Author + information
- Received August 1, 2018
- Revision received September 6, 2018
- Accepted September 11, 2018
- Published online December 17, 2018.
- Maayan Konigstein, MDa,b,
- Ori Ben-Yehuda, MDa,c,
- Pieter C. Smits, MDd,
- Michael P. Love, MDe,
- Shmuel Banai, MDb,
- Gidon Y. Perlman, MDf,g,
- Mordechai Golomb, MDa,
- Melek Ozgu Ozan, MSa,
- Mengdan Liu, MSa,
- Martin B. Leon, MDa,c,
- Gregg W. Stone, MDa,c and
- David E. Kandzari, MDh,∗ ()
- aClinical Trials Center, Cardiovascular Research Foundation, New York, New York
- bTel-Aviv Sourasky Medical Center, Tel Aviv, Israel
- cDivision of Cardiology, NewYork-Presbyterian Hospital/Columbia University Medical Center, New York, New York
- dMaasstad Ziekenhuis, Rotterdam, the Netherlands
- eUniversity of Manitoba, St. Boniface General Hospital, Winnipeg, Manitoba, Canada
- fHadassah Hebrew University Medical Center, Jerusalem, Israel
- gMedinol Ltd., Tel Aviv, Israel
- hPiedmont Heart Institute, Atlanta, Georgia
- ↵∗Address for correspondence:
Dr. David E. Kandzari, Piedmont Heart Institute, 95 Collier Road, Suite 2065, Atlanta, Georgia 30309.
Objectives The authors sought to investigate the impact of diabetes mellitus (DM) on outcomes following contemporary drug-eluting stent (DES) implantation in the BIONICS (BioNIR Ridaforolimus Eluting Coronary Stent System in Coronary Stenosis) trial.
Background Patients with DM are at increased risk for adverse events following percutaneous coronary intervention (PCI).
Methods A prospective, multicenter, 1:1 randomized trial was conducted to evaluate in a noninferiority design the safety and efficacy of ridaforolimus-eluting stents versus zotarolimus-eluting stents among 1,919 patients undergoing PCI. Randomization was stratified to the presence of medically treated DM, and a pre-specified analysis compared outcomes according to the presence or absence of DM up to 2 years.
Results The overall prevalence of DM was 29.1% (559 of 1,919). DM patients had higher body mass index, greater prevalence of hyperlipidemia and hypertension, and smaller reference vessel diameter. One-year target lesion failure (cardiac death, target vessel myocardial infarction, or ischemia-driven target lesion revascularization) was significantly higher among diabetic patients (7.8% vs. 4.2%; p = 0.002), mainly due to higher target lesion revascularization (4.5% vs. 2.0%; p = 0.002). Rates of cardiac death, myocardial infarction, and stent thrombosis did not statistically vary. Among 158 patients undergoing 13-month angiographic follow-up, restenosis rates were 3 times higher in diabetic patients compared with nondiabetic patients (15.2% vs. 4.7%; p = 0.01). Clinical and angiographic outcomes were similar between ridaforolimus-eluting stent– and zotarolimus-eluting stent–treated patients.
Conclusions Despite advances in interventional therapies, and the implementation of new-generation DES, diabetic patients still have worse angiographic and clinical outcomes compared with nondiabetic patients undergoing PCI.
Despite initially successful revascularization, the presence of diabetes mellitus (DM) has been consistently predictive of increased risk for repeat revascularization, myocardial infarction (MI), and death following percutaneous coronary intervention (PCI) (1–3). Drug-eluting stents (DES) have significantly reduced restenosis rates in patients with and without DM compared with bare-metal stents (4); however, controversy still exists regarding the relative effectiveness of DES in the treatment of patients with DM in comparison to those without DM. Specifically, whereas selected studies suggest DES have achieved similar rates of clinical restenosis among patient populations with and without DM (4,5), other trials continue to report DM as a major risk factor for DES failure (6,7). Furthermore, a pooled analysis of 18 randomized controlled trials of DES showed that patients with DM remained at increased risk for target lesion failure (TLF) and target vessel revascularization (TVR) following PCI with DES; however, DM was a risk factor for repeat revascularization at 1 year only in patients with complex lesions (8). Therefore, the influence of DM on outcomes following revascularization with DES used in current clinical practice remains uncertain.
