Objective: The purpose of this study was to investigate the safety and efficacy of multivessel beating heart revascularization in a high-risk group of patients with severe left ventricular dysfunction as well as to provide intermediate survival and quality of life data. Methods: Our prospectively updated database was queried to extract all patients with left ventricular ejection fraction ≤30% who underwent beating heart revascularization. Standard demographics, clinical profiles and outcomes were collected. Outcomes were compared with Society of Thoracic Surgeons (STS) benchmarks for all coronary artery bypass grafting (CABG) patients. Telephone interviews were conducted and survival and quality of life data were tabulated. In addition, morbidity and mortality outcomes were compared with a concurrent cohort of patients with similarly impaired left ventricular function who underwent conventional coronary artery bypass. Results: One hundred off-pump coronary artery bypass grafting patients were identified and follow-up was 93% complete in these patients. Mean age was 67±10.5 years and mean ejection fraction was 26±4%. Twenty-one percent were females. Balloon counterpulsation support was used liberally in the perioperative period. Patients received a mean of 3.5 grafts with 83% internal mammary artery use. Observed mortality was 3% with a predicted mortality of 5.3%. Observed to expected ratio was 0.56. Incidence of adverse events compared favorably with both that reported in the STS for all CABG patients regardless of left ventricular function, and also to a concurrent CABG cohort. One-year survival was 85%. Freedom from cardiac readmission was 88% and freedom from angina was 83%. No patient required repeat percutaneous or surgical intervention. Conclusions: We conclude that multivessel off-pump revascularization in patients with severe left ventricular dysfunction is a safe and effective alternative to conventional grafting. Long-term follow-up is mandatory to confirm these encouraging intermediate outcomes.
Off-pump coronary artery bypass grafting (OPCAB) is increasingly being applied as an alternative to conventional cardiopulmonary bypass-supported myocardial revascularization. Indeed, the last harvest of the Society of Thoracic Surgeons database demonstrates that the proportion of isolated coronary bypass procedures performed off pump has been increasing in the last few years (6.9% of cases in 1999 to 18.6% in 2001). Despite the rising enthusiasm for the clinical use of OPCAB, there has been a reluctance to apply this approach to patients with severe left ventricular dysfunction for fear of hemodynamic instability and increased morbidity and mortality.
It is well known that patients with impaired left ventricular function undergoing revascularization on cardiopulmonary bypass have increased mortality and morbidity when compared with patients with normal left ventricular function . It has been speculated that extracorporeal circulation may exacerbate myocardial damage in compromised left ventricles as a result of: (a) activation of inflammatory mediators; (b) non-physiologic ventricular geometry of the empty heart impeding collateral flow to ischemic areas ; and (c) worsened preservation of interventricular septal movement . In view of these factors and because patients with depressed left ventricular function often have several accompanying comorbidities, we hypothesized that beating heart revascularization may impact favorably on the suboptimal early and intermediate outcomes associated with cardiopulmonary bypass-supported revascularization in this ill group of patients.
We interrogated our prospectively updated database (CAOS, Intelligent Business Solutions, Clemmons, NC) to identify all patients who underwent OPCAB at our institution between January 1, 1999 and July 31, 2001, and who had a preoperative left ventricular ejection fraction (LVEF) of 30% or less as measured by echocardiography, nuclear imaging and/or ventriculography.
The incidence of adverse events among this high-risk group was compared with the Society of Thoracic Surgeons (STS) Database benchmarks for all CABG patients [unpublished data from STS US Cardiac Surgery Database]1 as the outcome benchmarks for patients with severe left ventricular dysfunction were not available to us. In addition, we compared the morbidity and mortality outcomes to a concurrent group of patients with similarly impaired left ventricular function who underwent conventional on-pump revascularization during the same period. The OPCAB patients and their referring cardiologists were contacted for follow-up data. All data were collected and defined using the STS Cardiac Surgery Database definitions [unpublished data from STS US Cardiac Surgery Database]2.
