Failure of a Novel Balloon-Expandable γ-Emitting (103Pd) Stent to Prevent Edge Effects
Background Balloon-expandable β-particle–emitting (32P) stents inhibit within-stent neointimal hyperplasia but induce lumen narrowing beyond the stent margins, ie, the so-called “edge effects.”
Methods and Results We prospectively investigated the performance of novel stents impregnated with the γ-emitting isotope 103Pd, designed to reduce edge effects, in 24 rabbits. The stents had a length of 18 mm and were mounted on 20-mm-long delivery balloons for deployment. Angiograms were obtained immediately and 1 month after direct implantation of control and 1-, 2-, and 4-mCi 103Pd stents into the iliac arteries without predilatation or postdilatation. Late lumen loss was measured with quantitative angiography. Neointimal hyperplasia and vascular remodeling were evaluated by histomorphometry. Late lumen loss was inhibited within 103Pd stents (control 0.18 mm, 1 mCi 0.08 mm, 2 mCi 0.05 mm, and 4 mCi −0.03 mm, P<0.05 all activities versus control). Conversely, late lumen loss occurred at the edges of 103Pd stents, correlating with areas of high balloon/artery ratios and vessel overstretch injury. Edge effects were primarily due to neointimal hyperplasia but were also caused by negative vessel remodeling at high stent activities.
Conclusions Edge effects after implantation of radioisotope stents can occur independently of the isotope chosen for stent impregnation.
Received May 25, 2001; revision received August 17, 2001; accepted August 20, 2001.
It was previously demonstrated that smooth muscle cell proliferation and neointima formation are markedly inhibited within β-particle–emitting stents in vitro and in a rabbit restenosis model.1–3 Clinical studies have shown an effective inhibition of late loss due to neointima formation within balloon-expandable β-particle–emitting stents (32P) after treatment of patients with coronary artery disease.4–6 These studies also showed, however, that β-particle–emitting stents can induce “edge effects” (restenosis at the stent ends). A number of factors have been suggested to be responsible for edge effects: balloon injury beyond the irradiation zone (geographic miss), radiation doses too low to prevent restenosis (dose falloff at stent ends), inhibition of cell migration within the stents resulting in accumulation of neointimal cells at the stent ends, rheological or inflammatory stimuli, etc. Early studies with coronary 32P stents suggested that edge effects may be caused by the dose falloff at the stent margins.5 32P stents with higher activities, however, even increased the incidence of edge restenosis.6
γ-Emitting isotopes provide a shallower dose gradient in tissue than β-particle emitters and are therefore thought to be more effective in overcoming edge effects caused by a rapid dose falloff at the stent margins. The purpose of this study was to evaluate the performance of a novel balloon-expandable stent impregnated with 103Pd, a weak γ-radiation–emitting isotope, in a rabbit iliac artery overstretch model.
Radioisotope Characteristics, Stent Manufacture, Shielding, and Dosimetry
103Pd (half-life 17 days) decays by electron capture and emits low-energy x-rays (21 keV) and Auger electrons. The dose rate constant for 103Pd is 23 mrem/h per mCi at 10 cm. Homogeneous 103Pd coating of BX stents (Cordis) and activity measurements were performed by MDS Nordion in Kanada, Ontario. Dosimetry calculations comparing dose rates at the stent edges of 103Pd stents with those of 32P stents were performed in Karlsruhe, Germany (K.S.). In all coated stents, the measured activity was ±10% of true activity, axial dose uniformity was ±15%, and the environmental isotope retention value was >98% (<2% activity washoff in ultrasonic bath). The 103Pd stents were shielded with a lead-impregnated acrylic cylinder. Radiation protection for manipulating the stents was achieved with lead-impregnated surgical gloves. Reading of the skin radiation dose of the animals after stent implantation indicated no significant radiation dose outside the body.
