To the Editor In a recent article, Dr Miller and colleagues 1 reported an increased risk of 1-year mortality and bleeding among patients supported by an intravascular microaxial left ventricular assist device (LVAD) vs an intra-aortic balloon pump (IABP) who had all experienced an acute myocardial infarction complicated by cardiogenic shock (CS) and were undergoing percutaneous coronary intervention. This retrospective, propensity-matched analysis was performed in more than 800 patients. Similar conclusions (ie, increased mortality and major bleeding in the LVAD vs the IABP group) were previously reported by 2 large retrospective studies across the US involving 3360 propensity-matched patients with CS and mechanical circulatory support (MCS) 2 and more than 48 300 patients undergoing percutaneous coronary intervention supported by MCS. 3 Although these findings 1 are important, we must strongly emphasize that the focus on outcome among this critically ill population should be shifted from device-stratified outcome to outcome according to patient selection and case management. Because this study population was highly heterogeneous-including stratification of severity, phenotypes and related pathophysiology, complication rates, and intensive care unit management-throughout different institutions, we agree with the authors that all of the findings should be cautiously contextualized by the retrospective nature of the analyses. Indeed, no information was provided on the shock stage classification 4 of patients with CS for the LVAD vs IABP groups nor on the device selection. 1 Because LVAD allows a higher level of support, it may likely have been used in patients with more severe CS and worse organ function, or more severe CS levels may have been associated with intrinsic or congestionrelated coagulation dysfunction.The importance of a standardized heparin anticoagulation strategy-based on parallel activated partial thromboplastin time and anti-Xa assay measurements-in microaxial flow pumpsupported patients to avoid (major) bleeding complications and hemolysis has recently been emphasized. 5 In none of 3 the above-mentioned retrospective analyses was information provided on the anticoagulant drug used, the tests for monitoring the anticoagulant, nor the anticoagulation target. Moreover, major bleeding was not defined. Therefore, we believe that a detailed pathophysiological assessment-including organ function, CS severity, and a more standardized approach toward managing these devices in patients in the intensive care unitis mandatory before any firm conclusions can be drawn regarding complications associated with the device itself.