Author + information
- Received May 6, 1999
- Revision received September 11, 2000
- Accepted October 16, 2000
- Published online February 1, 2001.
- Adam B Greenbaum, MD∗,* (, )
- Robert A Harrington, MD∗,
- Michael P Hudson, MD∗,
- Cynthia M MacAulay, MS∗,
- Robert G Wilcox, MD†,
- Maarten L Simoons, MD, FACC‡,
- Lisa G Berdan, PA-C, MHS∗,
- Alan Guerci, MD, FACC§,
- Dennis V Cokkinos, MD, FACC∥,
- Michael M Kitt, MD¶,
- A.Michael Lincoff, MD, FACC#,
- Eric J Topol, MD, FACC#,
- Robert M Califf, MD, FACC∗,
- E.Magnus Ohman, MD, FACC∗,
- for the PURSUIT Investigators
- ↵*Reprint requests and correspondence: Dr. Adam B. Greenbaum, Henry Ford Hospital, K-2 Cardiac Catheterization Laboratory, 2799 West Grand Boulevard, Detroit, Michigan 48202
We aimed to evaluate the benefits of the glycoprotein (GP) IIb/IIIa antagonist, eptifibatide, after patients with acute coronary syndromes (ACS) were admitted to hospitals that approach revascularization for ACS through early transfer to tertiary referral centers.
Across a variety of hospital settings, GP IIb/IIIa inhibition, after patients were admitted to the hospital for non-ST segment elevation ACS, is associated with a reduction in death or myocardial infarction (MI) before and during a percutaneous coronary intervention.
The outcomes of 429 patients from 153 sites in the Platelet glycoprotein IIb/IIIa in unstable angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) trial, who were transferred during study drug infusion (“transfer patients”), were compared with those of 1,987 patients who either remained in the hospital at those sites or were transferred after study drug termination (“nontransfer patients”).
The baseline characteristics of transfer and nontransfer patients were similar. Patients receiving eptifibatide were transferred less frequently than those receiving placebo (16% vs. 20%, p = 0.014). Transfer patients underwent more procedures and experienced a greater 30-day incidence of death or MI, as compared with nontransfer patients (21% vs. 12%, p = 0.001). Eptifibatide was associated with a reduction in death or MI through 30 days, independent of transfer status (2.5% absolute reduction), as well as for those transferred (5.5% absolute reduction).
For patients with ACS admitted to community hospitals, eptifibatide is associated with a reduced need for transfer and improved clinical outcomes.
Disruption of atherosclerotic plaque and subsequent platelet aggregation are common in both acute coronary syndromes (ACS) and percutaneous coronary interventions (PCI) (1–3). Randomized trials have shown that platelet glycoprotein (GP) IIb/IIIa receptor inhibition reduces the ischemic complications of percutaneous revascularization for unstable angina (4–7). Other trials have shown that the benefits of GP blockade are in effect during the period before PCI (5,8,9).
The Platelet glycoprotein IIb/IIIa in unstable angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) trial was designed to test whether potent platelet inhibition with eptifibatide would reduce death and myocardial infarction (MI) in patients with ACS. Among PURSUIT patients, those undergoing revascularization within the first 72 h experienced a 5.1% absolute reduction in events with the use of eptifibatide, and GP IIb/IIIa inhibition reduced the frequency of the composite end point both before and after the procedure (9).
Although the data supporting the use of GP IIb/IIIa antagonists during ACS are simple to extrapolate for centers with interventional facilities, the data are less clear for centers without such facilities (75% of U.S. hospitals and an even greater percentage worldwide), many of which employ a strategy of revascularization through transfer to tertiary referral centers. For these centers, the benefits of GP IIb/IIIa administration before a decision is made regarding revascularization or documentation of a percutaneously approachable lesion must be weighed against potential complications associated with use of the drug during patient transfer, as well as in a broader spectrum of the patient population.
In PURSUIT, a percentage of patients were transferred to tertiary care centers because of worsening clinical status or for additional diagnostic or interventional procedures. This study explores the strategy of therapy and early transfer to referral centers of patients with ACS initially admitted to nontertiary care centers, and it specifically examines the treatment effect of eptifibatide within this strategy. We hypothesized that eptifibatide administration might have incremental benefit in patients subsequently transferred because of their high risk profiles and the likelihood of a subsequent coronary intervention, without the increased complications associated with pretransfer use.