Animal Care and Surgical Procedure
All experiments were performed in accordance with the guidelines for animal research established by the American Heart Association and were approved by the local animal care committee. Female New Zealand White rabbits (n=24) weighing 2.7 to 3.2 kg were used for the study. Anesthesia was performed with acepromazine (0.1 mg/kg), ketamine (35 mg/kg), and xylazine (3 mg/kg) SC. Both femoral arteries were exposed, and a 4F pediatric sheath was inserted into each artery. Heparin 500 U IA was given via the sheath, and retrograde angiograms were obtained. Fifteen nonradioactive control BX stents (length 18 mm) and 33 BX stents impregnated with 1.1 (n=11), 2.2 (n=12), and 4.04 (n=10) mCi of 103Pd premounted on a balloon delivery system (balloon length 20 mm) were implanted into both iliac arteries. The 3.0-mm stent delivery balloon was inflated to 8 atm under fluoroscopic control. Control stents and radioactive stents were deployed opposite to each other in the common iliac artery without predilatation or postdilatation of the artery. All rabbits received aspirin 40 mg PO every 3 days for 1 month.
The angiographic films were analyzed with a quantitative coronary analysis program (Medis). Reference diameters and lumen diameters (proximal to, within, and distal to stents) of the iliac arteries were measured before and after stent implantation and at the 1-month follow-up. Balloon-to-artery (B/A) ratios were measured proximal to, within, and distal to the stents. Late lumen loss was calculated as minimal lumen diameter of the artery poststent minus minimal lumen diameter follow-up. To compare the edge effects of the proximal with those of the distal stent ends, the data were corrected for vessel tapering of the proximal to the distal end of the iliac artery.
Tissue Collection and Fixation
The rabbits were killed by a lethal dose of KCl and sodium pentobarbital (120 mg/kg) after 1 month. The abdominal aorta was cannulated, and the iliac arteries were flushed with physiological saline solution for 3 minutes. For in situ pressure fixation, the iliac arteries were infused with 10% formalin at 100 mm Hg for 10 minutes. The arterial specimens were dehydrated in graded alcohol solutions and were then embedded in methylmethacrylate and serially sectioned from the proximal to the distal end. The stented cross sections were stained with metachromatic stains, and the proximal and distal edges with hematoxylin-eosin and elastin stains.
The neointimal cross-sectional areas were measured with computer assistance with a compound microscope integrated to a digitizing tablet through a drawing tube attachment for histomorphometric analysis. Sections were graded semiquantitatively for arterial wall injury (score 0 to 3), inflammation, intimal fibrin deposition, density of intimal smooth muscle cells, and adventitial fibrosis. Neointimal area was expressed as percentage of lumen area. Vessel area was defined as the cross-sectional area encompassed by the adventitia.
The data are expressed as mean±SD and were compared by use of the StatView software package. The paired t test was applied for comparisons of the stents. Univariate regression analysis was performed to compare B/A ratios with angiographic late lumen loss.
In-stent lumen loss was inhibited in a dose-dependent fashion within 103Pd stents compared with nonradioactive control stents. The results for each stent activity are summarized in Figure 1. The lumen diameter at the distal stent edge was significantly decreased with radioisotope stents (control 1.5±0.3 mm, 1 mCi 1.1±0.3 mm, 2 mCi 1.0±0.2 mm, and 4 mCi 0.8±0.3 mm; P=0.002, 4 mCi versus control). B/A ratio, however, was also higher at the distal stent edge (1.3±0.1) than at the proximal stent edge (1.0±0.1, P<0.001). Figures 2 and 3⇓ depict the correlation between B/A ratios and late lumen loss after 1 month (control and 103Pd stents). Figure 4 shows the angiogram of the iliac arteries at 1-month follow-up with edge effects after the implantation of the 103Pd stents.
