Mechanisms of Very Late Bioresorbable Scaffold …Manuscript received July 25, 2017; revised manuscript received September 6, 2017, accepted September 6, 2017. SEE PAGE 2345 ABBREVIATIONS

Listen to this manuscript’s

audio summary by

JACC Editor-in-Chief

Dr. Valentin Fuster.

J O U R N A L O F T H E A M E R I C A N C O L L E G E O F C A R D I O L O G Y V O L . 7 0 , N O . 1 9 , 2 0 1 7

ª 2 0 1 7 B Y T H E A M E R I C A N CO L L E G E O F C A R D I O L O G Y F O U N DA T I O N

P U B L I S H E D B Y E L S E V I E R

I S S N 0 7 3 5 - 1 0 9 7 / $ 3 6 . 0 0

h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . j a c c . 2 0 1 7 . 0 9 . 0 1 4

Mechanisms of Very LateBioresorbable Scaffold Thrombosis
The INVEST Registry
Kyohei Yamaji, MD, PHD,a,b Yasushi Ueki, MD,a Geraud Souteyrand, MD, MSC,c Joost Daemen, MD, PHD,d

Jens Wiebe, MD,e Holger Nef, MD,f Tom Adriaenssens, MD, PHD,g Joshua P. Loh, MBBS,h Benoit Lattuca, MD,i

Joanna J. Wykrzykowska, MD, PHD,j Josep Gomez-Lara, MD, PHD,k Leo Timmers, MD, PHD,l Pascal Motreff, MD, PHD,c

Petra Hoppmann, MD,m Mohamed Abdel-Wahab, MD,n Robert A. Byrne, MB, BCH, PHD,e Felix Meincke, MD,o

Niklas Boeder, MD,f Benjamin Honton, MD,p Crochan J. O’Sullivan, MD, PHD,q Alfonso Ielasi, MD,r

Nicolas Delarche, MD,s Günter Christ, MD,t Joe K.T. Lee, MD,a,u Michael Lee, MD, PHD,v Nicolas Amabile, MD, PHD,w

Alexios Karagiannis, PHD,x Stephan Windecker, MD,a Lorenz Räber, MD, PHDa

ABSTRACT

Fro

Ca

CH

da

Ca

Le

Na

Me

Inv

Ca

deoD

BACKGROUND Very late scaffold thrombosis (VLScT) occurs more frequently after bioresorbable scaffold (Absorb

BVS 1.1, Abbott Vascular, Santa Clara, California) implantation than with metallic everolimus-eluting stents.

OBJECTIVES The purpose of this study was to elucidate mechanisms underlying VLScT as assessed by optical

coherence tomography (OCT).

METHODS The INVEST (Independent OCT Registry on Very Late Bioresorbable Scaffold Thrombosis) registry is an

international consortium of investigators who used OCT to examine patients with VLScT.

RESULTS Between June 2013 and May 2017, 36 patients with 38 lesions who had VLScT underwent OCT at 19 centers.

VLScT occurred at a median of 20 months (interquartile range: 16 to 27 months) after implantation. At the time of VLScT,

83% of patients received aspirin monotherapy and 17% received dual-antiplatelet therapy. The mechanisms underlying

VLScT were (in descending order) scaffold discontinuity (42.1%), malapposition (18.4%), neoatherosclerosis (18.4%),

underexpansion or scaffold recoil (10.5%), uncovered struts (5.3%), and edge-related disease progression (2.6%).

Discontinuity (odds ratio [OR]: 110; 95% confidence interval [CI]: 73.5 to 173; p < 0.001), malapposed struts (OR: 17.0;

95% CI: 14.8 to 19.7; p < 0.001), and uncovered struts (OR: 7.3; 95% CI: 6.2 to 8.8; p < 0.001) were more frequent in

the thrombosed than the nonthrombosed scaffold regions. In 2 of 16 patients with scaffold discontinuity, intercurrent

OCT before VLScT provided evidence of circularly apposed scaffold struts with minimal tissue coverage.

CONCLUSIONS The leading mechanism underlying VLScT was scaffold discontinuity, which suggests an unfavorable

resorption-related process, followed by malapposition and neoatherosclerosis. It remains to be determined whether

modifications in scaffold design and optimized implantation can mitigate the risk of VLScT. (Independent OCT Registry on

Very Late Bioresorbable Scaffold Thrombosis [INVEST]; NCT03180931) (J Am Coll Cardiol 2017;70:2330–44)

© 2017 by the American College of Cardiology Foundation.

m the aSwiss Cardiovascular Center Bern, Department of Cardiology, Bern University Hospital, Bern, Switzerland; bDivision of

rdiology, Kokura Memorial Hospital, Kitakyushu, Japan; cDepartment of Cardiology, University Hospital of Clermont-Ferrand,

U Clermont-Ferrand, Clermont-Ferrand, France; dDepartment of Cardiology, Thoraxcenter, Erasmus Medical Center, Rotter-

m, the Netherlands; eDeutsches Herzzentrum München, Technische Universität München, Munich, Germany; fDepartment of

rdiology, University of Giessen, Giessen, Germany; gDepartment of Cardiovascular Medicine, University Hospitals Leuven,

uven, Belgium, and Department of Cardiovascular Sciences, Catholic University Leuven, Belgium; hDepartment of Cardiology,

tional University Heart Centre, Singapore; iDepartment of Cardiology, University Hospital Caremeau, Nimes, France; jAcademic

dical Center, University of Amsterdam, Amsterdam, the Netherlands; kHospital Universitari de Bellvitge, Institut d’

estigació Biomèdica de Bellvitge, Universitat de Barcelona, L’ Hospitalet de Llobregat, Barcelona, Spain; lDepartment of

rdiology, University Medical Center Utrecht, Utrecht, the Netherlands; mKlinik und Poliklinik Innere Medizin I, Klinikum rechts

r Isar, Technische Universität München, Munich, Germany; nHeart Center, Segeberger Kliniken GmbH, Bad Segeberg, Germany;

epartment of Cardiology, Asklepios Klinik St. Georg, Hamburg, Germany; pDepartment of Interventional Cardiology, Clinique

https://clinicaltrials.gov/ct2/show/NCT03180931

https://s3.amazonaws.com/ADFJACC/JACC7019/JACC7019_fustersummary_02




http://crossmark.crossref.org/dialog/?doi=10.1016/j.jacc.2017.09.014&domain=pdf

https://doi.org/10.1016/j.jacc.2017.09.014

AB BR E V I A T I O N S

AND ACRONYM S

3D = 3-dimensional

ACS = acute coronary

syndrome(s)

BMS = bare-metal stent(s)

BVS = bioresorbable vascular

scaffold

DAPT = dual-antiplatelet

therapy

DES = drug-eluting stent(s)

EES = everolimus-eluting

stent(s)

OCT = optical coherence

tomography

PCI = percutaneous coronary

intervention

PSP = pre-dilation, sizing, and

post-dilation

QCA = quantitative coronary

angiography

J A C C V O L . 7 0 , N O . 1 9 , 2 0 1 7 Yamaji et al.N O V E M B E R 7 , 2 0 1 7 : 2 3 3 0 – 4 4 OCT in Very Late Scaffold Thrombosis

2331

F ully bioresorbable coronary scaffolds wereintroduced to overcome the limitations ofmetallic stents associated with permanent

caging of the arterial wall aiming to restore vesselphysiology, as well as to mitigate the long-termrisk of device-related adverse events, including reste-nosis and stent thrombosis (1). Early results from6 randomized controlled trials comparing the Absorbbioresorbable vascular scaffold (BVS) with metalliceverolimus-eluting stents (EES) showed similarclinical outcomes up to 1 year (2–7). More recently,longer-term follow-up of the ABSORB II trial suggestedan increased risk of scaffold thrombosis at 3 years (8).In addition, the AIDA (Amsterdam Investigator-Initiated Absorb Strategy All-comers Trial), whichcompared BVS with metallic EES in an all-comerpopulation, reported a higher risk of target-vesselmyocardial infarction and early, late, and verylate device thrombosis at a median follow-up ofapproximately 2 years after BVS implantation (9).

