Outcomes of early versus delayed invasive strategy in older adults with non-ST-segment elevation myocardial infarction

We evaluated the 3-year clinical outcomes following early invasive (EI) and delayed invasive (DI) strategies in older adults with non-ST-segment elevation myocardial infarction (NSTEMI) undergoing successful new-generation drug-eluting stents (DESs) implantation to reflect current real-world practice. Overall, 2437 older adults (age, ≥ 65 years) with NSTEMI were recruited from the Korea Acute Myocardial Infarction Registry-National Institute of Health. They were divided into two groups: EI (n = 1750) and DI (n = 687). The primary clinical outcome was the occurrence of major adverse cardiac and cerebrovascular events (MACCEs), defined by all-cause death, recurrent MI, any repeat coronary revascularization, and stroke. The secondary clinical outcome was stent thrombosis (ST). After multivariable-adjusted and propensity score-matched analyses, the primary and secondary clinical outcomes were not significantly different between the EI and DI groups. Even after the analysis was confined to those having complex lesions, these major clinical outcomes were similar between these two groups. The EI and DI strategies in older adults with NSTEMI receiving new-generation DES showed comparable results. Clinical Trial Registration: URL: http://cris.nih.go.kr/cris/en/; Unique identifier: KCT0000863.


Results
Baseline characteristics. Figure 1 shows the flow chart of this study. Table 1 shows the baseline, laboratory, angiographic, and procedural characteristics of the study population. The mean values of left ventricular ejection fraction (LVEF), peak creatine kinase myocardial band (CK-MB), and peak troponin-I, and the number of current smokers, and the prescription rates of ticagrelor, angiotensin converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) as discharge medications, multivessel disease and patients with pre-PCI thrombolysis in myocardial infarction (TIMI) flow grade 0/1 were higher in the EI group than in DI. In contrast, the patients who had Killip class ≥ 3, had reduced renal function (estimated glomerular filtration rate [eGFR], < 60 mL/min/1.73 m 2 ), and received clopidogrel as discharge medication; mean value of serum creatinine and mean number of deployed stents; the use of intravascular ultrasound/optical coherent tomography/ fractional flow rate were higher in the DI group than in EI (Table 1).
Clinical outcomes. The in-hospital mortality and 3-year major clinical outcomes are summarized in Table 2 and Fig. 2. In-hospital all-cause death (hazard ratio [HR] 1.581 (95% confidence interval [CI] 0.861-2.904; p = 0.140), cardiac death (CD, HR 1.924; 95% CI 0.899-4.117; p = 0.092) and non-CD (HR 1.031; 95% CI 0.368-2.892; p = 0.954) were not significantly different between the EI and DI groups. After multivariableadjusted analysis, the 3-year major adverse cardiac and cerebrovascular events (MACCE, adjusted HR [aHR] 1.159; 95% CI 0.960-1.398; p = 0.125), all-cause death (aHR 1.180; p = 0.192), CD (aHR 1.229; p = 0.228), non-CD (aHR 1.116; p = 0.564), recurrent MI (re-MI, aHR 1.040; p = 0.881), any repeat revascularization (aHR, 1.171; p = 0.327), stroke (aHR 1.099; p = 0.713), and stent thrombosis (ST [definite or probable], aHR 2.058; 95% CI 0.690-6.143; p = 0.196) rates were not significantly different between the EI and DI groups. (Table 2). These results were confirmed after PS-matched analysis. After PS-matched analysis, the primary and secondary clinical outcomes were not significantly different between the EI and DI groups ( Table 2). For further assessment of major clinical outcomes between the EI and DI groups, we compared these major clinical outcomes by limiting the study population to patients with complex lesions ( Table 3). The number of patients with complex lesions in each group was more than 50% (EI vs. DI = 51.3% vs. 56.2%, p = 0.028) (Fig. 3). The MACCE rates were similar between the EI and DI groups (aHR 1.034; 95% CI 0.810-1.320; p = 0.787) ( Table 3). The ST (definite or probable) rates were also similar between the EI and DI groups (aHR 2.662; 95% CI 0.531-13.35; p = 0.234). Additionally, the all-cause death, CD, non-CD, re-MI, any repeat revascularization, and stroke rates were not A total of 4,741 patients with NSTEMI who underwent successful new-generation DES implantation www.nature.com/scientificreports/ significantly different between the two groups after adjustment (Table 3). Figure 4 shows the subgroup analysis for MACCE. The results of subgroup analysis using Cox logistic regression model revealed that in the all subgroups except for those showing significant p-for-interaction demonstrated comparable MACCE rates in this study. Table 4 shows predictors for all-cause mortality in the total study population, which includes reduced LVEF (< 50%, aHR 1.

