Author + information
- Jeffrey J. Cavendish, MD*,
- William F. Penny, MD, FACC*,‡,
- Michael M. Madani, MD†,
- Shahin Keramati, MD*,
- Ori Ben-Yehuda, MD, FACC*,
- Daniel G. Blanchard, MD, FACC*,
- Ehtisham Mahmud, MD, FACC*,
- Anthony Perricone, MD† and
- Sotirios Tsimikas, MD, FACC*,* ()
- ↵*Reprint requests and correspondence:
Dr. Sotirios Tsimikas, Vascular Medicine Program, University of California-San Diego, 9500 Gilman Drive, BSB 1080, La Jolla, California 92093-0682, USA.
The Symmetry Bypass Connector (St. Jude Medical, St. Paul, Minnesota) is a nitinol, star-shaped device that was designed to facilitate placement of sutureless aorto-saphenous anastomoses during off-pump coronary artery bypass graft surgery (CABG). Although the device is approved for clinical use in Europe and the U.S., its short- and long-term safety and efficacy are not established. We report on 5 of 121 patients undergoing CABG who presented with an acute coronary syndrome two to five months following placement of this device. In each patient, all saphenous vein grafts (SVGs) placed (n = 11) with the device were totally occluded (n = 6) or compromised by ostial stenoses (n = 5). Treatment consisted of repeat CABG in one patient and percutaneous coronary intervention (PCI) in four patients with cutting balloon atherotomy and stenting. Following PCI, two of four patients presented again within two months with near-occlusive ostial restenosis in all stents placed. Intracoronary ultrasound showed severe neointimal hyperplasia, but only at the proximal interface of the device and stent. One patient was treated with brachytherapy in two SVGs but had a recurrence four months later and was treated with drug-eluting stents in both restenotic segments. Recalcitrant neointimal hyperplasia is postulated to be involved in the pathogenesis of anastomotic device stenosis, possibly similar to in-stent restenosis. Prospective randomized clinical trials are needed to assess the clinical safety and efficacy of this device. Pending such studies, consideration should be given in limiting its use to cases of unacceptably high risk of stroke during aortic cross-clamping. Dual antiplatelet agents, evaluation for ischemia, and close follow-up are warranted in patients that have already received the device.
Vascular anastomoses of the proximal aorta are usually performed by manual suturing techniques. Recent increases in the use of “off-pump” coronary artery bypass graft surgery (CABG), initiated to reduce neurological complications due to cross-clamping of the aorta, have resulted in a need to develop devices to facilitate rapid and efficient proximal and distal anastomoses. The Symmetry Bypass Connector (St. Jude Medical, St. Paul, Minnesota), the first device to be approved for such use, consists of a nickel-titanium (nitinol) connector (Fig. 1), an aortic cutter to create the aortotomy, and a delivery device to implant the saphenous vein graft (SVG) to the aortic wall (1,2). After choosing an appropriately sized device, the saphenous vein is loaded onto the device and fastened by hooks (Fig. 1). A hole is then made in the aortic wall using an aortic cutter that at the same time retrieves the aortic wall plug. The device is then deployed in the aortotomy, first by releasing the nitinol struts in the internal aortic wall and then the external wall whereby the device expands in a star appearance (Fig. 1B). This allows the creation of a proximal aorto-saphenous anastomosis without cross-clamping or side-biting the aorta, in a shorter time than it takes to make a traditional hand-sewn anastomosis. Following implantation, the device has a star-shaped angiographic appearance at the aorto-ostial junction (see the accompanying videoat http://www.cardiosource.com/jacc.html).
This device obtained European Certification in May 2000 and U.S. Food and Drug Administration approval in May 2001, and over 30,000 devices have been implanted as of November 2002 (3). Although initial results during implantation of this device have been reported, both short- and long-term clinical outcomes are not well known. We report on five patients presenting with an acute coronary syndrome (ACS) due to severe proximal stenoses or flush occlusion of all eleven SVGs placed using the Symmetry aortic connector device. From January 2002 to December 2002, a total of 320 devices were implanted in 121 patients at the University of California-San Diego and Veterans Administration Medical Center from a total of 263 CABG surgeries.
