Preintervention Arterial Remodeling as an Independent Predictor of Target-Lesion Revascularization After Nonstent Coronary Intervention
An Analysis of 777 Lesions With Intravascular Ultrasound Imaging
Background—Pathological and intravascular ultrasound (IVUS) studies have documented arterial remodeling during atherogenesis. However, the impact of this remodeling process on the long-term outcome after percutaneous intervention is unknown.
Methods and Results—We used preintervention IVUS to define positive and negative/intermediate remodeling in a total of 777 lesions in 715 patients treated with nonstent techniques. Positive remodeling (lesion external elastic membrane area greater than average reference) was present in 313 lesions; intermediate/negative remodeling (lesion external elastic membrane area less than or equal to reference) was present in the other 464. Baseline clinical and angiographic characteristics were similar, except for a slightly higher percentage of insulin-dependent diabetic patients (10.2% versus 6.1%; P=0.054) in the negative/intermediate-remodeling group. Angiographic success and in-hospital and short-term complications were comparable in the 2 groups. There was no significant correlation between remodeling (as a continuous variable) and final lumen area (r=0.06) or final lesion plaque burden (r=0.17). At 18±13 months of clinical follow-up, both groups had similar rates of death and Q-wave myocardial infarction: 3.4% and 2.5% for the negative/intermediate-remodeling group versus 2.7% and 2.7% for the positive-remodeling group. However, the target-lesion revascularization (TLR) rate was 20.2% for the negative/intermediate-remodeling group versus 31.2% for the positive-remodeling group (P=0.007), and remodeling, as a continuous variable, was strongly correlated with probability of TLR (P=0.0001). By multivariable logistic regression analysis, diabetes (OR=2.3), left anterior descending artery location (OR=1.8), and remodeling (OR=5.9) were independent predictors of TLR.
Conclusions—Positive lesion-site remodeling is associated with a higher long-term TLR after a nonstent interventional procedure. Thus, long-term clinical outcome appears to be determined in part by preintervention lesion characteristics.
During de novo stenosis formation, positive remodeling of the diseased arterial wall occurs to compensate for the accumulation of atherosclerotic plaque. This compensatory mechanism has been noted in primates with diet-induced atherosclerosis and in human coronary arteries, where it has been seen pathologically and by use of high-frequency epicardial echocardiography and intravascular ultrasound (IVUS).1 2 3 4 5 6 7 In the majority of lesions, the development of angiographically detectable coronary artery disease depends on relative rates of plaque deposition and positive remodeling and/or the ultimate failure of the remodeling response. However, results from more recent studies, the majority of which used IVUS in vivo, have emphasized that in an important minority of lesions, inadequate or negative arterial remodeling contributes to lumen compromise.8 9 10 11 12 13 14 Suggested reasons for inadequate or negative arterial remodeling include the following: (1) positive remodeling may fail from the outset; (2) early positive remodeling may be followed by late negative remodeling (ie, arterial shrinkage); and (3) certain types of plaque elements (ie, calcium or dense fibrous tissue) may limit the extent of positive remodeling.
We hypothesized that lesion-specific differences in the direction and magnitude of the vascular remodeling process may influence the long-term arterial response to catheter-based coronary intervention.
Lesion and Patient Population
The Washington Hospital Center IVUS Database was probed to identify de novo native nonostial coronary lesions that were treated with nonstent interventions and had complete high-quality preintervention and postintervention IVUS imaging with limited amounts of lesion-site and reference-segment calcium so that baseline remodeling characteristics could be determined, with ≥12 months of available follow-up. A total of 777 lesions in 715 patients were identified. There were 531 males (74.3%) and 184 females (25.7%) with a mean patient age of 61±11 years. Target-lesion location was left main artery in 13 cases (1.7%), left anterior descending artery in 331 (42.6%), left circumflex artery in 100 (20.6%), and right coronary artery in 273 (35.1%).
