Background
The discrepancy between planimetered mitral valve area (MVA) and mean diastolic pressure gradient (MDPG) has not been studied extensively in patients with mitral stenosis. The purpose of the present study was to investigate differences in characteristics and outcomes after mitral valve replacement (MVR) between low- and high-MDPG groups in patients with very severe mitral stenosis (VSMS). The hypothesis was that the low-MDPG group would have different characteristics and would be associated with poor clinical outcomes after MVR.
Methods
In total, 140 patients who underwent isolated MVR because of pure VSMS (planimetered MVA ≤ 1.0 cm 2 ) were retrospectively reviewed, and follow-up echocardiography was performed for ≥12 months after MVR. Patients were divided into two groups according to preoperative MDPG (low gradient [LG], <10 mm Hg; high gradient [HG], ≥10 mm Hg). Strain and strain rate analysis was performed using speckle-tracking echocardiography of the left ventricle before MVR in a subgroup of 56 patients.
Results
There were 82 patients (59%) in the LG group and 58 patients (41%) in the HG group. The LG group was older and demonstrated a higher prevalence of female gender, diabetes mellitus, and atrial fibrillation ( P < .05 for all). When comparing the LG and HG groups, the left atrial volume index was larger (105.1 ± 51.9 vs 87.8 ± 42.9 mL/m 2 , P < .001), and strain rate during isovolumic relaxation of the left ventricle was lower (0.17 ± 0.08 vs 0.29 ± 0.09 sec −1 , P < .001) in the LG group. After MVR, the percentage left atrial volume index reduction after MVR was significantly smaller in the LG group (−29.9 ± 15.1% vs −43.5 ± 16.4%, P < .001). Persistent symptoms after MVR were more common in the LG group compared with the HG group ( P = .004), even though preoperative functional class was similar between the groups.
Conclusions
Compared with those with HG VSMS, patients with LG VSMS were older, more often female, and more frequently had diabetes mellitus and atrial fibrillation. They also had greater impairment of isovolumic relaxation, less favorable left atrial reverse remodeling, and a greater risk for persistent symptoms after MVR. These data might suggest other concurrent mechanisms for left atrial enlargement and symptom development in LG VSMS, such as atrial fibrillation and diastolic dysfunction, as well as valvular stenosis.
Mitral stenosis (MS) is the narrowing of the mitral valve orifice, and very severe MS (VSMS) is defined as mitral valve area (MVA) ≤ 1.0 cm 2 . Hemodynamic severity is usually characterized by two-dimensional (2D) planimetry and calculated from diastolic pressure half-time. However, diastolic pressure half-time overestimates MVA in patients with impaired left ventricular (LV) compliance, because it is dependent on the degree of mitral obstruction as well as the compliance of the left ventricle. Mean diastolic pressure gradient (MDPG) is reliably assessed by Doppler echocardiography, but this is not considered the best marker of MS severity, because it is dependent on MVA along with other factors that influence transmitral flow rate, heart rate, cardiac output, and associated mitral regurgitation. Therefore, the severity of MS is determined by integrating all parameters, such as MVA, MDPG, and pulmonary artery (PA) systolic pressure.
MVA ≤ 1.0 cm 2 usually corresponds to an MDPG of >10 mm Hg at a normal heart rate in patients with MS. In clinical practice, discrepancy between planimetered MVA and MDPG is not uncommon in patients with VSMS, suggesting the presence of low-gradient (LG) VSMS. However, few studies have evaluated the clinical implication of this discrepancy in patients with VSMS. Therefore, we investigated the differences in characteristics and outcomes after mitral valve replacement (MVR) between low- and high-MDPG groups of patients with VSMS. The main objectives of the study were to assess the mechanism of LG VSMS and to evaluate its influence on the outcome of LG VSMS.
