MULTIVARIABLE MODEL FOR PREDICTION OF RISK OF SIGNIFICANT COMPLICATION DURING DIAGNOSTIC CARDIAC-CATHETERIZATION

Although adequate data exist on the frequency of procedural-related complications during diagnostic cardiac catheterization, little information is available for the quantitative assessment of risk factors for significant complication. We analyzed 58,332 procedures in the 1990 SCA&I Registry Database in order to assess 1) the independent and cumulative significance of easily obtainable pre-procedural factors predictive of major complication and 2) the likelihood of a significant complication given the presence of these risk factors. A model was developed on a random sample of 38,888 patients and validated in the remaining 19,444 patients. Twelve variables were identified as having independent predictive behavior for significant complication with excellent agreement in the validation sample. The current model is an accurate and reliable instrument for stratifying risk of a significant complication during diagnostic catheterization. (C) 1993 Wiley-Liss, Inc.