Prediction of Restenosis After Coronary Balloon Angioplasty

Patients

Patients were included in the study if they met the following criteria: (1) successful PTCA of a single coronary artery, defined as a decrease of ≥30% in angiographic diameter stenosis after PTCA, a residual diameter stenosis <50% (by visual assessment), and normal blood flow in the dilated artery (TIMI grade 3); and (2) reference diameter of the target vessel ≥2.5 mm.

Exclusion criteria included the following: (1) prior PTCA at the same site; (2) acute myocardial infarction <2 weeks before PTCA; (3) prominent coronary spasm; and (4) sharp bends and curves proximal to the segment of interest, precluding the passage of the ICUS catheter. Informed consent was obtained from all patients, and the protocol was approved by the local ethics committees of all participating centers. The centers and investigators participating in the study are listed in the “Appendix.”

Procedure

Standard PTCA was performed from the femoral artery approach. Balloon size was chosen with a nominal diameter to match the angiographic lumen diameter of an adjacent segment that appeared normal on the angiogram. The number of balloons and the number and pressure of the inflations were left to the discretion of the operator and were based on angiography only. After completion of the procedure, when the angiographic result met the inclusion criteria, the dilated artery was studied by ICUS (see below). If, in spite of a satisfactory initial result, the operator could not accept the angiographic appearance of the lesion after completion of the ICUS study (because of deterioration that had occurred during the time required for the ICUS study), by protocol a second ICUS study was performed after additional balloon inflations, and only this second study was used for analysis. After the ICUS study, QCA was performed. Follow-up coronary angiography was performed after 6 months. If symptoms recurred within 6 months, coronary angiography was performed earlier. If angiography was performed within 4 months but the index lesion was not redilated, the protocol required that angiography be repeated at 6 months.

QCA

The angiograms after the ICUS study and at 6-month follow-up were recorded after intracoronary injection of isosorbide dinitrate or nitroglycerin. At least two orthogonal projections were selected for analysis, and these projections were repeated at follow-up. To standardize the method of data acquisition and to ensure exact reproducibility of the angiographic studies, measurements were taken as described previously.^{17 18} For calibration purposes, the catheter tip was cut off for later measurement with a microcaliper. All angiograms were analyzed by a central core laboratory (Cardialysis, Rotterdam, Netherlands) using the computer-assisted CAAS-II technique, which has been described and validated previously.^{19 20} Restenosis, the primary end point of the study, was analyzed in two ways: as a binary definition of ≥50% diameter stenosis in the dilated lesion at follow-up and as a continuous variable, both as absolute MLD and as reduction in absolute luminal diameter in the dilated lesion at follow-up (“late loss”).

ICUS Imaging Protocol

All ICUS studies were performed with three mechanical, single-element, 30-MHz systems: (1) the Insight system (Cardiovascular Imaging Systems Inc) using 4.3F catheters, (2) the Intrasound system (Du-MED) using 4.1F catheters; and (3) the Boston Scientific system using 3.5F catheters with a Hewlett-Packard ICUS console.

After completion of the PTCA procedure, if the patient fulfilled the criteria mentioned above, intracoronary nitroglycerin or isosorbide dinitrate was administered and an ICUS catheter was introduced over the guidewire left in place after PTCA. Settings for gain and filters were adjusted to result in an optimal visual image. The entire procedure was stored on VHS or super VHS videotape, including annotations and audio recordings. The ICUS catheter was advanced to the dilated segment for close examination of vessel morphology and disruptions. At the narrowest site, 10 mL of saline and x-ray contrast flushes through the guiding catheter were performed to facilitate the recognition of disruptions and to confirm that the narrowest point on angiography was being examined. Additional flushing at other sites was left to the discretion of the operator. The timing of this flushing and of all other relevant events during the ICUS study was marked by voice recordings on the videotape. After the morphology was carefully explored, the transducer was advanced to a site distal to the lesion, and a slow, manual pullback maneuver was performed until the transducer reached the guiding catheter.

After the ICUS study, both the catheter and the guidewire were removed and a final angiogram (for quantitative analysis) was made.