The BIONICS (BioNIR Ridaforolimus Eluting Coronary Stent System in Coronary Stenosis) trial was a prospective, randomized, multicenter study comparing ridaforolimus-eluting stents (RES) and zotarolimus-eluting stents (ZES) in patients undergoing PCI (9). This study allowed broad clinical and angiographic inclusion criteria considered representative of many patients treated in clinical practice and demonstrated noninferiority between the DES groups regarding 1-year TLF. As a pre-specified analysis, we sought to investigate the impact of medically treated DM on outcomes following revascularization with these contemporary DES.
The BIONICS trial was a prospective, randomized, single-blind, multicenter trial comparing RES (EluNIR, Medinol Ltd., Tel Aviv, Israel) and ZES (Resolute Integrity or Resolute Onyx, Medtronic, Minneapolis, Minnesota) in patients undergoing PCI. Inclusion criteria have been described previously (9). In brief, patients with ischemic heart disease undergoing planned stent implantation were eligible for enrollment. Angiographic inclusion criteria included a reference vessel diameter between 2.5 mm and 4.25 mm, with a maximum of 2 lesions per vessel in up to 2 major coronary arteries. Calcified lesions requiring atherectomy were permitted, as were chronic total occlusions, planned 1-stent bifurcations, bypass graft stenoses, and bare-metal in-stent restenosis. Patients with recent (<24 h) ST-segment elevation MI, left ventricular ejection fraction <30%, active stent thrombosis (ST), creatinine clearance <30 ml/min, prior PCI within 12 months, and those unlikely to adhere to dual antiplatelet therapy were excluded. The study was approved by the institutional review board or ethics committee at each enrolling site, and eligible patients signed written informed consent before the interventional procedure.
Definition of DM
The presence or absence of medically treated DM was determined by the enrolling physician based on information collected from the patients and their medical records. Patients were considered to have DM if they were previously diagnosed and treated with insulin or noninsulin DM medical treatment. Patients who were diagnosed, but not medically treated (diet only), were considered nondiabetic.
The EluNIR stent has been described in detail elsewhere (9). Briefly, it features a cobalt alloy platform with 87-μm-thick struts with adaptive cells capable of differential lengthening to provide uniform drug distribution in variable vessel anatomy. A proprietary copolymer, 7-μm thick, is circumferentially coated on the stent. The polymer permits controlled elution of ridaforolimus, which is an analog of sirolimus. EluNIR is available in diameters ranging from 2.5 mm to 4.0 mm and in lengths from 8 mm to 33 mm. The comparator ZES (Resolute Integrity or Resolute Onyx) is available in diameters from 2.25 mm to 4.0 mm and in lengths ranging from 8 mm to 38 mm.
Randomization, interventional procedures, and adjunctive drug therapy
Patients were blinded to treatment assignment and randomized to RES or ZES in a 1:1 fashion. Randomization was stratified according to the presence or absence of medically treated DM, presentation with acute coronary syndromes versus stable angina, and enrolling site. Dual antiplatelet therapy use was mandatory pre-procedure and for a minimum of 6 months following the procedure. Anticoagulation during the procedure was prescribed according to local standards. Clinical events were assessed during hospital stay and at 30 days, 12 months, and 2 years after the index procedure.
Data management and core laboratories
All data were submitted to a central data coordinating facility (Cardiovascular Research Foundation, New York, New York). An independent clinical events committee adjudicated all primary and secondary clinical endpoints blinded to stent type. An independent data safety monitoring board was responsible for scheduled review of the clinical safety data and could recommend study discontinuation or modification. Coronary angiograms performed at baseline and at any time during the follow-up period were reviewed by an independent core laboratory.
