Management and outcomes of patients presenting with STEMI by use of chronic oral anticoagulation: results from the GRACE registry

Definitions

To characterize the onset of a myocardial infarction (MI) as a hospital outcome, it had to have occurred >24 h after hospital presentation, to have been confirmed by electrocardiographic changes or elevation of cardiac biomarkers, and to have involved one of the following scenarios: (a) patients with an admission diagnosis of unstable angina who developed an MI; (b) patients who were diagnosed with an MI after coronary artery bypass graft surgery (CABG) or a percutaneous coronary intervention (PCI) provided they had qualified for enrolment before the CABG or PCI; or (c) patients who had presented with an MI and were diagnosed as having had recurrent MI.

Patients receiving chronic OAC were defined as those routinely taking a vitamin K antagonist at home and within 7 days of hospital presentation.

All patients in whom cardiac catheterization was performed were included in the catheterization and/or PCI (Cath/PCI) group, regardless of whether a PCI was performed. Hours to Cath/PCI was defined as the time to the earliest procedure. The medical therapy group was defined as those who did not undergo Cath/PCI. Major bleeding in hospital had to be life threatening, require a transfusion of ≥2 units of packed red cells, cause a ≥10% decrease in haematocrit, or result in death. Six-month mortality encompassed all deaths occurring during the 180 days from the time of discharge from hospital.

Medications

At admission, OAC users were more likely than OAC non-users to be taking beta-blockers, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (ACEI/ARB), and statins, whereas aspirin and thienopyridine use was similar in both groups (Table 3).

During hospitalization, fewer of the OAC users received aspirin, beta-blockers, statins, or thienopyridines compared with OAC non-users (Table 3). Vitamin K antagonist therapy was maintained in 52.4% of OAC patients during hospitalization and in 65.1% at discharge, compared with 4.6 and 5.3%, respectively, of OAC non-users. Use of antiplatelet and anticoagulant drugs administered in the first 24 h of admission among OAC users is shown in Figure 1A.

Open in a separate window

Figure 1.

(A) Antiplatelet and anticoagulant agent administration in the first 24 h among 556 STEMI patients on chronic OAC with status known for all three agents. (B) Selected medication combinations at discharge from hospital among STEMI patients with status known for all three medications.

Categories are mutually exclusive.

At discharge from hospital, OAC non-users were most likely to be given aspirin and thienopyridines (52.8%) or aspirin alone (36.3%), whereas medical therapy for OAC users was more varied (Figure 1B).

Interventions and clinical outcomes

The rate of Cath/PCI during hospitalization was lower among OAC users than among OAC non-users (Table 4). Of the patients who had a Cath/PCI, just over half of the OAC users had the procedure performed in the first 24 h compared with almost two-thirds of OAC non-users.

Data were available for propensity matching in 319 patients because the strongest covariates associated with chronic OAC use were available only in the most recent version of the case report form. Non-OAC counterparts with similar model probabilities were found for 303 of these patients. In multiple logistic regression, the three strongest factors associated with chronic OAC use were medical history of atrial fibrillation, venous thromboembolism, and prosthetic valve replacement; other factors included age, presentation in Killip class III, and medical history of transient ischaemic attack, congestive heart failure, percutaneous coronary intervention, coronary artery bypass graft, and diabetes mellitus. The model c-statistic was 0.92.

After propensity matching, the three strongest factors in univariate analysis were similar among OAC users and non-users: 55.1% of OAC users and 55.7% of non-users had a history of atrial fibrillation; 6.3 and 3.6% had a prosthetic valve replacement; and 11.2 and 12.2% had a history of venous thromboembolism (full data not shown). In propensity-matched analyses, there was no significant difference in the overall frequency of in-hospital major bleeding in OAC users and non-users (Table 5), while recurrent MI and atrial fibrillation occurred more frequently in OAC users. Unadjusted in-hospital mortality rates were higher among OAC users than non-users (15.2 vs. 7.4%, p<0.0001; full data not shown), but were similar after propensity matching (15.2 and 13.2%; Table 5). Among discharged patients, OAC users had a higher unadjusted mortality rate at 6 months than non-users (13.4 vs. 4.3%, p<0.0001; data not shown), whereas propensity-adjusted differences were not statistically significant (14.7 vs. 11.8%, p=0.47; Table 5). Adjusted 6-month stroke rates were lower in OAC users (Table 5). No other adjusted differences were found.