Methods
Study Population
We retrospectively reviewed 259 patients who underwent isolated MVR because of rheumatic MS at Severance Cardiovascular Hospital from January 2004 to December 2013. Among them, patients who did not undergo follow-up echocardiography <12 months after MVR, patients with >1+ mitral valve regurgitation, >1+ aortic valve regurgitation, and/or more than mild aortic stenosis on preoperative or follow-up echocardiography, patients with MVA > 1.0 cm 2 , patients with cardiomyopathy or coronary artery disease requiring concurrent bypass surgery, patients with other combined congenital heart disease requiring concurrent surgical correction, and patients with atrial fibrillation (AF) with rapid ventricular rate (heart rate ≥ 100 beats/min) were excluded. MVA results were first screened using clinical reports, and all images were reviewed by two experienced echocardiographers, who were unaware of patients’ clinical data, to confirm the inclusion criteria. As a result, 140 patients who underwent isolated MVR because of VSMS (MVA ≤ 1.0 cm 2 by 2D planimetry) constituted the study population. During the recruitment period, 52 patients with VSMS were sent for percutaneous mitral valvotomy (PMV). The choice of surgery or PMV was determined by physicians. In general, PMV was performed in patients with favorable morphology for the procedure and Wilkins scores ≤ 8.
Twenty-eight patients (20%) and seven patients (5%) had histories of PMV and surgical open mitral valvotomy (OMV), respectively. The median interval between MVR and PMV or OMV was 15 years (interquartile range, 11–18 years), and all these patients underwent the procedure >1 year before the echocardiographic examination. MVR was considered in 139 patients who were not candidates for PMV because of unfavorable valve morphology for PMV or left atrial (LA) clot and in one patient because of failure of PMV. Concurrent tricuspid annuloplasty was performed in patients with more than mild TR (tricuspid regurgitation) and tricuspid annular dilation according to the American College of Cardiology and American Heart Association guideline. Tricuspid annular dilation is defined as >40 mm on transthoracic echocardiography or >70 mm measured by the distance between the anteroseptal and anteroposterior commissures on direct intraoperative measurement. Patients were divided into two groups according to preoperative MDPG (LG, <10 mm Hg; high gradient [HG], ≥10 mm Hg). New York Heart Association functional class (FC) was assessed before surgery and at follow-up. Echocardiographic image acquisition for speckle-tracking analysis was performed in subgroup of 56 consecutive patients who fulfilled the enrollment criteria from March 2010 to December 2013. This study was approved by the institutional review board of Yonsei University, Severance Hospital (Seoul, Korea).
Echocardiographic Measurement
Clinical and echocardiographic assessments were performed before MVR and 12 months after MVR. The echocardiographic images of the included patients were reanalyzed by two experienced echocardiographers who were blinded to patients’ histories. LV internal diameter, septal thickness, and LV posterior wall thickness were measured at end-diastole from the parasternal short-axis view. LV mass was calculated using the formula recommended by the American Society of Echocardiography, and LV mass was indexed to body surface area (BSA). LA volume was calculated from the parasternal long-axis view and apical four-chamber view using the prolate ellipse method and indexed to BSA. The percentage LA volume index (LAVI) change between preoperative period and follow-up was calculated. MVA was assessed using 2D planimetry.
The MDPG was measured from a continuous-wave Doppler signal across the mitral valve by tracing its envelope. The severity of TR was assessed using color flow imaging and regurgitant jet area. The calculated systolic PA pressure was defined as 4 × (maximum velocity of TR jet) 2 + right atrial (RA) pressure. RA pressure was estimated by measuring the inferior vena cava (IVC) diameter and its response to inspiration. IVC diameter ≤ 2.1 cm that collapses by >50% with inspiration suggests a normal RA pressure of 3 mm Hg, whereas IVC diameter > 2.1 cm that collapses by <50% with inspiration suggests a high RA pressure of 15 mm Hg. When IVC diameter and collapse did not fit this paradigm, a value of 8 mm Hg was used.
