Very late stent thrombosis is an infrequent yet potentially fatal complication associated with drug-eluting stents. We report the case of an 88-year-old man who sustained an ST-segment-elevation myocardial infarction 11 years after initial sirolimus-eluting stent implantation. Optical coherence tomograms of the lesion showed that the focal incomplete endothelialization of the stent struts was the likely cause; neointimal formation, neoatherosclerosis, and late stent malapposition might also have contributed. To our knowledge, this is the longest reported intervening period between stent insertion and the development of an acute coronary event secondary to very late stent thrombosis. The associated prognostic and therapeutic implications are considerable, because they illuminate the uncertainties surrounding the optimal duration of antiplatelet therapy in patients who have drug-eluting stents. Clinicians face challenges in treating these patients, particularly when competing medical demands necessitate the discontinuation of antiplatelet therapy. In addition to the patient's case, we discuss factors that can contribute to very late stent thrombosis.
Coronary artery stent thrombosis is associated with high morbidity and mortality rates. Very late stent thrombosis (VLST) raises particular concerns, because its underlying pathophysiology is not completely understood and because the optimal preventive strategies and durations of dual antiplatelet therapy after stent insertion are undetermined. We report a case of drug-eluting stent (DES) thrombosis that occurred 11 years after stent insertion and 5 days after the discontinuation of antiplatelet therapy, in the presence of an anterior ST-elevation myocardial infarction (STEMI). In addition, we discuss factors that can contribute to VLST.
Case Report
In January 2014, an 88-year-old man presented at our hospital with an anterior STEMI. In 2003, at age 77 years, he had undergone the elective implantation of a 2.75 × 13-mm Cypher® sirolimus-eluting coronary stent (Cordis, a Johnson & Johnson company; Miami, Fla) across an 80% stenosis in his proximal left anterior descending coronary artery (LAD). The lesion had been predilated with use of a 2.25 × 15-mm CrossSail® coronary dilation catheter at a pressure of 6 atm, and the stent had been deployed at a pressure of 14 atm without postdilation. The patient was intolerant of aspirin because of nausea and nonspecific arthralgia, so his condition had been managed by means of clopidogrel monotherapy after the first 12 months. Of note, he had discontinued clopidogrel 5 days before the current presentation in preparation for a colonoscopy—a routine that he had followed on multiple occasions with no adverse consequences.
At the current presentation, coronary angiograms revealed thrombotic occlusion of the sirolimus-eluting stent (SES). Optical coherence tomograms (OCT) of the LAD showed substantial proximal-edge malapposition of the SES, including exposed struts and superimposed thrombus (Fig. 1); substantial restenosis secondary to neoatherosclerosis in the mid segment (Fig. 2); and undisrupted neointimal formation in the distal segment of the stent (Fig. 3). We gave the patient bivalirudin and implanted 2 everolimus-eluting Xience Prime® stents (Abbott Vascular, part of Abbott Laboratories; Abbott Park, Ill) (3 × 23 and 3 × 8 mm, respectively, each deployed at a pressure of 12 atm). The postdilations were at pressures of 20 atm with use of 3.5-mm noncompliant balloons. Good stent apposition was confirmed by means of OCT (Fig. 4). The patient recovered uneventfully and was discharged from the hospital with instructions to take aspirin and clopidogrel for 12 months. He tolerated the aspirin well. As of his 18-month evaluation, he remained well.
Citation: Texas Heart Institute Journal 42, 5; 10.14503/THIJ-14-4550
Citation: Texas Heart Institute Journal 42, 5; 10.14503/THIJ-14-4550
Citation: Texas Heart Institute Journal 42, 5; 10.14503/THIJ-14-4550
Citation: Texas Heart Institute Journal 42, 5; 10.14503/THIJ-14-4550
Discussion
Our patient sustained an anterior STEMI as a result of VLST 11 years after the implantation of a Cypher SES. To our knowledge, this is by far the longest reported interval between stent insertion and the development of an acute coronary event consequent to VLST. Accordingly, this case is both unique and cause for alarm.
Defined as the development of thrombotic stent occlusion later than 12 months after stent insertion,1 VLST is an infrequent yet clinically important sequela of stent implantation.2 Accurate estimates of the incidence of VLST in DES are precluded by few long-term data, although in general it varies from 0.4% to 0.6% annually3 after 12 months.
Risk Factors. Although many lesion, patient, and procedural characteristics have been associated with early and late stent thrombosis, the specific risk factors for VLST are less well defined. Current smoking and longer stent and lesion length have been reported as risk factors.4 Higher numbers of stents per lesion and stent overlap are also prominent characteristics in patients with DES who sustain VLST.5 Of importance, the discontinuation of antiplatelet therapy in itself has not been shown to be a risk factor for VLST6; the crucial aspects of the underlying pathophysiology relate to a combination of delayed arterial healing, ongoing vessel inflammation, neoatherosclerosis, and late stent malapposition.
In early experience with the Cypher SES, investigators reported insufficient expansion of the stent struts, and this might have contributed to our patient's VLST.7 The nominal pressure for deployment of a Cypher SES is 11 atm. However, 14 atm of pressure yielded adequate stent expansion in only 15% of lesions in the Multicenter Ultrasound Stenting in Coronaries (MUSIC) study.8 Even at 20-atm balloon inflation, adequate expansion was observed in only 60% of instances, and minimal luminal diameter was consistently smaller upon intravascular ultrasonographic (IVUS) evaluation in vivo when compared with the manufacturer's sizing charts.9 The minimal luminal diameter of our patient's proximal LAD was 3.2 mm (Fig. 4), so a constrained 2.75-mm SES was likely to serve as a constant substrate for restenosis and stent thrombosis over time.
