Long-term Outcomes in Patients Undergoing Percutaneous Coronary Intervention with Drug-eluting Stents

Roberta Rossini; Giuseppe Musumeci; Alessandro Aprile; Orazio Valsecchi


Expert Rev Pharmacoeconomics Outcomes Res. 2010;10(1):49-61. 

In This Article

Safety of DES & Stent Thrombosis

Drug-eluting stents were quickly adopted into clinical practice after the publication of the randomized trials, which demonstrated their superiority over BMS in the reduction of restenosis and repeat TLR.[22,36] The use of DES in clinical practice has expanded to the majority of coronary lesion subsets (e.g., de novo complex lesions, long lesions and small vessels), to high-risk patients (e.g., multivessel angioplasty, patients with diabetes mellitus) and, more recently, to primary PCI for ST-segment elevation acute MI.[37,38]

In September 2006, at the World Congress of Cardiology, the results of a meta-analysis of the published randomized clinical studies comparing DES to BMS by Camenzind et al. were presented by Eduardo Camenzind at a late-breaking session. Camenzind reported that the incidence of the combined end points of total mortality and Q-wave MI were 38 and 16% greater for SES and PES, respectively, compared with BMS. "As a result", Camenzind concluded, "until safer second-generation DES are available, interventionalists should avoid the indiscriminate use of SES and PES".[39]

This presentation had an immediate effect in the field of cardiology. During the same World Congress of Cardiology session, Salim Yusuf, Chair in cardiology and director of the Population Research Institute at McMaster University (Ontario, Canada) called for a cessation of DES implantation. The media responded to the presentations of Camenzind and Yusuf with articles questioning the safety of DES. In 2006, as a consequence, DES use in the USA decreased almost 20% from as high as 89% at the beginning of the year and the financial impact on the device industry was substantial, jeopardizing future funding of next-generation DES research and development.[40] Owing to the relevant medical and economic implications of the controversy led to the release of two FDA advisories in 2006. Protracted dual antiplatelet therapy (12 months) is now advised after DES implantation.[41,42]


According to Camenzind, the pathophysiologic mechanism underlying stent thrombosis consisted of local changes in vascular biology, in combination with systemic alterations of coagulation pathways.[39] These factors, which are reminiscent of the pathophysiological mechanisms described in 1856 by Rudolf Virchow, are: an abnormal vessel wall lining (e.g., incomplete endothelialization), an abnormal blood flow pattern (e.g., slow flow) and altered blood constituents (e.g., increased blood thrombogenicity). Any of these elements alone or in combination favors intravascular thrombus formation.[43,44]

Abnormal Vessel Wall Lining Unlike BMS, where a new endothelial layer covers the stent struts, re-establishing a 'normal' coronary vessel wall lining, DES inhibit or may even abolish this physiological vessel wall healing, leaving the struts in direct contact with flowing blood and blood elements. The lower rate of endothelial stent coverage generates a long lasting, if not permanent, unhealed vessel wall surface favoring platelet adhesion and aggregation, which may eventually cause thrombus formation.[45,46]

Abnormal Flow Pattern Histologic[45,46] and intravascular ultrasound[47] data showed the presence of an inflammatory vessel wall response to DES implantation that occurs within 8 months. Coronary structural changes (vessel widening) may induce, according to the law of continuity,[48] slow flow velocities, thus promoting thrombogenesis.

Abnormal Blood Constituents Increased blood thrombogenicity plays a crucial role in favoring late stent thrombosis. The prothrombotic effect of the reduction or discontinuation of aspirin, ADP receptor inhibitors (e.g., clopidogrel), or both may be further potentiated by an ongoing systemic inflammatory reaction (e.g., fever, postoperative course, malignancy) or by dehydration.[48–50]

Interpreting the Data

After having pooled the published data from the four pivotal RCTs assessing the safety and efficacy of the SES (RAVEL, SIRIUS, The European SIRIUS [E-SIRIUS] and the Canadian SIRIUS [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.[39]

In contrast to the data presented by Camenzind and colleagues, Serruys et al. analyzed the same trials reaching different results: 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.[51]

