Current Affiliation: Stanford University, Department of Radiology, Stanford, CA, United States.
Impaired contractile reserve in severe mitral valve regurgitation with a preserved ejection fraction†
Article first published online: 31 AUG 2007
Published on behalf of the European Society of Cardiology. All rights reserved. © 2007 the Authors
European Journal of Heart Failure
Volume 9, Issue 9, pages 857–864, September 2007
How to Cite
McGinley, J. C., Berretta, R. M., Chaudhary, K., Rossman, E., Bratinov, G. D., Gaughan, J. P., Houser, S. and Margulies, K. B. (2007), Impaired contractile reserve in severe mitral valve regurgitation with a preserved ejection fraction. European Journal of Heart Failure, 9: 857–864. doi: 10.1016/j.ejheart.2007.05.013
This project has been supported, in part, through a grant from the Ottawa Heart Institute.
- Issue published online: 31 AUG 2007
- Article first published online: 31 AUG 2007
- Manuscript Accepted: 17 MAY 2007
- Manuscript Revised: 13 APR 2007
- Manuscript Received: 1 MAR 2007
- valvular disease;
- mitral valve regurgitation;
- myocyte physiology;
- left ventricular trabecular physiology;
- calcium transients;
- heart failure
Impaired contractile reserve in chronic MR results from load-independent, myocyte contractile abnormalities.
Investigate the mechanisms of contractile dysfunction in chronic mitral valve regurgitation (MR).
Mild MR was produced in eight dogs followed by pacing induced left ventricular (LV) dilatation over eight months. In-vivo LV dP/dt was measured at several pacing rates. Contractile function was measured in isolated LV trabeculae and myocytes at several stimulation rates and during changes in extracellular [Ca2+]. Identical studies were performed with six control dogs.
Chronic MR resulted in a preserved ejection fraction with decreased dP/dt (p<0.01). LV trabeculae demonstrated significantly lower developed force and a negative force–frequency relation with chronic MR (p<0.05). Myocytes exhibited a negative shortening-frequency relationship in both groups with a greater decline with chronic MR (p<0.001) paralleled by decreases in peak [Ca2+]i transients. Increases in extracellular [Ca2+] abrogated the defects in force generation in trabeculae from animals with chronic MR.
Even with a preserved EF, chronic severe MR results in a significant reduction in intrinsic contractile function and reserve. Functional impairment was load-independent reflecting a predominant defect in calcium cycling rather than impaired peak force generating capacity due to myofibrillar attenuation.
Chronic mitral valve regurgitation (MR)  produces changes in cardiac loading conditions resulting in volume overload. Persistent volume overload eventually leads to congestive heart failure with left ventricular (LV) pump dysfunction and an inability to maintain forward stroke volume . Progression to symptomatic congestive heart failure, caused by a cycle of cardiac remodeling and decreased pump function, is associated with a one-year survival of less than 50% .
In-vivo measurements of LV function demonstrated, in patients with severe chronic MR, a reduced velocity of shortening during ejection . Animal models of volume overload have shown that acute MR causes minimal effects on cardiac output . In a canine model, Carabello et al. demonstrated chronic MR was associated with an increase in end diastolic volume and LV mass and a decrease in the velocity of fiber shortening and end systolic elastance [6,7]. The contractile defects of chronic MR are not limited to global pump dysfunction. In studies by Urabe et al., cellular contractile abnormalities were identified by a decreased load-velocity relationship in isolated cardiac myocytes from dogs with chronic MR that correlated with a decreased volume fraction of myofibrils within cardiac myocytes . In a separate study from the same laboratory, essentially normal myocyte relengthening dynamics were reported after correction for defects in shortening . Beyond these two studies and a relatively crude earlier study reporting decreases in intracellular calcium [Ca2+]i, there is little direct evidence defining a cellular mechanism for reduced contractility in chronic MR. Moreover, the few studies examining in-vitro contractile dysfunction have mainly focused on basal function with little attention to contractile reserve. Finally, virtually no studies have directly related in-vivo measures of contractility to in-vitro metrics that more clearly reflect intrinsic myocardial adaptations rather than load-dependent functional changes.
