|
|
||||||||
a Division of Cardiothoracic Surgery, Saint Vincent Mercy Medical Center, Toledo, Ohio
b Yvonne Viens, SGM, Research Institute, Saint Vincent Mercy Medical Center, Toledo, Ohio
c Departments of Medicine and Surgery, University of Toledo College of Medicine, Toledo, Ohio
Accepted for publication March 27, 2009.
* Address correspondence to Dr Habib, Cardiovascular and Pulmonary Research, Yvonne Viens, SGM, Research Institute, St. Vincent Mercy Medical Center, 2222 Cherry St, MOB2, Suite 1250, Toledo, OH 43608 (Email: robert_habib{at}mhsnr.org).
Presented at the Poster Session of the Forty-fifth Annual Meeting of The Society of Thoracic Surgeons, San Francisco, CA, Jan 26–28, 2009.
| Abstract |
|---|
|
|
|---|
Methods: We compared late survival of 532 consecutive patients receiving sequential RA grafts (sequential RA group: 438 men; 462 with three-vessel disease) with that of a 4,131 contemporaneous internal thoracic artery (ITA) with saphenous vein (SV) multivessel CABG cohort (conventional group). Graft failure rates were determined from symptom-driven repeat angiography films in 122 sequential RA patients performed 2 to 4,317 days after surgery. Median survival sequential RA follow-up was 5.3 years (range, 0.5 to 12.3).
Results: The sequential RA patients received a total of 1,181 RA grafts (538 sequential [30 triple] and 75 single) along with 636 SV and 533 ITA. Overall RA graft failure (80 of 272; 29%) was intermediate to that for ITA (7 of 121; 5.8%; p < 0.001) and vein (54 of 133, 41.6%; p = 0.032) grafts. Sequential versus nonsequential RA failure did not differ (77 of 252 [31%] versus 3 of 20 [15%]; p = 0.202), while failure of the proximal (36 of 123; 29%) and distal (40 of 129; 31%) components of sequential RA grafts were essentially identical. A total of 69 deaths (6 operative; 1.1%) have occurred in the sequential RA cohort. Unadjusted 10-year sequential RA cohort survival was 76.2% overall, and 79.0% for the 454 primary isolated CABG subgroup. The risk-adjusted 10-year survival using a logit propensity score was substantially better for the sequential RA cohort versus the conventional CABG cohort (risk ratio [95% confidence interval] 0.61 [0.44 to 0.85]; p = 0.003).
Conclusions: Sequential RA grafting is a safe method for maximizing arterial revascularization and is associated with excellent 10-year survival that seems to be superior to conventional or ITA/SV CABG results. Also, the similar proximal and distal sequential RA patency mitigates concerns of a clinically significant effect of increased vasoreactivity of distal segments of RA conduits.
| Introduction |
|---|
|
|
|---|
The RA is ideally suited for use in sequential configuration and, consequently, it facilitates maximizing the number of constructed arterial grafts. Sequential grafting techniques have gained limited acceptance owing to technical issues, the dependence of multiple coronary beds on a single graft, and more challenging percutaneous approaches than single grafts in case of impending graft failure. Alternatively, there are compelling data supporting improved graft patency with sequential configurations in case of both saphenous vein (SV), especially in poor run-off coronary targets [6–8], and ITA grafts [9–11]. Sequential RA grafting has been described in reports focusing on all-arterial coronary revascularization with generally good results [4, 12–14]. Encouraged by these results, we extended our use of sequential RA grafting as a means of maximizing the number of arterial grafts. This study represents the initial report of the early and late outcomes of sequential RA grafts in a large contemporary CABG series.
| Patients and Methods |
|---|
|
|
|---|
Patients
The study population consisted of 532 consecutive surgical coronary revascularization patients (1996 to 2008) who received one or more sequential RA grafts at two Toledo, Ohio, area hospitals (Table 1). Patients with concomitant valve or aortic surgery were excluded. Primary isolated CABG was done in 454 of the 532 (85.3%) whereas the remaining 78 patients included 21 repeat CABG patients (3.9%), and 62 patients who underwent concomitant noncardiac (n = 21, including 16 carotid endarterectomy and 3 lung procedures) or nonvalvular cardiac surgery (n = 43; e.g., transmyocardial revascularization, atrial/ventricular septal defect repair, Maze procedures, ventricle wall repair). The annual volumes of sequential RA patients were 43, 24, 35, 42, 40, 44, 60, 66, 63, 37, 36, 30, and 12 for 1996 through 2008, respectively. A total of 6 surgeons performed these operations with widely varying frequency (n = 2, 10, 25, 32, 67, and 406), and with 1 surgeon accounting for 76% of all cases. Normothermic cardiopulmonary bypass was utilized in 527 patients (99%), whereas 5 had off-pump surgery. An isolated multigraft CABG comparison cohort was derived from the contemporaneous single ITA with additional saphenous vein (conventional: ITA/SV; n = 4,131) multivessel disease CABG population (Table 1).
