Background Public reporting of percutaneous coronary intervention (PCI) often uses periprocedural (30 days) mortality as a surrogate for procedural quality, though it is unclear how often death is attributable to the PCI. The cause of death among patients who died within 30 days of PCI in a national healthcare system was thus evaluated. Methods and Results We identified all patients who died within 30 days of PCI in the Veterans Affairs (VA) Healthcare System from October 2005 to September 2016. Causes of death were classified through a detailed chart review using definitions from the Academic Research Consortium. Of 115 191 patients undergoing PCI during the study period, 1674 patients died within 30 days of PCI (1.5%). A detailed chart review demonstrated that the majority of patients had an undifferentiated death not definitively attributable to a single cause (981, 59%), whereas a minority had a death directly attributable to a cardiovascular cause (467, 28%). The majority of cardiovascular deaths were unrelated to the interventional procedure (335, 72%). Cardiovascular deaths were more likely to occur in the inpatient setting (95%) compared with noncardiac (89%) or undifferentiated deaths (49%, P<0.001). Conclusions A minority of deaths occurring after percutaneous revascularization were definitively due to cardiac causes, with an even smaller proportion related to the PCI. With such a small proportion of deaths directly attributable to the PCI, these data suggest that 30-day mortality may be an inappropriate metric to assess procedural quality.
Keywords: cause of death; death; mortality; percutaneous coronary intervention; veterans.