Study endpoints and definitions
Study endpoints and definitions for the present analysis were based on those of the BIONICS study (9) The primary endpoint was TLF at 12 months, defined as the composite of cardiac death, target vessel–related MI, or ischemia-driven target lesion revascularization (TLR). Secondary clinical safety and efficacy endpoints included major adverse cardiac events (MACE) (cardiac death, MI, or ischemia-driven TLR); target vessel failure (TVF) (all-cause death, target vessel-related MI, or ischemia-driven target vessel revascularization. Definite or probable ST were defined according to the Academic Research Consortium criteria (10). Device success was defined as the achievement of <50% diameter stenosis of the target lesion (determined by the angiographic core laboratory) with the assigned study stent, and procedure success was defined as a final diameter stenosis <50% with the assigned stent and any adjunctive device, and with no in-hospital MACE. Periprocedural MI was defined according to the Society of Coronary Angiography and Interventions criteria (11), whereas spontaneous MI was defined according to the Third Universal Definition of Myocardial Infarction (12). Secondary angiographic efficacy endpoints at 13-month follow-up were determined by the angiography core laboratory. Binary restenosis was defined as a stenosis ≥50% of the lumen diameter of the target lesion. Restenosis patterns were characterized according to established criteria (13). All endpoints were evaluated at 30 days, 12 months, and 2 years.
Baseline characteristics of study patients were summarized in frequencies and percentages for categorical variables and by means and SDs for continuous variables. Patient-level categorical variables were compared between patients with and without DM by chi-square or the Fisher exact test. Continuous variables were compared by the Student’s t-test or Wilcoxon rank sum test for non-normally distributed data. Lesion-level variables were compared using generalized estimating equations to account for correlation of lesions within a subject. The 30-day, 1-year, and 2-year clinical events were summarized as Kaplan-Meier estimates and compared with the log-rank test. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated from a Cox proportional hazards model that included the presence or absence of medically treated DM as a covariate. Independent predictors of TLF at 12 months were examined using a multivariable stepwise Cox proportional hazards model with entry and exit criteria set at p = 0.1. The variables entered in the model were randomized treatment, presence or absence of medically treated DM, sex, age (≥65 vs. <65 years), body mass index, history of hypertension, hyperlipidemia, prior MI, prior PCI, prior coronary artery bypass grafting (CABG), renal insufficiency, smoking status (current vs. former/never), presence or absence of acute coronary syndromes, and lesion complexity (American College of Cardiology lesion class B2/C vs. others). A p value of 0.05 was established as the level of statistical significance for all comparisons. All statistical analyses were performed with SAS version 9.4 (SAS Institute, Cary, North Carolina).
Between March 2014 and August 2015, 1,919 patients were enrolled in the BIONICS trial (958 randomized to RES, 961 to ZES). The study population included 559 patients (29%) with and 1,360 (71%) without medically treated DM. Median follow-up was 733 days (interquartile range: 702 to 770 days). Baseline clinical and angiographic characteristics according to the presence or absence of DM are presented in Table 1. Patients with DM had higher body mass index (30.3 ± 6 kg/m2 vs. 28.5 ± 5 kg/m2; p < 0.0001), a greater prevalence of hyperlipidemia and hypertension (90.6% vs. 74.5% and 86.5% vs. 67.7%, respectively; p < 0.0001 for both), and prior revascularization (either CABG or PCI; p < 0.0001 for both). Reference vessel diameter was slightly smaller (2.69 ± 0.5 mm vs. 2.76 ± 0.5 mm; p < 0.001), and severe calcification was more frequent (14.2% vs. 10.9%; p = 0.02) among patients with DM.
Medical treatment at baseline and follow-up
Medical treatment of the study population at baseline and follow-up (30 days) is presented in Online Table 1. At baseline, antiplatelet treatment was similar in patients with and without DM, except for ticagrelor being more frequently used in patients without DM. Treatment with lipid-lowering medications, calcium channel blockers, and renin-angiotensin system blocking drugs was more frequent among patients with versus without DM. At 30 days and at 1-year follow-up, medical treatment was overall similar between groups except for ticagrelor and antianginal medications being more frequently used among patients without DM, whereas angiotensin receptor blockers and non-statin lipid-lowering drugs were used more often among patients with DM. At 2-year follow-up, patients with DM were more frequently treated with dual antiplatelet therapy compared to patients without DM (73.8% vs. 66.6%, respectively; p = 0.003).
