The Impact of Volume Reduction on Early and Long-Term Outcomes in Surgical Ventricular Restoration for Severe Heart Failure

Transcription

1 The Impact of Volume Reduction on Early and Long-Term Outcomes in Surgical Ventricular Restoration for Severe Heart Failure Nathan Wm. Skelley, BS, Jeremiah G. Allen, MD, George J. Arnaoutakis, MD, Eric S. Weiss, MD, MPH, Nishant D. Patel, MD, and John V. Conte, MD Department of Surgery, Division of Cardiac Surgery, and The Bloomberg School of Public Health, Johns Hopkins Medical Institutions, Baltimore, Maryland Background. Recent published results suggest no additive benefit to surgical ventricular restoration (SVR) when combined with coronary artery bypass grafting. However, there may still be a subgroup of patients with severe heart failure who can benefit from this procedure. We reviewed our institutional experience with SVR to determine early and late outcomes based on volume reduction. Methods. We retrospectively reviewed our SVR patients (January 2002 to April 2008) with follow-up to March Baseline comorbidities, operative data, and postoperative outcomes were assessed by chart review, phone calls, and mailings. Survival was modeled using the Kaplan-Meier method. Cardiac magnetic resonance imaging, myocardial perfusion scans, and echocardiography assessed cardiac function, candidacy for SVR, and volume reduction. Results. We reviewed 87 consecutive SVR patients (69 men). Mean age at operation was 61.1 years. Preoperatively, all patients had congestive heart failure, with 80 (92%) at New York Heart Association III/IV. All patients underwent preoperative viability studies. Three-vessel occlusion exceeding 50% was present in 69 (79%). After SVR, ejection fraction improved from to (p < 0.001). Preoperative and postoperative magnetic resonance imaging in 26 patients (30.0%) showed a 30.8% reduction in left ventricular end systolic volume index. At follow-up, 51 of 66 (77%) improved to New York Heart Association I/II. One intraoperative death occurred. Preoperative left ventricular end systolic volume index of 80 to 120 was associated with improved survival (73% at 3 years). Conclusions. SVR is a surgical option for appropriately selected patients with severe congestive heart failure. In these high-risk patients, SVR successfully increased ejection fraction and decreased symptoms. A left ventricular end systolic volume index of 80 to 120 may be the ideal range for SVR procedures. (Ann Thorac Surg 2011;91:104 12) 2011 by The Society of Thoracic Surgeons Congestive heart failure (CHF) is a major health problem in the United States, with more than 670,000 new cases annually [1]. In the absence of a surgically correctable condition, medical therapy is the first-line treatment. Cardiac transplantation is the gold standard for CHF patients without medical or surgical options; however, development of mechanical ventricular assist devices has made progress in the last decade [2, 3]. Even so, cardiac surgeons continue to pursue new surgical options for patients with CHF and left ventricular dysfunction after myocardial infarction (MI). Post-MI remodeling results in a dilated, thin-walled, and poorly functioning ventricle with impaired mitral leaflet coaptation. These changes are associated with decreased cardiac function and are the basis for surgical ventricular restoration (SVR). SVR describes a group of Accepted for publication Sept 27, Presented at the Fifty-sixth Annual Meeting of the Southern Thoracic Surgical Association, Marco Island, FL, Nov 4 7, Address correspondence to Dr Conte, Division of Cardiac Surgery, Johns Hopkins Medical Institutions, Blalock 618, 600 N Wolfe St, Baltimore, MD 21287; jconte@csurg.jhmi.jhu.edu. surgical procedures designed to counteract post-mi remodeling by reconstructing the anterior wall to optimize left ventricular size and shape [4]. Appropriate candidates for SVR have asynergy of 35% or more from a previous MI, left ventricular enlargement, anterior wall nonviability, and depressed ejection fraction (EF) [4, 5]. SVR has been shown to improve left ventricular function and New York Heart Association (NYHA) class in patients with anterior wall MI [2, 6]. Although the benefit of SVR alone has been demonstrated, the superiority of SVR with coronary artery bypass grafting (CABG) over CABG alone has been debated [2]. One small study reported an advantage with SVR and CABG, but lack of adequate controls and retrospective study design limit its applicability [7]. A recent multicenter prospective trial randomized patients to SVR plus CABG vs CABG alone, and concluded no discernible benefit to SVR in addition to CABG. However, concerns were raised regarding this trial [5]. There may still be a subgroup of patients who benefit from an adequate SVR procedure in conjunction with CABG. Therefore, we reviewed our single-institution experience with SVR for severe heart failure in this context to 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg SKELLEY ET AL 2011;91: VENTRICULAR RESTORATION FOR SEVERE CHF Abbreviations and Acronyms BMI Body mass index CABG Coronary artery bypass grafting CHF Congestive heart failure COPD Chronic obstructive pulmonary disease CPB Cardiopulmonary bypass EF Ejection fraction IABP Intraaortic balloon pump LOS Length of stay LVAD Left ventricular assist device LVEF Left ventricular ejection fraction LVESVI Left ventricular end-systolic volume index MI Myocardial infarction MRI