SEE PAGE 2345 RVD = reference vessel

diameter

VLScT = very late scaffold

thrombosis

Meta-analyses of these trials, which included morethan 5,000 patients with follow-up beyond 1 year,consistently showed that BVS is inferior to EES interms of target-lesion failure, myocardial infarction,and target-lesion revascularization in the absence ofdifferences in mortality. Of note, the relative risk ofscaffold thrombosis was more than 3 times higherwith BVS than with EES during extended follow-up,with a concerning 10-fold increased relative risk ofvery late scaffold thrombosis (VLScT) between 1 and2 years after BVS implantation (10–15).

Mechanisms underlying early scaffold thrombosisare largely consistent with observations previously

Pasteur, Toulouse, France; qDepartment of Cardiology, Stadtspital Triemli,

gamo Est, Bolognini Hospital, Seriate (BG), Italy; sDepartment of Cardiology,t5th Medical Department with Cardiology, Kaiser Franz Josef Hospital, Vi

Medicine, Pamela Youde Nethersole Eastern Hospital, Chai Wan, Hong Ko

Queen Elizabeth Hospital, Kowloon, Hong Kong; wCardiology DepartmentxCTU Bern, and Institute of Social and Preventive Medicine (ISPM), Univers

Motreff have received speaker honoraria from Abbott Vascular, Terumo, an

research support from Acist Medical, Boston Scientific, St. Jude Medical, Med

Nef has received research grants and speaker honoraria from Abbott Vasc

Vascular. Prof. Loh has received speaker honoraria from Abbott Vascular; a

Scientific. Dr. Wykrzykowska has received speaker honoraria and institution

consultant for Boston Scientific; and has received institutional research g

Biotronik. Dr. Byrne has received lectures fees from B. Braun Melsungen

support (to the institution) from Boston Scientific and Heartflow. Prof. C

Vascular. Dr. Amabile has received consulting and lecture fees from Abbott V

(to the institution) from Abbott, Biotronik, Bracco, Boston Scientific, Medtro

received research grants (to the institution) from Abbott Vascular, Sanofi, a

they have no relationships relevant to the contents of this paper to disclose

Manuscript received July 25, 2017; revised manuscript received September 6

reported with metallic bare-metal stents(BMS) and drug-eluting stents (DES),including procedural factors such as acutedevice underexpansion, malapposition,incomplete lesion coverage, and residualdissections (16,17). Conversely, the principalcauses underlying VLScT remain largely un-known, and as a result, the translation ofmechanistic insights into modifications of theimplantation procedure or device iterationsto address specific components in scaffolddesign, polymer composition, and degrada-tion is lacking. VLScT is a clinical manifesta-tion of particular concern because it isresponsible for a considerable proportion ofdevice-related myocardial infarctions andcould impact the need for extended-durationdual-antiplatelet therapy (DAPT) (18).

Because of its high spatial resolution (10 to20 mm), optical coherence tomography (OCT)has become the imaging modality of choice forthe investigation of coronary device failuresincluding thrombosis and is of particularvalue to investigate time-dependent changesafter scaffold implantation (19–21). Against
this background, the INVEST (Independent OCT Reg-istry on Very Late Bioresorbable Scaffold Thrombosis)registry was conducted by an independent interna-tional consortium of investigators to investigate themechanisms underlying VLScT as assessed by OCT inthe largest worldwide cohort to date.
METHODS

STUDY POPULATION. Patients enrolled in theINVEST registry had to fulfill the following criteria: (1)

Zurich, Switzerland; rCardiology Division, ASST Ber-

CH F Mitterand, Pau Université Cedex, Pau, France;

enna, Austria; uCardiology Division, Department of

ng; vCardiology Division, Department of Medicine,

, Institut Mutualiste Montsouris, Paris, France; and

ity of Bern, Bern, Switzerland. Drs. Souteyrand and

d Medtronic. Dr. Daemen has received institutional

tronic, Pie Medical, ReCor Medical, and Terumo. Dr.

ular. Prof. Adriaenssens is a consultant for Abbott

nd a research grant (to the institution) from Boston

al grants from Abbott Vascular. Dr. Abdel-Wahab is a

rants from Abbott Vascular, St. Jude Medical, and

AG, Biotronik, and Boston Scientific; and research

hrist has received speaker honoraria from Abbott

ascular. Prof. Windecker has received research grants

nic, Edwards, St. Jude, and Terumo. Prof. Räber has

nd Regeneron. All other authors have reported that

.

, 2017, accepted September 6, 2017.

Yamaji et al. J A C C V O L . 7 0 , N O . 1 9 , 2 0 1 7

OCT in Very Late Scaffold Thrombosis N O V E M B E R 7 , 2 0 1 7 : 2 3 3 0 – 4 4

2332

presentation with very late (>1 year) definite scaffoldthrombosis after BVS (Absorb BVS 1.1, AbbottVascular, Santa Clara, California) implantation; (2)investigation by OCT at the time of scaffold throm-bosis; and (3) sufficient quality of the OCT recordingsto allow for core laboratory analysis.

VLScT was defined according to the AcademicResearch Consortium criteria (22). OCT analyses andquantitative coronary angiography (QCA) of the indexBVS implantation and interim follow-up (if available)and for VLScT (all cases) were performed at the corelaboratory at the Department of Cardiology, BernUniversity Hospital, Bern, Switzerland.

Written informed consent for further use of health-related data was obtained, if applicable (dependingon country-specific legislation). The study compliedwith the Declaration of Helsinki and was approvedby the institutional ethics committee at BernUniversity Hospital. The study was registered atclinicaltrials.gov (NCT03180931).OCT ACQUISITION AND ANALYSES. OCT imageswere acquired with commercially availablefrequency-domain OPTIS or C7 systems (LightLab/St.Jude Medical, Westford, Massachusetts) or Lunawave(Terumo Corporation, Tokyo, Japan). Frame rateswere 100, 158, or 180 frames/s, and pullback speedranged from 18 to 40 mm/s depending on the consoleand settings used. OCT pullbacks were assessed atevery frame (at 0.1-, 0.125-, 0.2-, and 0.25-mm in-tervals) using a proprietary software (QCU-CMSversion 4.69, LKEB, Leiden University, Leiden,Netherlands).

Details of OCT assessments were described previ-ously (19,23). In brief, the lumen contour was drawnby semi-automated detection using QCU-CMS soft-ware, following the endoluminal contour of the neo-intima with manual correction wherever required.The scaffold area was measured by joining the middlepoints of the signal-poor core of the abluminal side ofthe struts, excluding discontinued struts. In case ofthrombus with low attenuation, the lumen/scaffoldcontour could still be drawn. In case of high attenu-ation, the lumen/scaffold contour was extrapolatedwhen the lumen/scaffold contour was visible in atleast 3 quadrants.

Uncovered struts were identified in the absence ofa homogenous regular tissue coverage over the entirestrut. If the coverage of apposed struts could not beassessed because of the presence of attenuatingthrombus that precluded the evaluation of neo-intima, the struts were considered covered andcategorized as “apposed struts behind thrombus.”If the apposed struts were not covered byneointima but rather by irregular tissue, struts were

categorized as “apposed uncovered with super-imposed thrombus.” Malapposed struts were definedas struts for which the abluminal surface was clearlyseparated from the vessel wall. Neointimal thicknesswas measured from the abluminal border of apposedstrut core to the lumen. Discontinued struts weredefined as isolated malapposed struts that could notbe integrated in the expected circularity of the de-vice in at least 1 cross section or those with an abruptloss in longitudinal scaffold continuity betweenadjacent 2 frames (Figure 1). Neoatherosclerosis wasdefined as the presence of a fibroatheroma, irre-spective of fibrous cap thickness, or fibrocalcificplaque within the neointima. In cases where strutswere not clearly assessable because of a highlyattenuating layer of thrombus, struts were catego-rized as not analyzable.

To evaluate the relation between focal structuralabnormalities and thrombus formation, we parti-tioned the scaffolded portion into regions with thepresence and absence of intracoronary thrombus.The former was identified as any irregular mass in-side the lumen or attached to the neointima, with asharp border and various degrees of attenuation.When this extended 1.2 mm, this was termed athrombus region. The specific cutoff value of 1.2 mmwas determined in consistency with a previous pub-lication (19). Thrombus-free regions were used as acontrol.