Accordingly, the purpose of these studies was to examine the hypothesis that, despite a preserved in-vivo ejection fraction, there is a load-independent defect in myocardial contractile reserve in chronic MR. At the same time, we sought to examine whether defects in contractile reserve, if present, were most consistent with defects in calcium homeostasis or myofibrillar attenuation, as previously suggested . For these studies, we employed a recently described, clinically relevant, canine model of chronic MR with significant LV hypertrophy yet a preserved resting ejection fraction .
Upon arrival at the University, eight mongrel dogs (weight=21.5±1.0 kg) underwent a physical exam and were observed for a minimum period of one-week. One day prior to surgical experimentation, the animals were sedated and an echocardiographic analysis was performed [see echocardiographic methods]. After an overnight fast, the animals were sedated with acepromazine (0.1–0.2 mg/kg] and anesthetized with intravenous propofol (5–10 mg/kg). Animals were intubated and maintenance anesthesia was provided with a 1.5% isoflurane and oxygen mixture. The animals were instrumented with 7.5-Fr pulmonary artery catheter, a 7-Fr manometer-tipped LV catheter, and a specially adapted permanent pacemaker (Prodigy 8162, Medtronic, Inc., Minneapolis, MN) connected to a right atrial pacing lead, as previously described . Baseline measurements of LV pressures, positive dP/dt, and negative dP/dt were obtained. Following a 2-minute stabilization period, the atrial pacing rate was sequentially incremented to the following rates: 120, 140, 160, 180, 200, and 220 beats/min (bpm) with 2 min of pacing at each rate. LV pressures and dP/dt were measured at each rate when a steady state had been achieved. The pacer was then deactivated to the native rhythm.
2.1. Creation of MR
During visualization via transesophageal echocardiography, a flexible grasping forceps was introduced and advanced across the aortic valve to grasp the chordae tendinae of the posterior MV and induce a structural insult to the sub-valvular mitral apparatus, as previously described . Following a 30-minute stabilization period, the Millar catheter was reinserted and the above in-vivo pacing protocol and measurements repeated. All catheters were removed, vessels were repaired with 7–0 proline sutures, and the skin closed with 3–0 Vicryl sutures. The animals were monitored in the vivarium under the supervision of a veterinary physician and staff in accordance with the guidelines of the American Physiological Society. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1985).
2.2. Study protocol
All animals were monitored clinically and with serial echocardiograms at two week intervals to track changes in LV geometry for a period of up to 10 months. Beginning one month after chordal rupture, an incremental rapid pacing protocol was initiated as previously described . Full echocardiographic and invasive hemodynamic measurements with in-vivo force-frequency challenges were performed at four distinct time points: prior to creation of experimental MR [Baseline], immediately following the creation of MR (Acute MR), at the conclusion of the five to seven month incremental pacing protocol (Peak pacing), and six weeks after rapid pacing had been stopped (Chronic MR). At the conclusion of the Chronic MR measurements, the heart was arrested in diastole using a cold, blood-based, high-potassium perfusion solution and explanted for in-vitro studies using isolated LV trabeculae and isolated LV cardiac myocytes.
2.3. Echocardiographic measurements
All data reported was obtained using a 5 MHz trans-thoracic probe [S4] connected to a Hewlett Packard Sonos 5500. All data were obtained independently, in triplicate, at each individual time point with no significant difference between the measured parameters ensuring reproducibility of the measurements. LV geometries were obtained using M-mode short axis images at the mid-papillary muscle level. The ejection fraction (EF) was calculated using M-mode LV dimensions and the following formula:
where LVEDD represents LV end-diastolic dimension and LVESD represents LV end-systolic dimension. Color Doppler long-axis imaging of the LV and LA was used to image the regurgitant flow. The severity of MR was measured using the proximal isovelocity surface area method (PISA).
