Effects of Intracoronary Radiation on Thrombosis After Balloon Overstretch Injury in the Porcine Model
Background—The main complications of PTCA remain thrombosis and restenosis. Recent studies have demonstrated reduction in the neointimal hyperplasia after intracoronary radiation (IR) with doses of 10 to 25 Gy of ionizing radiation delivered by either β- or γ-emitters to injured vessels. The purpose of this study was to examine the effect of ionizing radiation on the thrombosis rate (TR) of injured porcine coronary arteries.
Methods and Results—Thirty-four juvenile swine (63 coronary arteries) were subjected to overstretch balloon injury followed by IR with doses of 0 to 18 Gy of either β- or γ-radiation. Two weeks after treatment, tissue sections were perfusion-fixed, stained with hematoxylin-eosin and Verhoeff–van Gieson’s stain, and analyzed for presence of a thrombus, thrombus morphology, and neointima formation by computer-assisted histomorphometry techniques. Although the overall TR increased dose-dependently from 0 to 18 Gy prescribed dose, luminal thrombi decreased. Thrombus area also decreased with increasing radiation dose, whether assessed at the prescription point or at the luminal surface, which corresponded to decreased intimal area. Furthermore, luminal thrombi present after IR tended to consist mostly of fibrin and thus were less organized than in controls.
Conclusions—These results suggest that IR induces thrombosis but does not necessarily compromise the lumen. Strategies for reducing TR may further decrease intimal area as well as increasing the safety of this therapy.
Percutaneous transluminal coronary angioplasty (PTCA) has been used successfully to treat atherosclerotic coronary artery disease but has been accompanied by a rate of restenosis of 30% to 50% due to neointimal hyperplasia and unfavorable remodeling.1 One of the parameters considered key for the inflammatory process, which precedes the formation of neointima, is the presence of platelet thrombi.1 2 3 Although pharmacological therapies have proved to be largely ineffective at reducing restenosis after PTCA in coronary arteries, some therapies using antagonists of the platelet adhesion molecule GP IIa/IIIb have been associated with reduction of thrombotic events—and possibly also neointima—after intervention.1 A promising new avenue for the reduction of restenosis comes from recent animal studies and pilot clinical studies using intracoronary radiation (IR).4 However, it is of primary interest to devise means by which to reduce the attendant complications of this therapeutic modality and to reduce the dose of radiation required to achieve a reduction in restenosis.
Doses of external-beam radiation similar to those used to treat balloon-injured coronary arteries in animal models result in increased thrombosis and delayed wound healing.5 6 There have been anecdotal reports of subacute thrombosis (<30 days) and late thrombosis (>30 days) after IR despite treatment with antithrombotic agents in the feasibility clinical studies.7 This finding is paradoxical, because thrombosis has been linked to increased neointimal hyperplasia,1 2 3 whereas IR reduces neointimal hyperplasia.4 To address this paradox and to study this phenomenon in detail, we set out to examine the overall thrombosis rate (TR), the luminal thrombosis rate (LT), the nonluminal thrombosis rate (NLT), and the thrombus area (TA) after treatment with doses of IR ranging from 0 to 18 Gy of either β- or γ-radiation prescribed to the adventitia. These issues were addressed in a retrospective study in the porcine model of balloon overstretch injury (BI) at 14 days after treatment. This time point was chosen because previous studies had shown that the efficacy of catheter-based IR at suppressing intima formation at this time point persisted to 6 months8 9 and because the thrombosis observed clinically appeared to be subacute.7
BI Procedural Details
The animals (51 domestic juvenile swine, weight 30 to 45 kg [Thomas Morris, Inc, Reiserstown, Md]) were subjected to BI in the left anterior descending, left circumflex, or right coronary artery (a total of 76 arteries) as described previously.8 10 Briefly, the animals were sedated with a combination of ketamine 25 mg/kg and xylazine 2 mg/kg IM, then intubated and ventilated with oxygen (2 L/min), nitrous oxide (2 L/min), and isoflurane 1% (1.5 L/min) with a Harvard respirator. All animals were pretreated with aspirin 325 mg and ticlopidine 250 mg BID 24 hours before the procedure and on each subsequent day until death (14 days).
