Twenty four patients (20 males, 83%) who were diagnosed with STEMI and underwent primary percutaneous coronary intervention (PCI) were enrolled from Aug. 2007 to Mar. 2008 in the Cardiovascular Hospital of Yonsei University Medical Center. The study protocols were approved by the institutional ethics committee of Yonsei University College of Medicine, and written consent was obtained from all patients before the procedure. Emergency coronary angiography and primary PCI were performed by the femoral approach according to standard techniques in each of the patients. Using a guiding catheter for magnification and calibration, quantitative coronary angiography was performed immediately after the procedure. OCT was performed in infarct-related arteries treated with DES at the second stage within 72 hours of the primary PCI. STEMI was defined as continuous chest pain lasting > 30 min, with arrival at our hospital within 12 hours from the onset of the symptom, with ST-segment elevation ≥ 0.1 mV in two or more contiguous leads on 12-lead electrocardiogram and elevation of serum cardiac enzymes such as troponin and/or creatine kinase-MB (CK-MB). The luminal diameter proximal to the culprit lesion was required to be between 2.5 mm and 4.0 mm due to the maximal diameter of the currently available occlusion balloon. Exclusion criteria were poor left ventricular function (< 30% of ejection fraction), hemodynamic instability, significant left main coronary artery disease, renal insufficiency with baseline serum creatinine > 1.5 mg/dL, a single remaining vessel, highly tortuous vessels, and ostial or very proximal lesions (< 15 mm from ostium).

All patients received an intravenous bolus injection of 5,000 U of heparin. In primary PCI in STEMI, 200 mg of aspirin was given before the procedure and 600 mg of clopidogrel was given during the surgery. After the PCI, both drugs were continuously administered for at least 12 months. The use of glycoprotein IIb/IIIa inhibitors was left to the discretion of the operator.

An intravascular ultrasound (IVUS) was performed during primary PCI if needed. However, it was done concurrently with OCT within 72 hours and the data was analyzed.

IVUS assessment was performed using a commercially available IVUS system (Boston Scientific/Scimed, Natick, MA, USA) with automatic motorized 0.5 mm/s pullback. All IVUS procedures were performed after intracoronary administration of nitroglycerin 200 µg. Studies were recorded on a compact disk for offline analysis using Index (Echoplaque 2; INDEC System, Inc., Mountain View, CA, USA). Proximal reference cross-sectional area (CSA, mm²) and distal reference CSA (mm²) were measured at the most visually normal area (largest lumen with the least plaque burden) within 10 mm of the proximal or distal ends of the lesion. Malapposition was defined as the separation of at least 1 stent strut from the intimal surface of the arterial wall that was not overlapping a side branch and had evidence of blood flow (speckling) behind the strut.9

The OCT system used in this study consisted of a computer, a monitor display, an interface unit (Model M2 Cardiology Imaging System, LightLab Imaging, Inc., Westford, MA, USA), and a 0.014-inch wire-type imaging catheter (Image-Wire, LightLab Imaging, Inc., Westford, MA, USA). A motorized pull-back system at 1.0 mm/s was used, and OCT images were acquired at 15 frames/s. The axial resolution capacity of this system was 15 µm. All patients were given 5,000 IU of heparin intravenously before their respective procedures. A 6Fr or 7Fr guiding catheter was introduced into the coronary artery by a femoral approach, and nitroglycerin (200 µg) was administered through the guiding catheter. To remove the blood from the field of view and allow for clear images, an occlusion balloon catheter (Helios, Avantec Vascular Corp., Sunnyvale, CA, USA) and flushing with warm lac was used in an over-the-wire type method with the flush lumen and the lumen for crossing the imaging catheter. There were two metal markers at the tip and at the center of the occlusion balloon. The occlusion catheter was positioned proximal to the stent and the image catheter was positioned distal to the stent. During image acquisition, the occlusion balloon was inflated at 0.4 to 0.6 atm and Ringer lactate was infused at 0.5 to 1.0 mL/s. The image wire was pulled from distal to proximal, and continuous images were stored digitally for subsequent analysis.

