Relief of Mitral Leaflet Tethering Following Chronic Myocardial Infarction by Chordal Cutting Diminishes Left Ventricular Remodeling

Abstract

Background

One of the key targets in treating mitral regurgitation (MR) is reducing the otherwise progressive left ventricular (LV) remodeling which exacerbates MR and conveys adverse prognosis. We have previously demonstrated that severing two second–order chordae to the anterior mitral leaflet relieves tethering and ischemic MR acutely. The purpose of this study was to test whether this technique reduces the progression of LV remodeling in the chronic ischemic MR setting.

Methods and Results

A posterolateral MI was created in 18 sheep by obtuse marginal branch ligation. After chronic remodeling and MR development at 3 months, sheep were randomized to sham surgery (control group, n=6) versus second-order chordal cutting (n=12, half with anterior leaflet and half with bileaflet chordal cutting, both of which are techniques in clinical application). At baseline, chronic infarction (3 months), and follow-up at a mean of 6.6 months post-MI (sacrifice), we measured LV end-diastolic and end-systolic volume (EDV and ESV), ejection fraction (EF), wall motion score index (WMSi), and posterior leaflet (PL) restriction angle relative to the annulus by 2D and 3D a echo. All measurements were comparable among groups at baseline and chronic MI. At sacrifice, AntL and BiL chordal cutting limited the progressive remodeling seen in controls. LVESV increased relative to chronic MI by 109±8.7% in controls, versus 30.5±6.1% with chordal cutting (p<0.01) (LVESV=82.5±2.6ml in controls vs. 60.6±5.1ml and 61.8±4.1ml). LVEDV increased by 63±2.0% in controls vs. 26±5.5% and 22±3.4% with chordal cutting (p<0.01). LVEF and WMSi were not significantly different at follow-up among chordal cutting and control groups. MR progressively increased to moderate in controls but decreased to trace-to-mild with AntL and BiL chordal cutting (MR vena contracta 5.9±1.1mm in controls, vs. 2.6±0.1mm vs 1.7±0.1mm vs, p<0.01). BiL chordal cutting provided greater PL mobility (decreased PL restriction angle to 54.2±5.0° versus 83±3.2° with AntL chordal cutting, p<0.01).

Conclusions

Reduced leaflet tethering by chordal cutting in the chronic post-MI setting substantially decreases the progression of LV remodeling with sustained reduction of MR over a chronic follow-up. These benefits have the potential to improve clinical outcomes.

Mitral regurgitation (MR) more than doubles the risk of late death after both myocardial infarction and coronary revascularization ^1,2. Post-infarction, leaflet closure is restricted by tethering to displaced papillary muscles (PMs) ^3-7, with a prominent bend in the basal anterior leaflet (“seagull sign”) and markedly limited posterior leaflet motion (Carpentier functional classification Type 3b) ^8-10. It is therefore incompletely treated by annular reduction, especially when preoperative posterior leaflet angulation relative to the mitral annulus is increased^9,11,12.

Chordae tendinae inserting on the anterior mitral leaflet are commonly classified into two groups: finer marginal or first-order chordae that position the leaflet tips and prevent prolapse, and thicker basal, second-order strut chordae that insert symmetrically onto the central ventricular surfaces of both leaflets nearer their bases (Fig. 1)¹³ After inferior infarction, the PMs are displaced laterally, apically and posteriorly, pulling the leaflets down into the LV cavity, with the greatest leaflet distortion mediated by these basal chordae, leading to the recent proposal to cut a limited number of basal chordae, initially those attached to the anterior leaflet, to relieve the apical leaflet tenting and improve coaptation and MR ^14,15 (Fig. 2A, left). A more comprehensive approach will also involve the basal chordae inserting on the posterior leaflet in order to improve PL mobility (Fig. 2A, right, Fig. 2B).

Figure 1. Mitral valve chordae classification.

Chordae tendinae are classified as finer marginal or first-order chordae that position the leaflet tips and prevent prolapse, and thicker basal, second-order strut chordae that insert symmetrically onto the ventricular surface of both leaflet bases.

