|
|
||||||||
a Department of Cardio-Thoracic Surgery, Catharina Hospital, Eindhoven, The Netherlands
b Department of Education and Research, Catharina Hospital, Eindhoven, The Netherlands
c Center of Research on Psychology in Somatic Diseases, Department of Medical Psychology, Tilburg University, Tilburg, The Netherlands
d Department of Cardiothoracic Surgery, Cardiovascular Research Institute Maastricht—CARIM, Maastricht University Medical Center—MUMC, Maastricht, The Netherlands
Accepted for publication March 26, 2010.
* Address correspondence to Dr Soliman Hamad, Department of Cardiothoracic Surgery, Catharina Hospital, Michelangelolaan 2, Postbus 1350, 5602 ZA Eindhoven, The Netherlands (Email: aasmsn{at}cze.nl).
| ADULT CARDIAC SURGERY:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
| Abstract |
|---|
|
|
|---|
Methods: Data of patients who consecutively underwent isolated CABG between January 2003 and December 2007 were prospectively collected. The analysis included 5098 patients with preoperative sinus rhythm and no history of atrial fibrillation. Logistic regression analysis for early mortality and Cox regression analysis for late mortality were performed. Propensity score matching was performed to eliminate the effect of confounders.
Results: Median follow-up was 2.5 years. POAF was documented in 1122 patients (22.0%). Early mortality was more frequent in POAF patients (3.1%) vs non-POAF patients (1.6%, p = 0.002), but multivariate logistic regression analysis could not identify POAF as an independent predictor of early mortality (p = 0.169). This outcome did not change after adjusting for quintiles of the propensity score of POAF (p = 0.100). Multivariate Cox proportional hazard analyses demonstrated POAF was an independent predictor of overall and late mortality with hazard ratios of 1.35 (p = 0.012 and p = 0.039, respectively). Analyses after propensity score matching showed that patients with POAF had similar hazard ratios of 1.36 for overall mortality and 1.34 for late mortality (p = 0.009 and p = 0.042, respectively).
Conclusions: POAF is an independent predictor of overall and late mortality after isolated CABG but not of early mortality.
New-onset postoperative atrial fibrillation (POAF) is observed in about 20% to 40% of patients undergoing coronary artery bypass grafting (CABG) [1–12]. Postoperative AF after cardiac operations is associated with postoperative morbidities such as cerebrovascular accidents (CVA), infections (eg, septicemia, pneumonia and mediastinitis), and renal failure [1–6, 13, 14]. Previous studies have suggested that POAF after CABG is related to early [3, 4, 6, 9] and late mortality [4, 9,11,12]. Many of the predictors of POAF, of which age seems to be the strongest [1–3, 6, 7, 13, 14], also relate to early and late mortality after CABG [4, 5, 15, 16].
This study used multivariate analyses and propensity score matching to investigate if POAF is an independent predictor of early and late mortality after isolated CABG.
| Material and Methods |
|---|
|
|
|---|
Patients
The study included patients who underwent isolated CABG in a single center in the Netherlands (Catharina Hospital, Eindhoven) between January 2003 and December 2007. Demographic data, known risk factors for mortality, and in-hospital complications were prospectively collected in a computerized database. Only patients with documented preoperative sinus rhythm, without a history of AF were included. Follow-up mortality data were gathered using databases of health insurance companies and by information requests to general practitioners or municipal authorities.
Definitions
Early mortality was defined as death within 30 days postoperatively or death at anytime if the patient did not leave the hospital or a transfer to a tertiary hospital alive. Late mortality was defined as death from any cause that occurred at any time after hospital discharge. POAF was defined as any evidence of new AF by electrocardiography or continuous monitoring, which lasted at least 30 minutes, during the postoperative period in our hospital.
Operative Technique and Postoperative Policy
Our policy was to continue the use of β-blocking agents preoperatively until the day of the operation. All patients received short-acting anesthetic drugs to facilitate early extubation. Normothermic extracorporeal circulation was performed using nonpulsatile flow. Intermittent cold crystalloid cardioplegia (St Thomas solution) or intermittent warm blood cardioplegia was used to induce and maintain cardioplegic arrest, according to the surgeon's preference. Eleven percent of all operations were performed without the use of extracorporal circulation.
Postoperatively, all patients without a contraindication for β-blocking agents received metoprolol as AF prophylaxis, and one or more electrical cardioversions were done when indicated. Rhythm was continuously registered for at least 48 hours postoperatively. Afterwards, registration was performed at least 3 times a day or continuously in case of dysrhythmia. If AF persisted for more than 48 hours, an anticoagulant was started. The referring cardiologist determined AF treatment regimen after discharge, according to the guidelines of the American College of Cardiology/American Heart Association/European Society of Cardiology [17].