At 30 days, TVR was more frequent in the DM group (1.4% vs. 0.5%; p = 0.04), and there was no significant difference in the rates of TLF and ST between the DM and no DM groups (3.9% vs. 2.5%; p = 0.56 and 0.7% vs. 0.3%; p= 0.19, respectively) (Online Table 2).
One- and 2-year clinical outcomes are shown in Table 2 and Figures 1 and 2⇓⇓. At both 1 and 2 years, TLF, MACE, and TVF were higher in the DM group (7.8% vs. 4.2%; p = 0.002 for TLF at 1 year and 11.3% vs. 5.8%; p < 0.0001 for TLF at 2 years). Death and MI, were not significantly different between the groups. There was a total of 14 (0.7%) cases of ST in 2 years; early, late, and very late ST occurred in 8 (0.4%), 2 (0.1%). and 4 (0.2%) patients respectively. There was no significant difference in the rate of ST between patients with and without DM at all time points.
Insulin-treated patients versus noninsulin-treated patients
The DM group consisted of 181 patients (32.4%) treated with insulin and patients 378 (67.6%) treated with a noninsulin treatment. The comparison of 1- and 2-year clinical outcomes between insulin-treated patients, noninsulin-treated patients, and nondiabetic patients is presented in Table 3. At 1 year, TLF, MACE, ST, and TVR were all significantly more frequent among insulin-treated patients compared with noninsulin-treated patients and non-DM patients. A trend toward more MI events and all-cause mortality among insulin-treated patients was also observed. At 2 years, MACE and TVF were still more frequent among insulin-treated patients (both p = 0.01), with only a trend toward higher TLF (14.6% vs. 9.8%; p = 0.08) (Table 3). The comparison of TLF rates among insulin-treated patients, noninsulin-treated patients, and nondiabetic patients is presented in Figure 3. Of note, at 1 year, there was no significant difference between noninsulin-treated patients and patients with no diabetes, whereas at 2 years, the difference between these groups became significant (9.8% vs. 5.8%; p = 0.007, for noninsulin-treated patients versus nondiabetic patients, respectively).
Scheduled angiographic follow-up was performed in a total of 158 patients (8.0% of patients, 198 lesions). The comparison between 13-month angiographic parameters of patients with versus without DM is presented in Table 4 and Online Figure 1. Despite similar late lumen loss, in-stent binary restenosis (≥50% of the lumen diameter) rates were 3 times greater in patients with DM versus without DM (15.2% vs. 4.7%; p = 0.01).
Performance of RES versus ZES according to DM status
No differences were found in the performance of RES versus ZES regardless of DM status (the comparison between the stents is presented in Online Table 3). Device and procedure success rates were similar between RES- and ZES-treated patients in the DM as well as in the non-DM groups (pinteraction = 0.61 and 0.60 for device and procedure success, respectively). Rates of TLF at 30 days, 1 year, and 2 years were also similar between patients treated with RES versus patients treated with ZES both in the DM and no DM groups (pinteraction = 0.58, 0.91, and 0.53 for 30 days, 1 year, and 2 years, respectively). Among patients with DM who underwent angiographic follow-up (n = 55), in-stent percent diameter stenosis as well as binary restenosis were similar between RES- and ZES-treated patients (23.7 ± 21 vs. 21.2 ± 20; p = 0.63, and 17.1% vs. 12.0%; p = 0.58, respectively). Late lumen loss was also similar (0.27 ± 0.5 mm vs. 0.24 ± 0.4 mm; p = 0.79).
Independent predictors of TLF at 1 year
When performing a multivariable analysis for predictors of TLF, DM (HR: 1.71, 95% CI: 1.15 to 2.57; p = 0.009) together with prior CABG (HR: 1.97, 95% CI: 1.17 to 3.31; p = 0.01), complex lesions (American College of Cardiology lesion class 2B/C; HR: 1.59, 95% CI: 1.01 to 2.50; p = 0.047), and age (≥65 vs. <65 years, HR: 1.66, 95% CI: 1.11 to 2.49; p = 0.015) were identified as independent predictors of TLF at 1 year.
