The Impact of Atherosclerotic Arterial Remodeling on Percentage of Luminal Stenosis Varies Widely Within the Arterial System
A Postmortem Study
Abstract Luminal stenosis can be based on large atherosclerotic plaques in compensatory enlarged segments or on relatively little plaques in shrunken segments. In the present study, the contribution of plaque formation and remodeling to luminal narrowing was compared among six types of arteries prone to symptomatic atherosclerosis. Cross-sections (n=5195) were obtained at regular intervals from 329 arteries. For each artery, the cross-section that contained the least amount of plaque was considered to be the reference. For each cross-section, the percentage of lumen area decrease was expressed as a percentage of the lumen area at the reference site (luminal stenosis). Similarly, the area encompassed by the internal elastic lamina (IEL area) was expressed as a percentage of the IEL area at the reference site (relative IEL area). All cross-sections were categorized in three groups: relative IEL area >105% (enlargement), 95% to 105% (no remodeling), and <95% (shrinkage). The prevalence of enlargement (50% to 75%) was significantly higher compared with shrinkage (8% to 25%). Shrinkage was observed most frequently in the femoral arteries (25%) and infrequently in the renal arteries (8%). For all types of arteries, the relative IEL area correlated negatively with luminal stenosis (P<.001). Regression analysis of relative IEL area on luminal stenosis, however, showed significant differences in the first-order regression coefficients among artery types. On average, plaque increase was more compensated for by enlargement in the coronary, common carotid, and renal arteries compared with the arteries obtained from the lower extremities. Anatomic regional differences were observed in the impact of arterial wall remodeling on percent luminal stenosis in de novo atherosclerotic lesions.
- Received March 5, 1997.
- Accepted August 12, 1997.
Geometric remodeling in atherosclerosis comprises a change in total arterial circumference that, together with plaque growth, determines the lumen of the artery.1–5 Compensatory enlargement retards and paradoxical shrinkage accelerates luminal narrowing by plaque formation.3,5 The mechanisms that determine whether arterial segments shrink or enlarge are largely unknown. Compensatory enlargement in response to plaque accumulation has been demonstrated in vivo in the coronary artery,2,5 the femoral artery,3,4,6 and the carotid artery.7 Paradoxical shrinkage of the atherosclerotic artery has recently been described in the femoral3,8 and coronary5,9,10 circulation. Previously, we demonstrated that the decrease in luminal area cannot be attributed to plaque increase alone. Next to plaque accumulation, the variability in de novo atherosclerotic arterial remodeling was found to contribute significantly to luminal narrowing: compensatory enlargement initially prevents whereas paradoxical shrinkage accelerates narrowing of the lumen.3 It is unknown whether this relationship is characteristic of the femoral artery or whether it pertains to other clinically relevant arteries as well. Differences in the mode of remodeling among artery types may provide insight into physiological mechanisms that are related to the type and degree of atherosclerotic remodeling and may have consequences for therapeutic approaches. In the present study, the local remodeling response to plaque formation was compared among six types of arteries that are prone to develop atherosclerotic luminal narrowing. The impact of the mode and degree of remodeling and plaque accumulation on the percent luminal stenosis was examined in the coronary, common carotid, renal, common iliac, external iliac, and femoral arteries. We report that remodeling influences the lumen of all arteries with atherosclerotic lesions. However, the impact of remodeling on the percent luminal stenosis varies widely within the arterial system.
Postmortem, 20 hearts were removed from patients with causes of death unrelated to cardiac disease (13 men and 7 women, aged 71.2±12.4 years [mean±SD]). Within 24 hours after death the coronary arteries were pressure-fixed with 4% formalin at 80 mm Hg either in situ (n=10 hearts, 30 coronary arteries) or after dissection with ligation of all side branches (n=10 hearts, 30 coronary arteries) Coronary artery segments of 5 to 10 cm were dissected and cross-sections were obtained every 2 to 3 mm starting 1 cm distal from the ostium. One left anterior descending coronary artery and one left circumflex artery were lost due to severe calcifications when dissected.