A 59-year-old male patient with a history of hypertension and hypercholesterolemia developed accelerated angina over two weeks and sustained a non–Q-wave myocardial infarction (troponin I, 4.5 ng/dl). Coronary angiography revealed left main and severe three-vessel coronary artery disease (CAD). He underwent off-pump CABG in August 2002 with placement of a left internal mammary artery (LIMA) to the left anterior descending coronary artery (LAD). The SVGs to the right coronary artery (RCA), ramus intermedius (RI), and left circumflex obtuse marginal coronary artery (OM) were placed using the Symmetry device for proximal anastomoses. The patient had an uneventful postoperative course but presented in January 2003 with abrupt onset of chest pain and inferior ST-segment elevation. He was treated with aspirin, heparin, and intravenous nitroglycerin with rapid resolution of ST-segment elevation and symptoms. Emergent coronary angiography revealed that all three SVGs had >90% ostial stenoses at the junction of the Symmetry device, with the SVG to RCA also having a long proximal 40% stenosis (Fig. 2) as well as Thrombolysis In Myocardial Infarction (TIMI) II flow grade. Successful percutaneous coronary intervention (PCI) of all three SVGs was performed. The SVG-RCA stenosis was predilated with 3.0 × 20 mm CrossSail, 3.0 × 15 mm Cutting Balloon (Boston Scientific, Natick, Massachusetts), and 3.5 × 8 mm PowerSail balloons (Guidant, Santa Clara, California). Because of significant recoil, a 3.5 × 28 mm multi-link Penta (Guidant) coronary stent was deployed at 16 atm with acceptable result (Fig. 2B). The SVG-RI was dilated with the same CrossSail and Cutting Balloon, but owing to recoil, it was also stented with a 3.0 × 13 mm Penta stent (Fig. 2D). Three days later, PCI of the SVG-OM was performed by dilating the lesion with a PowerSail balloon but, again, because of recoil a 4.0 × 15 mm Penta was deployed (Fig. 2F).
The patient did well until March 2003 when he was re-admitted with unstable angina, and angiography revealed ostial restenosis of all three SVGs. The SVG-RCA had a 90% ostial in-stent restenosis and a 60% stenosis extending ∼25 mm beyond the stented portion (Fig. 3A). The SVG-OM also had an 80% ostial in-stent restenosis (Fig. 3C) and the SVG-RI was subtotally occluded at the ostium (Fig. 3E). Pre-PCI intracoronary ultrasound of the SVG-RCA revealed a well-expanded stent with focal, severe, neointimal hyperplasia at the proximal part of the stent at the interface of the device (Fig. 4A). The distal part of the stent had minimal neointimal hyperplasia (Fig. 4B). Beyond the stent, where no injury by the stent balloon had taken place, there was progression of the initial 40% stenosis to a 60% stenosis. The intravascular ultrasound (IVUS) showed moderate-severe neointimal hyperplasia (Fig. 4C). The in-stent restenotic lesion of the SVG-RCA was treated with a 3.5 × 15 mm Cutting Balloon, and the lesion beyond the stent was stented with a 4.0 × 28 mm Zeta (Guidant) stent (Fig. 3B). The SVG-OM was difficult to wire and the IVUS catheter would not pass through the ostium. The area was dilated with a 3.0 × 15 mm Power-Sail and 3.5 × 15 mm Cutting Balloon (Fig. 3D). Intracoronary brachytherapy was applied to both SVGs with a 40-mm source train (Novoste Corp., Norcross, Georgia). The subtotally occluded SVG-RI could not be wired, despite bringing the patient back two weeks later for another attempt.
The patient did well until September 2003 when he developed recurrent angina. Coronary angiography showed recurrent ostial restenosis with 80% ostial stenosis of the SVG-RCA and 99% stenosis of the ostium of the SVG-OM. The distal parts of the stents were patent. He was treated with 3.0 × 23 mm and 3.0 × 18 mm sirolimus-eluting Cypher (Cordis, Miami, Florida) stents of the SVG-RCA and SVG-OM, respectively, both post-dilated at the ostium to 4.0 mm with PowerSail balloons, with resolution of his angina.
