Risk scores for 30-day mortality after percutaneous coronary intervention: New insights into causes and risk of death

Objective: To determine the causes and risk of death after percutaneous coronary interventions (PCIs) and to compare the discriminatory ability of the New York State Risk Score (NYSRS) with the Mayo Clinic Risk Score (MCRS). Patients and Methods: We studied in-hospital and 30-day mortality after PCI in 4898 patients treated at Mayo Clinic in Rochester, Minnesota, from January 1, 2007, through December 31, 2010, to validate the NYSRS equation with recalibrated predicted probabilities of death. Results: Of the 4898 patients studied, 93 (1.9%) died during the index hospitalization, and 36 (0.7%) died within 30 days after discharge. For the in-hospital and 30-day mortality, respectively, the area under the receiver operating characteristic curve was 0.92 and 0.88 for the NYSRS and 0.93 and 0.90 for the MCRS, indicating excellent discrimination. The NYSRS model underpredicted event rates when applied in Mayo Clinic data (2.6% observed [127 of 4898 patients] vs 2.3% predicted [114 of 4898 patients]), even after recalibration. The instantaneous hazard over time revealed the highest risk of death in the first 3 days after PCI (daily probability, >0.2%), declined to 0.1% until about day 12, and then decreased below 0.1%. Cardiac causes (mainly myocardial infarction) dominated in the first week (83 of 85 deaths [97.6%]) and then decreased to 59.5% (25 of 42 deaths) between 8 and 30 days after PCI. Conclusion: The discriminatory ability of the NYSRS and the MCRS for in-hospital and 30-day mortality after PCI is roughly interchangeable. The risk of death is highest during the first 2 weeks and is dominated by cardiac causes of death.