Peripheral arteries of 29 donated corpses (17 men and 12 women, aged 80.1±7.9 years), were pressure-fixed within 24 hours after death with formaldehyde in situ (pressure, age+100 mm Hg). The right and left common carotid arteries (from the truncus brachiocephalicus and aortic arch to the carotid bifurcation), the femoral arteries, the common iliac arteries, the external iliac arteries, and the first 3 to 4 cm of the renal arteries (from the outer border of the aorta) were dissected. The femoral arteries were divided in 1-cm segments and the other arteries in 0.5-cm segments. The arterial cross-sections were numbered, ascending from proximal to distal. Five and three renal arteries were excluded because of the existence of multiple originating renal arteries and multiple cutting artifacts, respectively. One carotid artery was also excluded due to multiple cutting artifacts. Two common external iliac and femoral arteries were excluded because they had undergone surgery (aortobifemoro prosthesis). Two common iliac arteries were excluded due to aneurysm formation. Two totally occluded femoral arteries were also excluded from further analysis.
All cross-sections were stained with Lawson’s elastic tissue stain and studied under magnification. All microscopic images of the cross-sections were recorded on VHS videotape with a Sony videocamera (3 CCD) for further image analysis. A ruler was used for distance calibration.
Histological sections recorded on video tape were analyzed with a digital video analyzer as described previously.11 The lumen area and the area circumscribed by the internal elastic lamina (IEL area) were traced. The plaque area was calculated by subtracting the lumen area from the IEL area. For each artery type, cross-sections were pooled and the mean percent decrease of IEL area per centimeter (tapering) was calculated as follows: a linear regression analysis was performed between the IEL area and the distance. The slope of this regression line represents ΔIEL area (millimeters squared)/Δdistance (centimeters). The slope was divided by the mean IEL area at distance 0 cm to obtain a mean percent decrease of cross-sectional area of the vessel size over the distance. Values for percent vessel size tapering for the arterial segments under study compared with the vessel size at distance 0 cm were 6.3%, 3.1%, −4.1%, 1.1%, 3.7%, 0.6%, 0.2%, and 0.1% per centimeter for the left anterior descending, left circumflex, right coronary, common carotid, renal, common iliac, external iliac, and femoral arteries, respectively. Subsequently, measured values were corrected for tapering to make a valid comparison among the arterial cross-sections per artery. For the right coronary artery no correction for tapering was applied, since an increase of vessel size was observed in this group of arteries (ie, “reverse tapering”).
Calculations After Correction for Tapering
In each arterial segment, the cross-section that contained the least amount of plaque was chosen as a reference site, assuming that it had been least affected by remodeling.3,8 The reference site could be located proximally or distally within the artery. Percent lumen area stenosis was calculated by comparing the lumen area of every cross-section with the lumen area of the reference site3,8: (1−[lumen area/lumen area at the reference site])×100%. A positive value indicates luminal narrowing; a negative value indicates luminal dilatation or luminal “overcompensation.”
The mode and degree of atherosclerotic remodeling was calculated as (IEL area/IEL area at the reference site)×100%. A relative IEL area of >100% indicates compensatory enlargement, a relative IEL area of <100% indicates arterial wall shrinkage with respect to the reference site.
For each artery type, the mean plaque increase, mean IEL area increase, and mean lumen area decrease (compared with the reference site) were calculated. In addition, the mean percentage of plaque that encroached on the lumen was calculated as (mean lumen area decrease/mean plaque area increase)×100%.
All cross-sections were classified in three categories, based on type and degree of arterial remodeling as described previously8 (arbitrary division): (1) relative IEL area of >105% (enlargement), (2) relative IEL area ≤105% but >95% to 105% (failure of enlargement), and (3) relative IEL area <95% (shrinkage).
All values are expressed as mean±SD. A second-order polynomial regression analysis was performed between the relative IEL area and the percentage of luminal stenosis as described previously.3 For each regression line, the 95% confidence intervals (CIs) were calculated. Differences in variables between the remodeling groups were calculated using a one-way ANOVA with a posthoc Bonferroni correction. Differences in the prevalence of shrinkage and enlargement for the different artery types were calculated using the χ2 test. P<.05 was considered significant.
A total of 5195 cross-sections, obtained from 329 arteries, were analyzed. The lumen areas, IEL areas, and plaque areas of the cross-sections with the least (reference sites) and largest amount of plaque are shown in Table 1⇓. On average, 34±21%, 7±8%, 24±20%, 17±13%, 6±10%, and 21±20% of the area encompassed by the internal elastic lamina of the reference sites were occupied by plaque in the coronary, common carotid, renal, common iliac, external iliac, and femoral arteries, respectively.