Clinical Demographics and Follow-Up
The hospital charts of all patients were reviewed independently by a registered nurse to obtain clinical demographics and laboratory results. Risk factors for coronary artery disease that were tabulated included diabetes mellitus (only if treated medically), hypertension (only if treated medically), and hyperlipidemia (if treated medically or if serum cholesterol was ≥240 mg/dL). Non–Q-wave myocardial infarction (MI) was defined as a creatine kinase (CK)-MB enzyme elevation >5 times the upper limit of normal. Clinical follow-up was performed by telephone contact or office visit at 1, 3, 6, 12, 18, and 24 months after the procedure. The occurrence of major late clinical events was recorded, including death, Q-wave MI, and ischemia-driven target-lesion-site revascularization (TLR, whether percutaneous or surgical). These events were all source documented.
All cineangiograms were analyzed by use of a computer-assisted, automated edge-detection algorithm (ARTREK, Quantitative Cardiac Systems) by a core laboratory that was blinded to the IVUS and clinical findings. With the outer diameter of the contrast-filled catheter used for calibration, minimal lumen diameter (MLD) in diastole before intervention was measured from multiple projections, and results from the single “worst” view were recorded. Reference-segment diameter was averaged from user-defined 5-mm-long angiographically normal segments proximal and distal to the lesion but between any major side branches. Lesion length was measured as the distance (in millimeters) from the proximal shoulder to the distal shoulder in the projection with the least amount of foreshortening. A focal stenosis had a length ≤10 mm. For the purposes of the present study, ostial lesions were within 3 mm of the coronary ostia or <3 mm distal to a major proximal side branch. An eccentric target lesion appeared to have 3 times as much plaque on one side of the lesion as on the other. Calcification was identified as readily apparent radiopacities within the vascular wall at the site of the stenosis. An angulated segment contained a bend >45° within 5 mm of the lesion. Ectasia was a lumen >20% larger than the user-defined reference segments. Angiographic success was defined as a final diameter stenosis <50% in the absence of major complications. These represent standard qualitative and quantitative analyses and definitions; all have all been published previously.15
IVUS Imaging Protocol
All IVUS imaging studies were performed before any intervention and only after intracoronary administration of 200 μg of nitroglycerin. To perform the imaging run, the IVUS catheter was advanced ≥10 mm distal to the lesion, the video recorder turned on, the transducer pullback device activated, and the entire artery imaged to the aorto-ostial junction. The IVUS studies were performed with 1 of 2 commercially available systems. The first system (InterTherapy/Cardiovascular Imaging Systems Inc) incorporated a single-element, 25-MHz transducer and an angled mirror mounted on the tip of a flexible shaft that was rotated at 1800 rpm within a 3.9F short monorail polyethylene imaging sheath to form planar cross-sectional images in real time. The second system (Boston Scientific Corporation/Cardiovascular Imaging Systems Inc) incorporated a single-element, 30-MHz beveled transducer within either a 3.2F short monorail or 2.9F long monorail common distal lumen–imaging catheter. With both systems, the transducer was withdrawn automatically at 0.5 mm/s to perform the imaging sequence.16 The use of a motorized transducer pullback device and sheath-based imaging catheters permitted the transducer to move at the same speed as the proximal end of the imaging core. IVUS studies were recorded on 0.5-in high-resolution Super VHS tape for offline analysis.
Validation of plaque composition and measurements of external elastic membrane (EEM) cross-sectional area (CSA), lumen CSA, and plaque and media (P&M) CSA by IVUS have been reported previously.17 18 19 Computer planimetry (TapeMeasure, Indec Systems) was used to measure lesion-site and reference-segment EEM CSA and lumen CSA. We measured EEM CSA, the area encompassed by the ultrasonic media-adventitia border, by tracing the leading edge of the adventitia; this has been shown to be a reproducible measure of total arterial CSA. Because media thickness cannot be measured accurately, P&M CSA (EEM CSA minus lumen CSA) was used as a measure of atherosclerotic plaque.20 Cross-sectional narrowing (CSN; P&M CSA divided by EEM CSA) has also been called the plaque burden or the percent plaque area. The lesion site selected for analysis was the image slice with the smallest lumen CSA; if there were several image slices with an equally small lumen, the image slice with the largest EEM CSA and P&M CSA was analyzed. The proximal and distal reference segments were the most-normal-looking cross sections within 10 mm proximal or distal to the lesion but before any side branch. This method of selecting the lesion site and reference segments to study remodeling has been published previously.12 13
Definitions of Remodeling
For the purposes of the present analysis, we determined remodeling by dividing the target-lesion EEM CSA by the average of the proximal and distal reference-segment EEM CSA to create a remodeling index. Positive remodeling was defined as a remodeling index >1 and intermediate/negative remodeling as a remodeling index ≤1.0.
