Interact CardioVasc Thorac Surg 2009;9:431-436. doi:10.1510/icvts.2009.207134
© 2009 European Association of Cardio-Thoracic Surgery
Institutional report - Cardiac general |
Surgery of left ventricular aneurysm: a propensity score-matched study of outcomes following different repair techniques
Zhe Zheng1,
Hongguang Fan1,
Wei Feng,
Shiju Zhang,
Xin Yuan,
Liqing Wang,
Yunhu Song and
Shengshou Hu*
Department of Cardiovascular Surgery and Research Center for Cardiovascular Regenerative Medicine, Cardiovascular Institute and Fuwai Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences, 167A Beilishi Road, Xi Cheng District, Beijing 100037, People's Republic of China
Received 10 March 2009;
received in revised form 18 May 2009;
accepted 20 May 2009
1 The first two authors contributed equally to this work. 
 The study was funded by the National Science and Technology Pillar Program during the Eleventh Five-Year Plan Period (2006BAI01A09) and the Key Projects of Beijing Municipal Science and Technology Commission (D0906004040391).
*Corresponding author. Tel.: +86-10-88398847; fax: +86-10-88396051.
E-mail address: shengshouhu{at}yahoo.com (S. Hu).
 |
Abstract
|
|---|
To evaluate early and late outcomes of modified left ventricular reconstruction (VR) and linear repair (LR) of post-infarct left ventricular aneurysm (LVA). A total of 514 patients were consecutively operated on for LVA with modified VR technique in 145 and LR in 352 patients. Using the propensity score-matching method, we selected 202 patients (101 LRs vs. 101 VRs) with similar pre- and intra-operative characteristics and compared their clinical outcomes. After matching, the two groups of patients were similar with regard to baseline data. The increment of left ventricular ejection fraction (LVEF) in VR group was more significant than that in LR group. Operative mortality was 2.0% (2.0% LRs vs. 2.0% VRs, P=NS). There was a statistically significant difference between LR and VR patients in MACCEs (29.7% LRs vs. 13.9% VRs, P<0.01) and hospital readmissions (51.5% LRs vs. 30.7% VRs, P<0.01). Overall long-term mortality and cardiac mortality were the same between the two groups (mortality: 11.9% LRs vs. 11.9% VRs, P=NS; cardiac mortality: 9.9% LRs vs. 7.9% VRs, P=NS). The technique of repairing LVA did not affect the early and later mortality.
Key Words: Left ventricular aneurysm; Linear repair; Left ventricular reconstruction; Propensity score
|
1. Introduction
|
|---|
Left ventricular aneurysm (LVA) is a serious complication after acute myocardial infarction (MI), which can lead to heart failure. Despite recent progress in revascularization therapies, large transmural MI often results in the formation of dyskinetic or akinetic LVA. LVA is followed by an enlarged ventricular cavity and abnormal ventricular shape that permits maximal conversion of tension generated by the myocardium into cavitary pressure, and then congestive heart failure, arrhythmia and thrombogenesis.
Surgical treatment of LVA has been employed for a long time with certain success. From the first case of surgical resection of LVA that was reported by Likoff and Bailey in 1955 [1], many types of procedures were developed in the last half-century. The standard repair of LVA firstly described by Cooley et al. in 1958 consists of aneurysmectomy and simple linear closure of the ventricular wall [2]. In the following decades, in addition to the traditional linear repair (LR), alternative ventricular reconstruction (VR) techniques have been attempted in order to improve clinical outcomes [3, 4]. However, it is still controversial in relation to the optimal techniques for patients with postinfarction dyskinetic LV. Some reports [5, 6] showed that VR techniques had better clinical outcomes, whereas others [7–9] reported inconsistent or opposite outcomes. It should be noted that all the previous studies were retrospective and the assignment to each of the surgical techniques is somewhat biased. Propensity score methodology is commonly used to reduce the impact of selection bias. The aim of our study was to compare the early and late clinical outcomes of LR vs. geometry reconstruction in a propensity score-matched cohort of LVA patients.
|
2. Methods
|
|---|
2.1. Patient selection
From November 1995 to December 2005, a total of 514 patients were consecutively operated on for dyskinetic LVA after anterior MI. After exclusion of 17 patients treated with other repair techniques, such as off-pump overlapping and endocardial patch reconstruction, the remaining 497 patients were divided into two groups depending on surgical techniques: the LR group (n=352) and the VR group (n= 145). The two techniques were used at the same time in our center. The operation indications were enlarged dyskinetic left ventricle accompanied by angina and/or heart failure after MI. All surgical procedures were performed by five surgeons. The types of aneurysm closure, patch, no-patch or linear, were used depending on the personal preference of each surgeon. The left ventricular function and geometry were assessed preoperatively by transthoracic echocardiography. Every patient accepted cardiac catheterization with left ventriculography and selective coronary angiography routinely. The study was approved by the Ethics Committee of Fuwai Hospital, Beijing.
2.2. Data collection
Fig. 1 shows the study strategies. The ongoing data collection began from 2002 and was described elsewhere by our study group [10]. A pre-specified case report form (CRF) was used for the data collection. Definitions of preoperative characteristics conform to those of STS database (www.sts.org). Trained and dedicated staff was in charge of the institutional database collection. Two professional typists worked independently. All data were double checked and any discrepancy or error was corrected.
View larger version (10K):
[in this window]
[in a new window]
|
Fig. 1. Patient selection and propensity score matching. LR, linear repair; VR, ventricular reconstruction.
|
|
2.3. Surgical technique
An incision parallel to the left anterior descending artery was made in the infarcted anterior wall segment. Any thrombus was carefully removed. Standard linear closure involved longitudinal suture readaption of the left ventricular walls buttressed with polytetrafluoroethylene strips. For modified VR, an endoventricular purse-string suture was placed with a 1-0 prolene line suture. The suture was placed in the scarred tissue above the junctional zone between normal myocardium. The suture was tied and created an opening about 2 cm. The ventricular chamber was reduced and kept in satisfactory geometry. The next closure was similar with standard LR, however, the level of suture was adjusted continually to keep the ventricle from being distorted during closure and the left anterior descending was kept intact. If concomitant coronary revascularization was to be carried out, all distal anastomoses were performed during the same aortic cross-clamp period and proximal anastomoses were done during rewarming period (Fig. 2).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2. Technique of modified ventricular reconstruction (a) Exposure of the lesions and opening of the anterior wall. (b) Endoventricular continuous circular suture (Fontan stitch). (c) The ventriculotomy was closed with 1-0 prolene line as heavy horizontal mattress sutures buttressed in polytetra-fluoroethylene felt and reinforced by continuous sutures. (d) The cone shape of the cardiac apex could be restored after the suture was drawn tight and tied.
|
|
2.4. Follow-up
According to institutional follow-up protocol, patients discharged alive were required to visit our outpatient clinic six months after surgery, and then once every year. Thereafter, patients were followed up through two cross-sectional investigations in 2007 and 2008, respectively. In addition to the routine follow-up, all patients who survived to discharge were contacted again through telephone and mail by the research staff before 30 May 2008. If the patients reported adverse events, the medical records of the patients in outpatient clinic of Fuwai Hospital were double checked. If patients visited another hospital, patients were required to send their copies of medical records by mail. All patients were followed up with no missing patients. The mean follow-up for the whole study population was 58.0±27.0 months.
2.5. Statistical analysis and propensity score matching
There were two primary study endpoints including early mortality defined as all-cause death within 30 postoperative days and late mortality defined as all-cause death during the follow-up period ( 10 years). The secondary endpoints were cardiac death, MACCEs (the combination of death, Q-wave MI, stroke, and repeated revascularization) and hospital readmission due to nonfatal MI, stroke, angina pectoris, arrhythmia and heart failure.
Continuous variables are presented as mean±standard deviation (S.D.) and discrete variables were expressed by frequency percentages. Baseline characteristics considered clinically important were included. Comparative analysis of pre- and post-matched covariates based on  2 and independent Student t-test. If variables were not normally distributed, nonparametric tests were used. Postoperative survival of two-groups was compared in the Kaplan–Meier curve. Cox analysis was used to identify independent risk factors for MACCEs. A manual forward and backward variable selection procedure was used to help select the final model. A P-value of >0.05 indicates retention calibration of the model. Hazard ratios (HR) were calculated and 95% confidence intervals (CI) were given. Survival curves for all-cause mortality and readmission were estimated by Kaplan–Meier analysis.
The baseline covariates distributed between LV group and VR group were not the same. To account for this imbalance, a non-parsimonious regression model (c statistic=0.839) was built to determine the propensity score; covariates in the model included all baseline patient characteristics. Using an SPSS macro function, we matched 101 LR patients with 101 VR patients with similar propensity score. To help minimize any residual selection bias, propensity score was calculated as discussed above and a greedy matching algorithm was used to match patients on the logit of the propensity score with a caliper width of 0.5 of the standard deviation (S.D.) of the logit of the propensity score to generate a subset of patients for each surgery group which had similar baseline characteristics (Table 1).
