A total of 929 consecutive patients underwent percutaneous coronary intervention (PCI) with a DES from April 2005 to December 2007 at Yeungnam University Medical Center for the treatment of coronary artery disease, and who provided consent were enrolled in this retrospective study. Patients were treated with one of the following DESs; Cypher (sirolimus-eluting coronary stent [SES], Cordis, Warren, NJ, USA), Taxus (paclitaxel-eluting coronary stent [PES], Boston Scientific, Natick, MA, USA) or Endeavor (zotarolimus-eluting coronary stent [ZES], Medtronic, Santa Rosa, CA, USA). On-label indications were for symptomatic ischemic heart disease by de novo lesions ≤ 30 mm long in native coronary arteries of diameter ≥ 2.5 to ≤ 3.5 mm (Cypher), lesions ≤ 28 mm long in arteries of diameter ≥ 2.5 to ≤ 3.75 mm (Taxus), and lesions ≤ 35 mm long in arteries of diameter ≥ 2.25 to ≤ 4.2 mm, including those with diabetes mellitus (Endeavor) [10-12]. Off-label use was defined as DES use, according to the manufacturer’s instructions, for a lesion or patient subset not extensively studied in randomized trials and without FDA approval, and included acute MI (< 48 hours), very small vessel (< 2.5 mm), long stenotic lesions (> 30 mm), chronic total occlusions, bifurcation lesions, ostial lesions (< 5 mm from orifice), left main coronary artery diseases, multivessel diseases, saphenous vein graft lesions, and lesions with thrombus [6,9]. The 929 patients were divided into two groups according to the on-label group (n = 449) and the off-label group (n = 480) indications.

Decisions regarding the types of DESs used were made by respective operators. Prior to PCI, all patients received a 300 mg of aspirin and 300 or 600 mg of clopidogrel if not already on a maintenance dose or unknown. Unfractionated heparin boluses of 5,000 U were initially administered intravenously and then additional heparin boluses were given to maintain a targeting activated clotting time of 250 to 300 seconds during PCI. Coronary angiography and stenting were performed using routine methods. Daily 100 mg of aspirin and 75 mg of clopidogrel were continued for at least 6 months postimplantation regardless of the stent type. Durations of dual antiplatelet agent administration were determined by physician’s discretion based on the recommendations of the American College of Cardiology/American Heart Association [13].

Baseline characteristics and laboratory findings were collected. Left ventricular ejection fraction (LVEF) was measured by transthoracic echocardiography during index hospitalization. Procedural data, including the type and number of stents implanted, target lesion sites, and quantitative coronary angiography data (lesion length, stent diameter, and stent length) were assessed by the operator.

The endpoints of this study were composite of major adverse cardiac events (MACEs), which were defined as all-cause deaths, MI, ischemic-driven target vessel revascularization (Id-TVR), and ST. All-cause death was defined as the sum of cardiac and noncardiac deaths, which meant death that was not primarily due to cardiac causes. MI was defined as the presence of at least two of the following: anginal pain, elevation of creatine kinase to more than 3 times the upper limit of normal, and electrocardiographic changes suspicious for MI. Id-TVR was defined as intervention for chest pain or a positive test for ischemia (exercise stress test, stress echocardiogram, resting electrocardiographic evidence of ST-segment deviation in more than 1 lead, or a radionuclide study showing a reversible defect) [14]. ST was classified as ‘definite’ or ‘probable’ as proposed by the Academic Research Consortium, according to which definite ST refers to the presence of thrombus in a stent documented by angiography, and probable ST means any MI related to documented acute ischemia in the territory of an implanted stent without angiographic confirmation, or sudden or unexplained death within 30 days of implantation [15].

Overall clinical outcomes were determined by comprehensively reviewing medical records at routine outpatient visits or by telephone interview during the 5-year follow-up period. A committee independently adjudicated all clinical events. Clinical follow-up data was available for all 929 study subjects.

Results were reported as mean ± standard deviation for numerical variables and as frequencies and percentages for categorical variables. The independent t test was used to compare group continuous variables, and the chi-square test or Fisher exact test was used to compare categorical variables. Incidences of MACE, all-cause death, Id-TVR, MI, and ST in the two groups were determined using Kaplan-Meier event curves, and compared using log-rank tests. Multivariate logistic regression analysis was used to identify predictors of MACE in the off-label group. SPSS version 18.0 (SPSS Inc., Chicago, IL, USA) was used for the statistical analysis, and statistical significance was accepted for p values < 0.05.

Of the 929 study subjects, 710 patients were completely monitored for 5 years (follow-up rate, 76.4%). The baseline clinical characteristics of the on- and off-label groups are shown in Table 1. The groups were similar in terms of age, sex, previous medical history, smoking, hypertension, and low density lipoprotein cholesterol level. However, the on-label group had a lower prevalence (24.9% [n = 112] vs. 34.2% [n = 164], p = 0.002) and higher LVEF value (56% ± 11% vs. 52% ± 12%, p = 0.001). More patients in the off-label group were taking cilostazol (8.1% [n = 39] vs. 4.9% [n = 22], p = 0.048), but no significant intergroup difference was found for using of other periprocedural medicines, including other antiplatelet drugs (aspirin, clopidogrel, and ticagrelor), β-blocker, nitrate, or 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor.

Four hundred eighty patients (51.6%) were treated using off-label indications (Table 2) and 22.1% of the patients had more than two off-label indications. The most two common off-label indications were acute MI (n = 221, 46.0%) and chronic total occlusion (n = 381, 20.6%). Stent types and sites of implantation are listed in Table 1. In terms of procedural data, marked differences were observed between the two groups; the off-label group had a longer mean lesion length, a larger mean stent length, a smaller mean stent diameter, and more stents. These differences seemed to be natural consequences of following off-label indications.

Over the first year postimplantation; total MACEs (8.2% [n = 38] vs. 3.7% [n = 16], p = 0.005), Id-TVR (5.0% [n = 24] vs. 1.6% [n = 7], p = 0.004), and ST (1.7% [n = 7] vs. 0.3% [n = 1], p = 0.042) were significantly higher in the off-label group, and rates of all-cause death and MI were non-significantly higher in off-label group. Similar tendencies were observed over the 5-year postimplantation period; total MACEs (17.5% [n = 84] vs. 9.4% [n = 34], p < 0.001), Id-TLR (13.1% [n = 63] vs. 5.8% [n = 29]), and ST (2.1% [n = 9] vs. 0.3% [n = 1], p = 0.021) (Table 3).

MACE-free survival curves for the two groups at 1 year and at 5 years are shown in Fig. 1. For both periods, the on-label group had significantly higher event-free survival rates (p = 0.001). Multivariable analysis showed that multivessel disease (hazard ratio [HR], 2.00; 95% confidence interval [CI], 1.09 to 2.49; p = 0.004) and diabetes (HR, 1.65; 95% CI, 1.25 to 3.22; p = 0.017) were significant risk factors of long-term MACE in the off-label group (Table 4).