The Korean TRI working group (TRI-WG) registry was performed at nine university hospitals. These hospitals have high-volume patients and facilities for primary PCI and on-site cardiac surgery. Data were collected at each site by trained study coordinators with a standardized case report form. The study protocol was approved by the Institutional Review Board from each site, all patients provided written informed consent. Standardized definitions were used for all patient-related variables and clinical diagnoses. The study protocol was approved by the Ethics Committee at each participating institution.

A total of 4,785 patients who underwent PCI from January to December 2009 were retrospectively enrolled in the TRI-WG registry. A total of 936 eligible STEMI patients who underwent primary PCI with DESs were enrolled, patients were excluded if they had any of the following conditions: cardiogenic shock on admission, Killips grade IV cardiac function, hemodynamic supporting devices (intra-aortic balloon counterpulsation or extracorporeal membrane oxygenation), received bare metal stents, or received balloon angioplasty. A total of 689 patients were finally enrolled for this study and divided into two groups according to vascular access (transradial approach group, n = 220; transfemoral approach group, n = 469). To adjust baseline potential confounders, propensity score matched (PSM) analysis was performed using logistic regression model. After PSM, two well-balanced and well-matched groups (171 pairs, n = 342, C-statistic = 0.792) were generated and analyzed.

Diagnostic angiography and PCI were performed through either the femoral or radial artery after administration of unfractionated heparin (70 to 100 U/kg) to maintain activated clotting time of > 250 seconds in patients during the procedure. Stents were deployed after balloon angioplasty. Cilostazol or platelet glycoprotein (GP) IIb/IIIa receptor blocker was used at the discretion of individual operator. A successful PCI was defined as the achievement of an angiographic residual stenosis of less than 30% in the presence of thrombolysis in myocardial infarction (TIMI) blood flow grade 3. During in-hospital period, patients received cardiovascular beneficial medications including β-blockers, renin angiotensin system inhibitors, calcium channel blockers, and statins. After discharge, patients were encouraged to stay on the same medications as they received in the hospital.

Cardiovascular risk factors and past history (age, sex, hypertension, dyslipidemia, smoking, diabetes mellitus, prior myocardial infarction [MI], chronic heart failure [HF], prior cerebrovascular disease, and peripheral arterial disease) were based on self-reports of patients. Assess site complications were defined as vascular complications that occurred in the access vessel such as hematomas (minor hematoma was defined as < 4 cm of hematoma induration and major hematoma ≥ 4 cm), dissections, perforations, AVFs, and PSAs. Bleeding complications were defined as those previously described by Mehran et al. [8]. Procedure times were measured from the start of the puncture to the removal of the guiding catheters. Fluoroscopy times were measured automatically using fluoroscopes. Clinical endpoints were defined as those previously described by Cutlip et al. [9]. All deaths were considered as cardiac ones unless a non-cardiac cause was documented. Recurrent MI (Re-MI) was defined as recurrent symptoms of new ST-segment elevation or re-elevation of cardiac markers to a level of at least twice of the normal upper limit. Target lesion revascularization (TLR) was defined as an ischemia-induced PCI of the target lesion due to restenosis or re-occlusion within the stent or 5 mm in and adjacent of the distal or proximal segment. Target vessel revascularization (TVR) was defined as a clinically driven PCI of the target lesion or any segment of the coronary artery containing the target lesion. Total major adverse cardiac events (MACEs) was defined as the composite of total death, non-fatal Re-MI, and TVR. All participants were required to visit the outpatient Department of Cardiology at the end of the first month, every 3 to 6 months after the PCI procedure, and whenever angina-like symptoms occurred. Cumulative incidence of individual and composite clinical outcomes at 1-year of follow-up was compared between the two groups.

The primary study end-points were bleeding and vascular complications, procedural complications, and 30-day clinical outcomes after procedure. The secondary study end-points were cumulative incidence of individual and composite clinical outcomes during 1-year of clinical follow-up.