Magnetic resonance imaging NYHA New York Heart Association STICH Surgical Treatment for Ischemic Heart Failure SVR Surgical ventricular restoration identify patient subgroups based on left ventricular volume who benefit from SVR in addition to CABG. Material and Methods Study Design After Institutional Review Board approval, we retrospectively reviewed our prospective database for patients who underwent SVR (January 2002 to April 2008). Eligibility criteria for SVR included previous Q-wave MI, CHF (NYHA II-IV), retained left ventricular basilar function, documented anterior left ventricular wall nonviability, and asynergy (akinesia/dyskinesia) of 35% or more. Nonviability is preferably demonstrated using cardiac magnetic resonance imaging (MRI); however, nuclear myocardial perfusion scanning demonstrating a similar area of nonperfusion was deemed acceptable. Echocardiography assessed valvular function and identified akinetic/ dyskinetic wall segments. Patients were excluded if SVR was performed as a salvage operation, as documented in the medical record at the time of initial procedure. We also examined subgroups that met the criteria specified by the Surgical Treatment for Ischemic Heart Failure (STICH) surgical committee for a successful STICH operation [8, 9]; specifically, an EF increase of 10% or more and left ventricular end systolic volume index (LVESVI) reduction of 30% or more. We further stratified patients by preoperative LVESVI, as a measure for patient selection, and the presence of biventricular pacing to ascertain the impact of postoperative restoration of mechanical synchrony. Cardiac function was evaluated using MRI, echocardiography, and cardiac catheterization. Clinical data collection included demographics, preoperative cardiac function, preoperative viability studies, postoperative cardiac function, complications, and quality of life data. Quality of life data was assessed by a single investigator (N.W.S.) using the Short Form Health Survey. All-cause cumulative mortality was the primary end point; 30-day mortality was also examined. Secondary end points included cardiac function and NYHA class. Mortality data were retrieved from autopsy reports, death certificates, physicians records, and the Social Security Death Index. Operative Technique Our surgical technique has been previously described [4]. Briefly, cardiopulmonary bypass was initiated using standard arterial and venous cannulation. The left ventricle was vented through the right superior pulmonary vein and aortic root. CABG was performed using internal mammary arteries or saphenous veins, or both, followed by mitral valve repair or replacement, if necessary. Finally, SVR was performed through a ventriculotomy in the distal anterior wall parallel to the left anterior descending coronary artery. The ventriculotomy was extended distally to the apex and proximally as needed. Retention sutures were placed to maintain adequate exposure. Thrombus was removed and the left ventricle inspected for the extent of scar. In most patients, an intraventricular balloon was inserted into the left ventricle to prevent oversizing or undersizing the chamber. The balloon size was chosen to approximate optimal left ventricular end-systolic volumes indexed to body surface area (LVESVI 25 to 30 ml/m 2 ). A purse-string stitch (Fontan stitch) was then placed around the border of the intraventricular balloon to achieve the desired size and define the margins of the new anterior wall. Ventricular defects exceeding 2 to 3 cm were reconstructed with a Dacron (DuPont, Wilmington, DE) polyester patch (Invista, Wichita, KS); otherwise a linear closure was performed. The intraventricular balloon was deflated and removed once 50% of the defect was closed. Statistical Analysis Analyses were performed with STATA 9.2 software (StataCorp LP, College Station, TX). Means are presented with standard deviations and medians with interquartile ranges. Comparisons were performed using the 2 test for categoric data, as well as one-way analysis of variance and the t test for continuous data, as appropriate. For all analyses, p 0.05 (two-tailed) was considered significant. Survival was modeled using the Kaplan-Meier method, and the log-rank test assessed significance. Results Patient Population During the study period, 95 patients underwent SVR; of these, 8 were excluded because SVR was performed as a salvage operation, as documented prospectively in the medical record at the time of the procedure. Thus, 87 patients comprised the study cohort. Age at operation was years, with 69 men (79%). As defined by a body mass index exceeding 25 kg/m 2, 71 patients (81.6%) were overweight or obese. All patients had CHF and at least one previous MI. Preoperatively, 80 of 87 patients ADULT CARDIAC