After placement of an introducer sheath in the right carotid artery by surgical cutdown, each animal received a single dose of heparin (150 U/kg), and coronary angiography was performed and recorded on cine film (Phillips Cardiodiagnost). Coronary overstretch injury was performed with an angioplasty balloon 30% larger than the reference vessel diameter, positioned in the proximal segments of the left anterior descending, left circumflex, and right coronary arteries, and inflated to 10 atm 3 times for 30 seconds in each artery. After the completion of the injury, the angioplasty balloon was withdrawn, a final angiography was performed to assess vessel patency and degree of injury, and the vessels were randomly assigned to radiation or sham treatment. After irradiation, the delivery catheter and the guiding catheters were removed, and the carotid cutdown was repaired. Nitroglycerin ointment (1 inch) was administered topically, and the animals were returned to routine care.
Radiation Procedural Details
After the BI, a closed-end lumen catheter with marker was positioned over the injured site for use to deliver radiation dose. The position of the catheter was checked with fluoroscopy to ensure adequate coverage at both ends of the injured site. The 4 different radiation sources examined as part of this study were 192Ir ribbon from Best Medical (n=17 arteries), 186/188Re wire from Soreq (n=10 arteries), 90Y wire from Schneider (n=18 arteries), and 133Xe gas balloon from Cook Cardiology (n=9 arteries). The first 2 were noncentering γ- and β-sources, respectively, and the remaining 2 were centered β-sources. Dosimetric calculations were carried out by use of the TG43 algorithm11 with data generated by Monte Carlo calculations following TG60 recommendations.12 Dose prescription points were those used in previous animal studies, 2 mm from source center for 192Ir ribbon,8 2 mm from source surface for 186/188Re wire,13 1 mm from balloon surface for 90Y wire,14 and 0.25 mm from balloon surface for 133Xe gas15 (Table 1⇓). All sources were left in the catheter for a period sufficient to deliver the prescribed doses of 0 to 18 Gy (2 to 20 minutes). In some cases, the 0-Gy dose was delivered by administration of a dummy wire, and in others, no wire was administered; no significant quantitative or qualitative differences in extent of injury or thrombosis were observed between these 2 types of arteries (data not shown).
The doses described above of the various isotopes used in this study were also recalculated to reflect dose at the luminal surface. The calculations of luminal dose were different depending on whether the isotope was delivered via a centering system (133Xe, 90Y) or a noncentering system (192Ir, 186/188Re). Essentially, noncentering systems had maximal and minimal doses, whereas the centering systems did not. Consequently, we calculated the luminal dose on the basis of the average of the minimum and maximum doses for 192Ir and 186/188Re and used a simple multiple for 133Xe and 90Y (Table 1⇑).
Tissue Analysis Protocol
After an angiogram was obtained under anesthesia, the animals were killed 14 days after radiation or placebo treatment by injection of a euthanasia solution (390 mg/mL pentobarbital sodium, 1% propylene glycol, 29% ethyl alcohol, 2% benzyl alcohol). The hearts were perfusion-fixed, and the injured segments were dissected free from the heart. Serial 2- to 3-mm transverse segments were processed and embedded in paraffin. Cross sections (4 μm) were stained with hematoxylin and eosin (H&E) and Verhoeff–van Gieson elastin (VVG) stain. An experienced observer blinded to the treatment group examined all histological sections. Each specimen was evaluated for the presence of thrombus, neointima formation, and morphological appearance of the cells within the media, adventitia, and neointima.
Histomorphometric analysis was performed on each segment with evidence of medial fracture. The histopathological features were measured by use of a computerized PC-compatible image analysis program (Optimas 6; Optimas, Inc). VVG-stained sections were magnified at ×25, digitized, and stored in a frame-grabber board (DAGE-MTI). Thrombi were identified and designated as luminal or nonluminal (mural), depending on their predominant location in the artery: LTs were those thrombi in which >75% of the thrombus was present in the lumen, whereas NLTs were those thrombi in which >75% of the thrombus was present either between the media and the adventitia or completely in the adventitia. Area measurements were obtained by tracing the thrombus perimeter (TA, mm2), neointima perimeter (intimal area, IA, mm2, defined by the borders of the internal elastic lamina, lumen, media, and external elastic lamina), and external elastic lamina (vessel area, VA, mm2). The total TA was calculated by summing individual TAs. The arc length of the medial fracture (FL), traced through the neointima from one dissected medial end to the other, was used as a measure of the extent of injury. To correct for extent of injury, the ratios IA/FL and TA/FL were calculated. To correct for thrombosis as a function of vessel size, the ratio TA/FL was calculated.
