Impact of ischemia on left atrial remodeling and dysfunction in swine models of mitral regurgitation

Abstract

Left atrial (LA) dysfunction is one of the predictive factors of worse outcomes after mitral valve surgery for mitral regurgitation (MR). We aimed to investigate the effect of MR etiology on progression of LA remodeling in swine MR models. MR was induced in 14 Yorkshire pigs using catheter-based procedures. Seven pigs underwent simultaneous occlusions of the left circumflex artery and the diagonal branch, which resulted in ischemic mitral regurgitation (IMR group). The other seven pigs underwent chordal severing to induce leaflet prolapse simulating degenerative mitral regurgitation (DMR group). Changes in LA volume and function were assessed at baseline, 1 mo, and 3 mo using echocardiography and hemodynamic evaluations. Histopathological assessments were conducted to evaluate LA hypertrophy and fibrosis. At 3 mo, quantitative MR severity was comparable and severe in both groups. Despite the similar degree of MR, minimum LA volume index increased significantly more in the IMR group (IMR: 11.9 ± 6.4 to 73.2 ± 6.4 mL/m², DMR: 10.7 ± 6.4 to 29.5 ± 6.4 mL/m², P_interaction = 0.004). Meanwhile, increase in maximum LA volume index was similar between the groups, resulting in lower LA emptying function in the IMR group (IMR: 60.1 ± 3.1 to 29.4 ± 3.1%; DMR: 62.4 ± 3.1 to 58.2 ± 3.1%, P_interaction = 0.0003). LA reservoir strain assessed by echocardiography was also significantly lower in the IMR group. Histological analyses revealed increased LA cellular hypertrophy and fibrosis in the IMR group. In conclusion, ischemic MR is associated with aggressive remodeling and reduced emptying function compared with the MR due to leaflet prolapse. Earlier intervention might be necessary for ischemic MR to prevent LA remodeling.

NEW & NOTEWORTHY We show different LA structural and functional remodeling patterns between ischemic MR and MR due to leaflet prolapse. Severe ischemic MR was accompanied by extensive LA remodeling, which may be associated with poor clinical outcomes. Our data suggest that detailed structural and functional LA remodeling assessment is important for managing IMR and to determine the presence of LA ischemia.

INTRODUCTION

Mitral regurgitation (MR) is the second most common valvular disease following aortic stenosis, and surgical or transcatheter repair is the gold standard of treatment. Left atrial (LA) enlargement can be caused by volume overload from MR and elevation in left ventricular (LV) diastolic pressure, leading to a vicious cycle in which LA enlargement causes mitral annular enlargement and exacerbates MR (1). Furthermore, LA enlargement and dysfunction are independent prognostic factors of mortality and cardiac events in patients with MR (2). However, the influence of MR etiology on LA remodeling and prognosis is not fully understood. Some studies comparing surgical outcomes of degenerative and ischemic MR concluded that MR etiology has little impact on outcomes (3, 4), whereas others suggested worse postoperative outcomes in patients with ischemic MR (5). It still remains a matter of debate as to the timing (moderate or severe) and method (repair, replacement, or transcatheter) of intervention for treating ischemic MR (6). Importantly, none of these studies has considered LA remodeling as an indicator of postsurgical outcome. Current European guidelines include LA remodeling for surgical indication only for degenerative MR, whereas American guidelines do not mention LA remodeling (7, 8).

It is estimated that 1%–42% of myocardial infarction (MI) is accompanied by atrial infarction (9). Considering established evidence for the impact of myocardial ischemia on LV structural remodeling (10, 11), atrial infarction might also trigger LA remodeling. We hypothesized that degenerative and ischemic MR show different patterns of LA remodeling progression, and LA ischemia could play an important role in this remodeling process. The aims of this study were 1) to evaluate the changes in LV and LA structural and functional remodeling in clinically relevant MR models and 2) to analyze the pathophysiology to understand the cause of LA remodeling.

MATERIALS AND METHODS

Study Design

All animal procedures were approved by the Institutional Animal and Use Committee at the Icahn School of Medicine at Mount Sinai, and the care of all animals complied with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 85-23, Revised, 1996). The animals were acclimatized at the facility for at least 72 h before being enrolled in experiments.

