EPICOR (long-tErm follow-up of antithrombotic management Patterns In acute CORonary syndrome patients - NCT01171404) is a prospective, international, observational, real-life practice, cohort study. The rationale, design, definitions, site selection, and baseline patient characteristics have been published previously.14–16 Briefly, 10 568 patients hospitalized for an acute coronary syndrome, with or without ST-segment elevation, within 24hours of symptom onset and who survived until hospital discharge were enrolled in 555 hospitals in 20 countries in Northern, Southern, and Eastern Europe and Latin America between September 2010 and March 2011. Patients were excluded from the study if they had ‘secondary’ acute coronary syndrome, any condition or circumstance that might limit completion of follow-up, serious comorbidities considered likely to limit life expectancy to less than 6 months, and previous enrolment in EPICOR or another clinical trial. All patients gave informed consent. Medical treatments for acute coronary syndrome, diagnostic and therapeutic procedures, and clinical events during the acute phase (pre- and in-hospital) were recorded using electronic case report forms. Patients were followed up by telephone calls for up to 2 years after hospital discharge. Vital status, hospitalizations, cardiovascular and bleeding events, and changes in medication were recorded for each call.

The definitions used in EPICOR have been presented elsewhere.14,16 A diagnosis of non—ST-segment elevation myocardial infarction required the presence of chest pain/discomfort, lack of persistent ST-segment elevation, left bundle branch block or intraventricular conduction disturbances, and elevation of cardiac biomarkers (creatinine kinase-isoenzyme MB and troponins) with at least 1 value above the 99th percentile of the upper reference limit. Unstable angina was defined as the presence of angina symptoms at rest or on minimal exercise, and transient ST-T changes, and no significant increase in biomarkers of necrosis but objective evidence of ischemia by noninvasive imaging or significant coronary stenosis on angiography. Cardiovascular events included myocardial infarction, heart failure, arrhythmia, unstable angina, ischemic stroke, and transient ischemic attack. Bleeding events included all kinds of bleeds.

Two management strategies were defined for patients with NSTEACS: a) CR, which included patients who underwent any kind of CR (either percutaneous or surgical) during the index admission, and b) MM, for those discharged without CR. According to the reasons for MM, 3 subgroups were predefined: a) patients who did not undergo diagnostic coronary angiography (CAG–); b) patients who underwent CAG and had significant (at least 1 stenosis > 50% in 1 coronary artery) coronary artery disease (CAD) but did not undergo CR (CAG+, CAD+), and c) patients who underwent angiography and had no significant CAD (CAG+, CAD–).17

Baseline characteristics, hospital management, and in-hospital outcomes for patients with NSTEACS were compared according to the initial management strategy. Comparisons were made between CR and MM or across the 3 MM subgroups using chi-square tests. In a second step, we investigated the independent predictors of selection for angiography or MM. We used univariate logistic regression models to assess any association between angiography or MM and individual covariates. To identify the strongest independent predictors, we used multivariable logistic regression. We forced the inclusion of geographical region (Northern Europe, Eastern Europe, Southern Europe, and Latin America) and type of hospital (regional, nonuniversity general, university general, and private) into the model. Additionally, we fitted a random-effect at the hospital level to account for within-hospital clustering of events. We used a forward stepwise variable selection with a P-value cut-off of .05 to select a final model. Finally, the impact of MM on 2-year outcomes was studied. Comparisons of clinical outcome rates (mortality, cardiovascular events, and bleeding events) during follow-up between the management groups were done by fitting a Cox proportional hazards model for time to death or time to first event, censored at 2 years postdischarge. In our minimally adjusted Cox models, we adjusted for age, sex, geographical region, type of hospital (as described above), and a random-effect (shared frailty) term at the hospital level. In our fully adjusted models, we additionally adjusted risk factors associated with 1-year mortality identified from our previous publication.18

A total of 5625 NSTEACS patients were enrolled at hospital discharge. Data on in-hospital management strategies were available for all except 34 (0.7%) of these. Of the remaining 5591 patients, 4405 (78.8%) underwent CAG (Figure 1). Of these, 3954 patients (70.7%) had CAD, and 3285 (58.8%) underwent CR in hospital. Therefore, a total of 2306 patients (41.2%) were medically managed. Most MM patients (51.4%, n = 1186) did not undergo CAG during hospitalization (21.2% of total population), 451 (19.6% of MM, 8.1% of total population) lacked significant CAD, and 669 (29.0% of MM, 12.0% of total population) had significant CAD, but CR was not attempted (Figure 1).

