Intracoronary Optical Coherence Tomography and Histology at 1 Month and 2, 3, and 4 Years After Implantation of Everolimus-Eluting Bioresorbable Vascular Scaffolds in a Porcine Coronary Artery ModelClinical Perspective
An Attempt to Decipher the Human Optical Coherence Tomography Images in the ABSORB Trial
Background—With the use of optical coherence tomography (OCT), alterations of the reflectance characteristics of everolimus-eluting bioresorbable vascular scaffold (BVS) struts have been reported in humans. In the absence of histology, the interpretation of the appearances of the struts by OCT remains speculative. We therefore report OCT findings with corresponding histology in the porcine coronary artery model immediately after and at 28 days and 2, 3, and 4 years after BVS implantation.
Methods and Results—Thirty-five polymeric BVS (3.0×12.0 mm) were singly implanted in the main coronary arteries of 17 pigs that underwent OCT and were then euthanized immediately (n=2), at 28 days (n=2), at 2 years (n=3), at 3 years (n=5), or at 4 years (n=5) after implantation. All BVS-implanted arteries in these animals were evaluated by histology except for 5 arteries examined at 2 years with gel permeation chromatography to assess the biodegradation of the polymeric device. Fourteen arteries with BVS from an additional 6 pigs were examined by gel permeation chromatography at 1 (n=1), 1.5 (n=2), and 3 (n=2) years. Corresponding OCT and histology images were selected with the distal and proximal radiopaque markers used as landmarks. At 28 days, by OCT, 82% of struts showed sharply defined, bright reflection borders, best described as a box-shaped appearance. Histologically, all struts appeared intact with no evidence of resorption. At 2 years, by OCT, 60±20 struts were discernible per BVS with 80.4% of the strut sites as a box-shaped appearance. Despite their defined appearance by OCT, by histology, these structures appeared to be composed of proteoglycan, with polymeric material being at such low level as to be no longer quantifiable by chromatography. At 3 years, by OCT, recognizable struts decreased to 28±9 struts per BVS: 43.7% showed dissolved black box; 34.8%, dissolved bright box; 16.1%, open box; and 5.4%, preserved box appearance. Histology shows that connective tissue cells within a proteoglycan-rich matrix replaced the areas previously occupied by the polymeric struts and coalesced into the arterial wall. At 4 years, by OCT, 10±6 struts were recognizable as either dissolved black or dissolved bright box. In histology, these struts are minimally discernible as foci of low-cellular-density connective tissue. Relative to the prediction of histological type by OCT appearance, the preserved box appearance of OCT corresponds well with 2-year histology (86.4%), whereas the dissolved bright and black box appearances correspond to 3-year histology (88.0% and 90.7%, respectively). Struts indiscernible by OCT correspond to the integrated strut footprints seen at 4 years (100%).
Conclusions—Struts that are still discernible by OCT at 2 years are compatible with largely bioresorbed struts, as demonstrated by histological and gel permeation chromatography analysis. At 3 and 4 years, both OCT and histology confirm complete integration of the struts into the arterial wall.
Bioresorbable vascular scaffolds are a novel approach to the interventional treatment of coronary artery disease, providing short-term vascular scaffolding combined with drug-delivery capability while avoiding the long-term limitations of metallic drug-eluting stents. Such limitations of metallic drug-eluting stents include retardation of the growth of healthy endothelium over stent struts1,2 or endothelial dysfunction resulting in abnormal endothelial responses to acetylcholine.3,4 Permanent metallic stenting may also preclude surgical revascularization, jail side branches, prevent late luminal enlargement, and impair noninvasive imaging of coronary arteries with multislice computed tomography and magnetic resonance. Recently, an everolimus-eluting bioresorbable vascular scaffold (BVS; Abbott Vascular, Santa Clara, Calif) was tested in the first-in-humans ABSORB study with a series of 30 patients. In this trial, BVS demonstrated excellent long-term clinical results up to 2 years with a low major adverse cardiac event rate of 3.4%.5 The device consists of a backbone of poly-l-lactide (PLLA) coated with poly-D, l-lactide (PDLLA), which contains and controls the release of the antiproliferative drug everolimus (Novartis, Basel, Switzerland).