Figure 2. Chordal cutting techniques.

2A: Left: Anterior leaflet basal chordal cutting can eliminate the anterior leaflet bend, improve coaptation, and reduce MR; the marginal chordae prevent prolapse. Right. Bileaflet chordal cutting eliminates the anterior leaflet bend and improves PL leaflet mobility with decreased PL angle relative to the annulus.

2B: Basal chordae to both anterior and posterior leaflets (AL, PL), with anterior leaflet bend indicating tethering in an echocardiographic long-axis view

Since chordae are known to preserve left ventricular (LV) - mitral continuity and function in the normal heart ^16-19, some concerns have been raised regarding any possible influence of the chordal cutting technique on regional and global LV function ²⁰. For this reason, we began with a minimal approach by severing only the two most centrally positioned basal chordae to the anterior leaflet responsible for its bend that decreases coaptation surface. This technique was effective in acute ischemic MR with a follow up of up to 11 months ¹⁴ and acutely in a chronic ischemic MR model without any negative effect on LV remodeling and function ¹⁵. We also recently demonstrated in normal ovine beating hearts the absence of negative effects of this minimal approach on segmental and global LV function evaluated invasively (elastance, preloadrecruitable stroke work, dP/dt) and noninvasively (3D LV ejection fraction (EF) and volumes, and wall motion)²¹. Since this technique is already being tested clinically in multiple surgical centers ^22-26, it is important to test whether or not chordal cutting exacerbates long-term LV remodeling when applied to treat ischemic MR in a chronic myocardial infarction (MI), which is the purpose of this study.

We tested the minimal approach (anterior leaflet basal chordae) and a more comprehensive approach with bileaflet chordal cutting which severs all the secondary chordae to both leaflets, as in the current clinical application of David, Borger, and colleagues ^23,25,26. These approaches were tested in a chronic ischemic MR sheep model with long-term follow-up using 2D, 3D and Doppler echocardiography to quantify LV remodeling and MR.

METHODS

The chronic infarction model of Llaneras et al. ^27,28 produces chronic ischemic MR after 8 weeks of LV remodeling post-MI (occlusion of second and third circumflex marginals). In that model, 2D and 3D echo were repeated to evaluate LV remodeling, EF, wall motion score, mitral valve geometry and MR at three stages: acute infarction with only trace MR; chronic infarction with moderate MR, 3 months post acute MI; and at sacrifice at a mean of 6.6 months post MI or 3.6 months post chordal cutting or sham surgery (Fig. 3).

Figure 3. Study Design.

The overall study design involves repeated 3D and 2D echo to evaluate LV remodeling and MR with chronic infarction and moderate MR, and at sacrifice, 6.6 months post MI (3.6 months after intervention).

Animal Studies

Eighteen Dorsett hybrid sheep (40-50 kg) anesthetized with thiopentothal (0.5 ml/kg), intubated and ventilated at 15 ml/kg with 2% isofluorane and oxygen and given glyco-pyrrolate (0.4 mg IV) and prophylactic vancomycin (0.5 g IV) underwent sterile left thoracotomy, with procainamide (15 mg/kg IV) and lidocaine (3 mg/kg IV followed by 2 mg/min) given 10 minutes before coronary ligation. After baseline imaging, the pericardium was opened and the second and third circumflex obtuse marginal branches ligated to infarct the inferoposterior wall. Imaging was repeated and the thoracotomy closed. After a mean of 3 months, each animal had a second thoracotomy under general anesthesia to evaluate the ischemic MR due to mitral leaflet tethering at the chronic MI stage.

