The primary endpoint was the 30-day incidence of death, MI, urgent revascularization and major hemorrhage, which occurred in 9.2% of patients in the bivalirudin arm and 10.0% of patients in the heparin-IIb/IIIa arm. The trial had a non-inferiority design and this was satisfied for both the primary endpoint and the key secondary triple ischemic endpoint. Hemorrhage was reduced by 41% from 4.1% to 2.4%.

What was your role in the REPLACE-2 Trial?
It was a large, multicenter study with 233 sites. I was one of the co-investigators at NYU. The principal investigator at our site was Frederick Feit.

What were some of the patient criteria for inclusion?
The inclusion criteria were broad. Basically, you had to be over 21 years of age, have a lesion suitable for coronary intervention, and not be undergoing primary or rescue PCI for an acute MI.
Patients with unstable angina, recent infarction, and multivessel intervention were included. Patients already on a IIb/IIIa inhibitor or those who had received abciximab within the prior 7 days, eptifibatide within 12 hours, or low molecular weight heparin were not eligible.

Why was this trial designed as a non-inferiority trial and not as an equivalency trial?
Many of the recent trials in coronary intervention have been done as non-inferiority trials, the TARGET trial (Tirofiban and Reopro Give Similar Efficacy Outcomes Trial) being the best-known example. Due to the potential advantages of lower cost, shorter infusion time, and less bleeding compared to the heparin-IIb/IIIa combination, the trial was designed with a non-inferiority design.

The trial had an average ACT of 358 secs for the Angiomax (bivalirudin) arm and 317 secs for the heparin-IIb/IIIa arms. Interventions were performed in the 1990s using ACTs that were that high and there were significant bleeding complications. In 2003, using weight-based doses of heparin and keeping ACTs around 200 means there are little in the way of bleeding complications.
The mean ACT 5 minutes after the boluses in the heparin-IIb/IIIa arm was 317 seconds using a weight adjusted heparin dose of 65 U/kg. It™s difficult to criticize the heparin dosing, as this dose is the average of the doses administered in ESPRIT1, (Enhanced Suppression of Platelet Receptor GP IIb-IIIa using Integrilin Therapy) which was 60 U/kg, and EPISTENT2 (Evaluation of IIb/IIIa Platelet Inhibitor for Stenting), which was 70 U/kg. There™s really no data to support working at an ACT of 200. The ACTs from REPLACE-2 are in line with those from ESPRIT and EPISTENT. The mean ACT of 273 seconds from ESPRIT was at an unspecified timepoint, not specifically at 5 minutes. For comparison, by 30 minutes in REPLACE-2, the ACT had fallen to 276 seconds. If one looks at the ACT values from EPISTENT, they are also in the 300 to 320 second range. In addition, the hemorrhagic event rate in the IIb/IIIa arm of REPLACE-2 is actually lower than the bleeding rates in both EPISTENT and ESPRIT.

Would the fact that there was provisional use of GPIIb/IIIa inhibition in the bivalirudin arm, while being universally included in the heparin arm, skew bleeding and transfusion data?
The heparin plus IIb/IIIa arm was felt to be the standard treatment for coronary intervention. That™s the arm associated with the lowest rates of death, MI, and urgent revascularization in ESPRIT and EPISTENT. Bleeding in those trials was not significantly higher with the addition of a IIb/IIIa inhibitor to heparin. There was no heparin monotherapy arm in REPLACE-2, because it was not felt to be ethical, based on the results of previous studies. There was considerable debate as to whether there should be a combination bivalirudin/IIb/IIIa arm. Partly based on economic reasons, but also based on the fact that the results with bivalirudin alone in CACHET (Comparison of Abciximab Complications with Hirulog [now Angiomax®] Ischemic Events Trial) and REPLACE-1 were excellent, the trial was designed as a two-arm study.

Was there any bleeding difference related to which GPIIb/IIIa inhibitor was used?
Overall, major bleeding was reduced from 4.1% in the heparin-IIb/IIIa arm to 2.4% for bivalirudin, which was statistically significant. The rate was essentially the same regardless of which IIb/IIIa inhibitor a patient received 4.0% for abciximab and 4.1% for eptifibatide. The rates of transfusion, thrombocytopenia, and major organ bleed were also reduced in the bivalirudin arm compared to the heparin-IIb/IIIa arm. I have not seen the data yet for these events for each of the IIb/IIIa inhibitors.