Inaccurate deployment has been observed during SES implantation, and patients thus affected consequently have more need for target-vessel revascularization and a higher prevalence of myocardial infarction at one year.10 In our patient, the mismatch between the length of the predilation catheter and the SES might have resulted in longitudinally inaccurate placement. However, it is unclear whether this contributed substantially, because most adverse consequences of suboptimal deployment are observed early.
Mechanisms of Thrombosis. Incomplete endothelialization of stent struts is the primary precipitant of stent thrombosis. Other factors are late stent malapposition secondary to delayed positive remodeling, strut penetration into a necrotic core, and chronic vascular inflammation and hypersensitivity reaction to the metal struts.11,12
Neoatherosclerosis has an increasingly recognized role in stent restenosis and VLST.13 As a maladaptive endothelial response to stent implantation, neoatherosclerosis results in the evolution of a fibrous neointima into new in-stent atherosclerotic plaques. The mechanism underlying neoatherosclerosis might be a pathogenetic link with chronic inflammation.14 As a group, DES tend to be associated with greater and earlier development of neoatherosclerosis than are bare-metal stents.15 Specifically, SES tend to be associated with more rapid neoatherosclerotic changes than are paclitaxel-eluting stents, perhaps because of a difference in the polymer coating on the stent-strut surface. Indeed, SES have been shown to promote the formation of lipid-rich yellow neointima, which correlate with unstable plaques16 that have a higher potential of rupture and thrombotic sequelae. Finally, the incidence of thin-cap fibroatheromas (TCFA) that contain neointima seems to increase with time, as does the presence of red thrombi associated with atheromatous plaque rupture. All these factors suggest that neoatherosclerosis substantially contributes to VLST.
Optical Coherence Tomography and Neointimal Characterization. Both IVUS17 and OCT have been used clinically to characterize neointima. Aided by virtual histologic evaluation during IVUS, Kang and colleagues18 identified necrotic cores and dense neointimal calcification that indicated neoatherosclerosis at a mean follow-up duration of 11.1 ± 7.8 months in their DES cohort. However, the diagnostic capacity of IVUS is hampered by signal interference from metal stent struts. More recently, by virtue of superior spatial resolution, OCT has better enabled clinicians to view and evaluate endothelial response after stent implantation. Kang and associates19 used OCT to identify prevalent TCFA-containing neointima, in-stent neointimal rupture, and intraluminal thrombi within DES. The investigators found that 52% of the lesions had at least one TCFA-containing neointima, 58% had in-stent neointimal rupture, and 58% had intraluminal thrombi at 32.2 months. Of more importance was an association between clinical instability and OCT findings of instability. In the evaluation of previously implanted stents, OCT enables clinicopathologic correlation and provides diagnostic and prognostic information beyond that of conventional angiography.
Our patient had no high-risk clinical features that would have predisposed him to the development of VLST: he was a nondiabetic nonsmoker who had a single stent implanted across a relatively short diseased arterial segment. The focal delayed endothelialization seen on OCT would imply its role in the pathogenesis of our patient's VLST, and SES underexpansion might have been contributory. However, it is unclear why this had not occurred at other times, when antiplatelet therapy was temporarily withheld. It is likely that the exposed struts had been covered by fibrous neointima previously; as neoatherosclerosis developed over time, they became exposed consequent to neoatherosclerotic plaque rupture—hence the STEMI.
This process appears to be dynamic and to evolve over time, and we are concerned that previously successful temporary cessation of antiplatelet therapy does not necessarily predict immunity from future adverse events. This case highlights the deficiencies in the overall understanding of stent thrombosis and our ability to risk-stratify our patients and treat them accordingly. Ongoing research and increasing sophistication in coronary imaging might clarify the incidence and natural history of stent thromboses. Until then, the treatment of patients who have early-generation DES remains a matter of clinical judgment and will need ongoing refinement as new evidence surfaces.
Optical coherence tomogram shows exposed stent struts (asterisk) and associated thrombi (arrow) in the proximal edge of the sirolimus-eluting stent. Luminal area, 6.14 mm2; mean luminal diameter, 2.78 mm; minimal luminal diameter, 2.29 mm; and maximal luminal diameter, 3.24 mm.
Optical coherence tomogram at the mid segment of the sirolimus-eluting stent shows neoatherosclerotic plaque (asterisk) with possible superficial erosion and superimposed thrombi. Luminal area, 2.55 mm2; mean luminal diameter, 1.8 mm; minimal luminal diameter, 1.61 mm; and maximal luminal diameter, 1.94 mm.
Optical coherence tomogram shows homogeneous fibrous neointima (asterisk) covering the distal sirolimus-eluting stent struts. Luminal area, 3.72 mm2; mean luminal diameter, 2.17 mm; minimal luminal diameter, 1.88 mm; and maximal luminal diameter, 2.42 mm.
Optical coherence tomogram shows good stent apposition after high-pressure dilation. Luminal area, 8.07 mm2; mean luminal diameter, 3.2 mm; minimal luminal diameter, 2.78 mm; and maximal luminal diameter, 3.55 mm.
Contributor Notes
From: Eastern Heart Clinic, Prince of Wales Hospital, Randwick 2031, Australia
