Author + information
- Received March 9, 1998
- Revision received June 6, 1998
- Accepted August 6, 1998
- Published online December 1, 1998.
- Zoran Popović, MDa,
- Milutin Mirić, MD, PhDa,
- Siniša Gradinac, MDa,
- Aleksandar N Nešković, MDa,
- Ljiljana Jovović, MD, PhDa,
- Ljiljana Vuk, MDa,
- Milovan Bojić, MD, PhDa and
- Aleksandar D Popović, MD, PhD, FESC, FACCa,* ()
- ↵*Address for correspondence: Dr. Aleksandar D. Popović, Dedinje Cardiovascular Institute, Milana Tepića 1, 11040 Belgrade, Yugoslavia
Objectives. This study sought to assess the effects of partial left ventriculectomy (PLV) on left ventricular (LV) performance in a series of consecutive patients with nonischemic dilated cardiomyopathy.
Background. Reduction of LV systolic function in patients with heart failure is associated with an increase of LV volume and alteration of its shape. Recently, PLV, a novel surgical procedure, was proposed as a treatment option to alter this process in patients with dilated cardiomyopathy.
Methods. We studied 19 patients with severely symptomatic nonischemic dilated cardiomyopathy, before and 13 ± 3 days after surgery, and 12 controls. Single-plane left ventriculography with simultaneous measurements of femoral artery pressure was performed during right heart pacing.
Results. The LV end-diastolic and end-systolic volume indexes decreased after PLV (from 169 to 102 ml/m2, and from 127 to 60 ml/m2, respectively, p < 0.0001 for both). Despite a decrease in LV mass index (from 162 to 137 g/m2, p < 0.0001), there was a significant decrease in LV circumferential end-systolic and end-diastolic stresses (from 277 to 159 g/cm2, p < 0.0001 and from 79 to 39 g/cm2, p = 0.0014, respectively). Ejection fraction improved (from 24% to 41%, p < 0.0001); the stroke work index remained unchanged.
Conclusions. The PLV improves LV performance by a dramatic reduction of ventricular end-systolic and end-diastolic stresses. Further studies are needed to assess whether this effect is sustained during long-term follow-up and to define the role of PLV in the treatment of patients with dilated cardiomyopathy.
The reduction of left ventricular (LV) systolic function in patients with congestive heart failure is related to an increase of LV volume. Although an increase in end-diastolic volume improves ejection performance through the Frank–Starling mechanism, this process becomes self-sustaining, leading to progressive LV dilation and associated fibrosis, a phenomenon described as ventricular remodeling (1). Partial left ventriculectomy (PLV), which consists of resection of the LV posterolateral wall confined by the two papillary muscles, has been proposed to interrupt this process (2–4). This procedure changes LV geometry by making LV more elliptical, leading to a decrease of LV volume and an increase of relative LV thickness, thus decreasing LV wall stress (5). This enables the LV muscle to continue shortening beyond the point in which it would otherwise stop, owing to the development of end-systolic stress, which the LV force cannot counteract (6). Usually, the PLV is combined with additional surgical procedures, such as mitral or tricuspid valve repair, to eliminate additional volume load of cardiac chambers (2,7). It may be hypothesized that PLV might provide early hemodynamic benefit that would persist through a longer period, enabling sustained symptomatic improvement and increased exercise capacity (4,8). However, data on hemodynamic effects of PLV are sparse and limited to a small number of patients (5,7,8). Therefore, the aim of this study was to assess the effects of PLV procedure on LV performance in a series of consecutive patients with nonischemic dilated cardiomyopathy.
Materials and methods
The study group consisted of 22 consecutive adult patients who underwent PLV from October 1996 to May 1997. All patients had severely symptomatic chronic heart failure due to nonischemic dilated cardiomyopathy despite optimal pharmacologic therapy, with a mean symptomatic period of 39 ± 36 months. All patients had severe LV dilation (echocardiographic LV end-diastolic dimension >70 mm) with normal coronary arteries, normal valve morphology, and negative history of alcohol abuse or antracycline therapy. Preoperative endomyocardial biopsy was performed previously in 11 patients; in these patients, the cause of dilated cardiomyopathy was considered to be idiopathic in 7 patients, due to previous myocarditis in 3 patients, and peripartal cardiomyopathy in 1 patient. The investigational status of PLV was explained to all patients, who signed informed consent both for the catheterization studies and for the operation.
