of PMR-A and PMR-P. A larger number of patients and a study adjusting for many patient backgrounds would be needed for a rigid comparison. We would expect such adjusted studies to detect more significant differences. Because of these limitations, we understand that the results and conclusions of the study should be evaluated with caution. Large-scale prospective studies are needed to increase the level of evidence for the findings of this study.

CONCLUSIONS

Long-term survival for patients who underwent RA plus bilateral PMR-A was promising. Furthermore, patients with significantly higher BNP had lower survival after valve repair for FMR. RA plus bilateral PMR-A resulted in left ventricular reverse remodelling despite severe preoperative mitral valve leaflet tethering. These findings provide a more accurate understanding of the patient risk factors and support the efficacy of the surgical subvalvular intervention with PMR-A for FMR.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

ACKNOWLEDGEMENTS

We express our deepest gratitude to Professor Kunihiko Takahashi of the Biostatistics, Tokyo Medical and Dental University, for invaluable advice for the paper's statistical analysis and Dr. Ryo Uchimido of the Intensive Care Medicine, Tokyo Medical and Dental University, for his unconditional support for the study.

Funding

The authors declare that they have not received specific grants for this study from the public, commercial or non-profit funding agencies.

Conflict of interest: none declared.

Data Availability

The data underlying this article will be shared upon reasonable request to the corresponding author.

Author contributions

Keiji Oi: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Resources; Software; Validation; Visualization; Writing—original draft; Writing—review & editing. Hirokuni Arai: Conceptualization; Project administration; Resources; Supervision; Validation; Writing—review & editing. Eiki Nagaoka: Investigation; Project administration; Resources; Validation. Tatsuki Fujiwara: Investigation; Project administration; Resources; Validation. Kiyotoshi Oishi: Investigation; Project administration; Resources; Validation. Masashi Takeshita: Investigation; Project administration; Resources; Validation. Tatsuhiko Anzai: Data curation; Formal analysis; Methodology; Software; Validation; Writing—review & editing. Tomohiro Mizuno: Project administration; Resources; Validation.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks Ichiro Hayashi and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• ALCA

  Anterior leaflet closing angle

 
• ALOA

  Anterior leaflet opening angle

 
• BNP

  B-type natriuretic peptide

 
• FMR

  Functional mitral regurgitation

 
• LVDd

  Left ventricular end-diastolic diameter

 
• LVDs

  Left ventricular end-systolic diameter

 
• LVEF

  Left ventricular ejection fraction

 
• MR

  Mitral regurgitation

 
• PLCA

  Posterior leaflet closing angle

 
• PLOA

  Posterior leaflet opening angle

 
• PMR

  papillary muscle relocation

 
• PMR-A

  Papillary muscle relocation anteriorly

 
• PMR-P

  Papillary muscle relocation posteriorly

 
• RA

  Restrictive annuloplasty