| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Surgery Unit, Magna Graecia University of Catanzaro, Naples, Italy
Accepted for publication June 6, 2007.
* Address correspondence to Dr Onorati, Viale dei Pini, 28, Napoli, 80131, Italy (Email: frankono{at}libero.it).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentReferences |
|---|
Methods: In all, 516 consecutive isolated coronary artery bypass graft surgery patients admitted between January 2003 and February 2007 were grouped according to the harvesting technique of LIMAs (387 pedicled versus 129 skeletonized). Intraoperative transit time flow results, hospital outcome, perioperative troponin I leakage, and echocardiographic results were analyzed. Maximum, mean, and minimum flows, pulsatility index, and graft flow reserve were stratified according to the type of surgery (on pump and off pump) and subgroups at risk (severe left ventricular hypertrophy).
Results: The transit time flow results were unsatisfactory in 4 patients (0.8%), all 4 of whom underwent redo anastomosis. Maximum, mean, and minimum flows and pulsatility index values were comparable between skeletonized and pedicled LIMAs (p = not significant), for both on-pump and off-pump surgeries (p = not significant). Graft flow reserve was comparable between the two groups (p = not significant), as well as transit time flow results in severe ventricular hypertrophy. Troponin leakage, postoperative echocardiographic recovery, and hospital outcome were comparable between the two groups (p = not significant at all time points). Only superficial wound complications proved higher in the pedicled group (2.6% versus skeletonized group: 0%; p = 0.05).
Conclusions: Pedicled LIMA carries a higher risk for minor wound complications. Neverthless, the two techniques showed similar in-vivo functional behavior and hospital outcome. Therefore, there is no reason to prefer one technique over the other.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Although the survival advantage of LIMAs compared with other grafts is accepted worldwide, no clear benefits of one harvesting technique over the other have been definitely demonstrated. In particular, a recent review by Athanasiou and coworkers [3] has shown the only accepted evidence, grade A, is the induction of a significant sternal devascularization with bilateral pedicled IMAs. Moreover, despite a number of studies that have addressed the superiority of the skeletonized LIMA, little is known about the in-vivo functional behavior of skeletonized and pedicled grafts. Only one study analyzed transit time flow (TTF) results of skeletonized and pedicled LIMAs [9]. Few studies have described the free-flow capacity of the graft, often after intragraft injection of papaverine [5, 6, 10]. No study has reported on the graft flow reserve (GFR) of these conduits.
Finally, some authors have shown that the risk of postoperative hypoperfusion syndrome, a rare perioperative condition associated with low cardiac output, left venticular failure, and cardiac arrest, is higher when pedicled LIMAs are used [10, 11]. The major pathophysologic factor for that syndrome seems to be a disproportion between LIMA flow and myocardial demand due to severe venticular hypertrophy [10, 11]. Intraoperative TTF analysis is, therefore, able to detect early such a dreadful postoperative complication [9].
Therefore, the aim of this study was to systematically analyze our experience with TTF and in-hospital results of pedicled versus skeletonized LIMAs in first-time myocardial revascularization, stratifying the results by the surgical technique employed (on pump and off pump), and associated risk factor (severe ventricular hypertrophy).
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Five hundred-twenty consecutive patients, admitted between January 2003 and February 2007 for first-time elective CABG, were enrolled and grouped to receive pedicled LIMA (group A, 389 patients) or skeletonized LIMA (group B, 131 patients) according to the surgeon’s choice. Additional cardiac or vascular surgical procedures and severe systemic comorbidities (dyalisis, hepatic failure, cancer, autoimmune disease) were exclusion criteria. Because the different geometry for the construction of a Y graft or T graft influenced TTF results, coronary revascularization with these grafts were excluded from the analysis.
Angiography
Preoperative coronary angiography of each patient was reviewed by two cardiologists, blinded toward the study, and Thrombolysis in Myocardial Infarction study (TIMI) scores were calculated.
Surgery
All surgical procedures were performed by the same surgeons (A.R., P.M., A.dV., F.O.). In all patients, CABG was performed through a median sternotomy, and the left anterior descending artery always grafted using the LIMA.