Three-dimensional distances from strut to strutwere calculated. Mean neointimal thickness ofcovered struts was calculated for every 0.1 mm ofdistance from a target discontinued strut and coveredapposed struts to compare the neointimal thicknessin the adjacent regions of discontinued, malapposed,and uncovered struts.

OCT pullbacks (frame by frame) were assessed by 2analysts (K.Y. and Y.U.) and validated by a third an-alyst (L.R.). Cohen’s kappa was calculated to deter-mine the interobserver reproducibility of the scaffoldstrut assessment.ASSESSMENT OF THE MECHANISMS OF VLScT. Theprincipal mechanism of the VLScT was independentlyassessed by 4 investigators (K.Y., Y.U., N.A., L.R.).In case of more than 1 plausible mechanism, in-vestigators ranked the mechanism from 1 (mostplausible) to 3 (least plausible). Predefined mecha-nisms included scaffold discontinuity, malapposition,evagination (outward bulging of the vessel wall be-tween scaffold struts more than one-fourth of lumendiameter), uncovered struts, neoatherosclerosis,restenosis without neoatherosclerosis (neointimalhyperplasia >70% by visual estimate), scaffoldunderexpansion/device recoil (minimal scaffold

https://clinicaltrials.gov/ct2/show/NCT03180931

FIGURE 1 Cases of Scaffold Discontinuity as Assessed by OCT

Representative images are presented from a total of 16 patients (A-P) with scaffold discontinuity as a leading cause of VLScT. OCT ¼ optical coherence tomography;

VLScT ¼ very late scaffold thrombosis.


2333

area <50%), edge dissection, and edge-related diseaseprogression. The leading feature (ranking 1) wasdefined as the most prevalent in direct contact withthrombus. A final consensus meeting was held todetermine the final plausible cause of VLST, as re-ported in Figure 2.

The Fleiss kappa was calculated to assess theinterobserver agreement among the 4 assessors withrespect to the highest-ranked reason and putativecauses independent of the ranking.STATISTICAL ANALYSIS. Categorical variables arepresented as absolute values and percentages and

FIGURE 2 Representative Images and Frequency of the Single or Highest-Ranked Mechanism Underlying VLScT

Mechanism underlying very late scaffold thrombosis include (A) scaffold discontinuity, (B) malapposition, (C) neoatherosclerosis, (D) underexpansion, (E) under-

coverage, and (F) edge related progression. Asterisks indicate light attenuation due to lipid accumulation within the neointima of the scaffold. VLScT ¼ very late

scaffold thrombosis.



2334

were compared with Fisher exact test. Continuousvariables are expressed as mean � SD and comparedwith the Student t test. Thrombus and control regionswere defined in each scaffold. Associations of regiontype with frame-level OCT outcomes were estimatedusing generalized linear mixed models. Random in-tercepts were included at the patient level to accountfor the nonindependence of measurements from thesame patient. Statistical analyses were performedwiththe use of R version 3.4.1 (R Foundation for StatisticalComputing, Vienna, Austria) and Python version 3.6.0(Python Software Foundation, Beaverton, Oregon).All statistical analyses were 2-tailed. Values of p< 0.05were considered statistically significant.

RESULTS

BASELINE CHARACTERISTICS. Between June 2013and May 2017, 36 patients (38 lesions) with definite

VLScT underwent OCT before emergency percuta-neous coronary intervention (PCI) at 19 centers acrossEurope and Asia. Patient and lesion characteristicsare summarized in Tables 1 and 2. The mean durationbetween BVS implantation and VLScT was 21.9 �8.0 months (range 12 to 43 months). The majority ofpatients (72%) presented with acute coronary syn-drome (ACS) at the time of the index BVS implanta-tion. At the time of VLScT, patients received eitheraspirin monotherapy (83%) or DAPT (17%), althoughDAPT status was unknown in 1 patient. OCT wasperformed after thrombus aspiration in 66% of le-sions and after balloon dilation in 18%.

The mean scaffold diameter was 3.15 � 0.34 mm,with a total scaffold length of 29.5 � 15.7 mm. Multi-ple scaffolds were used in 24% of lesions. Intra-coronary imaging was used in 11 lesions (30.6%) at theindex BVS implantation. Pre-dilation was performedin 88% of lesions with a balloon diameter equal to the

TABLE 1 Patient Characteristics (N ¼ 36)

Time between index PCI and scaffold thrombosis,months

21.9 � 7.97

Patient age at time of index PCI, yrs 52.7 � 10.7

Male 33 (92)

BMI, kg/m2 26.2 � 3.78

Hypertension 17 (47)

Family history of coronary artery disease 13 (36)

Current smoker 15 (42)

Dyslipidemia 17 (47)

Diabetes mellitus 8 (22)

Diabetes mellitus, oral or insulin 4 (11)

Previous myocardial infarction 11 (31)

Previous PCI 13 (36)

Left ventricular ejection fraction, % 51.8 � 10.2

Number of diseased vessels

Single-vessel disease 23 (64)

2-vessel disease 8 (22)

3-vessel disease 5 (14)

Clinical presentation at time of index PCI

Stable angina 10 (28)

Unstable angina 3 (8)

Non–ST-segment elevation myocardial infarction 9 (25)

ST-segment elevation myocardial infarction 14 (39)

Clinical presentation at time of VLScT

Unstable angina 3 (8)

Non–ST-segment elevation myocardial infarction 8 (22)

ST-segment elevation myocardial infarction 25 (69)

Medications at discharge after index PCI

Dual-antiplatelet therapy 35 (97)

Aspirin 35 (97)

Clopidogrel 13 (36)

Prasugrel 12 (33)

Ticagrelor 11 (31)

Oral anticoagulation 1 (3)

Statin 34 (94)

Medications at time of presentation with VLScT

Dual-antiplatelet therapy 6 (17)

Aspirin 29 (83)

Clopidogrel 2 (6)

Prasugrel 1 (3)

Ticagrelor 3 (9)

Oral anticoagulation 3 (9)

Statin 28 (85)

Values are or mean � SD or n (%). Renal insufficiency was according to institu-tional criteria. Values were missing for BMI in 1 patient; smoking habit in 1 patient;left ventricular ejection fraction in 5 patients; medications for oral anticoagulation,b-blocker, and angiotensin-converting enzyme inhibitor or angiotensin receptorblocker at discharge in 1 patient; and medications at the time point of VLScT in1 patient.

BMI ¼ body mass index; PCI ¼ percutaneous coronary intervention;VLScT ¼ very late scaffold thrombosis.

TABLE 2 Lesion Characteristics (N ¼ 38)

Target vessel

Left anterior descending 16 (43)

Left circumflex 5 (14)

Right coronary artery 17 (45)

Bifurcation 8 (21)

Calcified lesion 5 (13)

Chronic total occlusion 2 (5)

Intracoronary imaging at time of index PCI

OCT 9 (25)*

Intravascular ultrasound 2 (6)

Number of scaffolds used 1.29 � 0.57

Scaffold overlap 9 (24)

Total scaffold length 29.5 � 15.7

Mean scaffold diameter 3.15 � 0.34

Pre-dilation 28 (88)

Pre-dilation balloon diameter, mm 2.81 � 0.46

Post-dilation 21 (60)

Maximal nominal post-dilation balloon diameter, mm 3.45 � 0.441

Maximal nominal post-dilation balloon diameterrelative to nominal scaffold diameter

0.00 mm 6 (18)

0.25 mm 4 (12)

0.50 mm 8 (24)

1.00 mm 2 (6)

Post-dilation balloon pressure, atm 16.6 � 4.30

$16 atm 11 (34)

Procedural variables at time of VLScT

Thrombus aspiration before OCT 25 (66)

Balloon dilation before OCT 7 (18)

Treatment for VLScT

Balloon dilation only 7 (18)

Additional metallic drug-eluting stent implantation 28 (74)

TIMI flow grade (pre)

0 25 (66)

1 2 (5)

2 3 (8)

3 8 (21)

TIMI flow grade (post)

3 38 (100)

Values are n (%) or mean � SD. One patient had VLScT in 2 lesions simultaneouslyat 18 months after BVS implantation, and another patient had VLScT in the leftcircumflex artery lesion at 16 months and in a right coronary artery lesion at32 months after BVS implantation. Values were missing for intracoronary imagingat the index PCI in 2 lesions, scaffold dimension in 2 lesions, pre-dilation in6 lesions, pre-dilation balloon diameter in 1 lesion, post-dilation in 3 lesions, post-dilation balloon dimension in 1 lesion, post-dilation balloon pressure in 3 lesions,and thrombolysis before procedure in 4 lesions. *Data on 4 OCT pullbacks at theindex BVS implantation were not available for analysis.