A two-tailed P-value of <0.05 was considered statistically significant. All analyses were carried out by using SPSS statistical software (SPSS version 15.0, SPSS Inc, Chicago, IL).
|
3. Results
|
|---|
3.1. Baseline patient characteristics
Before propensity score matching, there were 431 men and 66 women, with a mean age of 59.1±9.6 (31–80) years. Three-vessel diseases were presented in 59.0%, left main disease in 17.9%. The mean left ventricular ejection fraction was 36.4%, and 35.3% of the patients were in New York Heart Association (NYHA) functional class III or IV. Mean left ventricular end-diastolic diameter (LVEDD) was 59.9 mm. Compared with patients operated with LR, patients operated with VR had more unstable angina pectoris (UAP), more coronary vessels involved, larger LVA size, larger LVEDD and lower left ventricular ejection fraction (LVEF). After propensity score-match, the covariate distributions were similar between the two groups (Table 1).
3.2. Clinical outcomes
The early and late clinical outcomes are shown in Table 2. In propensity score matched groups baseline disease characteristics, comorbidities, cardiac history and anatomy were similar. The improvement of LVEF in VR group was significantly higher than those in the LR group (Fig. 3). Operative and long-term mortalities were the same between the two groups (2.0% and 11.9%). Cardiac mortality was 8.9% (9.9% LRs vs. 8.9% VRs, P=NS). The incidence of MACCEs was significantly higher in patients operated with LR (29.7%) than those operated with the VR (13.9%, P<0.01). The rates of hospital readmission was also significantly higher in the LR group (51.5%) than the VR group (30.7%, P<0.01). Cox analysis confirmed that LR was an independent risk factor for MACCEs (HR: 2.02, 95% CI= 1.07–3.80, P=0.031) and hospital readmission (HR: 1.86, 95% CI=1.15–3.02, P=0.012). Fig. 4 shows the Kaplan– Meier curves for clinical outcomes in these patients matched for propensity score. No significant difference in late overall survival, freedom from cardiac death was observed between the patients operated with LRs or VRs (log rank P=0.90 and P=0.788, respectively). There were apparent difference between LR and VP patients in freedom from MACCEs (log rank P=0.035) and freedom from readmission (log rank P=0.030).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 4. Kaplan–Meier curves for outcomes in a cohort of patients matched for propensity scores who underwent LR and VR. (a) Freedom from death; (b) Freedom from cardiac death; (c) Freedom from MACCEs-Major adverse cardiac and cerebrovascular events (including death, myocardial infarction, cerebrovascular events, and repeated revascularization); (d) Freedom from readmission.
|
|
3.3. Adverse cardiac and cerebrovascular events analysis
In the prematch cohort of 497 patients, compared with 2.8% of patients with VR, 11.4% of patients with LR presented ischemic stroke (P<0.01). After matching, in the postmatch cohort of 202 patients, 8.0% patients presented ischemic stroke in LV group, and 1.0% in modified VR group (P=0.02). Hemorrhagic stroke occurred in one patient totally. The rate of arrhythmia of the two groups were similar before matching, however, after matching LR group showed higher rate of arrhythmia in LR (7.0%) than that in VR (2.0%) and the P-value was 0.08. The rates of other complications, such as angina, MI, heart failure and revascularization did not different between the two groups (Table 3).
|
4. Discussion
|
|---|
In this study, we found that the difference of postoperative improvement of LVEF between the two operative techniques was outstanding. Repair of LVA using reconstructive techniques resulted in a significantly greater postoperative increase in LVEF than those repaired with the LR. This is consistent with the findings from some previous studies [7, 8]. The conical pattern of normal left ventricular shape is well known. The underlying spatial configurations are closely linked to the helical ventricular myocardial band comprised of a surrounding wrap of the basal loop with transverse fibers and an apical loop of reciprocal oblique fibers forming a spiral vortex at the apex [11]. VR takes care of physiological shape, volume, revascularization and residual akinesia and can potentially achieve a positive remodeling and equalization of geometrical outcome.