In the present study, we investigated the impact of DM on clinical and angiographic outcomes of patients undergoing PCI using contemporary DES in a largescale, randomized, controlled trial. The main findings of the study are the following: First, despite the advances in interventional therapies and the implementation of new-generation DES, patients with DM still present higher rates of TLF and MACE, driven by higher rates of TLR. Second, insulin-treated DM patients are at higher risk of adverse outcomes compared with noninsulin-treated DM patients, and at 2 years, both have a higher risk compared with patients with no DM. Third, RES and ZES presented similar performance among patients with and without DM.
The typical pattern of diabetic coronary artery disease is characterized by a greater atherosclerotic burden and diffuse disease, with smaller reference vessel diameter and poor collateral formation compared with nondiabetic patients (14,15). Hence, outcomes of coronary revascularization in diabetic patients have been inferior to nondiabetic patients. Specifically, PCI in diabetic patients is associated with an increased rate of restenosis, repeat revascularization, ST, and mortality (1,3). The higher rates of repeat revascularization and mortality after PCI in diabetic patients are caused by both stent failure, mainly in-stent restenosis, and disease progression. In diabetic patients, the leading cause of restenosis is a pronounced and accelerated intimal hyperplasia in combination with more vascular inflammation, endothelial dysfunction, and the direct growth factor–like effect of insulin on vascular smooth muscle and neointimal cells (16,17). Furthermore, the typical coronary artery disease pattern in DM, characterized by the presence of long and diffuse lesions with high grades of calcification and tortuosity, and the unfavorable location of the lesions, contribute to a higher risk of restenosis after PCI (16).
The introduction of DES reduced the rate of restenosis in diabetic and nondiabetic patients compared with bare-metal stents (4,18,19). Previous studies evaluating the outcomes of DES implantation in diabetic compared with nondiabetic patients reported conflicting results (4,5,20). The largest study to date, which included patients treated with DES from 18 randomized controlled trials, demonstrated that DM is a strong predictor of TLR and TVR; however, this finding was confined to patients with complex lesions in whom DM was associated with an approximate 80% increase in TLR and TVR at 1 year compared with those without DM. By contrast, the 1-year rates of TLR and TVR in patients with and without DM and more simple lesions were favorable (about 95% freedom from revascularization at 1 year) and nearly identical (8).
Our study, evaluating the performance of modern DES in patients treated for more complex clinical indications and lesion anatomy than most randomized controlled trials, indicates that DM persists as a risk factor for DES failure, even after adjusting for other important known clinical and angiographic risk factors including lesion complexity. Moreover, no significant interaction was found between DM and lesion complexity for the primary outcome of TLF (data not shown). Importantly, whereas TLF, MACE, TLR, and TVR were higher in DM patients compared with non-DM patients, the rate of MI was similar between groups, with only a trend toward more target vessel MI among patients with diabetes. Moreover, the rate of ST was also similar between patients with and without DM, but significantly higher in patients with DM patients treated with insulin. A recent study evaluating the outcomes of patients undergoing PCI with second-generation DES (21) also showed similar rate of MI in patients with and without DM, regardless of insulin treatment, and higher risk of ST that was limited to DM patients treated with insulin.
Insulin-treated patients showed worse clinical outcomes compared with noninsulin-treated patients and patients with no DM. This finding is in line with previous reports on bare-metal stents (22) and DES (20,21). Specifically, Tada et al. (20) demonstrated a significant increase, not only in the rate of TLR, but also in the composite endpoint of death, MI, and stroke among insulin-treated patients compared with patients with noninsulin-treated DM. Pi et al. (21) reported an increased risk of cardiac death, any revascularization, and ST in insulin-treated patients compared with DM patients not treated with insulin and to patients without DM. In our study, the increased rate of TLF and ST among insulin-treated diabetic patients, especially during the first year following DES implantation, probably indicates a more accelerated intimal hyperplasia, greater degree of vascular inflammation and/or endothelial dysfunction, and increased plaque vulnerability (23,24) in insulin-treated diabetic patients, and highlights the need for DES with even more potent antirestenosis efficacy.
