DEVELOPMENT AND VALIDATION OF A SIMPLIFIED PREDICTIVE INDEX FOR MAJORCOMPLICATIONS IN CONTEMPORARY PERCUTANEOUS TRANSLUMINAL CORONARY ANGIOPLASTY PRACTICE
Se. Kimmel et al., DEVELOPMENT AND VALIDATION OF A SIMPLIFIED PREDICTIVE INDEX FOR MAJORCOMPLICATIONS IN CONTEMPORARY PERCUTANEOUS TRANSLUMINAL CORONARY ANGIOPLASTY PRACTICE, Journal of the American College of Cardiology, 26(4), 1995, pp. 931-938
Objectives. This study was designed to determine the preprocedural ris
k factors for major complications (emergent coronary bypass surgery, m
yocardial infarction or death) of coronary angioplasty and to derive a
nd validate a simplified index that predicts patients' a priori risk o
f complications. Background. Previous studies of risk factors for comp
lications after coronary angioplasty may not be generalizable to curre
nt, broad-based angioplasty practice. Furthermore, to our knowledge a
clinically useful predictive index has not been derived and independen
tly validated. Methods. From data collected prospectively for the Regi
stry of the Society for Cardiac Angiography and Interventions for 1992
, multivariable logistic regression was used to determine which variab
les were independently associated,vith complications in 10,622 first a
ngioplasty procedures. Stepwise regression and receiver operating char
acteristic curves then were used in this registry to develop a predict
ive index for complications that was validated using 5,250 first angio
plasty procedures in the 1993 registry. Results. Predictors of major c
omplications were multivessel disease, unstable angina, recent myocard
ial infarction, type C lesion or left main angioplasty, shock, age, ge
ographic region and absence of previous coronary bypass surgery. The d
erived predictive index consisted of the first six of these variables
plus aortic valve disease and classified patients into four risk group
s: low (1.3% complications), moderate (2.8%), high (12.7%) and very hi
gh (29.7%) risk. This index demonstrated consistent reliability and di
scriminatory ability when applied to the 1993 data. Conclusions. Predi
ctors of major complications identified in selected populations also a
pply currently in broad-based practice. From these variables, a predic
tive index can stratify patients into risk groups before angioplasty,
thus aiding in risk assessment, resource allocation and risk adjustmen
t.