This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rud Andersen, H. Articles by Falk, E. Search for Related Content PubMed PubMed Citation Articles by Rud Andersen, H. Articles by Falk, E. Pubmed/NCBI databases Medline Plus Health Information Angioplasty

(Circulation. 1996;93:1716-1724.)
© 1996 American Heart Association, Inc.

Articles

Remodeling Rather Than Neointimal Formation Explains Luminal Narrowing After Deep Vessel Wall Injury

Insights From a Porcine Coronary (Re)stenosis Model

Henning Rud Andersen, MD, PhD; Michael Mæng, MS; Martin Thorwest, MS; Erling Falk, MD, PhD

From the Department of Cardiology and Institute of Experimental Clinical Research, Skejby University Hospital, Aarhus, Denmark.

Correspondence to Henning Rud Andersen, MD, PhD, Department of Cardiology, Skejby University Hospital, Brendstrupgaardsvej, DK-8200 Aarhus N, Denmark.

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Background Oversized balloon dilatation of normal porcine coronary arteries usually heals without stenosis formation.

Methods and Results With the purpose of developing a stenotic model and examining the mechanisms of luminal narrowing after angioplasty, we produced a circumferential deep vessel wall injury by inflating and withdrawing an oversized chain-encircled angioplasty balloon in the left anterior descending coronary artery (LAD) of 20 pigs. Three pigs died and did not complete the study. In 8 pigs (group 1), serial coronary arteriography was performed. The lumen diameter (mean±SD) before dilatation was 3.4±0.4 mm; after dilatation, 4.2±0.6 mm; and at follow-ups 2 and 4 weeks later, 1.6±0.4 mm (P<.0001). In 9 pigs (group 2) examined postmortem 3 weeks after dilatation, histology revealed that the injury was deep (out to adventitia) in all arteries and completely circumferential (360°) in all but two arteries. Adventitia was markedly thickened as a result of neoadventitial formation. Injury correlated strongly with neointimal formation (middle LAD, r=.71, P=.00001), but neither injury nor neointima correlated with lumen size (r=.14, P=.46 and r=.34, P=.07, respectively); ie, neointimal formation did not explain late luminal narrowing. Lumen size, however, did correlate strongly with vessel size (r=.74, P=.000005). The late loss in lumen diameter observed angiographically in group 1 substantially exceeded that caused by neointimal formation seen by histology in group 2.

Conclusions The chain-encircled angioplasty balloon produced a circumferential deep vessel wall injury that healed by luminal narrowing. In this porcine model, arterial remodeling was more important than neointimal formation in late luminal narrowing.

Key Words: angioplasty • remodeling • pathology • balloon • stenosis

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
Renarrowing after otherwise successful coronary angioplasty remains a major problem, occurring in 30% to 50% of treated arteries within 3 to 6 months.^{1 2 3} An injury-related exuberant healing response mediated by SMC proliferation and matrix synthesis (neointimal formation) traditionally has been considered essential in the development of late restenosis. However, antiproliferative agents have not been effective in reducing restenosis in clinical trials¹ despite their proven effectiveness in reducing the proliferative response after arterial injury in rats and rabbits.⁴ The reasons could be that neointimal formation does not, as anticipated, play a critical role in the development of human restenosis and that injury-response processes in smaller animals differ significantly from those in humans.

Contrary to results obtained in rat arterial injury models, those obtained in pig coronary models seem to reflect the pathogenesis of human restenosis much better,^{4 5} and the pig is now widely used in experimental restenosis studies. The focus of research in these porcine models has also been SMC-dependent neointimal formation.^{6 7 8 9 10 11 12} Clinical data obtained in human coronary arteries indicate, however, that neointimal formation may not be as important as arterial remodeling in postinterventional restenosis^{13 14 15 16} ; this phenomenon also has been documented experimentally in rabbit and pig peripheral arteries.^{17 18 19} Recently, postinterventional remodeling was described in pig coronary arteries, but significant late luminal narrowing did not occur in that model.²⁰ Therefore, we decided to develop a stenotic coronary (re)stenosis model in the pig and describe the processes responsible for the final size of the lumen late after intervention, focusing on both neointimal formation and arterial remodeling. The former was achieved by creating a circumferentially orientated deep vessel wall injury that healed by luminal narrowing; the latter, by serial angiographic examinations and extensive microscopic evaluation of the injured vessels.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
Twenty 6-month-old, 75-kg mixed Danish Landrace–Yorkshire pigs were used in the study. Eight pigs (group 1) from an ongoing survival study underwent serial angiographic examinations to document in vivo changes in lumen size. Nine pigs (group 2) underwent postmortem angiography and histomorphometric examination. Three pigs died and did not complete the study. The pigs were handled according to institutional guidelines and the American Heart Association guidelines for animal research. The pigs were fed a normal nonatherogenic laboratory chow diet.

Chain-Encircled Balloon
A 0.7-mm rough silver chain was mounted on the surface of a 3.5-mm PTCA balloon near its middle. The chain encircled the inflated balloon and was secured with thin threads fastened to the catheter at the proximal and distal ends of the balloon (Fig 1). Inflated at 8 atm, the total diameter (balloon plus chain) was 4.9 mm. The diameter of the deflated chain-encircled balloon was 2 mm, and the deflated balloon was introduced into the coronary artery by a specially constructed guide catheter.

View larger version (97K): [in this window] [in a new window]   Figure 1. The chain-encircled balloon. A 0.7-mm silver chain is mounted on the surface of a 3.5-mm PTCA balloon. The chain encircles the inflated balloon and is secured with thin threads fastened to the catheter at the proximal and distal ends of the balloon.

In Vitro Experiments
The intention was to produce a deep and completely circumferential (360°) injury. The reliability of the developed tools and techniques in producing such an injury was tested in pig hearts obtained from the local slaughterhouse.

