Background A potentially adverse vascular response to overlapping drug eluting stents (DES) has been suggested in current research.
Objective To evaluate the impact of baseline disease severity at the site of stent overlap.
Methods and results This is a substudy of ODESSA, a prospective, randomised controlled trial designed to evaluate healing of overlapping stents. 71/77 patients with a total of 86 overlapping stents were studied: 25 sirolimus, 24 paclitaxel, 26 zotarolimus-eluting stents; and 11 bare metal stents (BMS). Patients were categorised into high-grade stenosis (HGS, ≥70% diameter stenosis) and low-grade stenosis (LGS, <70%) at the site of stent overlap. Angiography and intravascular ultrasound were performed after stent deployment and repeated at 6 months, together with additional optical coherence tomography. Images were analysed by an independent core laboratory. End points were binary restenosis, percentage neointimal hyperplasia (%NIH), mean lumen and stent areas and degree of strut coverage/apposition at overlapping stents at 6 months. Stent overlaps occurred in 49 HGS and 37 LGS. Restenosis was found in 5/6 HGS versus 0/5 LGS treated with overlapping BMS (p=0.01) and 4/43 HGS versus 0/32 LGS treated with overlapping DES. There was a trend towards higher %NIH at BMS overlap in HGS versus LGS (p=0.07). DES overlaps had lower lumen and stent areas and similar %NIH in HGS versus LGS. Any uncovered or malapposed struts occurred more often in overlapping DES at LGS than at HGS (59.4% vs 32.6%, p=0.03).
Conclusions Overlapping DES in normal-appearing coronary segments showed a higher incidence of uncovered or malapposed struts, while restenosis occurred exclusively in overlapping stents at HGS. These findings should be considered when deploying overlapping stents.
- Optical coherence tomography
- overlapping stents
- coronary stent
- Coronary stenting
Statistics from Altmetric.com
Overlapping stent deployment is a routine part of clinical practice when treating diffuse severely stenotic segments, tandem lesions, or post-stent dissections or residual stenosis due to misplacement of the initial stent. Hence, the dual layers of struts may be located in segments with high-grade stenosis (HGS) or low-grade stenosis (LGS). Overlapping bare metal stents (BMS) have shown higher rates of repeat revascularisation1 in a pooled analysis of retrospective studies, but no prospective investigation has been performed to date. Preclinical data have suggested a potentially adverse vascular response to drug-eluting stents (DES) at the site of overlap, impairing arterial healing and promoting inflammation as compared with overlapping BMS.2 However, the impact of the underlying disease on outcomes of overlapping stents is unknown and decisions on the most appropriate location at which to overlap two stents remain mostly empirical.
Therefore, the aim of this study was to evaluate the 6-month vascular response to overlapping DES and BMS at locations of HGS and LGS using standard coronary angiography, intravascular ultrasound (IVUS) and optical coherence tomography (OCT). The high spatial resolution of OCT (up to 10 μm) enables detailed in vivo strut-level analysis of tissue coverage.3–5
ODESSA was a single-centre, prospective, randomised controlled trial designed to evaluate stent coverage of overlapping sites of DES and BMS in human coronary arteries using OCT.6 The study protocol was approved by the Ospedali Riuniti di Bergamo Ethics Committee, the study was registered at http://clinicaltrials.gov/ (identifier NCT00693030) and all patients provided written informed consent.
A total of 77 consecutive eligible patients with long coronary lesions were randomised in a 2:2:2:1 ratio to stent implantation with sirolimus- (SES, Cypher, Cordis, Miami, Florida, USA), paclitaxel- (PES, Taxus Libertè, Boston Scientific, Natick, Massachusetts, USA) and zotarolimus-eluting stents (ZES, Endeavour, Medtronic, Santa Rosa, California, USA) or BMS (Libertè, Boston Scientific), respectively. Only a single type of stent was allowed in each patient. Eligible patients were older than 18 years, presented with stable or unstable coronary syndromes and had angiographic evidence of long stenosis (>20 mm by visual estimate) in native coronary arteries, with 2.5–3.5 mm reference vessel diameter, requiring percutaneous coronary intervention (PCI) with deployment of overlapping stents. Exclusion criteria were left main coronary artery disease, active or recent (<72 h from symptom onset) myocardial infarction, previous stenting in the target vessel, left ventricular ejection fraction ≤30%, serum creatinine >2.5 mg/dl, no suitable anatomy for OCT (ostial lesions and extreme tortuosity) and inability to comply with dual antiplatelet therapy and follow-up requirements. PCI was performed by IVUS guidance and stent size selection for stents was done by angiographic assessment. Follow-up angiography, IVUS and OCT were performed at 6±0.5 months.