BVS ¼ bioresorbable vascular scaffold; OCT ¼ optical coherence tomography;TIMI ¼ Thrombolysis In Myocardial Infarction; other abbreviations as in Table 1.


2335

reference vessel diameter (RVD) in 68% of lesions.QCA measures at the time of the index BVS implan-tation are shown in Table 3. Appropriate BVS sizing(i.e., BVS 2.5 mm: RVD $2.5 mm and <2.75 mm; BVS3.0 mm: RVD $2.75 mm and <3.25 mm; BVS 3.5 mm:RVD $3.25 mm and #3.75 mm) was noted in 44% of

lesions. In the remainder, a >0.25-mm smaller BVS(compared with RVD) was used in 6.2% of lesions; a>0.25-mm larger BVS was used in 44% of lesions; andthe pre-RVD was <2.5 mm in 6.2%. Post-dilation wasperformed in 60% of lesions, with a maximal nominalpost-dilation balloon diameter of 3.45 � 0.44 mm. Apost-dilation balloon of equal or larger diameter thanthe scaffold was used in all lesions, whereas a

TABLE 3 Quantitative Coronary Angiography (N ¼ 32)

QCA pre-procedural

Lesion length, mm* 12.5 � 5.72

Reference diameter, mm 2.96 � 0.47

Minimal lumen diameter, mm 0.93 � 0.74

Diameter stenosis, % 69.4 � 22.7

Pre-dilation balloon diameter relative toreference vessel diameter

Equal to reference vessel diameter 21 (68)

Smaller than reference vessel diameter 6 (19)

None 4 (13)

Nominal scaffold diameter relative to referencevessel diameter

Appropriate 14 (44)

Reference vessel diameter <2.5 mm 2 (6)

Nominal BVS diameter < reference vesseldiameter � 0.25 mm

2 (6)

Nominal BVS diameter $ reference vesseldiameter þ 0.25 mm

14 (44)

QCA before BVS implantation




<30% 2 (6)

QCA post-procedural

Mean final balloon diameter, mm† 3.03 � 0.402

In-segment




<30% 27 (84)

In-scaffold

Scaffold length, mm 24.4 � 13.5




<30% 30 (94)

Acute recoil, mm 0.37 � 0.32

Acute recoil, % 12.1 � 10.5

Values are mean � SD or n (%). *Twenty-five lesions without total occlusion wereassessed. †Balloon diameter was assessed for every 1.5-mm-long subsegmentbetween the 2 radio-opaque markers.

BVS ¼ bioresorbable vascular scaffold; QCA ¼ quantitative coronaryangiography.



2336

post-dilation balloon size 1.0 mm larger than thescaffold was used in 2 lesions (5.7%). High-pressure($16 atm) post-dilation was performed in 34% of le-sions. Post-procedural in-segment percent diameterstenosis <30% was achieved in 84% of lesions.

MECHANISMS OF VLScT. A single mechanism un-derlying VLScT was identified in 23 cases (60.5%), 2 ormore coexisting mechanisms were identified in 14cases (36.8%), and no obvious reason for VLST wasidentified in 1 case (2.6%). The most plausible mech-anisms in each individual case as identified byconsensus among 4 investigators were, in descendingorder, scaffold discontinuity (n ¼ 16, 42.1%),

malapposition (n ¼ 7, 18.4%), neoatherosclerosis(n ¼ 7, 18.4%), underexpansion or scaffold recoil(n ¼ 4, 10.5%), uncovered struts (n ¼ 2, 5.3%), andedge-related disease progression (n ¼ 1, 2.6%)(Figure 2). When we considered coexisting findings,the leading mechanism underlying VLScT remainedscaffold discontinuity (n ¼ 20, 52.6%), followed bymalapposition (n ¼ 9, 23.7%), uncovered struts (n ¼ 8,21.1%), neoatherosclerosis (n ¼ 7, 18.4%), under-expansion or scaffold recoil (n ¼ 6, 15.8%), and edge-related disease progression (n ¼ 2, 5.3%). Agreementamong the 4 assessors (before the consensus meeting)with respect to the highest-ranked reason was mod-erate (k ¼ 0.67) and good (k ¼ 0.70) for the assessmentof various causes independent of the ranking.

In an exploratory analysis, there was no significantdifference in patient, lesion, or procedural charac-teristics or QCA findings between patients with le-sions with versus without scaffold discontinuity asthe leading mechanism underlying VLScT, exceptthat thrombus aspiration at the time of VLScT wasless often performed in lesions with scaffold discon-tinuity (Online Tables 1 to 3). In 38% of patients withscaffold discontinuity, neither thrombus aspirationnor balloon dilation was performed before OCT.Although no significant relationship between thetime course of the occurrence of the scaffold throm-bosis and the underlying mechanism was observed(p ¼ 0.69) (Online Figure 1), the duration between theindex BVS implantation and VLScT in 7 lesions withneoatherosclerosis as the leading mechanism under-lying VLScT was numerically longer than in thosewithout (26.9 � 11.3 months vs. 21.0 � 6.8 months;p ¼ 0.23). No significant differences in the underlyingmechanisms of VLScT were observed between lesionstreated for ACS versus non-ACS at the index proced-ure (p ¼ 0.40) (Online Figure 2) or according topresence or absence of optimized implantation tech-niques (including pre-dilation, sizing, and post-dilation [PSP]; p ¼ 0.50) (Online Figure 3).

QUALITATIVE AND QUANTITATIVE OCT ASSESSMENT.

Descriptive OCT assessments are summarized inTable 4. A total of 7,096 frames with 56,299 strutswere analyzed. Lumen area was assessed in 7,065frames (99.6%), excluding 31 frames with stronglyattenuating thrombus that precluded depicting thelumen contour in at least 3 quadrants. Scaffold areawas assessed in 5,708 frames (80.4%) in which scaf-fold contours were assessable in at least 3 quadrants.Scaffold expansion amounted to 85.6 � 28.6%, and anasymmetry index >0.3 was observed in 33 lesions(92%), whereas an eccentricity index <0.7 was pre-sent in 17 lesions (47%). Thrombus was identified in





TABLE 4 Quantitative Optical Coherence Tomography (N ¼ 38)

Lesion level

Number of frames 187.0 � 99.8

Scaffold length, mm 27.9 � 13.0

Mean reference area, mm2 6.93 � 2.33

Minimal lumen area, mm2 3.31 � 1.27

Area stenosis, % 50.2 � 18.7

Minimal scaffold area, mm2 5.63 � 1.70

Scaffold expansion, % 85.6 � 28.6

Asymmetry index 0.43 � 0.092

>0.3 33 (92)

Eccentricity index 0.71 � 0.076

<0.7 17 (47)

Frame level

Number of frames with lumen area 186 � 101

Mean lumen area, mm2 5.56 � 1.72

Number of frames with scaffold area 150 � 107

Mean scaffold area, mm2 8.20 � 1.65

Mean neointimal area, mm2 2.71 � 1.02

Mean incomplete scaffold apposition area, mm2 1.03 � 0.81

Number of frames with thrombus 75.1 � 56.3

Mean thrombus area, mm2 0.95 � 0.85

Total thrombus length, mm 11.1 � 6.33

Number of frames with neoatherosclerosis 20.8 � 41.6

Number of frames with visible struts 184 � 100

Strut level

Number of struts 1,482 � 978

Mean strut core area, mm2 0.0305 � 0.0063

Apposed struts covered by neointimal, % 87.9 � 14.7

Mean neointimal thickness, mm 0.348 � 0.100

Apposed uncovered struts, % 1.33 � 3.18

Apposed uncovered with superimposed thrombus, % 0.88 � 2.04

Apposed struts behind thrombus, % 4.56 � 5.18

Malapposed (nondiscontinued) struts, % 2.60 � 5.14

Mean incomplete scaffold appositiondistance, mm

0.262 � 0.137

Discontinuous struts (malapposed), % 1.56 � 3.17

Struts over side branch, % 0.34 � 0.46

Nonclassifiable struts, % 0.84 � 3.38

Values are mean � SD or n (%).


