Risk of Late-Acquired Incomplete Stent Apposition after Drug-Eluting Stent versus Bare-Metal Stent
A Meta-Analysis from 12 Randomized Trials
ABSTRACT: Background. Incomplete stent apposition (ISA) is an unusual finding of intravascular ultrasound (IVUS) that may occur both after drug-eluting stent (DES) or bare-metal stent (BMS) implantation and could be associated with late stent thrombosis. Controversy still remains about whether the risk of late-acquired ISA is increased after DES implantation. This meta-analysis aimed to clarify whether DES implantation is associated with an increased risk of late-acquired ISA. Methods. We performed a meta-analysis from 12 randomized trials that compared DES and BMS and included IVUS follow up: TAXUS II (n = 469), TAXUS IV (n = 187), TAXUS V (n = 213) and VI (n = 147), ASPECT (n = 81), DELIVER (n = 65), SIRIUS (n = 141), DIABETES (n = 140), ENDEAVOR II (n = 250), FUTURE I and II (n = 83), and SPIRIT-I (n = 58). In these trials, 1,834 patients (972 DES, and 862 BMS) underwent immediate and follow-up IVUS examination. Results. There was no heterogeneity among the trials (Q-test for heterogeneity: Chi2: 7.69; (p = 0.26), I2: 22%. Out of the 1,834 patients undergoing serial IVUS examination, 85 developed late-acquired ISA (4.6%). This incidence was significantly higher in DES compared with BMS (6.5% vs. 2.6%, respectively; odds ratio [OR] 2.48, 95% confidence interval [CI] 1.26 to 4.87; p = 0.008). That means that the risk of developing late-acquired ISA is 2.5 times higher after DES versus BMS implantation. No stent thrombosis occurred in the patients diagnosed with ISA over a period up to 12 months. Conclusion. DES implantation could be associated with an increased risk of late ISA in comparison with BMS. The clinical implication of late ISA in the long term remains to be clarified.
J INVASIVE CARDIOL 2008;20:417–422
Drug-eluting stents (DES) have demonstrated the capacity to reduce binary angiographic restenosis, and the need for subsequent revascularization procedures of the target vessel.1 Serial intravascular ultrasound (IVUS) studies have demonstrated that DES dramatically reduce the degree of neointimal hyperplasia compared to bare metal stent (BMS).2,3 DES do not increase the risk of stent thrombosis in comparison with BMS during the first year after stent implantation, at least under prolonged double antiplatelet therapy.4 Conversely, long-term follow up of randomized trials have shown that DES seem to increase the risk of stent thrombosis more than 1 year after implantation (very late stent thrombosis),5,6 but the reasons have not been elucidated yet.7,8
IVUS is of help in the evaluation of changes of vessel wall that occur after percutaneous coronary interventions. Late-acquired incomplete stent apposition (ISA) is a relatively unusual IVUS finding that may occur both after DES and BMS implantation, and could be associated with late stent thrombosis. Intravascular brachytherapy has an antirestenotic effect by inhibiting the proliferation of vascular smooth muscle cells similar to DES, and its use was associated with a high incidence of late thrombosis.9 Late-acquired ISA was proposed as a mechanism for late thrombosis after brachytherapy. Controversy still remains about whether the risk of late-acquired ISA is increased after DES implantation. In patients from two nonrandomized studies in “the real-world” practice, the reported incidence of late-acquired ISA was similar after DES and BMS, approximately 5%.10,11 The purpose of the present study was to ascertain whether DES are associated with an increased risk of late-acquired ISA. For this purpose, we have performed a meta-analysis from twelve randomized, controlled trials that compared DES and BMS, and included one IVUS evaluation.