Statistical Analysis
All statistical analyses were done with SPSS 15.0 software (SPSS Inc, Chicago, IL). Categoric variables were compared using the
2 test or the Fisher's exact test, if appropriate, and are presented as percentages. Continuous variables were compared using the t test and are presented as means ± standard deviations or median with range or interquartile range (IQR).
Univariate and multivariate logistic regression analyses were performed to evaluate predictors of early mortality. In addition, a separate analysis with a model stratified by quintiles of the propensity score of POAF was performed [18, 19]. This propensity score was calculated with a multivariate logistic regression model in which univariate predictors of POAF with a value of p < 0.10 were included. The area under the receiver operating characteristics curve of the multivariate model of predicting early mortality was calculated to evaluate the discriminatory capacity of this model.
Univariate and multivariate Cox proportional hazard analyses were performed to investigate predictors of late mortality. A model stratified by quintiles of the propensity score of POAF was used for a separate analysis of late mortality. Variables with a value of p < 0.10 in univariate analyses were included in the multivariate models.
Long-term survival curves were produced according to the Kaplan-Meier method, and differences between groups were compared with the log-rank test. Two Cox proportional hazard models were used to construct adjusted Kaplan-Meier survival curves for patients with and without POAF. In the first model, all independent predictors of overall mortality were included, and adjusted survival was calculated for the POAF and non-POAF groups. In the second model, survival was adjusted for quintiles of the propensity score of POAF. For both models hazard ratios (HR) for overall death after CABG were calculated. HRs and odds ratios (OR) with 95% confidence intervals (CI) and p values are reported. A value of p < 0.05 was used for all tests to indicate statistical significance.
| Results |
|---|
|
|
|---|
Characteristics of POAF and non-POAF patients are summarized in Table 1. Patients with POAF were older, more likely to have a preoperative CVA and hypertension, and to have lower mean creatinine clearance. Patients with POAF more often underwent CABG with the use of extracorporeal circulation, had a reoperation for bleeding, and received more perioperative blood transfusion. Patients with POAF presented more often with perioperative infarction, mediastinitis, postoperative in-hospital pneumonia, and CVA. A significant difference in New York Heart Association (NYHA) class and number of grafts was also present.
|
|
|
|
|
|
| Comment |
|---|
|
|
|---|
Early Mortality
Although the incidence of early mortality was almost twofold in the POAF group, POAF was not identified as an independent risk factor for early mortality in multivariate and propensity score adjusted analyses. The probability of developing POAF is strongly dependent of age [1–3, 6, 7, 13, 14]. This obviously explains why the mean age of the POAF patients is 4.5 years higher than the age in the non-POAF group. Age is also strongly associated with early mortality [5, 15, 16]. Correction for age (and other risk factors for early mortality) shows that POAF is not an independent risk factor of early death. Kalavrouziotis and coworkers [5] concluded the same in a large study on POAF in cardiac surgery patients after multivariate analysis and propensity score matching. In the study of Villareal and colleagues [4], POAF was a significant predictor of early death after adjusting for risk factors, but as suggested by Levy and colleagues [20], adjusting for age as a dichotomous variable might have been inadequate because age is such a powerful predictor of early death and POAF. A large number of studies [3, 5, 9, 12, 14] show significantly higher incidences of early death in patients with POAF after CABG or cardiac operations, but none of these studies identified POAF as an independent predictor of early mortality.
Overall and Late Mortality
A few other studies have evaluated the effect of POAF on long-term survival after CABG [4, 9, 11, 12]. Villareal and colleagues [4] reported an increased risk of overall and late death in patients with POAF (adjusted HR of 1.5 and 1.4, respectively) in a study of 6475 patients after first time isolated CABG. However, age was entered as a dichotomous variable in their multivariate model, which could have resulted in underestimation of the effect of age on mortality, thereby leaving a possible residual confounding effect. In the same study, a case matched analysis resulted in a HR of 3.4 for POAF regarding long-term mortality. However, only 20% of all cases of POAF were used in that particular analysis. The investigated patient group was not representative for the whole POAF group and therefore interpretation of the mentioned HR should be done cautiously.