In Vivo Testing
The pigs were sedated with 1000 mg ketamine (Ketalar) IM and 45 mg midazolam (Dormicum) IM, followed by 500 mg ketamine IV, intubation, and mechanical ventilation with 6 L/min N₂O and 3 L/min O₂. Anesthesia was maintained by infusion of ketamine and midazolam (400 and 12 mg/h IV, respectively). Fentanyl (Haldid) 0.5 mg/h was given against pain. Amiodarone (Cordarone, 900 mg) and ampicillin (Anhypen, 1 g) were used prophylactically to prevent fatal arrhythmias and infections, respectively. Before catheterization, heparin (20 000 U IV) and acetylsalicylic acid (500 mg IV) were given and later were supplemented with another bolus of heparin (20 000 U) just before the PTCA procedure. No further antithrombotic agents were given at any time until euthanitization.

After exposure of the right common carotid artery, a 14F introducer sheath was inserted for arterial access. A PTCA guide catheter was advanced to the aortic root, and selective coronary angiography was performed. The deflated chain-encircled angioplasty balloon was advanced into the LAD to a position where the diameter of the lumen was 3 mm (usually between the first and second diagonal branches). At that point, the chain, projecting 0.7 mm out from the surface of the inflated 3.5-mm balloon, was supposed to "cut" circumferentially into the vessel wall and "abrade" intima and media when withdrawn. The chain-encircled balloon usually stuck firmly in the LAD when inflated to 8 atm. When the catheter was pulled, the balloon suddenly yielded and moved quickly back through the LAD and out into the aorta, where the balloon was deflated and removed. Frequently, small pieces of tissue were found attached to the chain on the balloon after the procedure. The right carotid artery was ligated, the wound was closed, and the pig was returned to a recovery area.

Serial Coronary Arteriography (Group 1)
Coronary arteriography was performed before dilatation, 2 hours after dilatation, and at follow-up 2 (n=4) and 4 (n=4) weeks later to document changes in lumen size. The MLD was identified and measured at follow-up. All angiographic measurements were determined with end-diastolic frames. For calibration, a 2.00-mm silver ball was fixed to the tip of a guide wire and introduced into the distal LAD. The guide wire was withdrawn slowly during angiography. The angiographic diameter of the silver ball at the MLD site was used for calibration. Stenosis was calculated as stenosis=(1-MLD/predilated diameter)x100%.

Postmortem Procedures and Histomorphometric Analysis (Group 2)
Three weeks after dilatation, the 9 pigs in group 2 were anesthetized, and after thoracotomy, their hearts were removed and cannulas were inserted into the ascending aorta. The coronary arteries were gently flushed with 4% paraformaldehyde solution (pH 7.2) followed by infusion of a warm (37°C) barium gelatin mixture at 100 mm Hg. The contrast medium was hardened by placing the heart overnight in 4% paraformaldehyde at 4°C; ie, the arteries were fixed in the distended state. LAD was dissected free of the heart, fixed further by immersion in 4% paraformaldehyde, and then placed directly on radiographic films and x-rayed in different projections. Thereafter, the arteries were cut into cross sections at 3-mm intervals perpendicular to the direction of blood flow. All 3-mm segments were processed for microscopic examination. The final paraffin sections were stained with hematoxylin and eosin and a trichrome-elastin stain.

Area measurements were performed by point counting with a microscope (Olympus BHS) equipped with a video camera projecting the field of vision to a monitor. A computer with a stereological software package (GRID, Olympus) was connected to the monitor. This software can superimpose (on the monitor) a test system of regularly spaced points on the field of vision and defines the area (A_p) each point (P) represents. By counting the points hitting the area of interest, the software finds the total area (A_tot) by multiplying the number of hits by the area each point represents (A_tot=PxA_p).^{21 22 23} Point counting is a simple and reliable method for area measurements if enough points are counted. At least 60 points were counted for each area of interest, giving a coefficient of error <5% according to a published nomogram.²² The following parameters were measured (Fig 2).

View larger version (143K): [in this window] [in a new window]   Figure 2. A, B, Histological cross sections of coronary artery illustrating gap angle (g), neointima (neo), media (m), and blue-stained adventitia (a). Thick black-stained elastic laminae clearly define the outer and inner borders of tunica media that contains fine elastic fibers (vs no elastic fibers in neointima). Adventitia is thickened at the site of deep injury. C, D, Histological cross sections of two severely injured middle LAD segments at the same magnification (gap angle, 360° in C and 295° in D) illustrating the lack of a relationship between neointima and lumen size. Although much more neointima is present in D, the lumen is larger because the artery is larger. A thickened and collagen-rich neoadventitia is clearly seen in both C and D. Magnification x13 (A, C, and D) and x26 (B). Trichrome-elastin stain, rendering collagen blue, muscle red, and elastin black.

1. Injury score—the part of the circumference measured in degrees where tunica media had been abraded and adventitia exposed (gap angle; see Fig 2B). The sides of the gap angle were drawn from the center of the lumen, defined as the cross point of two lines drawn perpendicular to each other, with the lumen divided into four areas of equal size (determined visually by superimposing a transparent sheet marked with a cross on a projected image of the vessel). The gap angle was measured directly with a protractor. An injury score of 360° indicates a completely circumferential deep injury.

2. Adventitia—area between periadventitial tissues (adipose tissue and myocardium) and external elastic lamina.

3. Vessel size—area circumscribed by the external elastic lamina. The vessel wall is by definition composed of an intimal, a medial, and an adventitial layer. Of note, the vessel size as defined in this study does not include the adventitial component of the vessel wall.

4. Media—area between internal and external elastic lamina (normal media) and, when the internal elastic lamina is missing, areas of remnants of medial tissue (well-organized SMCs with intervening thin elastic fibers).

5. Neointima—area between the lumen and the internal elastic lamina and, when the internal elastic lamina is missing, the area between the lumen and remnants of media tissue or external elastic lamina.