Ultrasound Image Analysis

All images were analyzed by a committee of five observers (see “Appendix”) unaware of the results of angiography and follow-up examinations. Qualitative analysis of the images was performed independently by each member of the committee on a dedicated form, and all cases were reviewed in a panel discussion to reach consensus on all diagnostic parameters. The analysis consisted of two parts. First, the entire dilated segment was analyzed qualitatively (see below). Second, a single video frame in the cross section with the smallest lumen area (visually) was indicated as the narrowest site of the dilated lesion. This cross section was used for quantitative analysis (see below).

Qualitative Analysis

The following parameters (with definitions where appropriate) were assessed: image quality, classified as either acceptable or not acceptable; lesion morphology, classified as either eccentric or concentric (maximal plaque thickness/minimal plaque thickness ≥2 or <2, respectively) (this was performed subjectively because in the complex post-PTCA ultrasound images of dilated lesions, the accepted criteria for eccentricity cannot be applied); lesion type, classified as soft (less echogenic than adventitia), hard (as echogenic as adventitia), or calcific (at least one quadrant of calcium present)^{12 21} ; calcium (defined as dense reflections with acoustic shadowing^{9 11} ), with the number of deposits, their location relative to the narrowest point, their cumulative arc (in hours) in the dilated segment, and their depth (superficial, defined as no tissue between the deposit and the lumen; deep, all other depths; or both); rupture of the vessel wall, defined as a radial tear, ie, perpendicular to the vessel-wall layers; and dissection of the vessel wall, defined as a tear parallel to the vessel-wall layers. Regarding these disruptions, the following characteristics were noted: (1) location relative to the narrowest point; (2) the total circumferential arc in hours (with a minimum of 1); (3) maximal depth, classified as either partial (some plaque remaining intact between the rupture and the underlying media) or complete (extending through the plaque completely); and (4) the presence of calcium adjacent to the tear. Thrombus was not diagnosed by ICUS because the investigators did not feel confident making this diagnosis with existing criteria.^{22 23}

Quantitative Analysis

Measurements on the cross section designated as the narrowest point in the panel discussion were performed independently by two observers (W.E.M.K. and G.P.). After calibration by use of the markers in the ICUS image, the circumference of the lumen (including extensions created by ruptures or dissections) and of the media/adventitia border were traced manually by use of computer-assisted planimetry²⁴ (Fig 1⇓). If part of the media-adventitia interface was invisible owing to shadowing behind calcium deposits, the tracing was extrapolated from the adjacent sectors of the circumference. In cross sections with more than one quadrant of acoustic shadowing, the tracing was extrapolated from neighboring cross sections, as described by others.²⁵ All tracings were performed according to the consensus interpretation of the image. From this, the following parameters were derived: lumen area, vessel area (area encompassed by the media-adventitia border), plaque area (defined as Vessel Area−Lumen Area), and percent obstruction (defined as [Plaque Area/Vessel Area]×100). The mean value of the two independent measurements for each parameter was entered in the database.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Left, Original ultrasound cross section of a dilated lesion. The catheter is shown as a black circle in the center of the image, with an artifact from the guidewire between 12 and 1 o'clock. Right, Tracings have been added to the lumen and the adventitia-media border (Media B. area), demonstrating rupture (from 1 to 2 o'clock) and dissection (from 2 to 4 o'clock) of plaque and the method of obtaining quantitative parameters.

Interobserver variation of the measurements (199 baseline ICUS studies) was as follows: minimal lumen area, SD=0.66 mm², CV=12.6%; plaque area, SD=1.34 mm², CV=15.7%; vessel area, SD=1.43 mm², CV=10.4%; and percent obstruction, SD=5.22%, CV=8.6%. Consistent errors were 0.02 mm² for lumen area, −0.20 mm² for vessel area, −0.22 mm² for plaque area, and −0.89% for percent obstruction. Exclusion of lesions with more than one quadrant of calcium did not significantly alter the reproducibility.

Statistical Analysis

All variables were related to measures of restenosis in univariate analysis by use of the unpaired Student's t test, ANOVA, and χ² tests where appropriate. Variables significantly related to restenosis in univariate analysis were analyzed in multivariate analysis. For late loss, stepwise linear regression analysis was performed. For categorical restenosis, stepwise logistic regression analysis was performed.