Impact of Cath/PCI in OAC users

OAC users who underwent Cath/PCI during hospitalization were significantly younger, more likely to be male, to be in Killip class I, and have a lower GRACE risk score than OAC users who were treated with medical therapy only, whereas history of CABG did not differ between the two groups (Table 6).

Major bleeding occurred at a higher rate in OAC users who had Cath/PCI compared to those treated medically, but hospital mortality was higher for patients treated medically. Among patients discharged alive from hospital, 20.0% of medically treated patients died within 6 months compared with 6.9% of those who underwent Cath/PCI (Table 6).

Choice of reperfusion therapy

An analysis of data from GRACE found that the use of evidence-based therapies and PCI has increased over time in the STEMI population, and this has been matched with statistically significant decreases in the rates of death, cardiogenic shock, and heart failure compared to fibrinolysis.¹¹ A concern specific to emergency PCI is that of access-site bleeding complications, particularly among patients receiving OAC. We found a significantly higher frequency of major bleeding among patients who underwent cardiac catheterization, but whether this was directly related to the degree of anticoagulation (international normalized ratio, INR) or to the more frequent of use of anticoagulant and antiplatelet therapies in the Cath/PCI group is not known. Overcoagulation is likely to have occurred and may explain these findings.

We analysed the data for either diagnostic catheterization or catheterization followed by a PCI assuming an intention to perform a PCI, because in real life, not all patients who present with STEMI have a coronary anatomy suitable for this procedure. We wanted to emphasize this principle in a challenging population, in which comorbidities and the administration of chronic OAC may have played a significant role in deciding who should undergo Cath/PCI and when it should be performed.

Choice of adjunctive pharmacotherapy and cardiac catheterization access site

With regard to adjunctive pharmacotherapies (e.g. aspirin, thienopyridines, glycoprotein IIb/IIIa inhibitors, heparins, and direct thrombin inhibitors) in OAC patients presenting with STEMI, there is no evidence to suggest that one drug is better or safer than another because these patients have traditionally been excluded from clinical trials involving such treatments (Table 7).^12–32 Neither do we know whether different loading or maintenance doses of these drugs should be used.

The same applies to the choice of access site for cardiac catheterization. Two recent trials provide evidence that transradial catheterization is associated with a lower bleeding risk, but they too excluded patients on OAC.^16,24 It would appear reasonable, however, to infer that patients on OAC should preferably undergo transradial catheterization.

In the absence of evidence, the use of anticoagulation in the cardiac catheterization laboratory varies widely for OAC patients with STEMI: while some operators may choose the direct thrombin inhibitor bivalirudin because of its overall lower bleeding risk, others elect to use unfractionated heparin because its effects can be reversed with protamine in the event that bleeding occurs.

Discharge medications and outcomes

Another central problem is the systematic definition of a discharge medication plan for patients who require single or dual antiplatelet therapy in addition to OAC due to their long-term bleeding risk. These patients are treated with great variability. Various retrospective and prospective studies have addressed the role of triple therapy (aspirin plus thienopyridine plus vitamin K antagonist) and bleeding risk.^33–35 A multicentre, randomized, open-label trial is underway to assess whether the combination of OAC and clopidogrel reduces the risk of bleeding and is not inferior to triple therapy with respect to the prevention of thrombotic complications after PCI in patients on OAC.³⁶

We thank the patients, physicians and study coordinators participating in GRACE, and the staff at the Center for Outcomes Research, including Gordon FitzGerald for statistical support, Linda Chase and Sophie Rushton-Smith. Further information about the project, along with the complete list of participants, can be found at www.outcomes.org/grace.