2337

2,852 frames (40.2%), with an average thrombuslength of 11.1 � 6.3 mm and area of 0.95 � 0.84 mm2.Neoatherosclerosis was observed in 15 patients (792frames, 11.2%).

Struts were apposed and covered by neointima in87.9 � 14.7% of lesions, with a mean neointimathickness of 0.34 � 0.10 mm. Uncovered struts(1.3 � 3.2%) or struts without neointimal coverage butwith superimposed thrombus (0.88 � 2.04%) wererarely observed. When the overlaying thromboticlayer did not allow a differentiation betweenthrombus and neointima but only the assessmentof apposition, struts were categorized as apposedstruts behind thrombus (4.56 � 5.18%). Malapposed(nondiscontinued) struts were observed in 2.60 �5.14% of lesions. Strut discontinuity was observed in1.56 � 3.17% of lesions. The intra-observer and inter-observer variabilities, respectively, for the assess-ment of uncovered struts (k ¼ 0.96 and k ¼ 0.96),malapposed struts (k ¼ 0.92 and k ¼ 0.88), and dis-continued struts (k ¼ 0.83 and k ¼ 0.87) were good.

Lesions with scaffold discontinuity as a leadingmechanism underlying VLScT had reduced scaffoldexpansion (73.2 � 20.9% vs. 94.4 � 30.5%; p ¼ 0.02)and less neointimal thickness (296 � 55 mm vs. 385 �109 mm; p ¼ 0.002) (Online Table 4). The neointimalthickness (of covered struts) increased as a functionof the 3-dimensional (3D) distance from discontin-uous struts (Figure 3).

OCT performed at the time of the index procedurewas available in 4 patients, an interim OCT wasavailable in 2 patients, and both OCT at the time of theindex procedure and an interim OCT were available in1 patient. In 2 cases with the finding of strut discon-tinuity as the mechanism underlying VLScT, interimOCT evidenced fully apposed and covered struts(Figures 4A and 4B), whereas discontinued struts atthe time of VLScT were found to already be malap-posed at the time of the index procedure in 1 case(Figure 4C).OCT FINDINGS IN SCAFFOLD REGIONS WITH AND

WITHOUT THROMBUS. Detailed analysis of OCTfindings derived from regions with thrombuscompared with those without thrombus (controls) aresummarized in Table 5. Among 38 lesions, no thrombusregion was identified in 3 lesions (i.e., consecutivethrombus length <1.2 mm). Regions with thrombuscompared with those without thrombus had a smallerminimum lumen and scaffold diameter and a smallerneointimal thickness. The mean and maximal incom-plete scaffold apposition distance and area weresignificantly larger in thrombotic regions than in con-trol regions. Regions with thrombus as opposed tothose without were characterized by a 110 times higher

proportion of discontinued struts, a 17 times higherproportion of malapposed struts, and a 7 timeshigher proportion of apposed uncovered struts. The2-dimensional scaffold strut maps in Figure 5 providea visual distribution of accumulated adverse scaffoldstrut characteristics (discontinuation, malappositionand uncovered struts, or neoatherosclerosis). OnlineVideos 1 and 2 show representative 3D re-constructions illustrating the leading cause of VLScTand the corresponding scaffold strut maps.

DISCUSSION

The INVEST registry, which represents the largestcohort and first international multicenter consortium


http://jaccjacc.acc.org/video/2017/2828_VID1.mp4



FIGURE 3 Neointimal Thickness of Covered Struts

0

50

100

150

200

Neoi

ntim

al T

hick

ness

(μm

)

250

300A

0 1 23D-Distance from Discontinued Struts (mm)

3 4 5

Neoi

ntim

al T

hick

ness

(μm

)

B

3D-Distance from Malapposed Struts (mm)

0

50

100

150

200

250

300

0 1 2 3 4 5

Neoi

ntim

al T

hick

ness

(μm

)

C

0

50

100

150

200

250

300

03D-Distance from Covered Struts (mm)

1 2 3 4 5

Neointimal thickness of covered struts across 3-dimensional (3D) distance from

(A)discontinuedstruts, (B)malapposedstruts, and (C)coveredstruts.3Ddistance

was calculated between each target strut and all other struts. Mean of neointimal

thickness was calculated according to distance from the target struts.



2338

to investigate the mechanisms underlying VLScT, hasthe following salient findings (Central Illustration).First, scaffold discontinuity (a resorption-relatedphenomenon not encountered with metallic DES)was the most common mechanism underlying VLScT(42%), followed by strut malapposition (18%) andneoatherosclerosis (18%). Second, the abnormalscaffold strut findings of discontinuity and malap-position were strongly associated with the presenceof thrombus, with scaffold strut maps providingadditional evidence in support of a causal relation-ship. Third, the majority of VLScT patients receivedonly aspirin monotherapy at the time of the adverseevent. Fourth, only 1 of 4 lesions had been treatedaccording to current guidelines criteria for appro-priate BVS implantation.

As opposed to early (and late) device thrombosisthat can be easily related to procedural and antith-rombotic treatment issues, the phenomenon of VLScThas been encountered somewhat unexpectedly dur-ing a time period when long-term benefits of BVSwould have been expected to emerge over permanentmetallic stents because of vascular restoration.Recently, the relative risk of VLScT was increased10-fold with the Absorb BVS compared with metallicEES, which represents an important clinical concernwith potentially serious adverse outcomes. To date, atotal of 27 definite VLScT cases have been reported in7 randomized controlled trials comparing Absorb BVSwith metallic EES (8,9,24–28). The present study ex-tends these clinical observations by providing thedetailed OCT findings of 36 patients with 38 lesionstreated with BVS in routine clinical practice andtherefore offers important insights into the patho-physiological mechanisms underlying VLScT.

SCAFFOLD DISCONTINUITY AS LEADING MECHANISM

UNDERLYING VLScT. Scaffold discontinuity representsthe loss of the circular scaffold strut lining and occursas result of device degradation, that is, after thestructural device integrity is no longer maintained(12 months in Absorb BVS) (29). Discontinuous strutsembedded in neointima have been reportedfrequently (i.e., 25% of lesions show any discontinuityat 2 years [30] and 42% at 3 years [31]) and were notassociated with clinical sequelae in the ABSORB B andABSORB Japan studies (personal communication fromProf. Takeshi Kimura, Japan, July 31, 2017). Discon-tinuous struts can penetrate into the lumen and cause(late acquired) malapposition, thus exposing thehighly thrombogenic remnant scaffold material tothe blood flow, with subsequent activation of thecoagulation cascade. Indeed, discontinuous scaffoldstruts have been associated with VLScT in small case

FIGURE 4 Matched OCT Cross Sections in Patients With Scaffold Discontinuity at the Time of Scaffold Thrombosis

Arrowheads indicate scaffold struts. (A) This patient underwent OCT at 14 months’ follow-up and experienced VLScT at 34 months after implantation of a BVS.

Asterisks indicate calcium deposit to validate matching. (B) This patient underwent OCT at 10 months’ follow-up and experienced VLScT at 18 months after BVS

implantation. (C) This patient underwent OCT at the index BVS implantation and experienced VLScT at 16 months after BVS implantation. BVS ¼ bioresorbable vascular

scaffold; FUP ¼ follow-up; other abbreviations as in Figure 1.