6. Lumen—area circumscribed by the intima/neointima-lumina border.

To evaluate and eliminate the effect of differences in preinterventional vessel size on the above parameters, LAD was divided into proximal (cross sections 2 through 6), middle (cross sections 7 through 11), and distal (cross sections 12 through 16) 1-cm-long segments.

Postmortem angiographic diameter stenosis was determined as follows: stenosis=(1-MLD/normal distal diameter)x100%.

The distal and not the proximal segment was consistently used as normal reference lumen because LAD proximal to the balloon inflation site was damaged by the withdrawing procedure. Consequently, by use of a smaller lumen as reference, the real stenosis severity was consistently underestimated. If side branches were present in the histological sections, it was not possible to demarcate the areas of interest exactly, and such sections were excluded from analysis.

Statistical Analysis
Data derived from in vivo angiographic examination before dilatation, after dilatation, and at follow-up were first analyzed by one-way ANOVA and subsequently by a two-sided paired t test. Data derived from the study of all histological sections were pooled according to segment (proximal, middle, and distal LAD) and analyzed separately. Linear regression was performed to determine the correlation between two continuous data (Pearson's correlation test for independent data). Additionally, the point of maximal histological luminal narrowing was identified in each LAD, and mean values of measurements performed here (n=9) were compared with corresponding mean values of measurements performed 3 mm upstream and downstream of the narrowest point in each artery (two-sided paired t test). Data are presented as mean±SD. A value of P<.05 was considered significant. The reproducibility of measurements of adventitial area and gap angle was assessed by double measurements in 20 histological sections and calculation of the interobserver and intraobserver variabilities.²⁴

	Results

Top Abstract Introduction Methods Results Discussion References

 
In Vitro Experiments
Different balloon and chain sizes were tested in 31 coronary arteries postmortem. It was confirmed that a deep and circumferential injury could be produced by inflating and withdrawing a chain-encircled angioplasty balloon within a coronary artery. With this procedure, intima and media were abraded, and adventitia was exposed. To produce a deep and circumferential injury, it was not enough just to inflate the chain-encircled balloon; it was also necessary to move the balloon by pulling the catheter. The pulling procedure created a rather abrupt distal border between normal and abraded media, whereas the proximal border frequently was irregular, with detached medial flaps projecting into the lumen. Three different chain sizes were tested without any obvious differences in circumferential extent or deepness of injury produced, and the thinnest silver chain (diameter, 0.7 mm) was used for the subsequent in vivo experiments.

Coronary Angiography
In vivo coronary angiography (group 1) revealed that the lumen diameter before dilatation was 3.4±0.4 mm; after dilatation, 4.2±0.6 mm; and at follow-up, 1.6±0.4 mm (P<.0001, ANOVA; see the Table). All coronary arteries had an acute gain 2 hours after dilatation, and all arteries had a narrowed lumen at follow-up (see the Table). The late loss in lumen diameter was 2.6±0.7 mm. The mean angiographic stenosis was 53% compared with the vessel diameter before dilatation. Postmortem coronary angiography (group 2) revealed an MLD of 1.3±0.4 mm, corresponding to a 32% diameter stenosis compared with the normal noninjured vascular segment located 13.3±4.3 mm further distal in LAD (see the Table). In group 2, major side branches departed between the MLD site and the distal reference segment in six angiograms. Therefore, the diameter of the distal reference segment (1.9±0.3 mm; the Table) was smaller than one would expect from tapering alone. Only eight pigs underwent postmortem angiographic evaluation because of incomplete contrast injection in one heart.

View this table: [in this window] [in a new window]   Table 1. Coronary Arteriographic Lumen Diameter

Injury
Histology revealed that the injury was deep (out to adventitia) in all arteries, but the degree of circumferential injury varied greatly along the abraded arteries. Seven of the nine arteries contained at least one segment in which tunica media had been totally removed the entire circumference around; ie, a 360° circumferential and deep injury was produced in all but two cases.

Neointima
Like the degree of injury produced, the resultant healing response varied greatly along the abraded arteries. The newly formed tissue (neointima) consisted of spindle-shaped cells surrounded by a rather loose extracellular matrix without elastic fibers. The cells reacted positively for -smooth muscle actin by immunostaining (Dako 1A4). Neointima filled medial gaps and usually extended the entire circumference, also covering apparently normal tunica media and intact internal elastic lamina. At the base of medial gaps, thrombuslike material rarely was found. Where the external elastic lamina had been lacerated and within healed dissections between medial flaps and adventitia, a few small blood vessels containing radiographic contrast medium were sometimes seen entering neointima from the adjacent adventitia (neovascularization), but no capillaries were seen near the luminal surface. No obvious inflammatory cells were present within neointima.

Adventitia
Near sites of deep injury, adventitia stained strongly for collagen and was markedly thickened owing to neoadventitial formation (Fig 2). Infiltration with mononuclear inflammatory cells was always recognized. No consistent correlations were found between adventitial area and injury score, late vessel size, or late lumen size. The reproducibility of the adventitial measurements revealed interobserver and intraobserver variabilities of 9% and 8%, respectively.

Injury-Response Relationships
There was a strong positive correlation between injury score (gap angle) and resultant neointimal formation for all three vascular segments; the larger the injury, the more neointima was apparent 3 weeks later (Fig 3). There was, however, no consistent relationship between gap angle and maximal neointimal thickness (proximal LAD, r=.40, P=.12; middle LAD, r=.64, P=.0001; distal LAD, r=.26, P=.27). Late vessel size increased with increasing injury score for proximal and middle LAD segments (Fig 3), whereas the lumen stayed constant. For distal LAD segments, late vessel size did not correlate with injury score (r=.003, P=.99), and the lumen decreased with increasing injury score (r=-.59, P=.005). Maximal neointimal thickness (0.77±0.26 mm) and vessel size (7.05±4.50 mm²) correlated strongly (r=.77, P=.005) in the 11 histological sections from all three arterial segments (proximal, middle, and distal) in the seven arteries in which media had been totally stripped off. The reproducibility of the gap angle measurements revealed interobserver and intraobserver variabilities of 1.9% (4°) and 1.7% (2°), respectively.

