Author + information
- Received January 9, 2012
- Revision received May 21, 2012
- Accepted May 27, 2012
- Published online September 1, 2012.
- Etienne Puymirat, MD⁎,⁎ (, )
- Guillaume Taldir, MD⁎,
- Nadia Aissaoui, MD⁎,
- Gilles Lemesle, MD†,
- Luc Lorgis, MD‡,
- Thomas Cuisset, MD§,
- Pierre Bourlard, MD∥,
- Bruno Maillier, MD¶,
- Gregory Ducrocq, MD#,
- Jean Ferrieres, MD, PhD⁎⁎,
- Tabassome Simon, MD, PhD†† and
- Nicolas Danchin, MD, PhD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Etienne Puymirat, Division of Coronary Artery Disease, Hôpital Européen Georges Pompidou, 75015 Paris, France
Objectives This study sought to assess the impact of invasive strategy (IS) versus a conservative strategy (CS) on in-hospital complications and 3-year outcomes in patients with non–ST-segment elevation myocardial infarction (NSTEMI) from the FAST-MI (French Registry of Acute Coronary Syndrome).
Background Results from randomized trials comparing IS and CS in patients with NSTEMI are conflicting.
Methods Of the 3,670 patients in FAST-MI, which included patients with acute myocardial infarction (within 48 h) over a 1-month period in France at the end of 2005, 1,645 presented with NSTEMI.
Results Of the 1,645 patients analyzed, 80% had an IS. Patients in the IS group were younger (67 ± 12 years vs. 80 ± 11 years), less often women (29% vs. 51%), and had a lower GRACE (Global Registry of Acute Coronary Events) risk score (137 ± 36 vs. 178 ± 34) than patients treated with CS. In-hospital mortality and blood transfusions were significantly more frequent in patients with CS versus IS (13.1% vs. 2.0%, 9.1% vs. 4.6%). Use of IS was associated with a significant reduction in 3-year mortality and cardiovascular death (17% vs. 60%, adjusted hazard ratio [HR]: 0.44, 95% confidence interval [CI]: 0.35 to 0.55 and 8% vs. 36%, adjusted HR: 0.37, 95% CI: 0.27 to 0.50). After propensity score matching (181 patients per group), 3-year survival was significantly higher in patients treated with IS.
Conclusions In a real-world setting of patients admitted with NSTEMI, the use of IS during the initial hospital stay is an independent predictor of improved 3-year survival, regardless of age. (French Registry of Acute Coronary Syndrome [FAST-MI]; NCT00673036)
- acute myocardial infarction
- invasive strategy
- non–ST-segment elevation myocardial infarction
- percutaneous coronary intervention
Patients presenting with non–ST-segment elevation acute coronary syndromes (NSTE-ACS) constitute a very heterogeneous population. Using more sensitive markers, more patients are now characterized with non–ST-segment elevation myocardial infarction (NSTEMI), which makes this condition a public health problem. Multiple trials have evaluated different clinical strategies especially regarding the impact of systematic coronary angiogram (CAG) and subsequent revascularization of clinically stabilized patients with NSTE-ACS (1–8). Two general approaches have emerged. The first named “early invasive” or “routine invasive” strategy, involving routine early CAG followed by revascularization when appropriate. The second is the “conservative” or “selective invasive” approach, with initial pharmacological management and CAG followed by revascularization for recurrent or stress-induced ischemia only (9–11).
Four relatively recent large randomized trials, the FRISC (Fragmin and Fast Revascularization During Instability in Coronary Artery Disease) II trial, the TACTICS–TIMI (Treat Angina With Aggrastat and Determine Cost of Therapy With an Invasive or Conservative Strategy–Thrombolysis In Myocardial Infarction) 18 trial, the RITA (Randomized Intervention Trial of Unstable Angina) 3 trial, and the ICTUS (Invasive Versus Conservative Treatment in Unstable Coronary Syndromes) trial have compared both strategies (4–7). Overall, these trials showed a significant reduction in the combined endpoint of death and myocardial infarction (MI) with the invasive approach, especially in high-risk patients. Therefore, the American Heart Association/American College of Cardiology and European Society of Cardiology guidelines support routine invasive management in intermediate to high-risk patients with NSTE-ACS (10,11).