|
|
|
Graft Patency
Symptom-driven repeat angiography films were obtained for 122 of 532 sequential RA patients, given the particular interest in assessing the performance of the sequential radial graft. The corresponding angiography data for the conventional ITA/SV patients were not analyzed. These data were derived from multiple sources, including cardiac surgery and cardiac catheterization databases. Angiography reports were reviewed and the relevant data entered into a dedicated database. For this study, a coronary graft was considered to be an anastomotic failure in case of complete or 100% occlusion, stenosis of 75% or greater, or in case of extensive conduit narrowing or string sign.
Sequential Radial Artery Use Propensity Model
The sequential RA and ITA/SV cohorts exhibited significant demographic and risk factor differences (Table 1). To minimize possible confounding of these differences on outcome comparisons, we used propensity score adjustment where sequential RA grafting was considered as treatment [4]. The propensity score was derived from a nonparsimonious logistic multivariate model based on all factors in Table 1 and applied to patients from the two groups. Here, given the near-universal left ITA use and the variance in number of arterial grafts (5 or fewer), we modeled arterial grafting through a single numerical variable (number of arterial grafts). The increase in arterial grafts is primarily determined by the number of RA grafts but also incorporated the infrequent use of the right ITA. Coronary disease and number of grafts were incorporated into the model through a completeness of revascularization index, defined as the difference between the number of grafts and vessel disease [4]. The completeness of revascularization index was greater than 0 for all sequential RA patients as a minimum of 3 grafts was completed. Time of surgery was also entered as a continuous month of series variable (January 1996 = 1) to account for the varying frequency of sequential RA CABG over time. Expectedly, the resulting propensity scores were distinctly different (mean ± SD: 0.492 ± 0.304 sequential RA versus 0.056 ± 0.118 ITA/SV; p = 0.0000). The propensity model C-statistic (area under the receiver operating characteristic curve) was 0.932 ± 0.006, indicating excellent discrimination.
Data Analysis and Statistical Methods
Continuous data were expressed as mean ± SD unless otherwise stated. Univariate comparisons were done with the
2 or Fisher's exact test for categorical variables and the unpaired t test for continuous variables. Unadjusted Kaplan-Meier survival plots were derived up to 10 years and were compared with the log rank (Mantel-Cox) test. Next, the logit of the sequential RA use propensity score model was used to derive the risk-adjusted comparisons. Early deaths (less than 30 days) were excluded to avoid violation of the proportional hazard assumption. For the sequential RA population, we also determined the effects of explanatory variables on survival by multivariate Cox proportional hazard analysis. Model selection was first done with backward elimination (Wald statistic), and variables significant at the p less than 0.1 level were retained in the model and confirmed using forward stepwise selection (SPSS, version 15.0 software; SPSS, Chicago, IL).
| Results |
|---|
|
|
|---|
|
|
|
|
|
|
| Comment |
|---|
|
|
|---|
Sequential SV coronary grafts have been extensively described. Their use is based on SV conduit conservation, and on a presumed favorable hemodynamics manifested as increased graft flow rates and flow velocities due to decreased resistance of sequential versus single grafts. O'Neil and colleagues [21] showed that sequential SV grafts anastomosed to two targets have half the resistance of a single graft. Less turbulence was noted at a side-to-side anastomosis than at an end-to-side anastomosis [22]. Long-term graft patency has been shown to be flow dependent owing to an inverse relationship between graft flow rate and the development of graft intimal hyperplasia [23]. High flow velocities have been documented in the proximal side-to-side anastomosis of sequential SV [24]. These mechanistic considerations have been reflected in the clinical arena by a number of reports documenting improved patency in a sequential versus single SV graft configuration [7]. Kieser and colleagues [8) reported a better patency of the proximal side-to-side anastomosis compared with the distal end-to-side anastomosis of sequential SV grafts as well as the end-to-side anastomosis of nonsequential SV grafts. Yet, they also found a better patency of a nonsequential SV end-to-side anastomosis compared with the distal end-to-side anastomosis of a sequential SV and recommended single over sequential SV grafts. A number of investigators have suggested that a grafting strategy aimed at placing the distal anastomosis of the SV to a target with the best run-off while placing the proximal anastomosis to a less ideal target optimizes the flow characteristics resulting in superior patency [7, 8, 21]. Others have not emphasized this, and found that placement of the proximal side-to-side anastomosis to a better and bigger target vessel did not decrease flow through the distal anastomosis to a smaller coronary bed [25]. Survival data in patients revascularized predominantly with sequential SV were equivalent to patients receiving single SV at 10 years in a multivariate analysis [26].
