| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Joseph B. Whitehead Department of Surgery, Division of Cardiothoracic Surgery, Carlyle Fraser Heart Center, Cardiothoracic Surgery Clinical Research Unit, University School of Medicine, Atlanta, Georgia
b Department of Biostatistics, Rollins School of Public Health, Emory, University School of Medicine, Atlanta, Georgia
Accepted for publication June 5, 2007.
* Address correspondence to Dr Lattouf, Emory University School of Medicine, Emory Crawford Long Hospital, 6th Floor, Medical Office Tower, 550 Peachtree St NE, Atlanta, GA 30308. (Email: omar.lattouf{at}emoryhealthcare.org).
Presented at the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29–31, 2007.
| Drs Lattouf, Puskas, and Guyton disclose that they have a financial relationship with Medtronic; Drs Puskas and Thourani with Boston Scientific.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Methods: Emory Hospitals’ prospective database was retrospectively reviewed for 11,413 consecutive, isolated, primary coronary revascularization procedures between January 1997 and May 2005. Patients were divided into four groups: OPCABG 1 to 3 grafts (n = 3,187), OPCABG 4 to 7 grafts (n = 1,305), ONCABG 1 to 3 grafts (n = 3,279), and ONCABG 4 to 7 grafts (n = 3,642). A propensity score for surgery type was estimated from 39 risk factors. Multivariable logistic regression examined independent impact of surgery type and number of vessels grafted on outcomes. Computed interactions determined whether the effect of surgery type on risk-adjusted outcomes was consistent across groups.
Results: Patients requiring 4 to 7 grafts had adjusted odds of receiving ONCABG 2.92 times higher than patients requiring 1 to 3 grafts (p < 0.001). The OPCABG patients had adjusted odds ratios of 0.53 for death (p = 0.007), 0.42 for stroke (p < 0.001), 0.51 for major adverse cardiac events (p < 0.001), and 0.71 for renal failure (p = 0.05) as compared with ONCABG patients. The interaction between OPCABG and number of vessels grafted was not statistically significant.
Conclusions: This study demonstrates that surgeons tend to perform OPCABG for patients requiring 1 to 3 grafts and ONCABG for those requiring 4 to 7 grafts. Off-pump CABG is associated with reduced adjusted risk of adverse outcomes compared with ONCABG. This benefit is consistent for patients requiring 1 to 3 or 4 to 7 grafts.
In the last decade, interest in beating-heart techniques has experienced resurgence in an attempt to further decrease the morbidity associated with coronary artery bypass graft surgery (CABG) without jeopardizing the benefits. Technologic advancements with improved monitoring, hemodynamic support, and utilization of newly introduced positioning and stabilizing devices have significantly facilitated the performance of beating-heart surgery compared with prior experience of 30 years ago [1–3]. Published results of clinical trials have demonstrated improved coronary revascularization outcomes without cardiopulmonary bypass (off-pump CABG [OPCABG]) compared with coronary revascularization with cardiopulmonary bypass (On-pump CABG [ONCABG]) [4–6]. The increasing evidence based on multiple studies, that OPCABG is a safe and effective method for treating coronary artery disease, has led surgeons to broaden its application in their practice.
Although retrospective reviews have demonstrated that OPCABG is associated with decreased risk-adjusted morbidity and mortality compared with ONCABG [7, 8], the overall rate of OPCABG technique utilization remains relatively low. In 2004, US surgeons performed approximately 20% of all coronary bypass operations off pump. The continued interest in off-pump techniques is further supported by more recently published randomized prospective trials that have demonstrated improved outcomes after OPCABG [4–6, 9], thus giving added support to the notion of increased patient safety without the utilization of extracorporeal circulation. To date, no study has addressed or specifically focused on outcomes among patients who required extensive coronary revascularization (4 or more vessels bypassed). Heretofore, it is not known whether the number of grafts required impacts the decision of surgical technique (OPCABG versus ONCABG) or whether risk-adjusted outcomes are similar for OPCABG and ONCABG in patients receiving either fewer grafts (1 to 3) or a greater number of grafts (4 to 7). The present study was designed to assess in-hospital outcomes for patients having coronary revascularization to determine whether multivessel OPCABG is as safe as few-vessel OPCABG and to compare the results to similar cohorts of patients who underwent ONCABG.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Figure 1 shows the trends in surgical procedure type over time by vessel group. While emergency coronary artery bypass cases were included, those with redo coronary revascularization and those having revascularization combined with valve-related procedures or other cardiac surgery were excluded. Patient status, elective versus urgent or emergent, was not included in the propensity score adjustment. The propensity score is designed to address the potential selection bias associated with treatment assignment by accounting for preoperative variables that may predict treatment assignment. By far the most significant predictor of treatment assignment (off pump versus on pump) was surgeon identity. Additionally, variables that are likely associated with status are included in the propensity score. Thus, at worst, the only disadvantage to excluding status is a very negligible loss of predictive accuracy for the propensity score, which itself is already highly discriminative without status (concordance index of 0.91). Nevertheless, we examined the association between status and surgery type and saw that 81.9% of emergent cases were performed on-pump, compared with only 61.6% of elective surgeries and 44.9% of urgent surgeries (p < 0.0001). However, death rates were much worse in emergent patients with on-pump surgery (7.4%) versus off-pump surgery (1.0%, p = 0.014). Therefore, we believe that the exclusion of status does not change our conclusions.