The objective of this analysis from the FAST-MI (French Registry of Acute Coronary Syndrome) (12) is to assess the impact of the invasive strategy (IS) compared with the conservative strategy (CS) on 3-year outcomes in patients presenting with NSTE-ACS in real-world clinical practice.
We selected patients with NSTEMI from FAST-MI, the methodology of which has been previously described in detail (12). Briefly, the primary objective was to evaluate practices for MI management in “real life” and to measure their impact on the medium- and long-term prognosis in patients admitted to the intensive care unit with acute MI (within 48 h). This registry results from a prospective multicenter (223 centers) study, including 3,670 patients. Patients were recruited consecutively from intensive care units over a period of 1 month (from October 2005), with an additional month for diabetic patients. Participation in the study was offered to all French institutions, university teaching hospitals, general and regional hospitals, and private clinics with intensive care units in the capacity to receive ACS emergencies.
We included men or women aged over 18 years, who were admitted within 48 h after symptom onset for an acute MI characterized by the elevation of troponin or creatine phosphokinase-myocardial band associated with at least one of the following elements—symptoms compatible with myocardial ischemia, new pathological Q waves, ST-T changes compatible with myocardial ischemia—and who agreed to take part in the study.
The main exclusion criteria were: 1) iatrogenic MI, defined as MI occurring within 48 h of a therapeutic procedure (bypass surgery, coronary angioplasty, or any other medical or surgical intervention); 2) ACS diagnosis invalidated in favor of another diagnosis; and 3) patients with unstable angina and no increase in cardiac biomarkers.
Participating in the registry was not supposed to change the therapeutic approach of the cardiologist in any way. The registry was conducted in compliance with Good Clinical Practice guidelines, French law, and the French data protection law. The protocol was reviewed by the Committee for the Protection of Human Subjects in Biomedical Research of Saint-Antoine University Hospital and the data file of FAST-MI was declared to the Commission Nationale Informatique et Liberté.
Definition of NSTEMI and invasive strategy
NSTEMI was defined as the presence of cardiac markers with symptoms compatible with myocardial ischemia and no ST-segment elevation on the index or qualifying electrocardiogram; patients with left bundle branch block were considered to have ST-segment elevation myocardial infarction (STEMI), unless the block was known before the current episode. IS was defined as early CAG (i.e., during the initial hospital admission), whether it was followed (or not) by revascularization. Patients with CS did not undergo coronary angiography and received only medical therapy.
Data collection and follow-up
Baseline characteristics were collected prospectively. All data were recorded on computerized case record forms by dedicated research technicians sent in each of the centers at least once a week. The research technicians were also asked to ensure that recruitment was consecutive. Follow-up data were collected through contacts with the attending physicians, the patients, or their family. If missing, vital status was assessed from the registries of the patients' birthplaces. Follow-up at 1 year was 99% complete and at 3 years 97% complete. Bleeding was classified as major or minor according to the TIMI (Thrombolysis In Myocardial Infarction) criteria (13). Use of any transfusion during the hospital stay was also recorded. Major adverse cardiac events (MACE) were defined as death, nonfatal MI, or stroke. For combined endpoints, only the first occurrence is considered.
Statistical analysis was performed using SPSS software (version 18.0, SPSS Inc., Chicago, Illinois). For quantitative variables, mean ± SD were calculated. In addition, medians with interquartile ranges were calculated when appropriate. Discrete variables are presented as percentages. Comparisons were made with chi-square or Fisher exact tests for discrete variables, and by unpaired Student t tests, Wilcoxon sign-rank tests, or 1-way analyses of variance for continuous variables. Survival curves were estimated using the Kaplan-Meier estimation and compared using log-rank tests. Multivariate analyses of predictors of in-hospital endpoints were made by using backward, stepwise multiple logistic regressions. Correlates of survival were determined using a multivariate backward stepwise Cox analysis. Variables listed in Table 1 were included in the models. In addition, propensity scores for getting IS rather than CS were calculated using multiple logistic regressions and were used to build 2 cohorts of patients matched on the propensity scores: 1 score was calculated for the whole population (C-statistic: 0.913); and 1 additional score was calculated in the population of hospital survivors (C-statistic: 0.927). The variables used for the first model were those listed in Table 1, with the addition of type of institution (with or without catheterization laboratory); use of the mobile intensive care system; and use of antithrombotic medications, statins, beta-blockers, and ACE inhibitors during the first 48 h following admission. The same variables were used for the model in hospital survivors, with the addition of discharge medications and main in-hospital complications (reinfarction, stroke, major bleeding, or transfusion).
Comparisons between propensity score-matched cohorts used paired Student t tests for continuous variables and McNemar tests for discrete variables. For all analyses, a p value <0.05 was considered significant.
Of the 3,670 patients included in the registry, 1,645 (45%) presented with NSTEMI.
Baseline characteristics are described in Table 1. Eighty percent of patients had an IS, and 20% were managed conservatively. Both groups differed in many respects. In particular, patients with CS were significantly older, had higher GRACE (Global Registry of Acute Coronary Events) scores, lower left ventricular ejection fractions, and higher creatinine levels.
During the first 48 h, patients treated invasively received significantly more antiplatelet agents, beta-blockers, and statins; whereas patients treated conservatively received more diuretics and antiarrhythmic drugs (Table 2).
In patients treated invasively, the average time from admission to CAG was 2.35 ± 2.96 days. CAG was performed ≤24 h in 38%, between 48 and 72 h in 37%, and after 72 h in 25%. Overall, CAG performed during the first 24 h was used mainly in younger patients, men, and those with higher cardiovascular risk. Hospitalization stay was considerably shorter in the IS group (CAG ≤24 h) (7.8 ± 7.8 days vs. 10.4 ± 8.7 days, p < 0.001). Finally, 65% of the patients with IS had a percutaneous coronary intervention (PCI), and 71% had either a PCI or coronary artery bypass graft.
In-hospital complications and 3-year clinical outcomes
In-hospital complications are described in Table 3. In-hospital mortality and blood transfusions were significantly more frequent in patients with CS. Major and minor bleeding, however, were not significantly different between the 2 strategies.
At 3 years, rates of death, cardiovascular death, combined death or nonfatal MI, and combined MACE or revascularization (by PCI or coronary artery bypass graft) were significantly lower in patients with IS (Fig. 1). Using Cox multivariate analysis, the use of IS was associated with a reduced risk of death (hazard ratio [HR]: 0.44, 95% confidence interval [CI]: 0.35 to 0.55), cardiovascular death (HR: 0.37, 95% CI: 0.27 to 0.50), combined death or nonfatal MI (HR: 0.50, 95% CI: 0.40 to 0.61), MACE (HR: 0.54, 95% CI: 0.43 to 0.66), and MACE or revascularization (HR: 0.61, 95% CI: 0.50 to 0.75) (Fig. 1).
In the IS group, survival of patients who had CAG followed by PCI or coronary artery bypass graft was significantly higher than that of patients with CAG only, followed by no revascularization (86% vs. 77%; HR: 1.52, 95% CI: 1.15 to 2.01) (Fig. 2). The statistical significance, however, did not persist after multivariate adjustment.