A 46-year-old male patient with history of hypertension, hyperlipidemia, and tobacco use was diagnosed with severe left main and three-vessel CAD in August 2002. Off-pump CABG was performed with a LIMA to the LAD and SVGs to RCA and OM using the Symmetry device. He presented in January 2003 with rest angina, non-specific ST-T wave changes, and elevated troponin I (1.7 ng/ml). Coronary angiography showed a patent LIMA to the LAD. However, flush occlusion of the SVG-OM and a 90% ostial stenosis of the SVG-RCA were noted. The PCI of the SVG to the RCA was performed with pre-dilation using a 3.0 × 15 mm NC Ranger balloon (Boston Scientific) at 18 atm. Because of significant recoil, a 3.5 × 16 mm Express2 stent (Boston Scientific) was placed and post-dilated with a 4.0 × 15 mm NC Ranger balloon. Because the SVG-OM was totally occluded, stenting of the left main and first diagonal was performed with standard techniques. The patient had an uneventful hospital course. An exercise treadmill test performed five months later showed no ischemia.
A 58-year-old male patient with chest mantle radiation for Hodgkin's disease in 1983, hyperlipidemia, and PCI of an unknown vessel in 1993, underwent coronary angiography in October 2002 for angina and an abnormal stress test, which revealed a high-grade ostial left main stenosis but no significant angiographic disease in the LAD, left circumflex, and RCA. Pre-operative angiography of the LIMA revealed that it was atretic and, therefore, an unsuitable bypass conduit. As a result, he underwent CABG with SVGs to the LAD and OM using the Symmetry device. Postoperatively, he was stable for two months but then developed rapidly progressive angina. During exercise treadmill testing, he experienced severe chest pressure at 3 min with ST-segment elevation in the precordial leads. Sublingual nitroglycerin relieved his symptoms and ST-segment elevation. Urgent coronary angiography showed 99% ostial stenosis of the SVG-LAD with TIMI I flow grade and a totally occluded SVG-OM. There was a tight stenosis and tethering of the mid-LAD at the distal anastomosis. The PCI of the SVG-LAD was performed by dilating the lesion with a 2.5 × 20 mm OpenSail balloon (Guidant), and a 3.0 × 8 mm Zeta stent was placed. A second stent could not be passed more distally at the proximal aorto-saphenous segment. A 3.0 × 8 mm Zeta stent was also placed in the left main trunk to provide flow to the OM, and a 2.5 × 24 mm S660 stent (Medtronic AVE, Minneapolis, Minnesota) was placed at the mid-LAD. He was subsequently discharged from the hospital uneventfully.
The patient returned within two months with unstable angina, and coronary angiography revealed total occlusions of the stents in the SVG-LAD and in the native mid-LAD and 70% in-stent restenosis of the left main stent, despite treatment with both aspirin and clopidogrel. The patient required a redo CABG with manual suturing of the SVG-LAD and SVG-OM. He did well postoperatively and was discharged home.
A 69-year-old male patient with hypertension and hyperlipidemia was admitted with unstable angina and underwent diagnostic angiography in October 2002. The left main coronary artery had an 80% ostial stenosis and the RCA had a 90% stenosis. High-grade right internal carotid stenosis was discovered by duplex ultrasound and confirmed with angiography. The patient subsequently underwent combined CABG and right internal carotid endarterectomy. A LIMA was placed to the first diagonal artery because the LAD was a small vessel and SVGs were placed to the LAD and RCA using the Symmetry device. He did well until April 2003 when he was re-admitted with unstable angina with dynamic electrocardiographic changes, but ruled out for a myocardial infarction. Angiography showed that the native RCA was now occluded and filling via left-to-right collaterals. The LIMA to the first diagonal was patent. However, both SVGs were flush occluded. Redo CABG was performed with an SVG-LAD. His postoperative course was uneventful.
An 83-year-old male with hypertension and hyperlipidemia was admitted with an acute coronary syndrome and sustained a non–Q-wave myocardial infarction in December 2002. Coronary angiography revealed severe diffuse disease of the LAD involving the bifurcation with the first diagonal artery and an 80% proximal stenosis of the RCA. He underwent off-pump CABG with placement of a LIMA graft to the LAD. Using the Symmetry Aortic Connector device, two SVGs were placed to the right posterior descending and first diagonal arteries. Postoperatively the patient did well for four months. However, he developed severe dyspnea leading to congestive heart failure and required hospitalizations in April, July, and August 2003. He was ultimately diagnosed with severe ischemic mitral regurgitation. Coronary angiography showed ostial occlusions of both SVG grafts, a patent LIMA to the LAD and high-grade disease of the native RCA. Percutaneous coronary intervention of the RCA was performed by placing a 3.0 × 23 mm Cypher stent (Cordis, Johnson & Johnson) at the ostium of the RCA.