The reason for such different results, although comparing the same trials, might be due to the different end points used. In Camenzid's analysis only Q-MIs were included, whereas Serruys and colleagues included all MIs. Thereafter, in Serruys' analysis, also periprocedural MIs were analyzed, which were probably higher in the BMS group due to the higher rate of TLR. The rate of Q-wave MI was higher in DES than in BMS in Camenzid's and in Serruys' analyses (1.6 vs 0.6%; p = 0.06, and 2.1 vs 1.3%; p = 0.20, respectively). The incidence of non-Q MI, however, which was assessed only by Serruys and colleagues, was higher in BMS group (5.0 vs 4.3%, p = 0.48). Moreover, it is noteworthy that stent thrombosis offers different interpretations in DES and BMS. In the aforementioned analysis, in the DES group, the incidence of very-late primary stent thrombosis (not a sequela of reintervention) was prominent, and the occurrence of stent thrombosis after TLR was not observed mainly because of the very low incidence of TLR in the DES group. In the BMS group, nine late and six very late stent thrombosis events were documented, of which ten occurred after reintervention. Including stent thromboses that follow interim revascularization of the target vessel has a major impact on late stent thrombosis rates. As a result, thrombosis events that are subsequent to the treatment of in-stent restenosis in the control arm (e.g., with in-stent implantation of first-generation DES or brachytherapy) may counterbalance the spontaneously occurring late stent thrombosis in the DES arm. Moreover, because of the small sample size of many of these trials, even a small number of patients excluded from analysis (e.g., lost to follow-up, revoked patient consent or follow-up out of the predefined time-window) may have a large impact on comparative outcomes. Finally, it must be underlined that Camenzind performed a study-based analysis, whereas Serruys performed a patient-based analysis. All these methodological issues illustrate the complexities entailed by the analysis, presentation and comparison of such trials and databases.

Since then, several registries and meta-analyses comparing DES and BMS published, with contradictory results.

In 2005, Ong et al. highlighted that late angiographic stent thrombosis occurred with an incidence of at least 0.35% and possibly up to 0.72% after DES implantation.[52] Furthermore, they observed that it could also occur not only in temporal relation to complete cessation of antiplatelet therapy, but could also occur shortly after clopidogrel is stopped but aspirin administration continued, and unexpectedly, remote from clopidogrel cessation when patients were clinically stable on long-term aspirin therapy.

Central to the controversy over the relative safety of DES compared with BMS in terms of stent thrombosis was the Basel Stent Cost–Effectiveness Trial – Late Thrombotic Events (BASKET-LATE) analysis.[53] This study was, in part, a cost–effectiveness analysis of DES versus BMS in a relatively small number of patients in a single center. The 6-month outcomes showed no significant differences between DES and BMS with regard to cardiac death or MI, with a trend toward reduced TVR with DES (4.6 vs 7.8%, respectively; p = 0.08) and a significant reduction in major adverse cardiac events with DES (7.2 vs 12.1%; p = 0.02) driven by the reduced rate of TVR. The BASKET-LATE analysis assessed outcomes in patients free of major events at 6 months who were followed up for a year after discontinuation of clopidogrel (i.e., 7–18 months). Although the overall 18-month rates of cardiac death/MI did not differ between DES and BMS patients, there was a significant increase during the additional 12-month follow-up in patients who received DES (4.9 vs 1.3%). Thus, after the discontinuation of clopidogrel, the benefit of DES in reducing TVR was maintained but had to be balanced against an increase in late cardiac death or nonfatal MI, possibly related to late stent thrombosis.

Daemen's study in 2007 was the first to address the occurrence of stent thrombosis more than 1 year after DES implantation. During a follow-up period of up to 3 years in a large group of patients treated with the unrestricted use of DES, stent thrombosis occurred with an incidence density of 1.3 per 100 person-years and a cumulative incidence of 2.9% at 3 years. The incidence of late stent thrombosis did not diminish, but continued at a steady rate of 0.6% per year during the first 3 years. ACS at presentation and diabetes were independent predictors of overall stent thrombosis. Early and late stent thrombosis occurred with both types of DES, but late stent thrombosis was more frequently observed with PES than with SES.[28]

A meta-analysis of 17 randomized trials comparing DES and BMS reported by Nordmann et al. found no difference in cardiac or all-cause mortality between SES and BMS over 4 years, or between PES and BMS over 3 years, but it noted an increased risk of noncardiac mortality in year 2 (OR: 2.74) and a marginally increased risk in year 3 (OR: 2.04) with SES versus BMS.[54] The most frequently cited causes of noncardiac death were stroke, lung disease and cancer (lymphoma and malignancies of the lung, prostate, pancreas, kidney and rectum). It remains difficult to know what to make of these data, although it seems clear that thrombotic events were not responsible for any excess in noncardiac mortality.

The Swedish Coronary Angiography and Angioplasty Registry (SCAAR) study group reached contradictory conclusions in two different analyses with extended population and prolonged follow-up. They first assessed the long-term outcome in 19,771 patients who underwent stent implantation in Sweden in 2003 and 2004, and conducted a follow-up analysis of death and MI, using other national registries.[55] At 6 months, there was a trend toward a lower unadjusted event rate in patients with DES than in those with BMS. However, after 6 months, patients with DES had a significantly higher event rate (adjusted RR: 1.20). At 3 years, mortality was significantly higher in patients with DES (adjusted RR: 1.18), and from 6 months to 3 years, the adjusted relative risk for death in this group was 1.32. The increase in event rate was observed only after the first 6 months. Although no details on long-term use of clopidogrel were available, most patients were prescribed dual antiplatelet treatment for 6 months after implantation of DES but for only 1–3 months after implantation of BMS. Therefore, the early gain and late loss of clinical events in the group with DES might have been related to better protection with clopidogrel in the early phase and a prolonged need for such protection after 6 months.

