Electromechanical coupling between the atria and mitral valve

Julia C. Swanson, Gaurav Krishnamurthy, John-Peder Escobar Kvitting, D. Craig Miller, Neil B. Ingels Jr.


Anterior leaflet (AL) stiffening during isovolumic contraction (IVC) may aid mitral valve closure. We tested the hypothesis that AL stiffening requires atrial depolarization. Ten sheep had radioopaque-marker arrays implanted in the left ventricle, mitral annulus, AL, and papillary muscle tips. Four-dimensional marker coordinates (x, y, z, and t) were obtained from biplane videofluoroscopy at baseline (control, CTRL) and during basal interventricular-septal pacing (no atrial contraction, NAC; 110–117 beats/min) to generate ventricular depolarization not preceded by atrial depolarization. Circumferential and radial stiffness values, reflecting force generation in three leaflet regions (annular, belly, and free-edge), were obtained from finite-element analysis of AL displacements in response to transleaflet pressure changes during both IVC and isovolumic relaxation (IVR). In CTRL, IVC circumferential and radial stiffness was 46 ± 6% greater than IVR stiffness in all regions (P < 0.001). In NAC, AL annular IVC stiffness decreased by 25% (P = 0.004) in the circumferential and 31% (P = 0.005) in the radial directions relative to CTRL, without affecting edge stiffness. Thus AL annular stiffening during IVC was abolished when atrial depolarization did not precede ventricular systole, in support of the hypothesis. The likely mechanism underlying AL annular stiffening during IVC is contraction of cardiac muscle that extends into the leaflet and requires atrial excitation. The AL edge has no cardiac muscle, and thus IVC AL edge stiffness was not affected by loss of atrial depolarization. These findings suggest one reason why heart block, atrial dysrhythmias, or ventricular pacing may be accompanied by mitral regurgitation or may worsen regurgitation when already present.

  • markers
  • stiffness

mitral valve closure may be aided by a brief contraction of myocytes in the annular third of the anterior leaflet (AL) (15). This contraction is also believed to be the cellular basis for the substantial AL stiffening observed during isovolumic contraction (IVC) (5). Such stiffening, observed primarily as a stiffening transient in the annular third of the AL (8), subsides by isovolumic relaxation (IVR) (9). Thus the AL stiffens at the onset of each beat then relaxes back to IVR stiffness values as left ventricular (LV) ejection proceeds.

Electrophysiological studies have shown that the AL is in electrical continuity with the left atrium (LA) (4, 16, 17) and likely depolarizes with each beat as a consequence of left atrial depolarization (3). Without a properly timed sequence of atrial and ventricular excitation (such as with heart block, arrhythmias, or LV pacing), AL stiffening may be perturbed, potentially resulting in abnormal valve function. Thus the present study tested the hypothesis that AL IVC stiffening is abolished when atrial depolarization does not precede ventricular systole in the beating ovine heart.


All animals received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by the National Institutes of Health (U.S. Department of Health and Human Services, NIH Publication 85–23, Revised 1985). This study was approved by the Stanford Medical Center Laboratory Research Animal Preview Committee, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International, and conducted according to Stanford University policy.

Although a new cohort of animals was used in this experiment, the methodology used has been previously described (6, 8) and thus will only be briefly outlined here.

Surgical preparation.

Ten adult Dorsett-hybrid sheep (48 ± 3 kg) had 13 miniature radioopaque tantalum markers surgically implanted into the subepicardium to silhouette the LV along four equally spaced longitudinal meridians (ventricular Markers in Fig. 1, left). On cardiopulmonary bypass, 35 additional radioopaque tantalum markers were sutured in place to delineate the mitral valvular-ventricular complex, one at the tip of each papillary muscle, 16 around the mitral annulus (annular markers in Fig. 1, left), 16 on the atrial aspect of the anterior mitral leaflet (Fig. 1, Right), and one on the central edge of the central scallop of the posterior leaflet (PL). A single tantalum loop was used for each leaflet marker. Through the open LA a pacing wire was secured in the basal interventricular septum and externalized through the right ventricle (Fig. 2). Animals were then weaned from cardiopulmonary bypass, stabilized, and transferred to the catheterization laboratory for open-chest data collection.

