Effects of stent generation on clinical outcomes after acute myocardial infarction compared between prediabetes and diabetes patients

We investigated the effects of stent generation on 2-year clinical outcomes between prediabetes and diabetes patients after acute myocardial infarction (AMI). A total of 13,895 AMI patients were classified into normoglycemia (group A: 3673), prediabetes (group B: 5205), and diabetes (group C: 5017). Thereafter, all three groups were further divided into first-generation (1G)-drug-eluting stent (DES) and second-generation (2G)-DES groups. Patient-oriented composite outcomes (POCOs) defined as all-cause death, recurrent myocardial infarction (Re-MI), and any repeat revascularization were the primary outcome. Stent thrombosis (ST) was the secondary outcome. In both prediabetes and diabetes groups, the cumulative incidences of POCOs, any repeat revascularization, and ST were higher in the 1G-DES than that in the 2G-DES. In the diabetes group, all-cause death and cardiac death rates were higher in the 1G-DES than that in the 2G-DES. In both stent generations, the cumulative incidence of POCOs was similar between the prediabetes and diabetes groups. However, in the 2G-DES group, the cumulative incidences of Re-MI and all-cause death or MI were significantly higher in the diabetes group than that in the prediabetes group. To conclude, 2G-DES was more effective than 1G-DES in reducing the primary and secondary outcomes for both prediabetes and diabetes groups.


Discussion
The primary findings of this study are as follows: (1) in both prediabetes and diabetes groups, the cumulative incidences of POCOs, any repeat revascularization, and ST were higher in the 1G-DES than that in the 2G-DES; (2) in the diabetes group, the cumulative incidences of all-cause death and CD were higher in the 1G-DES than www.nature.com/scientificreports/ that in the 2G-DES; (3) in the normoglycemia group, the cumulative incidence of ST was higher in the 1G-DES than that in the 2G-DES; and (4) in two different stent generations, the cumulative incidence of POCOs was similar between the prediabetes and diabetes groups. However, in the 2G-DES group, the cumulative incidences of Re-MI and all-cause death or MI were higher in the diabetes group than that in the prediabetes group. Hyperglycemia, elevated free fatty acid level, and increased amount of circulating glucosylated serum products can accelerate atherosclerosis and vascular injury in patients with diabetes by inducing endothelial dysfunction and vascular inflammation 16 . Although previous reports demonstrated that the higher rates of repeat revascularizations and mortality after PCI in patients with diabetes are caused by restenosis and disease progression 4-6 , comparative clinical outcomes between prediabetes and diabetes were not well illuminated especially, between 1G-DES and 2G-DES. Some recent reports showed that prediabetes is associated with poorer clinical outcomes including cardiovascular mortality and patients with prediabetes and diabetes have similar higher risk profiles compared with normoglycemia 13,14,17 .
Although DES improved outcomes of high-risk patients by reducing the rate of restenosis as compared with BMS 18,19 , ST remains a major concern after the DES implantation, especially in diabetes 4 . Relative superiority between the 1G-DES and 2G-DES in patients with AMI and diabetes remains controversial, and most previous studies were not performed during the prediabetes stage 10,20,21 . In our study, the cumulative incidence of POCOs was significantly higher in the 1G-DES than that in the 2G-DES in both prediabetes and diabetes groups. Moreover, in two different stent generations, the cumulative incidence of POCOs was similar between the prediabetes and diabetes groups (Table 3). In a substudy of the multicenter BIO-RESORT (BIOdegradable Polymer and DuRable Polymer Drug-eluting Stents in an All COmeRs PopulaTion) trial 13 , comparative clinical outcomes were similar between prediabetes and diabetes (11.1% vs. 10.5%). Von Birgelen et al. 22 reported the results of the BIO-RESORT Silent Diabetes Study. In their study, the cumulative incidence of major adverse cardiac events was different between patients with prediabetes (5.5%) and normoglycemia (3.0%) (Log-rank, Table 1. Baseline clinical, laboratory, and procedural characteristics. Values are means ± SD or numbers and percentages. The p values for continuous data were obtained from the analysis of variance. The p values for categorical data were obtained from the chi-square or Fisher's exact test. PCI percutaneous coronary intervention, BMS bare-metal stents, 1G first-generation, 2G