A total of 14 female Yorkshire pigs at 6–7 wk old (19.5 ± 2.3 kg) were included in one of the two MR groups: the ischemic mitral regurgitation (IMR) group, which included seven pigs subjected to ischemic MR (12), and the degenerative mitral regurgitation (DMR) group, which included seven pigs subjected to MR induced by mitral chordal severing (13). For all procedures, animals were premedicated using intramuscular Telazol (tiletamine-zolazepam, 8.0 mg/kg). After intubation, they were ventilated with 100% oxygen. General anesthesia was maintained with intravenous propofol (10 mg/kg/h) throughout the procedure. Buprenorphine (0.03 mg/kg) was given every time before and postprocedure for analgesia. No drugs were given after 24 h except for intramuscular antibiotics (cefazoline, 25 mg/kg), which were given for 3 days after procedures. Echocardiographic and hemodynamic data were obtained at baseline, 1 mo, and 3 mo after the MR induction. For invasive measurements, Swan-Ganz catheter was delivered to the pulmonary artery and the right heart pressures and cardiac output (thermodilution) were measured during breath hold. A high-fidelity pressure catheter (Millar Instruments, Houston, TX) was inserted into the LV via the right carotid artery to collect LV pressure measurements. Direct left atrial pressure was measured by transseptal approach using a 5-Fr catheter through the venous access. At 3 mo, after right and left heart catheterization, pigs were euthanized, and tissues were processed for histopathological assessments. Three age-matched normal pigs underwent the same echocardiographic and hemodynamic assessments and were euthanized for histopathological assessments. This was a retrospective analysis of two different MR models that have been produced in our laboratory. All animals that underwent the same experimental protocols were included.

Ischemic MR Induction: MI Creation

The protocol for percutaneous MI induction has been described previously (12). At the beginning of each MI creation experiment, 75 mg amiodarone and 10 mEq potassium acetate were mixed into a 1-L saline bag and continuously infused at a rate of 300 mL/h. Five thousand units of heparin were added after vascular sheath insertion. Intramuscular (50 mg) and intravenous (10 mg) amiodarone was added before coronary occlusion. After the procedure, 500 mL saline drip with 75 mg amiodarone and 10 mEq potassium acetate was administered overnight in the recovery room. For MI, a 7-Fr guiding catheter was advanced to the left coronary artery. Coronary angiogram before coronary occlusion identified LA branches that bifurcated from the proximal to mid-left circumflex artery (LCX) in all IMR pigs (Supplemental Fig. S1; all Supplemental material is available at https://doi.org/10.6084/m9.figshare.19253396.v1). After coronary angiography, 0.014-in. coronary wires were advanced, one in the left circumflex artery (LCX) and another in the first diagonal branch. Over-the-wire coronary balloons were delivered and inflated simultaneously at the proximal LCX and the diagonal branch to induce myocardial ischemia. To achieve transmural MI, balloon occlusion was maintained for >90 or 60 min with an embolic coil implantation at the same location as balloon inflation.

Degenerative MR Induction: Chordal Cutting

As previously reported (13), the DMR model was created by chordal severing using a myocardial biopsy catheter. Briefly, a bioptome (BIOPAL 7, Cordis, Miami, FL) was advanced to the LV and the chordae tendineae of the mitral valve apparatus were grasped under echocardiographic and fluoroscopic guidance. The chordae tendineae were severed by retracting the bioptome into the guide sheath. Sequential chordal severing with MR assessments were continued until angiographic contrast reflux entered the pulmonary veins, and/or the echocardiographic MR jet reached the LA roof, and/or the hemodynamic LV end-diastolic pressure increased to 150% of the baseline. All pigs in this group met these criteria except for one pig, which presented unclear visualization of pulmonary vein flow. Animals were treated once with furosemide (50 mg iv) post-MR induction.