TABLE 5 OCT Findings in Regions With Versus Without Thrombus

Full Scaffold Region Thrombus Region Control Region Difference or Ratio p Value

No. of patients (frames) analyzed 35 (6,318) 35 (2,394) 35 (3,924) Difference

Minimal lumen diameter, mm 2.32 (2.18 to 2.47) 2.31 (2.16 to 2.45) 2.34 (2.19 to 2.48) �0.03 (�0.05 to �0.01) 0.002

Minimal scaffold diameter, mm* 2.93 (2.80 to 3.05) 2.87 (2.75 to 3.00) 2.96 (2.84 to 3.08) �0.09 (�0.11 to �0.07) <0.001

Average neointimal thickness, mm* 0.287 (0.252 to 0.321) 0.268 (0.234 to 0.303) 0.297 (0.263 to 0.331) �0.03 (�0.03 to �0.02) <0.001

No. of patients (frames, struts) analyzed 35 (6,277, 50,449) 35 (2,360, 15,928) 35 (3,917, 34,521) Ratio

Percentage of struts

Uncovered, apposed† 0.49 (0.19 to 1.08) 1.34 (0.54 to 2.82) 0.18 (0.07 to 0.40) 7.34 (6.17 to 8.77) <0.001

Malapposed‡ 1.44 (0.66 to 2.82) 4.43 (2.03 to 8.62) 0.27 (0.12 to 0.55) 17.0 (14.8 to 19.7) <0.001

Discontinued‡ 0.56 (0.25 to 1.07) 1.96 (0.83 to 3.92) 0.02 (0.01 to 0.04) 110 (73.5 to 173) <0.001

No. of patients (frames, struts) analyzed 35 (6,277 to 50,449) 35 (2,360 to 15,928) 35 (3,917 to 34,521) Ratio

Percentage of frames with

$30% uncovered, apposed struts† 0.77 (0.26 to 1.70) 2.02 (0.72 to 4.26) 0.17 (0.05 to 0.39) 12.4 (7.86 to 20.4) <0.001

$30% malapposed struts‡ 2.76 (1.15 to 5.40) 7.64 (3.05 to 15.0) 0.35 (0.13 to 0.79) 23.3 (16.9 to 33.0) <0.001

$30% discontinued, malapposed struts‡ 0.87 (0.30 to 1.84) 2.46 (0.77 to 5.47) 0.02 (0.00 to 0.07) 118 (48.1 to 390) <0.001

No. of patients (frames) analyzed 26 (1,086) 26 (609) 26 (477) Difference

ISA area, mm2§ 0.429 (0.261 to 0.594) 0.488 (0.317 to 0.657) 0.333 (0.154 to 0.510) 0.156 (0.043 to 0.269) 0.007

Average ISA distance, mm§ 0.140 (0.099 to 0.180) 0.153 (0.111 to 0.194) 0.119 (0.076 to 0.162) 0.034 (0.009 to 0.059) 0.009

Maximal ISA distance, mm§ 0.234 (0.166 to 0.301) 0.264 (0.195 to 0.332) 0.186 (0.114 to 0.256) 0.079 (0.037 to 0.120) <0.001

Values are estimates (95% confidence interval) unless otherwise indicated. *Denominator: all frames with lumen and scaffold area. †Denominator: all apposed struts. ‡Denominator: all measured struts.§Denominator: all frames with malapposed area.

ISA ¼ incomplete scaffold apposition; other abbreviations as in Table 1.


2339

FIGURE 5 Strut Maps of Scaffolded Regions From OCT Analysis Showing Spatial Distribution of Discontinuity, Malapposition, and Uncovered

Struts for All Lesions

The longitudinal extensions of neoatherosclerosis, scaffold overlap, and thrombus regions are shown as bars. Unwrapped scaffold circumference was plotted along the

x-axis, and longitudinal scaffold extension was plotted along the y-axis. Each dot represents a strut. Lesions are presented according to decreasing scaffold length and

not case number. Online Videos 1 and 2 show representative 3D reconstructions illustrating the leading cause of VLScT and the corresponding scaffold strut maps.



2340

series reported by our group (23) and others (16,30).The INVEST registry extends these initial observationsand identifies scaffold discontinuity as the leadingmechanism underlying VLScT (42%). The strong

association between strut discontinuity and presenceof thrombus with colocalization on the scaffold strutmaps supports a causal relationship of this finding inthe absence of a control group.



CENTRAL ILLUSTRATION Very Late Bioresorbable Scaffold Thrombosis: Underlying Causes

11%

OthersUnderexpansion

or scaffold shrinkage

Neoatherosclerosis

Malapposition

Scaffolddiscontinuity

11%

18%

18%

42%

Yamaji, K. et al. J Am Coll Cardiol. 2017;70(19):2330–44.

Frequency of the 4 leading mechanisms by means of optical coherence tomography analysis is presented, including a representative example.

“Others” includes uncovered struts (5.3%), edge-related disease (2.6%), and unknown (2.6%).


2341

The mechanisms that lead to discontinuous strutsresulting in malapposition (i.e., as opposed todiscontinuity embedded in neointima, the expectedconsequence of strut resorption) remain poorly un-derstood. Malapposed (circularly aligned) strutswithout neointimal encapsulation might represent aprecursor of discontinued struts. Indeed, in 1 casewith serial OCT, strut discontinuity was preceded bymalapposition at the time of device implantation(Figure 4C). Patient baseline, procedural, and lesioncharacteristics did not differ between patients with

VLScT due to scaffold discontinuity compared toother causes in an exploratory analysis. The onlyemerging difference at the time of VLScT was athinner neointimal layer among cases with scaffolddiscontinuity. A 3D analysis investigating the rela-tionship between neointimal thickness as a functionof the distance from discontinuous struts indicatedthat neointimal thickness decreased in the vicinity ofdiscontinuous struts. Of note, 2 patients with serialOCT provided evidence that scaffold discontinuitycan be preceded by a state of fully apposed and



2342

covered struts (Figures 4B and 4C). This observation isof concern because fully covered and apposed scaf-fold struts are believed to represent a benign findingthat suggests optimal deployment and arterialhealing.

The potential conditions that allow scaffold strutsegments to migrate into the lumen despite thepresence of tissue coverage might be related to me-chanical or biological phenomena or a combinationthereof. Thus, physical forces toward the lumencenter such as vessel contraction or external shearstress (i.e., vessel curvature) might come into play. Inaddition, the tissue covering scaffold struts couldplay a critical role to ensure long-term strut fixation.Although a mature neointima with a critical thicknessis expected to ensure full encapsulation of scaffoldstrut fragments, a very thin and incomplete neo-intima or coverage by tissue other than neointima(i.e., fibrin or organized thrombus) might allow forsubsequent strut penetration into the lumen. In thisregard, it is noteworthy that OCT cannot differentiatecompetent neointima from organized thrombus orfibrin. According to a recent in vitro study (32), BVShave a higher risk for thrombus or fibrin depositionthan their metallic counterpart. Indeed, BVS wasassociated with more platelet adherence and sub-stantially delayed stent strut coverage after 28 daysas assessed by electron scan microscopy (33). Delayedhealing with fibrin apposition rather than neointimalformation was a hallmark of the adverse arterialhealing response after implantation of first-generation paclitaxel-eluting stents (34). Becausethe biodegradation process of BVS results in an in-flammatory peak reaction approximately 1 year afterimplantation, a potent inflammatory response insome individuals leading to a necrotizing vasculitiswith subsequent penetration of scaffold remnantsmight represent an alternative biological explanationfor the phenomenon of scaffold discontinuity.Further research is required to investigate the deli-cate interplay between the resorption process and thequality and amount of tissue coverage to mitigate therisk of scaffold discontinuity–related VLScT. Severalongoing modifications in scaffold design are focusedon reducing strut thickness, modulating resorptiontime, and testing different polymer and non-polymeric materials to reduce the inflammatoryresponse and potentially mitigate the risk of scaffolddiscontinuity and scaffold thrombosis.

MALAPPOSITION. Strut malapposition unrelatedto scaffold discontinuity was the second- leadingmechanism underlying VLScT (18.4% of lesions).Malapposition consistently emerged as a principal

finding associated with very late metallic DESthrombosis in several OCT registries (19–21). Althoughit is impossible to distinguish between persistentand late acquired malapposition in the absence of abaseline OCT investigation, it is noteworthy that wedid not observe coronary evaginations, an OCT findingpreviously correlated with positive vessel remodeling.According to a recent substudy of the ABSORB II trial,positive remodeling occurred in 29% of Absorb BVS–treated lesions (35). The dynamic of the remodelingresponse appears slower than the loss of structuralintegrity, so that the fragmented device mightaccommodate the coinciding growth of the vessel walland might not lead to late acquired malapposition,a strong correlate for very late stent thrombosiswith first-generation DES. Along these lines, persis-tent scaffold malapposition was more frequentlyobserved than late acquired scaffold malapposition(81% vs. 16%) in the pivotal ABSORB trial (36).