Mariscalco and colleagues [9] calculated an adjusted HR of 2.13 for POAF on all-cause mortality after isolated CABG in a study in 1832 patients. Although data were prospectively collected, 8% of the patients were lost to follow-up, which may have influenced the study outcome. In their study, only patients with a history of permanent AF were excluded, which implies that potentially, patients with preoperative paroxysmal and persistent AF were included. By including such patients, it might be possible that part of the effect of POAF on survival could be attributed to preoperative AF. Recently, Mariscalco and Engström [11] described 7621 patients who underwent a CABG procedure in whom multivariate analysis found a HR of 1.22 for POAF. In contrast to our study, patients operated on without the use of extracorporal circulation were not included. They did not include gender in the multivariate model; gender was an independent risk factor in different studies [21, 22] as well as in our study. Not including gender in the analysis could have biased the results.
A recent report by Ahlsson and colleagues [12] indicated POAF was an age-independent risk factor for late mortality after CABG (HR, 1.56) in a study with 1419 patients. Only POAF, age, diabetes, and lower left ventricular ejection fraction were significant predictors of late death in multivariate analysis in their study.
We were not able to retrieve the causes of death in our study, which is an important limitation. The study of Mariscalco and colleagues [9] showed that embolic events might play an important role in postoperative death in patients with POAF. This is consistent with our findings of a significantly higher incidence of postoperative in-hospital CVA in patients with POAF.
We did not separately study the effect of therapeutic interventions for POAF on mortality. The effect of those interventions for POAF on survival might be small, because almost every patient receives β-blocking agents postoperatively, and most of the patients with POAF return to sinus rhythm [13, 23] and can be weaned from antiarrhythmic drugs.
The in-hospital follow-up period was relatively short, with a median of 5 days (IQR, 3 to 6 days). Onset of POAF after discharge was not registered, and therefore, the incidence of POAF may have been underestimated. However, most of the first episodes of POAF happen within the first 4 to 6 days after cardiac operations, with an incidence peak on the second day [1, 13, 14, 24]. Duration of POAF was not documented in our database, and therefore, its effect on mortality could not be analyzed.
As in all observational studies, there is a risk for residual confounding of measured and unmeasured variables. By adjusting for quintiles of the propensity score of POAF, we attempted to minimize the effect of residual confounders. Multivariate analysis and propensity score adjusted methods both produced very similar results for both early and late survival. Furthermore, the area under the receiver operating characteristic curve of the multivariate logistic regression model for predicting early mortality indicates an excellent discriminatory capacity of this model.
Although postoperative use of β-blocking agents, sotalol and amiodarone decrease the incidence of POAF [25, 26], it remains to be investigated whether the use of these agents improves overall survival after CABG. More aggressive surgical and preventive measurements in high-risk patients might possibly favor long-term survival after CABG, but well-performed prospective studies to support this theory are currently not available.
In conclusion, postoperative atrial fibrillation is an independent predictor of overall and late mortality after CABG but was not identified as an independent predictor of early mortality.
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
G. H. Almassi, S. A. Pecsi, J. F. Collins, A. L. Shroyer, M. A. Zenati, and F. L. Grover Predictors and impact of postoperative atrial fibrillation on patients' outcomes: A report from the Randomized On Versus Off Bypass trial J. Thorac. Cardiovasc. Surg., January 1, 2012; 143(1): 93 - 102. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Bramer, A. H. M. van Straten, M. A. Soliman Hamad, K. C. van den Broek, J. G. Maessen, and E. Berreklouw New-Onset Postoperative Atrial Fibrillation Predicts Late Mortality After Mitral Valve Surgery Ann. Thorac. Surg., December 1, 2011; 92(6): 2091 - 2096. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Bramer, A. H. M. van Straten, M. A. Soliman Hamad, E. Berreklouw, K. C. van den Broek, and J. G. Maessen Body mass index predicts new-onset atrial fibrillation after cardiac surgery Eur J Cardiothorac Surg, November 1, 2011; 40(5): 1185 - 1190. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Polat, B. Mert, E. B. Polat, C. Yucel, and V. Erentug Impact of Postoperative Atrial Fibrillation After CABG Ann. Thorac. Surg., June 1, 2011; 91(6): 2024 - 2024. [Full Text] [PDF] |
||||
![]() |
S. Bramer, A. H. M. van Straten, M. A. Soliman Hamad, E. Berreklouw, E. J. Martens, and J. G. Maessen Reply Ann. Thorac. Surg., June 1, 2011; 91(6): 2024 - 2025. [Full Text] [PDF] |
||||
![]() |
T. B. Albacker Invited Commentary Ann. Thorac. Surg., August 1, 2010; 90(2): 449 - 450. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |