Atherosclerotic Arterial Remodeling in the Superficial Femoral Artery
Individual Variation in Local Compensatory Enlargement Response
Background In previous studies on atherosclerotic arterial remodeling, compensatory enlargement of the artery in response to plaque accumulation was inferred from pooled data based on one cross section per artery. We assessed local arterial remodeling individually by analyzing 45 artery segments at 0.5-cm intervals over a length of 10 to 15 cm.
Methods and Results Twenty patients were studied by 30-MHz intravascular ultrasound (IVUS) before balloon angioplasty of the superficial femoral artery (370 cross sections), and 25 femoral artery segments were studied postmortem (551 cross sections). In each cross section, the area surrounded by the internal elastic lamina (IEL area) and the plaque area were measured. The IEL area was larger in the cross section with the largest plaque area than in the cross section with the smallest plaque area (32.5±13.0 and 32.0±11.5 mm2 versus 28.9±9.7 [P=NS] and 26.7±10.1 [P<.05] mm2 for IVUS and histology, respectively [mean±SD]). A significant positive correlation was found between plaque area and IEL area for the pooled data (r=.61 and r=.47 and slope=1.07 and 0.90 for IVUS and histology, respectively; both P<.001). In 12 of 20 and 16 of 25 individual arterial segments, however, no significant correlation was observed between plaque area and IEL area for IVUS and histology, respectively. A large variation was found in the correlation of the regression of plaque to IEL area (IVUS, r=−.40 to .89; histology, r=−.13 to .91) and slope (IVUS, −0.28 to 1.29; histology, −0.18 to 1.32).
Conclusions In the majority of atherosclerotic femoral arteries, significant compensatory enlargement could not be determined. It is inferred that arterial remodeling in response to plaque formation may vary among individuals.
Arteries remodel during the development of atherosclerotic lesions. To preserve the luminal area, coronary1 2 3 4 5 and femoral6 7 arteries may undergo compensatory enlargement in response to plaque accumulation. In most studies, the positive correlation in cross sections between the plaque area and the area encompassed by the internal elastic lamina (IEL area) was inferred from data pooled from many patients.1 2 3 4 However, pooled data may obscure a heterogeneous response in individual arteries. We recently reported that in the femoral artery, in contrast to compensatory enlargement, a paradoxical type of arterial wall remodeling, ie, shrinkage, may be observed locally.7 In the present study, we hypothesized that compensatory enlargement is not a general finding in individual arterial segments.
To acquire insight into individual variability of local arterial remodeling of de novo lesions, we assessed local changes in arterial size in response to focal plaque formation in 10- to 15-cm atherosclerotic femoral artery segments using in vivo intravascular ultrasound (IVUS) and postmortem histology.
Twenty-five femoral arteries were taken from donated corpses (11 men, 14 women; age, 79.9±8.1 years [mean±SD]). Four arteries were obtained within 24 hours postmortem, filled with glycolmethacrylate Technovit 7001 under a physiological pressure of 90 mm Hg, and fixed in formalin 4%, pH 7.4. Twenty-one femoral arteries were taken from corpses that had already been fixed in formalin 4%, pH 7.4.
To minimize any influence of anatomic tapering of the artery, segments 10 to 15 cm long were selected from 12 cm proximal to the adductor hiatus to 3 cm distal to the adductor hiatus. No major side branches originated in this area. The femoral arteries were decalcified with EDTA 10% in 5 days and dehydrated in an alcohol sequence ranging from 70% to 100%. From 7 arterial segments embedded in Technovit 7100 hydroxyethylmethacrylate, histological cross sections were obtained every 0.5 cm and stained with Verhoeff’s elastic tissue stain. Eighteen arterial segments were embedded in a mixture of liquefied polyethylene glycol 1000 and polyethylene glycol 400 in a 4:1 ratio, cut into 0.5-cm-long parts, stained with Lawson’s elastic tissue stain, and studied under magnification. All microscopic images of the cross sections were recorded on super VHS videotape with a Sony video camera (3 CCD) for further image analysis. A ruler was used for distance calibration. A total of 580 histological cross sections were obtained. Twenty-nine cross sections were excluded because of cutting artifacts. Thus, quantitative analysis was performed on 551 histological cross sections.