Editorial see p 2234
Clinical Perspective on p 2300
So far, it is unknown which diagnostic modality is most suitable to confirm complete bioresorption. In the ABSORB trial, multiple modalities of invasive imaging were used, for instance, intravascular ultrasound gray-scale with additional analysis of echogenicity and radiofrequency backscattering and optical coherence tomography (OCT). OCT was able to demonstrate serial changes in the optical properties of the struts over time, probably reflecting the bioresorption process.5,–,8 At 2 years, OCT analysis demonstrated that 34.5% of the struts were no longer discernible. Alterations of the struts in their optical appearance at follow-up was categorized into 4 subgroups: preserved box, open box, dissolved bright box, and dissolved black box.8 According to these subgroups, strut appearance in serial comparative analyses (7 BVS) changed from 6 months to 2 years: preserved box, n=0 and n=9; open box, n=143 and n=68; dissolved bright box, n=225 and n=185; and dissolved black box, n=56 and n=25, respectively.5,8 Although the absence of visible struts in OCT was considered a sign of complete bioresorption, there is no histological correlate to confirm this. Furthermore, the correlations of differential appearances by OCT and histology have not yet been investigated.
Here, we report the appearance of BVS on OCT with corresponding histology in a porcine coronary artery model immediately after and at 28 days, 2 years, 3 years, and 4 years after implantation. This study was initiated in 2005 as a preamble to the completed and currently ongoing clinical studies.
Experimental studies received protocol approval from the Institutional Animal Care and Use Committee and were conducted in accordance with American Heart Association guidelines for preclinical research and the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, 1996). Four Yorkshire landrace swine and 14 Yucatan miniswine were implanted via femoral access according to standard procedures with either a 1.1:1 (0, 28 days) or 1.2: 1 (>2 years) ratio of balloon to artery for implantation.9 Each pig received a single 3.0×12.0-mm BVS implant in 1, 2, or 3 of the 3 main coronary arteries. The prototypes of BVS systems used in the porcine study were of similar construct as those used in the ABSORB clinical trial but were slightly different in terms of chemical purity and crystallinity.
Thirty-five arteries implanted with BVS in 17 pigs were examined by OCT immediately after the procedure (4 scaffolds), at 28 days (4 scaffold), and at 2 (7 scaffolds), 3 (8 scaffolds), and 4 years (12 scaffolds) after implantation. Of these, 30 arteries were subsequently examined by histology immediately after the procedure (4 scaffolds), at 28 days (4 scaffolds), and at 2 (2 scaffolds), 3 (8 scaffolds), and 4 years (12 scaffolds) after implantation (Figure 1). At 2 years, 5 BVS-implanted arteries in 2 animals were designated for quantification of polymer degradation by gel permeation chromatography (GPC) analysis following in vivo evaluation by OCT. In addition to these, 14 BVSs implanted in 5 animals were evaluated for polymer degradation at 1 (3 scaffolds), 1.5 (6 scaffolds), and 3 (5 scaffolds) years after implantation. To report the inflammatory response induced by BVS, additional histological data were included in this report from 7 Yucatan minipigs with 14 BVSs euthanized at 1 month, 7 Yucatan pigs with 11 BVSs at 6 months, 4 Yucatan minipigs with 7 BVSs at 24 months, and 2 Yucatan minipigs with 3 BVSs at 48 months.