Evidence of similar patency of single and sequential ITA grafts mitigate concerns regarding the graft durability of the ostensibly more technically challenging sequential anastomotic technique or graft angulation and kinking [11]. Dion and associates [9] reported comparable acute early patency of sequential and single ITA grafts. Sterkenburg and associates [10] reported an early patency of triple sequential ITA grafts of 96.1% in 46 patients. Dion and colleagues [11] reported a 96% sequential ITA patency in 161 asymptomatic patients restudied at a mean interval of 7.4 years. Interestingly, a diamond proximal anastomosis was associated with a worse patency than a parallel proximal anastomosis. No differences in patency, however, were noted between proximal and distal anastomoses [11]. Their 10-year sequential ITA survival was 72%, yet there are no reports comparing late survival with sequential versus single ITA grafts.
The earliest report of sequential RA appeared in 1997. Weinschelbaum and colleagues [12] described their experience with sequential RA as part of their all-arterial revascularization strategy, and reported excellent perioperative outcomes with near perfect predischarge graft patency. Muneretto and colleagues [13] used sequential RA in composite configuration from the left ITA to facilitate all-arterial revascularization, and reported 1 RA graft occlusion in 100 patients. Their 2-year actuarial survival did not differ compared with ITA/SV CABG. Nakajima and colleagues [14] reported generally good angiographic results with off-pump all-arterial revascularization using sequential RA, but also raised concerns about coronary steal and graft flow reversal and recommended careful preoperative graft arrangement to minimize these complications. Their predominant inflow pattern was from the left ITA, which is in contradistinction to the present study in which 98% of the sequential RA were anastomosed to the aorta. Our repeat angiography results in symptomatic restudied patients indicated the following regarding radial graft patency: (1) sequential and nonsequential RA graft patencies are statistically similar; (2) patencies of the proximal versus distal anastomoses of sequential RA grafts are essentially identical; and (3) the overall patency of sequential RA grafts did not differ when the sequential graft involved a single versus multiple coronary vessel systems. The latter mitigates concerns that coronary steal phenomena in sequential grafts increases when coronary circulations become interconnected such as in the case of sequential grafts involving multiple coronary systems. These results stand in contrast to the results of Gaudino and colleagues [27] who found that the distal aspect of a RA graft was significantly more prone to spasm with the concomitant clinical finding of significantly less durable graft patency constructed from distal segments of RA compared with its proximal counterpart.
The flow dynamics reported in sequential SV grafts potentially predicts a similar favorable effect with RA. We did not observe an improved sequential versus single RA graft patency in our series, which may derive from the fact that only symptomatic patients were restudied and possibly reflects our operative strategy. This strategy was not aimed at targeting the biggest target with the highest degree of stenosis as the most distal anastomosis. Rather, the only criterion was to maximize the number of target vessels grafted with arterial conduits regardless of size and quality. Whether applying the former strategy leads to improved sequential RA patency rates, as for SV, remains untested. We are also unable to comment on differences between diamond versus parallel anastomoses, as the anastomotic technique was not available for query. We also failed to appreciate coronary steal phenomena between adjacent coronary targets connected by sequential grafts. Similarly, we did not encounter graft flow reversal as described by Nakajima and associates [14], which may be due to the 98% aortocoronary construction of our sequential RA as opposed to their RA inflow from the left ITA.
A multivariate analysis of long-term mortality after sequential RA CABG identified the well-known risk factors of diminished ejection fraction, diabetes mellitus, and advanced age. Surprisingly, we also found that male sex was associated with worse survival after sequential RA CABG compared with women. This finding remains unexplained but parallels the findings in the Bypass Angioplasty Revascularization Investigation (BARI) study [28], which revealed that women had a significantly lower risk of death at 5 years after CABG and after PTCA. In a previous analysis, we showed that women seem to have an earlier benefit from RA grafts with better survival seen earlier postoperatively compared with men [3]. In addition, Desai and colleagues [29] found, as we have, that although the RA graft failure was equivalent among men and women (8.6% and 5.3%, respectively at 1 year postoperatively), women had a much higher SV failure rate compared with men (23% and 12%, respectively). Increasing the proportion of substantially better patent RA grafts compared with SV in women as opposed to the less substantially better durability of RA grafts vis a vis SV in men may conceivably contribute to the relative improved long-term survival in women.
