Background: The goal of this study was to compare angiographic interpretation of coronary arteriograms by sites in community practice versus those made by a centralized angiographic core laboratory.
Methods and results: The study population consisted of 2013 American College of Cardiology-National Cardiovascular Data Registry (ACC-NCDR) records with 2- and 3- vessel coronary disease from 54 sites in 2004 to 2007. The primary analysis compared Registry (NCDR)-defined 2- and 3-vessel disease versus those from an angiographic core laboratory analysis. Vessel-level kappa coefficients suggested moderate agreement between NCDR and core laboratory analysis, ranging from kappa=0.39 (95% confidence intervals, 0.32-0.45) for the left anterior descending artery to kappa=0.59 (95% confidence intervals, 0.55-0.64) for the right coronary artery. Overall, 6.3% (n=127 out of 2013) of those patients identified with multivessel disease at NCDR sites had had 0- or 1-vessel disease by core laboratory reading. There was no directional bias with regard to overcall, that is, 12.3% of cases read as 3-vessel disease by the sites were read as <3-vessel disease by the core laboratory, and 13.9% of core laboratory 3-vessel cases were read as <3-vessel by the sites. For a subset of patients with left main coronary disease, registry overcall was not linked to increased rates of mortality or myocardial infarction.
Conclusions: There was only modest agreement between angiographic readings in clinical practice and those from an independent core laboratory. Further study will be needed because the implications for patient management are uncertain.
Keywords: angiography; catheterization; percutaneous coronary intervention; program evaluation.