| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;84:434-443
© 2007 The Society of Thoracic Surgeons
a Division of Cardiovascular Surgery, Mayo Clinic College of Medicine, Rochester, Minnesota
b Department of Biostatistics, Mayo Clinic College of Medicine, Rochester, Minnesota
Accepted for publication April 11, 2007.
* Address correspondence to Dr Ngaage, Department of Cardiothoracic Surgery, Cardiothoracic Centre, Castle Hill Hospital, Cottingham, East Yorkshire, HU15 6JQ, United Kingdom (Email: dngaage{at}yahoo.com).
Presented at the Fifty-third Annual Meeting of the Southern Thoracic Surgical Association, Tucson, AZ, Nov 8–11, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
Methods: From 1993 through 2002, 36% of 2,821patients with mitral regurgitation had preexisting AF. A cohort of these was matched with controls in sinus rhythm (SR) for age, gender, and ejection fraction. Follow-up was by questionnaire. Outcomes were compared between 231AF and 229 SR patients, and patients with different types of preoperative AF.
Results: Patients with preoperative AF were more symptomatic and frequently had cardiomegaly, heart failure, and higher mean pulmonary artery systolic pressure. Operative mortality was higher for AF patients (2% vs 0, p = 0.05). More AF patients had late adverse cardiac events and stroke (63% vs 31%, p < 0.0001). Five- and ten-year survival was, respectively, 95% and 88% for SR patients compared with 90% and 70% (p = 0.01) for the AF group. By multivariate analysis, preoperative AF was not a predictor of long-term survival but was an independent risk factor for late adverse cardiac events and stroke.
Conclusions: Preoperative AF is a marker for increased surgical risk of mitral regurgitation repair, and a risk factor for late adverse cardiac events and stroke. Although the independent contribution of AF to late survival is uncertain, preoperative AF increases postoperative morbidity independently; therefore, corrective intervention would be expected to benefit patients in this regard.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
Atrial fibrillation has been identified as a marker of severe cardiac disease and a risk factor for decreased long-term survival in nonsurgical series [4–6]. Quader and associates [7] have also reported excess late mortality after coronary artery bypass grafting in patients with preoperative AF. In order to identify the multiple impacts of preoperative AF on clinical outcomes after cardiac surgery, we performed a matched cohort comparison between patients with preoperative AF and those in preoperative sinus rhythm (SR) undergoing various cardiac operations. We have previously reported the adverse effects of preexisting AF on clinical results after aortic valve replacement [8], and coronary artery bypass grafting [9].
The present study compares clinical outcomes after repair of mitral valve regurgitation between patients with preoperative AF and matched controls in SR. Our objectives were to determine the prevalence of AF in patients undergoing repair of nonrheumatic mitral regurgitation, and to assess the influence of preoperative AF on early morbidity and mortality, late adverse cardiac and cerebrovascular events, and long-term survival.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
From published reports, we calculated a sample size of 200 per group to detect a 10% difference at 90% power. We selected 250 patients with preoperative AF using the random option of the survey select procedure of the SAS statistical analysis system (SAS Institute Inc, Cary, NC), and identified the best matched control cohort in preoperative SR based on age, gender, and left ventricular ejection fraction using the Greedy method [10]. The selection process is represented in Figure 1.
|
Definitions
The international consensus on nomenclature and classification [11] recently classified AF into initial event, paroxysmal, persistent, and permanent types. Data collection during the study period utilized the simpler schema proposed by Cox [12] and widely used in the surgical literature [13]. However, the similarity in the arrhythmia patterns that define the different types of AF in both Cox’s and the international classification, enabled reclassification. In this study, we used the following classifications: (1) paroxysmal or persistent atrial fibrillation is preoperative AF that is recurrent; (2) permanent atrial fibrillation is preoperative AF that is present at all times.
Statistical Analysis
The primary endpoints were (1) late mortality and (2) adverse cardiac events and stroke. Adverse cardiac events were defined as cardiac death, myocardial infarction, heart failure, and heart rhythm-related intervention. Secondary endpoints included postoperative length of hospital stay, cardiac-related hospital readmission, and noncerebral bleeding-thromboembolic complications.
