In Response We appreciate comments by Poelaert1 regarding our article titled “Is Compliance With Surgical Care Improvement Project Cardiac (SCIP-Card-2) Measures for Perioperative β-Blockers Associated With Reduced Incidence of Mortality and Cardiovascular-Related Critical Quality Indicators After Noncardiac Surgery?”2 Poelaert1 suggests that our findings show that “initiating beta-blockers in the preoperative setting of noncardiac surgery may increase the risk of cardiac failure.” Importantly, a limitation of our study was the inability to explore the potential effects of initiating β-blockers in the preoperative setting, and whether their initiation was associated with a higher risk for cardiac failure or all-cause mortality. To be precise, our retrospective finding was that, compared to continuation, “perioperative withdrawal of β-blockers was associated with increased risk for mortality within 48 hours after noncardiac surgery and with decreased risk for need of vasopressor during the early postoperative period and a shorter stay in postanesthesia care unit.” Indeed, we had no specific data regarding how many individuals among the 343,533 cases without β-blocker therapy before admission then received β-blocker therapy during the perioperative period. Hence, we restricted our Surgical Care Improvement Project Cardiac-2 definition of preoperative/perioperative β-blocker therapy, as well as analysis and conclusions, to the patient population receiving β-blocker therapy before admission. Poelaert1 also asserts that our study did not explore reasons for discontinuation of β-blocker therapy and, thus, is limited in terms of its findings and generalizability. In particular, he highlights the importance of considering type of β-blocker therapy when it comes to studying their effects on outcome after noncardiac surgery. In support of this statement, Poelaert1 claims that, due to their intrinsic sympathomimetic activity (ISA), continuation of some earlier generation β-blockers such as metoprolol can be potentially beneficial in the perioperative setting, while preoperative discontinuation of third-generation β-blockers such as labetalol and carvedilol, which lack ISA and have α1-antagonistic effect, may reduce risk of adverse events and mortality. Indeed, we agree that when it comes to variation in β-blocker therapy and studying potential effects on perioperative outcomes, these differentiating pharmacokinetic (cytochrome P450 2D6-dependent metabolism) and pharmacodynamic (relative β1 receptor selectivity, β1/β2 receptor affinity, and ISA) properties should be taken into consideration.3 Unfortunately, given that β-blocker therapy was a binary variable in our data source, we could not ascertain the various β-blocker therapies, and whether this factor was associated with their discontinuation in the postoperative period. Nevertheless, in our study, >85% of cases had a history of hypertension, and >30% had a history of coronary artery disease. It is highly likely that the majority of these individuals received metoprolol therapy given the past and current recommendations of guidelines for β-blocker therapy4 and the cost-effective prescribing of β-blocker therapy in the United States.5 Therefore, we believe that it is reasonable to consider the observed relationship between β-blocker withdrawal and mortality and cardiovascular-related critical quality indicators in our study as potentially related to withdrawal of metoprolol therapy. Of note, metoprolol does not have ISA, but, as previously acknowledged, its pharmacokinetic and pharmacodynamic properties may significantly influence its efficacy and safety relative to perioperative cardiovascular morbidity and mortality.3 Finally, Poelaert1 highlights the lack of information on type of surgery in our study. As clarification, we detailed extensively on the type of surgery relative to factors relevant to the study focus, including classifying risk of surgery type according to Revised Cardiac Risk Index suggested by Lee et al6 and modified by Kristensen et al.7 As described in the study, we used this composite risk index in our statistical modeling, both to account for surgical and patient characteristics while avoiding overfitting due to the limited number of perioperative events and to enable assessment of interaction between demographic and clinical risk factors (including type of surgery) and patterns of β-blocker use relative to all-cause mortality and cardiovascular complications. We thank Poelaert1 for his interesting comments, and while beyond the scope of our study, we agree that further research is warranted to explore how differences among types of β-blocker therapy, the de novo initiation of perioperative β-blocker therapy, and withdrawal or continuation of other cardiovascular medications are associated with mortality and cardiovascular-related risk indicators in noncardiac surgery patient cohorts. Miklos D. Kertai, MD, PhDDivision of Cardiothoracic AnesthesiologyDepartment of AnesthesiologyVanderbilt University Medical CenterNashville, Tennessee[email protected] Mary Cooter, MScDivision of Cardiothoracic Anesthesiology and Critical Care MedicineDepartment of AnesthesiologyDuke University Medical CenterDurham, North Carolina Richard J. Pollard, MDDepartment of Anesthesia Critical Care and Pain ManagementBeth Israel Deaconess Medical CenterHarvard Medical SchoolBoston, Massachusetts Mark Stafford-Smith, MD, FRCPCDivision of Cardiothoracic Anesthesiology and Critical Care MedicineDepartment of AnesthesiologyDuke University Medical CenterDurham, North Carolina