Discussion
The main findings of this prospective, observational study were: (1) after multivariable-adjusted and PS-matched analyses, MACCE, all-cause death, CD, non-CD, re-MI, any repeat revascularization, stroke, and ST (definite or probable) rates were similar between the EI and DI groups; (2) even after limiting the study population to patients who had complex lesions, the primary and secondary clinical outcomes were not significantly different between the EI and DI groups. Theoretically, through the EI strategy, the operator could find significant lesions earlier in patients with NSTEMI and could have the opportunity for early revascularization, salvage of ischemic myocardium, and facilitation of earlier discharge from a facility 2,10 . In contrast, DI strategy may provide adequate time for optimal medical treatment in order to decrease thrombus burden and improve plaque stability 10 . In the recent European guideline, the recommended diagnostic and interventional strategies for older patients and younger patients are the same (class I and level of evidence B) 1 . However, the optimal timing of PCI in NSTEMI remains a subject of debate. The clinical presentation of NSTE-ACS in older person is atypical 11,12 and the electrocardiographic changes are less frequent in older than in younger patients 7,12 . Despite the significant decrease in mortality and morbidities of ACS because of evidence based therapy 13 , these improvements in ACS treatment strategy have not equally improved outcomes for older adults 2 . Regarding these characteristics 2,7,11,12 in older people, the information dealing with the preferred treatment option between the EI and DI strategies could be important for the interventional cardiologist. In the old report, EI strategy showed significantly improved clinical outcomes compared with conservative treatment in elderly patients with NSTE-ACS 14 . However, these studies were not performed in the era of new-generation DES and that did not compare clinical outcomes between the EI and DI strategies 14,15 . Furthermore, since the available data on this subject is limited 8 , the comparative results between the EI and DI strategies in older patients with NSTEMI are limited. Hence, in this study, we investigated the long-term clinical outcomes between the EI and DI strategies in older adults with NSTEMI undergoing successful new-generation DES implantation. In our study, the major clinical outcomes were not significantly different between the EI and DI groups after adjustments (multivariable or PS-matched) during a 3-year follow-up period. Table 1. Baseline clinical, laboratory, angiographic and procedural characteristics. Values are means ± standard deviation or median (interquartile range) or numbers and percentages. The p values for continuous data were obtained from the unpaired t-test. The p values for categorical data from chi-square or Fisher's exact test. LVEF left ventricular ejection fraction, BMI body mass index, SBP systolic blood pressure, DBP, diastolic blood pressure, MI myocardial infarction, PCI percutaneous coronary intervention, CABG coronary artery bypass graft, HF heart failure, CK-MB creatine kinase myocardial band, Hs-CRP high sensitivity C-reactive protein, eGFR estimated glomerular filtration rate, HDL high-density lipoprotein, LDL low-density lipoprotein, GRACE Global Registry of Acute Coronary Events, BBs ß-blockers, ACEIs angiotensin-converting enzyme inhibitors, ARBs angiotensin receptor blockers, LAD left anterior descending artery, LCx left circumflex artery, RCA right coronary artery, ACC/AHA American College of Cardiology/ American Heart Association, TIMI thrombolysis in myocardial infarction, GP glycoprotein, IVUS intravascular ultrasound, OCT optical coherence tomography, FFR fractional flow reserve, IABP, intra-aortic balloon pump, ECMO, extracorporeal membrane oxygenation, ZES zotarolimus-eluting stent, EES everolimuseluting stent, BES biolimus-eluting stent.