2.4. Isolated trabeculae studies
The cardiac apex was dissected from the explanted heart and placed in Krebs-Haenseleit buffer (KHB) solution containing 20 mmol/L 2,3-butanedione monoxime (BDM) and 0.25 mmol/L CaCl2 gassed with 95% 02-5% CO2. Thin (<500 μm), non-branching, free-running trabeculae from the LV free wall were carefully removed with a small cube of tissue on either side and mounted in a specially-designed tissue bath as previously described . The KHB solution was replaced by BDM-free KHB solution and the bath CaCl2 was increased stepwise from 0.5 to 2.5 mmol/L over 15 min. Trabeculae were equilibrated for 45 min while being stimulated continuously at 0.5 Hz with 3 ms pulses at an energy 20% above threshold. After returning to a slack length, the muscles were stretched to a length (Lo) where a developed force was first identified. After defining the length where maximal force development occurred (Lmax), muscle length was set at 80% Lmax (0.8×(Lmax−Lo)). After an additional 30-minute equilibration period, steady state twitches were established (0.5 Hz, 37.0 °C, 2.5 mmol/L Ca2+), and the frequency of stimulation was increased sequentially from 0.5 Hz to 1.0 Hz, 1.5 Hz, 2.0 Hz, and 2.5 Hz. Measurements of developed force (dF), positive dF/dt, negative dP/dt, and the relaxation time to 50 (RT50) and 90 (RT90)% were obtained at each stimulation frequency. At the conclusion of in-vitro force-frequency experiments, responses to graded increases in bath [Ca2+] (1.75, 2.5, 5.0 and 10.0 mM) were assessed. The isometric preload of 80% Lmax was maintained throughout the study.
2.5. Isolated myocyte studies
Immediately following cardiectomy, the LV free wall was dissected from the heart and its arterial supply cannulated for perfusion digestion as previously described . Briefly, this included a 10 min non-recirculating rinse calcium, a 30–40 recirculating digestion with collagenase, and a 10 min non-recirculating rinse. The digested midmyocardium was minced, filtered through a 300 μm nylon mesh, centrifuged (25 g) and placed in a KHB resuspension solution containing 10 mmol/L taurine, 1% (wt/vol) BSA and 200 μM CaCl2. All isolated myocyte experiments were conducted within 6 h of isolation.
Using an inverted microscope, quiescent, rod-shaped isolated myocytes were field stimulated and contractions measured with a video-based edge detection technique as described previously . The myocytes were field stimulated at 0.2 Hz until contractions reached a steady state. Measurements of fractional cellular shortening , the average rate of shortening (+ dS/dt), and the time to 50% relaxation (RT50) were obtained. The stimulation frequency was then increased and steady state measurements obtained at the following rates: 0.5, 1.0, 1.5, 2.0, and 2.5 Hz. A subset of myocytes from the Chronic MR animals and age-matched controls were loaded with Fluo-3AM and [Ca2+]i transients were measured as previously described .
2.6. Statistical analysis
The dependent variables such as LVEDD, LV Mass, EF, and MR severity, were treated as continuous variables for all analyses. Means, standard deviations, and number of observations are presented for each variable. The experimental unit was each individual animal. The experiment used a repeated measures design with each animal evaluated at up to five periods (baseline, Acute MR, Peak Pacing, and Chronic MR). The null hypothesis was that there would be no difference between periods. Prior to analysis, all data were tested for normality using the Shapiro-Wilk test. The data was significantly non-normal for all variables. In order to apply ANOVA methods, a ‘normalized-rank’ transformation was applied to the data. The rank-transformed data was analyzed using a generalized linear model ANOVA for repeated measures followed by multiple comparisons to detect significant mean differences between ‘periods’. Multiple pair-wise comparisons used the Dunn-Bonferroni adjustment to maintain an experiment-wise type I error of 0.05 or less. Differences between means [rejection of the null hypothesis] were considered significant if the probability of chance occurrence was <0.05 using two-tailed tests.