Echocardiography

All pigs underwent two- and three-dimensional echocardiography at baseline, 1 mo, and 3 mo after model induction using Philips iE-33 ultrasound system (Philips Medical Systems, Andover, MA) (14). All parameters were evaluated in accordance with American Society of Echocardiography/European Association of Cardiovascular Imaging recommendations (15). LV and LA volumes and ejection fraction were measured in three-dimensional echocardiographic images using a semiautomated border detection software (Philips QLAB 3DQ Advanced, Philips Medical Systems, Andover, MA). The two-dimensional echocardiographic studies included parasternal long-axis and short-axis views to measure LV and LA dimensions. Two-dimensional four-chamber images were acquired from the subxiphoid window, and cross-sectional images of the LV were obtained from the right intercostal spaces. Reliable delineation of the entire LA endocardial border was obtained through long-axis view from the right intercostal space and was used for LA strain analysis. LV and LA strains were measured using the speckle-tracking algorithm for LV using the QLAB CMQ function. In each view, the myocardium was automatically divided by the software into six segments. The adequacy of tracking was verified and was adjusted manually if needed to achieve optimal tracking. The peak LV longitudinal and circumferential strains were measured from an average strain curve representing all analyzed segments from apical four-chamber and midpapillary short-axis views, respectively. The peak LA longitudinal strain was measured at the end of the reservoir phase, and calculated by averaging values of all LA segments. Total LA emptying function was defined as the difference between maximum and minimum LA volumes relative to the maximum volume. Body surface area (in m²) was calculated as described previously (16), and volume parameters were divided by the body surface area to calculate volume indexes. Regurgitant volume and effective regurgitant orifice area were calculated using the proximal isovelocity surface area (PISA) method based on the current guideline (17). In the four-chamber view, the color baseline was shifted toward the jet direction to obtain an aliasing velocity of 30–40 cm/s. PISA radius was measured from the aliasing velocity to the level of the valve orifice. Regurgitant fraction was calculated using regurgitant volume and forward stroke volume at the LV outflow tract.

Histopathological Assessments

At the end of the study protocol, the animals were euthanized, the hearts were explanted, and LA posterolateral walls were processed for histopathological analyses. Formalin-fixed tissues were embedded in paraffin, sectioned (8 μm), and mounted on microscope slides. The slides were then stained with fluorescein isothiocyanate-labeled wheat germ agglutinin to assess cell area as an indicator of hypertrophy, and with picrosirius red to detect collagen as an indicator of LA fibrosis. The LA tissues from three normal pigs were also examined as control. Paraffin-embedded sections were deparaffinized through 100% xylene and graded ethanol series. For picrosirius red staining, sections were stained with Bouin’s solution for 15 min, hematoxylin for 10 min, 0.1% Sirius Red in saturated aqueous solution of picric acid for 1 h. Sections were washed with 0.1 N hydrochloric acid, dehydrated with 70% ethanol, cleared with 100% xylene, and mounted. For wheat germ agglutinin immunostaining, antigen was retrieved with proteinase K for 15 min. Sections were blocked with 3% bovine serum albumin in tris-buffered saline with Tween-20 for 30 min, labeled with wheat germ agglutinin solution (Invitrogen, Waltham, MA) diluted 1:200 in phosphate-buffered saline (5 μg/mL) for 1 h and with DAPI solution diluted 1:1,000 in phosphate-buffered saline with Tween 20 (1 μg/mL), and mounted. Incubation for labeling was done under light protection.

For fibrosis analysis, seven, six, and two LA tissues were available for DMR, IMR, and control groups, respectively, and for hypertrophy analysis, seven, six, and three LA tissues were available. Fibrosis was quantified in 28, 24, and 8 fields (4 fields from each heart), respectively, for DMR, IMR, and control groups using ×100 images. Cell size (hypertrophy) was evaluated in seven, six, and three fields (1 field from each heart), respectively, for DMR, IMR, and control groups using ×100 images. All samples were stained at the same time. Quantitative analyses for fibrosis and hypertrophy were performed using image processing software (ImageJ, National Institute of Health, Bethesda, MD) with the same threshold. In the qualitative assessment of fibrosis, we followed the description by Frangogiannis et al. (18). Replacement fibrosis was defined as large amount of collagen accumulation, which occupies myocardium area accompanied by elimination of nucleus, whereas reactive fibrosis was defined as interstitial distribution in myocardium or limited accumulation in the perivascular area. Both patterns of fibrosis were included in the quantitative measurements.

