Author + information
- Received December 28, 2015
- Revision received February 3, 2016
- Accepted February 18, 2016
- Published online June 13, 2016.
- Jing Kan, MBBSa,
- Zhen Ge, MDa,
- Jun-Jie Zhang, PhDa,
- Zhi-Zhong Liu, PhDa,
- Nai-Liang Tian, MDa,
- Fei Ye, MDa,
- Sui-Ji Li, MDb,
- Xue-Song Qian, MDc,
- Song Yang, MDd,
- Meng-Xuan Chen, MBBSe,
- Tanveer Rab, MDf and
- Shao-Liang Chen, MDa,∗ ()
- aDivision of Cardiology, Nanjing First Hospital, Nanjing Medical University, Nanjing, China
- bDivision of Cardiology, Jintan People Hospital, Jintan, China
- cDivision of Cardiology, Zhangjiagang People Hospital, Jiangsu, China
- dDivision of Cardiology, Yixin People Hospital, Wuxi, China
- eDivision of Arts and Science, Emory College of Arts and Science, Emory University, Atlanta, Georgia
- fDivision of Cardiology, Department of Medicine, Emory University School of Medicine, Atlanta, Georgia
- ↵∗Reprint requests and correspondence:
Dr. Shao-Liang Chen, Nanjing First Hospital, Nanjing Medical University, Nanjing, China 210006.
Objectives The present study aimed to analyze the incidence of SF and its correlation with clinical events after DES implantation and the outcome of re-intervention for symptomatic in-stent restenosis (ISR) induced by stent fracture (SF).
Background SF is associated with a high rate of clinical events after the implantation of drug-eluting stents (DES). However, the chronological rate of SF and the effect of SF on clinical outcomes from a large patient population remain underreported.
Methods A total of 6,555 patients with 16482 DES in 10751 diseased vessels and surveillance angiography between November 2003 and January 2014 were prospectively studied. The primary endpoints included the incidence of SF, in-stent restenosis (ISR), target lesion revascularization (TLR), and definite stent thrombosis (ST) at the end of follow-up before and after propensity score matching. Clinical outcomes after TLR were also followed up.
Results The SF rate was detected in 803 (12.3%) patients, 3,630 (22.0%) stents, and 1,852 (17.2%) diseased vessels. SF increased over time. SF was associated with higher unadjusted rates of ISR (42.1%), TLR (24.8%, n = 379), and definite ST (4.6%) compared with stents without fracture (10.7%, 6.6%, and 1.03%, all p < 0.001), and the differences remained significant after propensity score matching (all p < 0.05). There was no significant difference in any-cause or cardiac mortality between patients with and without SF. After 1,523 days of follow-up since the first surveillance angiography, repeat ISR was detected in 90 of 379 (23.8%) stents after reintervention, and 6 (7.5%) stents required repeat TLR.
Conclusions SF is more frequently observed after DES implantation. TLR was required in almost one-fourth of fractured stents. Increased events in the SF group did not translate into a difference in mortality compared with the non-SF group. Reintervention was associated with acceptable clinical results.
The association of stent fracture (SF) with in-stent restenosis (ISR), stent thrombosis (ST), and subsequent target lesion revascularization (TLR) is of concern (1,2), particularly in the modern era of drug-eluting stents (DES). The incidence of SF varies from <1% to >8% (3) depending on the time point measured after the index procedure, DES types, fracture definitions, and imaging tools used for the analysis. Previous reports demonstrated that the majority of SFs occur within the first year after the stenting procedure (1–3). Globally, cobalt chromium-based second-generation DES, which feature a thin-strut and flexible design, have completely replaced first-generation DES, which have a stainless steel platform (1–3). However, the exact incidence of SF and its correlation with clinical events remain unclear. The present study aimed to investigate the incidence of SF after index stenting procedure, its correlation with clinical events, and clinical outcomes after repeat intervention for SF-related TLR.
Study population and exclusion criteria
The NFMD (Nanjing First Hospital Medical Data) database, established in November 2003, in Nanjing, China, is a prospective multicenter (4 centers), all-comers registry of patients who undergo percutaneous coronary intervention (PCI). By January 2014, this database included 10,077 PCIs (2,793 staged PCIs) for 8,602 patients. Routine follow-up angiography at 9 to 12 months post-procedure was encouraged for all patients. The current study complied with the Declaration of Helsinki regarding investigations in humans and was approved by the institutional ethics committees at the 4 participating centers. There was no industry involvement in the design, conduct, or analysis of the study. All of the study patients gave written informed consent for the procedure and the follow-up protocol.