Sheep were then randomized into three groups:

• -Control group (n=6) underwent cardiopulmonary bypass and left atrial incision but without any chordal or leaflet modification.
• -Anterior leaflet (AntL) second-order chordal cutting group (n=6): After cardiopulmonary bypass, left atrial incision was performed and the anterior mitral leaflet everted through the annulus. All the basal chordae to the anterior leaflet were cut proximal to their leaflet insertion. After repair of the atrial incision, rewarming and defibrillation, normal circulation was restored (Fig 2A, left).
• -Bileaflet (BiL) second-order chordal cutting group (n=6): After cardiopulmonary bypass and left atrial incision, all the second-order chordae to the anterior leaflet were severed, as were as all the second-order chordae to the posterior leaflet after eversion of that leaflet to identify the basal chordae arising from the PMs or the posterior wall (Fig. 2A, right).

3D LV volume and MR evaluation

Thirty rotated LV apical views were acquired (5 MHz epicardial probe, Acuson Sequoia C512) with suspended respiration, as previously described and validated against sonomicrometry ^5,29,30. 3D LV volumes were obtained using endocardial borders from 9 views ²⁹. MR stroke volume was calculated as LV ejection volume minus aortic outflow volume directly measured by flowmeter ³¹. Regurgitant fraction was calculated as (MR stroke volume)/(forward aortic + MR stroke volumes). MR was also graded using the vena contracta width in the 2D longaxis view as the narrowest portion of the color jet at or just downstream from the regurgitant orifice. The largest diameter during systole was measured in at least 3 cardiac cycles and averaged. Vena contracta width ≤2 mm was considered trace to mild, ≥5mm as moderate, and >2mm and <5mm as mild to moderate ^32-34.

2D wall motion analysis

Scoring of regional contractility

Using three short-axis views, the LV was divided into 17 segments according the American Society of Echocardiography recommendations and a score allocated to each segment according to its contractility as (0) normokinetic, (+1) hypokinetic, (+2) akinetic, or (+3) dyskinetic. A global wall motion score was derived from the algebraic sum of these values, and the wall motion score (WMS) index was obtained by dividing WMS by the total number of scored segments (n=17) ³⁵ .

Mitral valve geometry evaluation

Mitral valve configuration was assessed in midsystole using the parasternal long-axis and apical 4 chamber views. The posterior leaflet (PL) angle reflecting restricted closure due to tethering was calculated because of its prognostic significance recently demonstrated by Magne ⁹ _et al.:PLA=sin^-1 (CD/PLL), where PLA = PL angle, CD = coaptation distance from annular line to leaflet coaptation point, and PLL = PL annulus–to–tip length.

Acquisition of digital cineloops and analysis of segmental wall motion score and MR stroke volume were performed by a single experienced operator blinded to the treatment group. Reproducibility of echocardiographic measurements was tested by two independent observers and gave a variability of 1.5% of the mean.

Statistical analysis

Echocardiographic measurements were compared among stages and sheep by 2-way ANOVA for repeated measures. Significant ANOVAs were explored by paired t-tests (chronic ischemia vs. sacrifice), with significance at p≤0.01 (Bonferroni-corrected). Differences in progression among groups were explored by unpaired t-test with significance at p<0.05. All data were reported as mean ± SEM. Statistical analysis used Stat View 7.0.

RESULTS

Experimental sheep survival rate

Twenty-one animals survived out of 33 after the initial myocardial infarction by obtuse marginal ligation. During the second surgery (extracorporeal circulation), 3 animals died out of the remaining 21: one could not be weaned from bypass and 2 others died a few days after surgery due to pericardial tamponade. The remaining 18 sheep completed the study: 6 with sham surgery and 6 each with anterior or bileaflet chordal cutting.

Progression of LV remodeling (Fig. 4, Table)

Figure 4. LV remodeling progression.

All measurements were comparable among groups at baseline and chronic MI. At sacrifice, AntL and BiL chordal cutting limited the progressive remodeling seen in controls with less increase of LVEDV (upper left) and of LVESV (upper right) but with no adverse effect on LV EF (bottom left) or WMSi (bottom right).

Table.

Echocardiographic Measurements.