|
Interventions, Surgeons, and Surgical Technique
Each patient underwent one surgical session consisting of OPCABG or ONCABG, performed at the discretion of any of 14 faculty surgeons, who varied in their adoption of off-pump surgery. Off-pump CABG was performed with one of several commercially available cardiac positioning and coronary artery stabilizing devices, using techniques previously described [5]. Conventional ONCABG was performed with standard techniques, utilizing roller head pumps, membrane oxygenators, cardiotomy suction, arterial filters, cold antegrade and retrograde blood cardioplegia, and moderate systemic hypothermia (30°C to 34°C). Patients who were converted intraoperatively from OPCABG to ONCABG or from ONCABG to OPCABG were entered into the database and analyzed according to the operation they ultimately received.
Data Management and Statistical Analysis
All data for consecutive patients were prospectively entered into a computerized cardiac surgical database, utilizing the data fields and definitions of the STS National Adult Cardiac Database. Data were managed by local database staff, warehoused in locked, secure facilities, and protected by computer passwords and firewalls.
Before analysis, 39 candidate risk factors were selected and deemed potentially important predictors of surgical group assignment (OPCABG or ONCABG) and clinical outcomes. Table 1 lists these risk factors. Diabetes mellitus was classified into three categories: none, noninsulin dependent, and insulin dependent. Further, the STS database management has developed a proprietary mathematical model that allows estimation of mortality risk based on preoperative patient characteristics. In this paper, we have benchmarked our morbidity and mortality results on this widely reported and literature-validated STS predicted risk model [10].
|
The retrospective nature and low event rates of the study prompted us to use propensity scoring statistical methods to "balance" patients with respect to the effect of their preoperative risk factors on their probability of group assignment (OPCABG versus ONCABG) [8, 12]. Importantly, the number of distal vessels bypassed in each individual patient was not used to estimate the propensity score, but was used in the final outcome models, because its direct influence on the outcomes was of primary interest. As such, 39 risk factors (including surgeon identity) and 13 binary indicators of data missingness were used nonparsimoniously in a multivariable logistic regression (MLR) model to predict surgery type assignment. The resulting estimated probability of being treated with OPCABG, conditional on the risk factor values, is the propensity score. The propensity score was included as a covariate of risk in the final MLR models of outcomes [8, 12].
This study is designed to identify whether the number of grafts interacts with surgery type to confer some benefit or detriment to patients. In evaluating number of grafts for its effect on outcome, we had to exclude the potential that other associated covariates may have influenced the choice of surgery type (by propensity scoring). After this selection bias is addressed, the "pure" effect of the number of grafts can be examined, along with its potential interaction with surgery type. Therefore, we have not excluded the possibility that the surgeon’s choice is influenced by number of grafts; rather, we have adjusted for all other measurable determinants of surgeon choice so that number of grafts may, in an unbiased fashion, be evaluated with respect to outcome.
To address the primary study endpoints, separate MLR models were constructed that regressed the dichotomous outcomes (death, cerebrovascular accident, MI) and their composite (major adverse cardiac events, MACE) as a function of five risk factors: surgery type (OPCABG versus ONCABG), vessel group (1 to 3 or 4 to 7), the interaction of surgery type and vessel group, surgeon identity, and lastly, propensity score. The diagnosis of postoperative MI was made in accordance with the published STS definition of postoperative MI. Surgeon identity was modeled as a random effect as it is assumed that patients are "clustered" under surgeons in a random manner. The treatment of surgeon as a random effect allows for the generalization of the model inferences to other populations.