Methods
Trials included in the meta-analysis. We searched Medline to identify all randomized clinical trials that compared DES with BMS and included IVUS follow-up in the entire population or as an IVUS sub-study. We additionally reviewed abstract supplements of major scientific meetings of the European Society of Cardiology, American College of Cardiology, American Heart Association, Transcatheter Cardiovascular Therapeutics, and Paris Coronary Revascularization Course until May 2007. The search was performed using the keywords “elute” and “stent”. We first identified 1,951 reports, but after excluding non-english, as well as non-human and non-randomized studies, 127 articles were reviewed. From this sample, only 85 were studies or substudies of randomized trials that compared DES with BMS. We included only trials with IVUS substudies, which evaluated paclitaxel-, sirolimus-, zotarolimus-, and everolimus-eluting stents. Finally, we identified 12 randomized trials with IVUS findings that provided data on late ISA phenomenon. The Taxus® stent (Boston Scientific Corp., Natick, Massachusetts) was investigated in the TAXUS-II,12 TAXUS-IV,13 TAXUS-V and TAXUS-VI studies.14–17 DES with non-polymeric release of paclitaxel were evaluated in ASPECT (Supra-G Stent, Cook Inc., West Lafayette, Indiana)18 and DELIVER (ACHIEVE placitaxel-coated stent, Guidant Corp. and Cook’s proprietary paclitaxel nonpolymeric coating process) trials.19,20 The sirolimus-eluting Cypher™ stent (Cordis Corp., Miami Lakes, Florida) was investigated in the RAVEL, SIRIUS, and DIABETES.21–23 The zotarolimus-eluting Endeavor stent (Medtronic, Inc., Santa Rosa, California) was evaluated in the ENDEAVOR-II trial.24,25 Finally, the everolimus-eluting stent (currently the Xience V stent; Guidant Corp.) was studied in the FUTURE-I, FUTURE-II, and SPIRIT-I trials.26,27 Overall, 1,834 patients were included in these trials (972 allocated to DES and 862 to BMS).
Definitions and statistical analysis. ISA was defined in most trials as a ≥ 1 strut clearly separated from the vessel wall, with evidence of blood speckling behind the stent struts without overlapping side branches. ISA was usually classified in most trials into 3 categories: 1) resolved: ISA present at baseline but no longer present at follow up; 2) persistent: ISA present both after the procedure and follow up; and 3) late-acquired: ISA not present after the index procedure but present at follow up.
The review was conducted according to the Quality of Reports of Meta-Analyses of Randomized Clinical Trials (QUOROM) recommendations.28 Odds ratios (OR) for ISA and 95% confidence intervals (CI) were calculated by comparing DES with BMS rates using raw data for each study and for the pooled population (intention-to-treat basis). Heterogeneity among studies was assessed by the Q-test and the level of inconsistency was also assessed by the I2 test. Review Manager 4.2.1 (2006 Cochrane Collaboration, Oxford, United Kingdom) was used to perform the analysis. This program calculates both heterogeneity tests, Q and I2, based on the assumption that the second one gives a better measure of the degree of inconsistency between trials, especially because it is not susceptible to the number of trials included in the meta-analyses and also to the low statistical power of Q. This statistic describes the percentage of total variation across studies due to heterogeneity rather than chance. I2 can be readily calculated from basic results obtained from a typical meta-analysis as I2 = 100% x (Q - df)/Q, where Q is Cochran’s heterogeneity statistic and df the degrees of freedom. Negative values of I2 are put equal to zero so that I2 lies between 0% and 100%. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity. A value greater than 50% may be considered substantial heterogeneity.29,30 Although heterogeneity was not found, several sources of heterogeneity were anticipated a priori, therefore, the Der Simonian and Laird random effects model was used to estimate summary measures and their 95% CI, making an adjustment to the study weights according to the extent of variation based on the inverse variance approach. The effect of each study was weighted for its number of patients. The correlation test of Begg and the regression-based test of Egger were used to assess the possible existence of publication bias. Null hypothesis was rejected by a type I error less than 0.05 (a< 0.05).