For this analysis, the exclusion criteria (Figure 1) included: 1) bare-metal stent use (n = 186); and 2) poor quality of angiographic images (n = 12). A total of 6,555 patients with 10,751 diseased vessels (with 16,482 DES implanted) and repeat angiography (>30 days after stenting procedure) were included in this analysis.
All interventions were performed using standard techniques. Pre-dilation, post-dilation, and the use of intravascular ultrasound (IVUS) (Boston Scientific, Marlborough, Massachusetts) or optical coherence tomography (OCT) (S7, St. Jude Medical, St. Paul, Minnesota; since 2013) were decided on the basis of the operator's discretion. After the procedure, the patients were advised to continue on aspirin (100 mg daily) for life unless there were contraindications. Clopidogrel (75 mg daily) was also prescribed for at least 1 year after stent implantation.
Quantitative angiographic analysis
Coronary angiography was performed after the intracoronary administration of 0.2 mg nitroglycerin. Quantitative angiographic analysis was performed before and after stenting and during the follow-up angiography using a guide catheter to calibrate the magnification and a validated automated edge detection algorithm (CASS 5.7, Pie Medical Imaging, Maastricht, the Netherlands). The analyses were performed independently by 2 experienced observers who were blinded to the clinical information. The target lesion for measurement of the minimal luminal diameter included 5 mm margins proximal and distal to the stent and the stent itself. ISR was defined as a percent diameter stenosis of >50% within the stent at the time of follow-up within the stented segment or within 5 mm proximal or distal to the stent segment. The angiographic ISR patterns were classified I to IV according to Mehran's classification (4). A hinge motion lesion was defined as having a ≥16° difference in the angle between diastole and systole before the procedure.
Study endpoints and definitions
The study’s primary endpoints were the chronological incidence of SF, ISR, clinically driven target lesion revascularization (TLR), and definite/probable ST at the end of follow-up (January 2015). A clinically driven TLR was defined as treatment for recurrent angina pectoris before the scheduled follow-up angiography. The timing and diagnostic certainty of ST were assessed according to the Academic Research Consortium definition (6).
The angiographic diagnosis of SF required an independent analysis and the agreement of 2 independent cardiologists (J.K., J.-J.Z.). Angiographic SF using the stent boost technique was classified as types I∼IV according to Popma’s classification (5) as shown in Figure 1. For suspected SF during surveillance angiograms, IVUS or OCT was used. The prevalence of SF at the patient, vessel, and stent level was calculated.
The data are presented as mean ± SD or median (interquartile range) and percentages. Categorical variables were compared between groups using the chi-square test or Fisher exact test as appropriate. Continuous variables were compared between groups using Student unpaired t test or the Mann-Whitney U test on the basis of the distribution. Multivariable logistic regression analysis was used for the clinical variables, lesion, procedural, and DES platform-based analysis of the risk factors for SF. The survival curve at the patient level was estimated using the Kaplan–Meier method and was compared using log-rank analysis. For calculation of the SF rate, the patients were grouped by individual stents. The statistical analysis was performed using SPSS 17.0 (SPSS Inc., Chicago, Illinois).
Considering the significant differences in the baseline clinical and lesion characteristics between patients with and without SF, propensity score matching (PSM) was used to compare adverse events at the end of the clinical follow-up between SF-based groups to provide an unbiased estimation of treatment effects. All clinical and angiographic variables were included in the PSM analysis. The patients in the SF group who had an estimated logit within 0.5 standard error of the selected patients in the non-SF group were eligible for matching. If more than 1 patient in the non-SF group met this criterion, we randomly selected 1 patient for matching according to nearest rule. Estimates of the adjusted differences in risks are presented with 95% confidence intervals.
Comparison of baseline clinical variables at the patient level
Of the 6,555 patients, the 803 (12.3%) patients who experienced SF were more likely to have high levels of serum LDL, a more frequent family history of cardiovascular disease, a history of prior myocardial infarction (MI)/coronary artery bypass graft, and decreased left ventricular function (Table 1).