Twenty patients (14 men, 6 women; age, 65.4±8.2) were studied by IVUS before routine balloon angioplasty of the superficial femoral artery to treat disabling claudication. Six patients did not take any medication before hospitalization. In the remaining 14 patients, the following medications were taken: carbasalaatcalcium (n=6), acenocoumarol (n=6), β-blockade (metoprolol [n=5] or bisoprolol [n=1]), calcium antagonists (amlodipine [n=1], nifedipine [n=1], and diltiazem [n=1]), ACE inhibitor (captopril [n=2]), and antidiabetic treatment (insulin [n=1] and glibenclamide [n=1]). Only de novo atherosclerotic lesions were studied. Informed consent was obtained from all patients. A 4.1F IVUS mechanical catheter was used (30-MHz transducer rotating at 1000 rpm; axial resolution, 100 μm; lateral resolution, 200 μm; DuMED). The resulting images were displayed on a monitor by means of a video-scanned memory and recorded on super VHS videotape.
In all patients, a series of cross-sectional images was recorded during pullback of the IVUS catheter. To localize the IVUS catheter, a radiopaque ruler and a distance sensing device8 9 were used as reference during fluoroscopy in 17 and 3 patients, respectively. When vasospasm was suspected angiographically, 0.2 mg nitroglycerin was administered through the introducer sheath (n=4). Nitroglycerin was not administered routinely. All stenosis sites visualized with IVUS were detected previously with duplex measurements. Therefore, it was unlikely that these stenoses were caused by vasospasm elicited by the IVUS catheter.
To study individual differences in the relation between plaque area and the area encompassed by the echo-lucent media, ultrasound images were selected every 0.5 cm in arterial segments of ≈10 cm length. A total of 400 IVUS images were obtained from the stenosed femoral segments. Thirty IVUS images were excluded for further analysis owing to the presence of side branches or excessive calcification. Thus, quantitative analysis was performed on 370 IVUS cross-sectional images.
Histological sections and IVUS images recorded on videotape were analyzed with a digital video analyzer as described previously.10 In short, during image acquisition, video signals from a VHS videotape were converted with a frame grabber (Brand) to 512×512×8-bit digital image data and stored on a personal computer. In the IVUS image, we traced the lumen area and the IEL area. The latter is the area encompassed by the interface between the echo-dense intimal layer and the echo-lucent media. The circumferential outlines were processed to produce a smoothed and closed contour. The cross-sectional lumen area, IEL area, and plaque area were determined automatically after the two contours were completed. Plaque area was calculated by subtracting the lumen area from the IEL area. Intraobserver and interobserver variabilities in lumen area and plaque area measurements were assessed in 22 arteries. No significant observer bias was present (paired differences were not significantly different from zero). In lumen area measurements, observer variation (SD of the paired differences) was smaller than in lesion area measurements (0.6 and 1.1 mm2 and 1.2 and 1.9 mm2 for intraobserver and interobserver variations for lumen and lesion areas, respectively). Intraobserver and interobserver variations in measurements of percent area obstruction were 2.6% and 5.6%, respectively.
All measured values are presented as mean±SD. The relation between plaque area and IEL area was studied by linear regression analysis, first on the pooled data and then separately on each individual arterial segment. Paired Student’s t test was used to compare measured values in the cross section with the largest plaque area and the cross section with the least amount of plaque obtained from the same arterial segment. A value of P<.05 was considered statistically significant.
The IVUS measurements in patients with claudication corresponded well to the postmortem observations (Table 1⇓). In a paired comparison in the individual artery, in the cross section with the largest amount of plaque, the IEL area enlarged from 3.6 to 5.3 mm2 and the lumen area decreased from 11.3 to 6.6 mm2 compared with the cross sections with the least amount of plaque for IVUS and histology, respectively (Table 1⇓). Regression analysis for all pooled data revealed a positive correlation between plaque area and IEL area in the femoral artery (Fig 1⇓).
Individual Regression Analysis
In the individual artery, a large variation in correlation (IVUS, r=−.40 to .89; histology, r=−.13 to .91) and slope (IVUS, −0.28 to 1.29; histology: −0.18 to 1.32) between plaque area and IEL area was observed (Table 2⇓). In only 8 of 20 and 9 of 25 arterial segments, the correlation between plaque area and IEL area was significant (P<.05) for IVUS and histology, respectively (Fig 2⇓, top). In the remaining arterial segments, no significant correlation was found (P>.05; Fig 2⇓, bottom).