Results
Characteristics of the trials included. Table 1 shows the main clinical and angiographic baseline characteristics of the trials included in the meta-analysis. No publication bias was detected (Begg’s test: p = 0.453; Egger’s test: p = 0.619).
Incidence of late-acquired incomplete stent apposition. The results showed no heterogeneity among the trials (Q-test for heterogeneity: Chi2: 7.69; (p = 0.26), I2: 22%). All the 1,834 patients included underwent an IVUS follow up at 6–12 months (972 allocated to DES and and 862 allocated to BMS). Out of these, 85 (4.6%) developed late-acquired ISA. The incidence of late-acquired ISA was significantly higher in patients allocated to DES (6.5% vs. 2.6% in those allocated to BMS; OR 2.48, 95% CI 1.26 to 4.87; p = 0.008) (Figure 1). That means that the risk of developing late-acquired ISA was 2.5 times higher after DES versus BMS implantation. No cases of late-acquired ISA were reported with either everolimus-eluting stents (FUTURE I-II and SPIRIT-I trials), or with zotarolimus-eluting stents (ENDEAVOR-II trial). After excluding those studies evaluating everolimus- and zotarolimus-eluting stents, the incidence of late-acquired ISA was 8.3% with DES and 3.2% with BMS. The OR for late-acquired ISA was 17.2 (95% CI 2.3 to 130; p = 0.006) and 1.96 (95% CI 1.14 to 3.35; p = 0.01) in sirolimus and paclitaxel trials, respectively (Figure 2). The increased risk of late-acquired ISA with paclitaxel stents occurred mainly with moderate-release paclitaxel-eluting stents (OR: 2.45 [95% CI 1.08 to 5.56; p = 0.03]), whereas no significant increase in the incidence of late-acquired ISA was observed with slow-release paclitaxel-eluting stents (OR: 1.57 [95% CI 0.8 to 3.1; p = 0.19]) (Figure 3). Finally, after excluding the DIABETES study, that exclusively included diabetic patients, the incidence of late-acquired ISA was also increased in patients allocated to DES (OR: 2.08 [95% CI 1.23 to 3.53; p = 0.006]).
The rate of post-procedural ISA and resolved ISA did not differ between DES and BMS (Figure 4), but the rate of ISA present at follow up (persistent plus late-acquired ISA) was significantly higher in patients allocated to DES (13.6% vs. 6.7%, respectively; OR 2.29, 95% CI 1.35 to 3.91; p = 0.002) (Figure 5).
Finally, no stent thrombosis occurred in the patients diagnosed with ISA over an average period of 12 months.
Discussion
The reported incidence of late-acquired ISA after BMS in nonrandomized studies was 4–5%,10,31 which is not very different from nonrandomized studies with DES (5.1–12%).11,32 The results of randomized trials have also been inconclusive. The main finding of our study is that DES are associated with a higher risk (2.5 times higher) of late-acquired ISA in comparison with BMS.
In the present meta-analysis from 12 randomized trials with IVUS data, the incidence of late-acquired ISA was different depending on the type of DES. No late-acquired ISA was present in patients treated with everolimus- and zotarolimus-eluting stents. Stents with non-polymeric-release of paclitaxel showed a very low rate of late-acquired ISA (only 1 case with a high dose of paclitaxel in the ASPECT trial). Our study also showed that late-acquired ISA increased with moderate-release paclitaxel stents, but not with the commercially available slow-release paclitaxel-eluting stent. These findings are consistent with the recently published meta-analysis by Weissman et al14 from the TAXUS IV, V and VI trials in which sirolimus-eluting stents also had a significantly higher incidence of late-acquired ISA in comparison with BMS. Hong et al32 also reported a higher, but not statistically significant, incidence with sirolimus compared to paclitaxel-eluting stents in a nonrandomized study (13.2% vs. 8.4%; p = 0.12).