3 106 SKELLEY ET AL Ann Thorac Surg VENTRICULAR RESTORATION FOR SEVERE CHF 2011;91: (92%) were at NYHA III/IV and 69 (79%) had more than 50% occlusion of 3 or more coronary arteries (Table 1). Median follow-up was 683 days (interquartile range, 129 to 1108 days), and 63 patients (72.4%) had more than 1 year of follow-up. Preoperative Viability Studies and Volumetric Measurements Preoperative viability measurements were performed in all patients: gadolinium-enhanced cardiac MRI studies were done in 79 (90.8%) and myocardial perfusion scans in 8 (9.2%). Cardiac MRI provided volumetric data preoperatively in 58 patients (66.7%) and postoperatively in 29 (33.3%); 27 patients (31.0%) had both. For the 27 patients with preoperative and postoperative volume data, Figure 1 depicts the actual change in left ventricular volume achieved in each patient. A paired t test demonstrated a significant decrease in postoperative left ventricular volume. Table 1. Preoperative Characteristics Variable Mean SD or No. (%) Patient total 87 (100) Age at operation, y Female sex 18 (20.7) Body mass index kg/m 2 56 (64.4) 30 kg/m 2 15 (17.2) Race Caucasian 72 (82.8) Black 8 (9.2) Other 7 (8.0) Comorbidities/acuity Diabetes mellitus 35 (40.2) Hypertension 62 (71.3) Pulmonary hypertension 43 (49.4) Hyperlipidemia 64 (73.4) Smoking history 58 (66.7) COPD 13 (14.9) Renal insufficiency 18 (20.7) Coronary artery disease 85 (97.7) 3-vessel 69 (79.3) 2-vessel 10 (11.5) 1-vessel 6 (6.9) Pre-op atrial fibrillation 14 (16.1) Pre-op NYHA III 53 (61.0) Pre-op NYHA IV 27 (31.0) Pre-op pacemaker 11 (12.6) Pre-op syncope 12 (13.8) Previous cardiac surgery 31 (35.6) Previous cardiac stent 45 (51.7) Previous MI 87 (100) Recent MI 30 days 15 (17.2) COPD chronic obstructive pulmonary disease; MI myocardial infarction; NYHA New York Heart Association; SD standard deviation. Fig 1. Comparison of preoperative and postoperative left ventricular end systolic volume index (LVESVI) values for individual patients. Absolute differences 30 are depicted by the solid line and differences 30 are depicted by the dotted line. Operative Data SVR with CABG was done in 77 patients (88.5%; Table 2), and less than 30% required concurrent mitral valve interventions (20 repairs, 6 replacements). An intraventricular balloon was used in 72 (82.8%), and a Dacron patch repair was necessary in 33 (37.9%). Six (6.9%) required reoperation for bleeding and 39 (44.8%) required transfusions during the operation. There was 1 intraoperative death and 3 perioperative deaths. Survival and Outcomes Kaplan-Meier survival analysis revealed 51.7% overall 4-year survival (Fig 2). Biventricular pacing was used in 23 patients postoperatively (12 new postoperatively) and Table 2. Operative Data Variable No. (%) or Mean SD (n 87) Mitral valve repair 20 (23.0) Mitral valve replacement 6 (6.9) Mean cross-clamp time, min Mean CPB time, min Intraoperative IABP 35 (40.2) Transfusion 39 (44.8) Revascularization 77 (88.5) 5 grafts 4 (4.6) 4 grafts 9 (10.3) 3 grafts 35 (40.2) 2 grafts 18 (20.7) 1 graft 11 (12.6) Patch reconstruction 33 (37.9) Intraventricular balloon-sizing device 72 (82.8) CPB cardiopulmonary bypass; IABP intraaortic balloon pump; SD standard deviation.

4 Ann Thorac Surg SKELLEY ET AL 2011;91: VENTRICULAR RESTORATION FOR SEVERE CHF 107 ADULT CARDIAC Fig 2. Kaplan-Meier survival is shown for the entire cohort. did not improve survival (Fig 3). The percentage of LVESVI reduction was examined for its effect on survival (Fig 4). A reduction exceeding 30% was used for stratification because the STICH surgical therapy committee defined it as the minimum acceptable reduction. Our data show a 5-year survival of 58% in patients with a reduction of 30% or more and 39% in those with less than 30% reduction. This analysis did not reach statistical significance and was severely limited by the number of patients who had both preoperative and postoperative MRIs to measure ventricular volumes (n 27). When stratified by preoperative LVESVI, however, a difference in survival was observed (Fig 5). During the study period, 35 of the 87 patients (40.2%) died, and most deaths occurred after 30 days post-svr (Table 3). Recurrent CHF occurred in 26 patients (29.9%). Complications are detailed in Table 3. Complications among the 35 patients who died included recurrent CHF in 18 (51.4%), pulmonary hypertension in 9 (25.7%), deep vein thrombosis in 5 (14.3%), gastrointestinal bleeding in 4 (11.4%), and stroke in 2 (5.7%). Postoperatively, 19 patients experienced new-onset atrial fibrillation. After SVR, 10 patients (11.5%) required a left ventricular assist device placed at a mean of 300 days: 3 (3.4%) within 30 days and 7 (8%) after 30 days. Fig 4. Kaplan-Meier survival is shown stratified by percentage of left ventricular end systolic volume index (LVESVI) reduction of 30% (dashed line) and of 30% (solid line). New pacemakers were placed in 12 (13.8%) and new defibrillators were placed in 15 (17.2%). Postoperative Cardiac Function Examination of cardiac function and symptoms revealed significant improvement after SVR. Whereas 80 patients were at NYHA III/IV preoperatively, only 15 were at NYHA III/IV postoperatively, an 81% improvement (p 0.001). LVESVI decreased by 30.8%, and significant improvements in left ventricular end-diastolic volume index were also observed. As well, EF increased from to (p 0.001). Among the 15 patients with recent MI ( 30 days) before the operation, the percentage change in EF was no different from the rest of the cohort. When stratified by percentage of LVESVI reduction, EF improvement, and preoperative LVESVI, there were many significant findings (Table 4). A greater improvement in EF was observed in patients with a percentage Fig 3. Kaplan-Meier survival is shown stratified by no biventricular pacemaker (solid line) preoperative pacemaker (dashed line), and postoperative pacemaker (dotted line). Fig 5. Kaplan-Meier survival is shown stratified by preoperative left ventricular end systolic volume index (LVESVI) of 80 ml/m 2 (dashed line), 80 to 120 ml/m 2 (solid line), and 120 ml/m 2 (dotted line).