Histological Assessment and Morphometry
Arteries with nonradioactive control stents showed healing responses different from those with 103Pd stents at 1, 2, and 4 mCi after 1 month. The control stents were covered with mature neointima, whereas the 103Pd stents with activities of 2 and 4 mCi were covered only with fibrin-platelet deposits without significant endothelialization. Medial necrosis was present after implantation of 2- and 4-mCi stents, but rarely after use of 1-mCi 103Pd stents or control stents. Injury and inflammation scores were generally higher in arteries after implantation of 2- and 4-mCi 103Pd stents compared with 1-mCi 103Pd or control stents. Injury scores and lumen areas of the vessels beyond the stent edges are summarized in the Table. All vessels tapered from the proximal to the distal stent end (vessel area 3.0±1.0 versus 1.3±0.6 mm2, P<0.001). Shrinkage of the vessel (negative remodeling), however, was noted only at the distal stent edges with high stent activities (vessel area in mm2: control 1.4±0.5, 1 mCi 1.44±0.4, 2 mCi 1.48±0.5, 4 mCi 1.0±0.4, P<0.05 for 4 mCi versus control). Edge effects were predominantly due to neointimal thickening present mainly at the distal stent edge compared with the proximal edge, but edge effects also increased with the activity of the stent because of vessel shrinkage (luminal occlusion: control 12%, 1 mCi 18%, 2 mCi 31%, 4 mCi 40%, P<0.001 for 2 and 4 mCi versus control).
Dose-Rate Comparison Between 103Pd Stents and 32P Stents
Edge effects have previously been reported to occur with 32P stents. Activity levels of 3 mCi for a 103Pd stent and 15 μCi for a 32P stent were chosen to compare dose-rate distribution between the 2 stents of equal BX velocity design (Figure 5). Dose falloff is more shallow with the 103Pd isotope than the 32P isotope: The dose rates are high within both radioisotope stents, particularly around the struts, but fall rapidly beyond the margins. An overstretch injury at a distance of 2 mm away from the stent margins will not be covered by a therapeutic dose rate delivered by either type of radioisotope stent.
We evaluated implants of a novel stent emitting γ-radiation (103Pd) in a rabbit iliac artery overstretch injury model and studied activity levels yielding radiation dose rates comparable to those of previously tested 32P stents. A 103Pd stent inhibits within-stent neointimal hyperplasia in a dose-dependent fashion, causes delayed healing within the stent, and increases the presence of inflammatory cells around the stent struts. Edge effects, previously described in clinical studies using 32P stents, were found after implantation of the 103P stent into the rabbit. All edge effects occurred in the dose-falloff zone at the stent ends and correlated with the severity of balloon overstretch injury of the arteries. In our model, edge effects were therefore located predominantly at the distal stent ends. It should be noted that the stent delivery balloons used in this study were longer than the stents and that these stents were implanted with a minimum of 1 mm balloon trauma zone beyond each stent end. The data show that neointimal hyperplasia beyond the stent ends can be induced by stent delivery balloons even after nonradioactive stent implantation. The combination of diffuse in-stent restenosis with lumen narrowing at the stent ends, however, does not appear angiographically as an edge effect. Although a number of factors are thought to be responsible for edge effects, recent data from clinical studies of catheter-based radiation therapy shed some light on the problem. Geographic miss, ie, predilatation or postdilatation outside an irradiation zone receiving 100% isodose, is the major cause of edge effects in intracoronary radiation therapy to prevent restenosis.7–9 With catheter-based radiation therapy, however, it is difficult to differentiate whether vessel injury alone outside the irradiation zone, radiation injury alone, or a combination of radiation dose falloff and vessel injury causes edge effects. This distinction can be made by investigating a “direct-stenting” procedure using radioactive stents in a controlled experimental setting. Our study shows that direct stenting without predilatation or postdilatation does not appear to be a sufficient strategy to prevent edge effects with balloon-expandable radioactive stents.