In addition, to specifically assess the long-term impact of IS, we analyzed 3-year outcomes in patients discharged alive. We also included in-hospital complications and medications prescribed at discharge in the multivariate model. IS was associated with a significant reduction in 3-year mortality (HR: 0.56, 95% CI: 0.43 to 0.72), cardiovascular death (HR: 0.45, 95% CI: 0.31 to 0.65), MACE (HR: 0.68, 95% CI: 0.53 to 0.86), and MACE or revascularization (HR: 0.71, 95% CI: 0.58 to 0.89).
A subgroup analysis of 3-year mortality as a function of IS according to age, class, sex, GRACE score at admission, and the use of optimal medical therapy at hospital discharge is presented in Table 4. The IS was associated with lower 3-year mortality in all subgroups, regardless of age (including in patients aged over 75 years), sex, and whatever medical therapy was associated. Although IS was associated with improved survival in all GRACE risk score categories, a significant interaction was observed, with a greater relative risk reduction in patients with lower risk scores.
Propensity score–matched cohorts
Propensity score matching was performed to build 2 matched cohorts of 181 patients each that had similar baseline characteristics (Table 5). Three-year survival was significantly higher in the patients who had an IS (62%) compared with those who had not (43%) (HR: 0.54, 95% CI: 0.40 to 0.74; p < 0.001) (Online Fig. 1). Similar results were found with cohorts restricted to hospital survivors and paired on a propensity score, including medical history, admission characteristics, hospital management and complications, and discharge medications. The populations had similar baseline characteristics and received similar discharge medications, including clopidogrel (Online Table 1), and 3-year survival was 64% versus 51% (HR: 0.62, 95% CI: 0.44 to 0.89; p = 0.008) (Online Fig. 2).
The present data from a nationwide registry indicate that, in patients presenting with NSTEMI, the use of IS at the acute stage is associated with improved early and 3-year outcomes. The association was consistent across patient subgroups and persisted after multivariate adjustments and propensity score matching.
Many randomized controlled trials and meta-analyses have assessed the effects of a routine invasive versus conservative approach in patients with NSTE-ACS (1–8). Overall, the results of these studies and of their meta-analyses showed a significant reduction in the combined endpoint of death and MI with the IS. The difference observed was driven mainly by an excess in MI in the conservative arms of the trials and with no significant effect on mortality after adjustment. Four large randomized controlled trials have dominated the debate on the routine use of IS in NSTE-ACS with conflicting results. The FRISC II trial showed a significant reduction in the combined endpoint of death and MI with the routine invasive approach but no significant mortality benefit (4). The difference was driven mainly by an excess in MI in the selective invasive group. The TACTICS–TIMI 18 trial showed similar results: a decrease in MI but no significant benefit in terms of mortality (5). In 2003, the RITA 3 trial failed to show any significant benefit for death or MI (6). Ultimately, in 2005, the ICTUS trial, which was carried out in patients with NSTEMI and excluded unstable angina patients, showed an increased MI risk in the routine invasive arm, with no difference in mortality (7). Interpretation of these study results is difficult because of important differences in methodology, especially concerning the rate of revascularization procedures in the conservative arms, the definition of ACS used, the medications at hospital discharge, and the duration of clinical follow-up. A recent meta-analysis based on individual 5-year follow-up data from FRISC II, RITA 3, and ICTUS showed a reduction in the combined endpoint death or MI using the routine IS. The consistent trends for death, however, were not significant (8). Our data show that, as used in real-world clinical practice in patients with NSTEMI (i.e., excluding unstable angina patients), the routine IS is also associated with a significant reduction in the combined endpoint of death and MI with a difference driven mainly by an excess in death in the CS group.
The relation between treatment effect and patient risk has been evaluated in subanalyses of several trials. Regardless of the risk score used, there was a consistently higher treatment benefit for the invasive approach in high-risk patients than in low-risk patients. The FRISC II and TACTICS–TIMI 18 trials as well as the 5-year follow-up of RITA 3 showed the greatest benefit of the routine invasive approach in high-risk patients (4,5,14). This resulted in a wide acceptance of the routine invasive approach in this subpopulation. Our results, however, show that the impact of a routine IS may exist even in lower-risk NSTEMI patients and might even be of a greater relative magnitude in these patients.