Fig. 1.

Radioopaque marker array. CC, commissurer-commissure; SL, septal-lateral orientation. (Not shown: anterior and posterior papillary muscle tip markers and marker on the middle scallop of the posterior mitral leaflet.)

Fig. 2.

Basal interventricular septal pacing sit. *Just inferior to the membranous septum.

Data acquisition.

Videofluoroscopic images (60 Hz) of all markers were acquired in multiple-beat runs using biplane videofluoroscopy. Data were collected with the heart in normal sinus rhythm (control, CTRL) and during basal, interventricular pacing (Medtronic, Minneapolis, MN) performed at a rate that suppressed the sinus node and resulted in ventricular systole that was not preceded by atrial depolarization (no atrial contraction, NAC). Marker coordinates from each view were merged to yield the 3-D coordinates of the centroid of each marker in each frame. Left ventricular pressure (LVP), left atrial pressure (LAP), aortic pressure, and electrocardiographic voltage signals were digitally recorded simultaneously during marker data acquisition and synchronized with the images.

Finite element models.

Three consecutive beats were selected for analysis for the CTRL and NAC runs in each heart. As previously described (5), cardiac cycle timing and finite-element models of the AL were individually defined for each beat using matched IVC and IVR pressure intervals for that beat. The initial geometry for each model was defined by the coordinate values for each leaflet, annular and papillary tip marker at the reference states used in previous analyses (5, 8). A leaflet surface was generated from leaflet and annular marker positions. Strut chordae originating from measured papillary tip positions were inserted into the leaflet at positions identified from postoperative anatomical photographs. Previously measured ovine leaflet thickness values were used in an orthotropic linear elastic-material model. Although the isolated leaflet demonstrates nonlinear behavior in biaxial testing, we have shown (7) that the material behavior of the leaflet in the beating heart is linear over a wide range of LVPs.

The model used the experimental displacements of the papillary muscle tips and leaflet edge and annular markers as displacement boundary conditions while allowing leaflet displacements to respond to the experimentally measured LVP and LAP values as a function of the circumferential (Ec) and radial (Er) elastic-modulus (stiffness) values entered into the model. The use of measured leaflet-edge displacements accounted for the effect of the primary chordae distribution. A heterogeneous model with three regions (annulus, belly, edge) having two independent material parameters (Ec and Er) for each region was used as previously described (8).

Inverse finite-element-analysis algorithm.

Successive stiffness values were automatically entered into the model for each of the six material parameters (Ec and Er for annulus, belly, edge), and a simulation was run using the spatial and pressure boundary conditions for IVC and IVR. Leaflet displacements in the model were compared with the actual experimental leaflet displacements, and stiffness values were iterated using an optimization algorithm and solver until minimum error (0.01%) was reached between the simulated and measured leaflet displacements. The resultant regional stiffness values (Ec and Er elastic moduli for each of the 3 predefined regions) were interpreted as the in vivo regional radial and circumferential AL stiffness values for that heart for CTRL and NAC at both the IVC and IVR time intervals.

Transient stiffness analysis.

Anterior mitral leaflet stiffness has been shown to vary over the cardiac cycle, with a drop in leaflet stiffness as systole progresses. To quantify the change in material parameters through systole, global leaflet stiffness values (Ec and Er) were identified by applying the inverse analysis to four time steps using a homogeneous leaflet model for both the CTRL and NAC runs. As in Krishnamurthy et al. (9), the four systolic time steps analyzed were IVC (ΔT1), end-IVC to midsystole (ΔT2), midpoint between end-diastole (ED) and end-systole, midsystole to IVR onset (ΔT3), and IVR (ΔT4).

Mitral valve geometry.

To evaluate the effect of NAC on mitral valve geometry, the mitral annular septal-lateral (S-L) dimension chord length between the mitral annular saddle-horn and the midlateral annular marker was calculated for each frame for three beats for each heart for CTRL and NAC. The time when mitral annular S-L dimension reduction was initiated, as the mitral valve was closing, was identified relative to ED for CTRL and NAC. The time of mitral valve closure was defined when the distance between the central edge markers on the AL and PL reached a stable minimum.

Statistical analysis.

All values are given as group means ± SD. Three-beat averages were used to characterize each variable for each heart and experimental condition (CTRL, NAC). The Shapiro-Wilk Normality Test (OriginPro 8.0; OriginLab, Northampton, MA) demonstrated that, at the 0.05 level, all variables were significantly drawn from normally distributed populations. Data were analyzed using Student's paired t-test (SigmaStat 3.5; Systat Software, San Jose, CA) to compare CTRL and NAC. The AL homogenous stiffness values at the four different time intervals during systole were characterized by the slope of the systolic stiffness vs. time and variations in AL homogeneous stiffness throughout the four time intervals during systole compared with a Student's t-test. A Bonferroni correction for multiple comparisons was performed and statistical significance noted.


Electrocardiographic tracings confirmed that ventricular depolarization preceded atrial depolarization for all NAC studies and that CTRL data were collected in normal sinus rhythm. Gross examination of all hearts after euthanasia confirmed that all markers were in the correct position, with the exception of one heart, which was missing one annular marker, which did not affect the study.

The hemodynamic changes associated with NAC are shown in Table 1. All NAC studies were paced at a rate of 110, except one was paced at 117 beats/min because lower rates did not suppress sinus node activity. The hemodynamic changes in Table 1 are those expected with increased chronotropy.

View this table:
Table 1.


Figures 3, A and B, and 4 and Table 2 show the effect of a loss of properly timed atrial excitation (NAC) on the onset of S-L dimension reduction. NAC delayed the normal onset of the presystolic decrease in S-L dimension relative to CTRL by 67 ms; CTRL S-L dimension decrease was initiated 40 ± 18 ms prior to ED, whereas the NAC S-L dimension reduction was initiated 27 ± 28 ms after ED (P < 0.001). Leaflet closure timing, relative to ED, was not altered by the delay in S-L dimension reduction with NAC (Table 2).

Fig. 3.

A: Left ventricular pressure (LVP) (mmHg) and mitral annular SL dimension (cm) during 3 cardiac cycles with the heart beating in normal sinus rhythm (control, CTRL) in a representative heart. Arrow indicates end-diastole (ED). *Time when SL dimension reduction was initiated. B: LVP (mmHg) and mitral annular SL dimension (cm) through 3 cardiac cycles with basal interventricular septum over-drive pacing (no atrial contraction, NAC) for the same heart as A. Arrow indicates ED. *Time when S-L dimension reduction was initiated.

Fig. 4.

Group mean (± SE) SL dimension for CTRL and NAC centered on ED.

View this table:
Table 2.

Onset of septal-lateral annular dimension reduction and leaflet closure

Figure 5, A and B, and Tables 3 and 4 show the effect of NAC on AL regional circumferential and radial stiffness. CTRL IVC circumferential and radial AL stiffness values were 43–64% greater than IVR stiffness values (P < 0.001) in all regions. With NAC, IVC stiffness of the annulus region of the AL decreased relative to CTRL by 25% (P = 0.004) in the circumferential and 31% (P = 0.005) in the radial directions, such that annular IVC AL stiffness was reduced to IVR stiffness values. NAC did not alter edge stiffness (P = ns).

Fig. 5.

A: anterior mitral leaflet stiffness. Circumferential (Ec) elastic modulus (stiffness), for CTRL and NAC at isovolumic contraction (IVC) and isovolumic relaxation (IVR). B: anterior mitral leaflet stiffness. Radial (Er) elastic modulus (stiffness), for CTRL and NAC at IVC and IVR.

View this table:
Table 3.

Circumferential leaflet stiffness

View this table:
Table 4.

Radial leaflet stiffness

Figure 6, A and B, shows group mean (±SD) global anterior mitral leaflet circumferential and radial stiffness for the systolic intervals ΔT1, ΔT2, ΔT3, and ΔT4. With CTRL, IVC stiffening quickly decreased to IVR values (linear regression slope Ec, −0.79 ± 0.3 N/ms-mm2; Er, −0.29 ± 0.1 N/ms-mm2). With NAC, whereas radial and circumferential stiffness was significantly different from CTRL (both P < 0.001), it was unchanged throughout systole (Ec −0.08 ± 0.1 N/ms-mm2; Er −0.01 ± 0.1 N/ms-mm2), demonstrating that delayed AL stiffening [such as might be expected if retrograde transduction through the atrioventricular (A-V) node took place and excited the AL cardiac muscle] did not occur.

Fig. 6.

A: global anterior mitral leaflet stiffness throughout systole. Circumferential (Ec) elastic modulus (stiffness) for CTRL (C) and NAC (N). See materials and methods for specifications of ΔT1–4. C vs. N P < 0.001 for Ec. B: global anterior mitral leaflet stiffness throughout systole. Radial (Er), elastic modulus (stiffness) for control and NAC. C vs. N P < 0.001 for Er.


The principal finding of this study is that early systolic stiffening of the anterior mitral leaflet requires excitation from the atrial side of the atrioventricular (A-V) node before each beat. [The term A-V node is defined here as a distributed region, reflecting that the slow A-V nodal pathway may not be confined only to Koch's triangle but also may extend around the A-V orifices (11).]

This finding, coupled with the observation that global leaflet stiffness remains constant throughout systole when ventricular depolarization was initiated from the ventricular side of the A-V node (Fig. 6, A and B), implies that the primary excitation pathway for early systolic leaflet stiffening is either 1) directly from the LA into the leaflet, or 2) indirectly from the LA, through the A-V node, and then into the leaflet. Previous in vitro studies, however, provide considerable support for the latter (indirect) pathway. Fenoglio et al. (4), studying isolated canine preparations, found that electrical stimulation of distal atrial wall was propagated through the atrial myocardium, delayed by the annulus, and then entered the AL. Conversely, when isolated preparations consisting of ventricular anterior wall or interventricular septum in continuity with the AL were tested, signal propagation into the leaflet did not occur; that is, no retrograde propagation took place from ventricle to leaflet. The findings of this present in vivo study are consistent with these in vitro findings. Curtis and Priola (3) measured leaflet depolarization coinciding with the QRS complex of the electrocardiogram and suggested that leaflet contraction might also be stimulated via a ventricular pathway. The lack of leaflet stiffening in the present study when ventricular contraction was not preceded by atrial excitation casts doubt on the existence of this pathway.

Loss of preceding atrial depolarization abolished early systolic stiffening of the annular third of the AL, reduced early stiffening of the AL belly, but did not affect early stiffening of the AL edge (Fig. 5, A and B, and Tables 3 and 4). This parallels the distribution of cardiac myocytes in the AL, dense near the annulus, progressively less dense in the belly, and absent at the edge (2, 4, 12). In a recent study, Krishnamurthy et al. (8) showed that administration of systemic β-blockade resulted in a loss of annular, to a lesser degree belly, but not edge AL stiffening during early systole; these results are very similar to those reported here. Thus the effects of both β-blockade and lack of excitation, both known to suppress myocyte contraction, are consistent with reduction in force development in AL cardiac myocytes on the basis of their regional distribution.