We examined the mortality in a propensity score matched cohort of LVA patients who underwent LR or VR, and found similar early mortality and late mortality between LR and modified VR groups, which was consistent with many previous reports [7, 9, 12]. However, patients in the LR group had higher rates of MACCEs and hospital readmission rate than those in the VR group. This indicates that the early and long-term mortalities were not determined by repair techniques, but other factors such as, mitral regurgitation, advancing age, coronary stenosis, heart function and ventricular arrhythmia [5]. Nevertheless, ventricular techniques may influence the rates of hospital readmission and MACCEs. Considering these two groups had similar baseline risk profiles, the main difference between the two groups was LVA repair techniques, which resulted in different geometries of left ventricle postoperatively [13]. LR deforms the left ventricular geometry and reduces the curvature of ventricular wall [14]. In fact, mural thrombosis is not uncommon in patients with LVA, but the relationship of mural thrombosis and its potential complications has not received enough attention. From a theoretical and intuitive standpoint, left ventricular geometric deformation potentially increases the probability of thrombosis. In addition, deformation of left ventricle can also induce arrhythmia. Studies have shown an association between left ventricular geometry and complex ventricular ectopic activity. Increased transmural dispersion of repolarization, which correlates to the interval from the peak to the end of the T-wave on the surface electrocardiogram, is linked to ventricular tachyarrhythmias [15].
In conclusion, the technique of repairing LVA did not affect the early and late mortality. However, VR technique results in a better LVEF improvement, fewer MACCEs and hospital readmission than in those who were operated with LR. Our study suggests that left VR is a better surgical choice for patients with postinfarction LVA.
|
Acknowledgements
|
|---|
We thank Yang Wang PhD, Biostatistical Unit, Fuwai Hospital, for his assistance in the data analysis.
|
References
|
|---|
- Likoff W, Bailey CP. Ventriculoplasty: excision of myocardial aneurysm; report of a successful case. J Am Med Assoc 1955;158:915–920.[Abstract/Free Full Text]
- Cooley DA, Collins HA, Morris GC Jr, Chapman DW. Ventricular aneurysm after myocardial infarction; surgical excision with use of temporary cardiopulmonary bypass. J Am Med Assoc 1958;167:557–560.[Abstract/Free Full Text]
- Jatene AD. Left ventricular aneurysmectomy. Resection or reconstruction. J Thorac Cardiovasc Surg 1985;89:321–331.[Medline]
- Dor V, Saab M, Coste P, Kornaszewska M, Montiglio F. Left ventricular aneurysm: a new surgical approach. Thorac Cardiovasc Surg 1989;37:11–19.[Medline]
- Lundblad R, Abdelnoor M, Svennevig JL. Repair of left ventricular aneurysm: surgical risk and long-term survival. Ann Thorac Surg 2003;76:719–725.[Abstract/Free Full Text]
- Sinatra R, Macrina F, Braccio M, Melina G, Luzi G, Ruvolo G, Marino B. Left ventricular aneurysmectomy; comparison between two techniques; early and late results. Eur J Cardiothorac Surg 1997;12:291–297.[Abstract]
- Doss M, Martens S, Sayour S, Hemmer W. Long term follow-up of left ventricular function after repair of left ventricular aneurysm. A comparison of linear closure versus patch plasty. Eur J Cardiothorac Surg 2001;20:783–785.[Abstract/Free Full Text]
- Vural KM, Sener E, Ozatik MA, Tasdemir O, Bayazit K. Left ventricular aneurysm repair: an assessment of surgical treatment modalities. Eur J Cardiothorac Surg 1998;13:49–56.[Abstract/Free Full Text]
- Antunes PE, Silva R, Ferrao de Oliveira J, Antunes MJ. Left ventricular aneurysms: early and long-term results of two types of repair. Eur J Cardiothorac Surg 2005;27:210–215.[Abstract/Free Full Text]
- Li Y, Zheng Z, Hu S. Early and long-term outcomes in the elderly: comparison between off-pump and on-pump techniques in 1191 patients undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2008;136:657–664.[Abstract/Free Full Text]
- Buckberg GD. Basic science review: the helix and the heart. J Thorac Cardiovasc Surg 2002;124:863–883.[Free Full Text]
- Tavakoli R, Bettex D, Weber A, Brunner H, Genoni M, Pretre R, Jenni R, Turina M. Repair of postinfarction dyskinetic LV aneurysm with either linear or patch technique. Eur J Cardiothorac Surg 2002;22:129–134.[Abstract/Free Full Text]
- Marchenko AV, Cherniavsky AM, Volokitina TL, Alsov SA, Karaskov AM. Left ventricular dimension and shape after postinfarction aneurysm repair. Eur J Cardiothorac Surg 2005;27:475–480; discussion 80.[Abstract/Free Full Text]
- Hutchins GM, Brawley RK. The influence of cardiac geometry on the results of ventricular aneurysm repair. Am J Pathol 1980;99:221–230.[Abstract]
- Saba MM, Arain SA, Lavie CJ, Abi-Samra FM, Ibrahim SS, Ventura HO, Milani RV. Relation between left ventricular geometry and transmural dispersion of repolarization. Am J Cardiol 2005;96:952–955.[CrossRef][Medline]
|
|
Top
Abstract
1. Introduction