Evaluation by OCT was performed with a commercially available OCT system (M2 System, LightLab Imaging, Westford, Mass). An OCT catheter (Helios proximal occlusion catheter) was initially advanced distal to the area of interest over a conventional coronary guidewire, which was then replaced with the OCT imaging wire (ImageWire). The OCT catheter was withdrawn manually proximal to the implanted segment; the balloon was inflated at low pressure; crystalloids were used to clear the blood; and the wire containing the imaging core was withdrawn at 1 mm/s. During image acquisition, coronary blood flow was replaced by continuous flushing of Ringer lactate at 0.5 to 1.0 mL/s with a power injector (Mark V ProVis, Medrad, Inc, Indianola, Pa). The highly compliant occlusion balloon remained inflated proximal to the lesion at 0.5 or 0.7 atm. Cross-sectional images were acquired at 15.6 frames per second. Immediately after OCT, all pigs were humanely euthanized. For qualitative and quantitative analyses of the OCT, images were selected at 1-mm intervals in the implanted segment, which was defined as the region between the metallic radiopaque markers. The struts sites were categorized according to the 4 morphological subcategories that have been applied in the clinical study.5,8 The first subset (preserved box) is described as a box appearance with sharply defined borders with bright reflection, and the strut body shows low reflection. The second subset is characterized as an open box (ie, luminal and abluminal “long-axis” borders thickened with bright reflection and short-axis borders that are no longer visible at follow-up). The third subset is categorized as a dissolved bright box (ie, partially visible bright spot with poorly defined contours and no box-shaped appearance). The fourth subset is a dissolved black box (ie, black spot with poorly defined contours, often confluent but with no box-shaped appearance). The other OCT end points related to the interaction of the struts and vessel wall (eg, apposition and coverage).5,8 Incomplete strut apposition was defined as clear separation of 1 strut from the vessel wall, with evidence of space (black appearance without light reflection) between the strut and vessel wall. Coverage of the strut by tissue was categorized as complete, incomplete, or not visible.5,8
Hearts were explanted from the thoracic cavity and flushed using pressure perfusion with saline followed by 10% buffered formalin in preparation for histology. Implanted arteries were carefully dissected from the heart, routinely processed, embedded in methyl methacrylate, and sectioned in duplicate at 5 μm according to published procedures.9 Three sections were taken from an implanted segment: the proximal section at 2 mm distal to the proximal metallic marker, the middle section at 6 mm distal to the proximal marker, and the distal section at 2 mm proximal to the distal marker. Sections were stained with hematoxylin and eosin and elastic van Gieson (2 and 3 years) or Movat pentachrome (4 years). Duplicate sections from selected arteries at 2, 3, and 4 years also were stained with Alcian Blue, Masson trichrome, and Von Kossa histochemical stains for the detection of acid mucopolysaccharides (proteoglycans), connective tissue, and calcification, respectively, and/or were immunohistochemically stained for detection of smooth muscle actin (SMA) to identify smooth muscle cells.1 In pigs that underwent both OCT and histology, the struts were counted, and these struts were classified according to histological findings, including the presence of calcification and cartilaginous metaplasia. Additionally, qualitative characterization of the vascular response around the BVS was performed (eg, inflammation).
To correlate OCT classification and histology, 1 observer (Y.O.) aware of the histological image selected the matched OCT cross sections at 2, 3, and 4 years using the distance from markers and anatomic landmarks such as side branches. To obtain an even number of struts in histology and OCT, the observer indicated the potentially coregistered strut sites. These matched OCT images with suggested strut sites were sent to 2 independent OCT analysts (H.M.G.-G., N.G.). Although an observer for matching purposes was unblinded to both OCT and histology, the 4 other observers (2 for histology and 2 for OCT) were blinded to the other techniques. Histological analysis was performed unblinded to the time interval after implantation, whereas OCT images were analyzed by the observers blinded to the time interval.
To characterize the degradation of polymer, the weight-average molecular weight (Mw), number-average molecular weight (Mn), polydispersity index (PDI), and content of polymer in the BVS were analyzed by GPC before stent deployment (T0) (n=5) and at 1, 1.5, 2, and 3 years after implantation. The polymer was extracted with chloroform, and the tissue residue was then removed by filtration with deactivated glass wool packing. Measurements were performed with 2 PLgel Mixed-D columns (Polymer Laboratories, Stretton, UK) equipped with a refractive index detector. The Mw, Mn, and PDI of polymer were calculated from the calibration curves obtained for polystyrene standards (Mw range, 3.6 to 380 kDa), whereas the amount of polymer was calculated from the calibration curves of PLLA standards.
Categorical data are expressed as percentages, and continuous variables are presented as mean with SD. Generalized estimating equations modeling was performed10 with SAS software version 9.1 (SAS Institute, Cary, NC) to take into an account the clustered nature of >1 scaffolds analyzed from same pigs, which might result in unknown correlations among measurements within these scaffold clusters. A log link function was used for the categorical and continuous variables; a logit link function was used for the binary variables. An exchangeable structure working correlation matrix was used for the analyses. All statistical tests were 2 tailed, and a value of P<0.05 was considered statistically significant. Agreement between observers was measured by calculating the Cohen κ value, which takes into account the proportion of agreements occurring by chance.
Examples of strut appearances assessed with OCT are shown in Figure 2. Immediately after implantation, all struts had a preserved box appearance (Figure 3) and were well apposed to the vessel wall. In 1 of the 2 pigs euthanized immediately after implantation, the BVS was deployed after heparinization but without antiplatelet therapy. In this animal, by OCT, a strut of 1 BVS was covered with highly reflective structures without shadowing, suggesting the presence of a white thrombus (Figure 3A and 3E).