Categoric variables are expressed as percentages and continuous variables as mean ± SD. Univariate analysis was performed using the 
2 test for categoric variables and the Wilcoxon rank test for continuous variables. The Cox proportional hazards models and logistic regression were utilized for multivariable analysis, and long-term outcomes were analyzed using the Kaplan-Meier method. Risk factors associated with each endpoint of interest were identified by the stepwise model selection procedure. If the primary variable of interest (preoperative AF versus preoperative SR) was selected, then this model is reported. Otherwise, if the preoperative AF versus preoperative SR grouping variable was not chosen, it was forced into the final model provided by the stepwise procedure. The variables included in the stepwise logistic regression and multivariable models are listed in Appendix 1. The SAS statistical analysis system (SAS Institute Inc) was used for data analysis. Statistical significance was defined as p less than 0.05 with a 2-tailed test.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
|
|
|
0.0001), subsequent heart rhythm-related intervention (Fig 4) in the form of cardioversion and (or) permanent pacemaker insertion (24% vs 9%, p = 0.002), and stroke (12% vs 5%, p = 0.03) were more frequently observed in the AF group. Notably, preoperative AF (odds ratio = 3.12, 95% confidence interval 1.99 to 4.97, p < 0.0001) was an independent risk factor for adverse cardiac events and stroke, as was cardiomegaly (odds ratio = 2.09, 95% confidence interval 1.32 to 3.30, p = 0.002).
|
|
Improvement in symptoms and change in left ventricular ejection fraction were similar between both groups of patients at late follow-up (mean 5 ± 3 years). Both groups reported a significant improvement in the New York Heart Association (NYHA) functional class. At the time of surgery, 66% of preoperative AF and 45% of preoperative SR patients were in NYHA functional class III/IV compared with 4% and 7%, respectively, at follow-up. Similarly, the mean ejection fraction of both groups decreased over time (see Table 2).
Principal Findings: Impact of Clinical Type and Duration of Preoperative Atrial Fibrillation
The clinical type of preoperative AF was known in 182 patients (79%). The clinical outcomes for the different clinical types of preoperative AF are shown in Table 3. Patients with permanent AF had a 170% increased risk of late mortality (risk ratio = 2.7, p = 0.02) compared with those with paroxysmal or persistent AF. The three-year and six-year survival were, respectively, 93% and 79% for patients with permanent AF compared with 97% and 91% for those with paroxysmal or persistent AF (Fig 5).
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
Long-Term Survival and Late Adverse Events
There was a substantial reduction in long-term survival for patients with preoperative AF. Studies investigating the impact of preoperative AF on late survival after mitral valve surgery have given conflicting results. Studies comparing small unmatched groups like the series of Jessurun and associates [18] (68 AF and 57 SR patients) and Obadia and colleagues [15], (96 AF vs 95 SR patients) did not find a significant survival difference between the two groups. Similarly, in a previous study of a different cohort, we [14] did not detect a significant difference in survival between 97 AF and 216 SR patients. In all these series, however, there was a trend toward decreased survival in patients with preoperative AF. These studies are hindered by sample size and limited follow-up interval, hence the power to detect significant survival disparity is decreased.
Studies with larger sample sizes have identified significant differences in long-term survival. Bando and colleagues [17] compared 363 patients with preoperative AF (163 had maze operation) with an unmatched cohort of 663 patients with SR and reported a survival advantage for the latter group. A similar result was reported by Lim [16] and Eguchi [19] and their associates in their large series.
The present investigation differs from previous studies in the case-matched design, which minimized differences in preoperative characteristics. We observed a reduced late survival among patients with preoperative AF compared with patients with SR after repair of MR. In addition, and importantly, we found a difference in late survival of patients with preoperative AF on the basis of the clinical type of AF. There were more late deaths among patients with permanent AF compared with those with paroxysmal AF. This finding is in keeping with nonsurgical reports that describe permanent AF as an advanced form of the disease [20] associated with worse survival [21].
Despite matching patients for known risk factors of age, gender, and ejection fraction, we found that other variables influencing survival were not evenly distributed between the groups. In a multivariable model, these factors rather than presence of preoperative AF were predictive of late death. In clinical practice, AF is frequently observed in patients with mitral valve disease who have congestive heart failure and cardiomegaly. These markers of advanced valve disease appear to be more important determinants of postoperative mortality than the accompanying arrhythmia itself.