3.1. In-vivo adaptations
As shown in Table 1, there was marked dilation and hypertrophy of the LV from baseline to Chronic MR. Acute MR was relatively mild in severity and associated with minimal changes in cardiac geometry. With incremental rapid atrial pacing over five to seven months, there was progressive LV dilation, substantial hypertrophy and worsening MR. These changes persisted by six weeks after cessation of rapid pacing (Chronic MR), however a modest decrease in LV ejection fraction returned towards normal by the end of the experimental protocol. Fig. 1 depicts the results of in-vivo assessments of contractile reserve via acute increases in atrial pacing rates during evolving Chronic MR. Prior to MR, there was a progressive increase in positive dP/dt as the pacing rate was incremented (slope=8.69 mm Hg/s, Fig. 1A). With Acute MR, the values of dP/dt at each pacing rate were significantly higher than the baseline measurements (p<0.01). The slope of the dP/dt versus rate relation with Acute MR however was decreased from Baseline (slope=5.64 mm Hg/s). By the Peak Pacing phase, dP/dt was depressed at the lowest pacing rate (p<0.01 vs. Baseline) and exhibited only a minimal increase at higher pacing frequencies (slope=3.88 mm Hg/s, p<0.01 vs. Baseline). Depressed contractility and contractile reserve persisted after cessation of pacing in Chronic MR, as indicated by the significantly lower values of dP/dt at each pacing rate (p<0.01) and the decreased responses to pacing (slope=1.65 mm Hg/s, p<0.01 vs. Baseline). Changes in in-vivo relaxation reserve, as reflected by − dP/dt, are presented in Fig. 1B. Although − dP/dt and relaxation reserve (change in − dP/dt with higher pacing rates) did not change significantly with Acute MR, Peak Pacing and Chronic MR were associated with slowed relaxation and impaired relaxation reserve, with − dP/dt reduced at each pacing rate (p<0.01).
|LVEDD (cm)||LV mass (g)||EF||MR severity (cm3/s)|
|Pacing||5.79±0.24 *||228±56 *||48±7 *||76.4±19.6 *|
|Chronic MR||5.61±0.30 *||261±54 *||57±5 *||89.0±33.6 *|
3.2. Isolated trabeculae studies
As shown in Fig. 2, when studied at a constant preload (80% Lmax), developed force and + dF/dt were lower at every rate of stimulation above 0.5 Hz in trabeculae isolated from hearts with Chronic MR compared with those obtained from normal control hearts. Similarly, − dF/dt was lower at all rates in trabeculae from Chronic MR animals. Moreover, these intergroup differences tended to be more apparent as stimulation frequency increased consistent with decreased contractile reserve in the myocardium from the Chronic MR animals.
Fig. 3 depicts the results of studies in which isolated trabeculae were exposed to graded increases in bath [Ca2+]. At each stimulation frequency (0.5, 1.0, 1.5 and 2.0 Hz) the reduced developed force (DF) in Chronic MR could be largely overcome by supraphysiological increases in bath [Ca2+]. In these experiments, the peak force development achieved at high extracellular [Ca2+] did not differ between the trabeculae from Chronic MR and normal control hearts.
3.3. Myocyte studies
As shown in Fig. 4, isolated myocyte fractional shortening was lower at every rate of stimulation above 0.5 Hz in hearts with Chronic MR compared with normal control hearts. Moreover, the slightly negative shortening vs. frequency relationship observed with the control group was more pronounced in the myocytes from the Chronic MR hearts. This decrease in contractile reserve was most apparent for + dS/dt where the control myocytes exhibited increases in shortening velocity and the Chronic MR myocytes exhibited decreases in shortening velocity with increased stimulation frequency (Fig. 4B). With respect to diastolic function, the RT50 did not differ between control and Chronic MR myocytes. Indeed, rate-dependent decreases in RT50 tended to be greater in the Chronic MR myocytes [Fig. 4C].
Rate-dependent changes in parameters derived from isolated myocyte [Ca2+]i transients are presented in Table 2. In normal control myocytes, there is a modest decrease in peak [Ca2+]i and a steady decrease in the time to 50% decay of the [Ca2+]i transient (TR50Ca) as stimulation frequency increases from 0.2 to 0.5 Hz. In the Chronic MR myocytes, peak [Ca2+]i is lower even at 0.2 Hz, and exhibits a precipitous further decrease at higher stimulation frequencies. In contrast, the rate-dependent decreases in the TR50Ca are nearly identical to those observed in the normal control myocytes. For both the normal control myocytes and the Chronic MR myocytes, the changes in the [Ca2+]i transient amplitude closely parallel the changes in isolated myocytes shortening observed with increases in stimulation frequency. However, analysis of the rate of increase in calcium fluorescence (+ dF/dt) and the time constants for the initial and terminal portions of the [Ca2+]i transient decay  demonstrated no significant difference between groups.