Interestingly, lumen narrowing beyond the stent edge increased with stent activity, ie, with radiation dose, in our study. The 2 highest stent activities induced negative vessel remodeling, indicating that there may be a threshold dose for the occurrence of vessel shrinkage after radiation therapy. Previous animal studies of radioactive 90Y stents producing high dose rates showed that these implants cause delayed healing within the stented lumen and atherosclerotic lesion formation.10 The clinical evaluation of β-particle–emitting stents indicated that only low to moderate radiation doses are needed to markedly reduce late loss within the stent body and that higher doses may not bring a benefit for patients.6 Radioisotope stents emitting low dose rates have already been reported to reduce neointimal hyperplasia without apparent edge effects.2 Several years ago, we did not observe edge effects from stainless steel stents bombarded with protons yielding mixed β-particle and γ-radiation (55Co, 56Co, 57Co, 57Ni, 52Mn, etc). Those Co stents delivered lower dose rates than 32P stents but remained active over a longer period.2,3
γ-Emitting isotopes appear to be as suitable as β-particle–emitting isotopes to prevent within-stent restenosis. Another γ-emitting isotope (133Xe), ion-implanted into metallic stents, was reported to reduce neointimal hyperplasia in a similar rabbit restenosis model.11 Despite the shallower dose gradient, however, γ-emitters may not be superior to β-particle emitters with respect to the prevention of edge effects. In addition, γ-emitters have the disadvantage of requiring more radiation protection than β-particle emitters.12,13
An approach to reduce barotrauma beyond the stent ends caused by a stent delivery balloon and possibly reduce edge effects may be a self-expanding design for a radioisotope stent. This strategy should be investigated further.
Limitations of the Study
Only 71% of the total dose was delivered after 30 days; thus, the 103Pd stents were still radioactive after removal for histopathological analysis. In general, novel coronary brachytherapy devices should be studied in animals for a period of ≥6 months to fully understand the potential therapeutic benefit or recognize side effects if the early follow-up results, ie, after 4 weeks, are promising. Further studies are needed to fully understand the causes of edge effects.
This study was supported in part by grants from Isostent, Inc, Santa Clara, Calif. The study was conducted during Dr Hehrlein’s tenure as a visiting professor at the Borgess Heart Institute, Kalamazoo, Mich.
Fischell TA, Kharma BK, Fischell DR, et al. Low-dose, β-particle emission from “stent” wire results in complete, localized inhibition of smooth muscle cell proliferation. Circulation. 1994; 90: 2956–2963.
Hehrlein C, Gollan C, Dönges K, et al. Low-dose radioactive endovascular stents prevent smooth muscle cell proliferation and neointimal hyperplasia in rabbits. Circulation. 1995; 92: 1570–1575.
Hehrlein C, Stintz M, Kinscherf R, et al. Pure β-particle–emitting stents inhibit neointima formation in rabbits. Circulation. 1996; 93: 641–645.
Wardeh AJ, Kay IP, Sabate M, et al. β-Particle–emitting radioactive stent implantation: a safety and feasibility study. Circulation. 1999; 100: 1684–1689.
Albiero R, Adamian M, Kobayashi N, et al. Short-and intermediate-term results of 32P radioactive β-emitting stent implantation in patients with coronary artery disease. Circulation. 2000; 101: 18–26.
Albiero R, Nishida T, Adamian M, et al. Edge restenosis after implantation of high activity 32P radioactive β-emitting stents. Circulation. 2000; 101: 2454–2460.
Serruys PW, Kay IP. I like the candy, I hate the wrapper: the 32P radioactive stent. Circulation. 2000; 101: 3–7.
Kozuma K, Costa M, Sabate M, et al. Three-dimensional intravascular ultrasound assessment of noninjured edges of β-irradiated coronary segments. Circulation. 2000; 102: 1484–1489.
Sabate M, Costa M, Kozuma K, et al. Geographic miss: a cause of treatment failure in radio-oncology applied to intracoronary radiation therapy. Circulation. 2000; 101: 2467–2474.
Laird JR, Carter AJ, Kufs WM, et al. Inhibition of neointimal proliferation with low-dose irradiation from a β-particle–emitting stent. Circulation. 1996; 93: 529–536.
Carter AJ, Jenkins S, Sweet W, et al. Dose and dose rate effects of beta-particle emitting radioactive stents in a porcine model of restenosis. Cardiovasc Radiat Med. 2000; 1: 327–335.