Figure 1.

Distribution of EPICOR NSTEACS patients according to initial revascularization strategy and clinical pathways leading to medical management. CABG, coronary artery bypass graft; CAD, coronary artery disease; CAG, coronary angiography; NSTEACS, non—ST-segment elevation acute coronary syndrome; PCI, percutaneous coronary intervention.

(0.15MB).

Patients who received MM were older and less likely to present with non—ST-segment elevation myocardial infarction, but more often had prior cardiovascular disease, comorbidities, and cardiovascular medications (Table 1). They also had more severe cardiac disease (Table 1). When characteristics were compared across the 3 predefined subgroups of MM patients, significant differences were found again, with a gradient from younger age and lower comorbidity and cardiovascular burden among CAG+ CAD– patients to older and sicker patients among CAG– patients. Significant regional differences were found in the rate of MM (data not shown).

The most important independent predictor of undergoing CAG during the index hospitalization (Table 1 of the Supplementary Material) was the presence of a catheterization laboratory in the hospital (odds ratio [OR], 46.8; 95% confidence interval [95%CI], 22.4-97.6). Non—ST-elevation acute myocardial infarction (OR = 1.72; 95%CI, 1.24-2.38) was associated with a higher probability of undergoing CAG compared with unstable angina as well as prior myocardial infarction (OR = 1.58; 95%CI, 1.07-2.32), while age > 75 years (OR = 0.38; 95%CI, 0.28-0.53), current smoking (OR = 0.67; 95%CI, 0.51-0.88), hemoglobin levels < 13g/dL (OR = 0.65; 95%CI, 0.48-0.78), prior myocardial infarction (OR = 0.56; 95%CI, 0.39-0.67), prior coronary artery bypass graft surgery (OR = 0.60; 95%CI, 0.38-0.94), prior heart failure (OR = 0.30; 95%CI, 0.19-0.49), and being on angiotensin-converting enzyme inhibitors at admission (OR = 0.70; 95%CI, 0.53-0.92) were associated with lower probabilities. Patients from Latin America (OR = 0.04; 95%CI, 0.02-0.11) and Eastern Europe (OR = 0.15; 95%CI, 0.06-0.35) had a lower probability of undergoing CAG than patients from Northern Europe.

Independent predictors of not undergoing CR (Table 2 of the Supplementary Material) among patients who underwent CAG and had significant CAD were prior cardiovascular disease (OR = 0.53; 95%CI, 0.42-0.67), prior coronary artery bypass graft (OR = 0.45; 95%CI, 0.32-0.63), age > 75 years (OR = 0.73; 95%CI, 0.55-0.98), and serum creatinine > 1.2mg/dL (OR = 0.76; 95%CI, 0.58-0.99) were marginally associated with lower probabilities while male patients had a higher probability (OR = 1.34; 95%CI, 1.04-1.72). Patients from Latin America (OR = 0.29; 95%CI, 0.18-0.48) and Eastern Europe (OR = 0.50; 95%CI, 0.33-0.87) had a lower probability of undergoing revascularization after CAG than patients from Northern Europe. Admission to private hospitals was associated with an increased probability of being revascularized during hospitalization (OR = 2.19; 95%CI, 1.14-4.20).

In general, MM patients less frequently received diagnostic and therapeutic procedures during hospitalization than did CR patients (Table 2). Although all antithrombotic drugs and most cardiovascular preventative treatments were prescribed in most patients, MM patients were less likely to receive them in hospital. Among those who underwent CAG, multivessel disease was significantly more frequent in CR than in MM patients as a whole but not in the subgroup of MM patients with significant CAD. Interestingly, the results of CAG triggered small changes in antiplatelet drugs both in CR and MM patients, with the exception of clopidogrel, which was withdrawn in a substantial proportion of MM patients at discharge (Table 2).

Patients who underwent MM had a higher incidence of in-hospital cardiovascular complications, mainly heart failure and atrial fibrillation, particularly among those who did not undergo CAG (Table 3). The 2-year postdischarge all-cause mortality rate was 7.0% in the whole cohort, with significant differences between CR and MM patients (4.4% vs 11%; P < .001) (Table 3, Figure 2A). A gradient in 2-year mortality was also found among MM patients, with patients who did not receive CAG showing the highest mortality (14.6%) and those without significant obstructive CAD the lowest (4.1%). Cardiovascular event rates at 2 years, including myocardial infarction, heart failure, arrhythmia, unstable angina, ischemic stroke, and transient ischemic attack, were also significantly higher in MM than in CR patients (15.4% vs 9.6%; P < .001), and were highest in those who did not receive CAG (17.4%) (Figure 2B). In contrast, bleeding events were numerically but not significantly lower in MM vs CR patients (3.4% vs 4.6%; P = .06) (Figure 2C). Among the MM subgroups, the difference in bleeding event rates was not significant but appeared lowest in those who underwent CAG and had no significant CAD. Using 70% stenosis as the cut-off point for CAD+ did not significantly change the results (data not shown). Compared with the results for the 50% cut-off point, there was a slight increase in the mortality rate in both the CAG+CAD+ and CAG+CAD– groups, as they were both composed of higher risk patients, with a small change in mortality gradient between the groups. On exclusion of the 190 patients who underwent revascularization after discharge (including 32 within the first month) from the analyses, no relevant differences were found in patterns of mortality or other event rates.

Figure 2.

Postdischarge event rates at 2 years according to management strategy. A: All-cause mortality. B: Cardiovascular events. C: Bleeding events. Cardiovascular events included myocardial infarction, heart failure, arrhythmia, unstable angina, ischemic stroke, and transient ischemic attack. Bleeding events included all kinds of bleeds. CAD, coronary artery disease; CAG, coronary angiography; CR, coronary revascularization; MM, medical management.

(0.18MB).

Lack of CAG was found to be an independent predictor of 2-year mortality, adjusted for age, sex, and postdischarge mortality predictors, as previously described in the EPICOR cohort18 (hazard ratio [HR] = 1.81; 95%CI, 1.23-2.65; P < .001). Among patients who underwent CAG, MM patients with significant CAD had an increased adjusted mortality risk (HR = 1.90; 95%CI, 1.23-2.95; P < .001), while those without significant CAD did not (HR = 0.68; 95%CI, 0.21-2.21; P < .001) (Table 4). Other predictors of 2-year mortality are shown in Table 3 of the Supplementary Material.

This study is based on registry data and is therefore subject to the limitations of observational studies, ie, potential bias and confounding. The role of patient preferences in the decision to undergo CAG and CR was not recorded, which may have had an additional influence on the outcomes that could not be measured. The analysis of hospital procedures alone excluded patients in which scheduled CAG or CR might have been performed. However, when we used wider time frames for CR–10 days (as in TRILOGY ACS) and 30 days–there were no significant changes in our results, confirming the consistency of our findings. As mentioned previously, although our multivariable analysis included a rigorous adjustment using a previously developed model for mortality prediction,18 unmeasured confounders, such as known CAD not amenable for CR, dementia, too sick for other medical reasons, or patient preferences, could have affected the apparent protective role of CAG and CR. In addition, clinical events during follow-up were not centrally adjudicated. Finally, although we attempted to show representative examples of real-life practice in each country, by careful selection of local centers, caution is warranted in generalizing the results.