The OCT findings in porcine animals at 28 days and at 2, 3, and 4 years are summarized with the human observation up to 2 years in Table 1, and representative OCT images are provided in Figures 3 through 7. At 28 days, 82% of struts presented a preserved box appearance, and 18% had an open box appearance (Figure 3). At 2 years, on average, 60±20 struts were discernible per BVS device. The endoluminal lining of the implanted segment was smooth and circular (Figure 4). Four fifths of the struts showed a preserved box appearance, 17.2% showed a dissolved black box, and only a few struts (2.4%) demonstrated open box appearance.
At 3 years, the number of recognizable struts decreased to 28±9 struts per BVS: 43.7% of discernible struts showed a dissolved black box appearance, 34.8% showed a dissolved bright box appearance, 16.1% showed an open box appearance, and 5.4% remained preserved box appearances (Figures 5 and 6). At 4 years, only 10±6 struts per BVS were recognized by OCT: 51.2% of struts were classified as dissolved bright box and 48.8% as dissolved black box appearance (Figure 7). At 2, 3, and 4 years, all visible struts were fully apposed and covered by tissue.
The qualitative analysis of histology is summarized in Table 2. Immediately after implantation, the indentation of the media by the struts was visible (Figure 3C and 3G). At 28 days, the struts were completely sequestered from the lumen by a thin, fibromuscular neointima and had well-defined and squared edges. The polymeric material was not stained with Alcian Blue (Figure 3K, 3N, and 3O).
At 2 years, 6 histological images containing 76 struts were evaluated. All struts were morphologically identical as open acellular regions with well-defined borders. In such strut footprints, these “preserved box” structures likewise had discrete borders and were composed of faintly hyaline material that stained positively with Alcian Blue, indicating that they were composed of acid mucopolysaccharides (proteoglycans; Figure 4E). Because of the high water content of the matrix replacing the polymer, there was likely processing-induced swelling that made these footprints appear falsely larger by histology compared with OCT. Minimal calcification was present around all struts (100%, 76 of 76; Figure 4H). Preexisting struts were completely sequestered within a fibromuscular neointima, with no to minimal inflammatory cells (macrophages, multinucleated giant cells) being immediately associated with the struts (Figure 4D).
At 3 years, 23 histological images with 195 struts were analyzed. One hundred eleven strut footprints were recognized as a defined accretion of hyaline (proteinaceous) material separated by extracellular matrix (proteoglycans) and cells (Figure 5). The hyaline material was identified as nonfibrillar glycoprotein by transmission electron microscopy (Figure 5H). The majority of cells integrating into these sites did not stain positively for SMA (Figure 5G). In this histological classification, 74.8% of struts (83 of 111) showed calcification (Figure 5F). The remaining 84 areas previously occupied with struts were recognized as regions without hyaline material but with connective tissue with low to moderately cellular density, which indicated complete, benign involution of the struts into the arterial wall (Figure 6). In this morphology, cells replacing preexisting struts were irregularly arranged, and minimal calcification that was partially to completely circumferential around preexisting struts was observed around 36.9% of struts (31 of 84). For both morphologies, just as at 2 years, macrophages and multinucleated giant cells were noted only occasionally.
At 4 years, 36 histological specimens with 239 strut footprints were analyzed. Although 5 strut footprints remained as a defined accretion of hyaline material similar to what was observed at 3 years, the rest of the strut footprints were minimally discernible in histology as regions of integrated connective tissue (Figure 7). Again, as for 3 years, only a paucity of cells within the strut footprints stained positively with SMA (Figure 7G). One hundred sixty-two struts presented as a circumscribed area of dense connective tissue with low cellularity in which cells are arranged in circumferential pattern. Of these, calcification was present in 16.0%, whereas cartilaginous metaplasia was found in 4.4%. Calcification when observed was minimal, typically being only scant linear foci associated with the connective tissue replacing struts. The remaining 72 strut footprints were recognized as poorly circumscribed regions of dense connective tissue with moderate to low cellularity in which cells were not regularly arranged. In this morphology, 16.7% of struts showed calcification (Figure 7H), and 4.2% demonstrated cartilaginous metaplasia. Overall, from 3 to 4 years, the frequency of strut footprints with calcification decreased significantly from 58.5% (114 of 195) to 17.2% (41 of 239; P<0.001).
Relative to inflammatory responses, the percentages of struts with granuloma were 13.8±25.1%, 3.96±6.93%, and 0.37±1.11% at 1, 6, and 24 months, respectively; the percentages of struts with giant cells were 34.8±20.5%, 14.7±18.9%, and 1.6±2.9%, respectively. At 3 and 4 years, no granuloma/ giant cells were observed in the area previously occupied with polymeric struts.
Comparison Between OCT and Histology
The comparison between OCT and histological classification at 2, 3, and 4 years in matched struts is summarized in the Table 3. Of 510 struts analyzed by histology, 49 struts sites were excluded. Sixteen struts sites were not visualized with OCT because they were located out of range, and 33 struts were of poor imaging quality. The κ value of interobserver agreement was 0.58. In histology, the struts were classified into 5 categories, as shown in Figure 2.
In the strut footprints with histological category 1 (open acellular regions with well-defined borders), which is uniformly observed by histology at 2 years after procedure, 91.9% of strut footprints were recognized as the preserved box appearance by OCT (Figure 4). In 4.8%, the strut sites were recognized as dissolved black box owing to blurring of OCT image by artifacts. In the strut footprints with histological category 2 (a region with hyaline material separated by extracellular matrix and cells, which was observed frequently at 3 years and rarely at 4 years), 81.5% of strut footprints were recognized as dissolved black boxes, whereas 8.3% and 7.4% of strut footprints were classified as dissolved bright and preserved box, respectively. In particular struts with preserved box appearance, histology showed that the strut footprints coalesced with the surrounding neointimal tissue, with demarcating lines of calcification delineating preexisting strut locations (Figure 5). In the strut footprints with histological category 3 (a region without hyaline material but with low to moderately cellular connective tissue observed at 3 years), 80.7% and 15.7% of strut footprints were classified as dissolved black and bright boxes, respectively. In struts with histological categories observed at 4 years, regions of strut sites were poorly discernible in OCT (category 4, 90.9% indiscernible; category 5, 92.3% indiscernible).
When predicting histological type from OCT classification, the preserved box corresponds to an open acellular region with hyaline material at 2 years (category 1) in 86.4% and less frequently to a region with accretion of hyaline at 3 years (category 2) in 12.1%. The open box appearance was observed uncommonly (8 struts, 1.7%) in this animal experimental model. The dissolved bright box appearance corresponds to 3-year histology: category 2 in 36% and category 3 in 52%. The dissolved black box also corresponds to 3-year histology but more frequently with category 2 (51.5%) than category 3 (39.2%). Indiscernibility with OCT accurately predicts 4-year histology (31.4% in category 4 and 68.6% in category 5).
GPC analysis demonstrated an expected decrease in the Mw, Mn, PDI, and content of polymer in the BVSs over time. The Mw and Mn of PLLA dropped by ≈100%, from 200 and 100 kDa at T0 to unquantifiable at 2 years after implantation, and the polymer mass loss increased by ≈20%, ≈60%, and ≈90% at 1, 1.5, and 2 years, respectively, becoming undetectable at 3 years (Figure 8). In addition, the polymer peak on the chromatograms obtained from BVS samples at 1, 1.5, and 2 years after stent implantation shifted to the right (lower Mw range) relative to the polymer peak on the chromatogram from samples at T0, consistent with polymer degradation. In 24-month chromatograms, only a very small peak with a magnitude (S:N, ≈10:1) close to the limit of quantification (0.3 mg/mL of PLLA, ie, 5.7% of the gravimetric weight [5.3 mg] of the stent at T0) was observed. It is likely that the observed peak is attributed mainly to the trace tissue species (extracted together with the polymer), which may be superimposed on the polymer peak on the significant shift of the latter to the right as the polymer has degraded to very low molecular weight. Taking into consideration the limit of quantification of the GPC method used, the measurement uncertainty at the limit of quantification levels (±20%), and the possible constant systematic error caused by the interference of the tissue species, which can give rise to a false signal, the BVS can be considered to be fully resorbed at 24 months after implantation. Complete degradation was confirmed at 3 years, with polymer being undetectable (limit of detection, 0.1 mg/mL of PLLA, ie, 1.9% of the gravimetric weight of the stent at T0 [5.3 mg]). Therefore, the structures visible by OCT and histology at 2, 3, and 4 years represent locations of bioresorbed BVS struts.