The study design and results of PURSUIT have been published previously (9). Patients presenting with ischemic chest discomfort at rest lasting ≥10 min within the previous 24 h were eligible for randomization, provided they had either creatine kinase, myocardial band fraction (CK-MB) elevation or electrocardiographic (ECG) evidence of ST segment or T-wave deviation. Patients were randomized to receive an intravenous bolus and infusion of placebo, a bolus of 180 μg/kg body weight followed by an infusion of 1.3 μg/kg per min of eptifibatide or a bolus of 180 μg/kg followed by an infusion of 2.0 μg/kg per min of eptifibatide. The study drug was infused for 72 h or until hospital discharge. Aspirin and intravenous or subcutaneous heparin were recommended for all patients. During eptifibatide administration, fibrinolytic therapy and other GP IIb/IIIa inhibitors were prohibited. All other treatment decisions, including medical therapy, hospital transfer, cardiac catheterization and revascularization, were left the discretion of the physician caring for the patient. If transfer to a tertiary referral center was desired, continuation of the study drug was recommended. If catheterization and angioplasty were performed, continuation of the drug for 24 h after the procedure, up to a maximum of 96 h, was also recommended.
The primary end point of PURSUIT was the composite of death from any cause or nonfatal MI within 30 days. Myocardial infarction was defined as new chest pain and ST segment elevation within 18 h of enrollment, new or repeat CK-MB elevation above the upper limit of normal after 18 h, CK-MB elevation above three times the upper limit of normal after PCI and CK-MB elevation above five times the upper limit of normal after coronary artery bypass graft surgery. Suspected infarctions were adjudicated by a Clinical Events Committee in blinded manner, and serial cardiac enzyme testing was mandated for all prolonged anginal episodes and revascularization procedures.
The primary safety end point was bleeding. Bleeding was classified as mild, moderate, severe or life-threatening, based on the Global Utilization of Streptokinase and TPA for Occluded arteries (GUSTO) scale (10), and as major or minor, based on the Thrombolysis In Myocardial Infarction (TIMI) scale (11).
In total, 10,948 patients from 726 participating hospitals in 28 countries were randomized between November 1995 and January 1997. Of the 9,461 patients in the two primary treatment arms, 2,438 were enrolled at 153 hospitals that transferred at least one patient to a tertiary referral center during study drug infusion. Of these patients, 429 were transferred during the study drug infusion (“transfer patients”); 464 were transferred after completion of the study drug; 1,534 completed the study drug infusion or their hospital course at the enrolling facility; and 11 had unknown transfer status and were therefore excluded from the analysis (Fig. 1). For the purposes of this study, which evaluated the relative efficacy of eptifibatide on clinical outcomes at centers employing a strategy of early transfer, we assumed that patients transferred after study drug termination had much less risk of adverse events than those transferred earlier. In addition, we predicted no continued benefit of drug therapy after discontinuation (12). The 464 patients transferred after study drug termination were therefore considered as the “nontransfer” group.
The patients’ baseline demographic and clinical characteristics were collected from case report form data and summarized for both transfer and nontransfer groups. The baseline variables of both groups were also stratified by randomized therapy (eptifibatide vs. placebo). Continuous variables are presented as the median value (interquartile range). For dichotomous variables, actual percentile values are expressed. Comparisons between transfer and nontransfer groups and between treatment groups were conducted by using either a conventional chi-square test or the Fisher exact test for dichotomous variables, and the Wilcoxon rank-sum test for continuous variables.
Because hospital transfer was not a randomized event, transfer and nontransfer groups may possess unequal distributions of known predictors for death and MI. We attempted to control for differences between the groups by using an established model of 30-day death or MI developed from the entire PURSUIT population (13). Significant predictors of death or MI in this population, among others, included age, gender, region of enrollment, the presence of ST segment depression on the ECG at enrollment and the presence of MI at enrollment. The primary outcome was analyzed both before and after adjustment for differences in the baseline variables between the transfer and nontransfer groups. To lessen the bias, all nontransfer patients who died before the median time of hospital transfer (two days) were also excluded from the analysis (n = 11 [8 received placebo, 3 received eptifibatide]) (Fig. 1).
To test for the differential effect of eptifibatide treatment between the two groups, the interaction between transfer status and randomized treatment was sampled and tested for statistical significance using logistic regression analysis. The results are also displayed as odds ratios (95% confidence intervals) from logistic models (to compare the relative treatment differences of eptifibatide in nontransfer and transfer patients).
Table 1displays the baseline characteristics of the 2,416 patients. A higher proportion of transfers occurred in North America as compared with other regions, and a history of PCI was associated with transfer. Fewer transfer patients had a history of congestive heart failure, and there was a trend toward transferring a higher percentage of patients with MIs at enrollment compared with unstable angina. For both the transfer and nontransfer groups, the baseline characteristics were similar for those randomized to eptifibatide and those randomized to placebo.