Late-acquired ISA after BMS usually occurs at stent edges, and the main mechanism involved is a positive remodeling, with greater increase in the external elastic membrane (EEM) than in persistent plaque and media (P&M).10,33 However, Miyazawa et al34 have recently shown that the main factor for late-acquired ISA after DES and intracoronary radiation was positive remodeling, whereas plaque regression was the predominant mechanism for ISA after BMS. A possible explanation is that nearly 70% of patients underwent directional atherectomy before BMS implantation, and such a debulking procedure is known to induce thrombus formation, which could be reabsorbed at follow up. The predictors of late-acquired ISA after BMS were primary stenting in acute myocardial infarction and directional coronary atherectomy before stenting.10 As with BMS, the main mechanism of late-acquired ISA after DES implantation is also positive remodeling.11 Hong et al32 compared lesions with late ISA treated with sirolimus- and paclitaxel-eluting stents. The increase in the EEM area was greater than the increase in the P&M area with both types of DES, although there was a smaller increase in the P&M area and intimal hyperplasia with sirolimus-eluting stents. These findings suggest that the main mechanism of late ISA for both DES could be considered positive remodeling. In the SIRIUS trial, Ako et al22 identified a predominant location of late-acquired ISA in the body of the sirolimus-eluting stents, whereas persistent ISA occurred mostly at the stent edges. This pattern was also found by Siqueira et al.11 A possible explanation is that whereas persistent ISA is related to technical aspects during the procedure, late-acquired ISA is commonly related to regional remodeling.11,33 Predictors of late-acquired ISA after DES were total stent length, primary stenting in acute myocardial infarction and chronic total occlusions.14,32
The clinical implications of ISA remain unclear. In the present meta-analysis, late-acquired ISA was not associated with stent thrombosis over a period of 12 months, and none of the randomized studies reported late stent thrombosis in any patient with late ISA. The same findings were described by Hong et al32 in an observational and retrospective study involving 881 patients (54 of them with late-acquired ISA) after a 10-month follow-up period. Hoffmann et al,35 from a pooled analysis of RAVEL, SIRIUS and E-SIRIUS trials, showed no significant differences in clinical event rates, including stent thrombosis, between patients treated with sirolimus-eluting stents with ISA versus those without ISA at 1 and 4 years. Conversely, in the study by Siqueira et al11 involving 195 patients (10 of them with late-acquired ISA) and longer follow up (median 24 months), 2 cases of very late stent thrombosis were reported. This association between late ISA and very late stent thrombosis is supported by Cook et al36 who identified late ISA in 10 of 13 patients who presented with very late stent thrombosis, demonstrating a very high prevalence of late ISA in patients with very late stent thrombosis.
From a pathological point of view, Virmani et al37 first described a patient who died of very late thrombosis after receiving 2 sirolimus-eluting stents and had experienced a localized hypersensitivity reaction and extensive vasculitis of the intima, media and adventitia consisting predominantly of lymphocytes and eosinophils. Necropsy showed aneurysmal dilation of the stented arterial segments with evidence of stent malapposition and thick fibrin thrombus between the stent and the arterial wall. Joner et al38 compared 23 DES cases (> 30 days after the index procedure) with 25 matched necropsies after BMS implantation. Fourteen of 23 DES patients who suffered late stent thrombosis showed greater delayed healing characterized by poorer endothelialization and persistent fibrin deposition compared with patients who had both patent DES and BMS. In all 14 patients with late stent thrombosis, delayed arterial healing was found as a key risk factor, but 2 patients presented with malapposition related to positive arterial remodeling or suboptimal stent deployment as additional pathologic risk factors. Future studies and longer clinical follow up of these patients are warranted to clarify the role of late ISA in the physiopathology of very late stent thrombosis.
A limitation of this work is that the diagnostic accuracy of IVUS in detecting ISA is limited, although it is assumed that IVUS detects significant ISA, missing only minor cases.39 Another limitation is that due to the follow up of the trials included (up to 12 months), the effect of late ISA over very late stent thrombosis has not been completely evaluated.