Vessel-level analysis of angiographic and procedural variables
More SFs were localized in the right coronary artery (RCA). Vessels with SF had more complex lesions, as defined by multiple-vessel disease, chronic total occlusion, and ostial and bifurcation lesions and by more severe tortuosity, angulation, and calcification. The treated vessels in the SF group had smaller reference vessel diameters and longer lesion lengths (Table 2).
Pre-dilation was performed in 73.0% of vessels that experienced SF, which was significantly higher than 59.7% in non-SF vessels (p < 0.001). Vessels with SF were more likely to have smaller and longer stents deployed at higher pressures (Table 3), leading to a lower stent/vessel ratio. Overlapping stents were observed in 81.6% of SF vessels compared with 34.4% of vessels in the non-SF group (p < 0.001). SF vessels were more frequently treated with post-dilation using larger noncompliant balloons (higher balloon/stent ratio) at higher inflation pressures, resulting in more edge dissections and loss of side branches (≥2.0 mm).
Incidence, patterns, and independent predictors of stent fracture
Of 10,751 diseased vessels and 16,482 implanted stents, the incidence of SF was 22.0% (n = 3,630, Figure 2) at the stent level and 17.2% (n = 1,852) at the vessel level. Variance of %-SF between 4 centers was <3.7%, which could be ignored. There was no significant difference in rate of SF either between male (17.4%) and female (16.8%; p = 0.517) or between acute MI (18.1%) and nonacute MI (16.8%; p = 0.122, data not included in Table 3). The patterns of SF were distributed nearly equally; pattern I/II was found in 46.6%, and pattern III/IV was found in 53.4%. The rate of SF within 1 year after a stenting procedure was 31.2%; this rate was higher than the rate at the 1- to 2-year follow-up (19.6%) but lower than the rate after the 2-year follow-up (49.1%; all p < 0.05).
Of 3,630 fractured stents, 2,963 cases of SF were diagnosed by angiography and the stent boost technique, and the remaining 667 SF could not be determined and were ultimately diagnosed by IVUS (n = 640) or OCT (n = 27).
On the multivariate analysis, stents in the RCA, stainless stents, stent length >25 mm, hinge motion, overlapping, stent/vessel ratio <0.8 (requiring post-dilation using a larger balloon at higher pressure), and multiple stents were the 7 independent predictors of SF (Table 4).
Clinical consequence of stent fracture
The angiographic follow-up duration was 340 days (SF group IQR: 36 to 33,905 days) and nonsignificant to 340 days (non-SF group range: from 35 to 3,395 days; p = 0.929) (Table 5).
Of 3,630 fractured stents in 803 patients, 1,528 (42.1%) SFs had ISR, with 5.1% having occlusion, 6.0% having aneurysm and 4.6% having definite ST, and TLR was required in 24.8% (n = 379); all of these rates were significantly different than the rates in stents without fracture (10.7%, 0.3%, 0.7%, 1.03%, and 6.6%, respectively; all p < 0.05) (Figure 3). Among fractured stents without ISR, the TLR rate after an additional median of 1,507 (276 to 3,502) days since the first surveillance angiography was 0.8%; this rate was not significantly different from the rate of 1.8% in asymptomatic ISR induced by SF.
In general, patients (n = 803) with any SF had higher rates of MI, TLR, target vessel revascularization (TVR), definite/probable ST, and target vessel MI leading to accumulative MACE and target lesion failure (TLF), rates 2 and 3 times as high as the rates in patients without SF (Table 5, Figure 4), respectively. However, the rates of any-cause or cardiac death in patients with SF did not differ from the rates in patients without SF. Most patients with SF were asymptomatic (Table 5).
By PSM, 684 pairs of patients were matched (Table 5). The results of the comparisons of MI, TLR, TVR, target vessel MI, and definite ST between patients with versus without SF were as follows: 4.5% versus 2.0% (p = 0.014), 22.2% versus 9.5% (p < 0.001), 23.4% versus 10.7% (p < 0.001), 3.5% versus 1.6% (p = 0.027), and 3.2% versus 1.3% (p = 0.020), respectively. In unmatched patients with SF group, the rate of cardiac death, MI, TLR, MACE, and TLF was 7.6%, 14.3%, 34.5%, 59.7%, and 41.2%, extremely higher than those in unmatched patients without SF (2.5%, 1.5%, 5.7%, 14.4%, and 8.1%, respectively; all p < 0.001) (derived from Table 5).