Table 2⇓ demonstrates the numbers and percentages of cross-sections that were classified in the different remodeling groups. On average, 51±18%, 17±14%, 31±26%, 28±18%, 19±18%, and 40±19% of the area encompassed by the internal elastic lamina of all cross-sections were occupied by plaque in the coronary, common carotid, renal, common iliac, external iliac, and femoral arteries, respectively. The prevalence of enlargement (relative IEL area >105%) was significantly higher compared with shrunken lesions (relative IEL area <105%) (for all artery types P<.01). Overall, 50% to 75% of the cross-sections revealed enlargement (relative IEL area >105%) of the artery in response to plaque accumulation (Fig 1⇓, Table 2⇓). Shrinkage of the artery (relative IEL area <95%) was observed in 8% to 25% of the arteries and most frequently in the femoral arteries (Fig 2⇓ and Table 2⇓). On the other hand, shrinkage was infrequent in the renal arteries (8%). For all artery types the lumen area, IEL area, and plaque area were smaller in those cross-sections with a relative IEL area >95% compared with those cross-sections with a relative IEL area >105% (P<.05). Table 3⇓ shows that the distributions of shrunken and enlarged cross-sections were similar for the three coronary arteries.
The mean percentage of plaque increase that resulted in luminal narrowing is shown in Table 4⇓: in the renal artery only 2% of the total plaque increase was at the expense of the lumen area. In the femoral artery, 46% of plaque increase encroached on the lumen.
Fig 3⇓ illustrates the relation between the mode and degree of remodeling (relative IEL area) and the percent luminal narrowing for all artery types. For all artery types a significant negative relation was observed between the relative IEL area and the percent luminal narrowing. Thus, remodeling influenced the lumen of all arteries with atherosclerotic lesions. However, the impact of remodeling on the percent luminal stenosis varied widely within the arterial system. The most negative first-order regression coefficient of the regression line between the relative IEL area and the percent luminal stenosis, without overlap in 95% CIs compared with the first-order regression coefficient of the regression line for other artery types, was observed in the carotid artery (Fig 3⇓). The smallest first-order regression coefficient of the regression line between the relative IEL area and percent luminal stenosis was observed for the renal arteries. A significant negative relation was also observed for the three separate coronary artery types: left anterior descending: y=−0.52x+0.012x2+128.3, 95% CI of the first-order regression coefficient=−0.58 to −0.44; left circumflex: y=0.41x+0.041x2+125.0, 95%CI=−049 to −0.34; right coronary artery: y=−0.68x+0.002x2+126.4, 95%CI=−0.61 to −0.74.
Until recently, plaque formation was considered to be the only determinant of atherosclerotic luminal narrowing. Recent postmortem and intravascular ultrasound studies studies, however, revealed that arterial remodeling is another important determinant of luminal narrowing in de novo atherosclerosis. The change in total arterial circumference relative to a reference cross-section ranges from excessive enlargement2 with an actual increase in lumen to arterial shrinkage contributing to lumen narrowing.3 In the present study the impact of remodeling on luminal stenosis was compared among six artery types that are known to develop symptomatic atherosclerotic plaques.
The present study demonstrates that remodeling influences the lumen of all arteries with atherosclerotic lesions. Shrinkage of the artery in response to plaque formation was less frequently (8% to 25%) observed compared with compensatory enlargement (50% to 75%). Differences, however, were observed among artery types: shrinkage is frequently observed in the femoral artery but is infrequent in the renal artery. The impact of remodeling on the percent luminal stenosis varies within the arterial system: the negative first-order regression coefficient of the regression line of lumen stenosis on relative IEL area, which reflects the average influence of failure of enlargement or even shrinkage on luminal narrowing, was largest for the carotid artery and least for the renal artery. In addition, on average, plaque increase appeared largely compensated for by arterial enlargement in all artery types but is subject to regional variations: in the renal artery, on average, only 2% of the plaque increase was found to encroach on the lumen versus 46% in the femoral artery.
In the present study, arterial segments were collected without foreknowledge of their percentages of luminal stenosis. Most cross-sections revealed minor or no luminal stenosis. It was therefore to be expected that shrinkage would be observed in the minority of cross-sections since shrinkage is mostly found at those locations with significant luminal stenosis (Fig 3⇑ and Table 2⇑). In a previous study we investigated the prevalence of shrinkage and compensatory enlargement at the culprit lesion of the femoral artery.8 Shrunken lesions were found to be present in 54% of the cases. An increased prevalence of failure of enlargement at locations with severe luminal narrowing in the coronary artery was shown by Haussman et al,12 which confirms our observations in the femoral artery. In addition, Nishioka et al5 observed shrinkage of the arterial wall in >25% (9 of 35) of coronary arteries before angioplasty (14% in this study of randomly selected subjects). Thus, the prevalence of shrinkage at the lesion site may increase if cross-sections with severe luminal narrowing are particularly selected.