Comparisons of TA, IA/FL, TA/VA, and TA/FL between control and irradiated arteries were made by either 1-way ANOVA with the Bonferroni correction for groups whose SD of the means was not statistically different (P>0.05 by Bartlett’s test) or by the Kruskall-Wallis test for groups whose SD of the means was statistically different (P<0.05 by Bartlett’s test). Differences in TRs were analyzed by χ2 analysis. Statistically significant differences between treatment groups were considered to be those with P<0.05.
TR and LT
In this retrospective study, we examined 51 animals in which 76 arteries were subjected to BI and subsequently treated as follows: 0 Gy (n=22 arteries), 5 Gy (n=10 arteries), 15 Gy (n=26 arteries), and 18 Gy (n=18 arteries). All arteries were examined 14 days after treatment; no arteries from animals that died prematurely before this time point were analyzed. We first analyzed the histological incidence of thrombosis (TR) as a function of radiation dose. This initial analysis did not discriminate between LT and NLT, nor did we apply a minimal TA as an inclusion criterion for this analysis. As shown in Figure 1⇓, increased TR was observed with increasing doses of IR from 5 to 18 Gy (P<0.05 versus 0 Gy). However, when arteries were evaluated for the presence of LT, as opposed to NLT, we found that LT decreased with increasing doses of IR (Figure 1⇓; P<0.05).
Representative photomicrographs from these experiments are shown in Figure 2⇓. These photomicrographs demonstrate typical LT (Figure 2⇓, A, C, E, and G) and NLT (Figure 2⇓, B, D, F, and H) in arteries subjected to BI alone (Figure 2⇓, A and B), BI+5 Gy (Figure 2⇓, C and D), BI+15 Gy (Figure 2⇓, E and F), or BI+18 Gy (Figure 2⇓, G and H), respectively. Most LTs observed at 0 and 5 Gy appeared to be acute or subacute, consisting of platelet and/or fibrin deposits that in many cases tended to occlude the vessel by ≥75% (Figure 2⇓, A and C). The insets in these panels demonstrate the presence of erythrocytes and granulocytes. In contrast, the thrombi present in arteries treated with 15 or 18 Gy tended to be NLTs (mural), consisting of a more organized fibrinous layer of thrombus along the surface of the injured artery (Figure 2⇓, F and H; see also insets). As can be surmised from Figure 1⇑, few LTs were observed at 15 or 18 Gy, and this is represented in Figure 2⇓, E and F. Furthermore, the LTs present at 15 and 18 Gy were immature and consisted mostly of fibrin in a disorganized pattern. In general, these thrombi were acellular, lacking in monocytes, lymphocytes, and granulocytes. Angiographic analysis confirmed the presence of total occlusions as quantified histomorphometrically (data not shown).
TA and IA
The extent of thrombosis has been reported to affect the extent of restenosis.1 2 3 As can be seen in Figure 3A⇓, the extent of intimal hyperplasia measured as IA corrected for medial FL16 decreased in a radiation dose–dependent fashion (P<0.05). These findings are in agreement with previous results from studies that used both β- and γ-emitters.4 At the same radiation doses, TA also decreased in a dose-dependent fashion (Figure 3A⇓; P<0.05). A plot of TA versus IA/FL extracted from these data demonstrated a linear relationship (r2=1.00; Figure 3B⇓).
It may be argued that analyses based on prescribed dose are not correct and that in fact the radiation dose to the luminal surface of the vessel affects thrombosis more directly. The isotopes used in these studies vary greatly with respect to type of radiation produced (β versus γ versus a mixture of the two), delivery system (catheter-based versus balloon-based), and centering versus noncentering. These variations can contribute to greatly differing doses at the luminal surface (Table 1⇑). Accordingly, we reanalyzed TA as a function of luminal dose. As seen in Figure 4⇓, we obtained a relationship similar to that observed with doses calculated at the prescription point.
We next compared TR and TA in arteries subjected to BI followed by IR using isotopes with similar prescription point doses (15 Gy) but with widely different luminal surface doses (133Xe and 186/188Re) compared with an isotope with a higher prescription point dose (18 Gy) but a lower luminal surface dose (90Y; Table 2⇓). In addition, sample sizes were nearly identical in all categories. This analysis demonstrated no significant differences in TR and TA between arteries treated with 133Xe and 186/188Re (Table 2⇓), although both groups exhibited significantly greater TR and significantly lower TA and IA/FL than did animals receiving 0 Gy (Table 2⇓). These similar parameters were observed despite the almost 2-fold difference in dose at the luminal surface (84 versus 150 Gy). Furthermore, the highest TR was observed with 18 Gy of 90Y, a dose that corresponds to only 41 Gy at the luminal surface (Table 1⇑).