Figure 2displays the relationship between randomized therapy and transfer status. Patients randomized to and receiving eptifibatide were less frequently transferred than those receiving placebo. Of the patients transferred, the median time to transfer was two days (range 1 to 3) from randomization, with no significant difference in the time to transfer between those receiving eptifibatide and those receiving placebo. The median time to transfer in the 464 patients (23%) transferred after study drug termination was seven days (range 5 to 12) after randomization. Fifty-six percent of the patients had the study drug discontinued before or on the day of transfer, whereas 44% had therapy continued for a median of one additional day (range 1 to 2). Transfer patients underwent more angiography and revascularization as compared with nontransfer patients; the majority of these procedures (83%) occurred on the day of or after transfer (Table 1).
Table 2shows the unadjusted 30-day incidence of death or MI by transfer status and by treatment with eptifibatide. Regardless of randomized therapy and the timing of events, the 30-day incidence of death or MI was greater in transfer patients than in nontransfer patients (p = 0.001). Eptifibatide was associated with a reduction in death or MI in both transfer and nontransfer patients, with a greater reduction in the composite end point in transfer patients as compared with nontransfer patients. The greater absolute reduction in death or MI seen in transfer patients as compared with nontransfer patients was not statistically different after adjustment for differences in baseline variables between the two groups (p = 0.54). Figure 3displays the relative 30-day benefits of eptifibatide as compared with placebo in both transfer and nontransfer patients.
Eptifibatide showed a trend toward reducing the 30-day incidence of death or MI, whether the study drug was discontinued before or on the day of transfer (19.1% vs. 24.4%, p = 0.321) or continued (16.0% vs. 22.3%, p = 0.288). However, the benefit of eptifibatide was significantly greater in those who continued the study drug (p = 0.04). Figure 4and Table 3display the cumulative incidence and relative timing of death or MI, respectively, between transfer and nontransfer patients randomized to eptifibatide or placebo. Among transfer patients, the majority of benefit from eptifibatide was seen before and on the day of transfer.
There were more bleeding events in transfer patients (Table 4). However, the incidence of severe or life-threatening bleeding was not different, nor was there any significant difference in the incidence of stroke. Furthermore, among transfer patients, there was no significant difference in the incidence of bleeding or stroke between those randomized to eptifibatide and those receiving placebo.
This study, evaluating GP IIb/IIIa inhibition in patients with non-ST segment ACS presenting to community hospitals, where the decision to transfer rested with the treating physician, introduces valuable information for treating patients in this setting. First, eptifibatide significantly reduced the frequency of hospital transfer. Second, transfer patients, in general, represent a group at high risk for adverse events. Finally, eptifibatide use surrounding transfer was associated with a reduction in death or MI through 30 days.
Patients are transferred from community hospitals to tertiary referral centers after admission with ACS for many reasons, ranging from elective transfers for procedures to urgent transfers for worsening clinical conditions. In PURSUIT, the decision to transfer patients was not a randomized event; many clinical factors influenced the decision. Although this type of study precludes direct evaluation of the reason for transfer, physicians had no knowledge of the GP IIb/IIIa therapy during this period, and significantly less patients receiving eptifibatide were transferred.
There has been much discussion regarding the role of GP IIb/IIIa inhibitors in the passivation of the coagulation cascade and stabilization of the ruptured plaque during ACS (14). Eptifibatide may have altered the clinical course of patients enough that physicians caring for the patients were less inclined to transfer those patients. As evidence mounts suggesting the benefits of reserving revascularization for those who cannot be medically stabilized (15–17), GP IIb/IIIa antagonism may prove integral in the initial management of patients with ACS in community hospitals. Furthermore, as health care resources become scarce and insistence on the stringent justification of expenses becomes more commonplace, preventing transfer and subsequent procedures with GP IIb/IIIa inhibition may also prove cost-effective.
Despite adjustment for differences in baseline variables between transfer and nontransfer patients, death or MI occurred more frequently in transfer patients than in patients admitted to those same centers and in those not transferred while on the study drug (21% vs. 12%), as well as in patients who remained at the site where they were enrolled in the overall trial (15%) (8). This suggests that transfer patients represent a higher risk group with recurrent ischemia or more apt to have major adverse events. These events were not solely related to revascularization procedures: of the 62 end point MIs occurring after transfer, 40% occurred far after the procedures. We also observed a slightly greater reduction in 30-day events with eptifibatide in transfer patients compared with nontransfer patients, suggesting an augmented benefit of GP IIb/IIIa blockade in patients with increased risk of short-term complications.