Finally, in patients undergoing routine angiographic follow-up, despite similar late lumen loss in diabetic and nondiabetic patients, restenosis of ≥50% was more frequent among diabetic patients. These findings may further support the presumed multifactorial pathophysiological process of stent failure among diabetic patients and the ongoing need for better treatment strategies aimed at diabetic patients.
Although enrollment of patients in the BIONICS trial was stratified for DM, the trial was not powered for this subgroup analysis, and therefore, it is possible that some of the differences noted in the study, such as the difference in the rate of MI between patients with and without DM, did not reach statistical significance due to the limited power of this analysis. Second, in this study, patients were randomized to implantation of either RES or ZES. Due to possible differences between new-generation stents, the findings of this study should be interpreted with caution when discussing the performance of other contemporary stents. Third, routine angiographic follow-up was performed in only a minority of the study population.
In this subgroup analysis of the BIONICS trial, diabetic patients presented higher rates of TLF and MACE driven by a greater incidence of TVR, indicating that modern DES still fail more frequently among patients with versus without DM and among insulin-treated versus noninsulin-treated diabetic patients. No differences were noted in the performance of RES and ZES among diabetic and non diabetic patients.
WHAT IS KNOWN? DM is associated with increased risk of repeat revascularization, myocardial infarction, and death following percutaneous coronary intervention.
WHAT IS NEW? Modern drug-eluting stents still fail more frequently among patients with versus without DM and among insulin-treated versus noninsulin-treated diabetic patients. The novel EluNIR ridaforolimus-eluting stents performed as well as zotarolimus-eluting stents among patients with DM.
WHAT IS NEXT? The study highlights the need for drug-eluting stents with even more potent anti-restenosis efficacy to be used among diabetic patients.
Dr. Ben-Yehuda, Ms. Ozan, and Ms. Liu are employees of the Cardiovascular Research Foundation, which received funding from Medinol Ltd. for the conduct of the trial. Dr. Smits has received consulting/speakers fees from Medinol, Abbott Vascular, AstraZeneca, Terumo, and St. Jude Medical; and research support from Abbott Vascular, St. Jude Medical, and Terumo. Dr. Perlman is an employee of Medinol. Dr. Leon has an equity relationship with Medinol. Dr. Kandzari has received consulting fees from Medtronic, Cardinal Health, Boston Scientific, and Micell; and research support from Medtronic, Abbott Vascular, Boston Scientific, Biotronik, and Medinol. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery bypass grafting
- confidence interval
- drug-eluting stent(s)
- diabetes mellitus
- hazard ratio
- major adverse cardiac event(s)
- myocardial infarction
- percutaneous coronary intervention
- ridaforolimus-eluting stent(s)
- stent thrombosis
- target lesion failure
- target lesion revascularization
- target vessel failure
- target vessel revascularization
- zotarolimus-eluting stent(s)
- Received August 1, 2018.
- Revision received September 6, 2018.
- Accepted September 11, 2018.
- 2018 American College of Cardiology Foundation
- Cutlip D.E.,
- Chhabra A.G.,
- Baim D.S.,
- et al.
- Kirtane A.J.,
- Ellis S.G.,
- Dawkins K.D.,
- et al.
- Kedhi E.,
- Genereux P.,
- Palmerini T.,
- et al.
- Kandzari D.E.,
- Smits P.C.,
- Love M.P.,
- et al.
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.
- Moussa I.D.,
- Klein L.W.,
- Shah B.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Mehran R.,
- Dangas G.,
- Abizaid A.S.,
- et al.
- Abaci A.,
- Oguzhan A.,
- Kahraman S.,
- et al.
- Kip K.E.,
- Faxon D.P.,
- Detre K.M.,
- Yeh W.,
- Kelsey S.F.,
- Currier J.W.
- Bernelli C.,
- Chan J.,
- Chieffo A.
- Kornowski R.,
- Mintz G.S.,
- Kent K.M.,
- et al.
- Pi S.H.,
- Rhee T.M.,
- Lee J.M.,
- et al.
- Abizaid A.,
- Kornowski R.,
- Mintz G.S.,
- et al.