Late Stent Thrombosis
A Nuisance in Both Bare Metal and Drug-Eluting Stents
Late stent thrombosis (ST) is for us an old foe that we have tracked repeatedly throughout the history of interventional cardiology. In 1991, we made headlines by publishing in the New England Journal of Medicine a rate of early and late ST of 20% among the first 151 patients having received a wall stent.1 A few years later, when some believed that they had discovered the universal panacea for restenosis (vascular brachytherapy), we were the first to report in the literature 6 cases of late ST in 100 patients having undergone brachytherapy after stent implantation.2 At that time, this seminal observation triggered a wave of observations; the cases of late ST after brachytherapy accrued month after month until the pioneer in the field, Ron Waksman, courageously admitted in an editorial that we were “sitting on a time bomb.”3 Five years later (2004), in the beginning of the drug-eluting stent (DES) era, we reported, together with Waksman’s group, the first 4 cases of late ST.4 In the following 2 years, the incidence of late ST, scrutinized by alerted clinicians, was publicized in reports that included >25 000 patients treated with DES; the incidence ranged from 0.2% in a postmarketing surveillance trial to 1.8% in a small series of multivessel stenting.5–10 Around this period, we realized that Bern and Rotterdam had somewhat diverging incidences of late ST, 0.4% and 0.9%, and we joined efforts to retrospectively assess the rate of late ST over a period of 4 years. A steady rate of 0.6% per year was detected without abatement over a follow-up period of 3 years.11 What makes the experience of these 2 groups unique is that both tertiary institutions embraced the technology of eluting stents at the time of their commercial introduction and decided to treat all comers with this new promising therapeutic approach. Of 8146 patients, we saw 61 patients coming back to our catheterization laboratory with symptoms and angiographic signs of late ST. Of these 61 patients, 3 died in hospital from late ST, and 44 sustained a myocardial infarction (MI). At the 6-month follow-up, an additional 2 patients had died and 2 sustained a reinfarction. In summary, 5 of 8146 patients died as a result of this dreadful complication of ST. These numbers were largely confirmed by the 3 years of follow-up of the Arterial Revascularization Therapies Study Part II, showing a similar annual increase of definite ST of 1.7%, 0.6%, and 0.4% at 1, 2, and 3 years, respectively, in a population with a majority of 3-vessel treatment and with an average stented length of 73 mm. By reporting these figures, we are not trying to minimize the phenomenon, and we recognize that we are describing only the tip of the iceberg by disregarding fatal MI and unexplained death as clinical surrogates of ST, but we do not believe that 0.6% per year could, by any means, be compared with the 6% late ST seen after vascular brachytherapy. In other words, we are not facing a new “vascular brachytherapy syndrome.”
Response by Camenzind et al p 1439
After this introduction on the history of ST, we will analyze more judiciously the contentions by Camenzind et al that DES increase deaths, as headlined by the European Society and World Congress of Cardiology (ESC) newsletter at the congress in Barcelona 2006.12
Are There Physiopathological Reasons to Have Late ST?
Camenzind and coauthors referred to the classic triad of Virchow (altered blood constituents, flow pattern, and endothelial lining). This is without a doubt appealing to the readers, but let us analyze critically their arguments and the validity of their analogical comparison when they equate a stenotic lesion covered by a DES to an atherothrombotic phenomenon in the general circulation.
