Atherothrombotic risk stratification after acute myocardial infarction: The Thrombolysis in Myocardial Infarction Risk Score for Secondary Prevention in the light of the French Registry of Acute ST Elevation or non‐ST Elevation Myocardial Infarction registries

Background Guidelines recommend using risk stratification tools in acute myocardial infarction (AMI) to assist decision‐making. The Thrombolysis in Myocardial Infarction Risk Score for Secondary Prevention (TRS‐2P) has been recently developed to characterize long‐term risk in patients with MI. Hypothesis We aimed to assess the TRS‐2P in the French Registry of Acute ST Elevation or non‐ST elevation MI registries. Methods We used data from three 1‐month French registries, conducted 5 years apart, from 2005 to 2015, including 13 130 patients with AMI (52% ST‐elevation myocardial infarction [STEMI]). Atherothrombotic risk stratification was performed using the TRS‐2P score. Patients were divided in to three categories: G1 (low‐risk, TRS‐2P = 0/1); G2 (intermediate‐risk, TRS‐2P = 2); and G3 (high‐risk, TRS‐2P ≥ 3). Baseline characteristics and outcomes were analyzed according to TRS‐2P categories. Results A total of 12 715 patients (in whom TRS‐2P was available) were included. Prevalence of G1, G2, and G3 was 43%, 24%, and 33% respectively. Clinical characteristics and management significantly differed according to TRS‐2P categories. TRS‐2P successfully defined residual risk of death at 1 year (C‐statistic 0.78): 1‐year survival was 98% in G1, 94% in G2, and 78.5% in G3 (P < 0.001). Using Cox multivariate analysis, G3 was independently associated with higher risk of death at 1 year (hazard ratio [HR] 4.61; 95% confidence interval [CI]: 3.61‐5.89), as G2 (HR 2.08; 95% CI: 1.62‐2.65) compared with G1. The score appeared robust and correlated well with mortality in STEMI and NSTEMI populations, as well as in each cohort separately. Conclusions The TRS‐2P appears to be a robust risk score, identifying patients at high risk after AMI irrespective of the type of MI and historical period.


Conclusions:
The TRS-2P appears to be a robust risk score, identifying patients at high risk after AMI irrespective of the type of MI and historical period.

| INTRODUCTION
Risk stratification tools enable personalized risk assessment and may help guide therapeutic decision-making. Guidelines recommend their use in acute myocardial infarction (AMI) to identify high-risk patients and to assist with short-term prognostication and therapeutic decision-making (eg, early invasive strategy). [1][2][3][4][5][6][7][8] Several scores have been developed, especially in patients at the acute stage of MI; however, they remain underutilized in clinical practice in part they require specific tools as well as a perception that the impact on treatment decisions is limited, or both. The Thrombolysis in Myocardial Infarction (TIMI) Risk Score for Secondary Prevention (TRS-2P) is a simple nine-point risk stratification tool, derived in patients with previous MI to predict recurrent cardiovascular (CV) events. [9][10][11] This score has the advantage of being very simple to use and may assist with decisions on long-term response to treatment. Recently, the TRS-2P was validated in a clinical trial of acute coronary syndrome (ACS) patients followed for~7 years. 12 To our knowledge, the TRS-2P score has never been evaluated in a routine-practice population, focusing on patients who are discharged after an AMI. The aim of the present study was to test its robustness in several historical cohorts of patients after AMI, using the French Registry of Acute ST Elevation or non-ST elevation Myocardial Infarction (FAST-MI) registries.  15 (Supporting Information Methods S1). The methods used for these registries have been detailed previously. [13][14][15] Briefly, their primary objectives were to evaluate the characteristics, management, and outcomes of AMI patients, as seen in routine clinical practice, on a country-wide scale.