5 108 SKELLEY ET AL Ann Thorac Surg VENTRICULAR RESTORATION FOR SEVERE CHF 2011;91: Table 3. Postoperative Outcomes and Complications Variables No. (%) or Mean SD (n 87) Outcomes Death from any cause 35 (40.2) Death 30 days after SVR 4 (4.6) Death 30 days after SVR 31 (35.6) Total LOS, d Postoperative LOS, d Postoperative IABP 7 (8.0) Complications Respiratory failure 8 (9.2) Renal insufficiency 24 (27.6) Arrhythmias 39 (44.8) Atrial fibrillation 33 (38.9) Ventricular tachycardia 6 (6.9) New pacemaker 12 (13.8) Intraoperative transfusion 39 (44.8) Bleeding requiring reoperation 6 (6.9) Deep-vein thrombosis 7 (8.0) Stroke 5 (5.7) Infection Peripheral 5 (5.7) Sepsis 4 (4.6) Urinary tract infection 5 (5.7) Pneumonia 7 (8.0) Sternal 7 (8.0) New defibrillator implantation 15 (17.2) LVAD within 30 days 3 (3.4) LVAD after 30 days 7 (8.0) IABP intraaortic balloon pump; LOS length of stay; LVAD left ventricular assist device; SD standard deviation; SVR surgical ventricular restoration. LVESVI reduction of 30% or more (12%) vs a reduction of 20% or less ( 1.7%). Patients with lower preoperative LVESVI had greater preoperative EF, but there was no difference postoperatively in EF or change in EF (Table 4). However, the postoperative LVESVI and change in LVESVI achieved after SVR differed significantly by preoperative LVESVI strata. Table 5 examines EF and LVESVI as a function of time postoperatively. After 1 year, the difference in EF compared with preoperative values remained significantly higher (p 0.001). Similarly, the LVESVI remained significantly reduced at 1 year (p 0.01). Quality of Life Quality of life assessments were available for 32 of 52 living patients. General level of health was reported as excellent in 10 patients (31.3%), very good in 10 (31.3%), good in 6 (18.7%), and poor in 6 (18.7%). Seventeen patients (53.1%) reported little to no limitations in physical activity. When stratified by preoperative LVESVI strata, a trend was observed in the percentage of patients rating their overall level of health as excellent or very good, with LVESVI below 80, 25%; LVESVI of 80 to 120, 86%; and LVESVI exceeding 120, 50% (p 0.06). The exact same trend was observed for patients rating their health-related physical activity limitation as none or a little. When stratified by percent LVESVI reduction 30% or more vs less than 30%, overall level of health rating as excellent or very good as well as little or no healthrelated physical activity limitations did not reach significance (85.7% vs 50%, p 0.098, for both). However, this stratification was again limited by patients for whom preoperative and postoperative MRI data were available. Comment SVR technique has evolved since Dr Vincent Dor [10] described the procedure in 1984, but the objectives have remained constant. SVR aims to revascularize ischemic myocardium, if necessary, reduce end diastolic pressure, reduce ventricular dyssynchrony, improve overall ventricular function, and have a positive effect on measurable variables of CHF and survival. SVR is appropriate for a few patients with ischemic cardiomyopathy. Contraindications to SVR include a viable anterior wall and coronary artery disease not amenable to revascularization. Hence, it is extremely important that patients are properly evaluated for infarction, ventricular enlargement, and most important, anterior wall nonviability. At our institution, we prefer cardiac MRI to prove nonviability because it also provides accurate ventricular volume measurements. However, we use nuclear myocardial perfusion scans if MRI is not possible. In this study, we present SVR outcomes from a singleinstitutional cohort of severe CHF patients. The acuity of the cohort is illustrated by the high percentage of patients with NYHA III/IV, EF of less than 0.30, 3-vessel coronary artery disease, history of MI, and pulmonary hypertension. Although our patients have more advanced disease than most published series, we demonstrate that SVR can be performed safely and effectively on patients who are appropriately selected. Most important, 100% of our patients underwent preoperative studies demonstrating myocardial nonviability and had a history of MI, the two most essential selection criteria for SVR. As a result, we report improvement across several measures of cardiac function after SVR. Our patients exhibited significant improvements in NYHA class, EF, and ventricular volumes. In those with follow-up imaging, improvements persisted at 1 year. As well, 51.7% survival of our (mostly NYHA III/IV) patients 4 years after SVR is improved compared with 40% for medically managed NYHA II-IV patients [11]. In addition, most of our patients report a postoperative health status of excellent or very good after SVR, with more than 50% reporting little to no health-related activity limitations. When stratified by percent LVESVI reduction, there was no statistically significant survival advantage. A trend toward improved survival for patients with more than 30% reduction seemed to emerge at 5 years; however, this comparison suffered from lack of statistical power. It is possible that with higher sample size, a true advantage would be evident. This would be consistent with our belief