	CHRONIC MI			SACRIFICE
	CONTROL	ANT LEAFLET	BI LEAFLET	CONTROL	ANT LEAFLET	BI LEAFLET
HEART RATE	99±1.5	99.1±2.7	98±3.2	102.8±1.6	103.3±4.0	101.5±1.9
LVEDV (mL)	76±5.3	71±2.5	76±3.4	[117.8±4.6*	93.6±5.3*	92.5±4.6*]Ψ
LVESV (mL)	40±2.4	44±3.9	48±1.6	[82.5±2.6*	60.6±5.1*	61.8±4.1*]Ψ
LV EF %	37.7±2.5	38±4.3	37±0.1	31.8±1.5	36.2±1.7	36.3±1.4
WMSi	.97±.01	.95±0.2	.93±0.9	.98±.1	.96±0.1	1.08±0.1
PL ANGLE °	80±3.5	79±2.7	80.5±1.1	[83.5±4.0	[83±3.2	54.2±5.0*]#]Ψ
MRSV (mL/beat)	9.6±3.2	9.1±1.8	9.2±1.2	[12±2.4*	[7±2.0*	5±1.4*]#]Ψ
MRRF %	25±3.2	27±1.0	24±2.6	[33.5±2.4*	[18±1.3*	11±3.1*]#]Ψ
MR Vena contracta mm	4.3±0.1	4.5±1.1	4.9±1.0	[5.9±1.1*	[2.6±0.1*	1.7±0.1*]#]Ψ

LVEDV: LV end-diastolic volume. LVESV: LV end-systolic volume. WMSi: wall motion score index

PL angle: posterior leaflet angle to the mitral annulus in midsystole.

MRSV: MR stroke volume. MRRF: MR regurgitant fraction.

Significant changes p≤0.01 (Bonferroni-corrected) are indicated for the comparisons:

^*sacrifice vs. chronic MI

^Ψchordal cutting groups vs. control at sacrifice

^#bileaflet vs. anterior leaflet chordal cutting at sacrifice

At baseline, the mitral leaflets closed at the annular level with normal LV size and function and no or trace MR. With acute infarction, LV dilatation was limited, and the leaflets still closed at the annular level, with only trace MR. After mean of 3 months post infarction, however, LV volumes were considerably higher. Chronic inferior ischemia produced mild bulging of the affected wall and displaced the PM tip away from the annulus; the leaflets remained apically tented with a discrete angulated bend in the anterior leaflet between its basal portion and the rest of the leaflet, becoming convex toward the LV, and a strongly restricted PL angle relative to the mitral annulus, forming almost a right angle (80±0.05°), with mild to moderate MR. At sacrifice, a mean of 3.6±0.6 months after the mitral valve interventions, AntL and BiL chordal cutting limited the progressive remodeling seen in sham–operated controls. LVESV increased relative to the 3-month chronic MI stage by 109±8.7% in controls vs. 33±7.2% and 28±5.0% with AntL and BiL chordal cutting (p<0.01). LVEDV increased by 63±2.0% in controls vs. 26±5.5% and 22±3.4% with AntL and BiL chordal cutting (p<0.01). There were no significant differences among control and chordal cutting groups at sacrifice in LVEF or WMSi (p=0.12 and p=0.12, respectively).

MV geometry and regurgitation (Fig. 5, Table)

Figure 5. Mid-systolic apical 2D echo images.

Control, Left: Top: Leaflet apical tenting relative to the annulus with a prominent bend in the basal anterior leaflet (“seagull sign”) and markedly limited posterior leaflet motion with PL angle relative to annulus of 80°. Bottom. Moderate MR with a central jet into the left atrium. AntL chordal cutting, Center. Top: More concave anterior leaflet (toward the LV) but with still restricted motion of the posterior leaflet and PL angle of 75°, (Bottom) with trace to mild MR. BiL chordal cutting, Right. BiL chordal cutting, Top: Less LV remodeling with a decrease of PL angle to 45,° more concave anterior leaflet and no MR (Bottom).