For continuous variables, differences between the two groups were evaluated by unpaired t test or Mann-Whitney rank test. Data were expressed as mean ± standard deviation of means. For discrete variables, differences were expressed as counts and percentages. They were analyzed with chi-square or Fisher exact test between the two groups when appropriate. To adjust for potential confounders, PSM analysis was performed using logistic regression model. We tested all available variables that could be of potential relevance, including age, male, cardiovascular risk factors (hypertension, diabetes, dyslipidemia, HF, chronic kidney disease, coronary artery disease, stroke, peripheral artery disease, history of MI or PCI, and current smoker), angiographic factors (target vessels [left main, left anterior descending, left circumflex, right coronary artery], lesion type [B2, C], chronic total occlusion [CTO], bifurcation, diffuse long lesion, and calcification), and medications (cilostazol, angiotensin II receptor blockers, angiotensin-converting enzyme inhibitors, calcium channel blockers, β-blockers, diuretics, and statin). Patients in the TRI group were then 1-to-1 matched to those in the TFI group according to propensity scores with the nearest available pair matching method. Subjects were matched with a caliper width of 0.05. The procedure yielded 171 well-matched pairs. Cox-proportional hazard models were used to assess the hazard ratios in the two groups, the entire population, and the PSM population. The outcome of both groups were censored at a fixed point of 1 year (365 days) to avoid any bias caused by different follow-up duration. For all analyses, a two-sided p value of less than 0.05 was considered as statistically significant. All statistical analyses were performed using SPSS version 20.0 (IBM Co., Armonk, NY, USA).

A total of 4,785 patients were enrolled in the Korean TRIWG registry, including 936 patients who were diagnosed with STEMI. Among the 936 patients, 247 patient did not fulfill the inclusion criteria. Finally, 220 patient underwent PCI via TRI group while 469 patients underwent PCI via TFI group (Fig. 1). The baseline demographics, clinical characteristics, and labora tory findings of the two groups are summarized in Table 1. There were no significant differences in age, left ventricular ejection fraction, previous MI, or previous PCI. However, the prevalence rates of hypertension, diabetes, previous stroke, and current smoking were significantly higher than those in the TRI group. PSM yielded 171 pairs with well-balanced baseline characteristics. There was no significant differences in baseline characteristics between the TRI and TFI groups (Table 1).

At baseline angiography, the TFI group had higher incidence rates of multivessel disease, left circumflex lesion, B2/C lesions, and CTO lesion than the TRI group. The TFI group used bigger sheath than the TRI group. In addition, more patients in the TRI group used closure devices than those in the TRI group. Image and physiologic study guided intervention and adjuvant balloon inflation were more frequent in the TRI group compared to those in the TFI group. Baseline angiographic and procedural characteristics after PSM were similar to each other between the two groups except that the incidence rates of smaller sheath use, image and physiology guided intervention, and adjuvant ballooning were higher in the TRI group. Contrast amount was higher in the TRI group than that in the TFI group. Procedure time and fluoroscopic time were also longer in the TRI group (Table 2).

Major in-hospital and access site complications are listed in Table 3. The incidence rates of procedure-related acute and in-hospital complications in the two groups were similar to each other. The incidence of vascular access site hematoma in the TFI group was significantly higher than that in the TRI group. Although the incidence of major hematoma was similar to each other between the two groups, the incidence of minor hematomas was higher in the TFI group than that in the TRI group. No major life-threatening bleeding complications were observed in either group. Furthermore, the incidence of anemia requiring transfusion was similar to each other between the two groups. After PSM, the incidence of minor hematoma remained more frequent in the TFI group. However, all other procedural and in-hospital complications were similar to each other between the two groups (Table 3).

In the entire patient population before PSM, the incidence rates of recurrent MI (particularly non-STEMI [NSTEMI]) and repeat revascularization including TLR and TVR were higher in the TFI group than those in the TRI group during the 12-month of follow-up. After PSM, the incident rates of major clinical outcomes including mortality, recurrent MI, TLR, and stent thrombosis rates during the 12-month of follow-up showed no disparity between the two groups. However, the incidence of TVR and MACE were higher in the TFI group (Table 4).

Results of multivariate logistic regression analysis of individual and composite major clinical outcomes during the 12-month of follow-up are shown in Table 5. The incident rates of TVR and MACE were higher in the TFI group after adjustment. In PSM patients, TRI strategy and adjuvant ballooning were independent predictors for lower incidence of repeat revascularization during the 12-month of follow-up (Table 5).