|Stimulation frequency (Hz)||Number||Calcium fluorescence||RT50Ca (s)||Number||Calcium fluorescence||RT50Ca (s)|
The purpose of this study was to determine if there are intrinsic, load-independent, contractile defects in hearts with chronic mitral valve regurgitation with a normal ejection fraction. For this inquiry, we utilized a new animal model that mimics the patient with chronic severe MR, LV dilation and a preserved ejection fraction . In this model, we observed defects in contractile reserve can be demonstrated with a simple in-vivo provocative maneuver (incremental pacing). Our in-vitro studies using isolated trabeculae and myocytes affirm impaired contractile reserve is the most reliable indicator of the intrinsic contractile defects present in myopathic hearts and demonstrate contractile defects are apparent regardless of preload and afterload. Our cell- and tissue-based studies further demonstrate that abnormalities in myocyte [Ca2+]i transients closely parallel defects in defects in basal and rate-dependent shortening and that high bath [Ca2+] can largely overcome defects in force generation in Chronic MR. These findings indicate, for the first time, that impaired contractility and contractile reserve in Chronic MR is mainly due to defects in myocyte [Ca2+]i homeostasis rather than reduced myofibrillar content.
Our model of mitral chordal disruption and temporary incremental pacing produces a gradual progression of cardiac remodeling and MR severity that mimics the patient with chronic severe MR, LV dilation and a preserved LV ejection fraction. Even though our model demonstrated a preserved resting EF, Chronic MR was associated with impaired basal contractility as measured by + dP/dt which is less load-dependent than EF. Abnormal in-vivo contractile reserve, manifested during a short-term incremental pacing challenge, were even more pronounced in Chronic MR and contrasted sharply with the increases in dP/dt at all pacing rates observed before and immediately after chordal disruption. Previous studies have also demonstrated preserved contractile function with acute MR . One implication of these findings is that in-vivo assessment of contractile reserve, via an incremental pacing or dobutamine challenge, may be a clinical useful means of clarifying the degree of intrinsic contractile dysfunction among patients with severe MR and preserved ejection fractions or other states with distorted in-vivo loading conditions.
One might wonder whether in-vivo defects in dP/dt in the setting of severe MR are due to an inability to pressurize the LV because of emptying into a compliant left atrium. However, in some animals in which initial chordal disruption inadvertently induced severe MR, we still observed in-vivo increases in dP/dt during Acute MR [data not shown]. More importantly, the strong concordance of our in-vivo findings with the defects observed during analogous load-controlled in-vitro studies suggest the abnormal in-vivo responses are indeed due to defects in intrinsic contractility and rate-dependent contractile reserve. Trabeculae obtained from animals with chronic severe MR demonstrated a significant reduction in dF and dF/dt at each frequency compared to the control group. Furthermore, chronic MR resulted in a decline in the force-frequency relationship. These results indicate with the elimination of preload and afterload variations, impaired contractility is demonstrated at the tissue level similar to the changes measured in-vivo following chronic MR. Similar to previous studies in isolated cardiac myocytes , unloaded myocytes from control and chronic MR animals demonstrated a negative shortening-frequency relationship, with significantly lower values obtained from the chronic MR myocytes. The rate of shortening (dS/dt), demonstrated a positive frequency relation in the control group and a prominent negative relationship in the chronic MR group. These findings are consistent with previous studies demonstrating reduced shortening magnitudes and decreased rates of shortening in hypertrophied/failing myocytes [12,15–19].
