Secondary Logo

Journal Logo

Re

Multivessel Percutaneous Coronary Intervention in Patients with ST-Segment Elevation Myocardial Infarction with Cardiogenic Shock

Secemsky, Eric A.; Yeh, Robert W.

doi: 10.1097/EDE.0000000000000884
Letters
Free

Smith Center for Outcomes Research in Cardiology, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, MA, Harvard Medical School, Boston, MA

Smith Center for Outcomes Research in Cardiology, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, MA, Harvard Medical School, Boston, MA, ryeh@bidmc.harvard.edu

The authors report no conflicts of interest.

Back to Top | Article Outline

To the Editor:

A recent article by Lee et al.1 compared multivessel percutaneous coronary intervention (PCI) (i.e., PCI of lesions >50% severity before hospital discharge) versus infarct-related artery PCI alone among patients with ST-elevation myocardial infarction (MI) and cardiogenic shock. The authors employed various statistical techniques to address potential confounding and concluded that multivessel PCI was associated with lower risk of in-hospital death and repeat revascularization.

Unfortunately, this observational analysis is likely flawed by a key methodologic issue. Categorizing patients based on whether they remain alive to undergo an additional procedure may introduce “immortal time bias.”2 Because a large fraction of patients (40%) were included in the multivessel PCI group based on undergoing a staged procedure before discharge, this time period allowed for patients to not be at risk for death. The rapid separation in mortality curves observed between groups, starting before many had likely undergone a staged procedure, is suggestive of this. This bias may partially be mitigated by restricting the analysis to only those with multivessel PCI during the index procedure or by conducting a landmark analysis starting after all patients had completed their staged procedures. Additionally, this question has been addressed in the recent randomized Culprit Lesion Only PCI versus Multivessel PCI in Cardiogenic Shock (CULPRIT-SHOCK) trial,3 which demonstrated no benefit and possibly harm with multivessel PCI.

To further illustrate this bias, imagine that all patients with MI received an unopened bag of Cheetos at presentation, which they could consume at any point during hospitalization. Because only those patients who initially survived could later consume their Cheetos, Cheetos consumption would likely be associated with improved survival. By conditioning the inclusion of 1 group upon a future event, survival due to future Cheetos consumption will differentiate instantaneously, as seen in the referenced study.

We have observed the publication of similarly biased analyses despite numerous attempts to call attention to this issue.4 We hope our tongue-in-cheek analogy, which has been popularized on social media,5 might increase awareness of the potential threat of immortal time bias.

Eric A. Secemsky

Smith Center for Outcomes Research in Cardiology

Department of Medicine

Beth Israel Deaconess Medical Center

Boston, MA

Harvard Medical School

Boston, MA

Robert W. Yeh

Smith Center for Outcomes Research in Cardiology

Department of Medicine

Beth Israel Deaconess Medical Center

Boston, MA

Harvard Medical School

Boston, MA, ryeh@bidmc.harvard.edu

Back to Top | Article Outline

REFERENCES

1. Lee JM, Rhee TM, Hahn JY, et al; KAMIR Investigators. Multivessel Percutaneous coronary intervention in patients with ST-segment elevation myocardial infarction with cardiogenic shock. J Am Coll Cardiol. 2018;71:844–856.
2. Suissa S. Immortal time bias in pharmaco-epidemiology. Am J Epidemiol. 2008;167:492–499.
3. Thiele H, Akin I, Sandri M, et al; CULPRIT-SHOCK Investigators. PCI Strategies in patients with acute myocardial infarction and cardiogenic shock. N Engl J Med. 2017;377:2419–2432.
4. Secemksy EA, Yeh RW. Complete vs incomplete revascularization during percutaneous coronary intervention and improved survival - the key is immortality. JAMA Cardiol. 2018;3:443–444.
5. Yeh RW. Good Example of “Immortal Time Bias”. Available at: https://twitter.com/rwyeh/status/949479476240551939. Accessed 5 January 2018.
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.