Their first argument suggests that implantation of DES with concomitant but transient administration of dual platelet therapy would generate at the time of the discontinuation a thrombogenic milieu. Using common sense, we have to point out that ≈99% of the patients worldwide discontinue their thienopyridine medication without experiencing ST within days. We were the first to link the interruption of aspirin with ST in DES.4 In that respect, the relationship between platelet therapy and thrombosis is not unique to DES. In a study reporting on 1236 patients hospitalized for acute coronary syndrome, those who stopped aspirin presented significantly more often with ST-segment elevation acute coronary syndrome. Twenty percent of these cases involved a bare metal stent (BMS) thrombosis on an average of 15.5 months.13 In fact, interruption of aspirin is a risk factor for every patient with atherosclerosis.14 “Noncompliance or withdrawal of aspirin treatment has ominous prognostic implications in subjects with or at moderate-to-high risk for coronary artery disease. Aspirin discontinuation in such patients should be advocated only when bleeding risk clearly overwhelms that of atherothrombotic events.” These are the conclusions of a recently published meta-analysis on the hazards of discontinuing or not adhering to aspirin among 50 279 patients at risk for coronary artery disease.14 Similarly, clopidogrel withdrawal is associated with proinflammatory and prothrombotic effects in patients with diabetes mellitus and coronary artery disease.15 The concern in trials has not been the discontinuation of the antiplatelet therapy but its premature discontinuation, and this concept of premature discontinuation has to be critically discussed.
It is relevant to note that the patients included in the FIM (First in Man) and the RAVEL (Randomized Study with the Sirolimus-Coated Bx Velocity Balloon Expandable Stent in the Treatment of Patients With De Novo Native Coronary Artery Lesions) trials were prescribed 2 months of ticlopidine or clopidogrel. The rationale for the 2-month prescription is that initially (July 1999, trial discussion at the headquarters of Cordis) the First in Man trial was planned as a 60-day safety trial with concomitant use of 60 days of ticlopidine. Later, the protocol was converted into a 4- and 6-month study with quantitative coronary angiography and intravascular ultrasound.
After completion of the European pivotal trial, the American pivotal trial, SIRIUS (Sirolimus-Eluting Stent in De Novo Native Coronary Lesions), was initiated, and for reasons unknown to the authors, the prescription of clopidogrel was prolonged to 3 months, although no thrombotic issue was observed in the RAVEL trial in the first 5 years of follow-up. The decision to prescribe 6 months of clopidogrel in the TAXUS-I trial (and in the subsequent trials) was not discussed in the original report. In other words, the duration of prescription of clopidogrel does not seem to be validated by scientific arguments; consequently, the concept of premature discontinuation is arbitrary as well. So far, there is no evidence-based medicine that prolonged dual-antiplatelet therapy could reduce late ST in DES-treated patients. Of note, one quarter of our patients with late or very late ST were on dual-antiplatelet therapy.11 The only evidence of a preventive effect of dual-antiplatelet therapy on late ST stems from the experience with brachytherapy in which prolongation of dual-antiplatelet therapy was effective in preventing late ST, although a rebound phenomenon in ST was observed after cessation of clopidogrel. At 15 months, the incidence of thrombotic occlusion was 15.9% in the Washington Radiation for In-Stent Restenosis Trial plus 6 months of clopidogrel (WRIST-PLUS) versus 13.5% in WRIST.12,16,17
The postmortem findings in patients with a DES constitute dramatic and compelling evidence that stented arteries show, in a limited number of patients, impaired reendothelialization with uncovered stent struts as a consequence. However, the eminent pathologist Dr Renu Virmani is the first to admit that in postmortem studies there is no common denominator and that the number of patients treated with DES in whom “uncovered stent struts” do not lead to ST is unknown but undoubtedly very large. Still today, the relative impact of the stent platform with a variety of strut thicknesses, polymers, and drugs on ST remains unclear. In her most recent presentation, Dr Virmani demonstrated a complete endothelialization of the struts with cells exhibiting cd31 (antigen surface marker of good endothelial functionality) at 14 days in rabbit iliac arteries stented with a DES eluting everolimus, a sirolimus analog with a sole and minimal alteration of the molecular structure of sirolimus in the binding domain, without any chemical modification of the mTOR binding domain. With such a subtle difference between sirolimus and everolimus, it becomes unclear whether we have to blame the drug, the polymer, or even the platform for the difference in reendothelialization.18