The pedicled LIMA was harvested together with the surrounding veins, muscle, and fascia. Low-voltage (20 mV) cautery was always used, and the side branches were clipped. Proximal dissection was always above the first rib. The width of pedicled LIMAs always ranged from 1 to 1.5 cm. To avoid any distortion to the graft, the pedicle was always secured with two epicardial stitches using 6-0 polypropylene on both sides after completion of distal anstomoses.
When the LIMA conduit was skeletonized, the technique was as follows. After having dissected the reflection of the mediastinal pleura from the endothoracic fascia, the mammary artery and both satellite veins were visualized. The fascia was incised medially to the medial mammary vein for the whole length of the vessel, and pulled down on the pleura. A blunt dissection of the artery from the chest wall was performed by means of low-voltage (10 mV) cautery or cold tip; the sternal and the anterior intercostal branches were occluded with hemoclips. Collaterals were always divided using low-voltage cautery, and its mammary side similarly occluded with small hemostatic clips.
Both pedicled and skeletonized LIMAs were cut at the distal end 1 cm proximal to bifurcation, and checked for a good pulsatile free flow (at least > 20 mL/minute) before their use. No vasodilators or other drugs were injected inside the LIMA at any time. Two LIMAs in group A and 2 skeletonized LIMAs in group B were not used as conduits because of unsatisfactory free flow after harvesting. These grafts were excluded from the study. Therefore, 387 patients in group A and 129 in group B were ultimately enrolled in the protocol. Distal anstomoses were always constructed using 8-0 polypropylene sutures.
A standard cardiopulmonary bypass circuit was used: a Dideco (Mirandola, Modena, Italy) tubing set, which included a 40 µm filter, a Stockert roller pump (Stockert Instrumente, Munich, Germany), and a hollow fiber membrane oxygenator (Dideco D903 Avant; Mirandola). Myocardial protection was achieved by using blood potassium cardioplegia as previously reported [12]. Of 516 patients enrolled, 116 (29.9%) of group A and 42 (32.5%) of group B underwent OPCABG. Exposure and stabilization were achieved with the Octopus-IV with or without the Starfish-II tissue stabilizer (Medtronic, Minneapolis, Minnesota). Intracoronary shunts were used routinely, with a success rate approximating 100%.
Flowmetric Analysis
In both groups, assessment of each graft was performed under stable hemodynamic conditions (mean arterial pressure at 70 to 80 mm Hg), generally 30 minutes after protamine administration. Flowmetry of the grafts was performed with a transit-time flowmeter (HT313 Transonic; Transonic Systems, Ithaca, New York). The curves were always coupled with the electrocardiographic tracing to correctly differentiate the systolic from the diastolic flow. The TTF measurements were interpreted as previously suggested by D’Ancona and colleagues [13]. The maximum, minimum, and mean flows were reported as mL/min, and pulsatility index (PI) as an absolute number [13]. Data from LIMA conduits were recorded and compared between the two groups.
Graft Flow Reserve
All patients with ongoing unstable angina despite intravenous nitrates with poor left ventricular ejection fraction (
40%) with or without severe narrowing of the left main stem or ST-segment elevation underwent intra-aortic balloon pump (IABP) before anesthetic induction, in order to improve myocardial protection during the induction of anesthesia and the early phases of the operation. In particular, 142 patients in group A (36.7%) and 51 (39.5%) in group B underwent preoperative IABP, and were, therefore, the objective of GFR study. According to the previous demonstration that IABP recruits GFR during assistance [14], mean flow and PI were recorded in all patients during IABP support and during temporary cessation. Both measurements were accomplished under stable hemodynamic conditions. The GFR was calculated from the mean flow assessed during 1:1 IABP support divided by mean flow at baseline (IABP off) [15].
Postoperative Care
Inotropes were started immediately after aortic cross-clamp removal with enoximone at a dosage of 5 µg · kg–1 · min–1 [12, 15]. The need to further increase inotropes was recorded. Inotropic support was defined as low dose when enoximone was administered at a dosage lower than or equal to 5 µg · kg–1 · min–1; medium dose was defined when enoximone was used at a dosage between 6 and 10 µg · kg–1 · min–1 or dobutamine was added at a dosage between 5 and 10 µg · kg–1 · min–1; high dose was when enoximone or dobutamine infusion was greater than 10 µg · kg–1 · min–1 or epinephrine at any dose was added.