MR progressively increased to moderate in controls but decreased to trace-mild (vena contracta ≤ 2mm) in 83% of chordal cutting sheep, with a decrease in MR regurgitant fraction from 27±1.0 % to 18±1.3 % (p<0.01) with anterior leaflet chordal cutting vs. 24±2.6% to 11±3.1% (p<0.01) for bileaflet chordal cutting. At sacrifice, the decrease of MR was more pronounced in the bileaflet vs. anterior leaflet chordal cutting group (p=0.04). This result was consistent with the greater PL mobility provided by the bileaflet chordal cutting technique (decreased PL restriction angle to 54.2±5° versus 83±3.2° with AntL chordal cutting, p<0.01). There was no prolapse or chordal rupture in either group. The severity of LV remodeling at sacrifice correlated well with the degree of MR (r²=0.83, p<0.01).

DISCUSSION

Previous results of in vitro and in vivo models of acute and chronic ischemic MR have shown that tethering and the resulting malcoaptation can be relieved by basal chordal cutting ^14,15. Those studies demonstrated no adverse effect on segmental and global LV function after a follow-up of up to 11 months after chordal cutting to the anterior leaflet done at the initiation of an inferior infarction model and done acutely in a chronic ischemic MR model ^14,15. To answer concerns regarding the possibility of decreased LV function and increased LV remodeling with this maneuver ²⁰, the subsequent logical step was to determine whether or not chordal cutting exacerbates long-term LV remodeling when applied to treat chronic ischemic MR. This study demonstrates that cutting the secondary chordae in the chronic post MI setting does not adversely affect longterm LV remodeling and in fact limits the otherwise progressive increase in LV volumes without such intervention.

LV remodeling regression

The significant decrease in LV volume progression with stable EF in the two chordal cutting groups vs. the control one is consistent with the recently published data by Beeri ^36,37 et al. showing that MR increases LV remodeling and volume in the post MI setting with an aggravation of the morphological, functional, cellular, and extracellular stigmata of remodeling. MR can cause as well as result from LV remodeling, and can therefore potentially exacerbate remodeling in a vicious cycle spiraling down to cardiac failure unless there is intervention; in contrast, repairing MR decreased LV remodeling progression with reduced LV volumes, better maintained contractility, and improved molecular correlates of remodeling³⁶.

Consistent with this result, the chordal cutting group showed decreases in the LV remodeling that was otherwise progressive, with an increase of LV ESV of 33±7.2% and 28±5.0% with AntL and BiL chordal cutting vs. 109±8.7% in the control group (p<0.01). This can be explained by the decrease in MR progression, which plays a role in reducing this remodeling, although this remains an area of active research and controversy.

Stabilization of the remodeling LV is therefore not only achieved by localized patching of the infarct zone³⁸ but also by chordal cutting to reduce ischemic MR.

Ant Leaflet vs. Bi Leaflet chordal cutting

Because post-MI LV remodeling displaces both papillary muscles (even the one which is not involved by the infarction) as the entire LV dilates we cut all the basal chordae arising from the 2 papillary muscles in the AntL chordal cutting group.

Interestingly, cutting the basal chordae to the Ant Leaflet or all the basal chordae to both leaflets does not increase LV remodeling. This could be explained by the intact marginal chordae continuing to assure mitral-LV continuity with a less taut and tethered valve which can coapt in a more normal configuration, while repairing MR in this controlled ovine model decreased LV remodeling progression, with reduced LV volumes.

This confirms the finite element experimentation by Kunzelman and Cochran ³⁹ who noted that stress borne by marginal chordae exceeds that carried by the basal ones for any strain, with almost twice as many marginal as basal chordal insertions; and their suggestion that “it may be possible surgically to remove basal chordae without seriously compromising mitral valve function”.

Furthermore, chordal tension may in fact decrease as, over time, diminished MR stabilizes or reduces LV volume as seen in our study, and the leaflet assumes a more normal, less taut configuration (decreased leaflet radius of curvature decreasing tension by Laplace's Law). In a smaller LV, total stress may be less even if a greater proportion must be borne by remaining chords. This improved MV configuration was already apparent in prior studies on chronic ischemic MR in which the anterior leaflet bend disappeared after the two most central basal chordae to the anterior leaflet were severed ¹⁵.