We are aware that DES may induce complex interactions between shear stress and inhibition of neointimal growth, resulting in a peculiar rheological profile that we described early on in Circulation,19 and we have attempted to differentiate delayed neointimal growth (delayed restenosis) from adaptive shear stress–induced growth, which is a long-term physiological process.20,21
But before discussing in more detail the impact of vessel wall remodeling on late ST, we should clarify that late acquired malapposition and stent underexpansion are 2 different entities with different possible consequences. In the early days of BMS, Colombo et al22 convinced the world that incorrect deployment and incomplete apposition of stents were major contributors to subacute ST. This older generation of interventional cardiologists was painstakingly educated to postdilate the stent to avoid this incomplete deployment. Today, the lesson of the past seems to have been forgotten by a new generation of interventional cardiologists who rely too much on the antirestenotic properties of DES and do not care enough for correct mechanical deployment of the “metallic endoprosthesis.” In an extensive and retrospective analysis of BMS patients treated in the previous decade, stent underexpansion (minimum stent area <5.0 mm2) was present in 20% of all restenotic lesions.23 In a comparable but smaller analysis of DES-treated patients, stent underexpansion (using the same definition) was observed in at least 67% of all cases.24
Incorrect deployment with early thrombosis in (non) eluting stents will remain an actuality.25 Although late acquired malapposition also is observed with BMS, this phenomenon has been documented more frequently in patients with DES.26 The question remains whether late acquired malapposition has an unfavorable prognosis with respect to late ST. Hoffman et al have compiled the intravascular ultrasound studies performed in the randomized trials (RAVEL, SIRIUS and E-SIRIUS) (personal communication). Intravascular ultrasound was available in 325 patients and the incidence of incomplete stent apposition at 6 months was 25% in the SES group versus 8.3% in the BMS group. The authors were unable to demonstrate any prognostic impact of incomplete stent apposition on death (2.2% with incomplete stent apposition versus 5.2% without incomplete stent apposition) and major adverse cardiac events (8.9% with incomplete stent apposition versus 12.6% without incomplete stent apposition) in patients treated with sirolimus-eluting stents. Late ST was observed in only 1 of 45 patients with incomplete stent apposition at a 6- to 8-month follow-up. Considering the fact that the era of trials comparing BMS and DES is coming to an end, such a comparison between DES and BMS on that topic might not be obtainable in the future.
The ESC Firestorm
At the ESC Barcelona 2006, some of us were caught off guard by a plenary session combining 3 critical presentations by Edoardo Camenzind, Salim Yousouf, and Alain Nordmann. After having pooled the published data from the 4 pivotal randomized controlled trials assessing the safety and efficacy of the Cypher stent (RAVEL, SIRIUS, E-SIRIUS, and C-SIRIUS), Camenzind and colleagues disclosed a rate of total death and Q-wave MI of 6.3% in the Cypher group versus 3.9% in the control group, with a value of P=0.03. It did not take too long for the lay media to report with sensationalism that DES were increasing the incidence of death and Q-wave MI by 66%. However, what was overlooked is that this meta-analysis was derived from data published in the literature at different time points of follow-up. Camenzind took 2 hard clinical end points, total death and Q-wave MI, and disregarded the non–Q-wave MI incidences, which were substantially lower in the Cypher group, did not fit with the general contention, and would have ruined the statistical foundation (a probability value would not have been reached with a composite of Q- and non–Q-wave MI). The TAXUS results of the perennial competitor Boston Scientific were not questioned because a meta-analysis derived from the randomized TAXUS family trial data in the literature did not show an alarming incidence of death and Q-wave MI.