The current evidence-based guideline for management of ACS/NSTEMI suggests clopidogrel should be used in most patients with ACS (unless there is a high probability of surgical disease).
The CREDO trial (Clopidogrel for the Reduction of Events During Observation Results), reported at the AHA last fall, greatly strengthens this recommendation to include early treatment at least 6 hours prior to possible PCI and continued at least a year after intervention with a resultant 20% day -1 and 38% one-year relative risk reduction of major cardiac events. Since bivalirudin was not studied against this currently recommended and highly effective treatment regimen, why do the investigators think widespread adoption of bivalirudin is indicated, as opposed to niche use possible in place of GP receptor inhibition?
In other words, if you did heparin alone, with everybody getting pretreatment with clopidogrel, would Angiomax still give you an advantage over heparin?
It™s not really known. It has not been studied in any of the trials. Bivalirudin is a much better thrombin inhibitor than heparin. If aggressive pretreatment more than 6 hours prior to the intervention was done with both Angiomax and heparin patients, I believe there would still be less bleeding and an improvement in ischemic events with bivalirudin when compared to heparin. A similar question is whether the IIb/IIIa inhibitors maintain their advantage over heparin monotherapy in terms of suppression of CK-MB events if all patients are aggressively pre-treated with clopidogrel.
What are some of the new questions that the REPLACE-2 trial data have opened up?
The inclusion criteria were basically elective or urgent PCI, and the trial included patients with complex intervention as well as unstable angina patients, as long as they were not receiving a IIb/IIIa inhibitor or felt by the investigator to require them for the intervention. One question not addressed is whether in patients receiving a IIb/IIIa inhibitor, bivalirudin improves the outcome compared to heparin in a cost-effective manner. Is there still an incremental advantage in replacing heparin with bivalirudin in terms of safety and efficacy?
A second question is whether aggressive pre-treatment with clopidogrel is more essential in patients not receiving a IIb/IIIa inhibitor. The trial data does not show evidence of differential effect with pre-treatment between the 2 arms. One must remember, though, that the time of clopidogrel therapy initiation was not randomized.

Are there any subgroup analyses being performed?
There are many pre-specified subgroup analyses being performed. Important subsets of patients are those with diabetes, the elderly, patients with acute coronary syndromes, the results with the different IIb/IIIa inhibitors, and gender. The results in all of the subgroups analyzed to this point are consistent with the overall trial results, in that non-inferiority was demonstrated with bivalirudin compared to the heparin-IIb/IIIa arm. This includes the large subgroup of 1600 patients with diabetes. Both the quadruple and triple ischemic endpoints were virtually identical in the bivalirudin and heparin-IIb/IIIa arms 8.6% vs 8.4% for the quadruple endpoint and 5.9% vs 5.8% for the ischemic complications. In groups associated with a higher than normal risk of bleeding, such as females and the elderly, there is a larger difference in the primary quadruple endpoint favoring bivalirudin, although the 95% confidence intervals still cross unity. This is most likely due to a decrease in bleeding.

There was an increase in myocardial infarction in the Angiomax group, but there also was a reduction in bleeding.
There™s a non-statistically significant difference in myocardial infarction. The numbers were 6.2% for overall infarction in the heparin- IIb/IIIa arm, and 7.0% in the Angiomax arm, which has a p-value of 0.23.
The rates of Q-wave MIs are identical at 0.4% and the mortality rates are also not significant at 0.2% and 0.4%, favoring bivalirudin. Some interventional cardiologists have been discussing whether there is a tradeoff with a decrease in bleeding for an increase in ischemic events. I don™t believe there™s a tradeoff. Bivalirudin with provisional abciximab satisfied its not-inferiority objective regarding the triple ischemic endpoint.

What about the bolus-only use of Angiomax as being practiced in some labs?
There is no data to support the use of bivalirudin as a bolus-only regimen. Based on the short 25-minute half-life of the drug, I think it™s prudent to start the infusion. The mean duration of therapy in REPLACE-2 was 47 minutes, which is almost 2 half-lives of the drug. We use bivalirudin routinely at our institution for PCI and always start the infusion.

Can you explain the quadruple primary endpoint in REPLACE-2?
The quadruple primary endpoint for the trial was the 30-day incidence of death, MI, urgent revascularization and in-hospital major hemorrhage. The problem with composite endpoints is that you are combining endpoints that are nowhere near equal in terms of their significance, yet are combined in an unweighted manner. The mortality rate of PCI is so low that a mortality trial would require too many patients. The main debate is whether adding hemorrhage to the previously used triple composite endpoint of death, MI and urgent revascularization is valid. I believe it is. In-hospital hemorrhagic events have been shown in many data sets to be strong predictors of mortality and associated with increases in hospital stay and cost.

How do you think the FDA will react to the REPLACE-2 trial data?
REPLACE-2 is a large clinical trial that met each of its prespecified endpoints. It is hard for me to speculate what the FDA will do, but I believe they will react favorably to the data and expand the package labeling of the drug to include patients similar to those studied in the trial “ patients undergoing elective or urgent PCI including coronary stent implantation.

Do you have any comment on the controversies and debates surrounding the release of the REPLACE-2 data?
REPLACE-2 was designed by leaders in the field of interventional cardiology, many of whom were also involved in the design of the pivotal IIb/IIIa inhibitor trials. The results are solid and consistent among the many prespecified subsets. The IIb/IIIa inhibitors are effective in terms of platelet inhibition and the prevention of ischemic events following PCI. They do so, however, at a cost bleeding, prolonged infusion times, and increased cost. They may be necessary when heparin is used as the antithrombin, because it has so many limitations as a thrombin inhibitor. REPLACE-2 demonstrated that you can achieve the same results with bivalirudin with less bleeding, shorter infusion time, and reduced cost. We use bivalirudin routinely at NYU with a low incidence of ischemic complications and a hemorrhagic and vascular complication rate of under 1%.