While no formal quality of life instruments (MLHF, SF-36) were used, we contacted each patient who underwent off-pump revascularization to assess for indicators of quality of life. In particular, we investigated the incidence of recurrent angina, hospital readmission, hospital readmission for cardiac causes, need for recatheterization and need for repeat CABG. We also enquired from the patient whether they would undergo surgery again knowing what it entailed. In addition to telephone follow-up, a second outcome measure for this study was all-cause mortality over a 2-year interval. For patients who could not be contacted directly, mortality through November 1, 2002 was ascertained using the National Death Index (US Department of Health and Human Services). Social security numbers were used for the initial query, then those without records were subjected to a second query using a combination of name, date of birth and state of last known residence. The sensitivity of the National Death Index to identify deaths is between 92 and 99% depending on which identifiers are available .
Data were analyzed with Sigma Stat 2.0 statistical software package (SPSS Science, Chicago, IL). Clinical information was summarized by means±standard deviation for continuous variables and absolute numbers with percentages for categorical variables. Results were compared using unpaired Student's t-test, and Chi square (χ2) tests, where appropriate. Yates correction for continuity was used in all Chi square analyses. Adverse outcomes were analyzed with Chi square analysis after Bonferroni correction for 10 comparisons. Kaplan–Meier techniques were used to determine actuarial survival and were compared using the Cox–Mantel Log rank statistic (Win STAT, R. Fitch Software, Staufen, Germany). A P value less than 0.05 was considered statistically significant.
During the study period, 1624 patients underwent coronary artery bypass grafting as an isolated procedure or accompanied by transmyocardial laser revascularization or carotid endarterectomy (Fig. 1 ). Of these, 911 (56%) were performed as off-pump procedures. Among these, 100 (11%) had an LVEF ≤30%. The latter constitutes the study group. During this time period, traditional coronary artery bypass grafting was performed on 110 patients with an ejection fraction ≤30%, these patients served as the comparison group.
Patients demographic profile is depicted in Table 1 . Among the OPCAB patients, mean age was 67±10.5 years and the typical patient was a Caucasian male. A greater proportion of patients were in NYHA class I or II and had ejection fractions in the 26–30% range. Hypertension, previous myocardial infarction and a history of smoking were prevalent comorbidities. Approximately half of the patients were receiving β-blockers, angiotensin converting enzyme inhibitors and diuretics but the duration of treatment with these agents was not available. The majority of patients presented with symptoms of congestive heart failure. For those patients without angina, efforts were made to determine contractile reserve on dobutamine stress echocardiography. Left main disease, defined as >50% stenosis, and triple vessel disease were rather prevalent in this population. The study and comparison groups were very well matched except for a higher incidence of diabetes in the conventional CABG cohort.
All patients underwent general endotracheal anesthesia with placement of continuous output Swan–Ganz catheter monitoring, transesophageal echocardiography (TEE) and arterial pressure monitoring. A thorough TEE evaluation was undertaken in each patient to assess wall motion abnormalities, presence and degree of mitral regurgitation, severity of atherosclerotic disease of the aorta, and right and left ventricular function. Significant atherosclerotic disease of the aortic arch and/or descending aorta prompted evaluation of the ascending aorta with epicardial echo probe. Presence of more than 2+ mitral insufficiency was a contraindication to off-pump revascularization. Those patients were placed on cardiopulmonary bypass and underwent conventional revascularization and mitral valve repair.
After harvest of conduits, the pericardium was opened. Extent of disease, epicardial anatomy (i.e. intramuscular), and size of target vessels as well as size of the heart were assessed. Presence of extensive calcification, deep intramyocardial vessels, small (<1.5 mm) targets, and/or significant cardiomegaly were relative contraindications to proceed with off-pump grafting.
After the decision had been made to undertake off-pump grafting, four deep pericardial sutures with snare protectors were placed in a straight line between the left inferior pulmonary vein and the inferior vena cava. The snare protectors served to avoid erosion of the taut sutures through the epicardium. The right pleura was routinely opened and pericardial sutures were not used on the right side of the heart.