Statistical analysis was performed with StatView 4.5 or SAS (both SAS Institute Inc). Data are presented as mean±1 SD. Categorical data were compared with Fisher’s exact test. Continuous variables were compared with unpaired t tests or regression analysis. The primary end point was the association of preintervention remodeling with TLR; remodeling was entered into the analysis both as a categorical variable (positive versus intermediate/negative) and as a continuous variable. Univariate variables with a P value <0.2 were entered into the multivariate models. Forward stepping was used to determine the independent predictors of TLR. In addition, the 1-year TLR-free survival rate in the group with positive versus intermediate/negative remodeling was plotted; statistical significance was tested with the Mantel-Haenszel χ2 test. A P value ≤0.05 was considered significant.
Preintervention intermediate/negative remodeling was found in 464 lesions in 420 patients and positive remodeling in 313 lesions in 295 patients. Figure 1⇓ shows 2 lesions, 1 with intermediate/negative and 1 with positive remodeling.
Clinical, Angiographic, and IVUS Findings
Clinical characteristics and lesion location are shown in Table 1⇓. Insulin-treated diabetics and patients with an elevated cholesterol level were more common among those with intermediate/negative- versus positive-remodeling lesions.
There were only marginal differences in angiographic morphology between intermediate/negative- and positive-remodeling lesions: 59.7% versus 61.1% eccentric lesions (P=NS); 7.9% versus 8.9% type C lesions (P=NS); and 23.2% versus 31.2% type B2 lesions (P=0.06).
Quantitative angiographic data are presented in Table 2⇓. There was a trend for intermediate/negative-remodeling lesions to be less severe (smaller preintervention MLD, larger preintervention diameter stenosis); after intervention, there was a trend towards greater residual stenosis with intermediate/negative remodeling.
Quantitative IVUS findings are also presented in Table 2⇑. Preintervention, lesions with intermediate/negative remodeling had larger reference EEM and P&M CSA; conversely, these lesions had smaller lesion EEM and P&M CSA and CSN. After intervention, lesions with intermediate/negative remodeling had a smaller lesion lumen CSA and CSN. There was no significant correlation between IVUS reference lumen CSA (r=0.06), IVUS final lesion lumen CSA (r=0.06), or IVUS final lesion CSN (r=0.17) and the remodeling index.
Device Use, Procedural Success Rates, and In-Hospital Complications
Lesions with positive remodeling were treated with directional atherectomy more often than lesions with intermediate/negative remodeling: 174 (55.6%) versus 184 (40.0%) (P=0.001). The reverse was observed with respect to balloon angioplasty alone: 75 (24%) versus 184 (40%) (P=0.001). Both intermediate/negative- and positive-remodeling lesions had the same rate of treatment with rotational atherectomy (23.1% versus 21.7%) and excimer laser angioplasty (4.1% versus 3.8%).
Angiographic success was comparable in the 2 groups: 98.5% versus 97.9% for intermediate/negative versus positive remodeling (P=1.0). Major in-hospital complications (death/Q-wave MI/urgent revascularization) occurred in 8 patients in each group (1.9% versus 2.7%, respectively; P=0.46). The presence of any CK-MB enzyme elevation after intervention was found in 111 (26.4%) versus 70 (23.7%) patients, respectively (P=0.56), and non–Q-wave MI occurred in 44 (10.4%) versus 33 (11.1%) (P=0.66).
At follow-up, both groups had similar rates of death and Q-wave MI (Table 3⇓). However, the TLR rate was significantly lower for the intermediate/negative-remodeling group: 76 (20.2%) versus 84 (31.2%) patients (P=0.001). TLR-free 1-year survival rate was significantly higher in the intermediate/negative-remodeling group (P=0.0017; Figure 2⇓). Additionally, logistic regression analysis showed that the probability of TLR correlated strongly with the remodeling index as a continuous variable (P=0.0001).