Statistical Analysis

Continuous variables are presented as means and SD and were compared using Student’s t test for histopathological assessments. The Pearson’s correlation test was used for all correlation analyses. To compare the echocardiographic parameters at the different time points, the repeated-measures ANOVA was used with “time effect” as a within-subjects factor (baseline, 1 mo, and 3 mo) and “group effect” as a between-subjects factor (IMR and DMR). The least squares means and SE were used to describe the results of this analysis. All analyses were conducted with JMP v. 12.0 (SAS Institute, Cary, NC), and P < 0.05 was the criterion for statistical significance.

RESULTS

Seventeen pigs including three normal controls completed the experimental protocol. Physical data at each time point and hemodynamic data at the final time point are shown in Table 1. Body weight and surface area at baseline and 1 mo were similar between the groups, whereas those at 3 mo tended to be lower in the IMR group. Heart rate at the time of echocardiography was comparable between the groups at all time points.

Table 1.

Physiological trends, hemodynamic parameters, and tissue weight at 3 mo

	Control	IMR	DMR	P Value, IMR vs. DMR
n	3	7	7
Physical data
Body weight, kg
Baseline		19.6 ± 3.1	19.4 ± 1.2	0.92
1 mo		24.3 ± 2.2	25.1 ± 3.0	0.58
3 mo	34.3 ± 3.8	35.4 ± 4.1	40.7 ± 6.8	0.08
Body surface area, m2
Baseline		0.51 ± 0.05	0.51 ± 0.02	0.95
1 mo		0.59 ± 0.04	0.61 ± 0.05	0.59
3 mo	0.75 ± 0.05	0.76 ± 0.06	0.83 ± 0.10	0.10
Heart rate, beats/min
Baseline		77.1 ± 6.4	83.6 ± 8.2	0.13
1 mo		83.0 ± 9.5	89.4 ± 10.9	0.26
3 mo	82.3 ± 8.1	94.3 ± 16.9	90.3 ± 13.9	0.62
Hemodynamic data
Cardiac index, l/min/m2	4.8 ± 0.5	4.0 ± 1.0	4.5 ± 0.6	0.26
Mean pulmonary arterial pressure, mmHg	16.0 ± 1.7	24.1 ± 10.8	22.1 ± 5.3	0.67
Mean pulmonary capillary wedge pressure, mmHg	6.0 ± 3.5	13.9 ± 6.5	9.4 ± 5.3	0.19
Mean LA pressure, mmHg	N/A	10.3 ± 5.4	9.1 ± 6.1	0.77
End-diastolic LV pressure, mmHg	10.6 ± 0.9	17.9 ± 7.8	15.4 ± 5.5	0.50
End-systolic LV pressure, mmHg	129.9 ± 11.6	112.1 ± 18.3	103.2 ± 12.1	0.30
Maximum LV pressure, mmHg	138.7 ± 12.5	123.9 ± 18.2	110.9 ± 10.6	0.11
Maximum dP/dt, mmHg/s	2,105 ± 656	1,856 ± 354	1,786 ± 509	0.77
Minimum dP/dt, mmHg/s	−2,668 ± 307	−2,148 ± 542	−2,260 ± 376	0.66
Tissue data
LA weight, g	15.4 ± 0.8	26.3 ± 8.1	21.2 ± 4.8	0.18
LV weight, g	99.6 ± 18.3	117.7 ± 25.4	123.4 ± 10.0	0.59
RA weight, g	13.0 ± 3.4	15.1 ± 7.7	13.4 ± 2.3	0.58
RV weight, g	39.9 ± 7.0	47.2 ± 23.1	45.1 ± 15.3	0.84
LA weight/both atria	0.55 ± 0.06	0.66 ± 0.05	0.61 ± 0.04	0.04
LA weight/total heart	0.09 ± 0.01	0.13 ± 0.03	0.10 ± 0.02	0.05

Values are means ± SD. DMR, degenerative mitral regurgitation; dP/dt, first derivative of pressure over time; IMR, ischemic mitral regurgitation; LA, left atrial; LV, left ventricular; n, number of samples; N/A, not available; RA, right atrial; RV, right ventricular.

Progression of MR

A total of 14 pigs successfully completed the study protocol following MR induction. After model induction, MR grade was assessed with echocardiography. Ischemic MR animals had less than mild MR post-MI, whereas DMR animals exhibited severe MR. Longitudinal evaluation of MR grade at 1- and 3-mo time points showed that the regurgitant volume index increased similarly in both groups and were severe at 3 mo (Fig. 1). Considering the reduced LV systolic function associated with MI in the IMR group, the regurgitant fraction was also calculated. At 1 mo, the regurgitant fraction was equivalent to mild in the IMR group, whereas it was moderate to severe in the DMR group. The degree of the MR progressed to moderate to severe at 3 mo in IMR group, which became similar to the DMR group (IMR: 25.3 ± 5.5 to 44.0 ± 6.0%; DMR: 42.5 ± 5.5 to 45.6 ± 5.5%, P_interaction = 0.18).

Figure 1.

Representative echocardiographic images of degenerative mitral regurgitation (DMR; A) and ischemic mitral regurgitation (IMR; B) models and the quantitative mitral regurgitation (MR) severity at 1- and 3-mo time points (C and D). The MR severity in the IMR group increased progressively, whereas that in the DMR group remained similar at 3 mo. The MR severity at 3 mo was comparable between the groups. Data show least-square means and SEs (IMR, n = 7; DMR n = 7, all females). Individual dot plots are available in Supplemental Fig. S2.

LV Remodeling and Function

Figure 2 shows the changes in LV dimensions and function during the follow-up for each group. LV end-diastolic volume index increased in both groups over 3 mo with a larger increase in the IMR group. However, there was no significant interaction between time and group effects. The increase in LV end-systolic volume index was more pronounced in the IMR group, whereas it was relatively mild in the DMR group. Accordingly, LV ejection fraction was preserved in the DMR group, whereas it was markedly reduced in the IMR group, resulting in a significant time × group interaction (IMR: 68.3 ± 2.2, 46.5 ± 2.2, and 42.9 ± 2.2%; DMR: 71.3 ± 2.2, 74.5 ± 2.2, and 68.9 ± 2.2% at baseline, 1 mo, and 3-mo, P_interaction < 0.0001). The LV strain analysis showed obvious disparities between the two groups (Fig. 2); both longitudinal and circumferential strains worsened over time in the IMR group, whereas they were preserved in the DMR group (P_interaction = 0.005 for longitudinal strain and P_interaction = 0.004 for circumferential strain). LV diastolic function over time was assessed using E/A from pulse wave Doppler of mitral inflow. The change was not significantly different between the groups (P _interaction = 0.25; Supplemental Fig. S3).

Figure 2.

The impact of mitral regurgitation (MR) etiology on left ventricular (LV) volume and function changes. Ischemic mitral regurgitation (IMR) was associated with significantly larger LV volumes and worse systolic function relative to degenerative mitral regurgitation (DMR). P values indicate the result of time × group interaction by the repeated-measures ANOVA. Data show least-square means and SEs (IMR, n = 7; DMR, n = 7, all females). Individual dot plots are available in Supplemental Fig. S3.

LV hemodynamic data and LV weight at 3 mo are shown in Table 1. Cardiac index was not significantly different but tended to be lower in the IMR group compared with the DMR group. Pulmonary arterial wedge pressure tended to be higher in the IMR group. End-diastolic LV pressure was elevated in both groups relative to normal animals, but there were no significant differences. Postmortem LV weight was similar between the groups.

LA Remodeling and Dysfunction

Figures 3 and 4 show the changes in LA dimensions and function. Maximum and minimum LA volume indexes increased in both groups; however, the increase in minimum LA volume index was more prominent in the IMR group, resulting in a significant interaction (P_interaction = 0.004). Accordingly, similar to LV ejection fraction, LA emptying function was preserved in the DMR group, whereas it was markedly reduced in the IMR group (P_interaction = 0.0003). LA strain also exhibited preserved LA function in the DMR group relative to the IMR group (P_interaction = 0.001). At 3 mo, there were no significant differences in mean LA pressure (Table 1). LA weight was not statistically different between the groups. However, its proportion to both atria and total heart weights was significantly higher in the IMR group.

Figure 3.

The impact of mitral regurgitation (MR) etiology on left atrial (LA) dimension changes. Ischemic mitral regurgitation (IMR) was associated with significantly larger minimum LA volume index. Conversely, LA diameter and height showed similar changes between the groups. P values indicate the result of time × group interaction by the repeated-measures ANOVA. Data show least-square means and SEs [IMR, n = 7; degenerative mitral regurgitation (DMR), n = 7, all females]. Individual dot plots are available in Supplemental Fig. S4. LAVI, left atrial volume index.

Figure 4.

Representative echocardiographic image of left atrial (LA) strain (A) and the impact of mitral regurgitation (MR) etiology on LA functional changes (B and C). LA reservoir strain was measured as a peak strain of the average waveform (arrow). Ischemic mitral regurgitation (IMR) was associated with significantly reduced LA emptying function and peak reservoir strain. P values indicate the result of time × group interaction by the repeated-measures ANOVA. Data show least-square means and SEs [IMR, n = 7; degenerative mitral regurgitation (DMR), n = 7, all females]. Individual dot plots are available in Supplemental Fig. S4.

Figure 5 shows the change in mitral valve annular dimensions. The mitral annulus was similarly enlarged in both commissure-commissure (P_interaction = 0.67) and anteroposterior (P_interaction = 0.18) directions in both groups. The tethering height increased over time in both groups, but the change was more pronounced in the IMR group (P_interaction = 0.02).

Figure 5.

The impact of mitral regurgitation (MR) etiology on mitral annular (MA) and valve dimension changes. Ischemic mitral regurgitation (IMR) was associated with significantly more severe leaflet tethering, whereas annular dilatation was not significantly different. P values indicate the result of time × group interaction by the repeated-measures ANOVA. Data show least-square means and SEs [IMR, n = 7; degenerative mitral regurgitation (DMR), n = 7, all females]. Individual dot plots are available in Supplemental Fig. S5.

Histopathological Assessments

LA fibrosis and cardiomyocyte hypertrophy were evaluated in seven animals in the DMR group and six animals in the IMR group. Picrosirius red staining showed diffuse fibrosis that included replacement fibrosis and perivascular fibrosis in the IMR group. In contrast, only a limited amount of perivascular fibrosis and no obvious replacement fibrosis was seen in the DMR group (Fig. 6, A–F). The IMR group had significantly increased LA collagen area fraction relative to the DMR group (Fig. 6G). LA fibrosis was strongly correlated to LA volumes (maximum LA volume index: r = 0.79, P = 0.0004; minimum LA volume index: r = 0.78, P = 0.0006), weight, and myocyte size (Supplemental Table S1).

Figure 6.

Left atrial (LA) fibrosis. Picrosirius red staining of LA posterolateral wall tissue with low (×12.5) and high (×100) magnifications in the degenerative mitral regurgitation (DMR; A and D), ischemic mitral regurgitation (IMR; B and E), and control pigs (C and F). Myocardium was replaced with collagen fiber in IMR. Perivascular fibrosis was seen in both IMR and DMR. Fibrosis was quantified in 28, 24, and 8 fields (4 fields from each heart), respectively, for DMR (n = 7), IMR (n = 6), and control groups (n = 2) using ×100 images (G). Significantly more fibrosis was observed in the IMR group. Data show means, SDs, and results of Student’s t test. Scale bar = 750 µm in A–C, and 100 µm in D–F.

Figure 7 shows the results of the wheat germ agglutinin staining, in which the size of cardiomyocytes was measured. The DMR group showed similar myocyte size distribution to the sham samples with a minor increase in the large myocyte fraction (Fig. 7D). Meanwhile, distribution of LA myocyte size in the IMR group was shifted to the larger side with a significant difference in average cell size (Fig. 7E).

