From the University of Ottawa Heart Institute, Ottawa, Ontario, Canada, and Duke Clinical Research Institute, Durham, North Carolina
Glycoprotein (GP) IIb/IIIa platelet receptor blockade reduces ischemic complications of percutaneous coronary intervention (PCI) with stenting.1-4 Several studies have suggested abnormal platelet activity and reactivity in the presence of diabetes.5,6 Clinically, a greater degree of benefit in patients with diabetes than in those without diabetes has been reported in one analysis of the Evaluation of Platelet IIb/IIIa Inhibitor for Stenting (EPISTENT) trial of abciximab (ReoPro®), a monoclonal antibody that blocks the GP IIb/IIIa receptor.7 Analysis of the Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy (ESPRIT) trial2 regarding the outcomes of patients with diabetes treated with eptifibatide, a small molecule GP IIb/IIIa inhibitor, is reported herein.
Patients and methods
The ESPRIT trial2 randomly assigned 2,061 patients undergoing nonurgent coronary stent implantation to treatment with either eptifibatide or placebo. Eptifibatide was administered as two 180-µg/kg boluses given 10 minutes apart followed by a 2.0-µg/kg per minute infusion for 18 to 24 hours. All patients were treated with aspirin and a thienopyridine, and heparin was given during the procedure to achieve an activated clotting time of 200 to 300 seconds. The primary end point was a composite of death, myocardial infarction (MI), urgent target vessel revascularization (TVR), or bailout GP IIb/IIIa treatment for thrombosis at 48 hours. Secondary end points included the individual and combined occurrences of death, MI, and TVR at 30 days, 6 months, and 1 year. Definitions of these end points have been previously published.1 All sites were blind to treatment for the 1-year follow-up period.
For this report, patients were classified as having diabetes based on their medical history and included diet-controlled patients as well as patients receiving oral hypoglycemic agents or insulin. This definition is the same one used in the report of outcomes by diabetes status by Labinaz and colleagues8; the definition of diabetes as reported in the primary outcome paper did not include diet-controlled patients.1 These definitions were as specified in the long-term follow-up and primary trial protocols, respectively. Diabetic and nondiabetic patients were then analyzed for differences in baseline and procedural characteristics, and for differences in outcomes at 48 hours and 30 days with either a Pearson chi-square test or Wilcoxon rank-sum test, depending on the type of variable. A P value of .05 or below was considered statistically significant. A log-rank chi-square test was used to compare event rates at 6 months and 1 year by diabetes status.
To determine if treatment effects were different by the presence or absence of diabetes, the covariates of treatment assignment, diabetes, and the interaction of the two were entered into a logistic regression model for events at 48 hours and 30 days. The covariates were then entered into a Cox proportional hazards model to determine the interaction at 6 months and 1 year.
In the ESPRIT trial, 2,061 patients were randomly assigned to either eptifibatide or placebo as an adjunct to stent PCI. Of these, 466 (22.6%) had diabetes and 1,595 (77.4%) were classified as not having diabetes. There was complete follow-up in 95.6% of nondiabetic patients and 94.4% of diabetic patients. One-year mortality status was available in 98.5% of nondiabetic patients and 98.3% of diabetic patients.
In the overall ESPRIT trial, and also the subgroups by presence or absence of diabetes, there were no significant differences in baseline clinical characteristics between patients assigned to eptifibatide treatment or placebo. The comparison
of baseline characteristics between patients by diabetes status showed several differences. Diabetic patients were more likely to be women (37.6% versus 24.2%; P = .001), to have hypertension (71.7% versus 55.0%; P = .001), to have a higher weight (89.0 kg versus 83.5 kg;
P < .001), and to have undergone a previous PCI procedure (26.8% versus 22.4%; P = .046). Angiographic and procedural characteristics were similar between diabetic and nondiabetic patients, except that the diabetic patients tended to have longer procedure times.
Patient outcomes by diabetes. The primary end point (composite of death, MI, urgent TVR, or thrombotic bailout) between diabetic and nondiabetic patients was not statistically different at 48 hours (6.7% versus 9.1%) or at 30 days (7.5% versus 8.9%). At 6 months, the incidence of the composite end point of death, MI, and any TVR was higher in patients with diabetes, having been reached in 20.9% of diabetic patients compared with 15.0% of patients without diabetes (P = .008). At 1 year, patients with diabetes continued to have an excess of composite end point events (24.5% versus 18.4%;
P = .008), resulting mostly from a higher incidence of TVR in diabetic patients (figure 1).
Effect of eptifibatide on patient outcomes by diabetes. The primary end point for patients assigned to eptifibatide compared with placebo occurred at 48 hours in 4.7% versus 8.6% in diabetic patients and 7.2% versus 11.0% in nondiabetic patients. The effect of eptifibatide on diabet-
ic outcomes at 1 year is shown in
figure 2. At all time points, the
benefit of eptifibatide treatment was approximately the same with respect to diabetes status (and by definition of diabetes). All analyses at all time points for an interaction
of eptifibatide treatment with diabetes documented no additional specific benefit (or lack thereof) in diabetic versus nondiabetic patients (table).