Among patients with preoperative AF, cardiac-related hospital readmission was more common than among those in preoperative SR despite otherwise appropriate medical therapy. The rates of noncerebral bleeds and (or) thromboembolism, heart failure, subsequent heart rhythm-related interventions, and stroke were also higher. These data are consistent with our findings for patients undergoing aortic valve replacement [8] and other reports [18, 22].
Although preoperative AF was not a predictor of late death, it had an important impact on other nonfatal cardiac events and stroke. The risk of developing a stroke and (or) heart failure late postoperatively was increased twofold by preoperative AF. There is a strong association between AF and congestive heart failure, stroke, repeat hospital admissions, and cardiac death [17, 23].
The clinical impact of AF ablation at the time of mitral valve repair should therefore be put into perspective. Our data suggest that the potential benefit of concomitant AF ablation is reduction of late postoperative morbidity. The survival benefit, as has been suggested by others [24, 25] is not certain because of the other risk factors commonly present with AF. A recent metaanalysis by Wong and Mak [26] shows that the restoration of sinus rhythm by the maze procedure reduces the rate of adverse events, but did not improve survival. Avoidance of AF after operation might improve survival by decreasing fatal complications of stroke, but much larger numbers of patients would be necessary to study this. Understanding the influence of preoperative AF on late outcome is further complicated by the fact that late AF occurs in approximately one-quarter of patients who have sinus rhythm preoperatively.
Study Limitations
The exclusive study of nonrheumatic MR and matching of patients on fundamental variables that affect prognosis makes this study relevant to contemporary practice. However, findings in this study should be interpreted in the context of its retrospective design and the difficulty in separating paroxysmal and persistent AF. Also, the inclusion of patients who had the Cox-maze procedure (a strategy used previously [17]) is a potential weakness, although analysis of the subgroup, excluding the Cox-maze patients, yielded essentially identical results as were found with the overall cohort of patients with preoperative AF; probably because of the small number of patients who had the maze procedure. This study was not designed and adequately powered to compare the outcome of the maze operation performed in a selective subset of patients with patients in sinus rhythm.
Late postoperative heart rhythm was not confirmed by electrocardiography in all patients in this study. In 22% of patients who had electrocardiograms within 6 months of the survey, there was a 100% concordance between the self-reported heart rhythms and documented rhythms. Electrocardiograms provide a "snapshot" assessment of heart rhythm, and may not detect paroxysmal AF.
Conclusion
Atrial fibrillation is common in patients undergoing surgery for nonrheumatic mitral regurgitation and frequently coexists with morphologic and functional changes indicative of advanced mitral disease. It has multiple adverse effects on prognosis after repair of mitral regurgitation and, therefore, is a marker for poor clinical outcome. Patients with preoperative AF have high rates of adverse cardiac events and stroke, and a substantial reduction in survival. Preoperative AF is an important predictor of adverse cardiac events and stroke but is not an independent predictor of late mortality. We believe that ablation of AF at the time of mitral valve surgery for nonrheumatic regurgitation is warranted in order to reduce late morbidity, but the benefit of ablation of AF on late survival is uncertain.
|
Appendix 1 |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
Variables Selected and Retained in the Multivariable Model
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
In a case-matched cohort of patients having had modern mitral valve repairs at the Mayo Clinic, the authors have shown us that the operative mortality was increased and the operative morbidity was increased in patients who had prior atrial fibrillation. Moreover, the paper suggests that both five- and ten-year patient survival were shortened as well. However, atrial fibrillation in and of itself did not show up as an independent risk factor for late death. Those who had permanent atrial fibrillation had a significantly higher risk of death than those with intermittent episodes. Preoperative atrial fibrillation did remain an independent risk factor for late adverse cardiac events and strokes. This is really the first large surgical study to define clearly the direction in which surgeons should aspire when repairing degenerative mitral valves in the presence of preoperative atrial fibrillation. The present study differs from other author studies in that there was a case-matched design. Over the years surgeons, and especially cardiologists, largely have ignored the ravages of atrial fibrillation when combined with mitral disease. Their focus has been on replacing the valve, pharmacologic rate control, and degree of anticoagulation to provide hemodynamic stability and protect from strokes. As you know, Dr Jim Cox brought us an operation that was highly effective for surgically treating atrial fibrillation, either for stand-alone or combined with valve surgery. Nevertheless, the adoption rate of the combined operation remained very small, and your center is one of the very few to do the combined operation, especially when there is a complex repair needed. However, several large centers did show that the combination of the Cox-maze could be effective with a repair. In fact, Gillinov at the Cleveland Clinic showed that at five years after the repair and a combined maze, 80% were out of atrial fibrillation. The length of preoperative atrial fibrillation, age, increased left atrial size, and higher left ventricular mass were negative factors. Thus, with the newer, quicker methods for intraoperative ablation, it seems incumbent that the treatment philosophy espoused this morning should be adopted widely. I have several questions for you, Dr Ngaage.
