Author + information
- Brian J. Potter, MDCM, MSc∗ ( and )
- Alexandra Bastiany, MD
- ↵∗Centre Hospitalier de l’Université de Montréal (CHUM), Cardiovascular Center–Hôtel-Dieu Hospital, Interventional Cardiology Service, 3840, rue Saint-Urbain, Pavillon Marie-Morin, 5-338, Montreal, Quebec H2W 1T8, Canada
We would like to commend LeMay et al. (1) for their work addressing the important clinical conundrum of whether to provide triple antithrombotic therapy (TATT) for patients presenting apical akinesis/dyskinesis following anterior ST-segment elevation myocardial infarction. This study takes advantage of a patient population with consistent follow-up within the same regional health system, managed at a high-volume academic center, and participating in a detailed prospective clinical database. In reviewing the report, however, we were left with a few questions; the answers to which might be of interest to other readers.
First, it is not clear to us from our reading of the paper how the propensity score was derived, for what clinical parameter propensity was determined (i.e., propensity for TATT vs. propensity for net adverse clinical events) (2), or how the propensity score was used to determine the net adverse clinical events odds ratio reported for warfarin therapy. Was this also part of the inverse-probability weighting multivariable regression analysis?
Second, and somewhat related, it would seem, from the data presented, that anticoagulation with warfarin for apical dysfunction is the exception rather than the rule at this particular institution, with fewer patients treated and with TATT patients having more apical dysfunction, worse ejection fractions, and a 3-fold higher rate of cardiogenic shock. As such, we are left to wonder whether this retrospective analysis suffers from intractable confounding, which would explain the apparent paradoxical increase in nonhemorrhagic events in this group.
Finally, we would ask the authors to comment on both the timing of adverse bleeding events prior to hospital discharge (post-procedure vs. post-initiation of warfarin) and the decision to include these in the primary analysis. It would stand to reason that most patients in this group did not have a therapeutic International Normalized Ratio until the last day or 2 of hospitalization. A “back of the envelope” calculation suggests that the exclusion of in-hospital events would make the difference in outcomes between the 2 groups considerably less dramatic. Would a landmarked analysis from the time of discharge have also achieved statistical significance?
- American College of Cardiology Foundation
- LeMay M.R.,
- Acharya S.,
- Wells G.A.,
- et al.
- Potter B.J.,
- Mansour S.,
- Lelorier J.