Author + information
- Rahman Shah, MD∗ ( and )
- Sunil V. Rao, MD
- ↵∗The Duke Clinical Research Institute, 508 Fulton Street (111A), Durham, North Carolina 27705
We appreciate the comments by Dr. Mahmoud and colleagues about our study. They argued that heterogeneity for major bleeding may have confounded our analysis due to an imbalanced use of radial access and glycoprotein IIb/IIIa inhibitors (GPI) between the unfractionated heparin (UFH) and bivalirudin groups. First, as we noted in our original paper (1), the heterogeneity appeared to originate from the single-center study (i.e., How Effective Are Antithrombotic Therapies in Primary Percutaneous Coronary Intervention [HEAT PPCI]). Once this study was removed from our sensitivity analysis, no significant heterogeneity (p = 0.123) remained among the 4 remaining (multicenter) trials. It is well known that a single-center study is typically an outlier when combined with multicenter trials in a meta-analysis (2). Therefore, it is strongly recommended that meta-analysts perform sensitivity analyses to demonstrate how their conclusions might be affected by the exclusion of single-center trials (2).
Second, the primary focus of our paper was to evaluate the effect of prolonged bivalirudin infusion on major bleeding and acute stent thrombosis. Prolonged bivalirudin infusion at the full percutaneous coronary intervention (PCI) dose was used only in the MATRIX (Minimizing Adverse Hemorrhagic Events by Transradial Access Site and Systemic Implementation of Angiox), EUROMAX (European Ambulance Acute Coronary Syndrome Angiography Trial), and BRIGHT (Bivalirudin in Acute Myocardial Infarction vs Heparin and GPI Plus Heparin Trial) trials. No heterogeneity was found among these trials (Q = 3.1; df = 3; p = 0.374; I2 = 3) for the prolonged infusion subgroup analysis. Radial access was used in the same number of patients in both the bivalirudin and UFH groups in these trials. In addition, the MATRIX trial investigators found no evidence of interaction between access route and allocation to the bivalirudin or UFH groups for major bleeding (p for interaction ≥0.64) (3). In the EUROMAX trial, bivalirudin was associated with a signiﬁcantly lower rate of major bleeding (odds ratio: 0.44; 95% confidence interval: 0.24 to 0.82), even when compared with UFH without routine GPI (4). Similarly, in the BRIGHT trial, bleeding (1 of the primary endpoints) was reduced with bivalirudin compared with UFH alone and UFH with GPI (4.1%, 7.5%, and 12.3%, respectively; p < 0.001). Furthermore, during sensitivity analysis, removing the BRIGHT-UFH+GPI data did not change our summary result for major bleeding (risk ratio: 0.31; 95% CI: 0.15 to 0.63; p = 0.001). Finally, current guidelines recommend a 70 to 100 U/kg bolus of UFH without GPI or a 50 to 70 U/kg bolus of UFH with GPI for primary PCI, which were used in these trials.
Thus, it is very unlikely that our conclusion about the effect of prolonged bivalirudin infusion on major bleeding was confounded by these factors. However, as we noted in our original paper, access to patient-level data would enable multivariate analysis, which would markedly strengthen this analysis.
Please note: Dr. Shah has reported that he has no relationships relevant to the contents of this paper to disclose. Dr. Rao has served as a a consultant for Terumo Interventional Systems, AstraZeneca, Merck, and Medtronic.
- American College of Cardiology Foundation
- Shah R.,
- Rogers K.C.,
- Ahmed A.J.,
- King B.J.,
- Rao S.V.
- Zeymer U.,
- van 't Hof A.,
- Adgey J.,
- et al.