As the posterior leaflet motion restriction also decreases the surface of coaptation ^8,9, we compared the efficacy and safety of bileaflet chordal cutting vs. anterior leaflet alone. The results confirmed that in the case of important LV remodeling a complete section of all the basal chordae, especially the ones inserted on the posterior leaflet, increased PL mobility and decreased PL angle relative to the mitral annulus. This angle recently emerged as a strong predictor of MR persistence after MV annuloplasty and of patient 3-year prognosis ⁹. In this experimental study, bileaflet chordal cutting resulted in a better improvement of mitral valve configuration (especially PL angle) and a slightly more important reduction in MR without increasing LV remodeling.

This is of importance and confirmed by recent human studies performed by Borger, David ²³ et al., who compared the efficacy of a full chordal cutting (arising from the ischemic papillary muscle inserted on both leaflets) + partial annuloplasty vs. partial annuloplasty alone. This study showed a significant decrease of recurrent MR at 2 years follow up (37% in annuloplasty vs 15% in chordal cutting group, p=0.03) without any decrease in the post op increase of EF (10±5% in chordal cutting vs. 11±6% in the control group; p=0.9). In that study as well as our experimental one, no late prolapse or additional chordal rupture was noticed.

Mitral regurgitation

The goal of this study was not primarily to analyze the longterm effect of chordal cutting on ischemic MR regression but rather to assure the safety of the procedure on the LV remodeling process in the perspective of ongoing clinical application of this technique.

MR progression did significantly regress in both chordal cutting groups vs. control with MR, going from moderate to trace-to-mild at 3.6 months after the MV repair. This decrease was mildly but significantly more pronounced in the bileaflet group vs ant leaflet group and correlated well with a more mobile PL, which assured better coaptation.

Limitations and future direction

The first limitation of our study is by definition that our conclusion is strictly only applicable to our chronic ischemic MR sheep model. However the MR mechanism (leaflet tethering) in our experimentation was substantially identical to that observed in clinical practice; and our results are consistent with other human studies ^23,25,26 showing the absence of increase of LV remodelling after chordal cutting.

Moreover in this study and in the human study of Borger et al. ²³, chordal cutting leaves still some degree of residual MR. We are therefore conducting an experimental ovine study using the same model of chronic ischemic MR to define the best strategy to completely eliminate MR completely at long-term follow up, with the association of annuloplasty and chordal cutting. The clinical spectrum of ischemic MR includes widely varying location and chronicity of ischemia, PM tip geometry, and potentially leaflet length. It would therefore also be reasonable to pursue future experimental studies of chordal cutting in models of more global chronic LV dysfunction, with anterior and posterolateral infarction, and of more severe MR, in which a combined approach addressing both chordae and annulus should be necessary. It is interesting to point out that in the recent paper of Agricola, Alfieri et al ⁴⁰, the authors described 2 patterns of leaflet tethering that are essentially determined by the type of LV remodeling (local inferior vs. global remodeling). The most common pattern in all instances, especially following annuloplasty, was PL restriction ^12,41with or without anterior leaflet restriction, suggesting the possibility that bileaflet chordal cutting, which most strongly mobilized the PL, will be also effective in decreasing MR in more global LV remodeling as well, in which both leaflets are restricted.

Another limitation was the noninvasive evaluation of global LV dysfunction. As these experiments required three thoracotomies, including one for MI creation and another with survival extracorporeal circulation for MV chordal cutting, and therefore entailed a high attrition rate (>45%), we did not add invasive PV loop studies to minimize operative time, procedures and further mortality.

Conclusion

Cutting secondary chordae in the chronic post-MI setting does not adversely affect long-term LV remodeling and limits progressive increases in LV volumes. This study confirms the long-term safety of this technique and its availability as a strategy to relieve ischemic MR.