| Data collection
Data on baseline characteristics, including demographics (age, sex, body mass index), risk factors (hypertension, diabetes, current smoking, hypercholesterolemia, family history of coronary artery disease), and medical history (MI, previous myocardial revascularization, stroke, heart failure, peripheral artery disease [PAD], chronic renal failure), were collected as previously described. [13][14][15] Information on the use of cardiac procedures, including use of percutaneous coronary intervention (PCI), use of medications (anticoagulants, antiplatelet agents, diuretics, beta-blockers, angiotensin-converting enzyme inhibitors (ACE-I) or angiotensin receptor blockers (ARB), and lipid lowering agents) in the first 48 hours and at-hospital discharge was collected.
Bleeding was classified as major or minor according to the TIMI criteria. 16 Regarding bleeding complications, four end points of interest were used: in-hospital major bleeding (defined as a fall in hemoglobin ≥5 g, fall in hematocrit ≥15%, intracranial hemorrhage, retroperitoneal bleeding), minor bleeding (defined as a fall in hemoglobin between 3 and 5 g/dL, fall in hematocrit between 10% and 15%), use of any transfusion during the hospital stay, and 1-year survival.
For all surveys, follow-up was centralized at the French Society of Cardiology.

| Statistical analysis
Each patient was assessed for the presence of any of the nine previ- Multivariable analyses of correlates of 1-year mortality were performed using Cox backward stepwise multiple logistic regression, using a threshold of 0.10 for variable elimination. Beside time period, variables included in the final models were selected ad hoc, based on their physiological relevance and potential to be associated with outcomes; they comprised age, gender, risk factors, comorbidities, type of MI, TRS-2P categories, year, and management. Sensitivity analyses were performed focused on patients discharged alive in the main analysis, inpatients with STEMI or NSTEMI separately, and in each of the three historical cohorts. Analyses were repeated using forward stepwise analysis to check the consistency of the results. Statistical analyses were performed using IBM SPSS 23.0 (IBM SPSS Inc., Chicago, IL). For all analyses, two-sided P values <0.05 were considered significant.

| Study population
A total of 12 715 patients (97%) had all nine variables included in the TRS-2P score available at discharge and were included in the main analysis. Prevalence of Groups 1, 2, and 3 was 43%, 24%, and 33%, respectively. Over the 10-year period, the overall risk of patients admitted for AMI decreased, with the proportion in Group 3 declining from 43% to 29% (P < 0.001; Figure S1). The distribution of the nine variables according to the TRS-2P categories is presented in Figure S2. TRS-2P successfully defined patients with high-, intermediate-, and low-CV risk profile (Table 1). GRACE score was 168 AE 36 in Group 3, 139 AE 31 in Group 2, and 127 AE 27 in Group 1 (P < 0.001); simple risk index (SRI) was 35 AE 17, 26 AE 13, and 20 AE 10 in Groups 3, 2, and 1, respectively. In addition, the risk for major bleeding defined by the CRUSADE score decreased from Group 3 to Group 1.
The rate of STEMI patients was higher in Group 1, while the rate of patients with heart failure at admission (Killip class ≥ 2) was higher in Group 3. Interestingly, biomarkers of inflammation (eg, C-reactive protein) increased from Group 1 to Group 3.

| Early management
Early management including medications and myocardial revascularization were significantly different according to TRS-2P categories (Table 2). Overall, Group 3 patients received less antiplatelet agents, statin, beta-blocker, ACE-I, or ARB during the first 48 hours after admission as at discharge compared with both Groups 1 and 2 (P < 0.001 for all). In Group 3 patients, the use of appropriate secondary prevention treatment (dual antiplatelet therapy and statins for all; ACE-I/ARB and beta-blockers as indicated) was lower especially in patients with renal dysfunction (42% vs 55%, P < 0.001) and older patients (<60 years: 60%; 60-74 years: 53%; ≥75 years: 45%; P = 0.001). In addition, the use of invasive strategy (coronary angiography with or without PCI) was lower in Group 3, in which the rate of multivessel disease was higher. Radial access was preferentially used in low-or intermediate-risk patients. Finally, a full myocardial revascularization strategy during hospitalization was more frequently used in both Groups 1 and 2 (P < 0.001).

