Reports on increased frequency of myocardial infarction and stent thrombosis after use of drug-eluting stents (DES) during long-term follow-up have recently raised significant concerns.1^–4 Delayed and impaired endothelialisation of stent struts, late endothelial dysfunction and hypersensitivity reactions have been considered to be the main mechanism for late vascular events.5 Incomplete stent apposition (ISA) may also contribute to late stent thrombosis as indicated by a pathology study.5 ISA has been reported in 4–5% of cases at follow-up intravascular ultrasound (IVUS) after bare-metal stent (BMS) implantation.6 7 The inhibition of vascular smooth muscle cell proliferation after implantation of DES as the mechanism of reduced restenosis may cause also an increased frequency of ISA at follow-up. Previous reports on DES have indicated that acquired ISA (ISA seen at follow-up but not immediately after stent placement) has no negative clinical sequel during a 12-month follow-up period after DES implantation.8^–10 However, recent studies on ISA were mainly limited to follow-up periods of 12 months after stent implantation. There is insufficient knowledge of whether the finding of ISA by IVUS has negative long-term clinical consequences. The real impact of ISA on clinical events may become obvious only after longer observation periods.3 4 Reliable long-term data are now available from early randomised trials on DES versus BMS and can be pooled to conduct analyses with greater power than those in previous studies.

This analysis aimed at identifying the frequency and predictors of ISA in sirolimus-eluting stents (SES) as compared with BMS and at defining the clinical impact of ISA during a longer follow-up period than in previous studies. Thus, this study evaluated the impact of ISA during a 4-year follow-up period in a relatively large patient group using pooled data from three randomised studies.

PATIENTS AND METHODS

Patients and lesions

The study group comprised patients with IVUS analysis at angiographic follow-up from the SIRolImUS-eluting stent in de novo coronary lesions (SIRIUS) trial,11 the European, multicentre, randomised, double-blind trial of the SIRolImUS-coated Bx VELOCITY stent in the treatment of patients with de novo coronary artery lesions (E-SIRIUS)12 and the RAndomised study with the sirolimus-eluting Bx VELocity-stent (RAVEL)13—three multicentre, randomised double-blind studies that evaluated the safety and effectiveness of the sirolimus-coated Bx VELOCITY stent (Cordis Corp, Warren, New Jersey, USA) compared with the uncoated Bx VELOCITY balloon-expandable stent (Cordis Corp) in de novo native coronary artery lesions. IVUS follow-up was performed 8 months after stent implantation in E-SIRIUS and SIRIUS and at 6 months in RAVEL. Inclusion criteria for each of the studies and the IVUS substudies have been reported elsewhere.11^–14

Considering all three studies, lesions of up to 32 mm length and vessel ⩾2.5 mm and ⩽3.5 mm in diameter were included. Multiple stent use was allowed as needed at the operator’s discretion in SIRIUS and E-SIRIUS. All patients received aspirin 82–100 mg/day indefinitely and either a 500 mg loading dose of ticlopidine or a 300 mg loading dose of clopidogrel followed by ticlopidine 250 mg twice a day or clopidogrel 75 mg/day. Dual platelet therapy consisting of aspirin and ticlopidine or clopidogrel was recommended for 2 months in E-SIRIUS and RAVEL patients and for 3 months in SIRIUS patients. Subsequently, only aspirin was recommended. The independent ethics committees of all participating centres approved the study protocols. All patients gave written informed consent before enrolment. The primary endpoint results of these studies have been reported elsewhere.11^–13

IVUS imaging and analysis

After administration of intracoronary glyceryl trinitrate, IVUS was performed with automated pullback at 0.5 mm/s using commercially available scanners. All IVUS procedures were recorded on VHS videotapes. The IVUS images were interpreted at two independent core laboratories. IVUS studies of the RAVEL and E-SIRIUS trials were analysed at Cardialysis, Rotterdam, the Netherlands. IVUS studies of the SIRIUS trial were analysed at Cardiovascular Core Analysis Laboratory, Stanford University, Stanford, California, USA. Analysis was performed blinded to the treatment protocols. Qualitative analysis involved review of all IVUS tapes for the presence of ISA using the same definition of ISA. ISA was defined as one or more stent struts clearly separated from the vessel wall with evidence of blood speckles behind the strut without overlapping side branches.

Quantitative analysis was performed using computerised planimetry at the reference segment and within the stented segment. In the reference segments, lumen area and vessel area were determined. In the stented segment, lumen area, vessel area and stent area were measured. Neointimal hyperplasia area was calculated as stent area − lumen area. Mean lumen, vessel and stent area as the average of all measurements throughout the stent length were calculated.