NEOATHEROSCLEROSIS. Neoatherosclerosis wasobserved as a mechanism underlying VLScT in 18.4%of lesions at 26.9� 11.3 months after BVS implantation.Otsuka et al. (37) showed that inflammation sur-rounding struts was greater in BVS at 6 to 36 monthsthan in metallic DES in a porcine model. A moreintense inflammation during the absorption processcould accelerate the recurrence of atheroscleroticlesions within neointima because of impaired endo-thelial function, reportedly a key trigger ofneoatherosclerosis after DES implantation (34). Theobservation of neoatherosclerosis occurring early afterBVS implantation challenges the concept of BVS-related neocap formation as the neocap may serve asa source for lipid accumulation (38). We and otherspreviously described neoatherosclerosis as one of theprincipal factors underlying very late metallic DESthrombosis (i.e., approximately one-third of cases);however, the time from the index procedure tothrombosis was much longer (5.2 � 2.8 years after theindex procedure) than in the present scaffold series.

IMPACT OF IMPLANTATION TECHNIQUE. Optimizedimplantation techniques, including PSP, have beenproposed to mitigate the risk of device-relatedadverse events after BVS implantation (39,40). Inour cohort, pre-dilation was performed in 88% oflesions, appropriate scaffold sizing in relation to thereference vessel diameter was applied in 44% oflesions, and post-dilation was performed in 60% oflesions, whereas high pressure ($16 atm) was usedin 34% of lesions. The appropriate BVS size togetherwith pre-dilation and post-dilation was documentedin 9 lesions (24%). PSP is expected to lowerthe frequency of underexpansion and acute

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: The most

frequent mechanism of VLScT of the ABSORB BVS bioresorbable

scaffolds is scaffold discontinuity, followed by malapposition

and neoatherosclerosis. This suggests that processes related to

resorption of scaffold material inside the coronary artery lumen

may contribute to the development of late thrombosis.

TRANSLATIONAL OUTLOOK: Future studies should inves-

tigate the relationships between device-specific resorption,

deployment technique, and duration of DAPT to optimize the

long-term outcomes and prevent VLScT of bioresorbable

coronary scaffolds.


2343

malapposition, whereas an interaction with neo-atherosclerosis is not expected. To what degree thescaffold strut discontinuity can be mitigated by PSPremains unknown. One can only speculate that animproved expansion and embedment of scaffoldstruts could result in an improved neointimalencapsulation and thereby less thrombosis-pronediscontinuity.

EXTENDED DAPT DURATION. Because the 2 leadingmechanisms underlying VLScT are associated withtransient strut material inside the lumen, extensionof DAPT until the time of complete scaffold resorption(i.e., 3 to 4 years) appears to be important to mitigatethe risk of scaffold thrombosis among patients withan Absorb BVS while carefully balancing the associ-ated risk of bleeding.

STUDY LIMITATIONS. First, in this observationalstudy, the absence of a control group of patientswithout scaffold thrombosis is notable. The inclusionof an adequate number of well-matched control pa-tients is challenging for any stent thrombosis registry,particularly OCT registry. Furthermore, it is notfeasible to include a control group given the world-wide registry and the widely varied time frame fromscaffold implantation to thrombosis. Second, wecannot exclude a certain degree of selection, biasbecause patients had to have undergone OCT at thetime of scaffold thrombosis. Third, the number ofpatients analyzed was relatively small; however, thisis the largest cohort to date assessing VLScT by OCTand included more VLScT cases than reported in allrandomized controlled trials combined. Fourth, strutdiscontinuity could theoretically represent a directconsequence of thrombus aspiration or balloon dila-tion before OCT imaging. The observation thatthrombus aspiration and pre-dilation before OCT wasperformed less frequently in cases with scaffolddiscontinuity speaks against this hypothesis (OnlineFigures 4 to 6), as does the fact that discontinuitywas almost exclusively observed in regions withthrombus (Table 5). Fifth, it is sometimes challengingto assess strut characteristics, including stent strut

coverage and apposition, in the presence of thrombuswith various degrees of light attenuation. Neverthe-less, OCT represents the most precise method forfailure analysis to date. Sixth, we used a qualitativerather than a quantitative approach to classifyuncovered struts; however, only 0.2% of coveredstruts showed a neointimal thickness <30 um, an(unvalidated) cutoff previously recommended toassess scaffold strut coverage (41).

CONCLUSIONS

Using intracoronary OCT imaging, the INVEST regis-try discloses multifactorial causes of VLScT. Theleading mechanism underlying VLScT was scaffolddiscontinuity, which suggests an unfavorableresorption-related process, followed by malap-position and then neoatherosclerosis. It remains to bedetermined whether modifications in scaffold designand optimized implantation techniques can mitigatethe risk of VLScT.

ADDRESS FOR CORRESPONDENCE: Dr. LorenzRäber, Department of Cardiology, Bern UniversityHospital, Freiburgstrasse, 3010 Bern, Switzerland.E-mail: [email protected].

RE F E RENCE S

1. Serruys PW, Garcia-Garcia HM, Onuma Y. Frommetallic cages to transient bioresorbable scaf-folds: change in paradigm of coronary revascu-larization in the upcoming decade? Eur Heart J2012;33:16–25b.

2. Sabaté M, Windecker S, Iñiguez A, et al.Everolimus-eluting bioresorbable stent vs. durablepolymer everolimus-eluting metallic stent in pa-tients with ST-segment elevation myocardialinfarction: results of the randomized ABSORB ST-

segment elevation myocardial infarction-TROFI IItrial. Eur Heart J 2016;37:229–40.

3. Serruys PW, Chevalier B, Dudek D, et al.A bioresorbable everolimus-eluting scaffold versusa metallic everolimus-eluting stent for ischaemicheart disease caused by de-novo native coronaryartery lesions (ABSORB II): an interim 1-year anal-ysis of clinical and procedural secondary outcomesfrom a randomised controlled trial. Lancet 2015;385:43–54.

4. Kimura T, Kozuma K, Tanabe K, et al., for theABSORB Japan Investigators. A randomized trialevaluating everolimus-eluting Absorb bioresorbablescaffolds vs. everolimus-eluting metallic stents inpatients with coronary artery disease: ABSORBJapan. Eur Heart J 2015;36:3332–42.

5. Gao R, Yang Y, Han Y, et al. Bioresorbablevascular scaffolds versus metallic stents in pa-tients with coronary artery disease: ABSORB Chinatrial. J Am Coll Cardiol 2015;66:2298–309.



mailto:[email protected]

http://refhub.elsevier.com/S0735-1097(17)39671-7/sref1
































2344

6. Ellis SG, Kereiakes DJ, Metzger DC, et al., forthe ABSORB III Investigators. Everolimus-elutingbioresorbable scaffolds for coronary artery dis-ease. N Engl J Med 2015;373:1905–15.

7. Puricel S, Arroyo D, Corpataux N, et al.Comparison of everolimus- and biolimus-elutingcoronary stents with everolimus-eluting bio-resorbable vascular scaffolds. J Am Coll Cardiol2015;65:791–801.

8. Serruys PW, Chevalier B, Sotomi Y, et al. Com-parison of an everolimus-eluting bioresorbablescaffold with an everolimus-eluting metallic stentfor the treatment of coronary artery stenosis(ABSORB II): a 3 year, randomised, controlled,single-blind, multicentre clinical trial. Lancet2016;388:2479–91.

9. Wykrzykowska JJ, Kraak RP, Hofma SH, et al.,for the AIDA Investigators. Bioresorbable scaffoldsversus metallic stents in routine PCI. N Engl J Med2017;376:2319–28.

10. Cassese S, Byrne RA, Ndrepepa G, et al. Ever-olimus-eluting bioresorbable vascular scaffoldsversus everolimus-eluting metallic stents: a meta-analysis of randomised controlled trials. Lancet2016;387:537–44.

11. Lipinski MJ, Escarcega RO, Baker NC, et al.Scaffold thrombosis after percutaneous coronaryintervention with ABSORB bioresorbable vascularscaffold: a systematic review and meta-analysis.J Am Coll Cardiol Intv 2016;9:12–24.

12. Collet C, Asano T, Miyazaki Y, et al. Latethrombotic events after bioresorbable scaffoldimplantation: a systematic review and meta-analysis of randomized clinical trials. Eur Heart J2017;38:2559–66.

13. Sorrentino S, Giustino G, Mehran R, et al.Everolimus-eluting bioresorbable scaffolds versuseverolimus-eluting metallic stents. J Am CollCardiol 2017;69:3055–66.