In general, distal anastomoses were performed first and the aorta was partially occluded only once. In the rare occasion, a proximal anastomosis was performed first if a severely diseased aorta required use of a facilitating anastomotic device. In most instances, the left internal mammary artery to left anterior descending artery (LAD) anastomosis was performed first unless a totally occluded right coronary artery was present, in which case the latter would be revascularized first. Full systemic heparinization and complete protamine reversal were used in most instances. No antifibrinolytic therapy (i.e. aminocaproic acid, aprotinin) was used. Intracoronary shunts were rarely used. No partial bypass circuits or adjunctive retracting devices were used.
For patients undergoing conventional revascularization, the operation was conducted under moderate hypothermia (32°C) with intermittent anterograde and retrograde cold blood cardioplegia and with only one cross-clamp application.
Intraoperative doppler graft flow assessment (Medi-Stim Butterly Flowmeter, Medtronic, MN) was performed on a selective basis at the discretion of the surgeon. Depending on the electrocardiographic (EKG) tracing, transesophageal wall motion assessment and visual and manual inspection of the graft, failure to obtain flows of greater than 15 ml/min, and pulsatility indices between 1 and 5 usually led to revision of the anastomosis (Fig. 2 ).
One hundred patients with LVEF ≤30% underwent OPCAB and were compared with 110 conventional CABG patients. Intraoperative details are summarized in Table 2 . When possible, arterial conduits were used (either the internal mammary and/or the radial artery was used in 86% of OPCAB and 89% of CABG cases). There was no difference in the use of balloon counterpulsation, prevalence of redo cases, or number of grafts performed. Off-pump cases were completed with a statistically significant decrease in operative time.
The incidence of adverse events in the study group is displayed in Table 3 . These are compared with STS database benchmarks for all CABG patients regardless of left ventricular function. Median postoperative length of stay was 7 days (interquartile (IQ) range 6–10 days) in the study group compared with 5 days (4–7 days) for all CABG patients in the STS database. The incidence of adverse events in the OPCAB patients closely mirrored that seen in the STS benchmark data. Mechanical ventilation greater than 24 h, however, was much more prevalent (26% in OPCAB patients compared with 5.5% of STS all CABG patients, P<0.001).
Comparison of OPCAB and conventional CABG patients is depicted in Table 4 . The only statistically significant difference was the development of postoperative renal failure, which was more prevalent in the CABG patients (16 vs. 3%, P=0.003). Six OPCAB patients were readmitted within 30 days of surgery. Reasons included sterile sternal dehiscence (1), sternal wound infection (2; 1 superficial, 1 deep), leg wound infection (1), near syncope (1), and recurrent angina (1). Of the eight CABG patients readmitted, three returned for wound complications (1 sternal abscess, 1 superficial sternal wound, and 1 leg wound), two for recurrent angina, one with heart failure, and two with pulmonary complaints.
Quality of life
Telephone follow-up was done for the OPCAB patients and was 93% complete (Table 5 ). Median follow-up time was 14.3 months (IQ range 8.3–20.1 months). Freedom from recurrent angina was 83%, freedom from hospital readmission was 78%, and freedom from readmission for cardiac causes was 88%. Two patients underwent recatheterization. No patient required repeat surgical revascularization or intervention. When asked if they would have surgery again knowing what the process entailed, 75% answered yes, 8% answered no, and 5% were unsure. Twelve patients could not be reached. This survey revealed 13 late deaths – 11 of which were cardiac related (one was due to stroke and the other cause was unknown by the respondent).
Survival follow-up was complete in 99% of both off-pump and on-pump patients. One OPCAB and two CABG patients were not American or Canadian citizens and were neither included in the National Death Index nor could be reached by phone. These patients were censored in the analysis. Crude in-hospital mortality for the OPCAB patients (those occurring during the same hospitalization as the surgery) was 3% while the predicted (STS) in-hospital mortality was 5.3% with an observed to expected (O/E) ratio of 0.56. Individual predicted STS mortalities ranged from 0.41 to 50.9%. The three in-hospital deaths were due to low output syndromes, one on the day of surgery and two on postoperative day 4. There were 12 (10.9%) in-hospital deaths in the CABG patients – 10 were due to cardiac causes and two to pulmonary emboli. Kaplan–Meier estimates of survival are depicted in Fig. 3 . Six- and 12-month survivals were 90 and 85% in OPCAB patients and 82 and 75% in the CABG patients (P=0.592).