The limitations of this study include its retrospective nature, the unavailability of the sizes of the RA graft, the size of the coronary artery target, and the degree of target vessel stenosis. It has become relatively well accepted that RA graft patency is inversely related to the target vessel stenosis, most likely owing to competitive flow through the target vessel resulting in relatively diminished flow in the RA conduit, leading to premature graft loss. We speculate that this relationship is more complex and, indeed, depends on the absolute cross-sectional area of the target vessel at the site of the stenosis and the absolute cross-sectional area of the RA conduit. A 50% stenosis in a 4-mm target vessel grafted with a 2-mm RA graft is quite different from a fluid mechanics perspective than a 50% stenosis in a 2-mm coronary target grafted with the same 2-mm RA graft. We speculate that the degree of competitive flow, and consequently long-term graft durability, will be determined by the relative cross-sectional target vessel area at the level of stenosis to that of the RA graft. The independence of SV patency as a function of target vessel stenosis [29] supports this contention, as the relatively larger size of SV vis a vis the coronary target makes competitive flow phenomenon less important than in the case of RA grafts, the size of which are much more equivalent to the coronary targets and thus much more sensitive to competitive flow issues. This issue will need to be addressed in specifically designed future studies.
The seminal result of the present study is the substantial, 64% overall, reduction in long-term (10-year) mortality associated with sequential RA compared with the conventional ITA/SV CABG—which may be even greater in women. Also, similar proximal versus distal sequential patency mitigates concerns of a clinically significant effect of increased vasoreactivity of distal RA segments. Accordingly, we strongly recommend consideration of sequential RA grafting as a safe method of choice for maximizing arterial grafts in the current era of drug-eluting stents in which the success of catheter-based revascularization approaches that of the conventional CABG operation with ITA and SV grafts [30].
| Appendix |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
P. Achouh, K. O. Isselmou, R. Boutekadjirt, C. D'Alessandro, J.-Y. Pagny, R. Fouquet, J.-N. Fabiani, and C. Acar Reappraisal of a 20-year experience with the radial artery as a conduit for coronary bypass grafting Eur J Cardiothorac Surg, January 1, 2012; 41(1): 87 - 92. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Sugimura, M. Toyama, M. Katoh, M. Kotani, Y. Kato, and K. Hisamoto Outcome of composite arterial Y-grafts in off-pump coronary artery bypass Asian Cardiovasc Thorac Ann, April 1, 2011; 19(2): 119 - 122. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. F. Tranbaugh, K. R. Dimitrova, P. Friedmann, C. M. Geller, L. J. Harris, P. Stelzer, B. Cohen, and D. M. Hoffman Radial Artery Conduits Improve Long-Term Survival After Coronary Artery Bypass Grafting Ann. Thorac. Surg., October 1, 2010; 90(4): 1165 - 1172. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. H. Habib Reply to the Editor J. Thorac. Cardiovasc. Surg., October 1, 2010; 140(4): 941 - 942. [Full Text] [PDF] |
||||
![]() |
A. Zacharias, T. A. Schwann, C. J. Riordan, S. J. Durham, A. S. Shah, M. Engoren, and R. H. Habib Late outcomes after radial artery versus saphenous vein grafting during reoperative coronary artery bypass surgery J. Thorac. Cardiovasc. Surg., June 1, 2010; 139(6): 1511 - 1518. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. A. Schwann and R. H. Habib Reply Ann. Thorac. Surg., May 1, 2010; 89(5): 1705 - 1705. [Full Text] [PDF] |
||||
![]() |
D. Nezic, A. Knezevic, S. Micovic, and M. Jovic Sequential Saphenous Vein Conduits for Coronary Artery Bypass Graft Surgery: Long-Term Survival and Angiographic Results Ann. Thorac. Surg., May 1, 2010; 89(5): 1704 - 1705. [Full Text] [PDF] |
||||
![]() |
D. D. Glower Invited Commentary Ann. Thorac. Surg., July 1, 2009; 88(1): 39 - 39. [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 |