Because the vessel size (VA) or extent of injury (indicated by FL) may affect the degree of thrombosis, we plotted TA/VA (Figure 5A⇓) or TA/FL (Figure 5B⇓) as a function of radiation dose. Both these plots were similar to that obtained by plotting TA versus radiation dose (Figures 3⇑ and 4⇑), and both demonstrated statistically significant reduction of TA/VA or TA/FL as a function of radiation dose.
We and others have been examining the endovascular administration of ionizing radiation and determined that this therapeutic modality reduces neointimal thickening, at least in part by suppressing the proliferative response and inducing favorable remodeling10 17 18 19 ; however, this is the first publication to present data related to the effect of intravascular radiation on thrombosis in the porcine BI model.
Previous studies on the vascular effects of external beam radiation have suggested that increased thrombosis is an adverse late healing effect.5 6 Our findings suggest that IR is associated only with increased TR, but the percent of LT decreases with increasing IR. This finding would suggest that the healing response is indeed delayed, especially because mural thrombosis may be a feature of early arterial healing.3 Furthermore, the thrombi present in irradiated arteries tended to be acellular, lacking cells thought to be involved in the healing response to arterial injury (monocytes, lymphocytes, and granulocytes).1 2 3 Decreases in the numbers or functions of macrophages may increase the residence of the thrombus and its components at the injured segment and may delay its organization. Our findings may suggest that the healing of irradiated arteries is delayed, and this may have adverse functional consequences.
However, we also found that TA decreases with increasing IR, whether the dose is calculated at the prescription point in the adventitia or at the luminal surface. Likewise, IA/FL decreases with increasing IR, and plotting IA/FL as a function of TA yielded a perfect correlation (Figure 3B⇑). This finding is in agreement with previous reports of an association between extent of thrombosis (ie, TA) and intimal hyperplasia.1 2 3
It is well known that the LTs are more important and clinically relevant in the short term, because they may propagate, leading to total occlusion, or may dislodge, leading to distal embolization. Conversely, the organized mural thrombi are not relevant in the short term but may contribute to the initiation and development of atherosclerosis and restenosis.20 21 This has been demonstrated in humans at the perianastomotic site of coronary bypass surgery, where new lesions of atherosclerosis form in ≈30% of all bypass grafts.22
One intriguing finding of these studies is that TA was more closely associated with dose at the prescription point in the adventitia than with luminal surface dose. Table 2⇑ shows that arteries treated with isotopes that have identical doses at the prescription point (133Xe and 186/188Re; 15 Gy) exhibit similar TRs despite having received widely different luminal surface doses. Arteries treated with 90Y, which was administered at a higher dose at the prescription point (18 Gy), exhibited a higher TR despite having a much lower luminal surface dose. Furthermore, a comparison of Figures 3⇑ and 4⇑ suggests that a better correlation exists between TA and prescription point dose as opposed to luminal surface dose. The efficacy and safety of β- versus γ-radiation have been discussed, especially because of the rapid dose falloff with β-sources.4 23 Our findings suggest that TR and TA may depend more on the radiation dose at the adventitia than at the luminal surface.
This study is subject to several limitations. This was a retrospective study that combined several isotopes and both β- and γ-emitters. However, there were no statistical differences between the 2 isotopes 133Xe and 186/188Re with regard to various parameters of thrombosis and restenosis at the 15-Gy dose, a dose that caused a significant reduction in IA/FL (Table 2⇑). Another limitation of this study is that the arteries were analyzed 14 days after injury, and thus it is unknown what effect IR has on early thrombosis. Nonetheless, these studies suggest that IR at therapeutic doses does not worsen thrombosis but in fact may reduce its impact, and they raise the possibility that one of the mechanisms that underlie the beneficial effect of IR after BI is due to this reduction of thrombosis.
This study demonstrated an effect of intracoronary radiation on thrombus formation and thrombus morphology. Higher doses of radiation were associated with increases in TR, but more often a nonobstructive NLT rather than an LT. In addition, the morphology of thrombi in porcine irradiated arteries appeared to be less organized than the pattern of thrombosis observed in nonirradiated injured arteries. These changes in TR and thrombus pattern after radiation therapy may influence the clinical strategy with antiplatelet therapy in patients undergoing intervention followed by vascular brachytherapy.