Variability of the impact of remodeling on the percent luminal stenosis was observed within the arterial system. The slope of the regression line between the degree of remodeling and the percent luminal stenosis is partially determined by the existence of variability of remodeling within an artery type. However, a less steep negative slope of the regression line of the relation of the relative IEL area versus luminal stenosis does not necessarily imply that remodeling would not influence the lumen. For instance, it may well be that for all cross-sections enlargement originally prevented luminal narrowing but that eventually the plaque encroached on the lumen: in that case, the slope of the polynomial regression line of relative IEL area versus percent luminal narrowing would initially be positive and eventually become parallel to the x-axis. The latter would reflect the model as described by Glagov et al.2 Thus, the type of remodeling has a twofold impact on the percent luminal narrowing: on one hand, the artery may enlarge thereby preventing luminal narrowing; on the other hand, shrinkage may accelerate luminal narrowing. If an artery is capable of developing both types of remodeled arteries, the impact of remodeling on the percent luminal narrowing will subsequently be more pronounced, which is expressed in the slope of the regression line of relative IEL area versus luminal stenosis.
Most plaque increase was compensated for by arterial enlargement in all artery types. Differences, however, were observed among the artery types. In the femoral artery, on average 54% of plaque increase was compensated for by arterial enlargement with or without luminal enlargement. This compensation was 74% for the coronary artery and 98% for the renal artery. In the renal artery, on average, only 2% of the plaque increase was at the expense of the luminal area. These numbers confirm the important role of remodeling in the process that leads to luminal narrowing. It should be emphasized, however, that these percentages are mean values and that large variations in remodeling response for each location and each individual may exist (Table 2⇑)4.
Previously, we reported that remodeling (enlargement or shrinkage) is a local phenomenon3 and that compensatory enlargement may be individually determined.4 The present study shows regional differences in the remodeling response. The peripheral arteries under study were dissected from the same individuals. Thus, individually related variables cannot explain the observed regional, artery-related variation of remodeling in response to plaque formation. The reason for the different remodeling responses among artery types is unknown. Two hypotheses are suggested. First, geometric variables that determine shear stress13–15 are probably related to plaque accumulation and luminal stenosis and may determine remodeling responses as well. Second, the common carotid, coronary, and renal arteries showed a larger tendency to enlarge in response to plaque formation compared with conduit arteries such as the iliac and femoral arteries (Tables 2⇑ and 4⇑). It may be speculated that compensatory enlargement is part of an autoregulatory mechanism that regulates blood flow to organs: a threshold for endothelium-dependent arterial enlargement in response to an increased shear stress may be lower for arteries supplying the brain, heart, and kidney compared with lower extremity arteries. Regional heterogeneity of endothelial responses to alterations in shear stress has recently been demonstrated by Walpola et al,16 which may support this hypothesis.
This is a descriptive study and we can, therefore, only speculate on mechanisms responsible for the differences in remodeling response among the artery types.
The choice of the reference site is of crucial importance. It may well be that the reference site itself has been shrunken or enlarged. However, the relative changes in IEL area were also reflected by absolute changes of lumen area, plaque area, and IEL area (Table 2⇑). In a previous study in the femoral artery,8 the IEL area of the reference sites did not differ among the three remodeling groups (relative IEL area <95%, 95% to 105%, and >105%), indicating that variation of the relative IEL area was due to variability of the values obtained at the lesion sites.
One reference site was selected per artery. In the present study all cross-sections were studied at regular intervals in contrast with previous studies in which only lesion sites with significant luminal narrowing were addressed, allowing the selection of one or two reference sites for each lesion.5,8 The lumen area and vessel area of one single reference site may not be representative for all other locations throughout the artery due to tapering or the presence of side branches.17 Before selection of the reference site, cross-sectional areas were corrected for tapering to make a valid comparison of values among cross-sections. Additionally, cross-sections that were located nearest to a branching point were therefore excluded for further analysis. Surprisingly, on average, no tapering was observed in the segments obtained from the right coronary artery that were studied. This observation of reverse tapering has recently been observed by others too.17
The generalized correction for tapering that was applied in this study merits careful consideration. The correction for tapering may be influenced by local3 as well as individual4 variation in remodeling response. This limitation may be overcome by studying de novo atherosclerotic remodeling locally using a reference site located near the cross-section under study.5,6,8
It is well accepted that arterial remodeling contributes to the development of luminal narrowing. Insight into the mechanism that is responsible for atherosclerotic compensatory enlargement and shrinkage may help to develop new therapeutic strategies to treat and prevent luminal narrowing. The finding of regional, artery-related differences in the atherosclerotic remodeling response may help to understand these underlying mechanisms.