The regional findings also support a hypothesis previously developed in Krishnamurthy et al. (8) concerning the mechanistic basis for the early systolic stiffening of the leaflet edge. We have previously shown edge stiffening to be unaffected by β-blockade or neural stimulation (5) and in the present study now show it to be unaffected by loss of atrial depolarization, as well. The hypothesis of Krishnamurthy et al. (8) suggests that a change in shape of the AL edge, brought about by the initial contact between the AL and PL during IVC, provides geometric stiffening of the AL edge, in that subsequent pressure changes acting on this new edge geometry produce reduced leaflet-edge displacements relative to those observed with the same pressure change later in systole. The present study provides new information about this possibility. Without excitation (i.e., NAC), leaflet myocytes are not activated, so it is possible that any early systolic leaflet stiffening observed reflects geometric stiffening, a term referring to the curling of the leaflet-edge region as it makes contact with the PLs during IVC, which reduces the radius of curvature of the edge region such that the displacement of the region is reduced for a given change in pressure. If this is the case, such geometric stiffening would be greatest at the edge, less in the belly (with presumably less geometric change during leaflet coaptation because it is further from the coapting edge), and nonexistent in the annular region, with its thicker tissue even more distant from the edge; this is exactly the behavior depicted in Fig. 5, A and B. Conversely, the data in Fig. 5, A and B, also support the concept that the leaflet myocyte contribution to early systolic leaflet stiffening is just the opposite, greatest at the annulus, less in the belly, and nonexistent at the edge, because reduction of early systolic (myocyte) force development by loss of depolarization is greatest near the annulus, less so in the belly, and zero at the leaflet edge, reflecting the distribution of leaflet myocytes.

When leaflet myocyte excitation was eliminated (NAC), leaflet stiffness remained constant throughout systole (Fig. 6, A and B), providing the first direct experimental support for our concept of basal leaflet tone (5). We have previously shown (57) that leaflet stiffness is orders of magnitude greater in vivo than in vitro and have attributed this increased stiffness to some form of contractile tissue in the leaflet that is active in the beating heart but not in excised tissue. In Itoh et al. (5) we proposed that leaflet stiffness is characterized by a relatively constant basal leaflet tone brought about by leaflet contractile tissue, punctuated by a leaflet-stiffening twitch that transiently increases leaflet stiffness above the baseline tone values at the onset of each beat. The evidence is strongly converging now on the concept that the leaflet-stiffening twitch is brought about by the myocytes in the leaflet that are excited at the beginning of each beat via a pathway requiring atrial excitation. The remaining question, then, is what underlies leaflet tone. Because global stiffness (Figs. 6, A and B) and IVR stiffness (Fig. 5, A and B) were unaffected by NAC, this suggests that the contractile elements in the leaflet that impart leaflet tone must not be sensitive to cardiac conduction/depolarization. We have proposed that valvular interstitial cells within the collagen AL matrix may be responsible for this basal tone stiffness (5), as such cells are ubiquitous in the leaflet, bind to collagen, and have control mechanisms that are independent of the cardiac conduction system. Stephens et al. (13) have recently provided experimental support for the feasibility of this proposal.

In addition to abolishing leaflet stiffening [and consistent with the findings of Timek et al. (14)], presystolic annular S-L dimension reduction was also abolished when excitation was initiated on the ventricular side of the AV node. S-L annular reduction still occurred but now after the onset of the LV pressure rise during IVC (Figs. 3, A and B, and 4, Table 2). This appears to suggest a direct excitation pathway in normally conducted beats from the LA to the annular muscle, distinct from the indirect excitation pathway from the LA to the AL muscle, because annular S-L reduction normally begins before the A-V nodal delay. Thus the late S-L reduction during IVC observed here with NAC most likely reflects the forces pressing inward on the mitral annulus from the contracting LV basal myocardium coupled to the annulus. This supports the concept, proposed by Bothe et al. (1), that presystolic mitral annular contraction is normally triggered directly from the LA, when the sparse annular muscle is capable of reducing annular dimensions because the LA afterload (LVP) is very low during diastole and then the LV myocardial/annular coupling holds the annulus in the contracted state in the face of the high LVP during systole.