These final models were used to assess the association between the outcomes and the predictors and to generate adjusted odds ratios (OR) of exposure to adverse postoperative events. Eight preplanned comparisons for each of the six MLR models were of interest: (1) OPCABG (1 to 3 grafts) versus OPCABG (4 to 7 grafts); (2) ONCABG (1 to 3 grafts) versus ONCABG (4 to 7 grafts); (3) OPCABG (1 to 3 grafts) versus ONCABG (1 to 3 grafts); (4) OPCABG (4 to 7 grafts) versus ONCABG (4 to 7 grafts); (5) OPCABG (1 to 3 grafts) versus ONCABG (4 to 7 grafts); (6) OPCABG (4 to 7 grafts) versus ONCABG (1 to 3 grafts); (7) all OPCABG versus all ONCABG; and (8) all 4 to 7 grafts versus all 1 to 3 grafts.
The data were managed and analyzed with SAS software (version 9.1; SAS Institute, Cary, North Carolina) and Stata software (version 9.0; Stata Corp, College Station, Texas). All statistical tests were evaluated using a two-tailed 0.05 level of significance. All comparisons and model terms were preplanned. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
|
|
for interaction p values). Because no significant interactions exist, the investigators concluded that OPCABG benefited patients in terms of death (p = 0.007, OR = 0.53), stroke (p < 0.001, OR = 0.42), MACE (p < 0.001, OR = 0.51), and renal failure (p = 0.05, OR = 0.71). Surgery type was not predictive of MI (p = 0.28, OR = 1.47) or infection (p = 0.14, OR = 0.80). Further, vessel group was not statistically predictive of any outcome (Tables 5 and 6).
|
|
Observed Versus Expected Mortality
Our data analysis indicated that the 1 to 3 OPCABG vessel group had an observed to expected (O/E) ratio of 0.519 as compared with the 1 to 3 ONCABG vessel group’s O/E ratio of 1.133. Likewise, in the 4 to 7 vessel group, the O/E ratio was 0.789 in the OPCABG group as compared with 1.214 in the ONCABG group (Table 7). Further analysis of the data in relation to the number of grafted vessels and surgery type correlated to O/E mortality ratio revealed consistently more favorable O/E ratio in each of the vessel groups treated off pump (Table 8).
Figure 2
further illustrates the relation between number of vessels grafted and the O/E ratio.
|
|
|
Surgeon identity
Fourteen different academic surgeons performed the 11,413 CABG surgeries analyzed in this report over 8 years. One hypothetical explanation for improved surgical outcomes after OPCABG versus conventional ONCABG might be that the surgeons performing OPCABG surgery are more skilled than those performing the older, conventional operation, rather than any inherent benefit of the technique itself. To address this theoretical concern, surgeon identity was included both in the propensity score and in the final MLR models as a possible confounder of outcomes. The final MLRs are adjusted for the random effect of surgeon identity and thus the odds ratios for group comparisons in Tables 5 and 6 reflect this adjustment.
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Retrospective studies as well as prospective randomized clinical trials have repeatedly shown improved outcome for the off-pump patients as compared with the on-pump patients. In our institution, where off-pump techniques have been utilized regularly for a decade, an increasing percentage of our patients have been revascularized using the off-pump techniques. In a prospectively randomized trial, Puskas and colleagues [5] have shown the noninferiority of the off-pump technique as compared with the on-pump approach.
In this study, we conducted a retrospective review of prospectively collected preoperative risk factors and postoperative outcomes on 11,413 consecutive patients who underwent isolated, primary coronary artery bypass surgery. Of interest to note, as indicated in Figure 1, was that our utilization of the OPCABG technique in the 1 to 3 vessels group exceeded the ONCABG technique in the year 2000, whereas in the 4 to 7 vessels group, it took 3 additional years before our OPCABG volume exceeded the ONCABG volume. One could postulate that this observation may be explained by the gradual maturation of experience and surgeon’s confidence, thus leading to increased application of the OPCABG techniques to the benefit of patients needing 4 or more vessels grafted.