second-generation, DES drug-eluting stents, LVEF left ventricular ejection fraction, BMI body mass index, CPR cardiopulmonary resuscitation, MI myocardial infarction, CABG coronary artery bypass graft, CVA cerebrovascular accidents, HF heart failure, CK-MB creatine kinase myocardial band, NT-ProBNP N-terminal pro-brain natriuretic peptide, Hs-CRP highsensitivity-C-reactive protein, HDL high-density lipoprotein, LDL low-density lipoprotein, ACEIs angiotensin converting enzyme inhibitors, ARBs angiotensin receptor blockers, CCBs calcium channel blockers, IRA infarct-related artery, ACC/AHA American College of Cardiology/American Heart Association, CAD coronary artery disease, SES sirolimus-eluting stent, PES paclitaxel-eluting stent, ZES zotarolimus-eluting stent, EES everolimus-eluting stent, BES biolimus-eluting stent, IVUS intravascular ultrasound, OCT optical coherence tomography, FFR fractional flow reserve. p = 0.07). As mentioned, despite the combination of new platforms, more biocompatible polymers were utilized in 2G-DES, the relative superiority between 1G-and 2G-DESs in patients with diabetes remains controversial [8][9][10][11] . In the SPIRIT V Diabetic Study 10 , everolimus-eluting stent (EES) was superior to paclitaxel-eluting stent (PES) for in-stent late loss at 9 months. The composite death, MI, and TVR rates were the same in the two groups at 1 year. Bavishi et al. 9 reported that EES showed significantly lower incidence rates of MACEs by 18% and ST by 46% as compared with the 1G-DES. Moreover, the EES showed a trend toward reduced incidence rates of target lesion revascularization (TLR) and TVR (p = 0.05). In this study, based on the cumulative incidences of POCOs, any revascularization rate was significantly higher in the 1G-DES than that in the 2G-DES group in www.nature.com/scientificreports/ both prediabetes and diabetes group. Therefore, the major clinical outcomes of our study could reflect the metaanalysis results of Bavishi et al. 's study 9 .
The overall rate of ST was also higher in the 1G-DES than in the 2G-DES in all three different glycemic groups (prediabetes [1.8% vs. 0.7%, log-rang p = 0.001], diabetes [2.1% vs. 0.9%, log-rank p = 0.007], and normoglycemia [1.5% vs. 0.5%, log-rank p = 0.009]). With regard to prediabetes, follow-up data on the comparative long-term effects of 1G-DES and 2G-DES implantation were limited. According to Bavishi et al. 's report 9 , EES reduced the incidence of ST by 46% (RR: 0.54, 95% CI 0.35-0.82) as compared with the 1G-DES in patients with diabetes. The cumulative incidence of ST also higher in the 1G-DES than that in the 2G-DES in patients with normoglycemia. Our result is consistent with the result of Nakatsuma et al. study 23 . This low cumulative incidence of 2G-DES may be related with relatively thin stent struts (50-90 μm) and improved ability for deliverability while maintaining an adequate radial strength 24 and more compatible and thromboresistant than those in the 1G-DES 25 . However, in our study, the occurrence of ST was high within 6 months after index PCI (Supplementary Fig. 1). Therefore, we cannot completely exclude the possibility that ST was associated with PCI procedure 26 . Even though IVUSguided 27 or functional flow reserve (FFR)-guided PCI 28 could reduce MACE rate, the number of PCI base on these intracoronary image-or functional study-based PCI were less than 30% in our study. Unfortunately, currently under the Korea's health insurance system, the reimbursement program for the use of IVUS, optical coherence of tomography, or fractional flow reserve during the PCI is very limited or absent 29 .
Interestingly, comparative clinical outcomes of the two different stent generations according to glycemic status showed some different results (Table 3). Different clinical outcomes among three different glycemic states (normoglycemia, prediabetes, and diabetes) were more prominent in the 2G-DES rather 1G-DES. According to advances in interventional skill, devices, and antiplatelet agents 5,6 , 2G-DES showed decreased incidences of all-cause death (aHR: 1.534; 95% CI 1.115-2.112; p = 0.009) and CD (aHR: 1.700; 95% CI 1.195-2.448; p = 0.003) compared with 1G-DES in diabetes group after adjustment ( Table 2). Bavishi et al. 9 showed that there was a trend towards reduction in all-cause mortality with zotarolimus compared to 1G-DES (6.3% vs. 7.2%, relative risk: 0.74; 95% CI 0.55-1.00; p = 0.05) in their meta-analysis. However, the cumulative incidences of all clinical outcomes were significantly higher in the diabetes than that in the normoglycemia group. These results may reflect hazardous effects of diabetes are sustained even in the era of 2G-DES.