Biochemical Analysis
Determinations of cardiac troponin I were conducted preoperatively and at 12, 24, 48, and 72 hours. Diagnostic kits provided by Beckman Coulter for troponin I were used (Access Immunoassay System; AccuTnI). Criteria for perioperative myocardial infarction were peak troponin I greater than 3.7 µg/L, and troponin I concentration greater than 3.1 µg/L at 12 hours or greater than 2.5 µg/L at 24 hours [12, 15].
Echocardiography
All studies were performed using transthoracic echocardiography with an Acuson Sequoia C256 (Acuson Corporation, Mountain View, California), always done by the same physician in a blinded manner, at hospital admission and before discharge. Left ventricle ejection fraction and wall motion score index were recorded.
Statistical Analysis
Statistical analysis was performed by the SPSS program for Windows, version 10.1 (SPSS, Chicago, Illinois). Continuous variables are presented as mean ± SD, and categorical variables are presented as either absolute numbers or percentages. Data were checked for normality before statistical analysis. Normally distributed continuous variables were compared using the unpaired t test, whereas the Mann-Whitney U test was used for those variables that were not normally distributed. Categorical variables were analysed using either the 
2 test or Fisher’s exact test. Comparisons were considered significant if the p value was less than 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
|
When wound healing complications were considered, no differences were recorded between the two groups in terms of deep sternal wound infection (group A, 5 of 387 [1.5%] versus group B, 1 of 129 [0.8%]; p = 0.533), although superficial wound infection was higher in the pedicled group (10 of 387 [2.6%] versus group B, 0 of 129; p = 0.05). None of these patients underwent bilateral IMA grafting.
No diffferences were recorded between the two groups either in terms of postoperative acute myocardial infarction (group A, 5 of 387 [1.3%] versus group B, 2 of 129 [1.6%]; p = 0.556) or need for postoperative IABP (group A, 11 of 387 [2.8%] versus group B, 6 of 129 [4.7%]; p = 0.232). No IABP-related complications were registered during the period of assistance, except for a patient belonging to group A who had transient limb ischemia, and recovered after IABP withdrawal. Perioperative inotropic support and troponin I proved to be similar in the two groups (Table 2). Intensive care unit stay (group A, 2.1 ± 1.6 days versus group B, 1.7 ± 0.7; p = 0.137) and hospital stay (group A, 6.7 ± 0.6 days versus group B, 6.9 ± 1.0; p = 0.606) also proved to be similar. When echocardiography was considered, except for patients having perioperative acute myocardial infarction, the two groups demonstrated comparable recovery of left ventricular ejection fraction and wall motion score index (Table 2).
|
All 3 patients undergoing graft revision because of unacceptable TTF results demonstrated an uncomplicated postoperative course with optimal postoperative electrocardiographic, echocardiographic, and biochemical (troponin I) results. Another patient belonging to group B demonstrated a diastolic pattern but low maximum (13 mL/min) and mean flows (6 mL/min), with a sufficient PI (4.2), compatible with the poor run-off of a severely diseased LAD. The anastomosis was not redone, but the patient had perioperative acute myocardial infarction requiring IABP assistance. The perioperative course was then uneventful, and the patient was discharged home in healthy condition on postoperative day 13.
When TTF values were considered, both pedicled and skeletonized LIMA showed comparable results (Table 3), for on-pump and off-pump surgery (Table 4). Furthermore, 193 patients (142 belonging to group A, 51 to group B) underwent preoperative IABP because of unstable angina despite maximal intravenous nitrates administration. To rule out differences in GFR between pedicled and skeletonized IMAs, these patients underwent intraoperative TTF analysis with either 1:1 IABP and temporary cessation. The 1:1 IABP support recruited GFR in all these patients, and no differences were detected in graft mean flow, pulsatility index, and GFR between the two groups (Table 5). Finally, when patients with preoperative severe ventricular hypertrophy at echocardiography were analyzed, no differences in LIMA to LAD TTF measurements were ruled out (Table 6).
