Atrial fibrillation is common after cardiac surgery and results in prolonged hospitalization. The purpose of this study was to evaluate the role of postoperative vasopressor use as a predictor of postoperative atrial fibrillation.
Study population consisted of 199 patients who underwent cardiac surgeries. The patients were divided in two groups and compared based on whether or not they had received vasopressor agents postoperatively. The baseline and clinical variables were analyzed as the possible predictors of postoperative atrial fibrillation using the multivariate stepwise logistic regression analysis and the Cox regression analysis.
Postoperative atrial fibrillation occurred in 49 of the 127 patients (39%) with vasopressor use and 10 of the 72 patients (14%) without vasopressor use (p<0.01). By univariate analysis, vasopressor use, left ventricular hypertrophy, type of surgery, diabetes, advanced age, left atrial diameter, and cross clamp time were predictors of development of postoperative atrial fibrillation. In multivariate logistic regression analysis, use of vasopressor agents was an independent predictor of postoperative atrial fibrillation (odds ratio 3.35, 95% CI 1.38-8.12, p=0.016). The other independent predictors of postoperative atrial fibrillation were valve surgery (odds ratio 2.88, 95% CI 1.31-6.35, p=0.002) and advanced age (odds ratio 10.73, 95% CI 10.37-11.10, p=0.0001). By Cox regression analysis, vasopressor use, valve surgery, advanced age, and ventricular pacing were predictors of time to atrial fibrillation. Among various vasopressor drugs, use of those with predominantly beta-adrenergic affinity was associated with higher incidence of postoperative atrial fibrillation (dopamine 44%, dobutamine 41% vs. phenylepherine 20%, p=0.001).CONCLUSIONS: Cardiac surgery patients who receive vasopressors postoperatively are more likely to develop postoperative atrial fibrillation, and vasopressor use is an independent predictor of atrial fibrillation.
Judicious postoperative use of vasopressors may reduce the incidence of atrial fibrillation after cardiac surgery.
Variables in the Final Logistic Regression ModelVariablesBP valueOdds Ratio95% Confidence IntervalVasopressor use−1.0940.0163.351.38-8.12Valve surgery−1.2450.0022.881.31-6.35Age0.0700.000110.7310.37-11.10
I.A. Khan, None.