My first question relates to a statistical matter. The study appears to have been powered sufficiently at over 200 patients per cohort for correct statistical analysis. However, all of your other information suggests that atrial fibrillation should have been a strong independent risk factor for mortality. There was a higher five- and ten-year death rate. Nevertheless, your data suggests that this was because of associated factors other than atrial fibrillation, especially in your secondary analysis. Despite your present analysis, do you think that if you had followed up all 1,020 patients in your repair group with mazes that death would have clearly risen as an independent risk factor?
The second question. Can you comment on what percent of patients revert to nonfibrillating rhythms with repairs alone? I believe the Mayo Clinic has done some studies in that area.
The third question. Have you analyzed the features of preoperative atrial fibrillation and tailored your maze operation that you are doing now based on those predictors for atrial fibrillation such as left atrial reduction, left-sided lesions only versus a full maze, or appendage closure?
And the last question. Are you now using alternative energy sources? Since you have classically done a cut and sew maze, are you using sources like cryoablation to perform these operations quicker?
I really think that this paper should be a guideline as to what we should do when repairing mitral valves in patients with atrial fibrillation. It is now left up to the rest of us to follow your lead. Thank you for the opportunity to discuss this paper.
DR NGAAGE: Thank you, Dr Chitwood. In answer to your first question about the power of the study: From previous reports we calculated a sample size of 200 patients in each group to detect a 10% difference in five-year survival between the groups at 90% power, and 149 patients in each group at 80% power, so this study was adequately powered.
AF (atrial fibrillation) was not a risk factor, and I was surprised that this was the case. The association of atrial fibrillation with other cardiac morbidities like cardiomegaly and congestive cardiac failure, which were more prevalent in the AF group and exhibited profound impact on survival, may explain why atrial fibrillation was not a risk factor. When we did not include cardiomegaly and congestive cardiac failure in the analysis, preoperative AF was a risk factor for late mortality. Also, in another paper in the JTCVS (Journal of Thoracic and Cardiovascular Surgery) press for patients with coronary artery disease where cardiomegaly and congestive cardiac failure was not as prevalent, preoperative AF turned out to be a risk factor for late mortality. The classical maze, that is the biatrial cut and sew maze, allows some reduction of the right and left atrial size and that may be of advantage in patients with cardiomegaly. Finally, at our institution now, we use the alternative energy sources like the radiofrequency ablating devices, and our early results are similar to that of classical maze.
DR MICHAEL MACK (Dallas, TX): Did you compare your subgroup of 90 patients with AF who did have a maze procedure with either those that did not have a maze procedure or with sinus rhythm patients to see whether doing a maze changed the prognosis in these AF patients?
DR NGAAGE: We did not compare the groups because at our institution the maze procedure at the time of this study was performed predominantly for patients who were younger, so there was a bias in the selection of patients. When we looked at the baseline variables there was great disparity between the groups, which precluded a meaningful comparison.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAppendix 1DiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Groh Off-pump epicardial high intensity focused ultrasound ablation of atrial fibrillation in patients undergoing structural heart repair MMCTS, June 18, 2009; 2009(0618): 3061. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Modi, E. Rodriguez, W. C. Hargrove III, A. Hassan, W. Y. Szeto, and W. R. Chitwood Jr. Minimally invasive video-assisted mitral valve surgery: A 12-year, 2-center experience in 1178 patients. J. Thorac. Cardiovasc. Surg., June 1, 2009; 137(6): 1481 - 1487. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Louagie, M. Buche, P. Eucher, J.-C. Schoevaerdts, M. Gerard, J. Jamart, and D. Blommaert Improved patient survival with concomitant Cox Maze III procedure compared with heart surgery alone. Ann. Thorac. Surg., February 1, 2009; 87(2): 440 - 446. [Abstract] [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 |
