Jiwei Gu1, Jan J Andreasen2, Jacob Melgaard3, Søren Lundbye-Christensen4, John Hansen3, Erik B Schmidt5, Kristinn Thorsteinsson6, Claus Graff3. 1. Department of Cardiothoracic Surgery, Aalborg University Hospital, Aalborg, Denmark; Department of Clinical Medicine, Aalborg University, Aalborg, Denmark; Department of Cardiovascular Surgery, Heart Centre of General Hospital, Ningxia Medical University, Yinchuan, Ningxia, Peoples Republic of China; Atrial Fibrillation Study Group, Aalborg University Hospital, Aalborg, Denmark. Electronic address: g.jiwei@rn.dk. 2. Department of Cardiothoracic Surgery, Aalborg University Hospital, Aalborg, Denmark; Department of Clinical Medicine, Aalborg University, Aalborg, Denmark; Atrial Fibrillation Study Group, Aalborg University Hospital, Aalborg, Denmark. 3. Department of Health Science and Technology, Aalborg University, Aalborg, Denmark. 4. Atrial Fibrillation Study Group, Aalborg University Hospital, Aalborg, Denmark; Unit of Clinical Biostatistics and Bioinformatics, Aalborg University Hospital, Aalborg, Denmark. 5. Department of Clinical Medicine, Aalborg University, Aalborg, Denmark; Atrial Fibrillation Study Group, Aalborg University Hospital, Aalborg, Denmark; Department of Cardiology, Aalborg University Hospital, Aalborg, Denmark. 6. Department of Cardiothoracic Surgery, Aalborg University Hospital, Aalborg, Denmark; Department of Clinical Medicine, Aalborg University, Aalborg, Denmark.
Abstract
OBJECTIVE: To investigate if electrocardiogram (ECG) markers from routine preoperative ECGs can be used in combination with clinical data to predict new-onset postoperative atrial fibrillation (POAF) following cardiac surgery. DESIGN: Retrospective observational case-control study. SETTING: Single-center university hospital. PARTICIPANTS: One hundred consecutive adult patients (50 POAF, 50 without POAF) who underwent coronary artery bypass grafting, valve surgery, or combinations. INTERVENTIONS: Retrospective review of medical records and registration of POAF. MEASUREMENTS AND MAIN RESULTS: Clinical data and demographics were retrieved from the Western Denmark Heart Registry and patient records. Paper tracings of preoperative ECGs were collected from patient records, and ECG measurements were read by two independent readers blinded to outcome. A subset of four clinical variables (age, gender, body mass index, and type of surgery) were selected to form a multivariate clinical prediction model for POAF and five ECG variables (QRS duration, PR interval, P-wave duration, left atrial enlargement, and left ventricular hypertrophy) were used in a multivariate ECG model. Adding ECG variables to the clinical prediction model significantly improved the area under the receiver operating characteristic curve from 0.54 to 0.67 (with cross-validation). The best predictive model for POAF was a combined clinical and ECG model with the following four variables: age, PR-interval, QRS duration, and left atrial enlargement. CONCLUSION: ECG markers obtained from a routine preoperative ECG may be helpful in predicting new-onset POAF in patients undergoing cardiac surgery.
OBJECTIVE: To investigate if electrocardiogram (ECG) markers from routine preoperative ECGs can be used in combination with clinical data to predict new-onset postoperative atrial fibrillation (POAF) following cardiac surgery. DESIGN: Retrospective observational case-control study. SETTING: Single-center university hospital. PARTICIPANTS: One hundred consecutive adult patients (50 POAF, 50 without POAF) who underwent coronary artery bypass grafting, valve surgery, or combinations. INTERVENTIONS: Retrospective review of medical records and registration of POAF. MEASUREMENTS AND MAIN RESULTS: Clinical data and demographics were retrieved from the Western Denmark Heart Registry and patient records. Paper tracings of preoperative ECGs were collected from patient records, and ECG measurements were read by two independent readers blinded to outcome. A subset of four clinical variables (age, gender, body mass index, and type of surgery) were selected to form a multivariate clinical prediction model for POAF and five ECG variables (QRS duration, PR interval, P-wave duration, left atrial enlargement, and left ventricular hypertrophy) were used in a multivariate ECG model. Adding ECG variables to the clinical prediction model significantly improved the area under the receiver operating characteristic curve from 0.54 to 0.67 (with cross-validation). The best predictive model for POAF was a combined clinical and ECG model with the following four variables: age, PR-interval, QRS duration, and left atrial enlargement. CONCLUSION: ECG markers obtained from a routine preoperative ECG may be helpful in predicting new-onset POAF in patients undergoing cardiac surgery.