Variables
All patients (n = 2437) SD Propensity score-matched patients (n = 1314)  www.nature.com/scientificreports/ been attributed to several age-related changes including vascular wall elasticity, coagulation and hemostatic system, and endothelial dysfunction [17][18][19] . Therefore, age related decline in organ function increases cardiovascular diseases 19 . Frailty is very common in older adults with cardiovascular diseases and frailty contributes valuable prognostic insights incremental to existing risk models and assists clinicians in defining optimal care pathways for their patients 20 . In elderly NSTEMI patients, frailty was independently associated with all-cause mortality at long-term   22 , increased frailty was independently associated with increased post-discharge all-cause mortality. More recent study showed that an assessment of both cognitive and physical conditions should be included in the comprehensive geriatric evaluation of hospitalized older STEMI patients 23 . Hence, Faubert et al. 24 emphasized that the management of NSTEMI in elderly patients must be individualized with regard to the patient's goals, comorbid conditions, overall health, and cognitive status. Mone et al. 25 showed the importance of thrombus aspiration in the treatment of STE-myocardial infarction (STEMI) in a group of high-risk patients such as elderly with frailty. Even though the primary and secondary clinical outcomes were not significantly different between the EI and DI groups, after adjustment, reduced LVEF, cardiogenic shock, IABP or ECMO, reduced renal function, and a high GRACE risk score were significant predictors for all-cause mortality in this study (Table 4). Hayıroğlu et al. 26 showed the mortality rate remains high despite IABP support in patients with ACS. Çinar et al. 27 reported that the incidence of in-hospital mortality was significantly greater in patients with a high age, creatinine, ejection fraction score compared with the intermediate or the low score group (p < 0.005) among patients with STEMI related cardiogenic shock.  www.nature.com/scientificreports/ To clearly estimate the long-term clinical outcomes, we performed additional analysis as shown in Table 3. Even after considering the patients with complex lesions, the 3-year major clinical outcomes were not significantly different (Table 3). Subgroup analyses for MACCE in group A and B (Fig. 4) showed that all subgroups except for those showing significant p-for-interaction had comparable MACCE rates.

SD Early invasive (n = 1750) Delayed invasive (n = 687) p value Early invasive (n = 657) Delayed invasive (n = 657) p value
We agree with the current guideline recommendations that suggest that the management of older patients should be based on ischemic and bleeding risks, estimated life expectancy, comorbidities, the need for noncardiac surgery, quality of life, frailty, cognitive, functional impairment, patient values and preferences, and the estimated risks and benefits of revascularization 1 . Our results showed that in the era of new-generation DES, the major clinical outcomes were not significantly different between the EI and DI strategies in older adults with NSTEMI after successful stent implantation during a 3-year follow-up period. Hence, we suggested that the current guideline 1,2 about the management of older patients with NATE-ACS with CAG and PCI needs to be reevaluated under the era of new-generation DES. In this study, although the population may have been insufficient to provide meaningful results, 20 tertiary high-volume University hospitals participated in the registry. Therefore, we believe that our results could provide helpful information to interventional cardiologists in terms of long-term effects of EI and DI strategies in older adults with NSTEMI undergoing successful implantation of new-generation DES.
This study had other limitations. First, even though this study is a prospective, observational registry, it is not a randomized controlled study; there may have been some selection bias. Moreover, the variables that were not included in the data registry might have affected the study outcome despite the multivariable and PS-matched analyses. Second, because we set the cut-off value of older adults at age ≥ 65 years in our study, our results could change according to different cut-off ages. Third, as mentioned, although bleeding is an important complication that occurs after PCI in older adults 14,15 , anti-platelet therapy after 1 year index PCI was different among the physicians; we could not include bleeding complication as an outcome parameter in our study during a 3-year follow-up period. This is a major shortcoming of our study. Fourth, the 3-year follow-up duration was insufficient to evaluate long-term adverse events. Finally, contrast induced nephropathy is an important factor and acute kidney injury can effect long-term outcomes 28 . A recent report demonstrated that acute kidney injury was an important independent prognostic factor (HR 2.244; 95% CI 1.077-4.676; p = 0.031) for 5-year mortality among patients with STEMI complicated by cardiogenic shock and treated with primary PCI 28 . However, because these variables (contrast induced nephropathy and acute kidney injury) were not included in the data registry, which could have caused significant bias.
In conclusion, in the era of new-generation DES, the major clinical outcomes were not significantly different between the EI and DI strategies in older adults with NSTEMI after successful stent implantation during a 3-year follow-up period. However, further randomized, large-scale, and long-term follow-up studies are needed to clarify the differences of the clinical outcomes between these two different reperfusion strategies in those patients.  www.nature.com/scientificreports/ implantation were included (Fig. 1). The types of new-generation DES used are listed in Table 1. The definition of older adults is controversial. In general, a person is considered old if their civil age is ≥ 60 or 65 years 30 31 and other report 32 showed that multimorbidity and polypharmacy are highly prevalent among adults aged ≥ 65 years, we set the cut-off value at ≥ 65 years for older adults in our study. These patients were divided into two groups: EI (n = 1750, 71.8%) and DI (n = 687, 28.2%) (Fig. 1 -2011-172). All patients included in the study provided written informed consent prior to enrollment. They were followed-up via face-to-face interviews, phone calls, or chart reviews and they completed a 3-year follow-up schedule. All clinical events were evaluated by an independent event adjudication committee. The event adjudication process has previously been described by the KAMIR investigators 29 .