View larger version (18K): [in this window] [in a new window]   Figure 3. Scatterplots illustrating the histological relationships between injury score (gap angle), lumen, vessel, and neointimal size in the middle LAD segments. A positive correlation exists between degree of injury and both neointimal formation and vessel size. Thus, with more extensive injury, more neointima is formed. However, no significant correlation exists between lumen size and neointima (r=.34, P=.07, middle). Consequently, neointimal formation does not explain luminal narrowing. However, a small lumen is associated with a small vessel (middle), and a small vessel is associated with a small amount of neointima (right). Thus, vessel size and not neointimal formation is most important for late luminal narrowing ([re]stenosis).

Neointima, Vessel Size, and Lumen
Neointimal formation and late vessel size correlated strongly in all three vascular segments; the larger the vessel was 3 weeks after intervention, the more neointima had been formed (Fig 3). Neointimal formation was expected to correlate inversely with late lumen size, but it did not (proximal LAD, r=.10, P=.69; middle LAD, r=.34, P=.07; and distal LAD, r=-.08, P=.73). Thus, neointimal formation did not explain luminal narrowing after deep vessel wall injury. In contrast, final vessel size correlated strongly with late lumen size in all three segments. That is, the smaller the vessel was at 3 weeks, the smaller its lumen was (Fig 3).

Histology of Stenosis
Fig 4 shows all histological measurements of lumen size, vessel size, and vessel wall components. In one artery (No. 5), a double lumen was seen in the distal segment owing to intramural vessel wall dissection. Consequently, this segment of the artery was excluded from analysis. Compared with normal distal vascular segments, dilated and abraded segments were enormously thick-walled. Alignment of the most stenotic segment from each LAD after the point of maximal stenosis and combination of all data (mean values, n=9) revealed more clearly the contribution of individual vessel wall components to vessel size, wall thickness, and luminal narrowing (Fig 5). At the point of maximal narrowing, the lumen was significantly smaller than it was 3 mm distally and 3 mm proximally, despite a similar magnitude of neointima formation and similar or less tunica media present. However, the size of the vessel at the point of maximal luminal narrowing appeared smaller than it was 3 mm distally and 3 mm proximally. Furthermore, the size of the vessel correlated significantly with the size of the lumen (r=.79, P=.02). Thus, late vessel size and not neointimal formation was the major determinant for luminal narrowing at the point of maximal stenosis.

View larger version (71K): [in this window] [in a new window]   Figure 4. Drawings illustrating the measured cross-section areas in nine LAD arteries 3 weeks after angioplasty. Cross sections were obtained at 3-mm intervals from the left main stem (LM) to the distal nondilated LAD. The contribution of individual vessel wall components to vessel wall thickness, vessel size, and lumen size are depicted. The injured arterial segments are enormously thick walled, mainly because of a thick mantle of neoadventitial tissue, whereas neointima does not contribute so much to vessel wall thickness. In one artery (No. 5), a double lumen was present in the distal segment that precluded proper measurements.

View larger version (58K): [in this window] [in a new window]   Figure 5. Drawing derived by aligning the nine most stenotic segments (one from each artery) after the point of maximal stenosis. Neointima is not particularly thick at the point of maximal stenosis; in fact, it tends to be thicker 3 mm longer distally in the vessel where the lumen is larger. At the point of maximal stenosis, the vessel appears constricted compared with adjacent injured segments. Area measurements performed at maximal stenosis and 3 mm proximal (prox.) and distal (dist.) to the stenosis are listed. Comparing measurements at stenosis with those 3 mm away from stenosis gave the following results: adventitia (proximal, P=.43; distal, P=.20), vessel size (proximal, P=.015; distal, P=.09), media (proximal, P=.05; distal, P=.12, neointima (proximal, P=.87; distal, P=.15), and lumen (proximal, P=.02; distal, P=.01). Summation of lumen plus media plus neointima does not exactly equal vessel size because vessel size was not calculated from the other measurements but was measured separately by point counting, and each measurement had a small inaccuracy (coefficients of error, <5%22 ).

Adventitia was enormously thickened as a result of neoadventitial formation, but there was no significant correlation between adventitial area and vessel size at site of maximal stenosis (r=.44, P=.33). Neoadventitial formation was not more pronounced at the site of maximal luminal narrowing than it was 3 mm distal and 3 mm proximal to the narrowing (Fig 5). In the normal reference segment more distal in the LAD, the lumen was 2.3±0.6 mm², media was 0.8±0.4 mm², vessel size was 3.0±0.7 mm², and adventitia was 1.4±0.7 mm². Comparisons between maximal stenosis and the normal reference segment showed that at maximal stenosis the lumen was significantly smaller (1.4±0.7 versus 2.3±0.6 mm², P=.002) and adventitia was significantly larger (7.5±3.7 versus 1.4±0.7 mm², P=.009), whereas vessel size was the same (4.1±1.6 versus 3.0±0.7 mm², P=.07).

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
Dilatation of normal porcine coronary arteries with an oversized balloon usually creates only a single and longitudinally oriented slitlike medial rupture that heals in the course of 2 to 4 weeks, giving rise to an eccentric lesion that hardly protrudes into the lumen; ie, a stenosis rarely develops.^{6 7 8 20} Oversized stenting was introduced to create a more exuberant healing response that sometimes but not always results in late luminal narrowing.^{9 10 11 25} By preventing recoil, stenting preserves the acute luminal gain much better than balloon dilatation alone,²⁶ and occasionally the lumen may paradoxically stay larger despite much more stent-induced injury, thrombosis, and proliferation.^{6 11}