14. Toyota T, Morimoto T, Shiomi H, et al. Verylate scaffold thrombosis of bioresorbable vascularscaffold: systematic review and a meta-analysis.J Am Coll Cardiol Intv 2017;10:27–37.

15. Ali ZA, Serruys PW, Kimura T, et al. 2-yearoutcomes with the Absorb bioresorbable scaffoldfor treatment of coronary artery disease: a sys-tematic review and meta-analysis of seven rand-omised trials with an individual patient datasubstudy. Lancet 2017;390:760–72.

16. Karanasos A, Van Mieghem N, vanDitzhuijzen N, et al. Angiographic and opticalcoherence tomography insights into bioresorbablescaffold thrombosis: single-center experience. CircCardiovasc Interv 2015;8:e002369.

17. Cuculi F, Puricel S, Jamshidi P, et al. Opticalcoherence tomography findings in bioresorbablevascular scaffolds thrombosis. Circ CardiovascInterv 2015;8:e002518.

18. Lemesle G, Tricot O, Meurice T, et al. Incidentmyocardial infarction and very late stent throm-bosis in outpatients with stable coronary arterydisease. J Am Coll Cardiol 2017;69:2149–56.

19. Taniwaki M, Radu MD, Zaugg S, et al. Mecha-nisms of very late drug-eluting stent thrombosisassessed by optical coherence tomography. Cir-culation 2016;133:650–60.

20. Souteyrand G, Amabile N, Mangin L, et al., forthe PESTO Investigators. Mechanisms of stentthrombosis analysed by optical coherence to-mography: insights from the national PESTOFrench registry. Eur Heart J 2016;37:1208–16.

21. Adriaenssens T, Joner M, Godschalk TC, et al.,on behalf of the Prevention of Late Stent Throm-bosis by an Interdisciplinary Global EuropeanEffort (PRESTIGE) Investigators. Optical coher-ence tomography findings in patients with coro-nary stent thrombosis: a report of the PRESTIGEConsortium (Prevention of Late Stent Thrombosisby an Interdisciplinary Global European Effort).Circulation 2017;136:1007–21.

22. Cutlip DE, Windecker S, Mehran R, et al., onbehalf of the Academic Research Consortium.Clinical end points in coronary stent trials: a casefor standardized definitions. Circulation 2007;115:2344–51.

23. Räber L, Brugaletta S, Yamaji K, et al. Very latescaffold thrombosis: intracoronary imaging andhistopathological and spectroscopic findings. J AmColl Cardiol 2015;66:1901–14.

24. Arroyo D, Gendre G, Schukraft S, et al.Comparison of everolimus- and biolimus-elutingcoronary stents with everolimus-eluting bio-resorbable vascular scaffolds: two-year clinicaloutcomes of the EVERBIO II trial. Int J Cardiol2017;243:121–5.

25. Windecker S, Asano T, Raber L, et al. TCT-49Two-year clinical outcome of everolimus-elutingbioresorbable scaffold vs. durable polymereverolimus-eluting metallic stent in patients withST-segment elevation myocardial infarction: re-sults of the randomized ABSORB ST-segmentelevation myocardial infarction–TROFI II trial.J Am Coll Cardiol 2016;68 Suppl:B20.

26. Kozuma K, on behalf of the ABSORB-JapanInvestigators. ABSORB Japan: 3-year clinical andangiographic results of a randomised trial evaluatingthe Absorb bioresorbable vascular scaffold vsmetallicDES in de novo native coronary artery lesions. Pre-sented at: EuroPCR 2017; May 16, 2017; Paris, France.

27. Gao R, on behalf of the ABSORB-ChinaInvestigators. Randomised comparison ofeverolimus-eluting bioresorbable vascular scaf-folds versus everolimus-eluting metallic stents inpatients with coronary artery disease: 3-yearclinical outcomes from ABSORB China. Presentedat: EuroPCR 2017; May 16, 2017; Paris, France.

28. Ellis SG, Kereiakes DJ, Stone GW, for theABSORB III Investigators. Everolimus-eluting bio-resorbable vascular scaffolds in patients withcoronary artery disease: ABSORB III trial 2-yearresults. Presented at: ACC 2017; March 20, 2017;Washington, DC.

29. Sotomi Y, Onuma Y, Collet C, et al.Bioresorbable scaffold: the emerging reality andfuture directions. Circ Res 2017;120:1341–52.

30. Onuma Y, Sotomi Y, Shiomi H, et al. Two-yearclinical, angiographic, and serial optical coherencetomographic follow-up after implantation of aneverolimus-eluting bioresorbable scaffold and aneverolimus-eluting metallic stent: insights fromthe randomised ABSORB Japan trial. Euro-Intervention 2016;12:1090–101.

31. Onuma Y, Serruys PW, Muramatsu T, et al.Incidence and imaging outcomes of acute scaffolddisruption and late structural discontinuity afterimplantation of the Absorb everolimus-eluting fullybioresorbable vascular scaffold: optical coherencetomography assessment in the ABSORB cohort BTrial. J Am Coll Cardiol Intv 2014;7:1400–11.

32. Waksman R, Lipinski MJ, Acampado E, et al.Comparison of acute thrombogenicity for metallicand polymeric bioabsorbable scaffolds: Magmarisversus Absorb in a porcine arteriovenous shuntmodel. Circ Cardiovasc Interv 2017;10:e004762.

33. VorpahlM,NakanoM,Perkins LEL, et al. Vascularhealing and integration of a fully bioresorbableeverolimus-eluting scaffold in a rabbit iliac arterialmodel. EuroIntervention 2014;10:833–41.

34. Nakazawa G, Finn A V, Vorpahl M, Ladich ER,Kolodgie FD, Virmani R. Coronary responses anddifferential mechanisms of late stent thrombosisattributed to first-generation sirolimus- andpaclitaxel-eluting stents. J Am Coll Cardiol 2011;57:390–8.

35. Serruys PW, Katagiri Y, Sotomi Y, et al. Arterialremodeling after bioresorbable scaffolds andmetallic stents. J Am Coll Cardiol 2017;70:60–74.

36. Gomez-Lara J, Radu M, Brugaletta S, et al.Serial analysis of the malapposed and uncoveredstruts of the new generation of everolimus-elutingbioresorbable scaffold with optical coherence to-mography. J AmColl Cardiol Intv 2011;4:992–1001.

37. Otsuka F, Pacheco E, Perkins LEL, et al. Long-term safety of an everolimus-eluting bioresorbablevascular scaffold and the cobalt-chromium XIENCEV stent in a porcine coronary artery model. CircCardiovasc Interv 2014;7:330–42.

38. Karanasos A, Simsek C, Gnanadesigan M, et al.OCT assessment of the long-term vascular healingresponse 5 years after everolimus-eluting bio-resorbable vascular scaffold. J Am Coll Cardiol2014;64:2343–56.

39. Puricel S, Cuculi F, Weissner M, et al.Bioresorbable coronary scaffold thrombosis:multicenter comprehensive analysis of clinicalpresentation, mechanisms, and predictors. J AmColl Cardiol 2016;67:921–31.

40. Ortega-Paz L, Capodanno D, Gori T, et al.Predilation, sizing and post-dilation scoring inpatients undergoing everolimus-eluting bio-resorbable scaffold implantation for prediction ofcardiac adverse events: development and internalvalidation of the PSP score. EuroIntervention2017;12:2110–7.

41. Nakatani S, Sotomi Y, Ishibashi Y, et al.Comparative analysis method of permanentmetallic stents (XIENCE) and bioresorbable poly-L-lactic (PLLA) scaffolds (Absorb) on opticalcoherence tomography at baseline and follow-up.EuroIntervention 2016;12:1498–509.

KEY WORDS bioresorbable coronaryscaffolds, coronary artery disease, stent,stent thrombosis

APPENDIX For additional OCT definitionsas well as supplemental tables, figures, andvideos, please see the online version of thisarticle.
































































































http://refhub.elsevier.com/S0735-1097(17)39671-7/sref24a












































































Documents

Mechanisms of Very Late Bioresorbable Scaffold …Manuscript received July 25, 2017; revised manuscript received September 6, 2017, accepted September 6, 2017. SEE PAGE 2345 ABBREVIATIONS