In this analysis, the 464 patients who were transferred after study drug termination were included in the nontransfer group. We considered the patients transferred after study drug termination from centers concomitantly transferring patients during study drug infusion to represent a dissimilar group, and one at lesser risk for adverse events than the group of patients transferred during study drug infusion. Compared with patients transferred during study drug infusion, late transfers experienced an overall lower rate of death or MI at 30 days (21.0% vs. 17.2%). Nonetheless, the unexpected finding of a reduced need for transfer associated with eptifibatide persists when all transfers are combined (34% vs. 40%, p = 0.005) and when centers originally excluded for only transferring patients after study drug termination are included (22% vs. 26%, p = 0.006). As no continued accrual of benefit would be expected after termination of GP IIb/IIIa inhibition, some dilution of the treatment effect with eptifibatide is observed (3.2% vs. 5.5% absolute reduction) when all transfers are analyzed together, possibly due to late procedures in these patients without GP IIb/IIIa coverage. However, a reduction in events with eptifibatide is still apparent, and a small benefit is seen, even in an analysis of only those 464 patients who were transferred after study drug termination (0.8% absolute reduction). Our findings suggest that transfer, whether planned or urgent, should be carried out while drug infusion is still ongoing.
In this study, the benefits of administration of eptifibatide by community hospitals before transfer came with no additional risk of serious adverse bleeding. The similar bleeding and stroke rates between those receiving eptifibatide and those receiving placebo occurred despite the high risk demographic data of those transferred and the high rates of invasive procedures after transfer.
To the best of our knowledge, this is the first known evaluation of the strategy of transfer from community hospitals to tertiary centers after patients were admitted for non-ST segment elevation ACS. In our study, transfer to tertiary centers was associated with more procedures and more adverse events. In studies of ST segment elevation acute MI, patients admitted to centers with on-site catheterization facilities similarly undergo angiography and revascularization more frequently than those admitted to centers without catheterization facilities, but with no difference in mortality at three years of follow-up (18). In retrospective comparisons of clinical outcomes between patients with acute MI transferred to tertiary centers for direct angioplasty and matched subjects who arrive directly to these centers, greater enzymatic infarct size and lower six-month ejection fractions are seen in those transferred, but again with no differences in six-month mortality (19,20). The clinical outcomes associated with transfer in our study are consistent with evaluations of multidisciplinary transfers from acute-care hospitals to tertiary centers in general, where higher mortality (>50% relative increases) is common despite adjustment for differences in disease severity (21,22).
Certain limitations must be taken into consideration when interpreting this study. Indications for transfer, and the practice thereof, were not addressed in the PURSUIT protocol. Eptifibatide was therefore not randomly assigned based on transfer status, and was administered in subgroups with unequal baseline characteristics. Although we adjusted for some of these variables, unrecognized confounders could contribute to the observed differences in event rates. Furthermore, transfer was a postrandomization event, possibly affected by study drug assignment or clinical events. Although some of these events were captured on case report forms, there is no model incorporating postrandomization events that predicts 30-day outcomes. Should one exist, uncollected postrandomization events would still confound the conclusions.
Another limitation of our study involves the definition of transfer groups. As we sought to evaluate a therapy initiated shortly after presentation, follow-up began at randomization. Although deaths before the median time to transfer were censored, pretransfer MIs were not. We chose not to exclude these MIs, because the effect of eptifibatide on transfer rates may be mediated through an effect on early MI. This study design allowed events that may have contributed to the decision to transfer to be included in the analysis, and it mimics commonly used clinical care strategies.
Eptifibatide use at community hospitals early after presentation with non-ST segment elevation ACS reduces the need for transfer and improves clinical outcomes in those who require transfer and in those remaining at the community hospital. Transfer, if necessary, should be performed in conjunction with continued GP IIb/IIIa inhibition, based on either the patient’s high risk profile or the likelihood of subsequent revascularization.
The authors thank John Daniel for editorial assistance.
☆ Funding for this manuscript was supplied by Cor Therapeutics, San Francisco, California.
- acute coronary syndromes
- creatine kinase, myocardial band
- electrocardiogram or electrocardiographic
- Global Utilization of Streptokinase and TPA for Occluded arteries trial
- myocardial infarction
- percutaneous coronary intervention
- Platelet glycoprotein IIb/IIIa in unstable angina: Receptor Suppression Using Integrilin Therapy trial
- Thrombolysis In Myocardial Infarction trial
- Received May 6, 1999.
- Revision received September 11, 2000.
- Accepted October 16, 2000.
- American College of Cardiology
- RESTORE Investigators
- Chesebro J.H,
- Knatterud G,
- Roberts R,
- et al.
- Akkerhuis K.M,
- Deckers J.W,
- Boersma E,
- et al.
- Boersma E,
- Pieper K,
- Wilcox R.G,
- et al.
- TIMI-IIIB Investigators
- Every N.R,
- Parsons L.S,
- Fihn S.D,
- et al.
- Liem A.L,
- van ’t Hof A.W,
- Hoorntje J.C,
- de Boer M.J,
- Suryapranata H,
- Zijlstra F
- Zijlstra F,
- van’t Hof A.W,
- Liem A.L,
- Hoorntje J.C,
- Suryapranata H,
- de Boer M.J