Figure 7.

Left atrial (LA) cellular hypertrophy. LA posterolateral wall tissue was stained with wheat germ agglutinin (magnification: ×400) in the degenerative mitral regurgitation (DMR; A), ischemic mitral regurgitation (IMR; B) and control groups (C). Cell size was evaluated, respectively, for DMR (n = 7), IMR (n = 6), and control groups (n = 3) using ×100 images (D). Data show means, SDs, and results of Student’s t test. IMR pigs exhibited significantly larger cell size compared with DMR pigs. Histogram shows the relative frequency of cardiomyocyte size. Distribution of LA myocyte size in the IMR group shifted to the larger side with significant difference in average cell size (E).

DISCUSSION

The main findings of this study were that 1) MR grade gradually progressed in the pig model of ischemic MR, whereas it remained unchanged in the model of degenerative MR; 2) LA functional and structural remodeling is more progressive in the ischemic MR model than the degenerative MR model; and 3) myocardial fibrosis and cellular hypertrophy were more severe in the ischemic MR model than degenerative MR model. These findings were observed under the similar MR grade at the 3-mo time point.

LA in chronic MR is affected by both pressure and volume overload (1). Under chronic MR conditions, volume overload has been proposed to be the most powerful determinant of LA enlargement (19). As a compensation, the LA becomes more enlarged and spherical as the MR grade progresses (20). In the present study, regurgitation volume index and regurgitation fraction, surrogate markers of volume overload, were equivalent in both groups at 3 mo. However, ischemic MR showed more progressive LA remodeling compared with degenerative MR. These results suggest that other factors besides volume overload due to MR influence LA structural and functional remodeling. The potential contributing factors include the degree of LV dysfunction, difference in MR jet profiles, and LA ischemia as discussed in the following.

Degree of LV Dysfunction

Ischemic MR is caused by progressive LV dilatation and papillary muscle displacement, which often occurs after posterolateral MI (21). In this study, broad posterolateral MI was created by occlusion of the proximal LCX and diagonal branches. As we reported previously (12), this method results in reproducible IMR. Accordingly, LV systolic function in the IMR group was significantly worse than the DMR group. However, end-diastolic LV pressure was not significantly different between the groups, although it tended to be slightly higher in the IMR group. Based on these data, LA afterload associated with LV dysfunction, i.e., LV diastolic pressure, was likely similar between the groups. Thus, LV dysfunction may not be the primary cause of difference in LA remodeling patterns between the groups. Recent clinical trials demonstrated different results in efficacy of transcatheter edge-to-edge repair for functional MR (22, 23). Grayburn et al. (24) discussed the concept of “proportionate/disproportionate MR” as a way to distinguish patients who would benefit from this intervention. The phenotypes were defined by the ratio of effective regurgitant orifice area to LV end-diastolic volume. Unfortunately, the present study was unable to investigate the difference in LA remodeling between proportionate and disproportionate MR due to the small sample size in the IMR group. Whether disproportionate MR results in more severe LA remodeling compared with proportional MR requires future investigation.

Difference in MR Jet

One of the major differences between these two types of MR is the direction of the regurgitant jet: in ischemic MR, it is generally a central jet associated with malcoaptation, whereas in degenerative MR, it is an eccentric jet due to leaflet prolapse. It remains unclear whether the LA wall confronted by the jet is more vulnerable to remodeling. Since both LA diameter and height showed similar increases between the groups during 3-mo follow-up, the influence of jet direction on LA shape seems minimal. However, because the direction of the eccentric jet was variable in the DMR group, further studies are needed to evaluate the impact of difference in MR jet profile. For this purpose, three-dimensional echocardiographic assessments of mitral annulus and valve might offer better morphological characterization.