In contrast to these data, we had access to the patient-level–based data for all 4 trials (RAVEL, SIRIUS, E-SIRIUS, and C-SIRIUS) with a complete follow-up of 4 years.26a These data had been transferred to 2 independent academic statisticians for verification. Previously, the data management had been performed by 2 independent central research organizations located in Rotterdam and Harvard. In these trials, the clinical events had been adjudicated by independent clinical event committees, whereas the incidence of ST had been readjudicated by the Harvard Central Research Institute according to new definitions of ST (Table 1). The definitions were formulated before the ESC congress by a consortium of interventional cardiologists from both sides of the Atlantic, representatives of the Food and Drug Administration, academic central research organizations (Harvard Central Research Institute, Duke Clinical Research Institute, Cardiovascular Research Foundation, and Cardialysis), and representatives from major stent manufacturers. Two meetings held in Washington in March 2006 and in Dublin in June 2006 aimed to reach a consensus on the new standard and to promote consistency for reporting to guarantee the transparency of the data. In the past, inappropriate comparisons and conclusions were based on various definitions with the potential to bias results by choosing definitions most favorable to those conducting analyses. In contrast to the data presented by Camenzind and colleagues, the actual rate of total death and all MI at 4 years was 11.4% in the Cypher group and 10.1% in the control group, with a value of P=0.4. Not being “P valuists,” we recognized that the Kaplan-Meier curves were nevertheless slightly diverging over time. Although not completely correct from a methodological point of view, Camenzind’s report at the ESC became a wakeup call for everybody (the device industry, principal investigators, clinical research organizations, The European Agency for the Evaluation of Medical Products, and the US Food and Drug Administration).
Puzzled by these observations, we started looking for heterogeneity of the treatment effect and discovered that the slight excess in mortality in the total population was due exclusively to the diabetic population. In the 428 diabetic patients enrolled in the 4 pivotal randomized trials, the 3-year survival was 96% in the BMS control group and 87.8% in the Cypher group, with a value of P=002. Because the nondiabetic Kaplan-Meier curves were nicely superimposed, our first reflex was to analyze the incidence of ST in these diabetic patients as a specific cause of mortality. The rate of definite or probable ST was 1.0% (2 of 195 patients) in the Cypher group versus 2.1% (5 of 233) in the control group. However, there was an imbalance between the Cypher (5.6%, 10 of 195) and control (1.7%, 4 of 233) groups in possible ST defined as any unexplained death from 30 days after the procedure. This intense scrutinizing exercise in comparing the death rate in these 428 diabetic patients boiled down to a difference of 6 unexplained deaths in the Cypher group and an excess of 3 cases of definite or probable ST in the control group, whereas the 3 categories of ST were very well balanced in the total population (Table 2).
However, a chronological analysis of the occurrence of these thromboses revealed some subtle differences in timing. In the DES group, the incidence of very late (n=23) primary ST (not a sequela of reintervention) was prominent, and the occurrence of ST after target lesion revascularization was not observed mainly because of the very low incidence of target lesion revascularization in the DES group. In the BMS group, 9 late and 6 very late ST events were documented, of which 10 occurred after reintervention. These incidences of late ST in BMS are not uncommon and have been previously documented in the literature in >9000 patients, with rates of late ST (>30 days) ranging from 0.4% and 0.8%.27–29
These numbers emphasize the fact that restenosis with BMS is not just a nuisance but has potentially severe consequences. In 2006, Chen and colleagues30 reported on the clinical presentation of 1186 consecutive episodes of in-stent restenosis in patients treated with BMS. Whereas 26.4% of the patients required hospitalization for unstable angina, an additional 9.5% presented with MI, and 0.7% died. Nayak et al31 reported a similar 10% incidence of MI resulting from in-stent restenosis, and an additional 12% presented with a 2-fold increase in troponin. Additionally, 2 recent studies proved a close correlation between the extent of restenosis and late mortality.32,33 In other words, by preventing 100 restenoses per 1000 patients (clinical restenosis reduced from ≈20% to 10%), DES could prevent 10 restenosis-related MIs (9.5% of 100 prevented restenosis), and we could surmise that a reduction of 10 per 1000 cases of restenosis-related MIs would be sufficient to offset an increase of 5 per 1000 in very late ST–related MIs, leading to similar late death and MI rates for DES and the BMS control.
How Do We Interpret the Data in Light of the Academic Research Consortium Definitions?
We have 3 options: there is a new problem, and the use of DES results in more (very) late thrombosis than BMS; there is no new problem, and the rate of (very) late ST after DES is similar to that of BMS; and there is a problem, and early and (very) late ST should be abolished. In the near future, we may consider 3 strategies (the Figure).
1. Pharmacological and long-term randomized trials will test the antithrombotic properties of current or novel dual-antiplatelet therapy in trials with death and MI as their primary end points and ST as their secondary end point.
