The CONSTANT registry was a prospective, open labeled, multi-center registry including single center randomized study. It included patients undergoing single lesion per single vessel percutaneous coronary intervention (PCI) with Resolute integrity DES. The NOBORI registry, on the other hand, is an open label, multi-center observational registry that also included patients undergoing PCI with Biolimus A9-eluting stent (Nobori DES) for all-comer patients per single coronary lesion. Patient's enrollment was all-comer diagnosed as stable angina and acute coronary syndrome who are indicated for coronary revascularization as targeted lesion which was significant (>70% by quantitative angiographic analysis) and a single lesion per single vessel, confirmed by stress test, imaging study or cardiac enzymes.

The primary endpoint was a composite of major adverse cardiovascular and cerebrovascular events (MACCEs), including cardiac death, MI, clinically driven target lesion revascularization and stroke. Deaths were classified as cardiac or non-cardiac, and death of any unidentified cause or in which a cardiac cause could not be excluded was classified as cardiac in this study. MI was classified as Q wave or non-Q wave, and was defined as a rise in creatine kinase enzyme concentrations above twice the normal upper limit. Re-interventions inside the implanted stent or within 5 mm proximal or distal to the stent were classified as target lesion revascularization (TLR). The definition of TLR had been registered according to the SIRIUS criteria. That is, TLR is defined as any “clinically-driven” repeat percutaneous intervention of the target lesion or bypass surgery of the target vessel. Clinically-driven revascularizations are those in which the patient has a positive functional study, ischemic electrocardiographic changes at rest in a distribution consistent with the target vessel, or ischemic symptoms, and an in-lesion diameter stenosis ≥50% by QCA. Revascularization of a target lesion with an in-lesion diameter stenosis ≥70% (by QCA) in the absence of the above-mentioned ischemic signs or symptoms was also considered clinically driven. In the absence of QCA data for relevant follow-up angiograms, clinical need for revascularization was adjudicated using the presence or absence of ischemic signs and symptoms. Repeated PCI to the same vessel with the exception of TLR was counted as non-target lesion (TL) target vessel revascularization (TVR). Target vessel failure was defined as all target vessel-related events, which included cardiac death, MI, thrombosis, and TVR. According to the Academic Research Consortium Classification, definite and probable stent thrombosis was considered stent thrombosis.11

Statistical analysis was performed using SPSS (version 18.0.0, SPSS Inc., Chicago, IL, USA) and R (version 2.8.0, R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are presented as mean ±standard deviation and were compared using a Student unpaired t-test. Categorical variables are presented as counts and percentages, and were compared using chi-square test. Control of confounders between both registries was undertaken by propensity score matching (PSM). The propensity score is the probability of treatment assignment conditional on observed baseline characteristics. The propensity score allows one to design and analyze an observational (nonrandomized) study, so that it mimics some of the particular characteristics of a randomized controlled trial.12

Propensity scores were estimated by fitting a logistic regression model using the following variables for Nobori and Resolute stents: age over 65 years old, gender, diabetes mellitus, prior history of MI, PCI, and coronary artery bypass graft (CABG), clinical presentation of acute coronary syndrome, and angiographic coronary artery disease findings. Propensity scores yielded a C statistic of 0.635, indicating a good ability to differentiate between two groups of patients. Nearest neighbor matching with a caliper of 0.001 was used. The Hansen and Bowers balance test p value was 1.000, indicating good covariate balance. Table 1 shows a list of variables used to construct the propensity score. If a subject of the NOBORI group could not be matched to any subject of the CONSTANT group, that subject was discarded from the matched analysis. Finally, of 2083 patients in both groups, 1398 patients (67%) were matched. Finally, the baseline covariates were compared between the two groups with statistical tests for matched data. All p values are 2-sided, and p values <0.05 were considered to indicate statistical significance.

Studied patients were derived from our Nobori and CONSTANT Korea registries. In brief, 1225 patients were undergoing PCI with BES and 858 patients with ZES. The two groups differed significantly in terms of some potential confounders (Table 1). Particularly, patients who received BES were more likely to present with diabetes and to be managed as acute coronary syndromes. On the other hand, those who received ZES were more likely insulin-dependent diabetes and had previous history of MI. In coronary intervention, Nobori group had more left main and left anterior descending coronary artery involvement, and their lesions were more severely calcific, tortuous and thrombus with a need for more than one stent in comparison with Resolute group. The use of intravascular ultrasound evaluation IVUS and bifurcation side-branch stenting was higher in Nobori group, while multivessel involvement was higher in the Resolute group.

The two groups did not differ significantly in terms of age, gender, hypertension, hyperlipidemia, past history of coronary bypass surgery, previous percutaneous interventions, and history of cerebrovascular accidents. Nobori group had higher tendency to present with hyperlipidemia. After PSM (Table 2) was performed for the entire group (n=2083), there were 699 matched pairs of patients. In matched analysis, two propensity-matched groups (699 pairs, n=1398 patients) were generated and the baseline characteristics of these two groups were balanced. In this matched cohort, the mean age was 64.5 years, men accounted for 72.7%, 32.9% were diabetes mellitus, 61.2% were hypertensive, and 45.9% were hyperlipidemic. Clinically, 52.9% presented with acute coronary syndromes. 22.3% had a multivessel disease, 6.7% had severely calcific lesions, 8.7% of lesions were thrombus containing and 11.8% were bifurcated lesions. The mean total stent length was 22.4 mm and mean minimal stent diameter was 22.3 mm. In matched group, there was no significant difference between the Nobori and Resolute group in any other covariates.

Table 3 shows 12-months clinical outcomes after PSM. Among MACCEs components, there was no significant difference between Nobori and Resolute groups in terms of either target lesion revascularization (1.1% Nobori vs. 0.7% Resolute, p=0.579) or TVR [5 cases in either group (0.72%)]. Cardiac death was also similar, occurring in 0.7% in patients who received Nobori vs. 0.4% in those received Resolute (p=0.73), as was non-Q wave MI, 0.14% in both groups. There was neither Q wave MI nor stent thrombosis in the matched cohorts. Hemorrhagic stroke occurred once in the Nobori group 0.1%, and bleeding was also similar; four cases (0.6%) in either group. Non-TVR was similar between Nobori and Resolute [11% vs. 10%, respectively (p=1.000)].

We have some limitations about this article. First, this article is not a randomized controlled clinical trial; therefore, our comparison between new generation stents even after propensity matching to overcome the limitation should be carefully analyzed and applied individually for real world practice. Second, intravascular imaging tools such as intravascular ultrasound, and optical coherence tomography can unravel strong clue for final discrimination about information of neointimal coverage and failure of apposition between different designs and character of stents. However, because of limitation of registry data, we are not able to suggest about parameters for vascular imaging.