To obtain a more reliable estimate of long-term outcome and efficacy, the SCAAR study group extended the study population and the follow-up, comparing patients who received only one DES with those who received only one BMS, with a mean follow-up of 2.7 years.[56] There was no overall difference between the group that received DES and the group that received BMS in the combined end point of death or MI or the individual end points of death and MI, and there was no significant difference in outcome among subgroups stratified according to the indication for stent implantation. Patients who received DES in 2003 had a significantly higher rate of late events than patients who received BMS in the same year, but no difference in outcome among patients treated in later years was observed. The average rate of restenosis during the first year was 3.0 events per 100 patient-years with DES versus 4.7 with BMS (adjusted RR: 0.43).

The difference might be explained with the following considerations. The conclusion of the first SCAAR study was based on the total patient population for the 2003–2004 period, which included patients who received multiple stents.[55] In the second SCAAR study,[56] the same authors focused their attention to patients who received only one stent (either a DES or a BMS) during the index procedure. Moreover, a change in the outcome over time emerged, with an early outcome that became gradually worse in the BMS group and gradually better in the DES group. The most important change in clinical practice during the extended study period was an increase in the use of primary PCI for patients with ST-segment elevation MI. The proportion of such patients who received stents increased more in the BMS group than in the DES group. An increasing proportion of patients pretreated with clopidogrel, progressively higher balloon pressures and a gradual increase in the duration of clopidogrel treatment after the implantation of DES might have further contributed to the relatively lower rate of late events in the DES group in the second SCAAR study compared with the first one.

Owing to the relatively restrictive and nonuniform definitions of stent thrombosis and to the limited power to detect low-frequency events, Mauri et al. sought to increase the power to detect differences in stent thrombosis in data available from extended follow-up of randomized trials of DES and to evaluate the effect of stent thrombosis on late mortality.[57] A new standardized, hierarchical definition of stent thrombosis was implemented by the Academic Research Consortium in order to uniform evaluation of events in a pooled analysis of eight randomized trials of DES compared with their respective BMS. The new definition was based upon timing (acute, subacute, late and very late) and level of documentation (definite, probable and possible). The analysis involved 878 patients treated with SES, 1400 treated with PES, and 2267 treated with BMS. They concluded that the incidence of stent thrombosis did not differ significantly between patients with DES and those with BMS in randomized clinical trials, although the power to detect small differences in rates was limited.

However, the need for a categorization of stent thrombosis that illustrates that late stent thombosis is highly dependent on definition and adjudication. This 'unifying' definition improves the comparability across trials but will not necessarily guarantee a more accurate assessment of the incidence of late stent thombosis with DES. An example is that including stent thromboses that follow interim revascularization of the target vessel has a major impact on late stent thombosis rates (excluded according to study protocol definitions). As a result, thrombosis events that are consecutive to the treatment of in-stent restenosis in the control arm (e.g., with in-stent implantation of DES) may counterbalance the spontaneously occurring late stent thombosis in the DES arm.

Other registries and analyses using patient-level data from randomized trials consistently support the contention that the use of DES is not associated with increased risk of stent thrombosis, MI or cardiac death, but it is associated with a reduced need for reintervention.[54,58–60]

The discrepancies among different studies may be explained by several reasons (Table 4). First, given the relatively infrequent occurrence of death, MI and stent thrombosis, some of the trials were underpowered to detect small differences in event rates. Second, the majority of the studies that demonstrated the safety DES included all MIs as an end point, not only Q-MI. This might have contributed to counterbalance an increased risk of stent thrombosis in DES by including small periprocedural MI, more frequent in BMS due to the higher risk of in stent restenosis. Third, in meta-analyses, aggregate data from published reports, rather than data from individual patients were examined, thus leading to an imprecise estimate of the overall treatment effect and interstudy heterogeneity. Fourth, a uniform definition of late stent thrombosis was not used in all the trials performed. Fifth, the pivotal SES and PES trials did not include secondary thrombotic events occurring after a TLR. Sixth, in some studies, DES and BMS arms were compared by 'intention-to-treat' (which could maximize event rates in BMS arm), whereas in others, the comparisons were made between patients 'actually treated' with DES or BMS. Finally, detailed data regarding the use of antiplatelet medication throughout the follow-up period were often not available, precluding firm recommendations regarding the optimal duration of thienopyridine administration.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.