Remarkably, recent data from the GRACE registry document the presence of an inverse relation between patient risk and the rate of PCI (8). In daily practice, angiographic findings and referral practice may more substantially influence the decision to proceed to PCI than patients' risk status does. A similar paradox, with patients at lower risk having a higher use of IS is present in our study, within a population having increased cardiac biomarkers at baseline.
Many clinicians consider that IS should be used in most patients, but, in real-world practice, many centers tend to use a routine IS in younger patients, and a more selective approach in older individuals, in whom the presence of comorbidities is often perceived as a limitation to perform CAG (15). Recently, Damman et al. (16) have analyzed the impact of the initial management strategy (routine invasive or selective invasive) on the 5-year risk of cardiovascular death or MI, according to the patients' age categories in the FIR (FRISC II–ICTUS–RITA 3 trials) meta-analysis. Although the main objective of the study was to confirm (or refute) the positive role of IS in elderly patients (an age category often underrepresented in randomized trials), the most striking finding was that a routine IS in younger patients was not associated with a reduction in the composite endpoint of cardiovascular death or MI, whereas such a strategy appeared beneficial in patients older than 65 years, including those over the age of 75 years. The unfavorable results of the IS in the younger patients were most probably drawn by the results of ICTUS (17), because the benefits of the IS were homogeneous across age groups in FRISC II (18), and no interaction with age was found in the RITA-3 trial (14). When the combined endpoint of cardiovascular death or MI was broken down into its 2 components, the routine IS, however, was associated with an increased risk of MI, but a trend to lower cardiovascular mortality in the younger age group. A similar mortality trend was noted in our population with the lowest GRACE risk score, granting the fact that only very few of these patients had CS.
The optimal timing of CAG and subsequent intervention if indicated, that is, immediately after admission or after pre-treatment with optimal medical therapy, remains debated. Randomized studies that have specifically addressed the issue in patients with NSTE-ACS have yielded inconclusive results (19–22). Recently, a meta-analysis of 4 major trials, including 4,013 patients indicated that early intervention can be safely adopted without increased risk and with significant benefits in terms of recurrent ischemia (risk ratio: 0.59, 95% CI: 0.38 to 0.92; p = 0.002) and duration of hospital stay (by 28%, 95% CI: 22% to 35%; p < 0.001). As in our study, the timing of CAG (in IS group) did not significantly affect the risk of death (23).
Our study suffers the same limitations as all observational studies: namely, no causality can be asserted between parameters that are correlated. Comparisons between patients receiving an invasive and those receiving a conservative strategy were not randomized and, despite careful adjustments on a large number of potentially confounding variables, the results can only be considered indicative. In particular, we did not specifically record the reason for not performing CAG, and reasons other than any information we collected may have contributed to the fact that patients were treated conservatively. The use of propensity score matching, however, limits the biases inherent to observational data, by giving the opportunity to compare outcomes in cohorts of patients with similar baseline characteristics. Some patients may have died before coronary angiography could be performed and, therefore, might have been “unjustly” classified in the CS group; to avoid such a bias, however, we repeated our analyses in the population of patients who survived the hospital stay and observed a similar association between IS and late outcomes. Patients with left bundle branch blocks were classified in the STEMI population and, therefore, were not included in the current analysis, unless the left bundle branch block was known to pre-exist the current episode of acute MI. Finally, a few “late arrival” patients may have had STEMI but their electrocardiograms had evolved by the time they were admitted, resulting in their classification in the NSTEMI group.