Quantitative coronary angiography
Coronary angiography was performed at baseline, immediately after PCI and at follow-up in at least two orthogonal views after 200 μg intracoronary glyceryl trinitrate injection. Digital coronary angiograms were analysed offline at the Cardiovascular Imaging Core Laboratory, University Hospitals Case Medical Centre, using a validated automated edge detection system (CAAS II, PIE Medical, Maastricht, The Netherlands) and previously reported methodology.7 The site of stents overlap was determined and the diameter of stenosis (DS) was calculated at the same matched site at the pre-procedural angiogram. The population was stratified into two groups: ≥70%DS stenosis (HGS) or <70%DS (LGS) at the overlap site.
IVUS imaging was performed after intracoronary injection of 200 μg glyceryl trinitrate after implantation of stents and at the 6-month follow-up using the Atlantis SR Pro 40 MHz catheter and the iLab ultrasound console (Boston Scientific). Imaging was performed to include the stents and at least 5 mm segments proximal and distal to the stent using a motorised pullback at 0.5 mm/s. All IVUS data were digitally stored for subsequent independent core laboratory analysis. Quantitative volumetric IVUS analysis was performed using a validated semiautomated detection algorithm (Curad, version 4.32, Wijk bij Duurstede, The Netherlands) in a previously described methodology.8 Lumen and stent areas and volumes were semiautomatically measured and neointimal hyperplasia (NIH) calculated as the difference between stent volume and lumen volume.
Optical coherence tomography
OCT imaging was conducted at a 6-month follow-up and performed after a 200 μg intracoronary glyceryl trinitrate injection. A time domain OCT system (M2CV OCT Imaging System, LightLab Imaging, Westford, Massachusetts, USA) was used. The occlusive technique was used to completely flush blood from the artery.5 Briefly, an over-the-wire occlusion balloon (Helios Goodman, Advantec Vascular, Sunnyvale, California, USA) compatible with a 6 Fr guiding catheter (0.071″ inner diameter) was advanced distal to the stented segment over a conventional angioplasty guide wire. The guide wire was then replaced by the 0.019″ OCT Image wire (ImageWire, LightLab Imaging). The occlusion balloon was repositioned proximal to the stented segment and inflated at low pressure (0.4–0.7 atm) with simultaneous infusion of Ringer's solution at 37°C with a flow rate of 0.5–1 ml/s through the distal tip of the catheter. Images were acquired with an automated pullback at a rate of 1.0 mm/s, generating 15.6 frames per second. While multiple pullbacks were allowed for suboptimal images, final analysis was performed only on the most optimal continuous pullback. Images were digitally stored and submitted to the core laboratory for offline analysis. Measurements of OCT cross-sectional images were performed using a dedicated semiautomated contour-detection system (OCT system software B.0.1, LightLab) developed in collaboration with the core laboratory. OCT images were analysed by two independent observers. Area measurements were performed every five frames (ie, every 0.33 mm). Lumen, stent and NIH areas were calculated in a similar fashion to the IVUS methodology described previously.
OCT strut-level analysis was performed in every frame (ie, every 0.06 mm) and three categories of struts were classified based on quantitative strut assessment: covered, uncovered, malapposition. A covered strut was defined as a strut totally embedded (figure 1A). An uncovered strut was a strut partially embedded, without overlaying tissue and with the distance from strut surface to lumen contour <100% of the strut plus polymer thickness (figure 1B). Malapposition was defined as any strut with a strut-surface to lumen distance greater than the stent strut/polymer thickness plus blooming correction (figure 1C).
Study end points
The main study end point was the 6-month presence of uncovered and malapposed struts measured by OCT at the site of overlapping stents. Secondary end points were binary restenosis, which is defined as >50% DS by quantitative coronary angiography, and mean lumen area, mean stent area and %NIH by IVUS and OCT and malapposition evaluated by IVUS at the overlap segment.
All analyses were conducted with SAS, version 8.2 (SAS Institute). A p value <0.05 was considered significant. Continuous variables were expressed as mean±SD and categorical variables were expressed as frequencies. The p values were calculated using a t test if the variable was continuous. A χ2 test was used if the variable was categorical and Fisher's exact test was used when the sample sizes requirements for the χ2 tests were not met by the data.
