Formula-Based Relocation of the Papillary Muscle in Patients With Ischemic Mitral Regurgitation

Abstract

Background

We developed a formula to estimate the optimal distance between the posterior papillary muscle (PPM) and the midseptal fibrous annulus (D_P-A) to facilitate PPM relocation in patients with ischemic mitral regurgitation (IMR) undergoing mitral valve repair.

Methods

D_P-A was measured in 109 patients without valvular disease undergoing electrocardiogram-gated 3-dimensional computed tomography, and a predictive formula that best approximated this value was developed. This formula was evaluated in 10 patients (59.2 [SD, 10.7] years, 7 men) with moderate-severe IMR undergoing mitral valve repair using PPM relocation by 3 surgeons in 6 institutions from 2014 to 2023. Two arms of CV-3 suture fixed at the PPM were passed through the midseptal fibrous annulus to the annuloplasty ring and tied at the length estimated using the formula (35 [SD, 4] mm).

Results

Postoperative echocardiography demonstrated neither residual regurgitation nor leaflet tenting. Three patients died and 1 underwent mitral valve replacement during follow-up (71 [SD, 35] months), but IMR was well controlled, and left ventricular reverse remodeling and improved left ventricular function was observed in the other 6.

Conclusions

This formula facilitates mitral valve repair using PPM relocation in patients with IMR.

Visual Abstract

In Short.

• ▪A formula to estimate the optimal distance between the posterior papillary muscle (PPM) and midseptal fibrous annulus was developed based on 109 healthy individuals.
• ▪This formula was applied in 10 patients with moderate-severe ischemic mitral regurgitation undergoing mitral valve repair using PPM relocation.
• ▪Mitral regurgitation was well controlled in all except 1 patient during follow-up, suggesting that this formula facilitates mitral valve repair using PPM relocation in patients with ischemic mitral regurgitation.

Mitral valve repair is not superior to coronary bypass grafting alone or mitral valve replacement in patients with ischemic mitral regurgitation (IMR).1, 2, 3, 4, 5, 6 IMR is a disease not of the mitral valve but of the left ventricle (LV), and therefore, mitral regurgitation (MR) may be only a surrogate marker not causally related to clinical outcome. Therefore, “mitral complex” should be optimized in cases of IMR to increase the clinical benefit of mitral valve repair.

The primary mechanism underlying IMR is tethering of the mitral leaflet due to apical displacement of the posterior papillary muscle (PPM) tip. Kron and colleagues⁷ proposed relocating the PPM to the adjacent mitral annulus just posterior to the right fibrous trigone. Langer and Schäfers⁸ refined this technique by relocating the PPM toward the midseptal fibrous annulus (saddle horn) because the intervening distance (D_P-A) plays a key role in the pathogenesis of IMR. Their refinement included adjusting D_P-A under transesophageal echocardiographic guidance in the loaded, beating heart. However, this technique has some inherent limitations. The optimal length of “string” is subjective and depends on LV volume. An extra aortotomy is also required, and bleeding risk remains a major concern. This approach may be more appropriate when optimal D_P-A can be estimated preoperatively and fixed within the cardiac chamber.

We measured D_P-A in 109 individuals using 3-dimensional computed tomography (3D-CT) and determined standard D_P-A using a formula based on the patient’s baseline characteristics. Here, we introduce this formula and report clinical experience with its application.

Material and Methods

Ethics Statement

This multi-institutional collaborative study was centrally reviewed and approved by the Jikei University School of Medicine Ethics Committee (#35-383[12020]). After approval on March 11, 2024, the directors of each institution gave permission for use of data. Recent patients provided written informed consent, but the requirement was waived for earlier patients and opt-out information was given.

Establishment of Formula

Retrospectively selected for this study were 109 patients without valvular heart disease examined between May and December 2014 at the Cardiovascular Institute.

The patients were a mean age of 60 (SD, 11) years, and 74 (68%) were men. Their mean height, body weight, body surface area, and body mass index (BMI) were 165 (SD, 9) cm, 66 (SD, 14) kg, 1.7 (SD, 0.2) m², and 24 (SD, 4) kg/m², respectively. Indications for 3D-CT were chest pain in 66, electrocardiographic abnormalities in 27, preoperative screening in 5, postcoronary intervention follow-up in 4, and other causes in 7. Percutaneous coronary artery intervention was performed in 7 patients.

Exclusion criteria were cardiomegaly, abnormal LV wall motion, more than mild MR, prior cardiac surgery, atrial fibrillation/flutter, echocardiographic studies ≥4 months before CT, and inadequate visualization of PPM.

All patients underwent 3D-CT (Brilliance64; Philips Healthcare), and D_P-A was measured at end-systole using an Aquarius workstation (TeraRecon) (Figure 1). LV end-diastolic and systolic volumes (LVEDV and LVESV, respectively) were also measured. The means of 3 independent measurements were recorded. Mean D_P-A, LVEDV, and LVESV were 36 (SD, 5) mm, 97 (SD, 23) mL, and 41 (SD, 14) mL, respectively.