LA Ischemia

Posterolateral MI occasionally involves LA infarction. Cushing et al. (25) reported a large series of atrial infarction cases by autopsy study, which showed 17% incidence of atrial infarction in patients with MI. Of note, the atria receives its blood supply from the proximal regions of the right coronary artery and LCX (26). In the present study, the LA branch was involved in all IMR group animals by the occlusion of the proximal LCX. In picrosirius red staining, the IMR group showed diffuse replacement of myocardium with collagen tissue, which suggests ischemic change, whereas the DMR group showed only limited amount of perivascular fibrosis. Given the different features of LA fibrosis, we speculate that atrial infarction has exacerbated LA remodeling in ischemic MR pigs. This result is consistent with the previous report (27), which showed that fibrosis was more severe and LA remodeling was more prominent in the pig model of ischemic MR with LA infarction than the model without LA infarction. In addition to fibrosis, both groups showed LA cardiomyocyte hypertrophy. Although the IMR group was exposed to severe MR for a shorter period than the DMR group, the hypertrophy was more prominent. This might be a compensatory mechanism for reduced LA emptying function. It is also possible that a direct hypertrophic change is induced by LA infarction. It has been reported that some hormonal factors such as angiotensin and the inflammatory process are associated with LV concentric hypertrophy after MI (28).

A novel finding of our study is that ischemic MR is associated with more severe and progressive LA remodeling compared with degenerative MR with a similar degree of regurgitation. Furthermore, we demonstrated that these changes were accompanied by significantly more severe fibrosis and cellular hypertrophy in histology. Thus, if ischemic MR is left until it becomes severe, the progression will be accompanied by extensive LA remodeling, which is the key prognostic factor of poor prognosis in patients with chronic MR (2). To date, only a few retrospective, single-center studies have reported LA remodeling as a predictor of outcomes after mitral valve surgery (29, 30). These studies are limited to degenerative MR and there is little data available for defining the role of LA remodeling in ischemic MR. Our data suggest that more frequent follow-up on both MR degree as well as progression of LA remodeling is necessary for ischemic MR compared with degenerative MR. It is noteworthy that accumulation of these data will contribute not only to better understanding of LA remodeling but also to restratifying the timing of mitral valve interventions across the spectrum of patients with MR to improve clinical outcomes.

This study has several limitations. First, we did not evaluate LA electrophysiological changes in this study. This should be studied in the future experiments. Second, the observers were not blinded to group status in histopathological analyses. However, all tissue samples were stained at the same time, and quantitative measurements were performed using the same threshold in all groups. Thus, we believe that the results have limited observer bias. Third, there was no degeneration in the leaflets in the DMR group and significant MR was suddenly induced by chordal severing. However, the situation in which the LA is exposed to volume overload due to MR with leaflet prolapse is similar to degenerative valve diseases. Finally, although this study showed that MR etiology influenced LA remodeling, the impact on prognosis remains unclear. Further clinical studies are required.

In conclusion, ischemic MR due to the occlusion of the proximal LCX involving the LA branches is associated with more progressive LA remodeling than MR associated with leaflet prolapse. Our study suggests that in patients with ischemic MR, not only MR grade and LV parameters, but also the possibility of LA infarction should be considered when deciding the timing of surgical intervention. Careful evaluation of LA remodeling patterns might contribute to improved patient prognosis.

SUPPLEMENTAL DATA

Supplemental Table S1 and Supplemental Figs. S1–S5:https://doi.org/10.6084/m9.figshare.19253396.v1.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants R01HL139963 (to K.I.) and T32HL007824-23 (to R.M. and S.M.). T.S. was supported by Japan Heart Foundation/Bayer Yakuhin Research Grant Abroad.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

T.S. and K.I. conceived and designed research; S.A.M., F.R., A.J.R., S.W., T.K., and K.I. performed experiments; T.S., R.M., S.A.M., A.J.R., S.W., T.K., and K.I. analyzed data; T.S., R.M., S.A.M., F.R., S.W., T.K., and K.I. interpreted results of experiments; T.S., R.M., and K.I. prepared figures; T.S. drafted manuscript; T.S., R.M., S.A.M., F.R., A.J.R., S.W., T.K., and K.I. edited and revised manuscript; T.S., R.M., S.A.M., F.R., A.J.R., S.W., T.K., and K.I. approved final version of manuscript.