In the present study, we describe the preclinical experience using BVSs in a porcine coronary artery model. Twenty-eight days after BVS implantation, on OCT, 82% of the struts had the preserved box appearance, which corresponded to intact polymeric struts. Two years after BVS implantation, OCT images demonstrated that 80% of the struts still had a preserved box appearance. Nonetheless, the corresponding histological images and polymer degradation assessment with GPC showed that the polymeric struts could, for all intents and purposes, be considered to be fully resorbed and replaced by proteoglycan. Three years after BVS implantation, OCT images demonstrated that only 5% of the struts retained a preserved box appearance, whereas 44% had evolved to a dissolved black box appearance. In addition, at the 3-year follow-up, we observed on OCT dissolved black boxes that, on histology, reflected infiltration of the sites of preexisting struts by connective tissue cells. At 3 years, the absolute count of discernible struts was reduced by ≈50%, and the no-longer-discernible strut footprints on OCT seem to correspond in histology to complete integration into the surrounding arterial wall. Furthermore, at 4 years, only one sixth of the struts were discernible in OCT with a dissolved black or dissolved bright box appearance, which were minimally discernible in histology. Relative to prediction of histological type by OCT appearance, the preserved box appearance of OCT corresponds well with 2-year histology (86.4%), whereas dissolved bright and black box appearance corresponds well to 3-year histology (88.0% and 90.7%, respectively). Indiscernible struts in OCT can confirm the integrated strut footprints seen at 4 years (100%).
Histological Response After BVS Implantation
The present study suggests the following temporal progression after implantation of this PLLA/PDLLA BVS as viewed with histology. First, the polymeric struts are covered by a fibromuscular neointima. Second, struts are replaced by proteoglycan matrix, which corresponds to resorption demonstrated by GPC analysis. Third, there is inspissation of the proteoglycan matrix, and connective tissue cells infiltrate the region of the preexisting struts. Fourth, the areas of preexisting struts become fully integrated into the arterial wall and are difficult to discern. Strictly speaking, the process of “bioresorption” finishes at the second phase. Subsequent phases represent an “integration” process of the struts within the arterial wall; ultimately, the locations of preexisting struts become indiscernible on histology.
Of note, a thin rim of calcification around a majority of struts at the tissue interface was observed on histology at 1 month and at 2 and 3 years. This precipitation of calcium phosphate might be the result of benign, localized drop in pH caused by acidic polymer degradants at the strut-tissue interface.11 However, this is not unique to BVS or PLA but has been observed with bioprosthetic and synthetic materials.12,13 In the present study, the total amount of calcification is minimal and is demonstrated to be largely resolved at 4 years after implantation.
Assessment of the Bioresorption Process by OCT
In the first-in-humans trial using the BVS, the OCT classification of the 4 strut appearances was developed to characterize the process of bioresorption.5,8 Immediately after implantation, the polymeric struts appeared on OCT as clear-cut boxes lying on the vessel wall rather than being embedded (Figure 3). The disappearance of this box appearance was considered to be the initial stages of the bioresorption process because it was no longer seen at 6 months.5,8 The first optical change of this box appearance in the ABSORB trial was the “opening” of the extremities of the box in its short axis, which was considered a first sign of biochemical or histological alteration (Figure 1), observed in 34% of struts at 6 months and still in 14% at 2 years. The dissolved black box appearance and dissolved bright box appearance were considered to reflect further stages of bioresorption and vessel wall integration. The inability to discern the strut footprints on OCT was interpreted as a sign of complete bioresorption. At 2 years, 36.3% of the struts were indiscernible by OCT in humans. So far, no attempt has been made to interpret these human OCT observations by correlating OCT and histological observation in an animal model.
In contrast to what was initially thought after the first-in-humans BVS trial and according to the present animal study, the preserved box appearance does not exclude the resorption of polymeric struts. The polymer was only no longer accurately quantifiable by GPC at 2 years after BVS implantation, but the structures observed with OCT and histology represented tissue-derived regions corresponding to the location of the former BVS struts. In other words, the preserved box appearance with optical translucency is compatible with complete polymer dissolution.