Figure 1.

(A) Measurement of the distance (green line) between the posterior papillary muscle (PPM) (red circles) and the midseptal fibrous annulus (blue circles) using 3-dimensional computed tomography. (B) Relation between estimated distance between PPM and the midseptal fibrous annulus (D_P-A) calculated by the formula and the actual D_P-A measured using 3-dimensional computed tomography. Box and whisker plot: The horizontal line in the middle of each box indicates the median; the top and bottom borders of the box mark the 75th and 25th percentiles, respectively, the whiskers mark minimum and maximum of all the data, and the circles indicate outliers.

Within 3 months before the CT measurement, LV end-diastolic and systolic diameters (LVEDD and LVESD, respectively), and LV ejection fraction (LVEF) were also measured by echocardiography (means: 46 [SD, 4] mm, 28 [SD, 4] mm, and 0.68 [SD, 0.06], respectively). Sex differences in body size and LV size are summarized in Supplemental Table 1.

On Pearson multiple regression correlation analysis using the stepwise method, D_P-A was best approximated as follows:

$${\mathrm{D}}_{\mathrm{P}-\mathrm{A}}\left(\text{mm}\right)=15.528+\left(0.213\times \text{LVEF}\right)+\left(0.071\times \text{LVESV}\right)+\left(5.229\times \text{Sex}\ast \right)\left(\ast \text{male}=1,\text{female}=0\right)\left(\mathrm{R}=0.561,{\mathrm{R}}^2=0.315,\mathrm{P}<.001\right)$$

Age, sex, height, body weight, body surface area, BMI, LVEDD, LVESD, LVEDV, LVESV, and LVEF were used as covariates. Actual D_P-A determined by CT was correlated with estimated D_P-A calculated using the above formula as follows (Figure 1):

$$\left(\text{Actual}{\mathrm{D}}_{\mathrm{P}-\mathrm{A}}\right)=3.553{\mathrm{e}}^{-14}+1.000\times \left(\text{estimated}{\mathrm{D}}_{\mathrm{P}-\mathrm{A}}\right)\left(R=0.561,R^2=0.315,P<.001\right)$$

Technique

A Teflon-pledgeted CV-3 expanded polytetrafluoroethylene suture (W. L. Gore & Associates) was fixed at the tip of the PPM (Figure 2A) and marked at the estimated D_P-A (Figure 2B). The medial head was selected in cases where the PPM had several heads.⁹ This was performed through the left ventriculotomy in a patient undergoing LV thrombectomy. Both arms of the CV-3 suture were passed through the midseptal fibrous annulus from the LV side to the left atrial side. When mitral annuloplasty was performed, CV-3 sutures were also passed through the annuloplasty ring but left untied (Figure 2C). After completion of mitral valve repair, the CV-3 suture was tied at the marked point (Figures 3; Video 1).

Figure 2.

(A) A Teflon-pledgeted CV-3 expanded polytetrafluoroethylene suture was fixed at the tip of the posterior papillary muscle (PPM) (yellow arrow). (B) A CV-3 suture was marked at the estimated distance between PPM and the midseptal fibrous annulus (D_P-A) from the PPM (yellow arrow). (C) Both arms of the CV-3 suture were passed through the midseptal fibrous annulus from the left ventricular side to the left atrial side and passed through the annuloplasty ring. It was finally tied carefully at the marked point (yellow arrow). (D) Intraoperative transesophageal echocardiography demonstrated a well-functioning CV-3 suture improving leaflet tenting (yellow arrow).

Figure 3.

(A) Schematic diagram shows the CV-3 suture (arrow) running inside the left ventricle. (B) Schematic diagram shows the CV-3 suture passing through the annuloplasty ring.

Results

Study Population

We applied this formula in 10 patients with moderate to severe IMR due to tethering (mean tenting height: 11 [SD, 3] mm) undergoing elective mitral valve repair by 3 surgeons at 6 institutions between June 2014 and January 2023. The patients were a mean age of 59.2 (SD, 10.7) years, and 7 were men. Their mean body surface area was 1.71 (SD, 0.26) m² and 2 patients were on long-term hemodialysis. One had a history of coronary artery bypass grafting. Preoperative mean LVEF, LVEDD, LVESD, LVEDV, and LVESV were 0.37 (SD, 0.12), 62 (SD, 6) mm, 51 (SD, 7) mm, 180 (SD, 53) mL, and 115 (SD, 42) mL, respectively. D_P-A was estimated as 35 (SD, 4) mm using the formula. Details of the etiology of ventricular failure are summarized in Supplemental Table 2.

