Clinical prediction rule for atrial fibrillation after coronary artery bypass grafting
David Amar, MD*,*,
Weiji Shi, MS ,
Charles W. Hogue, Jr, MD ,
Hao Zhang, MD*,
Rod S. Passman, MD||,
Betsy Thomas, RN ,
Peter B. Bach, MD ,
Ralph Damiano, MD and
Howard T. Thaler, PhD
* Departments of Anesthesiology and Critical Care Medicine
Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center and Weill Medical College of Cornell University, New York, New York, USA
Anesthesiology
Cardiac Surgery, Washington University School of Medicine, St. Louis, Missouri, USA
|| Division of Cardiac Electrophysiology, Northwestern University School of Medicine, Chicago, Illinois, USA

View larger version (16K):
[in a new window]
|
Figure 1 Predicted probability of developing atrial fibrillation/atrial flutter (AF) by the point score. Circles = observed probability of post-coronary artery bypass graft (CABG) AF. Areas of the circles are proportional to the number of patients with each score value in the study population. Curve = predicted probability of post-CABG AF generated from a spline smooth of the logistic regression model based on point score. Vertical dash lines = cutoff of point scores of the three risk categories for AF.
|
|

View larger version (19K):
[in a new window]
|
Figure 2 Receiver-operator characteristic (ROC) curve of point score as a predictor of post-coronary artery bypass graft atrial fibrillation/atrial flutter as determined by logistic regression analysis. The area under the ROC curve for the model was 0.69. Every fifth point score value is labeled.
|
|

View larger version (7K):
[in a new window]
|
Figure 3 Nomogram for the proposed prediction rule for post-coronary artery bypass graft (CABG) atrial fibrillation/atrial flutter (AF). Point scores are computed by adding one point for each year of age, 12 points for prior history of AF, 3 points for P-wave duration >110 ms, and 10 points for low cardiac output. For any point score on the scale above the line, one can read the corresponding estimated probability of AF on the scale below the line.
|
|
|