In addition to these cases, of the entire cohort of 121 patients, 4 other patients underwent clinically indicated coronary angiography following CABG with the Symmetry device. One patient developed ST-segment elevation immediately postoperatively and had no-reflow in two SVGs and in the LIMA graft, apparently secondary to massive reperfusion injury. Both ostia of the SVGs at the site of the Symmetry device were patent. The other three patients had angiography at one, five, and seven months following device placement and one of one, two of two, and three of three SVGs placed with the device were completely patent. No patients who had conventional SVG anastomoses returned for repeat angiography within the same time frame.
This study documents that severe SVG compromise and recurrent ischemia may occur in the first six months following CABG when proximal anastomoses are performed with the Symmetry device. We report on five patients presenting with ACS following placement of the device whose symptoms were due to severe ostial anastomotic device stenoses (five SVGs) or total occlusion (six SVGs) of all eleven SVGs that were placed. Performance of PCI in four of these patients revealed highly recalcitrant lesions requiring high-pressure inflations with non-compliant or cutting balloons and adjunctive stenting due to recoil. Two of these patients developed recurrent ostial in-stent restenosis or occlusions within two months of PCI and required repeat PCI with brachytherapy and redo CABG, respectively. One patient also failed brachytherapy and was treated with drug-eluting stents. This study suggests that with rapid acceleration in the use of such devices, despite the lack of clinical outcomes data, anastomotic device stenosis may become a significant new disease entity in patients undergoing CABG with such devices.
Saphenous vein grafts are the most commonly used conduits for revascularization during CABG surgery. However, the use of SVGs has been associated with both short- and long-term failure and with an incidence of approximately 15% at one year following CABG (4,5). Etiologies for SVG failure include poor quality veins, subacute thrombosis, technical issues such as vein trauma or suturing technique, ischemia-reperfusion during vein explantation, and poor distal runoff. Proximal anastomotic stenoses of SVGs in the peri-operative period are generally due to pathological intimal hyperplasia, resulting from accumulation of smooth muscle cells and extracellular matrix, and may occur in 5% to 8% of cases (4,6,7).
Devices that more quickly and efficiently place aorto-saphenous anastomoses during off-pump CABG, by avoiding cross-clamping of the aorta, offer potential advantages over hand-sewn techniques. Although acute procedural success of this device has been reported (1,2,8), short- and long-term angiographic or clinical outcomes are lacking. Several reports have described ∼5% technical failure rate due to loading of the vein on the device and/or deployment of the device, persistent bleeding around the device, or unrecognized dissection flaps created during formation of the aortotomy (1,2,8). In addition, in-hospital occlusions during the postoperative state have been described, although the investigators believed these were due to thrombotic disorder rather than device-related (9,10).
During the time this study was in revision, two additional studies were reported confirming short-term failure with the Symmetry device. Carrel et al. (3)reported that 2 out of 10 patients who had six-month angiographic follow-up experienced significant ostial stenoses, whereas none of the hand-sewn devices had any stenosis. They also reported that 5 of the 13 patients (38%) with six-month angiographic follow-up showed proximal SVG stenoses (ranging from 30% to 90% stenosis) in a small randomized trial comparing the device to hand-sewn techniques. This trial has subsequently been halted owing to unacceptable rates of stenosis in the device group (3). Mack et al. (8)also reported that 4 of the 67 patients required repeat CABG or PCI owing to total occlusion of SVG implanted with the device. In addition, 6 of the 45 (13.3%) SVGs placed with the device were asymptomatically occluded in a subgroup undergoing computed tomography angiography. The investigators did not report the percentage of patients with non-occlusive stenoses, possibly because they are difficult to quantify with computed tomography angiography in its current state-of-the-art. In addition, an abstract reported that 7 (11%) patients had SVG failure, with 6 of the 7 at the neo-ostium in 72 patients receiving this device (11).
