Author Affiliations: Division of Cardiology, Department of Medicine and Therapeutics, Prince of Wales Hospital, Chinese University Hong Kong. There are no grants, financial support, or conflicts of interests related to the research in this paper. Manuscript submitted February 14, 2008, provisional acceptance given April 28, 2008, manuscript accepted June 3, 2008. Address for correspondence: Eugene Brian Wu, MRCP, MD, Division of Cardiology, Department of Medicine and Therapeutics, Prince of Wales Hospital, Chinese University Hong Kong, Ngan Shing Street, Shatin, Hong Kong. E-mail: firstname.lastname@example.org
ABSTRACT: We report on three cases of left main stem thrombosis complicating percutaneous coronary angioplasty. We focused on the issue of dealing with emboli in the non-culprit vessel, as embolism in the “virgin” vessel is often associated with severe cardiogenic shock. We describe the use of a FilterWire technique to protect the non-culprit vessel and a case in which the FilterWire was occluded by thrombus leading to shock. The third case illustrates the difficulty with FilterWire protection if stenting is required in the left main stem itself.
J INVASIVE CARDIOL 2008;20:E283–E287
Left main stem thrombus formation is a rare but dreaded complication of percutaneous coronary intervention (PCI). Thrombus in the left main coronary artery may cause emboli into the non-culprit vessel leading to ischemic insult to a virgin myocardial territory. This type of ischemic insult is particularly hazardous, as the culprit territory is often akinetic and transient myocardial stunning or infarction of the virgin territory may lead to severe cardiogenic shock. We report on our recent experience with three cases of left main thrombus.
Case 1. A 77 year-old male with diabetes, hypertension, and hypercholesterolemia presented with non-ST-elevation myocardial infarction (N-STEMI). His coronary angiogram showed a tight 99% proximal left anterior descending artery (LAD) lesion, an 80% mid-LAD lesion, and a mid-right coronary artery (RCA) clot-filled subtotal lesion, which was deemed to be the culprit (Figures 1A and B).
Procedure. A total of 5,000 units of heparin (at 70 units per kilogram of body weight) and a 15 mg bolus of abciximab were administered intravenously after guiding catheter placement. A Vista Brite Tip JR 3.5 guiding catheter (Cordis Corp., Miami Lakes, Florida) was placed in the right coronary artery (RCA) os via the right femoral approach. The lesion was wired using a Hi-Torque balanced middleweight universal guidewire (Abbott Vascular, Abbott Park, Illinois) and a 7 Fr Rebirth thrombectomy catheter (Goodman Co. Ltd., Nagoya, Japan) was used, followed by dilatation with a 2 x 15 mm Maverick balloon (Boston Scientific Corp., Natick, Massachusetts) at 10 atm. This was complicated by slow reflow in the posterior left ventrciular branch, likely due to distal embolization of thrombus. An EZ FilterWire (Boston Scientific) was deployed in the distal RCA and a 3 x 20 mm Taxus stent (Boston Scientific) was placed in the mid-right coronary lesion at 12 atm. Postdilatation was performed using a 3.5 x 8 mm noncompliant Maverick balloon. The FilterWire was retrieved, the vessel was rewired, and an intravascular ultrasound (IVUS) examination was performed by manual pullback with a Galaxy II, Atlantis 40 MHz catheter (Boston Scientific). A hazy lesion was angiographically apparent both before and after the stent, with 50% luminal stenoses distal to the stent. On IVUS, these areas demonstrated intraluminal thrombus. However, as the lesion was not flow-limiting, no further intervention was performed on the RCA at this juncture.