Stroke volume was calculated using the Doppler method with LV outflow tract diameter and velocity-time integral and indexed to BSA. Cardiac output was calculated as the product of stroke volume and heart rate. Cardiac index was defined as cardiac output divided by BSA. Mitral valve effective orifice area was determined using the stroke volume measured in the LV outflow tract divided by the velocity-time integral of the mitral valve transprosthetic velocity during diastole and divided by BSA. Because patients with more than mild mitral and aortic regurgitation were excluded, mean diastolic flow rate was defined as the ratio of LV stroke volume to diastolic filling time. Net atrioventricular compliance (Cn) was determined as follows: Cn (mL/mm Hg) = 1,270 × (planimetric MVA/E-wave downslope). Systolic mitral annular (Sm) and early diastolic mitral annular (Em) velocities were assessed using pulsed-wave Doppler tissue imaging of the septal mitral annulus from the apical four-chamber view. Pulmonary hypertension (PHT) was defined as a systolic PA pressure ≥ 35 mm Hg on echocardiography. Patients were stratified into three groups: no PHT if systolic PA pressure < 35 mm Hg, mild PHT if 35 mm Hg ≤ systolic PA pressure < 50 mm Hg, and moderate to severe PHT if systolic PA pressure ≥ 50 mm Hg.
Three to five cardiac cycles were recorded in apical four-chamber, two-chamber, and long-axis views using grayscale acquisition at a frame rate > 80 frames/sec. LV longitudinal peak systolic strain and strain rate were measured offline using EchoPAC PC version 113 (GE Vingmed Ultrasound AS, Horten, Norway). After manual tracking of the endocardial contour on an end-systolic frame, the software automatically tracked the motion through the rest of the cardiac cycle. The average values of peak systolic longitudinal strain and peak systolic strain rate of the left ventricle from all three views were then calculated as global longitudinal strain and global longitudinal strain rate, respectively. Peak global strain rate during early diastole and during isovolumic relaxation (SR IVR ) was also determined. Echocardiographic measurements were averaged for three beats in patients in normal sinus rhythm and for five beats in those with AF.
Statistical Analysis
The distributions of all relevant variables are reported as percentages or as mean ± SD. The groups were compared using χ 2 statistics for categorical variables and Student’s t test for continuous variables. Correlation between the variables was assessed with the Pearson correlation test. To determine independent associates for percentage change in LAVI, linear relationships were checked with univariate linear regression analysis. Variables that were statistically significant in univariate analysis and indexed effective orifice area were entered in the multiple linear regression model. P values < .05 were considered to indicate statistical significance.
Results
Comparison between LG and HG VSMS
There were 82 patients in the LG group (59%) and 58 patients in the HG group (41%). Comparison of patient characteristics and preoperative echocardiographic parameters between the two groups are listed in Table 1 .