In our study, in the 1G-DES, the primary and secondary end-points were similar between the prediabetes and diabetes groups. However, in the 2G-DES, the cumulative incidences of Re-MI (aHR: 1.393; 95% CI 1.135-2.043; p = 0.032) and all-cause death or MI (aHR: 1.224; 95% CI 1.023-1.524; p = 0.029) were significantly higher in the diabetes group than that in the prediabetes group. Although the precise mechanisms of the higher incidence of Re-MI in diabetes group are not fully known, one report 30 suggested that the association between diabetes and Re-MI may be related with a direct effect of diabetes. According to recent reports 14,31 , the cumulative incidence of Re-MI of the diabetes group was significantly higher than that of the prediabetes group (aHR: 1.884; 95% CI 1.201-2.954; p = 0.006 or aHR: 1.660; 95% CI 1.000-2.755; p = 0.020).
More than 50 high-volume university or community hospitals in South Korea participated in this study. The limited reports on the impact of stent generation on long-term clinical outcomes in AMI patients with prediabetes or diabetes were the motivation for the current study. Thus, we believe that our study may provide significant information to interventional cardiologists who perform PCI in patients with AMI with prediabetes or diabetes.
This study has several limitations. First, because the study population was obtained from the Korea AMI registry data, some data might be under-reported and/or missed. Second, it is necessary for diagnosing diabetes to check an HbA1c level ≥ 6.5%, FPG ≥ 126 mg/dL (7.0 mmol/L), and/or RPG ≥ 200 mg/dL (11.1 mmol/L) by repeat testing. If first glycemic status was diabetes and second was prediabetes, or first glycemic status was prediabetes and second was normoglycemia, and this is particularly important in patients with AMI, because such patients reveal hyperglycemia in acute phase. However, in this study, the definitions of prediabetes and diabetes were not based on repeat testing. Moreover, considering the limitation of HbA1c, any other diagnostic tests for diabetes including oral glucose tolerance test are needed for a finer classification. However, detailed information on this variable was not included in the KAMIR. Hence, the results of this study can be altered based on other diagnostic tests and which directly influences the assignment of participants, and this factor may have served as an important bias in this study 32 . Third, the duration and types of antidiabetic treatment are major determinants after PCI in patents with prediabetes or diabetes. However, this study was conducted based on discharge medications, and owing to limitation of registry study, we did not precisely know the adherence or non-adherence of enrolled patients to antidiabetic drugs during the follow-up period. Therefore, this may act as an important bias in this study. Fourth, 2G-DES consisted of durable-polymer-coated DES and biodegradable-polymer-coated DES. The number of biodegradable-polymer DES (BES) was highest in the normoglycemia group (prediabetes: 600/4438 (13.5%); diabetes: 536/4238 (12.6%); normoglycemia: 525/3191 (16.5%); p < 0.001) ( Table 1). Although this number was not significantly different between prediabetes and diabetes (p = 0.226), this division may be not reasonable and the composition of 2G-DES could be changed according to other types of utilized newergeneration DES. Therefore, other types of newer-generation DES could influence the outcome of our study. Fifth, although multivariate analysis was performed to strengthen our results, variables not included in the KAMIR may have affected the study outcomes. Sixth, the 2-year follow-up period in this study was relatively short in order to determine the long-term major clinical outcomes; therefore, data from studies with longer follow-up periods are required. Seventh, this study retrospectively enrolled the patients who underwent PCI from 2005-2015. The development of stent platform, potent antiplatelet drugs, and use of intracoronary imaging and improvement of procedural skills, all these factors substantially affect the clinical outcomes. Therefore, these factors could be also important bias of this study. Finally, although 2G-DES are considered the safest in the general population, this study confirms that in a select and growing population.
In conclusion, in this study, we observed that 2G-DES was more effective than 1G-DES in reducing POCOs, any repeat revascularization and ST in both prediabetes and diabetes group. Moreover, in two different stent www.nature.com/scientificreports/ generations, and the cumulative incidence of POCOs was similar between the prediabetes and diabetes groups. However, further studies regarding the most advanced DES technology joined with the most advanced antithrombotic regimen are needed to confirm these results.