Since the LAD lesion was very tight, we decided to proceed with LAD PCI at this point. An EBU Vista Brite Tip 3.5 guiding catheter was placed in the left main os. Initial angiography showed subtotal TIMI 2 LAD flow, thus the LAD, a small first diagonal, and a large second diagonal branch were wired with a Hi-Torque balanced middleweight universal guidewire. The mid-LAD and proximal LAD lesions were dilated with a 2 x 15 mm Maverick balloon at 12 atm. The same IVUS system was used with manual pullback on the LAD, which demonstrated significantly stenotic and diffuse disease all through the LAD from the os to the distal part of the vessel. After predilatation of the second diagonal os using the 2.0 x 15 balloon, a 3 x 33 mm Cypher stent (Cordis) was placed from the mid-to-distal LAD across the second diagonal os at 12 atm. This was complicated with loss of flow in the second diagonal, and attempts to recross the diagonal were unsuccessful. Therefore, a 3.5 x 8 mm Maverick noncompliant balloon was used to postdilate the stent at 16 atm. After postdilatation, a balloon-supported Hi-Torque Pilot hydrophilic wire was able to pass into the second diagonal, and kissing balloon was performed with the 2 mm balloon and the 3.5 mm balloon in the second diagonal and the LAD at 8 atm, respectively. The second diagonal wire was removed and the circumflex artery was wired instead. A 3.5 x 23 mm Cypher stent was placed overlapping the distal stent right up to the LAD os. However, considerable time was used to place the stent accurately at the os and the angiogram before stent deployment showed thrombus in the left main and the distal part of the undeployed stent (Figure 1C). After stent deployment, an activated clotting time (ACT) was measured at 379 seconds. Angiography after removal of the stent balloon showed less thrombus in the left main artery, but the small amount of thrombus continued its propagation despite the use of another 5 mg of intracoronary abciximab, leading to progressive thrombus propagation in the left main, the LAD ostium and the proximal part of the circumflex artery (Figure 1D). The proximal circumflex artery was retroflexed, and with much difficulty, a 6 Fr Export XT thrombectomy catheter (Medtronic, Inc., Minneapolis, Minnesota) was used in the proximal circumflex artery. This resulted in embolization of the thrombus into the distal circumflex and total occlusion to the distal circumflex and large second obtuse marginal arteries (Figure 1E).
The patient went into cardiogenic shock at this juncture and inotropes were used to sustain his blood pressure. The distal circumflex wire was redirected into the large second obtuse marginal (OM) branch, and the Export catheter retrieved some thrombus, but also further dislodged emboli into the distal part of the large OM. Final angiography showed that there were no significant lesions remaining in the circumflex territory, although the second OM was now stump-occluded by thrombus at its distal end (Figure 1F). There was no contrast extravasation or staining in the distal OM vessels. The patient’s blood pressure remained stable on inotropes, and he was admitted to the coronary care unit on 10 µg/kg/minute of adrenaline and an abciximab infusion.
The patient developed transient inferior ST-elevation 4 hours after the procedure and was brought back to the catheterization laboratory. Coronary angiography demonstrated patent left-sided stents and continued stump occlusion of the distal second OM. The distal right coronary thrombus area seemed to have progressed slightly, resulting in a 60% luminal stenosis. A Vista Brite Tip JR 3.5 guiding was used, and the lesion was wired with a Hi-Torque balanced middleweight universal guidewire, followed by direct stenting to both hazy areas proximal and distal to the first stent with 4 x 20 mm and 3.5 x 20 mm Taxus stents at 12 atm. There was slight embolization to a very distal branch of the PLV, but final angiography showed excellent patency of the stents without any thrombus or dissection.
Progress. The patient developed sudden cardiac arrest 6 hours after the second procedure with electrical mechanical dissociation and failed resuscitation. Periresuscitation echocardiography demonstrated pericardial effusion predominantly over the posterior wall suggestive of cardiac tamponade secondary to cardiac rupture from the circumflex territory. The post mortem autopsy demonstrated blood in the pericardium secondary to bleeding from the pericardial fat supplied by the circumflex artery suggestive of distal coronary artery rupture.
Case 2. A 65 year-old female with diabetes, hypertension and known coronary artery disease who had a Cypher stent placed to her mid-LAD in October 2004, presented to our unit with N-STEMI. Her electrocardiogram showed anterior deep T-wave inversion and her troponin-T level was 1.0 µg/l. She underwent coronary angiography, which showed an occluded mid-LAD just after a first diagonal branch at the previously stented site.