The linear regression lines of the individual arterial segments (Table 2⇑) and the regression line calculated for the pooled data (Fig 1⇑) are shown together in Fig 3⇓. This graph illustrates that in 19 and 21 arterial segments (95% and 84%, IVUS and histology, respectively), the slope of the individual artery linear regression was smaller than that calculated for the pooled data. Furthermore, Fig 3⇓ illustrates that in the cross sections with the least amount of plaque (0.0 to 11.0 and 1.2 to 13.3 mm2 [Table 2⇑]), the IEL area increased by 2.88 and 1.53 mm2 for each 1-mm2 increase in plaque area for IVUS and histology, respectively.
In the cross sections with the least amount of plaque (Table 1⇑), the IEL areas were classified into the two groups listed in Table 3⇓. The groups consist of cross sections obtained from arterial segments with and without a significant positive correlation between plaque and IEL area. In arterial segments with a significant positive correlation, the IEL area (ie, artery size) was smaller than in arterial segments without a significant correlation (P=.03 and P=.02 for IVUS and histology, respectively).
In addition to compensatory enlargement of the coronary1 2 3 4 5 and femoral6 7 arteries in response to plaque formation, we recently observed in the femoral artery that paradoxical arterial shrinkage may contribute to atherosclerotic luminal narrowing.7 Compensatory enlargement will retard but arterial wall shrinkage will accelerate luminal narrowing by plaque formation. In hemodynamically significant stenoses (>50% diameter stenosis), arterial shrinkage dominated compensatory enlargement.7
Compensatory enlargement has previously been inferred from pooled data based on one cross section per arterial segment.1 2 6 In the present study, we focused on individual arterial segments analyzed at 0.5-cm intervals. From a set of 55 arterial segments reported on earlier,7 we selected 45 segments on the basis of the large variation in plaque load over the length of the segment. The principal findings are as follows: (1) A significant positive correlation between plaque area and IEL area was observed in only 17 of 45 (39%) of the superficial femoral arteries; (2) in 40 of 45 arterial segments (89%), the slope of the individual regression line between plaque area and IEL area was smaller than the slope calculated from the pooled data; and (3) segments with a significant positive correlation between plaque area and IEL area were smaller than segments without a significant correlation.
Relation Between Plaque Area and IEL Area
From coronary postmortem measurements, Glagov et al1 inferred a model for compensatory enlargement in which the artery enlarges until, on average, ≈40% of the IEL area is occupied by plaque. In the early phase, even luminal overcompensation may be observed. When the plaque exceeds ≈40% of the IEL area, however, further plaque growth will encroach on the lumen.
Glagov et al1 and other investigators2 6 based their inference of compensatory enlargement on pooled data. In the present study, in which the IEL area was plotted versus plaque area, individual segments also were analyzed. For the pooled data, a slope of 1.07 and 0.88 was found between plaque area and IEL area for IVUS and histology, respectively. This is consistent with the slopes observed in previous studies.1 2 6 It is unlikely, however, that a slope of 1.07 for the IVUS data, obtained from patients to be treated for claudication, is based solely on an increase in the IEL area in response to plaque increase because it implies that on average, compensatory enlargement was >100% effective. Such effective compensatory enlargement would be incompatible with symptomatic disease. We hypothesized that pooling data from all subjects may mask an individual variation in the susceptibility of the artery to undergo compensatory enlargement and any relation between artery size and minimum plaque load. Fig 3⇑ illustrates both aspects. First, in 19 of 20 and 21 of 25 arterial segments, the slope of the regression line between plaque area and IEL area was smaller than for the pooled data, implying that the extent to which the artery enlarges in response to plaque formation differs for individual segments. Second, slopes of 2.88 and 1.53 were observed for the relation between plaque area and IEL area in the cross sections that contained the least amount of plaque for IVUS and histology, respectively, implying a large variation in vessel size. The cross section with the least amount of plaque is probably least affected by arterial remodeling and therefore best approximates the original lumen. Thus, two independent variables account for the positive relation between plaque area and IEL area in a pooled data set: compensatory enlargement and individual variation in vessel size. Therefore, a positive relation between plaque area and IEL area with a large slope cannot be attributed to compensatory enlargement solely without correction for differences in individual arterial size (see “Appendix”).
The IVUS studies were performed on patients with advanced, symptomatic atherosclerotic disease. It may be argued that the response of compensatory enlargement was overwhelmed by atherogenesis, implying the disappearance of any correlation between plaque area and IEL area. However, for the IVUS data, 222 of 370 images (60%) demonstrated <50% histological stenosis; ie, <50% of the IEL area was occupied by plaque. For the histological data, only 382 of 551 images (69%) demonstrated <50% histological stenosis. Thus, the early and advanced stages of the atherosclerotic process have been investigated in the present study.