Assessment of Integration Process by OCT
Although OCT might not be sensitive enough to evaluate the degradation process of the polymer, it may provide more detailed information on the integration process. In the present study, the transition from the preserved box appearance (at 2 years) to the full disappearance of the strut footprints on OCT or the presence of dissolved black box (at 3 years) appears to be related to the replacement of the preexisting struts by proteoglycan and dense connective tissue, respectively. Although the demarcation of the preexisting struts becomes undetectable with OCT imaging, by histology, the struts ultimately coalesce with the arterial wall. This suggests that changes in strut appearances on OCT mirror the integration process of the struts. In a few cases, a demarcating line of calcification seems to be the sole histological remnant of the preexisting strut, and it is still unclear whether this thin line of calcification would be seemingly detectable on OCT (Figure 7).
In humans, we have reported the changes in ultrasound backscattering using gray-scale assessment and backscattering of radiofrequency signal.5,7–8 We have also demonstrated the disappearance of the echogenic signature of the polymeric struts with time.14 The ultrasonic change in signal was characterized by the disappearance of brightness (gray-scale) and pseudodense calcium (virtual histology), whereas with light assessment, the black appearance (resulting from the homogeneity of the polymeric material) of the struts shortly after stent implantation is progressively replaced by light backscattering undistinguishable from the backscattering of tissue background.
Recently, Jiang et al15 demonstrated in in vitro and in vivo studies that a strong correlation exists between the rate of material degradation and the degree of inflammatory response to implanted material. In Swiss Webster mice with polymeric disks implanted into the peritoneum, the PLLA disk, which has longer degradation time compared with the mixed polymer of PLLA and PEG, has the least associated inflammation. However, in a rat study by Polimeni et al,16 when the PLA device implanted subcutaneously came in contact with bones, the polymer triggered a foreign body reaction, including multinucleated giant cells, macrophages, and lymphocytes. Experts on biodegradable polymer have repeatedly emphasized the importance of the locus of the implant on its rate of degradation and the inverse relationship between rates of biodegradation and the presence of an inflammatory process.15 In the present histological study, relative to the degree of completeness of bioresorption of BVS, there was no significant inflammatory response associated with BVS at 2, 3, or 4 years, although a likely short-term inflammatory response during the active resorption phase of the scaffolds was not evaluated as part of this study.
Differences Between Human and Porcine Coronary Artery
At the 2-year follow-up, the distribution of OCT strut appearances observed in the porcine animal model differed from the results of the first-in-humans trial (Table 1). More specifically, the open box appearance and dissolved bright box appearance observed in 21.7% and 62.0% of struts in the ABSORB trial were scarcely observed at 2 years in the present animal study (open box, 2.4%; dissolved bright box, 0%). Furthermore, in the ABSORB trial, the number of discernible struts per device was 38±30 at 2 years, whereas in the present animal model, 60±20 strut voids per device were still discernible at 2 years. This suggests a slower integration process of the strut voids filled with proteoglycans in porcine arteries than in human arteries.
One possible explanation is the absence of underlying atherosclerotic plaque with or without inflammatory component in the implanted lesion of the animal model. The vascular response of aged, human atherosclerotic arteries is distinct from that of juvenile, healthy animal arteries. In some human cases, the deployed polymeric struts are possibly apposed against highly inflammatory atherosclerotic plaque with activated macrophages and metalloproteinase activity. In a single cross section of OCT, heterogeneity of strut appearance (4 categories) is not uncommon.
It has been reported in the ABSORB trial that the lesion morphology of implanted segments affected the degree of “late recoil” seen at 6 months with the BVS.14 Calcified plaques resulted in significantly less late recoil (0.20±1.54 mm2 and 1.97±22.2%) than fibronecrotic plaques (1.03±2.12 mm2 and 12.4±28.0%; P=0.001 and P=0.001, respectively) or fibrocellular plaque (0.74±1.48 mm2 and 8.90±19.8%; P=0.001 and P=0.001, respectively). This suggests that the underlying fibronecrotic or fibrocellular plaque might have a determinant role in early alteration of polymeric struts in the first 6 months, which seems to evolve into a more favorable integration process at 2 years, as reflected by the striking difference between the OCT appearance in human at 6 months and 2 years.
Another possible explanation for the slower integration in the present porcine study than in the ABSORB trial is presumably the slower hydrolysis time of the polymeric strut constructions in this study compared with the BVS in the ABSORB trial. It is known that polylactide polymer with greater monomer content has a faster degradation rate. Polylactic acid degrades by hydrolysis, and this process is catalyzed at early stages of degradation by the presence of carboxyl moieties. This autocatalytic process is initiated either from the ends of each chain or from the presence of free residual monomer in the polymer matrix. The polymeric BVS coronary system used in the first-in-humans trial has greater monomer content in its polymer than the device used in the present animal study. These facts may explain the faster integration process observed in the first-in-humans trial than in the present animal study.