Among 77 patients randomised, 76 patients were eligible for analysis. One patient was excluded owing to coronary perforation, which necessitated a covered stent implantation. A complete series of angiograms was available for 75 patients. In four patients, no overlapping segment was identified by OCT or IVUS. Ultimately, we analysed 86 overlapping stents (25 SES, 24 PES, 26 ZES and 11 BMS) from 71 patients. Overlap sites were located in 49 coronary segments with HGS and 37 with LGS at the baseline stent implantation. As shown in table 1, initial clinical characteristics were similar in both groups, except for a slightly lower left ventricular ejection fraction (HGS group 53.9% vs LGS group 50.8%, p=0.009). There were no differences between groups in mean stent diameters (HGS group 3.2±0.3 mm vs LGS group 3.0±0.3 mm, p=NS), total stent lengths (48.5±14.3 mm vs 50.7±15.7 mm, p=NS), baseline angiographic diameter stenosis (77.5±16.4% vs 71.7±17.8%, p=NS) and reference vessel diameter (2.7±0.4 mm vs 2.8±0.5 mm, p=NS). Lengths of the overlapping segments were similar in IVUS (HGS 3.5±2.1 mm vs LGS 3.4±1.8 mm, p=NS) and OCT (3.5±2.1 mm vs 3.4±1.8 mm, p=NS). Proximal stent size was larger than the distal stent in all overlapping segments and the relative sizes of the stents in the overlapping segments were slightly larger in HGS (1.12±0.10 vs LGS 1.07±0.09, p=0.03).
Procedural characteristics at overlapping segments, such as predilatation (HGS 78% vs LGS 70%, p=NS), balloon to artery ratio (HGS 0.86±0.15 vs LGS 0.91±0.14, p=NS), stent (smaller stent) to artery ratio (HGS 1.01±0.11 vs LGS 1.03±0.16, p=NS), maximal pressure for second stent implantation (HGS 18.2±2.3 atm vs LGS 18.1±2.9 atm, p=NS), postdilatation (HGS 79.6% vs LGS 73.0%, p=NS) and postdilatation balloon diameter (HGS 3.1±0.3 mm vs LGS 3.1±0.4 mm, p=NS) were similar between two groups. Postdilatation maximal pressure was slightly higher in HGS than in LGS (19.0±2.5 atm vs 17.4±2.0 atm, p=0.04).
Angiographic restenosis at the site of overlap occurred in 5/6 (83.3%) in the HGS group treated with BMS and 4/43(9.3%) had restenosis in the HGS treated with DES (table 2). There were no restenoses in the LGS group treated with either DES or BMS. There was a trend towards smaller stent areas in DES overlapping HGS versus LGS (table 3), which probably contributed to the angiographic restenosis and reduced lumen area (table 2). OCT detected more NIH in HGS than in LGS for DES, albeit there were no statistical differences (table 4). The percentage NIH was similar between groups. Stent area was similar between BMS overlapping HGS versus LGS (table 3). There was a trend towards higher %NIH at BMS overlap site in HGS versus LGS group (table 4).
Comparison of the IVUS study at post-stent implantation and follow-up showed that there was no persistent malapposition at overlapping segments, but four segments with acquired malapposition. They were all SES overlapping stents, one of which was implanted in HGS and the others were implanted in the LGS (p=NS, HGS vs LGS). In one SES overlapping segment in HGS, malapposition was exclusively seen in the overlapping segment, not in consequent proximal and distal non-overlapping segments. In three SES overlapping segments in LGS, malapposition was exclusively seen in the overlapping segment in one case and it extended to the proximal and distal non-overlapping segments beyond the overlapping segment in the other two cases.
The segment with any uncovered or malapposed struts was most common in DES overlapping segments with LGS by OCT evaluation (table 5). Among the DES, PES showed the highest rate of uncovered or malapposed struts at the site of LGS (75.0%, table 5), while SES showed elevated but evenly distributed rates of uncovered or malapposed struts in both HGS and LGS segments. BMS showed a higher rate of uncovered or malapposed struts in LGS than in HGS, although there was no statistical difference (40% vs 0%, respectively, p=NS). The incidence of >5% and >10% uncovered or malapposed struts was higher in PES in LGS than in HGS, although there was no statistical significance.