Multivariable logistic regression analysis was used to determine the independent predictors of TLR. The following categories of variables were entered into the multivariate model: clinical (diabetes), lesion (left anterior descending coronary artery location), procedural (directional atherectomy and balloon angioplasty use), and IVUS (reference lumen CSA, final lumen CSA, final lesion CSA, and remodeling index). History of diabetes (OR, 2.3; 95% CI, 1.5 to 3.7; P=0.0004), left anterior descending coronary artery location (OR, 1.8; 95% CI, 1.1 to 2.9; P=0.0093), and an increasing remodeling index (OR, 5.9; 95% CI, 2.0 to 17.2; P=0.0011) were independent predictors of TLR.
Previous IVUS studies have detected a spectrum of baseline lesion-remodeling characteristics; however, these studies did not attempt to link preintervention remodeling to the long-term response to catheter-based intervention. The present study shows that the pattern of arterial remodeling of an atherosclerotic lesion may be an important predictor of long-term clinical outcome after nonstent coronary interventions.
IVUS Assessment of Remodeling
De novo stenoses have been classified into positive-, intermediate-, or negative-remodeling categories by comparison of the lesion site with the reference segment.13 In that study, positive remodeling was defined as a preintervention lesion site arterial area larger than the proximal reference, intermediate remodeling as a preintervention lesion site arterial area smaller than the proximal reference but larger than the distal reference, and negative remodeling as a preintervention lesion site arterial area smaller than the distal reference.13 In another study,12 the magnitude of arterial remodeling was determined by comparison of the target-lesion EEM CSA with a proximal reference, allowing for the maximum amount of arterial tapering. In various studies,12 13 21 native coronary artery lesions have been shown to exhibit positive remodeling in 35% to 54% and inadequate or negative arterial remodeling in 15% to 34%.
In the present study, lesions were classified as having characteristics of intermediate/negative remodeling (60% of the total cohort) versus positive remodeling (40%). There were more insulin-treated diabetic patients in the intermediate/negative-remodeling group, which confirms a previous report from our laboratory.22 Others have reported more smoking among patients with lesions with intermediate or negative remodeling,21 more hypercholesterolemia among patients with lesions with positive remodeling,21 more frequent unstable clinical presentation among patients with lesions with positive remodeling,23 24 and more fibrocalcific plaque elements in lesions with intermediate/negative remodeling.12 21 25 The latter 2 findings suggest that positive-remodeling lesions are “younger” and less stable, whereas intermediate/negative-remodeling lesions are “older” and more mature.
Remodeling and TLR
The present study suggests that the pattern of remodeling is not just a pathological curiosity. Lesions with positive remodeling had more revascularization events despite a larger final IVUS lumen CSA after the interventional procedure.
Recent studies23 24 26 27 have implicated positive remodeling in the pathogenesis of unstable coronary syndromes. The present study suggests that lesions with positive remodeling may also have a less-favorable long-term outcome after nonstent, catheter-based interventions. It has long been recognized that unstable angina and post-MI lesions have a higher rate of clinical and angiographic restenosis.28 Lesions with positive remodeling may be more biologically active. This increased biological activity may cause both an unstable clinical presentation and more late revascularization events. It has recently been shown29 that the early proliferative response to nonstent, catheter-based interventions is quantitatively related to late negative remodeling, the main mechanism of restenosis in these lesions.30 31
Conversely, lesions with intermediate/negative remodeling may not have the capacity to actively constrict (“negatively remodel”) late after intervention. This may also be due to the inherent characteristics of such lesions. They were unable to adapt to the initial growth of the atherosclerotic plaque with a positive-remodeling response during atherogenesis. Similarly, they may be less “proliferative” or “reactive” after injury associated with catheter-based intervention.32 In addition, a recent study33 suggested that TLR as a measure of clinical restenosis is less common in smokers, and 2 previous reports21 34 indicated that intermediate/negative-remodeling lesions are more common in smokers.