6 Ann Thorac Surg SKELLEY ET AL ;91: VENTRICULAR RESTORATION FOR SEVERE CHF Table 4. Cardiac Function Tests Percent LVESVI Reduction Variable 30% (n 17) 20% (n 7) p Value a ADULT CARDIAC Pre-op EF Post-op EF Change in LVEF, % Pre-op LVESVI (ml/m 2 ) Post-op LVESVI (ml/m 2 ) Change in LVESVI (ml/m 2 ) EF increase p Value a (n 39) (n 39) Pre-op EF Post-op EF Change in EF, % (n 29) (n 25) Pre-op LVESVI (ml/m 2 ) (n 17) (n 19) Post-op LVESVI (ml/m 2 ) (n 15) (n 12) Change in LVESVI (ml/m 2 ) Preoperative LVESVI 80 ml/m ml/m ml/m 2 p Value b (n 22) (n 21) (n 15) Pre-op EF Post-op EF Change in LVEF, % Pre-op LVESVI (ml/m 2 ) (n 8) (n 11) (n 10) Post-op LVESVI (ml/m 2 ) (n 8) (n 10) (n 9) Change in LVESVI (ml/m 2 ) a The p value is from Fisher exact test for categoric variables, t test for parametric continuous variables, and Wilcoxon rank-sum test for nonparametric continuous variables. b The p value is from one-way analysis of variance. EF ejection fraction; LVEF left ventricular ejection fraction; LVESVI left ventricular end-systolic volume index. that SVR imparts a survival advantage in the mid-to-late postoperative period. Stratification of survival, cardiac function, and anatomic data indicated the greatest improvement in patients with preoperative LVESVI of 80 to 120 ml/m 2. These data support the use of LVESVI as a tool for SVR patient selection. As well, although survival was higher for patients with a preoperative LVESVI of less than 80 ml/m 2 than for patients with an LVESVI exceeding 120 ml/m 2, the latter group was more likely to have improvement in LVESVI and there was a trend toward improved EF after SVR. The lower likelihood of patients with preoperative LVESVI of less than 80 ml/m 2 to achieve improvement in postoperative LVESVI may reflect difficulty in achieving significant volume reduction in patients who started with relatively smaller left ventricular volumes. In addition, stratification by postoperative biventricular pacing, which reflects restoration of mechanical synchrony after SVR, did not affect survival, as has been reported previously [2, 12]. The SVR literature largely consists of retrospective single-center or multicenter database reviews [10, 13, 14]. SVR proponents point to salutary observational outcomes. Skeptics identify the flaws of such studies and hypothesize that improvements could be solely attributable to concomitant revascularization. However, some patients demonstrate clinical improvement and relief of symptoms with SVR alone. To address this question, the STICH trial was designed to study the additive effect of SVR on revascularization in a

7 110 SKELLEY ET AL Ann Thorac Surg VENTRICULAR RESTORATION FOR SEVERE CHF 2011;91: Table 5. Cardiac Function Tests Over Time Variable Preoperative Postoperative 3 mon 3 6 mon 6 mon 1 yr 1 yr (n 87) (n 44) (n 6) (n 6) (n 19) EF a a a (n 54) (n 33) (n 4) (n 3) (n 9) LVESVI, ml/m a a a Significant (p 0.05) using paired t test. EF ejection fraction; LVESVI left ventricular end-systolic volume index. prospective randomized fashion. To control the oftenheterogeneous SVR population, specific study enrollment criteria were designed, with documentation of anterior wall nonviability by MRI being central among them. To account for variability in surgical technique, the surgical therapy committee set criteria by which an operation would be considered an appropriate STICH procedure. Rather than mandating which of the widely accepted techniques was to be used, they stated that any procedure reducing LVESVI by 30% and increasing EF by 10% was acceptable [8, 9]. The study concluded no significant advantage for SVR in addition to CABG. In light of these findings, some members of the heart failure community have concluded that there is no role for routine SVR. Despite the rigorous prospective randomized design, others have pointed out certain flaws in the study methodology and question applying STICH results in everyday practice [15], Although there were narrowly defined inclusion criteria, the final study had broad enrollment of heart failure patients to achieve acceptable sample size. Indeed, SVR may not be appropriate for many CHF patients. We believe SVR is only appropriate for a highly select group of severe CHF patients who meet specific eligibility criteria. These results suggest patients with preoperative LVESVI of 80 to 120 ml/m 2 achieve maximal benefit. STICH protocol modifications resulted in only 50% of STICH patients having demonstrated anterior wall nonviability and 13% having no history of MI. These changes imply that an operation designed to correct the effects of post-mi remodeling was performed in a fraction of patients without prior MI. All of our 87 patients had documented nonviability of the anterior wall and a prior MI before undergoing SVR. Furthermore, although an adequate SVR procedure was designated as an LVESVI reduction of 30% or more, the STICH trial noted an average decrease of 19%, whereas we report 30.8%. Our data, although retrospective and underpowered, suggest that there is likely a population of severe CHF patients who benefit from SVR, owing to observed differences in outcomes based on preoperative LVESVI and the percentage of LVESVI reduction. The concept that an appropriate SVR must be performed to achieve benefit is intuitive and is supported by our data. We also show that patients with LVESVI reduction of 30% or more achieved a greater increase in EF. Other studies have shown a link between overall survival and left ventricular volume in patients with past MI and severe heart failure [16, 17]. In addition, an analysis of preoperative characteristics reveals that our patient population is significantly different when compared with that of recently published series. Only 5% of STICH patients were NYHA IV preoperatively, whereas 31% of the patients in this study were NYHA IV. Furthermore, the STICH trial excluded patients with previous coronary artery