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Conversely, clinical results and angiographic patency rates appear to be comparable betwen the two techniques [3, 16, 18]. However, a recent review by Athanasiou and coworkers [3] showed that the only evidence from the current literature is a higher sternal devascularization with the pedicled technique, which was clearly demonstrated by the studies of Lorberboym and colleagues [22], Cohen and colleagues [7], and Parish and coworkers [8]. We found a higher incidence of superficial wound complications in patients undergoing pedicled LIMA harvesting, but comparable deep sternal wound infections in the two groups. It can be argued that the relatively low incidence of such complication in our practice may explain such discrepancy. Moreover, it can be hypothesized that the relatively small width of the pedicle (1 to 1.5 cm) and the low prevalence of bilateral mammary grafting in the study period may further account for these results. However, further studies are warranted to better define such topics.
It has to be kept in mind that very few studies have evaluated the functional behavior of the conduits, [4–6, 9], most focusing on the free-flow capacity [4, 5]. Both Deja and coworkers [4] and Wendler and coworkers [5] found skeletonized LIMAs to have higher free flow compared with pedicled conduits. Both studies measured flows after intra-arterial papaverine injection, which has been demonstrated to cause endothelial damage to the mammary artery [23]. Accordingly, we never used intraluminal papaverine on LIMA or on the other arterial conduits employed. Furthermore, Wendler and coworkers [5] analyzed only a limited number of grafts (40 patients in each group), initially showing comparable pedicled and skeletonized graft free flows, which reverse in favor of the skeltonized technique after papaverine injection; we do not know how this different behavior could be explained, especially when Deja and coworkers [4], who found similar response to papaverine, reported a comparable functional behavior when the acetylcholine-induced vasodilation was studied [4]. Finally, these studies did not report if differences existed between the two techniques in the caliber of LIMA, which seems to be the most important factor for free-flow measurements.
However, surgeons now have the possibility of studying the functional "in-vivo" behavior of the grafts, after completion of distal anastomoses, through the aid of TTF technology. The TTF method is fast, easy, and predictive of graft occlusion at short-term angiographies, and it is less affected by artifacts than are other techniques [13, 24]. In particular, not only are the absolute flow values but also the flow curve patterns and the PI needed to correctly use this technology. In particular, according to D’Ancona and coworkers [13], and based on our experience, we take down and redo the anastomoses in which systolic spiky patterns of the curves, low flow values, and PIs of 4 or more are reported by TTF, whereas low flows with preserved diastolic pattern and good PI, especially when associated with intraoperative findings of a small and severely diseased coronary bed, do not mandate graft revision. However, any unexpected and unexplained TTF finding should suggest graft revision.
Only Takami and coworkers [9] used TTF to compare pedicled and skeletonized LIMA, and found skeletonization to give higher mean flows, but comparable PI. Apart from this discrepancy, the study did not stratify the results by grafted vessel, did not reported the quality (run-off) of the grafted vessel (of paramount importance in TTF interpretation), and, most importantly, did not analyze GFR. Moreover, Takami and coworkers [9] found larger juxta-anastomotic diameters of skeletonized LIMAs, which can be, accordingly, the result of a different run-off of the grafted vessels in the two groups. On the contrary, we found comparable maximum, mean, and minimum flows and PIs in pedicled and skeletonized grafts, either after off-pump or on-pump CABG, with comparable TIMI scores of the grafted territories. Moreover, instead of papaverine injection, we chose IABP to study GFR by a more physiologic approach [14, 15], and found comparable GFR in pedicled and skeletonized grafts. These data confirm the reported comparable acetylcholine-induced vasorelaxation [4], and may explain the reported comparable long-term angiographic patency, survival, and freedom from cardiovascular events with these two techniques [3, 16, 18]. Accordingly, comparable in-vivo behavior of pedicled and skeletonized conduits resulted in similar in-hospital results in the two groups; although the optimal flow values we found in the skeletonized group must be also attributed to the use of low-current cautery, minimizing the risk of endothelial damage [25], and to the absence of a learning curve in our experience because only experienced surgeons employed skeletonization.
One potential disadvantage of the pedicled technique, often underestimated by the authors, is the possibility of a pedicled torsion on its axis, leading to severe hypoperfusion of a well-harvested pedicled graft. As in the case we reported, such a complication can be easily detected by TTF, and should always be suspected when unexpectedly poor TTF results are encountered. Obviously, extreme caution in securing a pedicled LIMA is mandatory.
Another supposed advantage of the skeltonized technique is the direct visual inspection of the vessel to identify vessel wall injuries, hematomas, and dissections, which, if unnoticed, may jeopardize the outcome [16]. First, it has to be considered that the pedicled harvesting technique is the only real "no-touch" technique [3, 10, 16, 18]. Similarly, although skeletonization has been demonstrated to preserve the mammary function [1-6, 10, 18, 19, 26], different pathologic studies show a tendency for skeletonized grafts to contain a larger number of lesions in the endothelium, microscopic intimal dissection and detachment, and injuries to the external elastic lamina [23, 27]. It is well known, however, that intimal dissections as well as any vessel wall injury and hematomas are responsible for, at least, arterial spasm, leading to reduced pulsation and free flow of the conduit, which may be easily recognized in the operating theater. Unsatisfactory pulsatile flow of the LIMA may negate its use as a graft, or at least induce the surgeon to inspect the graft and maybe skeletonize it if harvested as a pedicle, to exclude vessel wall damages. With such an approach, we excluded 2 pedicled LIMAs during the last 4 years of clinical practice.
Finally, it has been suggested that skeletonization may improve clinical results, reducing the risk of hypoperfusion syndrome due to a disproportion between LIMA flow and myocardial demand because of a severe venticular hypertrophy [10, 11]. Our results demonstrated that GFR was recruited in a comparable manner, regardless of the harvesting technique employed. Furthermore, patients with echocardiographic evidence of severe left ventricular hypertrophy, who are at higher risk for such catastophic complication, showed comparable TTF results and GFR with pedicled and skeletonized LIMAs. Our data are consistent with those of Athanasiou and coworkers [3], who demonstrated that skeletonization did not affect the incidence of hypoperfusion syndrome.
We conclude that skeletonized and pedicled LIMA showed a comparable in-vivo functional behavior, giving similar TTF results and recruitment of GFR, comparable in-hospital results, similar troponin I leakage, and similar postoperative echocardiographic improvement. Apart from the risk of higher sternal devascularization or the need for longer conduits, there is no reason to prefer skeletonized to pedicled mammary grafts.
Limitations of the Study
The main limitation of the study is the absence of randomization of the two surgical techniques. An ideal comparison between two treatments should be performed on the basis of a prospective randomized study. The study has been prospectively designed, but no real randomization has been performed, although such a goal may be very difficult to reach in current clinical practice. Furthermore, surgeons harvesting IMAs, and surgeons implanting LIMAs did not vary during the study period.
The study lacks angiographic control. Angiography is still the gold standard to evaluate long-term results of two different surgical strategies in coronary artery surgery. However, it was not the purpose of the study to assess graft patency, but to evaluate functional in-vivo behavior of the LIMA grafts—with the aid of intraoperative TTF as well as major perioperative clinical and subclinical results. Furthermore, it has been demonstrated that the TTF method is predictive of graft occlusion at angiographic follow-up [13, 24]. Finally, different and complex mechanisms affect graft patency during follow-up, first of all the progression of atherosclerosis and comorbid conditions. It can therefore be a mistake to only attribute graft patency to a single surgical variable, such as the harvesting technique.
Finally, this is a single-center study, which guarantees uniformity of the perioperative management of the patient population. On an intention-to-treat basis, we excluded patients undergoing sequential or T-graft and Y-graft construction, first of all because of the limited number of sequential LIMA grafting in our experience and also because the different hemodynamic rules underlying sequential and composite grafts might mislead TTF results. Certainly, larger randomized studies including sequential and composite graft construction may help to better define the differences in the functional properties of pedicled and skeletonized mammary grafts.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Foroughi, S.-A. Hassantash, M. Hekmat, A. Bolourian, M. Shahzamani, and Z. Ansari Do Different Techniques of Left Internal Mammary Graft Harvesting Really Affect Hospital Outcome? Ann. Thorac. Surg., September 1, 2008; 86(3): 1053 - 1053. [Full Text] [PDF]
|
|
|
|
|
|
F. Onorati, G. Santarpino, and A. Renzulli Reply. Ann. Thorac. Surg., September 1, 2008; 86(3): 1053 - 1054. [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 |