We propose that neointimal hyperplasia may be the underlying pathogenic mechanism, as noted previously in SVGs anastomosed with plain sutures (7), or as in in-stent restenosis (12). Although we did not perform IVUS on the first presentation, the patient in Case 1 did undergo IVUS on the second presentation of in-stent restenosis. This showed severe neointimal proliferation at the ostium at the site of the device (Fig. 4) but not within the remainder of the stent. The IVUS of the area beyond the stented site showed moderate neointimal hyperplasia but no evidence of atherosclerotic plaque. Owing to the difficulty of cannulating and even passing the wires in these ostial stenoses, IVUS could not be obtained in other sites prior to PCI. Additional studies will be needed before stent implantation to prove this. The neointimal process may be initiated and/or enhanced by injury to the SVG during placement onto the vein transfer sheath, in securing to the hooks, or due to size mismatch of the vein and the aortic connector. Because all the SVGs reported here were highly stenotic or totally occluded, thrombotic occlusion in the setting of a tight stenosis may have also occurred. However, poor distal anastomotic technique or poor distal runoff cannot be excluded in the six SVGs that were totally occluded. In fact, the time course of touchdown stenoses has also been noted to occur within the first six months of SVG implantation. It is also possible that the patients who develop SVG stenoses with these devices may have an underlying predisposition to developing anastomotic stenoses, similar to patients who develop in-stent restenosis.
Although the incidence of SVG failure following the use of this device is not known, the limited data currently available suggest that up to 10% of patients develop clinical symptoms and up to 38% may have evidence of some extent of angiographic stenosis in the first six months following CABG. However, the use of these devices has only recently accelerated, and neither complete angiographic nor clinical follow-up is available to ascertain the true incidence. In addition, how such devices compare to traditional suturing techniques remains to be determined in randomized trials.
In view of the dramatic presentation of these patients, this report suggests that close follow-up is warranted in patients in whom this device has been implanted. Consideration should be given to treatment with dual antiplatelet agents for at least six months and to evaluate for myocardial ischemia at two to six months. A low threshold in performing coronary angiography should be employed if symptoms or objective evidence of coronary ischemia is documented, as it seems unlikely that even recently occluded SVGs can be salvaged by PCI. In patients with compromised SVGs, the mode of revascularization needs to be determined. Accepted PCI techniques seem appealing but are challenging, as these cases depict, owing to difficulty engaging with guiding catheters and wire passage because of the ostial location of these lesions and in treating native lesions initially believed unfavorable for PCI. Redo CABG has also been used successfully in selected patients. Whether brachytherapy will have a role in cases where stenting is not applied initially or following in-stent restenosis remains to be determined. However, this case and a recent report by Traverse et al. (13)suggest that failures of brachytherapy may also occur. Drug-eluting stents may also be useful initially, as this case implies, although the long-term outcomes remain to be determined.
Finally, the long-term clinical safety and efficacy of this device should be evaluated thoroughly by prospective randomized clinical trials comparing this technique to traditional hand-sewn techniques. Pending such studies, strong consideration should be given in restricting its use to cases of unacceptably high risk of stroke during aortic cross-clamping (9).
☆ Spencer King, MD, acted as the Guest Editor for this paper.
- acute coronary syndrome
- coronary artery bypass graft surgery
- coronary artery disease
- intravascular ultrasound
- left anterior descending coronary artery
- left internal mammary artery
- left circumflex obtuse marginal coronary artery
- percutaneous coronary intervention
- right coronary artery
- ramus intermedius
- saphenous vein graft
- Thrombolysis In Myocardial Infarction
- American College of Cardiology Foundation
- Motwani J.G.,
- Topol E.J.
- Fitzgibbon G.M.,
- Kafka H.P.,
- Leach A.J.,
- Keon W.J.,
- Hooper G.D.,
- Burton J.R.
- Thiene G.,
- Miazzi P.,
- Valsecchi M.,
- et al.
- Reuthebuch O.,
- Kadner A.,
- Turina M.
- Lowe H.C.,
- Oesterle S.N.,
- Khachigian L.M.
- Traverse J.H.,
- Mooney M.R.,
- Pedersen W.R.,
- et al.