New Porcine (Re)stenosis Model
To increase the likelihood of obtaining a stenotic healing response without stenting (precludes late arterial remodeling), we maximized the circumferential extension of deep injury by retracting a chain-encircled balloon in LAD. This concept is opposite that suggested previously to minimize the healing response.^{27 28} Barath et al²⁷ and Bonan et al²⁸ introduced the "cutting balloon" and hypothesized that sharp and regular longitudinally oriented vessel wall incisions might limit the injury-induced healing response and reduce the risk of restenosis. We hypothesized that doing the opposite, maximizing the injury by creating an irregular circumferentially oriented deep vessel wall injury, would induce an exuberant circumferential healing response and increase the likelihood of getting the stenotic animal model we wanted. It worked; the healing response narrowed the lumen, and an angiographic stenosis developed 2 to 4 weeks later. We decided to evaluate 3-week-old lesions by histology because the neointimal healing response reportedly is maximal at that time.²⁹

Injury-Response Relationships
The depth of injury caused by oversized stenting¹⁰ and the circumferential length of medial rupture caused by oversized balloon dilatation^{7 8 20} were previously shown to correlate with resultant neointimal formation. Also, the present porcine model revealed a strong injury-neointima relationship. Of more interest, neointimal formation and lumen size did not correlate. Thus, neointimal formation did not influence the most important outcome parameter: the size of the lumen late after intervention.

Injury score also correlated with late vessel size. That is, the more a vessel is dilated, the more injury is produced and the more neointima is subsequently formed, but the enlarged vessel can accommodate a huge mass of neointima without luminal narrowing. Only if late luminal loss (includes both neointimal formation and vascular remodeling) exceeds the preceding acute luminal gain does a stenotic lesion develop. In the present porcine model, late luminal narrowing was caused predominantly by constrictive vascular remodeling resulting in a shrunken artery.

Like the degree of injury produced, the resultant healing responses varied greatly along the abraded arteries. The reason for that is not obvious but could be qualitative differences in local cell function that probably are related to the rapidity of complete reendothelialization, the degree of adventitial/perivascular damage, and the magnitude of inflammatory changes.¹⁸

Neointima, Remodeling, and Restenosis
Serial in vivo angiography revealed a late loss in lumen diameter of 2.6±0.7 mm, giving rise to a late MLD of 1.6±0.4 mm (see the Table). The mean MLD at 4 weeks' follow-up in group 1 was identical to that determined by postmortem angiography in group 2 (1.3 mm), indicating that similar results were obtained in the two groups. Histology revealed a minimal lumen area of 1.4 mm² surrounded by 1.8-mm² neointima (Fig 5), which corresponds to a calculated 0.3-mm-thick ring of neointima at the point of maximal narrowing. As shown previously, our histological procedure results in tissue shrinkage amounting to 20% linear reduction.³⁰ Therefore, the 0.3-mm-thick neointima seen by histology explains only 0.7 to 0.8 mm (2x0.3 mm+25%) of the 2.6-mm late loss in lumen diameter seen by in vivo angiography. That is, neointima accounts for no more than 30% of the late luminal loss, whereas constrictive vascular remodeling is responsible for at least 70% of the late loss in this (re)stenosis model. Compared with the normal predilated lumen size (3.4 mm), the calculated neointima thickness (0.7 to 0.8 mm) explained only a fraction of the observed loss in lumen (1.8 mm). Consequently, the vessel had shrunk, not only from immediately after the procedure to follow-up but also from before dilatation.

Remodeling is also important for late lumen size in the atherosclerotic rabbit restenosis model^{17 18} and some porcine models,^{19 20 31} and recent data indicate that remodeling also may play a very important role in human restenosis.^{13 14 15 16} Mintz et al^{13 14 15 16} used intracoronary ultrasound in humans and found that remodeling accounted for 65% of the late luminal loss (area). Thus, our (re)stenosis model mimics the human experience observed by intracoronary ultrasound. We are currently investigating the temporal interplay between acute remodeling (overdilatation and subsequent recoil), late remodeling (chronic enlargement or contraction), and growth (acute thrombus, neointimal formation, and neoadventitial formation) in this porcine (re)stenosis model with angiography, angioscopy, intracoronary ultrasound, and histology.

Neoadventitia and Restenosis
It was suggested recently that adventitial fibrosis could play an important role in restenosis by compressing the vessel (like scar contraction) or preventing compensatory enlargement during healing after angioplasty.³² We found impressive neoadventitial formation 3 weeks after intervention, and our data suggest that one mechanism of restenosis could be circumferential neoadventitial shrinkage, "strangulating" the artery to late luminal narrowing.

Remodeling also takes place in native atherosclerotic human arteries and may cause both compensatory vessel enlargement³³ and arterial wall shrinkage.³⁴ An important difference between native human atherosclerotic arteries and our injured pig coronary arteries is that severe adventitial damage is usually present in the latter. The adventitial damage in the pig model triggers neoadventitial formation that may be followed by neoadventitial shrinkage. We hypothesize that restenosis in human coronary arteries may be triggered by the same mechanism. During successful angioplasty in humans, the rigid atherosclerotic plaque is separated from the more compliant vessel wall components (plaque-free wall, media, and adventitia) that are stretched, and a tear usually extends deeply into the vessel wall.^{35 36} The adventitial and perivascular damage caused by stretch and tears may trigger neoadventitial formation and shrinkage (constrictive remodeling), which may cause late luminal narrowing. Some published illustrations of human restenosis do, in fact, reveal adventitial thickening at sites of deep vessel wall injury.^{37 38}

Study Limitations
A major limitation of this and most other porcine (re)stenosis models is the lack of preexisting vessel wall disease, like atherosclerosis in humans. This concern may, however, not be great for neointimal formation after arterial injury if, as suggested by some postmortem findings, the reparative response in diseased human coronary arteries originates from the plaque-free or less diseased vessel wall rather than the diseased part of the wall.³⁶ Our porcine model involves tissue removal (medial abrasion) and may mimic debulking procedures like atherectomy better than balloon angioplasty. Medial and even adventitial tissue and plaque-free normal vessel wall are frequently retrieved by atherectomy,³⁹ indicating that deep and circumferential injury may follow this procedure in humans. To minimize the influence of differences in predilatation vessel size in late luminal outcome, we evaluated proximal, middle, and distal LAD segments separately. However, minor differences in vessel size must have been present even within individual segments before dilatation, but they cannot explain the main findings of the present study. Finally, all data from the same vascular segment were pooled and analyzed statistically with the assumption of independence although only nine measurements per segment (one from each of the nine pigs) were totally independent. The results, however, were confirmed by the separate analysis of the independent mean values (Fig 5), indicating that the conclusions drawn are valid.