Clinical outcomes after treating SF-related symptomatic ISR
Of 379 fractured DES that required TLR in 193 patients, repeat DES implantation (for type III/IV ISR) was performed in 201 (53.3%) stents, and balloon angioplasty (for type I/II ISR) was performed in the remaining 178 (46.7%) stents; the rates were similar to the rates (45.2% and 54.8%) in 1,085 nonfractured stents that required TLR (non-SF-TLR group) in 356 patients. After 1,523 (375 to 3,650) days of follow-up, repeat ISR was detected in 23.8% (n = 90) stents, and 7.5% (n = 7) required repeat TLR; these results were not significantly different compared with the results in the non-SF-TLR group (22.5% and 6.9%, respectively).
The major findings of the present study on the basis of a larger patient population who underwent DES implantation are as follows: 1) the prevalence of SF at the stent level was 22.0%, and the prevalence increased over the 2-year follow-up; 2) given 6 independent predictors of SF (stents in the RCA, stainless stent, stent length >25 mm, hinge motion, overlapping, and multiple stents), stent/vessel ratio <0.8 (requiring post-dilation using a larger balloon at higher pressure) were strongly correlated with SF; 3) SF was associated with significantly increased angiographic (ISR, occlusion, and aneurysm formation) and clinical (TLR, MI, ST, MACE, and TLF) events after propensity score matching, which did not translate into significant differences in mortality among the 2 groups; and 4) balloon angioplasty for type I/II ISR and repeat implantation of DES for type III/IV ISR induced by SF had acceptable clinical results.
The reported incidence of SF is dependent on asymptomatic SF, different definitions and modalities for SF, different percentages of surveillance angiography, and longer follow-up durations, resulting in much higher SF rates by autopsy than are clinically reported (6,7). Theoretically, all pro-fracture factors facilitate the process of metal fatigue over time (8–12); metal fatigue is a mechanism leading to the postulation that more frequent SF might be detected over 1-year follow-up, which is supported by our finding that the SF rate after 2 years of follow-up reached 49.1%. Furthermore, we found that patterns of SF were almost equally distributed, and pattern II/III/IV SF more frequently lead to symptomatic ISR.
Previous studies (13–18) have reported several angiographic/procedural and clinical variables powerfully predicting the occurrence of SF. In general, RCA stents, stainless stents, longer stents or multiple stents, and overlapping stents were universally accepted to be correlated with SF (13–17). RCA stents, particularly in the proximal-to-middle segment of the RCA, were most commonly exposed to severe cardiac motion and angulation, a mechanical mechanism causing stent fracture (9,10). Once longer or multiple or overlapping stents were used, aggressive post-dilation was usually performed to achieve optimal angiographic results (11–18), which was consistent with our findings that patients with SF had more frequent multivessel disease requiring multiple and longer stents followed by post-dilation at a higher pressure. SF does not happen equally among different stents, and more frequent SF was observed with stainless steel stents (7) compared with cobalt-chromium platform stents (13), indicating less durability with stainless stents, which should partially explain why cobalt-chromium stents are associated with improvement of clinical outcomes (13,15–17). A recent study by Kuramitsu et al. (18) introduced the concept of hinge motion, which together with overlapping stent and tortuosity could predict the occurrence of SF; that concept is in line with our results. The severe angulation, tortuosity, and hinge motion reflect abnormal movement during the cardiac cycle, which puts stents under mechanical comprise. The present study also found that small stents (stent/vessel ratio <0.8) were not correlated with SF. However, the combination of a small stent and over post-dilation at higher pressure could predict the occurrence of SF; this finding is similar but not exactly the same as the finding of a previous study (17) and suggests the importance of the selection of a stent with an appropriate diameter.