Three patients who died after the operation (on postoperative days 12, 16 and 27) were excluded from the study, as their condition precluded the performance of the postoperative catheterization study. They were similar to the remaining 19 patients regarding age, sex, New York Heart Association class, preoperative left ventricular volumes and ejection fraction (LVEF), cardiac index, pulmonary artery pressure and LV wall stress (p > 0.05 for all). Therefore, the final study group consisted of 19 patients in whom both preoperative and postoperative catheterization studies were performed. All patients were treated with digoxin, captopril, amiodarone and furosemide both preoperatively and postoperatively. The dosage of medications (except furosemide) was kept constant over 1 month before preoperative catheterization. Five patients required intravenous inotropic support during the immediate preoperative period. However, no patient was on inotropic support during pre- or postoperative catheterization studies.
Twelve patients undergoing coronary angiography in whom no significant coronary artery disease or valvular heart disease was found served as control subjects. None of these patients were taking beta-blockers or digoxin at the time of the study.
The surgical procedure involved a diamond-shaped resection of the LV posterolateral wall confined by two papillary muscles (2,3). Three patients underwent mitral valve replacement with a mecahnical prosthesis with preservation of chordal apparatus. In all other patients, a commissural stitch was placed across the mitral valve by the transventricular approach. Additional mitral valvuloplasty procedures were performed in eight patients, also using the transventricular approach. De Vega annuloplasty for severe tricuspid regurgitation was performed in seven patients.
All patients underwent preoperative (15 ± 10 days before surgery) and postoperative (13 ± 3 days after surgery) catheterization studies. Control group was evaluated using the same protocol as for the study patients. Cardiovascular medications were not withheld on the day of the procedure. Patients received diazepam 3.5 mg intravenously immediately before the procedure. After recording baseline intracardiac pressures and cardiac output (measured by thermodilution method) and performing coronary angiography, patients entered the study protocol, previously reported by Mehmel et al. (9). A bipolar pacing catheter was placed in the right atrium in patients with sinus rhythm, or in the basal part of the right ventricle in patients with atrial fibrillation. Heart rate was set at 90 beats per minute, or to a heart rate slightly higher than the basal heart rate. In patients with basal heart rate of more than 100 beats per minute, pacing was not performed. During the postoperative catheterization study, pacing rate was set to match the preoperative heart rate. A 7F pigtail catheter was inserted into the left ventricle through an 8F femoral sheath. Single plane left ventriculography was performed in 30° right anterior oblique projection using 35-mm film at a rate of 50 frames/s, after the injection of 40 to 50 ml of nonionic contrast. Femoral artery pressure was recorded during left ventriculography using a side-arm of a femoral sheath at a paper speed of 25 mm/s, along with an electrocardiogram and a first derivative of arterial pressure on a Mingograf 7 strip-chart recorder (Siemens, Erlangen, Germany). For calibration purposes, a metal sphere with a diameter of 5 cm, placed in the position of the left ventricle, was filmed at the same focal length and image intensifier height as the ventriculogram.
Data were collected by evaluating the first two beats providing adequate LV opacification and the results were averaged. All extrasystolic and postextrasystolic beats were excluded. The LV volumes were calculated using the area-length method, and were corrected by the regression equation derived by Wynne et al. (10). The LV major axis was taken to represent the distance between aortic-mitral junction and LV apex. The minor axis was calculated as D=4A/(πL), where A= area, D= minor axis, and L= major axis. End-diastolic frame was defined as the frame with the largest ventricular silhouette, and end-systolic frame was defined as the frame with the smallest ventricular silhouette. To validate the use of single-plane ventriculography for LV volume measurements in this population, angiographic single-plane volume indexes were correlated with biplane volume indexes obtained by two-dimensional echocardiography (within 48 h) using Simpson’s biplane formula (Fig. 1).