2. Certain current active and passive coated stents claim to have lower rates of ST and will be compared in dedicated randomized trials with long-term follow-up with as a sole primary end point difference in ST with as secondary end point death and MI (eg, REVOLUTION, PROTECT [Prospective Reinfarction Outcomes in Thrombolytic Era Cardizem CD Trial], and TRIAS).
3. The current technology is viewed as imperfect, and the deficiencies should be amended by introducing more biocompatible absorbable coatings and new drugs with biological targets other than smooth muscle cell duplication (eg, thrombotic and inflammatory mechanisms) using dual elution of drugs or a prohealing approach such as the capture of endothelial progenitor cells with or without drug elution.
It is clear that abolishing neointimal hyperplasia is no longer the ultimate goal. Development of more biocompatible and bioabsorbable stents facilitating adequate endothelialization is expected in the near future.
Late ST exists in both DES and BMS. In an all-comer population, angiographic ST in DES seems to occur at a steady rate of 0.4% to 0.6% per year. Clinicians, regulators, and the device industry now realize that clinical surrogates for ST (death and MI) have to be incorporated into the long-term follow-up of the patient to capture late and very late ST patients who do not reach the catheterization laboratory to have their thrombosis angiographically confirmed. Fortunately, a first consensus on these angiographic and clinical end-point definitions has been reached under the umbrella of the Academic Research Consortium. Retrospective analysis of ST, applying the Academic Research Consortium definitions in randomized controlled trials, does not disclose a different rate of late ST between BMS and DES up to 4 years. However, the chronology and circumstances of occurrence seem quite different. In DES, late ST occurs later than in BMS and seems to appear as primary thrombosis, whereas in BMS, a certain number of late thromboses are related to repeat interventions of the target lesion. Dedicated research is warranted to further elucidate the role of endothelial dysfunction, malapposition, and prolonged antiplatelet therapy. Currently, the second generation of DES is attempting to resolve the issues discovered so far with the first generation of DES.
Costa MA, Sabate M, van der Giessen WJ, Kay IP, Cervinka P, Ligthart JM, Serrano P, Coen VL, Levendag PC, Serruys PW. Late coronary occlusion after intracoronary brachytherapy. Circulation. 1999; 100: 789–792.
Waksman R. Late thrombosis after radiation: sitting on a time bomb. Circulation. 1999; 100: 780–782.
McFadden EP, Stabile E, Regar E, Cheneau E, Ong AT, Kinnaird T, Suddath WO, Weissman NJ, Torguson R, Kent KM, Pichard AD, Satler LF, Waksman R, Serruys PW. Late thrombosis in drug-eluting coronary stents after discontinuation of antiplatelet therapy. Lancet. 2004; 364: 1519–1521.
Iakovou I, Schmidt T, Bonizzoni E, Ge L, Sangiorgi GM, Stankovic G, Airoldi F, Chieffo A, Montorfano M, Carlino M, Michev I, Corvaja N, Briguori C, Gerckens U, Grube E, Colombo A. Incidence, predictors, and outcome of thrombosis after successful implantation of drug-eluting stents. JAMA. 2005; 293: 2126–2130.
Kuchulakanti PK, Chu WW, Torguson R, Ohlmann P, Rha SW, Clavijo LC, Kim SW, Bui A, Gevorkian N, Xue Z, Smith K, Fournadjieva J, Suddath WO, Satler LF, Pichard AD, Kent KM, Waksman R. Correlates and long-term outcomes of angiographically proven stent thrombosis with sirolimus- and paclitaxel-eluting stents. Circulation. 2006; 113: 1108–1113.
Urban P, Gershlick AH, Guagliumi G, Guyon P, Lotan C, Schofer J, Seth A, Sousa JE, Wijns W, Berge C, Deme M, Stoll HP. Safety of coronary sirolimus-eluting stents in daily clinical practice: one-year follow-up of the e-Cypher registry. Circulation. 2006; 113: 1434–1441.