Angiographic luminal narrowing does not necessarily reflect the atherosclerotic process.18,19 Due to compensatory enlargement of the artery, plaque formation may occur without subsequent changes of the lumen diameter on the angiogram. The results of the present study indicate that angiography may seriously underestimate the extent of the atherosclerotic process: in the renal artery: on average, only 2% of the actual plaque increase encroaches on the lumen reflected by angiographic narrowing of the lumen. Angiographic luminal narrowing may therefore be considered as a tip of the iceberg for the extent of the atherosclerotic process: in the renal artery. For instance, on average 98% of the plaque mass is hidden beneath the luminal surface.
The type of atherosclerotic de novo remodeling (enlargement versus shrinkage) does not influence the immediate outcome of balloon angioplasty.8,20 The dilation mechanism differs, however, for the different types of remodeled arteries.8,20 In the long term, restenosis may be more frequently observed for those arterial lesions that were previously shrunken.20 Thus, the type of de novo atherosclerotic remodeling may be related to the development of restenosis after balloon angioplasty. The finding of regional variability in the type of remodeling in response to plaque formation may explain differences in patency outcomes after balloon angioplasty of, for instance, the femoral artery.21–23
In conclusion, the type of remodeling (enlargement/shrinkage) influences the lumen in all artery types that are prone to develop atherosclerotic lesions. The impact of remodeling on percent luminal stenosis differs among artery types and may therefore be considered as a regionally determined process.
This study was supported by the Dutch Heart Foundation (Grant 94-115). Mark Post, MD, is gratefully acknowledged for his critical comments on the manuscript.
Armstrong ML, Heistad DD, Marcus ML, Megan MB, Piegors DJ. Structural and hemodynamic responses of peripheral arteries of macaque monkeys to atherogenic diet. Arteriosclerosis. 1985;5:336–346.
Pasterkamp G, Wensing PJW, Post MJ, Hillen B, Mali WPTM, Borst C. Paradoxical arterial wall shrinkage contributes to luminal narrowing of human atherosclerotic femoral arteries. Circulation. 1995;91:1444–1449.
Pasterkamp G, Borst C, Post MJ, Mali WPTM, Wensing PJW, Gussenhoven EJ, Hillen B. Atherosclerotic arterial remodeling in the superficial femoral artery: individual variation in local compensatory enlargement response. Circulation. 1996;93:1818–1825.
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.
Steinke W, Els T, Hennerici M. Compensatory carotid artery dilatation in early atherosclerosis. Circulation. 1994;89:2578–2581.
Haussman 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(suppl):175A. Abstract.
Wensing PJ, Scholten FG, Buijs PC, Hartkamp MJ, Mali WPTM, Hillen B. Arterial tortuosity in the femoropopliteal region during knee flexion: a magnetic resonance angiographic study. J Anat. 1995;187:133–139.
Yamamoto T, Ogasawara Y, Kimura A, Tanaka H, Hiramatsu O, Tsujioka K, Lever MJ, Parker KH, Jones CJ, Caro CG, Kaijya F. Blood velocity profiles in the human renal artery by Doppler ultrasound and their relationship to atherosclerosis. Arterioscler Thromb Vasc Biol. 1996;16:172–177.
Walpola PL. Gotlieb AI, Cybulsky MI, Langille L. Expression of ICAM-1 and VCAM-1 and monocyte adherence in arteries exposed to altered shear stress. Arterioscler Thromb Vasc Biol. 1995;15:2–10.
Kim MH, Ziada KM, Crowe TD, De Franco AC, Boparai N, Tuzcu EM, Nissen S. Intravascular ultrasound assessment of coronary artery “tapering.” Circulation. 1996;94(suppl):I-134. Abstract.
Ozaki Y, Violaris AG, de Feyter P, Roelandt J. Role of underlying vascular remodeling mode in the mechanism of acute luminal gain and late restenosis after balloon angioplasty (BA) and directional coronary atherectomy (DCA). Circulation. 1996;94(suppl):I-134. Abstract.