All patients who underwent isolated first time revascularization since the gradual adoption of OPCABG techniques that began in 1997 were studied. The focus of this study was to investigate the outcome of the OPCABG patients requiring 4 or more vessel bypass operations as compared with those undergoing 4 or more grafts utilizing the ONCABG technique. Multivessel OPCABG patients were similar in age and had equivalent STS predicted risk of mortality as the ONCABG cohorts. Although the number of diseased vessels was equivalent between the two groups (3.79 vessels; Table 2), there was higher incidence of incomplete revascularization in the OPCABG group than in the ONCABG group (1.14 versus 1.20, p < 0.001). Although the predicted risk of mortality was equivalent in both multivessel OPCABG and ONCABG, the observed outcome in the OPCABG group demonstrated decreased incidence of postoperative mortality (Table 7). Additionally odds ratio analyses were statistically significant in favor of multivessel OPCABG in stroke and MACE and trended in favor of multivessel OPCABG in renal failure and infection. Of interest, the entire OPCABG group of 4,492 patients had a higher STS predicted risk as compared with the ONCABG group of 6,921 patients (Table 4), yet the observed postoperative stroke rate and MACE rate was statistically higher in the ONCABG group, thus establishing in this series the increased margin of safety of the off-pump techniques.
Further analysis of the data with specific attention to the patients who required few vessel bypasses (1 to 3 grafts) revealed there was a decreased incidence of postoperative renal failure, stroke, and MACE even though the STS predicted risk was higher in the 1 to 3 vessel OPCABG group (Tables 5 and 6). Examining the entire patient population, we found that more single-vessel patients were revascularized off pump rather than on pump. Of the 632 patients with single-vessel disease, 506 patients were treated with OPCABG whereas 126 were revascularized ONCABG. In the multivessel patients—those requiring 2 or more grafts—patients with two-vessel disease were more frequently revascularized off pump (1,045 OPCABG versus 777 ONCABG). In the three or more vessels patient population, there was overall more ONCABG utilization rather than OPCABG (2,490 OPCABG versus 6,057 ONCABG; Table 8). However, if one examines the trends in surgical procedure type over time by vessel group, one finds that there was gradually increasing shift toward OPCABG utilization in the 1 to 3 vessel group as well as in the 4 to 7 vessel group (Fig 1) By year 2000, 3 years after in the launching of our OPCABG program, 50% or more of the 1 to 3 vessel revascularization patients were treated by OPCABG; and by the year 2003, 6 years after the program initiation, 50% or more of the 4 to 7 vessels group were revascularized off pump. These findings allude to the gradual gaining of experience and confidence in the OPCABG techniques, as noted by the gradual shift from on-pump to off-pump techniques.
This dataset has confirmed the increased safety in the OPCABG as compared with ONCABG in the entire OPCABG group (1 to 7 grafts) as well as for the 1 to 3 grafts OPCABG as compared with the 1 to 3 grafts ONCABG group; and most importantly for the purpose of this study, comparison of the adjusted odds ratios for adverse outcomes between the 4 to 7 grafts OPCABG versus the 4 to 7 grafts ONCABG revealed statistically significant difference in favor of OPCABG in stroke, MACE, and infection. Additionally, trending toward significant difference was noted in death and, to a lesser extent, renal failure in favor of OPCABG. That was independent of 39 other risk factors (including surgeon identity) that were included in the multivariate and propensity score analyses used to adjust for preoperative differences between patients. In this paper, no effort was made to address the intermediate- or long-term outcome of OPCABG versus ONCABG experience. Further studies are currently ongoing to examine if the short-term morbidity and mortality benefits of the OPCABG techniques are maintained at the intermediate- and long-term levels.
In 2005, in the United States, in accordance with data published by The Society of Thoracic Surgeons, 144,967 patients underwent isolated coronary artery bypass operations [13]. Of these, 116,316 patients received ONCABG surgery (2.1% mortality rate) and 28,651 received OPCABG (1.6% mortality rate). It would be of great interest to examine the aforementioned STS database and examine whether our findings of this single institutional increasing trends for OPCABG utilization and the observed decreased morbidity and mortality are in line with nationally definable outcomes.