The introduction of enabling technologies in the mid-1990s paved the way for the dissemination of beating heart revascularization. Prior to this, OPCAB procedures had been largely in the realm of a few pioneering groups in South America [5,6] and limited to patients with one- or two-vessel (non-circumflex) disease.
Most of the literature addressing beating heart revascularization is focused on the application of this technology to low-risk patients. In fact, the six published randomized trials comparing coronary bypass grafting with and without cardiopulmonary bypass [7–11] did not include any patients with severe left ventricular dysfunction. In a recent prospective, randomized trial comparing OPCAB and CABG patients, Nathoe et al.  concluded that off-pump surgery was more cost effective than on-pump and that there was no significant difference in cardiac outcome between the two groups. Yet they too, excluded patients with poor left ventricular function. This may reflect a skepticism regarding the safety of off-pump surgery in patients with compromised left ventricular function. We feel that the greatest benefits derived from avoidance of extracorporeal circulation will become manifest not in low-risk individuals, but in patients with significant comorbidities, but this remains unproven.
Extensive literature exists evaluating the outcomes of conventional bypass grafting in patients with severe left ventricular dysfunction [13–15] but limited number of such patients have been submitted to off-pump revascularization. Sternik et al.  retrospectively compared 64 patients operated on without cardiopulmonary bypass with 53 who underwent conventional operation. Mean number of grafts was much lower in the off-pump group (1.9 grafts/patient) versus the conventional group (3.5 grafts/patient) with almost no patient receiving circumflex revascularization in the former group. The early and midterm survivals were better for the off-pump cohort but this probably reflects the lesser severity of disease present in the off-pump population. Another report from the same group  again investigating the safety of beating heart approaches in patients with compromised left ventricular function described 75 patients with ejection fraction less than 35%. There was a 2.7% mortality rate and 1.3% stroke rate. Again, a mean of 1.9 grafts were performed, suggesting less extensive disease or incomplete revascularization.
Arom et al.  compared in a retrospective fashion 45 OPCAB versus 132 conventional CABG recipients with ejection fraction less than 30%. The authors highlighted, as we do, that cardiac displacement is well tolerated in these patients. Fewer grafts were performed in the OPCAB group, which benefited from reduced postoperative blood loss. The incidence of adverse events in both groups was comparable for all variables studied. Meharwal and Trehan  published the largest study to date. They reported on 355 patients, who underwent OPCAB, with ejection fractions of 30% or less and an average age of 58 years. Mean number of grafts was 2.8 (versus 3.3 in the on-pump cohort) and mortality was 3.9%. The incidence of atrial fibrillation, prolonged ventilation, and hospital stay were reduced in off-pump recipients. This group of patients from India, appear to be younger and with less triple vessel disease than that reported in most CABG series in the recent literature.
In an effort to enhance the limited experience available describing patients with compromised left ventricular function undergoing beating heart coronary bypass, we report on the use of off-pump coronary bypass in 100 patients with severe (LVEF≤30%) ventricular dysfunction. Unlike the aforementioned studies, our population of patients had a very high incidence of triple vessel disease. Moreover, the mean number of grafts performed in our patients (3.5 per patient) is higher than that achieved in the literature and is suggestive of more complete revascularization. In contrast to these earlier studies, revascularization of the circumflex territory was routine. We were very pleased with the low incidence of adverse events which parallel the benchmark incidence of these events for all CABG patients, regardless of ejection fraction, in the STS database. We presume that the incidence of adverse events for patients with LVEF≤30 in the database would be much higher, but these data were not available to us. We did not fare so well with early extubation, probably a result of the high use of balloon pumps and inotropes in the early postoperative setting in combination with a high prevalence of active or recent smoking (77%) in our population of patients.