Conclusions
A porcine coronary (re)stenosis model is described in which a deep and 360° circumferential injury was produced. The resultant healing responses narrowed the lumen, and a stenosis developed, clearly seen angiographically 2 and 4 weeks later. The stenoses were caused predominantly by constrictive vascular remodeling rather than neointimal formation, and adventitial processes could be of importance for the final outcome. This porcine coronary injury model may prove useful in testing the effect of treatments not only on postinterventional neointimal formation but also on arterial remodeling, stenosis development, and final lumen size—the most important late outcome parameter.

	Selected Abbreviations and Acronyms

 

LAD = left anterior descending coronary artery MLD = minimal lumen diameter PTCA = percutaneous transluminal coronary angioplasty SMC = smooth muscle cell

	Acknowledgments

 
The study was supported by grants from Aarhus University Research Foundation, The Danish Heart Foundation, Danish Foundation for the Advancement of Medical Science, and Bristol-Myers Squibb, Denmark. Nycomed DAK donated the acetylsalicylic acid. We want to thank the staff at the Institute of Experimental Clinical Research for good working conditions and Ulrik Baandrup, MD, PhD, Peter Grønning Olesen, MS, and Niels Christian Emmertsen, MS, for assistance with the morphometric analysis and serial angiography examinations.

Received August 16, 1995; revision received October 23, 1995; accepted November 3, 1995.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Faxon DP. Restenosis after angioplasty: what have we learned from clinical trials? Cardiol Rev. 1993;1:209-217.
   
2. Kuntz RE, Baim DS. Defining coronary restenosis: newer clinical and angiographic paradigms. Circulation. 1993;88:1310-1323. [Free Full Text]
   
3. Serruys PW, Foley DP, Kirkeeide RL, King SB. Restenosis revisited: insights provided by quantitative coronary angiography. Am Heart J. 1993;126:1243-1267. Editorial. [Medline] [Order article via Infotrieve]
   
4. Muller DW, Ellis SG, Topol EJ. Experimental models of coronary artery restenosis. J Am Coll Cardiol. 1992;19:418-432. [Abstract]
   
5. Ferrell M, Fuster V, Gold HK, Chesebro JH. A dilemma for the 1990s: choosing appropriate experimental animal model for the prevention of restenosis. Circulation. 1992;85:1630-1631. Editorial. [Free Full Text]
   
6. Karas SP, Gravanis MB, Santoian EC, Robinson KA, Anderberg KA, King SB. Coronary intimal proliferation after balloon injury and stenting in swine: an animal model of restenosis. J Am Coll Cardiol. 1992;20:467-474. [Abstract]
   
7. Bonan R, Paiement P, Scortichini D, Cloutier MJ, Leung TK. Coronary restenosis: evaluation of a restenosis injury index in a swine model. Am Heart J. 1993;126:1334-1340. [Medline] [Order article via Infotrieve]
   
8. Humphrey WR, Simmons CA, Toombs CF, Shebuski RJ. Induction of neointimal hyperplasia by coronary angioplasty balloon overinflation: comparison of feeder pigs to Yucatan minipigs. Am Heart J. 1994;127:20-31. [Medline] [Order article via Infotrieve]
   
9. Schwartz RS, Murphy JG, Edwards WD, Camrud AR, Vliestra RE, Holmes DR. Restenosis after balloon angioplasty: a practical proliferative model in porcine coronary arteries. Circulation. 1990;82:2190-2200. [Abstract/Free Full Text]
   
10. Schwartz RS, Huber KC, Murphy JG, Edwards WD, Camrud AR, Vlietstra RE, Holmes DR. Restenosis and the proportional neointimal response to coronary artery injury: results in a porcine model. J Am Coll Cardiol. 1992;19:267-274. [Abstract]
    
11. Cox DA, Anderson PG, Roubin GS, Chou CY, Agrawal SK, Cavender JB. Effect of local delivery of heparin and methotrexate on neointimal proliferation in stented procine coronary arteries. Coron Artery Dis. 1992;3:237-248.
    
12. Waksman R, Robinson KA, Crocker IR, Gravanis MB, Cipolla GD, King SB. 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. [Abstract/Free Full Text]
    
13. Mintz GS, Kovach JA, Javier SP, Ditrano CJ, Leon MB. Geometric remodeling is the predominant mechanism of late lumen loss after coronary angioplasty. Circulation. 1993;88(suppl I):I-654. Abstract.
    
14. Mintz GS, Kovach JA, Pichard AD, Kent KM, Satler LF, Popma JJ, Painter JA, Morgan K, Leon MB. Geometric remodeling is the predominant mechanism of clinical restenosis after coronary angioplasty. J Am Coll Cardiol. 1994;23(suppl A):138A. Abstract.
    
15. Mintz GS, Matar FA, Kent KM, Pichard AD, Satler LF, Harvey M, Ditrano CJ, Prunka N, Popma JJ, Leon MB. Chronic compensatory arterial dilatation following coronary angioplasty: an intravascular ultrasound study. J Am Coll Cardiol. 1994;23(suppl A):139A. Abstract.
    