Methods
PCI procedure and medical treatment. CAG and PCI were performed via a transfemoral or transradial approach in accordance with the general guidelines 33 . Aspirin (200-300 mg) and clopidogrel (300-600 mg), ticagrelor (180 mg), or prasugrel (60 mg) were prescribed to the patients as loading doses before PCI. After PCI, all patients were recommended to take aspirin (100 mg/day) along with clopidogrel (75 mg/day), ticagrelor (90 mg twice a day), or prasugrel (5-10 mg/day) for at least 1 year. The access site, revascularization strategy, and selection of DES were left to the discretion of the individual operators.
Study definitions and clinical outcomes. NSTEMI was defined as the absence of persistent STE with increased levels of cardiac biomarkers and appropriate clinical context 1,2 . A successful PCI was defined as residual stenosis of < 30% and thrombolysis in MI (TIMI) flow grade 3 in the infarct-related artery. Glomerular function for estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation 34 . The GRACE risk score 35 was calculated for all the patients. Complex lesions were defined as PCI for unprotected left main coronary disease, multivessel PCI, multiple stents implantation (≥ 3 stents per patient), and those with the total length of deployed stent being over 38 mm 36,37 . The primary clinical outcome was the occurrence of major adverse cardiac and cerebrovascular events (MACCE), which was defined by all-cause death, recurrent MI (re-MI), any repeat coronary revascularization, including target lesion revascularization, target vessel revascularization (TVR), non-TVR, and stroke. According the American Heart Association/American Stroke Association guideline 38 , an acute cerebrovascular event resulting in death or neurological deficit for > 24 h or the presence of acute infarction demonstrated by imaging studies was defined as a stroke. An all-cause death was considered a cardiac death (CD) unless an undisputed non-cardiac cause was present 39  Statistical analysis. For continuous variables, the differences between the groups were evaluated using unpaired t-tests. Data are expressed as the mean ± standard deviation, or median (interquartile range). For discrete variables, the differences between the groups were expressed as counts and percentages and were analyzed using the chi-squared or Fisher's exact test. Univariate analysis was performed for all variables of EI and DI groups with the p-value set at < 0.05. Subsequently, we performed a multicollinearity test 42 between the included variables to confirm non-collinearity between them (Supplementary Table S1). Variance inflation factor (VIF) values were calculated to measure the degree of multicollinearity among the variables. A VIF of > 5 indicated a high correlation 43 . When the tolerance value was < 0.1 44 or the condition index was > 10 43 , the presence of multicollinearity was considered. The variables included in the multivariable Cox regression analysis were: male sex, age, LVEF, body mass index, systolic blood pressure, diastolic blood pressure, cardiogenic shock, symptom-todoor time, hypertension, diabetes mellitus, dyslipidemia, previous MI, previous PCI, current smoker, CK-MB, peak troponin-I, serum creatinine, eGFR < 60 mL/min/1.73 m 2 , high-density lipoprotein cholesterol, and GRACE risk score > 140. Moreover, to adjust for potential confounders, propensity score (PS)-matched analysis was performed using a logistic regression model. We tested all potentially relevant variables such as baseline clinical, angiographic, and procedural factors ( Table 1). The c-statistic for the PS-matched (PSM) analysis in this study was 0.724. Patients in the EI group were matched to those in the DI group (1:1) using the nearest available pair-matching method according to PSs. The subjects were matched with a caliper width of 0.01. This procedure yielded 1314 well-matched pairs (Table 1). Various clinical outcomes were estimated using a Kaplan-Meier curve analysis, and group differences were compared using the log-rank test. Statistical significance was defined as a 2-tailed p-value of < 0.05. All statistical analyses were performed using SPSS software v. 20 (IBM; Armonk, NY, USA).

Data availability
Data is contained with the article or supplementary material.