There may also be a mechanical explanation. Arterial expansion is an important mechanism of lumen enlargement after all catheter-based interventions. Assuming that restenosis is, in part, a proportionate response to vascular and perivascular trauma, lesions with positive remodeling (which already have increased their arterial dimensions significantly) may have more trauma associated with the procedure-induced arterial expansion. This is supported by the finding in the present study that the probability of TLR increases with increasing values of the remodeling index. Furthermore, baseline positive remodeling may be associated with a larger postintervention plaque burden, which might amplify the impact on lumen dimensions of late negative remodeling.35
This was a retrospective analysis. Clinical outcome (ie, TLR) may have potential bias due to patient selection at follow-up.33 The definition of lesion-site remodeling in the present study could have been influenced by the degree of reference-segment remodeling. In the overall cohort, greater plaque accumulation and positive remodeling at the reference segment could have artificially produced the appearance of intermediate/negative remodeling at the lesion site. Introduction of the remodeling index as a continuous variable both in the multivariate analysis and in the assessment of TLR-free survival helped to reduce the bias introduced by categorical definitions. Previous investigators have used other definitions of adequate, adaptive, or positive remodeling; some of these definitions differ from that used in the present study, and they differ among each other as well. Introduction of the remodeling index as a continuous variable may have helped to alleviate the discrepancies introduced by different arbitrary categorical definitions. Finally, the present study addresses the impact of preintervention lesion remodeling on the outcome of nonstent interventions. Stenting is now the predominant type of coronary intervention.
The baseline (preintervention) pattern of lesion-site remodeling appears to have important prognostic implications on long-term clinical outcome after catheter-based, nonstent coronary interventions. Positive remodeling appears to have a worse prognosis (more frequent need for subsequent revascularization) than intermediate/negative remodeling.
These findings suggest that positive-remodeling lesions should be targeted for treatment with interventions that have been shown to reduce restenosis, ie, stents. Conversely, negative-remodeling lesions may be suitable for a provisional stent-implantation strategy.
This study was supported in part by the Cardiology Research Foundation, Washington, DC.
- Received December 31, 1998.
- Revision received March 15, 1999.
- Accepted April 9, 1999.
- Copyright © 1999 by American Heart Association
Armstrong ML, Heistad DD, Marcus ML, Megan MB, Piegors DJ. Structural and hemodynamic responses of peripheral arteries of macaque monkeys to atherogenic diet. Atherosclerosis. 1985;5:336–346.
Stiel GM, Stiel LS, 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.
McPherson DD, Sirna SJ, Hiratzka LF, Thorpe L, Armstrong ML, Marcus ML, Kerber RE. Coronary arterial remodeling studied by high-frequency epicardial echocardiography: an early compensatory mechanism in patients with obstructive coronary atherosclerosis. J Am Coll Cardiol. 1991;17:79–86.
Losordo DW, Rosenfield K, Kaufman J, Pieczek A, Isner JM. Focal compensatory enlargement of human arteries in response to progressive atherosclerosis: in vivo documentation using intravascular ultrasound. Circulation. 1994;89:2570–2577.
Pasterkamp G, Borst C, Post MJ, Mali WP, Wensing PJ, Gussenhoven EJ, Hillen B. Atherosclerotic arterial remodeling in the superficial femoral artery: individual variation in local compensatory enlargement response. Circulation. 1996;93:1818–1825.
Berglund H, Luo H, Nishioka T, Fishbein MC, Eigler NL, Tabak SW, Siegel RJ. Highly localized arterial remodeling in patients with coronary atherosclerosis: an intravascular ultrasound study. Circulation. 1997;96:1470–1476.
von Birgelen C, Mintz GS, de Vrey EA, Kimura T, Popma JJ, Airiian SG, Leon MB, Nobuyoshi M, Serruys PW, de Feyter PJ. Atherosclerotic coronary lesions with inadequate compensatory enlargement have smaller plaque and vessel volumes: observations with three dimensional intravascular ultrasound in vivo. Heart. 1998;79:137–142.
Mintz GS, Kent KM, Pichard AD, Satler LF, Popma JJ, Leon MB. Contribution of inadequate arterial remodeling to the development of focal coronary artery stenoses: an intravascular ultrasound study. Circulation. 1997;95:1791–1798.