16. Mintz GS, Pichard AD, Kent KM, Satler LF, Popma JJ, Leon MB. Intravascular ultrasound comparison of restenotic and de novo coronary artery narrowings. Am J Cardiol. 1994;74:1278-1280. [Medline] [Order article via Infotrieve]
    
17. Kakuta T, Currier JW, Haudenschild CC, Ryan TJ, Faxon DP. Differences in compensatory vessel enlargement, not intimal formation, account for restenosis after angioplasty in the hypercholesterolemic rabbit model. Circulation. 1994;89:2809-2815. [Abstract/Free Full Text]
    
18. Lafont A, Guzman LA, Whitlow PL, Goormastic M, Cornhill JF, Chisolm GM. Restenosis after experimental angioplasty: intimal, medial, and adventitial changes associated with constrictive remodeling. Circ Res. 1995;76:996-1002. [Abstract/Free Full Text]
    
19. Post MJ, Borst C, Kuntz RE. The relative importance of arterial remodeling compared with intimal hyperplasia in lumen renarrowing after balloon angioplasty: a study in the normal rabbit and the hypercholesterolemic Yucatan micropig. Circulation. 1994;89:2816-2821. [Abstract/Free Full Text]
    
20. Nunes GL, Sgoutas DS, Redden RA, Sigman SR, Gravanis MB, King SB, Berk BC. Combination of vitamins C and E alters the response to coronary balloon injury in the pig. Arterioscler Thromb Vasc Biol. 1995;15:156-165. [Abstract/Free Full Text]
    
21. Gundersen HJG, Bendtsen TF, Korbo L, Marcussen N, Møller A, Nielsen K, Nyengaard JR, Pakkenberg B, Sørensen FB, Vesterby A, West MJ. Some new, simple and efficient stereological methods and their use in pathological research and diagnosis. APMIS. 1988;96:379-394. [Medline] [Order article via Infotrieve]
    
22. Gundersen HJG, Jensen EB. The efficiency of systematic sampling in stereology and its prediction. J Microscopy. 1987;147:229-263. [Medline] [Order article via Infotrieve]
    
23. Jorgen H, Gundersen G, Boysen M, Reith A. Comparison of semiautomatic digitizer-tablet and simple point counting performance in morphometry. Virchows Arch B Cell Pathol Incl Mol Pathol. 1981;37:317-325. [Medline] [Order article via Infotrieve]
    
24. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurements. Lancet. 1986:1:307-310.
    
25. van der Giessen WJ, Serruys PW, van Beusekom HM, van Woerkens LJ, van Loon H, Soei LK, Strauss BH, Beatt KJ, Verdouw PD. Coronary stenting with a new, radiopaque, balloon-expandable endoprosthesis in pigs. Circulation. 1991;83:1788-1798. [Abstract/Free Full Text]
    
26. Haude M, Erbel R, Issa H, Meyer J. Quantitative analysis of elastic recoil after balloon angioplasty and after intracoronary implantation of balloon-expandable Palmaz-Schatz stents. J Am Coll Cardiol. 1993;21:26-34. [Abstract]
    
27. Barath P, Fishbein MC, Vari S, Forrester JS. Cutting balloon: a novel approach to percutaneous angioplasty. Am J Cardiol. 1991;68:1249-1252. [Medline] [Order article via Infotrieve]
    
28. Bonan R, Joyal M, Untenberg C, Zeiher A, LaBlanche JM, Bertrand M, Specchia G, de Jaegere P, Ruygrok P. Multicentric experience with a cutting balloon for percutaneous transluminal coronary angioplasty: in-hospital results and restenosis rate. Eur Heart J. 1994;15(suppl):335. Abstract.
    
29. Santoian EC, Gravanis MB, Karas SP, Schneider JE, Anderberg KA, Scott NA, Cipolla G, King SB. Sequence of the reparative phenomena and development of smooth muscle proliferation following coronary angioplasty in a swine restenosis model. Clin Res. 1992;40:364A. Abstract.
    
30. Falk E. Coronary artery narrowing without irreversible myocardial damage or development of collaterals: assessment of `critical' stenosis in a human model. Br Heart J. 1982;48:265-271. [Abstract/Free Full Text]
    
31. Brott BC, Labinaz M, Culp SC, Zidar JP, Fortin DF, Virmani R, Phillips HR, Stack RS. Vessel remodeling after angioplasty: comparative anatomic studies. J Am Coll Cardiol. 1994;23(suppl A):138A. Abstract.
    
32. Isner JM. Vascular remodeling: honey, I think I shrunk the artery. Circulation. 1994;89:2937-2941. Editorial. [Free Full Text]
    
33. Zarins CK, Weisenberg E, Kolettis G, Stankunavicius R, Glagov S. Differential enlargement of artery segments in response to enlarging atherosclerotic plaques. J Vasc Surg. 1988;7:386-394. [Medline] [Order article via Infotrieve]
    
34. Pasterkamp G, Wensing PJW, Post MJ. Hillen B, Mali WPTM, Borst C. Paradoxical arterial wall shrinkage may contribute to luminal narrowing of human atherosclerotic femoral arteries. Circulation. 1995;91:1444-1449. [Abstract/Free Full Text]
    
35. Nobuyoshi M, Kimura T, Ohishi H, Horiuchi H, Nosaka H, Hamasaki N, Yokoi H, Kim K. Restenosis after percutaneous transluminal coronary angioplasty: pathologic observations in 20 patients. J Am Coll Cardiol. 1991;17:433-439. [Abstract]
    
36. Gravanis MB, Roubin GS. Histopathologic phenomena at the site of percutaneous transluminal coronary angioplasty: the problem of restenosis. Hum Pathol. 1989;20:477-485. [Medline] [Order article via Infotrieve]
    
37. Marx J. CMV-p53 interaction may help explain clogged arteries. Science. 1994;265:320. Editorial. [Free Full Text]
    
38. Essed CE, van den Brand M, Becker AE. Transluminal coronary angioplasty and early restenosis: fibrocellular occlusion after wall laceration. Br Heart J. 1983;49:393-396.[Abstract/Free Full Text]
    
39. Kuntz R, Hinohara T, Safian R, Selmon MR, Simpson JB, Baim DS. Restenosis after directional coronary atherectomy: effects of luminal diameter and deep wall excision. Circulation. 1992;86:1394-1399.[Abstract/Free Full Text]
    