| Variable | Overall patients ( n = 140) | Speckle-tracking analysis cohort ( n = 56) | ||||
|---|---|---|---|---|---|---|
| LG ( n = 82) | HG ( n = 58) | P | LG ( n = 33) | HG ( n = 23) | P | |
| Age (y) | 61 ± 9 | 51 ± 11 | <.001 | 63 ± 8 | 53 ± 12 | .001 |
| Female gender | 64 (78.0) | 36 (62.1) | .039 | 7 (21.2) | 9 (39.1) | .114 |
| Hypertension | 11 (13.4) | 3 (5.2) | .109 | 6 (18.2) | 1 (4.3) | .224 |
| Diabetes mellitus | 12 (14.6) | 1 (1.7) | .010 | 5 (15.2) | 1 (4.3) | .384 |
| AF | 60 (73.2) | 28 (48.3) | .003 | 26 (78.8) | 12 (52.2) | .036 |
| Medications | ||||||
| β-blocker | 19 (23.2) | 10 (17.2) | .394 | 8 (24.2) | 3 (13.0) | .299 |
| CCB | 15 (18.3) | 4 (6.9) | .052 | 7 (21.2) | 2 (8.7) | .282 |
| ACE inhibitor/ARB | 11 (13.4) | 6 (10.3) | .584 | 7 (21.2) | 3 (13.0) | .500 |
| Diuretic | 48 (58.5) | 21 (36.2) | .009 | 18 (54.5) | 8 (34.8) | .145 |
| Digoxin | 29 (35.4) | 17 (29.3) | .452 | 14 (42.4) | 10 (43.5) | .938 |
| BSA (m 2 ) | 1.6 ± 0.1 | 1.6 ± 0.2 | .070 | 1.6 ± 0.1 | 1.6 ± 0.2 | .154 |
| Heart rate (beats/min) | 67 ± 12 | 80 ± 20 | <.001 | 68 ± 12 | 77 ± 13 | .008 |
| Echocardiography | ||||||
| LVEDD (mm) | 48.4 ± 4.2 | 46.8 ± 5.5 | .046 | 48.7 ± 4.3 | 46.7 ± 5.0 | .112 |
| LVESD (mm) | 33.0 ± 4.4 | 32.1 ± 5.0 | .154 | 33.0 ± 4.1 | 31.8 ± 3.9 | .284 |
| LVEF (%) | 61.9 ± 8.1 | 61.1 ± 10.3 | .619 | 63.6 ± 6.2 | 62.2 ± 10.3 | .554 |
| LVMI (g/m 2 ) | 94.2 ± 21.1 | 81.0 ± 21.7 | <.001 | 98.0 ± 20.6 | 84.7 ± 21.5 | .025 |
| LAVI (mL/m 2 ) | 105.1 ± 51.9 | 87.8 ± 42.9 | .039 | 116.6 ± 59.1 | 82.7 ± 33.0 | .016 |
| MDPG (mm Hg) | 7.1 ± 1.8 | 14.5 ± 3.6 | <.001 | 7.3 ± 1.8 | 13.6 ± 3.2 | <.001 |
| Planimetered MVA (cm 2 ) | 0.88 ± 0.12 | 0.73 ± 0.17 | <.001 | 0.89 ± 0.10 | 0.74 ± 0.16 | <.001 |
| Mitral EOA (cm 2 ) | 0.95 ± 0.27 | 0.64 ± 0.22 | <.001 | 0.94 ± 0.26 | 0.63 ± 0.21 | <.001 |
| Mean diastolic flow rate (mL/sec) | 113.4 ± 34.1 | 106.3 ± 33.8 | .331 | 113.1 ± 34.1 | 103.3 ± 29.9 | .126 |
| Cn (cm 3 /mm Hg) | 5.7 ± 1.7 | 4.0 ± 1.2 | <.001 | 5.5 ± 1.7 | 3.9 ± 1.1 | <.001 |
| Systolic PA pressure (mm Hg) | 37.3 ± 8.9 | 50.8 ± 17.4 | <.001 | 39.2 ± 7.6 | 50.2 ± 17.0 | .002 |
| Tricuspid annular size (mm) | 40.6 ± 6.9 | 37.3 ± 7.0 | .006 | 42.0 ± 6.2 | 38.4 ± 8.1 | .066 |
| Sm (cm/sec) | 4.9 ± 1.3 | 4.9 ± 1.2 | .742 | 5.0 ± 1.6 | 5.0 ± 1.4 | .968 |
| Em (cm/sec) | 4.3 ± 1.4 | 4.3 ± 1.8 | .762 | 4.5 ± 1.6 | 4.4 ± 1.7 | .786 |
| More than moderate TR | 28 (34.1) | 15 (25.9) | .295 | 14 (42.4) | 8 (34.8) | .565 |
| Stroke volume (mL) | 58.4 ± 14.2 | 51.0 ± 13.9 | .005 | 57.6 ± 11.9 | 47.4 ± 11.8 | .003 |
| Stroke volume index (mL/m 2 ) | 37.8 ± 8.7 | 32.1 ± 8.0 | <.001 | 37.9 ± 8.0 | 29.2 ± 6.3 | <.001 |
| Cardiac index (L/min/m 2 ) | 2.5 ± 0.6 | 2.5 ± 0.7 | .747 | 2.5 ± 0.6 | 2.2 ± 0.6 | .069 |