Only three perioperative deaths (3% mortality) were observed despite an expected 5.3% mortality. The incidence of early hospital readmission was remarkably low and was mostly related to wound complications. No instances of acute graft occlusion occurred and no patient required percutaneous intervention or a reoperation.
The low perioperative mortality and incidence of adverse events in this series can be attributed to several factors. First is the achievement of multivessel revascularization. Completeness of revascularization has been found to improve early survival in young  and elderly  patients and we believe it to be a critical factor in patients with left ventricular dysfunction. Second, the liberal use of balloon pump support stabilizes hemodynamics and allows the cardiac mobilization necessary for exposure of lateral and inferior wall targets. It is of interest that Dietl et al.  documented an improved survival for the cohort patients with ejection fractions less than 25% who underwent conventional CABG with preoperative balloon insertion as compared with those who did not receive counterpulsation support. Finally, and perhaps most importantly, is the presence of a dedicated cardiac anesthesiologist constantly vigilant of hemodynamics and progress of the operation who can alert the surgeon of the need to return the verticalized heart to a more physiological position and to rapidly create a proximal anastomosis or to insert a shunt for hemodynamically compromising ischemia. Moreover, the experienced off-pump anesthesiologist is adept at using inotropes and pressors as necessary based on hemodynamics and continuous TEE evaluation of the degree of mitral regurgitation and can anticipate the next step in the operation and adjust vasoactive drips and bed positioning as necessary to ensure a smooth operation.
On follow-up, most of the patients were free from angina, rehospitalization and recatheterization at a median of 14.3 months postoperatively. When asked if they would go through the procedure again, the majority of patients responded affirmatively. The 88% 1-year survival and good quality of life achieved in this ill group of patients lends validity to the safety and efficacy of beating heart revascularization for patients with severely compromised left ventricular function. Though the survival curve of the off-pump patients appears to indicate slightly better early survival than the conventional CABG cohort, no statistical difference was demonstrated.
While the present study represents the largest series of patients with severe left ventricular dysfunction undergoing OPCAB with multivessel revascularization, several limitations of the study deserve mention. This study was a retrospective review and is prone to the limitations of this design. As with many other surgical undertakings, we doubt that a randomized trial evaluating the safety and efficacy of OPCAB in severe left ventricular dysfunction will ever be realized and the best alternative is the accumulation of large series of patients with prospectively collected data. A second, and probably most profound limitation is the introduction of selection bias. The decision to proceed with OPCAB or conventional CABG was the sole decision of the operating surgeon made at the time of operation based usually on anatomy or clinical findings. Recently, however, we approach all patients for revascularization as potential off-pump candidates and tend to select patients for OPCAB preferentially when we think they would benefit most from elimination of cardiopulmonary bypass. These tend to be the sicker, more morbid patients. Furthermore, a longer follow-up period and the addition of formal instruments to measure quality of life would be of great interest to enhance our understanding of the benefits of off-pump revascularization in this ill group of patients.
In conclusion, OPCAB is a safe and effective approach for revascularization of the patient with severe left ventricular dysfunction. Liberal use of intraaortic balloon pump is recommended and multivessel revascularization is achievable and desirable. To confirm the benefit of this approach, long-term follow-up is mandatory but we believe that every coronary surgeon should have equal expertise in on- and off-pump revascularization so that the bypass operation can be optimally tailored to any individual patient, regardless of anatomy or comorbidities.
The authors gratefully acknowledge Patricia Garland, RN for her indispensable assistance with maintenance and querying of the CAOS database.
- angina pectoris
- ventricular function, left
- left ventricular ejection fraction
- coronary artery bypass surgery
- coronary revascularization
- ventricular dysfunction, left
- internal thoracic artery
- balloon dilatation
- patient readmission
- surgical procedures, operative
- tissue transplants
- quality of life
- thoracic surgery specialty
- ejection fraction
- coronary artery bypass, off-pump
- adverse event
- perioperative period