The limitations of this study are the following. The number of specimens examined was small. The sampling rate of histological slices from each scaffold is insufficient to describe the 3-dimensional distribution of strut bioresorption/ integration. In addition, as discussed above, differences between the healthy porcine coronary artery and diseased human coronary artery could limit the ability to generalize the hypothetical concepts put forward in this report.
The present porcine animal study potentially elucidates the histological responses after implantation of the BVS: bioresorption and integration. OCT might be more sensitive for assessing the integration process rather than the polymer alteration. Ultimately, the absence of the BVS device footprint on OCT suggests complete integration of the struts into the arterial wall.
J.C. Powers and Drs Perkins, Kamberi, and Rapoza are employees of Abbott Vascular. Dr Virmani has been on the advisory boards for Medtronic Vascular and Valve, Abbott Vascular, and BVS; has been a consultant to W.L. Gore and Atrium Medical Corp; and has received honoraria from Medtronic, Abbott Vascular, W.L. Gore, Terumo, and Atrium Medical Corp. The other authors report no conflicts.
We thank Jim Yu for his intellectual input into the statistical analyses.
- Received December 9, 2009.
- Accepted July 26, 2010.
- © 2010 American Heart Association, Inc.
- Finn AV,
- Joner M,
- Nakazawa G,
- Kolodgie F,
- Newell J,
- John MC,
- Gold HK,
- Virmani R
- Hofma SH,
- van der Giessen WJ,
- van Dalen BM,
- Lemos PA,
- McFadden EP,
- Sianos G,
- Ligthart JM,
- van Essen D,
- de Feyter PJ,
- Serruys PW
- Serruys PW,
- Ormiston JA,
- Onuma Y,
- Regar E,
- Gonzalo N,
- Garcia-Garcia HM,
- Nieman K,
- Bruining N,
- Dorange C,
- Miquel-Hebert K,
- Veldhof S,
- Webster M,
- Thuesen L,
- Dudek D
- Bruining N,
- Tanimoto S,
- Otsuka M,
- Weustink A,
- Lighthart J,
- de Winter S,
- van Mieghem C,
- Nieman K,
- de Feijter PJ,
- van Domburg RT,
- Serruys PW
- García-García HM,
- Gonzalo N,
- Pawar R,
- Kukreja N,
- Dudek D,
- Thuesen L,
- Ormiston JA,
- Regar E,
- Serruys PW
- Ormiston JA,
- Serruys PW,
- Regar E,
- Dudek D,
- Thuesen L,
- Webster MW,
- Onuma Y,
- Garcia-Garcia HM,
- McGreevy R,
- Veldhof S
- Vorpahl M,
- Finn AV,
- Nakano M,
- Virmani R
Alterations of the reflectance characteristics of the everolimus-eluting bioresorbable vascular scaffold struts have been reported in humans. However, in the absence of histology, the interpretation of the appearances of the struts by optical coherence tomography (OCT) remains speculative. In the present animal study, the bioresorbable vascular scaffold (3.0×12.0 mm) was implanted in the coronary arteries of pigs that underwent OCT and then were examined by histology immediately or at 28 days or 2, 3, or 4 years after implantation. To assess the biodegradation, gel permeation chromatography was performed. At 2 years, OCT showed that 80.4% of the strut sites had a box-shaped appearance. On histology, these structures were composed of proteoglycan. By gel permeation chromatography, the polymeric material of bioresorbable vascular scaffold was no longer quantifiable, suggesting complete bioresorption. At 3 years, by OCT, most struts showed dissolved appearances (dissolved black, 43.7%; dissolved bright, 34.8%). Histology showed connective tissue cells within a proteoglycan-rich matrix, signifying the beginning of the integration process. At 4 years, only a few struts were recognizable by OCT, and on histology, struts indiscernible by OCT are also minimally detectable, which suggests complete integration. In summary, despite their defined appearance by OCT, struts at 2 years were largely bioresorbed, and at 4 years, struts indiscernible by OCT corresponded well to the completely integrated struts. OCT might be more sensitive to assess the integration process rather than the polymer alteration. This animal study will potentially serve as a guide for interpretation of OCT after bioresorbable vascular scaffold implantation in the clinical setting.