Previous studies confirmed the correlation of SF with ISR and subsequent events (7,15), in line with our findings. The majority of SFs were not silent. TLR was required in 24.8% of stents with SF; these results are similar to the results of studies by Lee et al. (19) and Chhatriwalla et al. (8). It has been suggested that SF causes uneven and impaired local drug delivery at the stent site. However, in our analysis, SF mostly occurred long after drug delivery had been terminated, indicating that ISR might have been caused by mechanical damage to a vessel wall rather than by the loss of drugs; this postulation is supported by the results of Halkin et al. (20). In our study, aneurysm was more common in patients with type III/IV SF, which is in agreement with a previous study that reported that aneurysms were mostly observed in patients with complete SF. We found that asymptomatic ISR caused by SF had a lower rate of TLR, demonstrating the importance of severe chronological mechanical damage to vessels in inducing narrow lumens in symptomatic patients. Because the baseline clinical and angiographic characteristics varied widely, propensity score matching was used in our analysis; in this analysis, SF remained associated with an increased rate of primary and secondary endpoints. From our results, the higher rate of MI, TLR, and ST in the SF group did not translate into significant differences in any-cause or cardiac mortality compared with the non-SF group. Possible explanations for this finding might include the following: a) the small patient sample size underpowered the significance in death; and b) the percentage of dual antiplatelet therapy was relatively high, in accordance with the recommendation that for patients with SF and asymptomatic ISR, extending dual antiplatelet therapy is recommended and appears safe (19).
Reintervention was associated with a significant reduction of MACE for patients with symptomatic SF-induced ISR (21), and this result was confirmed by this study. The recommendation for reintervention for SF-induced ISR was that new generation DES (cobalt chromium platform) should be selected as the first line option; Ito et al. (22) and Ohya et al. (12) reported significant increases in MACE, TLR, ST, and MI after implantation of a Cypher stent (Cordis, Johnson & Johnson, Fermont, California) for symptomatic SF-induced ISR. A few studies have compared the effect of restenting and ballooning with the recurrence of ISR (7). Our study showed acceptable results after balloon angioplasty for focal ISR; these results were in agreement with reports by Mitomo et al. (23). In that study, 50.4% of cases were treated with ballooning.
Bare-metal stents and poor quality of images were excluded from the current analysis, and this exclusion could result in underestimations of the SF rate. However, the high rate of surveillance angiography in this study might demonstrate the actual incidence of SF. Another limitation was the only use of Popma’s definitions of SF, which might report the incidence of SF differently compared with other definitions, and which may result in a lower rate of SF than the true incidence, even IVUS and OCT were used for some cases suspected to have SF. Finally, only a subset of patients was included (always true in matching) and the potential influence of unmeasured confounders could not be excluded.
WHAT IS KNOWN? SF is associated with increased clinical events. However, the actual incidence of SF varies from studies on the basis of different definitions.
WHAT IS NEW? We provide the actual rate of SF at patient and vessel level after implantation of a DES from a large patient population. The rate of cardiac death, myocardial infarction, revascularization, and stent thrombosis in SF patients was still significantly different to the rate in non-SF patients after propensity score matching.
WHAT IS NEXT? Further studies on mechanical movement of heart and vascular fluid dynamic stress are required to elucidate the mechanisms of SF.
Given the factor that stent fracture is associated with a higher rate of clinical events and that a stainless stent is 1 of the independent factors of stent fracture, more mechanical durable DES (cobalt chromium platform stent) should replace first generation, stainless platform stents. As stent fracture is more frequently detected in anatomically complex lesions (angulation, tortuousity, hinge motion), better understanding of abnormal distribution of the shear stress (24) before and post-stenting underscores the improvement of stenting techniques, which should have been analyzed in further studies.
Dr. Chen now is the fellow of the Collaborative Innovation Center for Cardiovascular Disease Translational Medicine and Clinical Medical Research Center of Jiangsu Province, China. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- drug-eluting stent(s)
- in-stent restenosis
- intravascular ultrasound
- major adverse cardiovascular event(s)
- myocardial infarction
- optical coherence tomography
- percutaneous coronary intervention
- propensity score matching
- quantitative coronary analysis
- right coronary artery
- stent fracture
- stent thrombosis
- target lesion revascularization
- Received December 28, 2015.
- Revision received February 3, 2016.
- Accepted February 18, 2016.
- American College of Cardiology Foundation
- Mehran R.,
- Dangas G.,
- Abizaid A.S.,
- et al.
- Nakazawa G.,
- Finn A.V.,
- Vorpahl M.,
- et al.
- Ohya M.,
- Kadota K.,
- Tada T.,
- et al.
- Schömig A.,
- Dibra A.,
- Windecker S.,
- et al.
- Kandzari D.E.,
- Rao S.V.,
- Moses J.W.,
- et al.
- Kuramitsu S.,
- Hiromasa T.,
- Enomoto S.,
- et al.
- Halkin A.,
- Carlier S.,
- Leon M.B.
- Molony D.,
- Timmins L.H.,
- Huang O.,
- et al.