Methods
Study population. A total of 45,322 patients with AMI who underwent successful stent implantation, including patients with DM aged ≥ 30 years at the onset of diabetes, from November 2005 to June 2015 in the KAMIR, were evaluated. Details of the registry can be found at the KAMIR website (http:// www. kamir. or. kr) 33 . Among them, patients with incomplete laboratory results (n = 9081, 20.0%), those who were lost to follow-up (n = 2175, 4.8%), those with unidentified blood hemoglobin (Hb) A1c and blood glucose level results (n = 13,931, 30.7%), those with different generations of stents were deployed in the same patients (n = 40, 0.1%), those who received dual antiplatelet therapy (DAPT) less than 12 months (n = 5438, 12.0%), and those who received BMS (n = 762, 1.7%). Finally, a total of 13,895 patients with AMI who underwent successful implantation were considered for inclusion. Patients were classified into normoglycemia (group A: 3673; 26.4%), prediabetes (group B: 5205; 37.5%), and diabetes (group C: 5017; 36.1%) ( Table 1). Subsequently, all three groups were further divided into 1G-DES (group A1, group B1, and group C1) and 2G-DES groups (group A2, group B2, and group C2) (Fig. 2). The study protocol was approved by the institutional review board of each participating center and the Chonnam National University Hospital Institutional Review Board ethics committee approved (approval number: CNUH-2011-172) the study protocol. The study has been performed in accordance with the ethical standards laid down in the 1975 Declaration of Helsinki. All persons gave their written informed consent prior to their inclusion in the study. All 13,895 patients completed the 2-year clinical follow-up by face-to-face interviews, phone calls, or medical chart review. All clinical events were evaluated by an independent event adjudi-    Percutaneous coronary intervention and medical treatment. Before PCI, all patients were administered loading doses of aspirin 200-300 mg and clopidogrel 300-600 mg; alternatively, ticagrelor 180 mg or prasugrel 60 mg was administered. PCI was performed via the femoral or radial approach after an intravenous bolus dose of heparin (50-100 U/kg) to achieve an activated clotting time of > 250 s. DAPT (a combination of aspirin 100 mg/day with clopidogrel 75 mg/day or ticagrelor 90 mg twice daily or prasugrel 5-10 mg/day) was recommended for > 12 months for patients who underwent PCI. Triple antiplatelet therapy (TAPT: cilostazol 100 mg twice daily in addition to DAPT) was left to the discretion of the individual operators. Diagnostic coronary angiography and PCI were performed using standard guideline 35 .

Study definitions and clinical outcomes.
Glycemic status was determined based on medical history and glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), and random plasma glucose (RPG) levels at the index hospitalization. According to the American Diabetes Association clinical practice recommendation 32 , prediabetes was defined as an HbA1c of 5.7-6.4% and an FPG of 100-125 mg/dL (5.6-6.9 mmol/L). Diabetes was categorized as either known diabetes defined as ongoing medical treatment for diabetes (insulin or antidiabetics), or newly diagnosed diabetes, defined as an HbA1c level ≥ 6.5%, FPG ≥ 126 mg/dL (7.0 mmol/L), and/ or RPG ≥ 200 mg/dL (11.1 mmol/L). If the admission electrocardiogram of patients who complained of chest pain showed ST-segment elevations in at least two contiguous leads of ≥ 2 mm (0.2 mV) in men, or ≥ 1.5 mm (0.15 mV) in women in leads V2-V3 and/or ≥ 1 mm (0.1 mV) in other contiguous chest leads or limb leads or new-onset left bundle branch block, the patients were considered to have STEMI 36 , whereas patients who did not show persistent ST-segment elevation with increased cardiac biomarkers and with appropriate clinical context were considered to have non-STEMI (NSTEMI) 37 . In cases of NSTEMI, an early invasive treatment strategy was defined as PCI within 24 h after admission 37 . A successful PCI was defined as a residual stenosis of < 30% and more than grade 3 flow in Thrombolysis In Myocardial Infarction flow for the infarct-related artery (IRA) after the procedure. The primary outcome of this study was the occurrence of POCOs, defined as all-cause death, Re-MI, or any coronary repeat revascularization 38 . The secondary outcome was definite or probable ST during the 2-year follow-up period. All-cause death was classified as CD or non-CD. Any repeat revascularization comprised target lesion revascularization, target vessel revascularization, and non-TVR. Re-MI, TLR, TVR, and non-TVR definitions have already been published previously 39,40 . The cumulative incidence of ST was defined by the current consensus 41 .
Statistical analyses. For continuous variables, differences between the two groups were evaluated with the unpaired t-test. Additionally, differences among the three glycemic groups were evaluated using analysis of variance or the Jonckheere-Terpstra test, whereas a post-hoc analysis of the two groups was performed using the Hochberg test or Dunnett T3 test 14 ; data were expressed as mean ± standard deviation. For categorical variables, intergroup differences were analyzed using chi-squared test or Fisher's exact test, as appropriate. Data were expressed as numbers and percentages 14 . The Kaplan-Meier method was used to estimate various clinical outcomes, and the log-rank test was used to compare intergroup differences ( Fig. 1 and Supplementary information). Variables with a p value of < 0.001 14 or < 0.05 42 in the univariate analysis and conventional risk factors of poor outcomes in the AMI population were considered potential confounding factors and were entered into the multivariate analysis. These included variables shown in Tables 2 and 3. For all analyses, two-sided values of p < 0.05 were considered statistically significant. All statistical analyses were performed using the SPSS software version 20 (IBM; Armonk, NY, USA) 14 .