BACKGROUND: We have shown previously that either echocardiographic indices of diastolic dysfunction or increased preoperative brain natriuretic peptide (BNP) predict postoperative atrial fibrillation (POAF). Because these 2 predictors of POAF have not been evaluated together, our goal was to further elucidate their concurrent role in patients undergoing noncardiac thoracic surgery.METHODS: We retrospectively identified 191 patients who had a preoperative transthoracic echocardiogram and serum BNP level collected as part of routine care before major lung or esophageal resection. Clinical and echocardiographic data were compared between patients who did or did not develop POAF (>5 minutes), and prognostic factors for POAF were identified.RESULTS: Univariate associations with POAF (41 of 191; 22% patients) included older age (P = .04), male sex (P = .01), hypertension (P = .03), increased body mass index (P = .01), and prolonged transmitral flow deceleration time (P < .0001), whereas BNP was not statistically significant (P = .07). Stepwise logistic regression analysis showed that both increasing transmitral flow deceleration time (continuous data log base 2 transformed; odds ratio, 16.05; 95% confidence interval, 3.74-68.96; P = .0002) and left atrial diastolic volume index (continuous data log base 2 transformed; odds ratio, 3.29; 95% confidence interval, 1.22-8.91; P were independent risk factors of POAF (area under the receiver operating characteristic curve = 0.73). There was no significant interaction between BNP and the 2 independent variables (P = .60, and P = .90), respectively.CONCLUSIONS: In a cohort of patients who had echocardiography and BNP measurements before undergoing major thoracic surgery, this study showed that when evaluated together greater preoperative left atrial diastolic volume index and transmitral flow deceleration time but not BNP levels were independent predictors for POAF.