The LG group was older (61 ± 9 vs 51 ± 11 years, P < .001), had a higher prevalence of female gender (78.0% vs 62.1%, P = .039), had a larger proportion with diabetes mellitus (14.6% vs 1.7%, P = .010), and had a larger proportion with AF (73.2% vs 48.3%, P = .003). Use of diuretics was more common in the LG group (58.5% vs 36.2%, P = .009). Compared with the HG group, the LG group had higher LV end-diastolic dimension (48.4 ± 4.2 vs 46.8 ± 5.5 mm, P = .046), LV mass index (94.2 ± 21.1 vs 81.0 ± 21.7 g/m 2 , P < .001), LAVI (105.1 vs 87.8 ± 42.9 ml/m 2 , P = .039), planimetered MVA (0.88 ± 0.12 vs 0.73 ± 0.17 cm 2 , P < .001), tricuspid annular size (40.6 ± 6.9 vs 37.3 ± 7.0 mm, P = .006), stroke volume (58.4 ± 14.2 vs 51.0 ± 13.9 mL, P = .005), and stroke volume index (37.8 ± 8.7 vs 32.1 ± 8.0 mL/m 2 , P < .001). Among the HG group, higher MDPG (14.5 ± 3.6 vs 7.1 ± 1.8 mm Hg, P < .001) and systolic PA pressure (50.8 ± 17.4 vs 37.3 ± 8.9 mm Hg, P < .001) were seen, along with higher heart rate ( P < .001). Cn was higher in the LG group (5.7 ± 1.7 vs 4.0 ± 1.2 cm 3 /mm Hg, P < .001). No significant differences were seen in BSA, LV ejection fraction, mean diastolic flow rate, and prevalence of more than moderate degree of TR.
Among the 56 patients who underwent speckle-tracking echocardiographic analysis, 33 had LG VSMS and 23 had HG VSMS. Figure 1 shows the comparison of speckle-tracking echocardiography–derived characteristics of the left ventricle between the LG and HG groups. There were no significant differences between the groups in systolic strain (−14.1 ± 3.2% vs −14.3 ± 2.8%, P = .795), systolic strain rate (−0.74 ± 0.25 vs −0.85 ± 0.11 sec −1 , P = .056), and strain rate during early relaxation (0.90 ± 0.22 vs 0.98 ± 0.25 sec −1 , P = .200). However, SR IVR (0.17 ± 0.08 vs 0.29 ± 0.09 sec −1 , P < .001) was significantly lower in patients with LG VSMS. SR IVR showed a negative correlation with planimetered MVA ( r = −0.333, P = .013) and a positive correlation with MDPG ( r = 0.414, P = .002). However, SR IVR did not correlate with LAVI ( r = −0.250, P = .066).
Table 2 shows the multivariate analysis to determine independent correlates for preoperative MDPG in all patients. Low MDPG was independently associated with higher LV mass index, larger MVA, low systolic PA pressure, and low Cn ( < .05 for all).
| Variable | β | P |
|---|---|---|
| LVEDD | 0.736 | .103 |
| LVESD | −0.825 | .178 |
| LVEF | −0.479 | .116 |
| LVMI | −0.261 | .005 |
| Planimetered MVA | −0.285 | .005 |
| Systolic PA pressure | 0.358 | <.001 |
| Sm | 0.104 | .226 |
| Mean diastolic flow rate | −0.131 | .126 |
| Cn | −0.242 | .017 |
Table 3 shows the simple correlation between LAVI and echocardiographic variables in patients with LG and HG VSMS on preoperative echocardiography. LV parameters, including end-diastolic dimension ( r = 0.311, P = .001), end-systolic dimension ( r = 0.306, P = .001), ejection fraction ( r = −0.196, P = .042), mass index ( r = 0.453, P < .001), and Sm ( r = −0.340, P = .003), MDPG ( r = 0.252, P = .009), planimetered MVA ( r = −0.384, P < .001), and systolic PA pressure ( r = 0.261, P = .006), demonstrated correlations with LAVI in LG VSMS, although they were all relatively weak. In the HG VSMS group, only planimetered MVA showed a weak correlation with LAVI ( r = −0.288, P = .033).
