Surgical candidacy and selection biases in nonemergent left main stenting: implications for observational studies

Abstract

Objectives: This study sought to characterize reasons for surgical ineligibility in patients undergoing nonemergent unprotected left main (ULM) percutaneous coronary intervention (PCI) and to assess the potential for these reasons to confound comparative effectiveness studies of coronary revascularization.

Background: Although both PCI and coronary artery bypass graft surgery are treatments for ULM disease, some patients are not eligible for both treatments, which may result in treatment selection biases.

Methods: In 101 consecutive patients undergoing nonemergent ULM PCI, mixed methods were used to determine the prevalence of treatment selection dictated by surgical ineligibility and to identify the reasons cited for avoiding coronary artery bypass graft surgery. We then determined whether these reasons were captured by the ACC-NCDR (American College of Cardiology-National Cardiovascular Data Registry) Cath-PCI dataset to assess the ability of this registry to account for biases in treatment selection. Finally, the association of surgical eligibility with long-term outcomes after ULM PCI was assessed.

Results: Treatment selection was dictated by surgical ineligibility in over half the ULM PCI cohort with the majority having reasons for ineligibility not captured by the ACC-NCDR. Surgical ineligibility was a significant predictor of mortality after adjustment for Society of Thoracic Surgeons (hazard ratio [HR]: 5.4, 95% confidence interval [CI]: 1.2 to 25), EuroSCORE (European System for Cardiac Operative Risk Evaluation) (HR: 5.9, 95% CI: 1.3 to 27), or NCDR mortality scores (HR: 6.2, 95% CI: 1.4 to 27).

Conclusions: Surgical ineligibility dictating treatment selection is common in patients undergoing nonemergent ULM PCI, occurs on the basis of risk factors not captured by the ACC-NCDR, and is independently associated with worse long-term outcomes after adjusting for standard risk scores.