The effect of a small molecule GP IIb/IIIa inhibitor on outcomes according to diabetes status for patients undergoing nonurgent coronary stent implantation has not previously been studied. Treatment with eptifibatide was associated with a reduction in the primary composite end point of death, MI, urgent TVR, or thrombotic bailout at 48 hours that was similar in both diabetic and nondiabetic patients. The benefit of eptifibatide on outcomes continued during the 1-year follow-up period and was similar between diabetic and nondiabetic patients.
Outcomes in diabetic versus nondiabetic patients. The incidence of adverse events in the diabetic group was relatively low, and actually was lower for diabetic patients than for those without diabetes at 48 hours. By 1 year, however, there were significantly more events in the diabetic patients. The incidence of death or MI was similar for all points during the 1-year follow-up. The difference between diabetic and nondiabetic patients was due to an increase in TVR for diabetic patients. This is consistent with the fact that restenosis is known to occur more often with diabetes.9,10
In a previous analysis of the ESPRIT study, there were 419 diabetic patients receiving oral hypoglycemic agents or insulin.2 In performing the 1-year analysis of the ESPRIT study we elected to obtain greater detail on the type of diabetes. As a result, we identified an additional 47 patients who were controlled by diet and were not included in the original analysis. Therefore, in order to provide a comprehensive analysis of all diabetic patients in the ESPRIT study, we included these diet-controlled patients in the current analysis, resulting in a total of 466 patients with diabetes. Statistical testing showed no interaction between the definition of diabetes and outcomes.
Other studies. The EPISTENT investigators studied outcomes in diabetic patients according to the use
of abciximab, a monoclonal antibody that binds to the GP IIb/IIIa receptor and to other integrins.7 Patients with diabetes undergoing stent implantation were found to have a statistically significant reduction in adverse outcomes with abciximab; as in ESPRIT, the overall degree of benefit was similar to that seen in patients without diabetes. The composite end point of death, MI, or TVR at 6 months was significantly lower in diabetic patients undergoing stenting if they received abciximab compared with stenting alone (13.0% versus 25.2%; P = .005). At 6 months, both the incidence
of death and MI were improved
with abciximab (12.7% versus 6.2%;
P = .041). At 1 year, there was a trend toward reduced mortality in patients with diabetes (4.1% versus 1.2%).
Subsequently, tirofiban, another small molecule GP IIb/IIIa inhibitor, was compared with abciximab as an adjunct to stent PCI in the Do Tirofiban and ReoPro Give Similar Efficacy Trial (TARGET).11 The 6-month composite end point of death, MI, or TVR was similar with abciximab compared with tirofiban in diabet-
ic patients undergoing coronary stent implantation (16.7% versus 15.2%, respectively), suggesting similar long-term outcomes regardless of the GP IIb/IIIa inhibitor chosen.
Target vessel revascularization. The significant effects of eptifibatide
in both diabetic and nondiabetic
patients appear to be principally
in the reduction of death and MI. The effects on reducing TVR appear to be much less important. Studies
of the effects of GP IIb/IIIa inhibi-tors on TVR according to diabetes status have reported conflicting results. In EPISTENT, the subgroup of patients having diabetes and undergoing coronary stenting had a 51% reduction in the rate of TVR with treatment (16.6% versus 8.1%;
P = .021). Conversely, there was no effect on reducing TVR among patients without diabetes (8.8% versus 9.0%). The earlier Evaluation in PTCA to Improve Long-term Outcome with Abciximab Glycoprotein IIb/IIIa Blockade (EPILOG) trial
of abciximab suggested a trend toward increased rates of TVR after abciximab treatment in patients with diabetes treated with balloon an-
gioplasty.12 Perhaps the most relevant trial is the TARGET trial11; in this study, rates of TVR at 6 months in patients with diabetes were
10.8% after abciximab treatment and 8.8% in those treated with tirofiban, suggesting no specific interaction between diabetes and abciximab treatment with regard to the TVR end point.
In this study of eptifibatide, patients with diabetes assigned to eptifibatide or placebo had almost identical rates of TVR at 6 months (13.8% versus 13.0%), which continued to be no different at 1 year. The reason for the reported differences between studies of the effects of platelet GP IIb/IIIa inhibition vis á vis TVR is unclear, but is most likely related to the small sample sizes inherent to subgroup analyses, the differences in trial designs or populations, or both.
This study analyzed the outcomes of patients with diabetes compared with those without diabetes in the ESPRIT trial. Eptifibatide treatment during nonurgent coronary stent implantation conferred similar degrees of benefits to both cohorts of patients in reducing ischemic complications associated with stent PCI.