Procedure. Subcutaneous enoxaparin 60 mg was given 3 hours before the procedure, followed by a 7.6 mg/hour infusion of eptifibatide. A 6 Fr Vista Brite EBU 3.5 guiding catheter was placed in the left main coronary ostium. An over-the-wire 2.0 x 15 mm Maverick balloon guided a Miraclebros 3g wire (Asahi Intecc Co. Ltd., Aichi, Japan) across the lesion, but the balloon itself was unable to cross. A Ryujin Plus 1.25 x 10 mm over-the-wire balloon (Terumo Corp., Shibuya-Ku, Tokyo, Japan) crossed the lesion, and after distal intraluminal position was confirmed with through-balloon contrast injection, the lesion was dilated to 10 atm and the Miraclebros wire was exchanged for a Hi-Torque balanced middleweight universal guidewire. An angiogram was taken with the balloon withdrawn into the left main, which showed adherent thrombus on the balloon now sitting in the left main stem (Figure 2A). The balloon was removed and the thrombus saddled the distal left main extending into both the proximal circumflex and anterior descending arteries (Figure 2B). Bailout 2,000 units of unfractionated heparin and 15 mg of intracoronary abciximab (ReoPro, Eli Lilly and Co., Indianapolis, Indiana) were administered. The ACT after this was > 999 seconds. An EZ FilterWire was deployed in the proximal circumflex artery (Figure 2C) and a 6 Fr Export XT thrombectomy catheter was used in the proximal LAD. Although some thrombus was aspirated, there was embolization both into the mid-LAD and the circumflex filter. The left main stem was now thrombus-free. The circumflex FilterWire was retrieved and another FilterWire was deployed in the diagonal artery. Export thrombectomy was performed in the LAD twice, but a significant thrombus mass remained in the LAD. The FilterWire was retrieved and replaced with a Hi-Torque balanced middleweight guidewire. The thrombus propagation was relentless and the thrombus size increased after FilterWire removal. A 2.5 x 15mm Multi-Link bare-metal stent (Abbott Vascular) was deployed, covering the thrombus from the mid-LAD into the diagonal branch, but a small thrombus remained proximal to this stent (Figure 2D). A 2.5 x 8 mm Liberté bare-metal stent (Boston Scientific) was therefore placed to cover this thrombus. Final angiography showed persistent total occlusion of the mid-LAD after the diagonal, but no thrombus in the left main, the proximal LAD or the diagonal branch (Figure 2E).
Progress. The patient had a postprocedural creatine kinase rise from 299 to 797 U/l, with a creatine kinase-MB (CKMB) rise from 14.2 to 40 U/l. She was discharged 6 days after the procedure and remained in Canadian Cardiovascular Society (CCS) Class 1 angina at clinical follow up.
Case 3. A 70 year-old male with hypertension presented 5 months previously with acute pulmonary edema after a silent inferior myocardial infarction. Coronary angiography then showed a nondominant occluded distal RCA and significant lesions in both the LAD and circumflex arteries. After a failed attempt to open the RCA, an Endeavor stent (Medtronic) was placed in the proximal LAD and a Taxus Liberté stent in the mid-left circumflex artery (LCx). He developed in-stent restenosis 5 months later and had a 3 x 8 mm Xience V stent (Abbott Vascular) placed in the proximal LAD overlapping the previous proximal edge of the LAD stent for proximal stent edge restenosis. The patient presented 6 days after the Xience V stent implantation with acute pulmonary edema and his electrocardiogram (ECG) showed hyperacute T-waves in the anterior leads.
He was brought to the catheterization laboratory urgently and coronary angiography showed subacute stent thrombosis with thrombus extending right up to the ostium of the LAD (Figure 3A).
Procedure. Three thousand units of heparin and 15 mg of abciximab were administered. A 7 Fr Vista BriteTip EBU 3.5 guiding catheter was placed in the left main coronary ostium and the LAD was wired with a Hi-Torque balanced middleweight universal guidewire. A Filter Wire was deployed in the proximal circumflex artery (Figure 3B) and a 6 Fr Export XT thrombectomy catheter was used in the proximal LAD. After Export thrombectomy, the patient’s systolic blood pressure dropped to 60 mmHg. Dopamine and epinephrine infusions were started. Angiography still showed heavy thrombus burden in the proximal LAD and thrombus filling the circumflex FilterWire leading to obstruction of circumflex blood flow (Figure 3C).