Daemen J, Wenaweser P, Tsuchida K, Abrecht L, Vaina S, Morger C, Kukreja N, Jüni P, Sianos G, Hellige G, Van Domburg R, Hess O, Boersma E, Meier B, Windecker S, Serruys P. Early and late coronary stent thrombosis of sirolimus-eluting and paclitaxel-eluting stents in routine clinical practice: data from a large two-institutional cohort study. Lancet. In press.
Camenzind E. Do drug-eluting stent increase death? ESC Congress News. Barcelona, Spain; 2006.
Biondi-Zoccai GG, Lotrionte M, Agostoni P, Abbate A, Fusaro M, Burzotta F, Testa L, Sheiban I, Sangiorgi G. A systematic review and meta-analysis on the hazards of discontinuing or not adhering to aspirin among 50,279 patients at risk for coronary artery disease. Eur Heart J. 2006; 27: 2667–2674.
Angiolillo DJ, Fernandez-Ortiz A, Bernardo E, Ramirez C, Sabate M, Jimenez-Quevedo P, Hernandez R, Moreno R, Escaned J, Alfonso F, Banuelos C, Costa MA, Bass TA, Macaya C. Clopidogrel withdrawal is associated with proinflammatory and prothrombotic effects in patients with diabetes and coronary artery disease. Diabetes. 2006; 55: 780–784.
Waksman R, Ajani AE, White RL, Pinnow E, Dieble R, Bui AB, Taaffe M, Gruberg L, Mintz GS, Satler LF, Pichard AD, Kent KK, Lindsay J. Prolonged antiplatelet therapy to prevent late thrombosis after intracoronary gamma-radiation in patients with in-stent restenosis: Washington Radiation for In-Stent Restenosis Trial plus 6 months of clopidogrel (WRIST PLUS). Circulation. 2001; 103: 2332–2335.
Waksman R, Bhargava B, White L, Chan RC, Mehran R, Lansky AJ, Mintz GS, Satler LF, Pichard AD, Leon MB, Kent KK. Intracoronary beta-radiation therapy inhibits recurrence of in-stent restenosis. Circulation. 2000; 101: 1895–1898.
Virmani R. Safety of Polymer Coating on Drug-Eluting Stents: What You Don’t Know Could Kill You. Washington, DC: Transcatheter Cardiovascular Therapeutics; Washington DC; 2006.
Aoki J, Colombo A, Dudek D, Banning AP, Drzewiecki J, Zmudka K, Schiele F, Russell ME, Koglin J, Serruys PW. Persistent remodeling and neointimal suppression 2 years after polymer-based, paclitaxel-eluting stent implantation: insights from serial intravascular ultrasound analysis in the TAXUS II study. Circulation. 2005; 112: 3876–3883.
Wentzel JJ, Janssen E, Vos J, Schuurbiers JC, Krams R, Serruys PW, de Feyter PJ, Slager CJ. Extension of increased atherosclerotic wall thickness into high shear stress regions is associated with loss of compensatory remodeling. Circulation. 2003; 108: 17–23.
Serruys PW, Di Mario C. Who was thrombogenic: the stent or the doctor? Circulation. 1995; 91: 1891–1893.
Castagna MT, Mintz GS, Leiboff BO, Ahmed JM, Mehran R, Satler LF, Kent KM, Pichard AD, Weissman NJ. The contribution of “mechanical” problems to in-stent restenosis: an intravascular ultrasonographic analysis of 1090 consecutive in-stent restenosis lesions. Am Heart J. 2001; 142: 970–974.
Takebayashi H, Kobayashi Y, Mintz GS, Carlier SG, Fujii K, Yasuda T, Moussa I, Mehran R, Dangas GD, Collins MB, Kreps E, Lansky AJ, Stone GW, Leon MB, Moses JW. Intravascular ultrasound assessment of lesions with target vessel failure after sirolimus-eluting stent implantation. Am J Cardiol. 2005; 95: 498–502.
Uren NG, Schwarzacher SP, Metz JA, Lee DP, Honda Y, Yeung AC, Fitzgerald PJ, Yock PG. Predictors and outcomes of stent thrombosis: an intravascular ultrasound registry. Eur Heart J. 2002; 23: 124–132.