End-diastolic pressure was defined as LV pressure at the peak of the R wave measured immediately before left ventriculography. End-systolic pressure was defined as the dicrotic notch of the femoral artery pressure tracing recorded during left ventriculography. The use of dicrotic pressure of the femoral artery was validated in 12 patients with ejection fraction <40%. We simultaneously measured central aortic and femoral artery pressures and found excellent agreement between the two measurements of end-systolic pressure (r = 0.98, p < 0.0001). Femoral artery dicrotic notch pressure was insignificantly lower than central aortic pressure (0.45 ± 2.91 mm Hg, p > 0.05).
The LV end-diastolic wall thickness was determined by measuring the distance between epicardial and endocardial surfaces of the LV anterior free wall in its middle third (11). The LV mass was determined using the approach of Rackley et al. (12); LV wall thickness at end-systole was calculated assuming constant LV mass (13). Midwall circumferential stress was calculated using Mirsky’s thick wall model (14): where a= midwall semimajor axis (cm); b = midwall semiminor axis (cm); h = LV wall thickness (cm); and P = ventricular pressure (mm Hg).
The LV sphericity was assessed as the LV major-to-minor axis ratio. Circumferential and long-axis shortening (%) were calculated as a change of minor and major axes, respectively, divided by the initial diastolic dimension, and multiplied by 100. The LV stroke work index (LVSWI) was determined by the equation where MSBP = mean systolic blood pressure (mm Hg); PCWP = mean pulmonary capillary wedge pressure (mm Hg); and SVI = stroke volume index (ml/m2) by left ventriculography.
All data are presented as mean ± SD. A paired ttest was used to compare patients’ data before and after surgery and unpaired ttest was used for comparisons between controls and patients, with Bonferroni method used for correction of multiple comparisons. Impact of mitral regurgitation and the coronary artery dominance on the improvement of LVEF following surgery was tested using the repeated measures analysis of variance. Correlations were tested with the use of linear regression. The p value ≤0.05 was considered significant.
Three out of 22 patients (13.6%) died after the operation on postoperative days 12, 16 and 27. Both pre- and postoperative clinical data are presented in Table 1. One of the patients with preoperative atrial fibrillation converted spontaneously to sinus rhythm postoperatively. Pulsus alternans disappeared postoperatively in all three patients in whom it was present preoperatively.
Volumetric and hemodynamic data
Figure 2shows representative LV end-systolic frames before and after PLV. Table 1summarizes the effect of PLV on LV volumes and pressures. After PLV, a significant decrease of both end-diastolic (from 169 ± 35 to 102 ± 25 ml/m2, p < 0.0001) and end-systolic volume indexes (from 127 ± 30 to 60 ± 21 ml/m2; p < 0.0001) were noted. Cardiac index increased (from 2.27 ± 0.86 to 2.85 ± 0.6 l/min/m2, p = 0.035), whereas end-diastolic pressure decreased (from 23 ± 11 to 15 ± 6 mm Hg; p = 0.011). Mean mitral regurgitation grade decreased from 2.1 ± 1.2 to 0.5 ± 0.8 (p = 0.0007).
LV geometry and stress
Table 2summarizes the effects of PLV on LV systolic and diastolic function. Although LV mass decreased after the operation (from 162 ± 26 to 137 ± 23 g/m2, p < 0.0001), there was also a significant decrease of the LV end-diastolic volume/mass ratio (from 1.05 ± 0.2 to 0.75 ± 0.15 ml/g, p < 0.0001). In addition, a significant increase occurred in the LV major-to-minor axis ratio in both systole (from 1.43 ± 0.17 to 1.98 ± 0.34, p < 0.0001) and diastole (1.34 ± 0.15 to 1.66 ± 0.22, p < 0.0001). Accordingly, there was a significant decrease of end-diastolic (81 ± 43 to 39 ± 16 g/cm2, p = 0.0014) and end-systolic circumferential stress (277 ± 46 to 159 ± 41 g/cm2, p < 0.0001).
The LVEF significantly improved from 24 ± 8 to 41 ± 12% (p = 0.0001) (Table 2); an increase in LVEF paralleled the decrease of end-systolic stress (Fig. 3). An increase in LVEF was due to the improvement in both minor axis shortening (from 11 ± 4 to 21 ± 8%, p < 0.0001) and major axis shortening (from 5.2 ± 3.0 to 6.9 ± 3.1%, p = 0.024). Figure 4represents the relationship between LVSWI and end-diastolic stress, a volume-independent measure of preload. Although there was no improvement in LVSWI (41 ± 18 vs. 39 ± 19 gxm/m2, p > 0.05), the work was performed at significantly lower preload.