Surgical Technique

In all patients, undersized annuloplasty was performed with a full ring sized 29 (SD, 3) mm. Additional coronary artery bypass grafting and tricuspid ring annuloplasty were performed in 6 patients each. Papillary muscle approximation was applied in the first 2 patients, but only 2 heads of the PPM were approximated in the second patient. LV thrombectomy, aortic valve replacement, closure of atrial septal defect, and maze procedure were performed in 1 patient each. Mean operation, cardiopulmonary bypass, and aortic cross-clamp durations were 345 (SD, 1140), 172 (SD, 54), and 111 (SD, 29) min, respectively. The procedure for PPM relocation alone took ∼7 minutes. Intraoperative transesophageal echocardiography after weaning from cardiopulmonary bypass demonstrated neither residual MR nor leaflet tenting in any patients (Figure 2D). Observed D_P-A was similar to the intended length.

Clinical Outcomes

All patients survived the operation, but 3 died of stroke or unknown causes at 24, 33, and 58 months postoperatively. One patient underwent mitral valve replacement 19 months postoperatively due to recurrent mitral leaflet tethering caused by a persistent increase in LVESD from 52 to 61 mm. The remaining 6 patients are currently undergoing outpatient care 71 (SD, 35) months after surgery, and MR is well controlled to mild or less in all patients. LV reverse remodeling was already observed at 1 year after surgery when 1 patient undergoing mitral valve replacement was excluded (n = 8). LVEF improved from 0.37 (SD, 0.13) to 0.45 (SD, 0.14) (P =.105), LVEDD and LVESD were significantly reduced from 63 (SD, 6) to 57 (SD, 7) mm (P = .003) and from 52 (SD, 8) to 45 (SD, 9) mm (P = .015), respectively (paired t test). No relevant stenosis was observed (mean pressure gradient: 4.0 [SD, 0.9] mm Hg). The follow-up rate for the 6 patients, excluding 3 who died and 1 who underwent aortic valve replacement, was 100%, with a mean follow-up period of 63 (SD, 28) months. At the latest follow-up, they were further improved to 0.45 (SD, 0.13) (P =.257), 54 (SD, 8) mm (P = .045), and 41 (SD, 9) mm (P = .060), respectively (Figure 4).

Figure 4.

Left ventricular (LV) reverse remodeling and improved LV function at the latest follow-up (FU) at 63 (SD, 28) months (M) after surgery in 6 patients. (A) LV ejection fraction (LVEF), (B) LV end-diastolic diameter (LVEDd), and (C) LV end-systolic diameter (LVEDs). (Preop, preoperative.)

An example of a case is shown in Video 2. A 63-year-old man with moderate IMR and reduced LV function (LVEDD, 66 mm; LVESD, 56 mm; LVEF, 0.31) underwent mitral valve repair using formula-based PPM relocation (estimated D_P-A= 36 mm). Although MR was well controlled, he died suddenly 58 months after the operation. Routine echocardiography performed 55 months after surgery showed markedly deteriorated LV function due to further enlargement of the LVESD (LVEDD, 64 mm; LVESD, 59 mm; LVEF, 0.17); however, MR was completely controlled due to PPM relocation.

Comment

Our formula to estimate D_P-A could facilitate less invasive, more objective, and predictable relocation of the PPM independent of LV volume load. Clinical application could be performed safely, and the midterm outcomes were favorable.

This study had some limitations. First, there was a risk of entanglement between CV-3 sutures and the anterior leaflet chorda, which could be prevented by exerting lateral traction on the chorda. In addition, the formula was developed from measurements in a small number of Japanese patients with small body size, and its applicability in other populations is unclear. Sex was included as a factor because there are marked sex-related differences in body size and cardiac geometry between men and women (Supplemental Table 1). In particular, D_P-A was more than 5 mm shorter in a small-sized female patient than in male patients. However, we have applied this formula in 2 other patients with dilated cardiomyopathy (LVEDD 78 and 81 mm, respectively) and confirmed positive findings.

This study also lacked a control group. However, in a large, sophisticated, randomized trial involving patients with severe IMR,⁵ the rate of recurrence of moderate or severe MR over 2 years was 58.8% after mitral valve repair in which subvalvular procedures were used in only 11.9% of cases. IMR was well controlled in all patients in our study, except 1 who underwent mitral valve replacement. Although 3 patients died during follow-up, no MR recurrence occurred.

The purpose of this study was not to examine the effectiveness of the formula but to test the safety of its clinical application. We therefore applied this formula in nonselected patients, and additional procedures may affect alleviation of functional MR. Nonetheless, we believe this formula-oriented relocation of the PPM may be useful after further modifications. This pilot study will form a basis for work by other groups, including those in Western countries, and the formula should be modified depending on the severity of IMR (eg, tenting height) and confirmed by a prospective study evaluating preoperative and postoperative D_P-A in a larger cohort in comparison with a control group over a longer follow-up period.