Hong MK, Mintz GS, Lee CW, Park DW, Park KM, Lee BK, Kim YH, Song JM, Han KH, Kang DH, Cheong SS, Song JK, Kim JJ, Park SW, Park SJ. Late stent malapposition after drug-eluting stent implantation: an intravascular ultrasound analysis with long-term follow-up. Circulation. 2006; 113: 414–419.
Wenaweser P, Rey C, Eberli FR, Togni M, Tuller D, Locher S, Remondino A, Seiler C, Hess OM, Meier B, Windecker S. Stent thrombosis following bare-metal stent implantation: success of emergency percutaneous coronary intervention and predictors of adverse outcome. Eur Heart J. 2005; 26: 1180–1187.
Van Belle E, Ketelers R, Bauters C, Perie M, Abolmaali K, Richard F, Lablanche JM, McFadden EP, Bertrand ME. Patency of percutaneous transluminal coronary angioplasty sites at 6-month angiographic follow-up: a key determinant of survival in diabetics after coronary balloon angioplasty. Circulation. 2001; 103: 1218–1224.
Response to Serruys and Daemen
Edoardo Camenzind, MD, P. Gabriel Steg, MD, William Wijns, MD
As always, the perspective provided by Serruys and Daemen is factual and informative. We agree that (very) late stent thrombosis (LST) with first-generation drug-eluting stents (gen1-DES) exists and may cause myocardial infarction (MI) or death. We agree that the global safety-efficacy balance can be assessed from individual patient-based analyses of data sets recently made available to selected investigators by the industry. Serruys et al suggest that diabetic patients have a less favorable safety profile, as predicted by our model. If confirmed, implications would be far-reaching because most physicians prefer gen1-DES in high-risk patients, such as those with diabetes, and are sometimes encouraged in this preference by healthcare payers and regulators. We believe that the biological processes behind LST are genuine to therapies such as brachytherapy and drug-device combinations that interfere with the healing response to injury. We still believe that discontinuation of antiplatelet therapy is an important trigger of ST. Events may occur remotely from the discontinuation of therapy because the risk for ST is restricted to patients with site-specific incomplete vascular healing and is modulated by systemic prothrombotic factors. Blanket indefinite dual-antiplatelet therapy may avoid a fraction of (very) LST at the expense of an increased risk of severe bleeding events and higher costs. Most important, we wish to highlight remaining concerns over the definitions of LST: (1) The proposed definitions consider any MI as a potential ST. This puts procedure-related MIs that are clinically silent and diagnosed by small enzyme leaks on a par with large Q-wave MIs caused by ST, thereby diluting the true thrombosis rates. (2) In the “revised” analysis, events resulting from the use of brachytherapy or gen1-DES for the treatment of BMS restenosis (control arm) were counted. This means that the very drug-device combination whose safety is being scrutinized now contaminates outcome in the control group. (3) Systematic angiographic follow-up leads to non-clinically driven redilatations. These unnecessary procedures are associated with procedural complications and implantation of new BMS or gen1-DES. These fresh implantations generate the potential for late complications, and in the “revised” analysis, these accrued events compensate for the late hazards that result from the primary use of gen1-DES in the “active” arm. Thus, the currently propagated “by intention-to-treat” analysis maximizes event rates in the BMS arm, some of which are of limited clinical relevance (procedure-related myonecrosis), whereas others are caused by gen1-DES (late crossover). This confuses the assessment of safety and likely results again in an underestimation of the imbalance in death and MI rates between BMS and gen1-DES. For the above-mentioned reasons, the outcomes of both analyses will necessarily be different. Until the “ideal” DES becomes available, clinicians are in need of post-hoc safety-efficacy analyses in various lesions or patients subgroups, as can be derived from pooled individual patient databases. It would be catastrophic if analysis efforts such as those initiated by the self-appointed Academic Research Consortium were wasted in denial or in minimization of the problem of (very) LST. When unanswered issues cannot be addressed from post hoc analyses of existing data, adequately designed prospective trials should be conducted.
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.