| Variable | LG VSMS ( n = 82) | HG VSMS ( n = 58) | ||
|---|---|---|---|---|
| Pearson correlation coefficient | P | Pearson correlation coefficient | P | |
| LVEDD | 0.311 | .001 | 0.194 | .148 |
| LVESD | 0.306 | .001 | 0.159 | .055 |
| LVEF | −0.196 | .042 | −0.214 | .110 |
| LVMI | 0.453 | <.001 | 0.108 | .426 |
| MDPG | 0.252 | .009 | 0.036 | .792 |
| Planimetered MVA | −0.384 | <.001 | −0.288 | .033 |
| Systolic PA pressure | 0.261 | .006 | 0.137 | .309 |
| Sm | −0.340 | .003 | −0.208 | .238 |
| Mean diastolic flow rate | 0.236 | .122 | 0.041 | .775 |
| Cn | 0.032 | .781 | 0.228 | .091 |
Echocardiographic studies were repeated 22.8 ± 16.1 months after MVR. Follow-up durations were similar between the groups. Comparison of postoperative patient characteristics and follow-up echocardiographic parameters between patients in the LG and HG groups are shown in Table 4 . In the LG VSMS group, persistent AF after MVR was more prevalent (56.1% vs 31.0%, P = .003), and tricuspid annuloplasty was performed more frequently (53.7% vs 25.9%, P = .001). Additionally, in the LG group, higher values of LAVI (73.2 ± 37.5 vs 48.0 ± 24.9 mL/m 2 , P < .001), systolic PA pressure (30.5 ± 7.3 vs 25.5 ± 10.1 mm Hg, P = .010), Sm (5.0 ± 1.0 vs 5.6 ± 1.2 cm/sec, P = .008), and Em (5.4 ± 1.3 vs 6.1 ± 1.6 cm/sec, P = .012) were seen.
| Variable | LG VSMS ( n = 82) | HG VSMS ( n = 58) | P |
|---|---|---|---|
| Follow-up duration after MVR (mo) | 21.6 ± 12.8 | 24.5 ± 19.9 | .304 |
| Persistent AF after MVR | 46 (56.1) | 18 (31.0) | .003 |
| Tricuspid annuloplasty | 44 (53.7) | 15 (25.9) | .001 |
| Maze operation | 17 (20.7) | 13 (22.4) | .811 |
| Echocardiography | |||
| LVEDD (mm) | 47.8 ± 3.8 | 46.5 ± 5.0 | .076 |
| LVESD (mm) | 32.1 ± 3.7 | 31.6 ± 4.5 | .490 |
| LVEF (%) | 64.3 ± 6.3 | 63.2 ± 6.1 | .333 |
| LVMI (g/m 2 ) | 93.5 ± 20.0 | 79.9 ± 23.5 | .010 |
| LAVI (mL/m 2 ) | 73.2 ± 37.5 | 48.0 ± 24.9 | <.001 |
| MDPG (mm Hg) | 3.6 ± 1.3 | 3.2 ± 1.2 | .066 |
| Indexed EOA (cm 2 /m 2 ) | 1.2 ± 0.3 | 1.3 ± 0.3 | .355 |
| Systolic PA pressure (mm Hg) | 30.5 ± 7.3 | 25.5 ± 10.1 | .010 |
| Sm (cm/sec) | 5.0 ± 1.0 | 5.6 ± 1.2 | .008 |
| Em (cm/sec) | 5.4 ± 1.3 | 6.1 ± 1.6 | .012 |
| More than moderate TR | 8 (9.8) | 5 (8.6) | .778 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