Losordo et al6 compared the IEL areas of cross sections obtained from an atherosclerotic femoral artery segment and a nondiseased, proximally obtained cross section and found an increase of the IEL area at the atherosclerotic cross section compared with the nondiseased site. In our IVUS data set, no significant increase was observed between the IEL area of a proximally or distally located cross section with the least amount of plaque and the cross section with the largest plaque load. An explanation for these conflicting results may be that in the present study, the IEL area of a reference site, ie, the cross section with the least amount of plaque, was compared with the IEL area in the cross section with the largest plaque load and a substantial luminal narrowing. In contrast, Losordo et al6 did not specifically use the cross section with maximal plaque load or maximal luminal narrowing for comparison. When a cross section with maximal plaque load and significant luminal narrowing is used, it is to be expected that in addition to compensatory enlargement, paradoxical shrinkage may be observed.7 Thus, the group of cross sections with the largest plaque load is apparently comprised of two subgroups of cross sections with different types of remodeling: compensatory enlarged and shrunken.
The present results demonstrate that in individual femoral artery segments, a local increase of the IEL area may be observed in response to plaque accumulation. However, in only 17 of 45 segments (39%), a significant correlation between plaque area and IEL area was observed, implying that both local and systemic factors may play an important role in the process of arterial enlargement. Thus, local variation in arterial wall remodeling (ie, compensatory enlargement or paradoxical shrinkage) may be observed despite the fact that on average, the artery enlarges or fails to enlarge. In some arterial segments, a slope of the regression line was calculated with a large 95% CI, suggesting that either a large variation in arterial size or a nonlinear relation was present. From the individual data point plots, we inferred that the former was generally the case.
It has been suggested that eccentric lesions, with a partly disease-free wall, are more prone to compensatory enlargement.1 In contrast to concentric lesions, the disease-free wall in an eccentric lesion can still show endothelium-dependent dilation11 in response to a flow increase.12 However, it is still unknown whether a relation exists between the type and degree of arterial remodeling and lesion eccentricity.
A difference in IEL area was observed between arterial segments that demonstrated a significant correlation between plaque area and IEL area and those that did not. The reason for the difference in arterial size among groups is unknown. Hemodynamic changes may occur sooner in small vessels compared with large vessels exposed to the same amount of plaque. Therefore, the compensatory adaptive mechanism of the vessel to enlarge may be enhanced in small vessels. Our findings are supported by Zarins et al,2 who observed that compensatory enlargement is more pronounced in the small distal coronary artery compared with the large proximal parts of the coronary artery. Another explanation is that the group without a significant correlation may comprise arterial segments in a later stage of atherosclerosis that may subsequently be maximally enlarged. However, for the cross sections with the smallest plaque area and IEL area, the percentage of IEL area occupied by plaque did not differ between arterial segments with and without a significant correlation between plaque and IEL area (Table 3⇑). This implies that the arterial segments were in the same stage of the atherosclerotic process.
A number of potential limitations of the present study need to be addressed. Only 4 of the postmortem arteries were pressure fixed. As a result, the lumen area and IEL area may be underestimated in 21 non–perfusion-fixed arterial segments. However, the results obtained with histology are corroborated by the results obtained with IVUS studies.
Arterial tapering might influence the differences in IEL area between cross sections in one segment. However, the IEL area in the most proximal and most distal cross sections did not differ significantly (30.0±11.9 versus 28.9±10.4 mm2 and 27.8±12.2 versus 26.4±10.8 mm2 for IVUS and histology, respectively).
The present study was performed in the superficial femoral artery. Therefore, these findings may not be generalizable, for example, to the coronary circulation.
Arterial remodeling occurs over time. In the present study, the IEL area was related to the plaque area through the use of different points in the same artery with the assumption that the different stages of the atherosclerotic process within one artery are representative for changes in vessel size over time. Future serial studies are needed to determine whether this assumption is correct.
Understanding of the mechanisms responsible for local and individual variabilities of arterial wall remodeling may aid in the development of new therapeutic strategies to prevent luminal narrowing by de novo atherosclerotic lesions. The remarkable local and individual variabilities in arterial remodeling in response to plaque accumulation may have a bearing on the early and late results of angioplasty.
A significant positive correlation between plaque area and IEL area was observed in the minority of the superficial femoral arteries. Arterial remodeling in response to plaque accumulation appears to vary among individuals. The mechanisms of local atherosclerotic arterial remodeling remain to be investigated.