The FilterWire was retrieved and another 5 mg of abciximab was administered. A new FilterWire was deployed again in the circumflex artery and further Export thrombectomy to the LAD was performed. Thrombus remained in the proximal LAD extending to the left main artery, and the patient developed severe respiratory distress. The FilterWire was replaced with a Hi-Torque balanced middleweight guidewire and a 3 x 28 mm endothelial progenitor cell-attracting antibody Genous stent (OrbusNeich, Wanchai, Hong Kong) was deployed covering the previous stented areas right up to the left main ostium at 18 atm. The final angiogram showed no thrombus in the coronary arteries and a patent stent (Figure 3D).
Progress. The patient was placed on noninvasive continuous positive airway pressure (CPAP) respiratory support and was admitted to the intensive care unit. The patient was transferred to a convalescent hospital for rehabilitation.
Left main thrombus is a rare but dangerous complication of PCI. Although at times prothrombotic substrates like heparin-induced thrombocytopenia1 or thrombocytosis2 can cause left main thrombus, procedural issues such as prolonged occlusive manipulation, inadequate anticoagulation and presence of thrombus before intervention can all cause progressive left main thrombosis. The danger of left main thrombus lies in the risk of ischemic injury to the virgin vascular territory by embolism or by occlusion. In patients where the LAD territory is already stunned by ischemia, further ischemic insult to the circumflex territory can cause devastating heart failure and cardiogenic shock (especially if the RCA is already occluded, as in Case 3).
The importance of adequate anticoagulation cannot be overemphasized, and the use of glycoprotein IIb/IIIa agents along with heparin, aiming for an ACT > 250 seconds, should be the goal. However, even with adequate anticoagulation, prolonged occlusive manipulation of balloons and IVUS catheters can still cause thrombosis, as in our first case.
Prevention is better than the cure. In retrospect, prolonged occlusive manipulation and prolonged contact of the balloon with in-stent thrombus caused left main thrombus in our cases. Perhaps in the patient who had a heavy thrombus load in the RCA, we should have employed a simpler strategy for the LAD stenting, or even considered delaying the LAD PCI until the next day. More rapid stenting and fewer balloon exchanges may have prevented these complications. It seems likely that the first patient died from a coronary wire perforation, and the importance of keeping the wire tip fixed in a good position, even when all hell is breaking loose, has been deeply impressed upon us.
In the first case, we deemed the circumflex artery to be too retroflex for FilterWire use, even the Export catheter failed to get past the first 10 mm of the artery. However, through our experience with Case 1, we were impressed by the need to protect the virgin territory and proposed using a FilterWire technique, as described as soon as we detected left main thrombus. This technique served us well in Case 2, but we encountered the problem of thrombus filling up the filter and occluding flow to the circumflex in Case 3. However, with a FilterWire exchange, we were able to continue the procedure.
Export thrombectomy seems to be relatively inefficient in removing thrombus from the left main and often results in pushing the thrombus downstream. We (Case 2), and others,3 have found stenting to be more effective in covering/removing thrombus. An alternative technique would be to place FilterWires in both the LAD and LCx and using a large balloon to dislodge the thrombus into the FilterWire net. However, we feel some concern that the thrombus load may exceed the FilterWire’s ability to trap thrombus, leading to emboli.
We could not keep the FilterWire in the LCx when stenting into the left main, and out of desperation in Case 3, we stented the LAD to the left main without circumflex protection. In better conditions, we might have stented the LAD to the ostium with a FilterWire in place, and then pulled back the stent balloon into the left main for one expansion to dislodge thrombus into the circumflex filter, and then exchanged it for a guidewire in the LCx and stented the left main into the LAD ostium.
These cases taught us the importance of nonocclusive protection of the virgin territory when PCI is complicated by left main thrombus. We found that the FilterWire can achieve this protection, but vigilance is needed for filter obstruction secondary to thrombus load. We found that Export thrombectomy can reduce the thrombus load, but rarely manages to remove all of the thrombus. Stenting remains a useful tool to cover residual thrombus.