Our data show that, as used in real-world clinical practice, IS in patients suffering from NSTEMI is associated with fewer in-hospital complications and higher 3-year survival rates. These real-world findings, which were confirmed in propensity score–matched populations, suggest that the use of IS during the initial hospital stay is an independent predictor of improved 3-year survival in patients with NSTEMI.
The authors are deeply indebted to all physicians who have taken care of the patients at the participating institutions, as well as to Nadine Roumier, from the International Clinical Trials Association (ICTA) contract research organization (Fontaine-lès-Dijon, France), and to the devoted personnel of the URCEST (Assistance Publique des Hôpitaux de Paris and University Paris 6) under the leadership of Elodie Drouet and Institut National de la Santé et de la Recherche Médicale (INSERM) U 1027 (Toulouse). Special thanks to Vincent Bataille, for his careful data management; to Benoît Pace (Société Française de Cardiologie) for his invaluable assistance in designing the electronic case report form; and to Geneviève Mulak, PharmD, and the personnel of the registry committee of the Société Française de Cardiologie, and particularly Frédérique Fortin. The authors also wish to thank all patients who agreed to participate in the registry for their invaluable help during the follow-up period.
For more information about the propensity score–matched cohorts, please see the online version of this paper.
FAST-MI is a registry from the French Society of Cardiology that is supported by unrestricted grants from Pfizer and Servier and an additional grant from the Caisse Nationale d'Assurance-Maladie-Travailleurs Salariés. Dr. Ducrocq has a relationship with Medtronic, Inc., AstraZeneca Pharmaceuticals, and Eli Lilly. Dr. Simon has received research grants from Pfizer and Servier for the FAST-MI registry; served as a board member for Bayer; and has given lectures to Sanofi and Lilly. Dr. Danchin has received research grants from AstraZeneca, Daiichi-Sankyo, Eli Lilly, GlaxoSmithKline, MSD, Novartis, Sanofi-Aventis, Servier, and The Medicines Company; and has served on advisory panels or received lecture fees from AstraZeneca, Boeheringer-Ingelheim, Bristol-Myers Squibb, Eli-Lilly, Menarini, Merck-Serono, Novo-Nordisk, Roche, Servier, and Sanofi-Aventis. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute coronary syndromes
- coronary angiogram
- confidence interval
- conservative strategy
- hazard ratio
- invasive strategy
- major adverse cardiac event(s)
- myocardial infarction
- non– ST-segment elevation myocardial infarction
- non–ST-segment elevation acute coronary syndromes
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Thrombolysis In Myocardial Infarction
- Received January 9, 2012.
- Revision received May 21, 2012.
- Accepted May 27, 2012.
- American College of Cardiology Foundation
- Bavry A.A.,
- Kumbhani D.J.,
- Rassi A.N.,
- Bhatt D.L.,
- Askari A.T.
- Cannon C.P.,
- Weintraub W.S.,
- Demopoulos L.A.,
- et al.,
- TACTICS-TIMI 18 Investigators
- Fox K.A.,
- Poole-Wilson P.A.,
- Henderson R.A.,
- et al.,
- RITA 3 Investigators
- Fox K.A.,
- Clayton T.C.,
- Damman P.,
- et al.,
- FIR Collaboration
- Anderson J.L.,
- Adams C.D.,
- Antman E.M.,
- et al.
- Hamm C.W.,
- Bassand J.P.,
- Agewall S.,
- et al.
- Rao A.K.,
- Pratt C.,
- Berke A.,
- et al.
- Bhatt D.L.,
- Roe M.T.,
- Peterson E.D.,
- et al.,
- CRUSADE Investigators
- Damman P.,
- Clayton T.,
- Wallentin L.,
- et al.
- Hirsch A.,
- Windhausen F.,
- Tijssen J.G.,
- et al.,
- ICTUS Investigators
- van't Hof A.W.,
- de Vries S.T.,
- Dambrink J.H.,
- et al.
- Katritsis D.G.,
- Siontis G.C.,
- Kastrati A.,
- et al.