Vessel Size and the Slope Between IEL Area and Plaque Area
Arteries may vary in size. In Fig 4A⇓, schematic cross sections of three arteries without plaque are shown. In Fig 4B⇓, the same arteries are depicted with 0.5-mm intimal thickening (plaque) and no change in IEL area. The relation between IEL area (vessel size) and plaque area is depicted in Fig 4C⇓, in which single measurements from the three different arteries have been pooled.
From Fig 4C⇑, it might be inferred that in the plaque area range of 5.5 to 8.5 mm2, the artery expands by 5.2 mm2 for every 1-mm2 increase in plaque area. This may be interpreted incorrectly as being a result of gross overcompensation in artery size in response to plaque accumulation. With more plaque accumulation, the slope of the regression line decreases and reaches 1.0 when total occlusion occurs without any change in IEL area. The slope of the regression line of IEL area versus plaque area becomes <1.0 if and only if plaque accumulates without 100% compensatory enlargement in vessels of the same original size. The influence of (over)compensatory enlargement is due to the substantial bias introduced by pooling single measurements from different vessels with originally different sizes (ie, without plaque). Implicit to the current interpretation of the regression line of IEL area versus plaque area by use of pooled data1 is the assumption that the arteries originally (before the onset of atherosclerosis) had the same size. This assumption probably is unfounded. The bias introduced by arteries of originally different size tends to amplify or introduce (Fig 4C⇑) compensatory enlargement, as inferred from the plot of IEL area versus plaque area.
A second approach was chosen to illustrate that the regression of plaque area versus IEL area is confounded by the arterial size. In addition to plaque area, two measures of arterial size were added as variable in a multivariate regression model: the mean IEL area of all cross sections of one artery and the IEL area of the reference cross section of that artery. These variables were determined for each artery and alternatively put into the multivariate regression model. For both the IVUS and histology data, we performed four subsequent regression analyses: (1) multivariate regression with plaque area and mean IEL area, (2) regression with plaque area, (3) multivariate regression with plaque area and reference IEL area, and (4) regression with plaque area, with the reference cross section excluded. For each regression, the F statistic was calculated. The difference between the F statistic from the multivariate and (plaque area) monovariate models quantifies the influence of arterial size on the regression. Table 4⇓ shows the results of these multiple regression analyses. For all four multivariate regression analyses, F statistics of plaque area and mean IEL area or reference IEL area were larger than for plaque area (P<.0001). Thus, arterial size as defined by the average IEL area per artery or by the IEL area of the reference cross section had more impact on the regression of IEL area on plaque area than plaque area itself.
This study was supported by a grant from the Netherlands Heart Foundation (grant 92,136).
- Received November 13, 1995.
- Revision received December 8, 1995.
- Accepted December 19, 1995.
- Copyright © 1996 by American Heart Association
Stiel GM, Stiel LSG, Schofer J, Donath K, Mathey DG. Impact of compensatory enlargement of atherosclerotic coronary arteries on angiographic assessment of coronary artery disease. Circulation. 1989;80:1603-1609.
Hausmann D, Mullen WL, Friedrich GJ, Fitzgerald PJ, Yock PG. Variability of remodeling in early coronary atherosclerosis: an intravascular ultrasound study. J Am Coll Cardiol. 1994;:72:175A. Abstract.
Losordo DW, Rosenfield K, Kaufman J, Pieczek A, Isner JM. Focal compensatory enlargement of human arteries in response to progressive atherosclerosis. Circulation. 1994;89:2570-2577.
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.
Gussenhoven EJ, van der Lugt A, van Strijen M, Li W, Kroeze H, The SHK, van Egmond FC, Honkoop J, Peters RJG, de Feyter P, van Urk H, Pieterman H. Displacement sensing device enabling accurate documentation of catheter tip position. In: Roelandt J, Bom N, Gussenhoven EJ, eds. Intravascular Ultrasound. Dordrecht, Netherlands: Kluwer Academic Publishers; 1993:157-166.
Pasterkamp G, Post MJ, Mali WPTM, Bom N, Borst C. From which segment of the artery is this IVUS cross-section? Two methods for localization: fluoroscopy and displacement transducer. J Am Coll Cardiol. 1993;21:192A. Abstract.
Hodgson JM, Marshall JJ. Direct vasoconstrictium and endothelium dependent vasodilatation: mechanism of acetylcholine effects on coronary blood flow and arterial diameter in patients with nonstenotic coronary arteries. Circulation. 1989;79:1043-1051.