Quantitative coronary angiography

With the guiding catheter used for magnification–calibration, lesion length, minimal lumen diameter of the lesion and diameters of the reference segments were measured before stenting and minimal lumen diameter of the lesion and stent after stenting and at follow-up.

Definition of clinical events during follow-up

Long-term clinical follow-up data were obtained as part of the routine clinical follow-up within the three randomised studies. Data were obtained either from outpatient visits or telephone interviews. Death was classified as cardiac versus non-cardiac. Myocardial infarction was defined as an increase of the MB fraction of creatine kinase to a value three times the upper limit of the normal range and target-lesion revascularisation was defined as repeat percutaneous or surgical intervention of the lesion stented during the study procedure. Major adverse cardiac event (MACE) was defined as death, myocardial infarction, and target-lesion revascularisation.

Stent thrombosis was defined according to the study protocols as the occurrence of acute symptoms in combination with angiographically documented thrombolysis in myocardial infarction (TIMI) flow 0 or 1 or the presence of flow-limiting thrombus (TIMI flow 1 or 2). Stent thrombosis was defined as subacute, if occurring within the first 30 days after stent implantation and as late if occurring later than 30 days after stent implantation.

Follow-up data over a period of 4 years were available for all patients of the SIRIUS, E-SIRIUS and RAVEL trials included in this analysis.

Statistical analysis

Statistical analysis was performed using an SPSS software program, version 14.0 (SPSS Inc).Continuous variables are presented as mean (SD) and categorical data are presented as frequencies. Continuous variables were compared with an unpaired or paired Student t test or Mann–Whitney U test. Categorical variables were compared using χ² statistics or Fisher exact test. Event-free survival was analysed with the Kaplan–Meier method, and the corresponding probability value was obtained from a log-rank test. Analysis was performed for SES and for BMS. Multiple stepwise logistic regression analysis was performed to assess independent predictors of ISA at follow-up. To evaluate the impact of ISA at follow-up on long-term clinical event rates, independent predictors of MACE, target lesion revascularisation and target vessel failure at 4 year follow-up were determined. Predictors of ISA and clinical events were determined for all patients and for the SES and BMS patients separately. A probability value <0.05 was considered significant.

RESULTS

Frequency of ISA at follow-up

IVUS analyses at angiographic follow-up were available from 179 patients of the SIRIUS trial (SES: n = 103; BMS = 76), 95 patients of the RAVEL trial (SES: n = 48; BMS: n = 47) and 51 patients of the E-SIRIUS trial (SES: n = 29, BMS: n = 22). Overall IVUS at follow-up was available in 325 patients, 180 patients with SES and 145 patients with BMS. ISA at IVUS follow-up was observed in 45 patients (25%) of the SES group and 12 patients (8.3%) of the BMS group (p<0.001).

Predictors of ISA at follow-up

Table 1 gives the baseline clinical characteristics and procedural characteristics of SES and BMS patients with and without ISA at follow-up.

Among the SES group patients there were no differences between those with and without ISA at follow-up apart from a lower frequency of diabetes, a higher frequency of Canadian Cardiology Society (CCS) class III or IV angina at stent implantation and a tendency for longer stent length in patients with ISA at follow-up. These differences were not seen in the BMS group patients.

Considering the angiographic baseline characteristics, SES patients with ISA at follow-up had slightly longer lesions than those without ISA at follow-up (table 2). Again this difference was not seen in the BMS patients.

No differences in IVUS findings after intervention (table 3) were determined between SES patients with and without ISA at follow-up considering only the 132 patients evaluated from the SIRIUS and E-SIRIUS study.

At IVUS follow-up external elastic membrane (EEM) cross-sectional area (CSA) was greater in SES and in BMS patients with ISA than in patients without ISA. In the 230 patients with IVUS after intervention and at follow-up, change in EEM (ΔEEM) CSA could be calculated. It was significantly greater in ISA patients of the BMS group and tended to be greater in ISA patients of the SES group than in the non-ISA patients.

Multivariate logistic regression analysis was performed to determine predictors of ISA at follow-up. The following variables were tested (all with p<0.2 in univariate analysis): diabetes, cigarette smoking, CCS group III or IV at stent implantation, reference vessel diameter, lesion length, lesion calcification, treatment group (SES vs BMS), number of stents implanted and total stent length. Analysis considering patients treated with SES and BMS, indicated stent type being a SES stent (odds ratio (OR) = 4.47, 95% CI = 2.09 to 9.56, p<0.001) and CCS class III or IV at stent implantation (OR = 2.54, 95% CI 1.33 to 4.86, p = 0.005) to be independent predictors of ISA at follow-up. Considering only the SES patients, CCS III or IV at stent implantation (OR = 4.69, 95% CI 2.15 to 10.23, p<0.001) and absence of diabetes (OR = 3.42, 95% CI 1.05 to 11.1, p = 0.041) were predictive for ISA at follow-up. No independent predictor of ISA could be defined considering the BMS patients only.

Long-term clinical impact of ISA at follow-up

Clinical follow-up was available for all 325 patients during a 4-year follow-up period. Event-free survival at 1 and 4 years was significantly higher in the SES group than in the BMS group. The comparison of clinical event rates including the rate of stent thrombosis between SES patients with ISA versus those without ISA demonstrated no significant differences at 1 and 4 years (table 4).

There was a trend for more myocardial infarctions in BMS and SES patients with ISA. However, only five of all 21 myocardial infarcts in the BMS and SES patients were related to the target vessel. Kaplan–Meier curves for MACE-free survival of ISA and non-ISA patients over a 4-year follow-up period were identical (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 MACE-free survival curves for patients with sirolimus-eluting stents and either complete stent apposition (non-ISA) or incomplete stent apposition (ISA).

Multivariate logistic regression analysis was performed to determine predictors of MACE at the 4-year follow-up. The following variables were tested (all with p<0.2 in univariate analysis): gender, diabetes, history of myocardial infarction, CCS group III or IV at stent implantation, lesion length, lesion angulation, pre-procedural thrombus, treatment group (SES vs BMS), number of stents implanted, total stent length. Considering all 325 SES and BMS patients, treatment group (SES vs BMS, OR = 0.27, 95% CI 0.14 to 0.54, p<0.001), number of stents implanted (OR = 1.06, 95% CI 1.03 to 1.11, p<0.001) and gender (male OR = 2.50; 95% CI 1.01 to 6.18, p = 0.046) were found to be predictive for MACE at the 4-year follow-up. Considering only the SES group, total stent length (per mm; OR = 1.06, 95% CI 1.02 to 1.12, p = 0.010) was the only predictor for MACE at the 4-year follow-up. ISA at follow-up was not predictive for MACE at the 4-year follow-up considering all SES and BMS patients or the SES group patients only. Similarly, ISA was not a predictor for target lesion revascularisation or target vessel failure during the 4-year follow-up period in either patient group.

DISCUSSION

The major findings of this study are the following: (a) the incidence of ISA at follow-up angiography is higher with SES than with BMS; (b) vessel area at follow-up is greater in lesions with ISA, indicating excess positive remodelling as the cause of ISA; (c) CCS III or IV angina pectoris at stent implantation and absence of diabetes are predictive for ISA at follow-up after SES implantation; (d) MACE is not more frequent in patients with SES and ISA than in patients with SES and non-ISA even during a 4-year follow-up period.

Late ISA

ISA at angiographic follow-up has been observed after placement of BMS.6 7 15 It is either due to unresolved ISA after stent implantation or late-acquired ISA. Several mechanisms of late-acquired ISA have been postulated. Positive remodelling of the vessel with an increase in total vessel area out of proportion to the increase in persistent plaque and media area has been thought to be the most important mechanism in addition to dissolution of thrombotic material behind the stent and cell necrosis.6 7 10 15 A greater EEM CSA for patients with ISA at follow-up than for those without ISA at follow-up could be confirmed in this study for SES and BMS patients.

In this study, ISA rates at follow-up were 25% for SES and 8.3% after BMS. The rates are high compared with previous reports for both the BMS and the SES group. ISA at follow-up has been reported in 4–5% of lesions after BMS and in 12–21% of lesions after DES implantation.6 7 9 10 16^–18 It is noteworthy, that the definition used for ISA included stents with only one strut being detached from the vessel wall as detected by IVUS. This definition might be very sensitive.

In this study, the rate of ISA at follow-up was significantly higher in the SES group than in the BMS group. This finding differs from those of previous reports. Previous studies on ISA after DES have either not included a comparison with a BMS group or demonstrated only a non-significantly higher rate of ISA after DES in comparison to SES.16 However, Degertekin et al have already indicated that ISA after DES is likely to be higher than after BMS.9 Hong et al have reported a rate of late-acquired ISA of 5.4% after BMS.6 Later the same authors reported a rate of late-acquired ISA of 12% after DES including SES and paclitaxel-eluting stents.17 The difference in ISA at follow-up may be explained by the difference in proliferative response. While positive vessel remodelling occurs after BMS as well as DES, the proliferative vascular response is suppressed after DES to fill the developing space.