Limitations

There are several limitations to this study. We included nontortuous coronary arteries, and a satisfactory angiographic result was required. These criteria may have resulted in the selection of favorable cases.

The late loss in MLD observed by QCA (0.17 mm) is low compared with other reports. With an average reference-segment diameter of 2.95 mm, a mean late loss of ≈0.26 mm is expected.⁴² Several factors may contribute to this discrepancy. First, although percent diameter stenosis was <50% visually in all lesions (inclusion criterion), there was in fact a >50% diameter stenosis on off-line QCA in 17 lesions (10%). In these cases, late loss is expected to be lower, and categorical 6-month restenosis (>50%) is actually pseudorestenosis. This may explain why a normal incidence (29.4%) of categorical restenosis was found in spite of the limited late loss. Second, it is possible that in the lesions with additional balloon dilations after the initial ICUS study, the early component of MLD loss was abolished, thereby lowering total 6-month loss.

We found a trend toward a greater percent area obstruction on ICUS in the restenosis group. With a greater number of observations, this relation would possibly reach statistical significance, consistent with the findings of others.^{28 29 30} It is doubtful, however, if the difference could reach clinical significance.

The use of an early generation of ICUS catheters may have influenced the results in two ways. First, image quality was not yet optimal. Second, the relatively large diameter of the catheters may have resulted in false-negative observations on vessel-wall dissections because the catheters may have “tacked” intimal flaps during the ICUS study. However, with a rupture incidence of 89% on ICUS, it is not likely that we have underestimated the degree of vessel-wall damage.Potentially, the patients who had additional dilations after the ICUS examination were a specific subgroup at high risk. By using the second, final ICUS examination for analysis, predictors present in the first ICUS examination may have remained undetected. However, it is unlikely that plaque composition is altered by additional dilations, and the incidence of disruptions remained high. In addition, we repeated the analysis of the relation of vessel-wall disruptions and restenosis after exclusion of this subgroup of patients, and this did not alter the findings.

We conclude that after angiographically and clinically successful balloon angioplasty, a larger lumen and vessel area and a smaller plaque area as determined by ICUS were associated with a larger angiographic MLD at follow-up. However, these relations were weak, and the study did not identify ICUS parameters predictive of restenosis after PTCA.

Participating Centers

Interuniversity Cardiology Institute of the Netherlands: Klaas Bom, PhD; Elma J. Gussenhoven, MD; Hans Rijsterborgh, PhD (†deceased).

Academic Hospital Free University, Amsterdam, Netherlands: Otto Kamp, MD (center coordinator); Jean G.F. Bronswaer, MD; Carel C. de Cock, MD; Cees A. Visser, MD.

Academic Hospital, University of Leiden, Netherlands: Lucas Savalle, MD (center coordinator); Bert Buys, MD (†deceased).

Academic Hospital, University of Maastricht, Netherlands: Frits W.H.M. Bär, MD (center coordinator); H. Quint.

Academic Hospital, University of Amsterdam, Netherlands: George K. David, MD; Karel T. Koch, MD; Wouter E.M. Kok, MD (center coordinator); Ron J.G. Peters, MD (principal investigator); Jan J. Piek, MD.

Academic Hospital, University of Utrecht, Netherlands: Piet W. Westerhof, MD; Cornelius Borst, MD; Gerard Pasterkamp, MD (center coordinator).

Sint Antonius Hospital, Nieuwegein, Netherlands: W. Jaarsma, MD (center coordinator); H.W. Thijs Plokker, MD; V. Radjan N. Panday, MD.

Thoraxcenter, Academic Hospital Dijkzigt, Erasmus University, Rotterdam, Netherlands: Pim J. de Feyter, MD; David T. Linker, MD; Carlo di Mario, MD (center coordinator); Patrick W. Serruys, MD.

Data Management

Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, Amsterdam, Netherlands: Martin H. Prins, MD.

Angiographic Core Laboratory

Cardialysis, Rotterdam, Netherlands: Eline Montauban van Swijndrecht, Edwin Mibbering, Patrick W. Serruys.

ICUS Analysis Committee

Otto Kamp, MD; Wouter E.M. Kok, MD; Carlo di Mario, MD; Gerard Pasterkamp, MD; Ron J.G. Peters, MD.