Aneurysmal change, which was defined as more than 1.5 times larger than the reference lumen area, was seen in one PES overlapping segment in HGS compared with one SES and one PES overlapping segment in LGS (p=NS, HGS vs LGS).
This study demonstrates that overlapping DES in minimally diseased or normal appearing coronary segments is associated with a high frequency of uncovered or malapposed struts. Conversely, angiographic restenosis occurred exclusively in overlapping stents deployed in coronary segments with ≥70%. These findings are preliminary and should be interpreted as hypothesis-generating, but suggest a differential vascular response to overlapping stents depending on the disease severity of the underlying coronary segment. Our study also confirmed the ability of high-resolution OCT for the evaluation of small neointimal detection.9
Overlapping BMS have been associated with a higher restenosis rate, possibly a result of excessive neointimal stimulation from the dual layers of struts.1 10 Our study provides further insights into the restenotic response of overlapping stents occurring exclusively at sites with severe disease and primarily due to a more pronounced NIH response. In the DES groups, on the other hand, restenosis appeared to be mainly due to poor stent expansion in HGS segments, although the post-dilatation was performed at overlapping site in almost 80% of cases.
DES overlapping at segments with LGS showed a high percentage of uncovered or malapposed struts. Indeed, uncovered or malapposed BMS struts were found only at LGS sites. The reduced strut coverage/apposition of DES has been previously considered to be a product of delayed endovascular healing at the site of overlap in animals.2 In this previous experiment, overlapping SES and PES showed different histopathological features, with PES inducing more eosinophils and greater fibrin deposition than SES, while SES was more often associated with giant cell peri-strut infiltration. Delayed healing was also shown in postmortem human in-stent restenosis after single PES implantation, with a greater amount of fibrin in PES than in SES.11 Drug and polymer properties have been implicated in these histological findings. Though only speculative, it may be that the higher incidence of uncovered/malapposed regions with dual layers of potent DES results from increased local drug or polymer concentrations, potentially reaching toxic levels in minimally diseased, thinner arterial wall segments. Because overlapping BMS were also responsible for uncovered/malapposed struts in LGS, though much less frequently than DES, mechanical factors associated with deploying two layers of metal should also be considered a causative factor in strut uncoverage or malapposition. Acknowledging that the long-term clinical impact of unfavourable strut coverage/apposition is yet to be demonstrated, the rates of uncovered/malapposed struts seen at LGS were high and corroborate previous concerns about the safety of overlapping DES, particularly the first-generation potent DES.2 11 12 Whether new stent platforms with thinner struts and a more conformable design will improve the rate of strut coverage and apposition at sites of LGS remains to be investigated.
These findings introduce a clinical dilemma to operators who must balance the risk of restenosis versus less coverage or apposition of overlapping stents depending on the underlying coronary disease. While determining the proper location for stent overlap might be challenging, the use of more aggressive high-pressure post-balloon inflations would help to mitigate the impact of disease severity on both restenosis and uncoverage rates in overlapping stents.
The sample size of this study was small, especially in the BMS group and three different types of DES were used for analysis. We did not perform OCT imaging immediately after stent implantation; therefore, we could not eliminate the possibility of strut-level baseline malapposition, although IVUS showed no persistent malapposition. Further extended malapposition from the overlapping segment to non-overlapping segment suggests the possibility of other concurrent factors that might influence the incidence of malapposition at the overlapping site. We expressed our data in arbitrary percentage cut-off values of uncovered/malapposed struts associated with stent thrombosis because of the lack of an established value in the clinical setting.
While OCT provides micron level resolution, it does not yet offer single-cell level resolution or functional tissue differentiation. Therefore, we cannot characterise the impact of disease severity on the type of tissue covering strut surfaces. Currently, accurate tissue characterisation is only possible with histology, which would preclude an in vivo assessment.
Funding Supported by Ospedali Riuniti Bergamo, Bergamo, Italy, University Hospitals of Cleveland and Case Western Reserve University, Cleveland, USA, with grant support from Medtronic Vascular, Santa Clara, USA, Boston Scientific Corporation, Natick, USA.
Competing interests MAC reports receiving consulting fees from Lightlab, Medtronic, Scitech, Cordis, Boston Scientifics and Abbott Vascular. GG reports receiving consulting fees from Boston Scientific and Volcano and receiving grant support from LightLab, Medtronic Vascular, Boston Scientific and Abbott Vascular.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the Ospedali Riuniti di Bergamo Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.