| Outcomes
In-hospital complications are described in (Table 3). The rate of re-MI, atrial fibrillation, stroke, and major and minor bleedings were higher in Group 3 patients. Mortality at 30 days was 9% in Group 3, 3% in Group 2, and 1% in Group 1 (P < 0.001).

| Subgroup analyses
Similar trends were found using TRS-2P score according to type of MI and year of survey (Tables S1-S6; Figure S2). The score appeared

| DISCUSSION
The main findings of this study are that the easily calculated TRS-2P appears to be a robust risk score, identifying patients at high-risk after AMI, irrespective of the type of MI and historical period. In addition, we showed that the rate of high-risk patients among those hospitalized for AMI decreased over the 10-year period from 2005 to 2015.

| Change in risk-profile
Several sources including registries specific to AMI and large databases, have shown a decrease in mortality over the past 20 years. [17][18][19][20][21] Using the FAST MI program, we previously reported that this decrease was correlated partially with a substantial change in patient risk profile, and not only with changes in management. Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blockers; CABG, coronary artery bypass grafting; CAG, coronary angiography; UFH, unfractionned heparin; LMWH, low-molecular-weight heparin; NSTEMI, non-ST-elevation myocardial infarction; PCI, percutaneous coronary intervention; STEMI, ST-elevation myocardial infarction; TIMI, thrombolysis in myocardial infarction; VD, vessel disease. Values are expressed as mean (AE SD) or number (percentage).   The TRS-2P score was, however, defined in a selected population in which for example, women and minorities made up a small proportion of the study population. [9][10][11] Recently, the TRS-2P score has been validated in routine practice using two large, independent integrated healthcare delivery systems in United States between 2008 and 2013 (ie, Cleveland Clinic and Geisinger Health System). 22 However, to our knowledge, this stratification tool has never been evaluated in a routine-practice population, focusing on patients who are discharged alive after an AMI. In our analysis, the TRS-2P score appears to be a robust risk score to

| Risk assessment and therapeutic intensification
Atherothrombotic risk assessment may be useful to identify high-risk patients who have the greatest potential to benefit from more intensive secondary prevention therapy such as antithrombotic or lipidlowering. In the TRA2P-TIMI 50 trial, the risk stratification tool identified a gradient of risk for recurrent events and distinguished a pattern of increasing absolute benefit with vorapaxar. [9][10][11] Similarly, using data from the IMPROVE IT trial, the TRS-2P score identified a strong gradient of risk for recurrent CV events; and an increasingly favorable relative and absolute benefit from the addition of ezetimibe to simvastatin therapy with increasing risk-profile. 12 Finally, this score could be evaluated to identify high-risk patients for new strategies as PCSK9-inhibitors or prolonged double antiplatelet therapy in ACS patients. 23,24 Yet, in clinical practice, the highest risk patients are paradoxically often the least intensively treated.

| Limitations
The TRS-2P score was designed to be a simple tool, using readily available clinical data. There are other previously identified risk indicators and other yet to be identified parameters that may provide additional refinement for stratification. However, the ability of this simple scoring system to identify differential treatment benefit for different classes of secondary prevent therapy supports its clinically utility. Our data are derived from an observational study of AMI patients admitted in ICUs while TRS-2P was defined in a population of stable patients with previous MI. In addition, our analyses were focused on the mortality at one-year while this risk stratification tool was developed for all CV-events at 3-year. The rate of CV death was not available. Finally, we cannot exclude that other factor than those collected in the surveys could also explain the evolution observed according to TRS2P categories.

| CONCLUSION
Atherothrombotic risk assessment may be useful to identify high-risk patients who have the greatest potential to benefit from more intensive secondary preventive therapy. Using a routine-practice population, TRS-2P appears to be a robust risk score, identifying patients at high-risk after AMI irrespective of the type of MI and historical period.