stenting, whereas more than 50% of our cohort underwent prior percutaneous revascularization. In addition, this study included a greater percentage of patients with hyperlipidemia, hypertension, diabetes, chronic renal insufficiency, and current smoking status. As a consequence, our cohort represents patients with greater preoperative risk factors than similar single and multiinstitutional studies and likely represents more closely a patient population in whom SVR would be beneficial [2, 13, 18]. These data illustrate that SVR can be safely performed in a complex and high-risk population. Our study is strengthened by having a single surgeon perform more than 95% of cases and by having the same treatment team evaluate preoperative and postoperative studies. All operations were performed in the same hospital, and the technique described in Methods has been used on all patients in our institutional history. This study does not have the rigorous prospective design of the STICH trial, but it does provide an interesting and necessary counterpoint. Even though SVR is not appropriate for many CHF patients, it can be beneficial in patients of high acuity and complexity who meet strict eligibility criteria. Hence, we advocate a specific reexamination of the high-quality STICH data using strict eligibility and volumetric outcomes criteria. Our study is limited by its retrospective design; this negatively affected our ability to obtain complete preoperative and postoperative cardiac MRI volumetric measurements for all patients. Therefore, our volumetric data are not complete for all patients. Our study was not protocol driven, and patient care was determined by clinical judgment. As well, we did not measure cardiac function for patients at well-defined predetermined postoperative times, so our follow-up imaging is limited to studies obtained for clinical purposes. Owing to our relatively small sample size, we were unable to compare SVR with and without CABG; more than 88% of patients had SVR with concomitant CABG. In this single institution review of outcomes after SVR, we have examined outcomes in a severe CHF population carefully selected to achieve benefit from SVR. In this

8 Ann Thorac Surg SKELLEY ET AL 2011;91: VENTRICULAR RESTORATION FOR SEVERE CHF cohort of very high acuity patients, 100% of whom had anterior wall nonviability, SVR can yield good clinical outcomes and freedom from CHF with 5-year follow-up. Furthermore, these data suggest that preoperative LVESVI of 80 to 120 ml/m 2 may be the ideal range over which patients derive functional benefit from SVR. Further studies on SVR are warranted to specifically examine subgroups of severe heart failure patients who are likely to derive benefit from this procedure. Dr Allen is the Hugh R. Sharp Research Fellow, and Drs Weiss and Arnaoutakis are Irene Piccinini Investigators in Cardiac Surgery. This work was supported by the National Institutes of Health (IH2T32DK ,ESW). References 1. American Heart Association. Heart disease and stroke statistics. Dallas: American Heart Association, Athanasuleas CL, Buckberg GD, Stanley AW, et al. Surgical ventricular restoration: the RESTORE Group experience. Heart Fail Rev 2004;9: John R, Kamdar F, Liao K, Colvin-Adams M, Boyle A, Joyce L. Improved survival and decreasing incidence of adverse events with the HeartMate II left ventricular assist device as bridge-to-transplant therapy. Ann Thorac Surg 2008;86: ; discussion Conte JV. Surgical ventricular restoration: technique and outcomes. Congest Heart Fail 2004;10: Athanasuleas CL, Buckberg GD, Conte JV, Wechsler AS, Strobeck JE, Beyersdorf F. Surgical ventricular reconstruction. N Engl J Med 2009;361:529 30; author reply Williams JA, Weiss ES, Patel ND, Nwakanma LU, Conte JV. Outcomes following surgical ventricular restoration for patients with clinically advanced congestive heart failure (New York Heart Association Class IV). J Cardiac Fail 2007;13: Prucz RB, Weiss ES, Patel ND, Nwakanma LU, Baumgartner WA, Conte JV. Coronary artery bypass grafting with or 111 without surgical ventricular restoration: a comparison. Ann Thorac Surg 2008;86:806 14; discussion U.S. National Institutes of Health. ClinicalTrials.gov. Comparison of Surgical and Medical Treatment for Congestive Heart Failure and Coronary Artery Disease (STICH). clinicaltrials.gov/show/nct Buckberg GD, Athanasuleas CL. The STICH trial: misguided conclusions. J Thorac Cardiovasc Surg 2009;138: e Dor V. Left ventricular reconstruction: the aim and the reality after twenty years. J Thorac Cardiovasc Surg 2004;128: O Connor CM, Velazquez EJ, Gardner LH, et al. Comparison of coronary artery bypass grafting versus medical therapy on long-term outcome in patients with ischemic cardiomyopathy (a 25-year experience from the Duke Cardiovascular Disease Databank). Am J Cardiol 2002;90: DiDonato M, Toso A, Dor V, et al. Mechanical synchrony: role of surgical ventricular restoration in correcting LV dyssynchrony during chamber rebuilding. Heart Fail Rev 2004;9: Menicanti L, Castelvecchio S, Ranucci M, et al. Surgical therapy for ischemic heart failure: single-center experience with surgical anterior ventricular restoration. J Thorac Cardiovasc Surg 2007;134: Athanasuleas CL, Buckberg GD, Stanley AW, et al. Surgical ventricular restoration in the treatment of congestive heart failure due to post-infarction ventricular dilation. J Am Coll Cardiol 2004;44: Conte J. An indictment of the STICH trial: True, true, and unrelated. J Heart Lung Transplant 2010;29: White HD, Norris RM, Brown MA, Brandt PW, Whitlock RM, Wild CJ. Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation 1987;76: Migrino RQ, Young JB, Ellis SG, et al. End-systolic volume index at 90 to 180 minutes into reperfusion therapy for acute myocardial infarction is a strong predictor of early and late mortality. The Global Utilization of Streptokinase and t-pa for Occluded Coronary Arteries (GUSTO)-I Angiographic Investigators. Circulation 1997;96: Jones RH, Velazquez EJ, Michler RE, et al. Coronary bypass surgery with or without surgical ventricular reconstruction. New Engl J Med 2009;360: ADULT CARDIAC DISCUSSION DR IRVING KRON (Charlottesville, VA): I would like to thank the authors, and, by the way, Mr. Skelley, you did a great job presenting. I agree entirely with the authors that the 19% reduction in ventricular volumes noted by the Surgical Treatment for Ischemic Heart Failure (STICH) investigators is just too small, and this probably explains why the STICH trial showed no benefit for left ventricular restoration. The fault is ours as surgical investigators, and you know how troublesome it is to involve people in surgical cardiac trials. We as surgeons have a poor history with this. In many centers, the people that we thought were too sick for randomization in STICH ended up getting a left ventricular restoration. In fact, the people we didn t think needed the operation ended up getting randomized, and no doubt that is why we are left with the results that are. That is really the issue, and the operation may not be available for us. I have two questions for the authors. The first is how their long-term outcomes compared to the cohort reported by the STICH trial? I would like to specifically ask about operative mortality and long-term freedom from heart failure. As the authors know, there is no difference between coronary bypass vs the restoration group. The second question is a bit more complicated and it is a question we all face. Insurance will no longer pay for this procedure. It has been proven to not be effective. So how does one offer this procedure to patients who we believe are deserving of the procedure and get reimbursement for it? I agree with you; hopefully the subgroup analyses will show a difference. In fact, the people who really should have gotten the restoration will prove to be better than the ones who didn t. Thank you very much for a great presentation. MR SKELLEY: Thank you, Dr Kron, for your comments. To address your first question regarding the clinical outcomes between our institution and the STICH trial, the 4-year survival for STICH is approximately 75%, whereas we report a 51.7% survival at 4 years. The important difference to note, which I highlighted in my presentation interpreting these numbers, is that the patient population at our institution was of a much higher risk stratification compared to the STICH trial. Our patients presented with a greater number of comorbidities at the time of the surgical ventricular restoration (SVR) procedure, we had a greater percentage of patients with diabetes, hyperlipidemia, hypertension, current smoker status, chronic renal insufficiency, previous percutaneous coronary intervention, and a body mass index (BMI) greater than 25. It is also important to realize that 31% of our patients were New York Heart Association IV preoperatively, while only 5% of the STICH trial patients were

9 112 SKELLEY ET AL Ann Thorac Surg VENTRICULAR RESTORATION FOR SEVERE CHF 2011;91: New York Heart Association IV. Although not mentioned during my presentation, it is also important to realize that the STICH trial had 4% of patients with severe mitral regurgitation (MR) preoperatively, 15% of our patient population had severe MR preoperatively and was reduced to 3% after SVR. Eleven percent of our patients received a left ventricular assist (LVAD) post-svr compared to less than 1% for the STICH trial. We had a similar percentage of patients receiving a pacemaker post-svr, though. We reported 13% and STICH reported 15%. Both studies reported 17% receiving a defibrillator postoperatively. When we stratified our patient population by this pacemaker performance, however, there was no statistically significant difference in survival. Your second question regarding the insurance aspect, as you noted, is an extremely complicated topic. Currently at Johns Hopkins, we are reimbursed for this procedure. Your question comes at a very unique time in health care reform, though, when studies like the STICH trial can dictate the reimbursement practices for medical procedures such as SVR. There has been a recent movement across all medical specialities to embrace evidence-based medicine. The evidence-based medicine model applies the scientific method to medical research and then attempts to help direct medical care based on these results; however, physician and hospital reimbursements are also being directed by evidencebased medicine protocols. As this trend continues, it will be even more important for the scientific community to perform quality and properly standardized research. The conclusions from the STICH trial demonstrate that there is no additional benefit with coronary artery bypass grafting (CABG) and SVR vs CABG alone. In an editorial published in the same edition of The New England Journal of Medicine by Dr Howard Eisen, he states, and I quote, On the basis of this trial, the routine use of SVR in addition to CABG cannot be justified. That is an accurate statement based on the results of the STICH trial, but as I showed in the presentation, there are many complications with the study that call statements like that