This article has been cited by other articles:

				S. Kawatsu, K. Oda, Y. Saiki, Y. Tabata, and K. Tabayashi External Application of Rapamycin-Eluting Film at Anastomotic Sites Inhibits Neointimal Hyperplasia in a Canine Model Ann. Thorac. Surg., August 1, 2007; 84(2): 560 - 567. [Abstract] [Full Text] [PDF]

				C. Deiner, E. Shagdarsuren, P. L. Schwimmbeck, P. Rosenthal, C. Loddenkemper, U. Rauch, M. Pauschinger, R. Dietz, H.-P. Schultheiss, R. Dechend, et al. Nf-{kappa}b and AP-1 activation is associated with late lumen loss after porcine coronary angioplasty and antiproliferative beta-irradiation Cardiovasc Res, July 1, 2007; 75(1): 195 - 204. [Abstract] [Full Text] [PDF]

				K D Krueger, A K Mitra, M G DelCore, W J Hunter III, and D K Agrawal A comparison of stent-induced stenosis in coronary and peripheral arteries J. Clin. Pathol., June 1, 2006; 59(6): 575 - 579. [Abstract] [Full Text] [PDF]

				H. Shimokawa and A. Takeshita Rho-Kinase Is an Important Therapeutic Target in Cardiovascular Medicine Arterioscler. Thromb. Vasc. Biol., September 1, 2005; 25(9): 1767 - 1775. [Abstract] [Full Text] [PDF]

				K. Wang, X. Zhou, Z. Zhou, N. Mal, L. Fan, M. Zhang, A. M. Lincoff, E. F. Plow, E. J. Topol, and M. S. Penn Platelet, Not Endothelial, P-Selectin Is Required for Neointimal Formation After Vascular Injury Arterioscler. Thromb. Vasc. Biol., August 1, 2005; 25(8): 1584 - 1589. [Abstract] [Full Text] [PDF]

				O. D. Defawe, R. D. Kenagy, C. Choi, S. Y.C. Wan, C. Deroanne, B. Nusgens, N. Sakalihasan, A. Colige, and A. W. Clowes MMP-9 regulates both positively and negatively collagen gel contraction: A nonproteolytic function of MMP-9 Cardiovasc Res, May 1, 2005; 66(2): 402 - 409. [Abstract] [Full Text] [PDF]

				M. Nakayama, M. Amano, A. Katsumi, T. Kaneko, S. Kawabata, M. Takefuji, and K. Kaibuchi Rho-kinase and myosin II activities are required for cell type and environment specific migration Genes Cells, February 1, 2005; 10(2): 107 - 117. [Abstract] [Full Text] [PDF]

				B. Somoza, M. C. Gonzalez, J. M. Gonzalez, F. Abderrahim, S. M. Arribas, and M. S. Fernandez-Alfonso Modulatory role of the adventitia on noradrenaline and angiotensin II responses: Role of endothelium and AT2 receptors Cardiovasc Res, February 1, 2005; 65(2): 478 - 486. [Abstract] [Full Text] [PDF]

				R. Wyttenbach, A. Gallino, M. Alerci, F. Mahler, L. Cozzi, M. Di Valentino, J. J. Badimon, V. Fuster, and R. Corti Effects of Percutaneous Transluminal Angioplasty and Endovascular Brachytherapy on Vascular Remodeling of Human Femoropopliteal Artery by Noninvasive Magnetic Resonance Imaging Circulation, August 31, 2004; 110(9): 1156 - 1161. [Abstract] [Full Text] [PDF]

				Y.-H. Chen, L.-Y. Chau, M.-W. Lin, L.-C. Chen, M.-H. Yo, J.-W. Chen, and S.-J. Lin Heme oxygenase-1 gene promotor microsatellite polymorphism is associated with angiographic restenosis after coronary stenting Eur. Heart J., January 1, 2004; 25(1): 39 - 47. [Abstract] [Full Text] [PDF]

				P. Urban, P. Serruys, D. Baumgart, A. Colombo, S. Silber, E. Eeckhout, A. Gershlick, K. Wegscheider, L. Verhees, R. Bonan, et al. A multicentre European registry of intraluminal coronary beta brachytherapy Eur. Heart J., April 1, 2003; 24(7): 604 - 612. [Abstract] [Full Text] [PDF]

				M. S Conte, G. A VanMeter, L. M Akst, T. Clemons, M. Kashgarian, and J. R Bender Endothelial cell seeding influences lesion development following arterial injury in the cholesterol-fed rabbit Cardiovasc Res, February 1, 2002; 53(2): 502 - 511. [Abstract] [Full Text] [PDF]

				A. H. M. Hassan, I. M. Lang, M. Ignatescu, R. Ullrich, D. Bonderman, P. Wexberg, F. Weidinger, and H. D. Glogar Increased intimal apoptosis in coronary atherosclerotic vessel segments lacking compensatory enlargement J. Am. Coll. Cardiol., November 1, 2001; 38(5): 1333 - 1339. [Abstract] [Full Text] [PDF]

				E. Jorgensen, H. Kelbaek, S. Helqvist, G. V. H. Jensen, K. Saunamaki, J. Kastrup, O. Havndrup, H. Bundgaard, J. Kyst Madsen, M. Christiansen, et al. Predictors of coronary in-stent restenosis: importance of angiotensin-converting enzyme gene polymorphism and treatment with angiotensin-converting enzyme inhibitors J. Am. Coll. Cardiol., November 1, 2001; 38(5): 1434 - 1439. [Abstract] [Full Text] [PDF]

				M. R. Ward, P. Kanellakis, D. Ramsey, J. Funder, and A. Bobik Eplerenone Suppresses Constrictive Remodeling and Collagen Accumulation After Angioplasty in Porcine Coronary Arteries Circulation, July 24, 2001; 104(4): 467 - 472. [Abstract] [Full Text] [PDF]