Furthermore, the rate of ISA is likely to be a function of lesion complexity. Lesion length, reference vessel diameter and lesion calcification were univariate predictors of ISA. In this study, CCS III or IV angina pectoris at stent implantation and absence of diabetes were found to be independent predictors of ISA in SES group patients. Resolution of thrombus material during follow-up in patients with CCS III or IV angina pectoris at stent implantation and less intimal hyperplasia in non-diabetic patients may be reasons for more ISA in these patients. The finding is in agreement with previous analyses. Hong et al reported total stent length, primary stenting in acute myocardial infarction, and chronic total occlusion to be predictors of late-stent malapposition after DES implantation.17 Tanabe et al reported from the IVUS substudy of TAXUS II that lesion length, unstable angina, and absence of diabetes mellitus were predictors of late stent malapposition after paclitaxel-eluting stent implantation.10 In contrast to the SES group, no predictors of ISA could be defined for the BMS group. Furthermore, implantation of SES was found to be an independent predictor of ISA considering both SES and BMS group patients. This finding indicates that the different vascular response after SES implantation compared with BMS implantation has an impact on the occurrence of ISA.

Clinical significance of ISA determined by IVUS

The impact of ISA detected by IVUS on clinical events has been a matter of concern. A retrospective study evaluated the postprocedural IVUS findings of 53 patients who developed stent thrombosis. ISA was discussed as a potential cause of stent thrombosis for this study group.19 However, other studies of patients with BMS have indicated that ISA has no negative impact on clinical event rates.6 7 15 Hong et al reported no difference in MACE between patients with or without stent malapposition 6 months after stent implantation during a mean 3-year follow-up period after BMS implantation including a total of 881 patients.6 After introduction of DES and the description of late stent thrombosis, late-acquired ISA has again become a matter of concern. A recent pathology study indicated that ISA is a predictor of stent thrombosis.5

However, previous IVUS studies on DES have suggested that compared with non-ISA, ISA after DES implantation has no negative impact on clinical event rates.9 10 16^–18 Ako et al reported no negative clinical events for any of 13 ISA cases after SES during a 12-month clinical follow-up.16 In a study on 13 patients who received SES and showed ISA at angiographic follow-up, all patients remained asymptomatic during a subsequent 12-month follow-up period.9 Tanabe et al reported on 229 patients who received paclitaxel-eluting stents as part of the TAXUS II study.10 No difference in adverse clinical event rates was seen between patients with or without ISA immediately after stent placement or at 6-month angiographic follow-up during a 12-month clinical follow-up period. In a study on 82 patients with late stent malapposition by IVUS at 6-month follow-up, no major adverse cardiac events were reported during a mean 10-month follow-up after detection of ISA.17 However, these IVUS studies were limited to relatively short observation periods after stent placement. Thus, the observation period covered largely the period of dual antiplatelet therapy.

The present study covered a much longer follow-up period after SES implantation. The definition of ISA has been used in this study as proposed requiring at minimum one stent strut being separated from the vessel wall as demonstrated by IVUS. The study results indicate that ISA may be associated with a higher rate of myocardial infarction indicated by a trend for more myocardial infarction in the ISA group than in the non-ISA group. There was a trend for more frequent myocardial infarctions in SES and BMS patients with ISA than in non-ISA patients. However, most myocardial infarctions were not related to the target vessel. Furthermore, ISA detected by IVUS, considering the currently used sensitive IVUS definition for ISA, appears to have no significant negative impact on the total number of clinical events or on stent thrombosis during a 4-year follow-up period.

Limitations

Patients included into this analysis were part of three randomised studies. Inclusion criteria were different for the three studies, resulting in an inhomogeneous patient subset. However, the inclusion criteria may still not represent routine clinical practice. Treatment with clopidogrel was required for at least 2 or 3 months, according to the original trial protocols, but no particular information was available on the actual duration of clopidogrel application. The primary end point of the included studies was not to determine the incidence of ISA at angiographic follow-up or to define the impact of ISA on clinical outcome. Thus, no power calculation for detection of ISA or clinical end points related to the presence of ISA was performed. Given the follow-up period of 4 years and the 325 patients included into this study, it surveys a total of 1300 patient-years, which is more than in previous studies on the clinical impact of ISA. The study may still have insufficient power to detect differences in MACE between ISA and non-ISA group patients. However, these limitations frequently apply to currently available data even if pooling from several randomised studies is performed as has been stated recently.20

IVUS was performed only at follow-up in the RAVEL study. Thus, this study evaluated only the frequency of ISA at angiographic follow-up. A differentiation of persistent ISA and late-acquired ISA was not possible. However, the impact on late clinical events of ISA determined at follow-up may be similar irrespective of whether it is persistent or late acquired.

The conclusions generated from this study cannot be generalised to other DES with different polymer basis, drug dosing or elution profile.

CONCLUSIONS

This analysis indicates that ISA at follow-up is more frequent after implantation of SES than BMS. However, during a 4-year follow-up ISA detected by IVUS appears to be a finding without significant negative impact on the total frequency of MACE.