This work was carried out with the support of the Cardiovascular Research Foundation and the Medlantic Research Institute, Washington Hospital Center, Washington, DC. The authors would like to acknowledge Sara D. Collins, Rufus Seabron, and Anthony Pierre (Cardiovascular Research Foundation, Washington, DC) for technical experimental assistance and Marc Kollum (Cardiovascular Research Foundation, Washington, DC) for help with figure preparation. The authors would also like to thank Dr Christian Haudenschild (Holland Laboratory, American Red Cross, Rockville, Md) for helpful discussions.
- Received April 19, 1999.
- Revision received July 16, 1999.
- Accepted July 20, 1999.
- Copyright © 1999 by American Heart Association
Fingerle J, Johnson R, Clowes AW, Majesky MW, Reidy MA. Role of platelets in smooth muscle cell proliferation and migration after vascular injury in rat carotid artery. Proc Natl Acad Sci U S A. 1989;86:8412–8416.
Bhargava B, Vodovotz Y, Waksman R. Intracoronary radiation therapy for prevention of restenosis. Indian Heart J. 1998;50(suppl I):120–129.
Condado JA, Waksman R, Gurdiel O, Espinosa R, Gonzalez J, Burger B, Villoria G, Acquatella H, Crocker IR, Seung KB, Liprie SF. Long-term angiographic and clinical outcome after percutaneous transluminal coronary angioplasty and intracoronary radiation therapy in humans. Circulation. 1997;96:727–732.
Waksman R, Robinson KA, Crocker IR, Gravanis MB, Cipolla GD, King SB III. Endovascular low-dose irradiation inhibits neointima formation after coronary artery balloon injury in swine: a possible role for radiation therapy in restenosis prevention. Circulation. 1995;91:1533–1539.
Waksman R, Robinson KA, Crocker IR, Wang C, Gravanis MB, Cipolla GD, Hillstead RA, King SB III. Intracoronary low-dose β-irradiation inhibits neointima formation after coronary artery balloon injury in the swine restenosis model. Circulation. 1995;92:3025–3031.
Bhargava B, Waksman R, Vodovotz Y, Chan RC, Collins SD, Seabron R, Pierre A, Lavie E. Intracoronary irradiation with 186/188rhenium following balloon overstretch injury reduces neointima but does not impair vasoreactivity of porcine coronary arteries. J Interventional Cardiol. 1999;12:263–270.
Waksman R. Endovascular beta radiation in the swine model. In: Waksman R, ed. Vascular Brachytherapy. 2nd ed. Armonk, NY: Futura Publishing Co; 1998:273–286.
Apple M, Waksman R. Xenon 133 gas-filled balloon. In: Waksman R, ed. Vascular Brachytherapy. Armonk, NY: Futura Publishing Co; 1998:569–578.
Robinson KA. Pig coronary artery model of post-angioplasty restenosis. In: Waksman R, King SB, Crocker IR, Mould RF, eds. Vascular Brachytherapy. Armonk, NY: Futura Publishing Co; 1996:30–40.
Hehrlein C, Stintz M, Kinscherf R, Schlösser K, Huttel E, Friedrich L, Fehsenfeld P, Kübler W. Pure β-particle-emitting stents inhibit neointima formation in rabbits. Circulation. 1996;93:641–645.
Antman EM, Braunwald E. Acute myocardial infarction. In: Braunwald E, ed. Heart Disease. Philadelphia, Pa: WB Saunders, Inc; 1997:1184–1288.
Fuster V, Vestraete M. Hemostasis, thrombosis, fibrinolysis and cardiovascular disease. In: Braunwald E, ed. Heart Disease. Philadelphia, Pa: WB Saunders, Inc; 1997:1809–1852.
Chesebro JH, Clements IP, Fuster V, Elveback LR, Smith HC, Bardsley WT, Frye RL, Holmes DR Jr, Vlietstra RE, Pluth JR, Wallace RB, Puga FJ, Orszulak TA, Piehler JM, Schaff HV, Danielson GK. A platelet-inhibitor-drug trial in coronary-artery bypass operations: benefit of perioperative dipyridamole and aspirin therapy on early postoperative vein-graft patency. N Engl J Med. 1982;307:73–78.