Our studies also provide new insights into the mechanism of impaired contractility and contractile reserve in the setting of Chronic MR with LV dilation and a preserved EF. Our observations that defects in isolated myocyte shortening are closely correlated with abnormalities in [Ca2+]i transients strongly support the hypothesis that abnormal calcium homeostasis is the primary mechanism for abnormal contractility and contractile reserve in Chronic MR. Interestingly, this association of abnormal rate-dependent contractile reserve with abnormal calcium homeostasis in Chronic MR is similar to our observations in failing human hearts with low in-vivo ejection fractions [11,13]. Previous studies demonstrating benefits of beta-adrenergic antagonists in Chronic MR  are also consistent with a central role for abnormal calcium homeostasis because previous studies suggest that beta-blocker therapy enhances expression of sarcoplasmic reticulum calcium ATPase [21–23]. Although previous studies in a similar animal model implicated reductions in myocyte myofibrillar content as a mechanism of impaired contractility , our finding that defects in force generation in isolated trabeculae can be overcome with increases in bath [Ca2+] do not support reduced myofibrillar content as a dominant mechanism for impaired contractility in Chronic MR. However, possible alterations in myofilament Ca sensitivity are suggested by the observation that cellular relengthening is markedly slowed at higher stimulation frequencies even though there is essentially normal rate-dependent decreases for the time to 50% decay of the in [Ca2+]i transient in Chronic MR.
Each animal used in this study did not follow the exact same time course of pacing. This variation in experimental study however did not limit our findings. Predetermined criteria of LV geometry and symptoms were used to define the absolute experimental length thereby eliminating bias. Nevertheless, the final state of Chronic MR with severe valvular regurgitation, significant LV dilation and preserved EF is not infrequent in clinical settings.
Although LVEF was significantly lower with chronic MR compared to baseline values, the EF values at this stage remained within the normal range. This is likely due to the high preload and low effective afterload tending to increase EF in the setting of severe MR. In a clinical setting, where the baseline EF may not be available for a paired comparison, analogous findings might be interpreted as an absence of significant contractile dysfunction. From this perspective, compared with more load-independent measures of contractile performance, our findings confirm the relative insensitivity of EF as a measurement of function in the setting of significant MR.
Given that in-vitro studies were performed at least sixnweeks after cessation of pacing and LVEF improved during this period, it is likely the intrinsic contractile dysfunction we observed is a reflection of the underlying myopathic state in a severely hypertrophied heart, rather than a specific residual effect of tachypacing alone. No measure of myocardial contractility is ideal. In-vivo dP/dt measurements in the LV are subjected to changes in preload and afterload. At the same time, in vitro studies could be affected by the cellular isolation procedure, the absence of preload in isolated myocytes, or isometric conditions used in isolated trabeculae. To compensate for these known limitations, unlike any previous study, we performed our functional assessments of contractility, contractile reserve and relaxation dynamics at cell, tissue and organ levels.
4.2. Clinical implications
The present studies provide insights into why patients with a relatively preserved LVEF often develop reductions in ejection fraction following mitral valve repair or replacement. Our data suggests that decreases in preload and effective afterload following technically successful valve surgery unmask intrinsic contractile defects present in these severely hypertrophied hearts. The more severely myopathic hearts with decreased contractile reserve are not able to compensate for these hemodynamic changes resulting in a post-operative reduction in ejection fraction and contractile dysfunction. Thus, dynamic testing on myopathic hearts by invasive or non-invasive techniques  may guide clinical decision making by better defining the contractile state of the myocardium. This approach to preoperative prognostication has already been validated in two small studies in patients with mitral regurgitation [25,26] and an analogous approach has been validated for patients with low gradient aortic stenosis . In patients with Chronic MR and dilated hearts, pre-operative assessment of contractile reserve might help optimize the timing of valvular replacement or establish the need for perioperative mechanical support or preoperative interventions designed to improve myocyte calcium homeostasis prior to elective surgical intervention.
Funding was provided by the University of Ottawa Heart Institute for the development of the animal model used in this manuscript.
- 2Heart Disease: A textbook of cardiovascular medicine 1988 Philadelphia WB Saunders
- 3Hosenpud J. Greenberg B. Overview of pathophysiology of clinical heart failure Congestive heart failure: pathophysiology, diagnosis, and comprehensive approach to management 1994 New York Springer-Verlag
- 19Cellular and molecular alterations in the failing human heart Basic Res Cardiol 1992 87 Suppl 1 331 342