Lim TT, Liang DH, Botas J, Schroeder JS, Oesterle SN, Yeung AC. Role of compensatory enlargement and shrinkage in transplant coronary artery disease: serial intravascular ultrasound study. Circulation. 1997;95:855–859.
Popma JJ, Bashore TD. Qualitative and quantitative angiography. In: Topol E, ed. Textbook of Interventional Cardiology. Philadelphia, Pa: WB Saunders; 1994:1052–1068.
Tobis JM, Mallery J, Mahon D, Lehmann K, Zalesky P, Griffith J, Gessert J, Moriuchi M, McRae M, Dwyer ML, Greep N, Henry WL. Intravascular ultrasound imaging of human coronary arteries in vivo: analysis of tissue characteristics with comparison to in vivo histologic specimens. Circulation. 1991;83:913–926.
Potkin BN, Bartorelli AL, Gessert JM, Neville R, Almagor Y, Roberts WC, Leon MB. Coronary artery imaging with intravascular high-frequency ultrasound. Circulation. 1990;81:1575–1585.
Schwarzacher S, Juegel H, Filardo SD, Lee DP, Yeung AC. Ruptured plaques have greater degree of remodeling than stable lesions: an IVUS study. Circulation. 1998;98(suppl I):I-145. Abstract.
Schoenhagen P, Ziada KM, Nissen SE, Tuzcu EM. Arterial remodeling in stable versus unstable coronary syndromes: an intravascular ultrasound study. Circulation. 1998;98(suppl I):I-368. Abstract.
Sabate M, Desphande NV, Ligthart JMR, van Domburg RT, de Feyter PJ. Remodeling of atherosclerotic coronary arteries in relation to the composition of the plaque. Circulation. 1998;98(suppl I):I-369. Abstract.
Von Birgelen C, Ge J, Klinkhart W, Welge D, Papatheodorou A, Liu F, Baumgart D, George G, Haude M, Erber R. Intravascular ultrasound assessment of coronary lesions following spontaneous plaque rupture: a prospective study in patients with prior unstable clinical episodes and typical angiographic signs. Circulation. 1998;98(suppl I):I-367. Abstract.
Kawagoe T, Sato H, Ishihara M, Noma K, Nishioka K. Significance of vascular remodeling as the culprit lesion of acute myocardial infarction: an intravascular ultrasound study. Circulation. 1998;98(suppl I):I-368. Abstract.
Holmes DR, Vlietstra RE, Smith H, Vetrovec G, Kent KM, Cowley MJ, Faxon DP, Gruentzig AR, Kelsey SF, Detre KM, Van Raden MJ, Mock MB. Restenosis after percutaneous transluminal coronary angioplasty (PTCA): a report from the PTCA Registry of the NHLBI. Am J Cardiol. 1984;53:77C–81C.
Mintz GS, Kimura T. Relationship between early positive remodeling, late negative remodeling, and intimal hyperplasia after coronary angioplasty or atherectomy: an intravascular ultrasound study. Circulation. 1998;98(suppl I):I-512. Abstract.
Mintz GS, Popma JJ, Pichard AD, Kent KM, Satler LF, Wong SC, Hong MK, Kovach JA, Leon MB. Arterial remodeling after coronary angioplasty: a serial intravascular ultrasound study. Circulation. 1996;94:35–43.
Kimura T, Kaburagi S, Tamura T, Yokoi H, Nakagawa Y, Yokoi H, Hamasaki N, Nosaka H, Nobuyoshi M, Mintz GS, Popma JJ, Leon MB. Remodeling of human coronary arteries undergoing coronary angioplasty or atherectomy. Circulation. 1997;96:475–483.
Cohen DJ, Rodriguez O, Chauhan MS, Kuntz RE. Does smoking reduce restenosis? The smoker’s paradox revisited. Circulation. 1998;98(suppl I):I-511. Abstract.
Weissman NJ, Chari R, Mendelsohn FO, Anderson WD, Breall JA, Tanguay J-F, Diver DJ, Gersh BJ. Patient and plaque characteristics associated with coronary artery remodeling: an intravascular ultrasound analysis. J Am Coll Cardiol. 1998;29(suppl A):124A. Abstract.