Cross-sectional OCT images were analyzed at 1-mm intervals (every 15 frames). Stent struts appeared as highly reflective surfaces that cast shadows on the vessel wall behind. The distances between the endo-luminal surface of the strut reflection and the vessel wall were measured by prolonging and joining the contours of the wall on either side of the strut shadow with a measurement line as perpendicular as possible to the strut and vessel wall (Fig. 1). When the strut was not fully attached to the vessel wall by visual estimation, the position of the stent strut to the vessel wall was measured by magnifying the individual stent strut to maximize accuracy. Malapposed stent struts were defined as struts with detachment from the vessel wall ≥ 160 µm for sirolimus-eluting stents (SES; Cypher select, Cordis, J&J, East Bridgewater, NJ, USA), ≥ 130 µm for paclitaxeleluting stents (PES; Taxus Liberte, Boston Scientific, Natick, MA, USA), and ≥ 110 µm for zotaroliums-eluting stents (ZES; Endeavor, Medtronic, Minneapolis, MN, USA).10 Stent overlapping segments and bifurcation lesions with major side branches were excluded from this analysis. Inter-observer and intra-observer variabilities in measured distances were assessed by the evaluation of 20 random cross-sectional images by two independent readers and by the same reader at two separate time points, respectively. The variations between measurements were calculated using the linear mixed model (one-way mixed and two-way mixed models).

Results are expressed as a mean ± SD or number (%). Comparisons of categorical variables were made using the Chi-square test and the student's t-test while one way ANOVA were used to compare continuous variables. The intra-observer and inter-observer agreements (reproducibility) were assessed by determining the mean ± SD of differences between observations and between observers, respectively. If the distributions were skewed, a non-parametric test was used. Because the observation of struts in the same stent is not independent of each other and the hierarchical nature of the data, mixed effect multiple logistic regression with malapposition as the outcome variable was applied to address random and fixed effects at the strut, lesion and patient levels. At the lesion level, stent type (SES versus non-SES), stent diameter, stent length, reference vessel diameter, minimal luminal diameter and maximum inflation pressure were considered. At the patient level, age, sex and diabetes mellitus were considered.11 All analyses were performed using the Statistical Analysis System software (SAS 9.1.3., SAS Institute, Cary, NC, USA; and R 2.8.1., R Foundation, Vienna, Austria) and p < 0.05 was considered statistically significant.

Baseline characteristics of patients are presented in Table 1.

The patients' mean age was 59 ± 10 years and they were predominantly male. Seventy percent of patients (16, 66%) had multi-vessel disease, with culprit lesions predominately in the left descending artery (14, 58%). SES was implanted in 7 patients (29%), PES in 7 patients (29%), and ZES in 10 patients (42%). Angiographic results are shown in Table 2. A pre-stenting balloon was applied in all patients. High pressures (15.0 ± 2.0 atm) were applied in all implanted stents but it was not statistically significant (Table 3). The quantitative angiographic analysis in each group is shown in Table 3.

Post-intervention IVUS measurements are represented in Table 3. Among the groups, IVUS parameters were not significantly different except for minimal stent CSA. The initial rate of malapposition measured by IVUS was 21%, and had a higher trend toward to SES.

The total number of measured struts was 4951, including 1463 (30%) in SES, 1522 (31%) in PES, and 1966 (39%) in ZES. The average distances between the lumen and stent were different, and the SES was significantly longer than the other stents (188 ± 46 in SES, 144 ± 18 in PES, 127 ± 34 in ZES, p = 0.00) (Fig. 2). The numbers of malapposition struts were 28% in SES, but only 11% in PES and 10% in ZES; the malapposition rates of SES were significantly higher than those of the other two stents (Fig. 3). The rate of malapposition of 5 patients who had malapposition in IVUS showed 32% in strut level by OCT but, 12% in 19 patients who had not and this is statistically significant (32 ± 5% vs. 12 ± 6%, p = 0.00) (Table 4).

If we consider the definition of stent malapposition in OCT as having malapposed struts over 10% in all struts, the malapposition rate was 67% (16/24) or if over 5%, 83% (20/24) or if any malapposition, 100% (24/24). Regarding sensitivity of OCT as 100%, that of IVUS could be calculated as 33%.

The intra-observer correlation coefficient between variability of observer1 was 0.981 (95% CI: 0.975-0.985) using the one-way mixed model, where patient effects are random. The inter-observer correlation coefficient between the variability of two observers using the two-way mixed model, where patient effects are random and observers effects are fixed, were 0.982 (95% CI: 0.977-0.986) and 0.983 (95% CI: 0.978-0.987), respectively. Multivariate predictors of malapposed stent struts on mixed effect logistic regression analysis are shown in Table 5. The use of SES was an independent predictor of malapposed struts together with large reference vessel diameter, small stent size and high maximal pressure.