Factors associated with the improvement of LVEF
The improvement of LVEF correlated with an increase of major-to-minor axis ratio at end-systole (r = 0.48, p = 0.038) and with the decrease of end-systolic stress (r = 0.49, p = 0.037). In a post hoc analysis, we also assessed the impact of mitral regurgitation and coronary artery dominance on LVEF improvement. To assess the effect of mitral regurgitation, we divided the patients into two groups: patients with mild (grade ≤2, n = 10) and severe (grade >2, n = 9) mitral regurgitation. The severity of postoperative mitral regurgitation was similar in both groups (0.60 ± 0.84 vs. 0.44 ± 0.88, p > 0.05) and the improvement of LVEF was similar in patients with mild (from 23 ± 8 to 41 ± 12%) and severe mitral regurgitation (from 25 ± 7 to 41 ± 13%) (p > 0.05).
To assess the effect of coronary artery dominance, we divided patients into groups with right coronary artery dominance (n = 12) and left circumflex artery dominance or co-dominance (n = 7). Interestingly enough, the improvement in LVEF was significantly greater in patients with right coronary artery dominance (from 25 ± 8 to 44 ± 3%), compared to patients with circumflex artery dominance (from 24 ± 8 to 34 ± 8%) (p = 0.017). In addition, a decrease of LV circumferential end-systolic wall stress was more prominent in patients with right coronary artery dominance (from 280 ± 53 to 152 ± 36 g/cm2) than in patients with circumflex artery dominance (from 233 ± 34 to 170 ± 50 g/cm2, p = 0.012).
Our data revealed that PLV performed in patients with nonischemic dilated cardiomyopathy reduced LV volumes, without a significant change in stroke volume, thereby improving LVEF. This was accompanied by a simultaneous increase of cardiac index and a decrease in LV end-diastolic pressure. Thus, LVEF, cardiac index and end-diastolic pressure, the most common parameters of LV function, all improved following PLV.
Our study is in accordance with previous preliminary reports showing that LVEF improves following PLV (7,8,15–17); the postoperative improvement of LVEF in these studies varied between 10% and 22%. Widely cited mechanism of this ejection improvement, which is also the proposed mechanism of the beneficial effect of PLV, is reduction of LV systolic stress (2,5,6). The PLV leads to a smaller LV minor axis diameter in its unstressed state, with no change of LV absolute wall thickness. Thus, relative wall thickness is increased, which decreases LV stress for any given level of intracavitary pressure. As the amount of ejection is inversely proportional to LV systolic stress, the net result is improved ejection performance. Although a previous study (5)reported a decrease of meridional systolic stress following PLV, this is the first study showing an inverse relation between the decrease of circumferential end-systolic stress and an increase of LVEF. Our data further strengthen this concept by showing the postoperative shift of LVEF-end-systolic stress relations in parallel to the LVEF-end-systolic stress regression line of control subjects. This indicates that the improvement in LVEF was largely due to the reduction of stress, rather than to the change of LV contractility (18).
Previously reported data are controversial regarding the effect of PLV on LV filling pressures: although Batista et al. (8)did not show reduction in LV end-diastolic pressure, two other studies, as well as our data, showed the reduction of LV filling pressures postoperatively (5,15). The LV end-diastolic pressure is influenced by passive myocardial properties, LV chamber volume and relative LV wall thickness (19). If passive myocardial properties (mostly determined by myocardial structure) do not change within a 2-week period, and relative LV wall thickness increases, it is most likely that a decrease in LV end-diastolic pressure reflects improvement in ejection performance, thus moving the end-diastolic point on a diastolic pressure/volume curve away from its steep portion. The change in myocardial stiffness cannot be completely ruled out, as Bellotti et al. (5)showed that it decreased 2 weeks after surgery. However, the assessment of LV end-diastolic properties in a period immediately following surgery must be cautiously interpreted, as they are sensitive to loading conditions, changes in ventricular interdependence, and pericardiectomy (19).