into question. Insurance companies, clinicians, government agencies, and patients will have to be increasingly vigilant in the new age of health care reform to insure that the necessary studies are carefully performed and interpreted so that procedures, like SVR, are still available to the patients who need them. DR KEVIN D. ACCOLA (Orlando, FL): Very nice presentation. I have a question, and perhaps Dr Conte would want to address this, and possibly editorialize somewhat as well, because I know he was very intricately involved in the STICH trial. We have done 127 ventricular restoration procedures with very good results and only 3 deaths, although we don t have any long-term follow-up. I applaud you for following these patients so closely. They demonstrated improved left ventricular function as well as their physical capabilities. Since this procedure has been blacklisted from the recent New England Journal article involving the STICH trial, have you altered your algorithm of which patients on whom you perform ventricular restoration? Can you possibly give us some guidance in regards to collecting data, as Dr Kron pointed out, from a surgical perspective so we can prove efficacy of this procedure? Those of us who have done a number of these procedures know the patients get better; they have better long-term physical capabilities with improvement in ejection fraction on follow-up echos. So again, I applaud your results and we need more of these studies in the surgical literature. for the enrollment issue this trial faced, and surgeons need to be more vigilant about getting patients appropriately enrolled in trials. The biggest problem I had with the conduct of the STICH trial is that the STICH trial we had at the beginning was not the STICH trial at the end. Because of poor enrollment, and you can go and check this out at the National Heart Lung Blood Institute (NHLBI) Web site, the entrance criteria in the STICH trial were changed so that you no longer had to document nonviability to allow patients to be enrolled. So what we are talking about, and quite honestly I think our paper points this out, is that the wrong patients were enrolled in the trial for the sake of completing the trial. Rather than looking to make sure the original question was answered, they made sure that enrollment was completed. And I think that that, quite honestly, calls into question the leadership of the trial and what their primary goals were. The right question was asked at the beginning but the right patients weren t enrolled. This isn t the group of patients that Vincent Dor came up with this operation for; they are very different. But unfortunately, the paper was published in the New England Journal before people who had knowledge were actually able to review it. The STICH surgical therapy committee decided that an appropriate STICH SVR would have a 30% reduction of volume and a 10% increase in EF percentage. They based their 30 cc volume reduction on papers over the 20 years in the history of this operation that actually showed some benefit. And while they didn t proscribe you had to do the operation a certain way, they said that, okay, do a Dor, do a linear closure or like Lynda Mickleborough does it, use a cerclage technique like Pat McCarthy does it, but just go and make sure that you achieve a 30% volume reduction. But in fact they didn t. So if they didn t do the right operation on the right patients, how can you expect that there would be a benefit? And in fact, you should probably have a worse result, because I think all of us that do coronary bypass surgery would say that if you added an unnecessary ventriculotomy to a revascularization operation, you are probably not going to have as good long-term outcomes because you are going to have arrhythmias and scar formation and you are going to decrease function. So the fact that the outcomes were equal and not worse probably shows that there was a successful outcome in those few SVR patients, appropriately operated-on SVR patients, in that series. You know, Kevin, I haven t done anything to change what we tell patients. I will tell you, referrals are far less than they used to be for this, but I don t necessarily think it is because of that. I think it is because of other options in the medical realm of things. They are broadening the use of biventricular (BiV) pacers to people who do not have widened QRS complexes, and in addition to that, there are some studies out there that will look at expanded use of BiV pacers and different medical therapies, including oral inotropes. So that that same small group of patients that should be candidates for this operation aren t being referred. SVR is not for everybody; no one ever said it was. But, unfortunately for the sake of completing the trial they enrolled the wrong patients and now many patients who were probably appropriate candidates will not be offered this procedure. DR HERMAN A. HECK, JR (New Orleans, LA): In those patients that you do SVR and coronary bypass who have the higher degrees of mitral regurgitation, do you feel that adding a ring is necessary, or does realignment of the papillary muscles suffice in those more extensive mitral regurgitant patients? DR CONTE: I would like to thank Irv and Kevin for their comments. You are absolutely right, Irv. I think we as surgical investigators in the STICH trial share some of the responsibility DR CONTE: I think it depends on the size of the annulus. We do reduce the degree, but if we have enlarged annular diameters, we put a ring in as well. I think it is complementary.