This study has several limitations. Its retrospective nature does not permit complete accounting for all sources of bias, despite advanced statistical methodology designed to correct for both treatment selection bias and potential confounders of outcomes in preplanned analyses. In addition, the database utilized in this study reported surgical cases according to the ultimate surgery type performed. This meant that patients whose coronary revascularization was initially attempted without cardiopulmonary bypass and who required conversion to cardiopulmonary bypass (typically due to hemodynamic instability) were included in the ONCABG group. This may disadvantage ONCABG in comparison of outcomes with OPCABG. Reciprocally, patients converted from ONCABG to OPCABG (usually owing to intraoperative discovery of severe aortic atherosclerosis) were included in the OPCABG group. Additionally, our data analysis did not address the extent of coronary calcification and the utilization of coronary endarterctomy in each group or the incidence of conversion from OPCABG to ONCABG in the occasional cases of severe calcified coronary arteries. Such patients’ potentially increased incidence of complications may disadvantage one method for revascularization in comparison with the other. The database does not allow reconciliation of these data to an intention-to-treat analysis. Fortunately, intraoperative conversion is an infrequent event, affecting approximately 2% of cases [5]. Finally, although the 14 surgeons who performed coronary revascularization in this study varied greatly in their interest in OPCABG (several rarely performed OPCABG, whereas several used OPCABG in the majority of their cases), Emory University has maintained a strong institutional interest in OPCABG since 1997. Thus, faculty surgeons in the present study may have average experience with OPCABG that exceeds the national norm, limiting the ability to generalize these results.
In conclusion, patients undergoing multivessel coronary artery bypass surgery are at increased risk of death, stroke, or MI, or the composite endpoint of death/stroke/MI compared with patients who need fewer diseased vessels. The OPCABG techniques have been shown here to provide a greater short-term safety margin in the few-vessels as well as in the many-vessels bypass group. The idea that patients who need multivessel revascularization grafts (4 to 7) are better served by undergoing on-pump surgical revascularization rather than off-pump revascularization can no longer be supported. In fact, our results attribute a greater short-term safety margin to the OPCABG techniques for patients who need extensive revascularization, and the perceived safety offered by cardiopulmonary bypass cannot be supported by our data (Fig 3).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
As with all single-institution studies, though, a central question is, do the results from this single institution apply to a wide range of centers? To investigate this, we used the STS database to examine the total number of patients undergoing surgical revascularization in calendar year 2005. We conducted an analysis that was essentially identical to the analysis reported by the authors with the exception that we used the STS models rather than the propensity matching that they used. We found virtually the same thing that they found on a national multi-institutional level.
Of course, it is quite important to reinforce the general application of your findings, but I am still a bit reluctant to make this recommendation to use off-pump surgery to the cardiac surgery community as a whole. Because of the increased technical challenges associated with off-pump surgery, it is likely that there probably is some bias, not only in their data but in the STS data as well. The more technically adept surgeons are still more likely to be the ones that are performing this operation. So I am reluctant to recommend off-pump surgery to the "average" cardiac surgeon, and I think you can see the dilemma here. That leads me to ask if you have come up with a way to account for technical skill in the equation when you evaluate off-pump versus conventional surgery? If not, how can we justify recommending the use to all? And if we don’t recommend the use to all, then what are you recommending from the results of your study?
DR LATTOUF: Doctor Edwards, thank you so kindly for your comments. We agree with you that there is a learning curve for the surgeon to become efficient and comfortable with the utilization of the off-pump technique—not only that the surgeon must become efficient and comfortable with it, but also that the entire operating room team, and particularly the cardiac anesthesiologist, has to become efficient and capable of addressing the minute-to-minute variation in the hemodynamics. And I believe, and it is my personal conviction, that the role that the cardiac anesthesiologist plays is exceptionally important to support the patient, because after all, the surgeon can control his or her abilities and you need the anesthesiologist to control his and her abilities.
We agree also that there are variations even within any one institution about the capability and the interest of the different surgeons in the utilization of this technique, and that is indeed reflected in our institution and probably in our results.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. D. Puskas, V. H. Thourani, P. Kilgo, W. Cooper, T. Vassiliades, J. D. Vega, C. Morris, E. Chen, B. J. Schmotzer, R. A. Guyton, et al. Off-pump coronary artery bypass disproportionately benefits high-risk patients. Ann. Thorac. Surg., October 1, 2009; 88(4): 1142 - 1147. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Stassano, L. Di Tommaso, M. Monaco, S. Iesu, G. Brando, S. Buonpane, G. Ambrosio, G. Di Benedetto, and P. Pepino Myocardial Revascularization by Left Ventricular Assisted Beating Heart Is Associated With Reduced Systemic Inflammatory Response Ann. Thorac. Surg., January 1, 2009; 87(1): 46 - 52. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. M. Lattouf, V. H. Thourani, P. D. Kilgo, M. E. Halkos, K. T. Baio, R. Myung, W. A. Cooper, R. A. Guyton, and J. D. Puskas Influence of On-Pump Versus Off-Pump Techniques and Completeness of Revascularization on Long-Term Survival After Coronary Artery Bypass Ann. Thorac. Surg., September 1, 2008; 86(3): 797 - 805. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