A possible reason for the improvement in LVEF is the elimination of mitral regurgitation (4). Improvement of LVEF following mitral valve repair for severe mitral regurgitation due to idiopathic dilated cardiomyopathy has been recently reported (20). In our patients, however, the degree of improvement in LVEF was not related to the correction of mitral regurgitation; in addition, the improvement of LVEF was higher in our study.
A potential disadvantage of PLV is the decrease of coronary perfusion of inferior LV segments due to the ligation of the dominant left circumflex artery or one of its major branches, which cross the ventriculectomy area. Our data indicate that an increase of LVEF was less dramatic in patients with dominant circumflex artery and that LV circumferential end-systolic wall stress decreased less in these patients, suggesting that myocardial scarring after PLV may have significant impact on LV properties. Postoperative coronary angiography in these patients frequently shows the discontinuity of the middle portion of the left circumflex artery with delayed filling of its distal branches. This may imply that a special technique may be considered in the presence of a large/dominant left circumflex artery.
The PLV was proposed as a procedure offering immediate hemodynamic benefit by LV unloading during systole, potentially serving as a bridge to heart transplantation (2,6). However, both the degree of improvement and its underlying mechanisms are not known. The present study demonstrates that an increase in LVEF correlates with a decrease in end-systolic stress and LV sphericity. Also, the decrease in LV wall volume does not increase LV end-diastolic pressure or stress to maintain the stroke work; on the contrary, the similar stroke work is performed on significantly lower levels of end-diastolic pressure and stress. If sustained, improved hemodynamics may have important therapeutic implications. A recent study indicated that chronic LV unloading with the use of LV assist device may reverse the remodeling process, as evidenced by the decrease of the LV volumes and improvement in myocardial histology (21). According to available data, it may be hypothesized that, at least in some patients, hemodynamic benefit may be sustained (8), implying the possibility of the improvement of exercise tolerance and increased life expectancy.
The major limitation of our study is that we did not perform biplane left ventriculography. Although all our patients had diffusely decreased LV kinetics preoperatively, the presence of the scar early after surgery probably diminished the motion of the posterior wall; thus, it is possible that we underestimated postoperative LV volumes. However, we have shown in our patients that LV volumes obtained by single-plane ventriculography correlated well with biplane echocardiographic measurements.
The discrepancy between angiographic and thermodilution stroke volume indexes, observed in several of our patients, may have been related to several factors, including tricuspid regurgitation (De Vega annuloplasty was performed in 7 patients), mitral regurgitation (detected in 17 patients) and atrial fibrillation (6 patients), which is known to impede exact matching between these two volumes.
Finally, for the similar reasons, we have assessed the degree of mitral regurgitation semiquantitatively by angiographic grading, rather than calculating the regurgitant volume.
Our data indicate that PLV improves LV performance by a dramatic reduction of ventricular end-systolic and end-diastolic stresses. Further studies are needed to assess whether this effect is sustained during long-term follow-up and to define the role of PLV in treatment of patients with dilated cardiomyopathy.
This study was presented in part at the 70thScientific Sessions of the American Heart Association, November 9–12, 1997, Orlando, Florida, and at the 47thAmerican College of Cardiology Annual Scientific Sessions, March 29–April 1, 1998, Atlanta, Georgia.
- left ventricle
- left ventricular ejection fraction
- left ventricular stroke work index
- partial left ventriculectomy
- Received March 9, 1998.
- Revision received June 6, 1998.
- Accepted August 6, 1998.
- American College of Cardiology
- Cohn J.N.
- ↵Carpentier A. Does surgical reduction of heart size reduce heart failure? Lancet 1997;350:456–00.
- Takeshita N.,
- Kawaguchi A.T.,
- Lima P.R.N.,
- et al.
- Mehmel H.C.,
- Stockins B.,
- Ruffmann K.,
- Olhausen K.,
- Schuler G.,
- Kuhler W.
- Sandler H.,
- Dodge H.
- Rackley C.E.,
- Dodge H.T.,
- Coble Y.D.,
- Hawley R.E.
- Ferzoco S.,
- Ahmad R.,
- Reul R.,
- et al.
- Parga J.R.,
- Avila L.F.,
- Moraes A.,
- et al.
- Glantz S.A.,
- Parmley W.W.
- Nakatani S.,
- McCarthy P.M.,
- Kottke-Marchant K.,
- et al.