Unlike the findings of Timek et al. (14) in a similar ovine model, leaflet closure times in the present study were unaffected by the lack of atrial systole (Table 2). Although presystolic mitral annular S-L dimension reduction has been shown to be important for valve closure (14), NAC eliminated such presystolic S-L dimension reduction (Fig. 4) without changing closure time. Two factors may be involved in this finding. First, the S-L dimension in NAC was always considerably reduced relative to CTRL before and immediately after ED, then virtually identical to CTRL during final valve closure. In their experiment, Timek et al. (14) reported that S-L dimension increased 8% with pacing. Therefore, a reduced S-L dimension was always present during NAC in the present study, thus the leaflets were brought equivalently close together by the same S-L dimension in both CTRL and NAC during the final closure process. Second, this may reflect the importance of developed LVP (dLVP)/dtmax during IVC in the closure process. NAC and CTRL had the same LVPmax, but dLVP/dtmax was significantly increased with NAC. Theories suggesting that LV systolic function is important to the process of valve closure have a long history (10, 14). Perhaps the increase in dLVP/dtmax with NAC helped close the valve, even though the mitral annular S-L dimensions were altered. This could be the equivalent of a regurgitation test during mitral valve repair surgery, which requires a very rapid influx of fluid through the LV cannula to close and seal the valve.

Finally, Timek et al. (15) showed that ablation of leaflet myocytes in the annular region of the ovine anterior mitral leaflet slowed valve closure by some 35 ms relative to controls. This has been interpreted to suggest that the leaflet twitch may help close the valve, as suggested as a possibility by Curtis and Priola (3). The data in the present study call this concept into question, in that, in each individual heart, the S-L annular dimension was very similar between CTRL and NAC with the closure time virtually identical, yet the myocytes were active in CTRL but inactive in NAC. Rather than helping to close the valve (which is still a possibility), perhaps the most important role for variable leaflet tone and twitch stiffening is to allow the anterior mitral leaflet to maintain its very complex shape as it encounters the changing demands of the heart over a wide range of systolic and diastolic loading conditions.

This study has some limitations. One limitation of the experimental design used for this study is the change in heart rate and thus hemodynamic parameters between CTRL and NAC. A-V sequential pacing could have been used for CTRL and thus assured matched heart rates; however, it was determined that the benefit of CTRL hearts in normal sinus rhythm outweighed this drawback. This limitation is also offset by the fact that the finding of interest requires comparison of stiffness within a single beat (IVC vs. IVR) at matched pressure changes. In CTRL all normal-sinus-rhythm beats demonstrated that all AL regions were stiffer during IVC compared with IVR. Similarly, all NAC beats demonstrated loss of annular AL IVC stiffening such that it resulted in no difference between IVC and IVR stiffness. We have previously shown (7) that leaflet stiffness is not a function of LVP, which alters the geometry of both the leaflet and the LV. Thus changes in leaflet geometry attributable to LV pacing should not influence the results of this study.

The three AL regions were determined on the basis of the marker positions that were spaced to span the area of the leaflet and thus were not constant from leaflet to leaflet on the basis of histological landmarks. However, because all the animals were of similar weight, relative heart size was similar and leaflet marker placement was relatively constant. Finally, the findings reported herein are from ovine heart, and further work is warranted before these findings can be applied more broadly.


This work was supported in part by the National Institutes of Health grant R01 HL67025, Western States Affiliate American Heart Association Fellowship to J. Swanson, Stanford Bio-X Graduate Student Fellowship to G. Krishnamurthy, US-Norway Fulbright Foundation, the Swedish Heart-Lung Foundation, and the Swedish Society for Medical Research to J.-P. Kvitting.


No conflicts of interest, financial or otherwise, are declared by the authors.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
View Abstract