- Email: [email protected]
Use of a Fenestration Should be Routine During the Fontan Procedure: PRO
Use of a Fenestration Should be Routine During the Fontan Procedure: PRO Scott M. Bradley, MD The physiologic effects of a fenestration are to decrease “Fontan pressure” and systemic oxygen saturation, but to improve both cardiac index and systemic oxygen delivery. Several nonrandomized studies have shown that a fenestration improves clinical outcomes by decreasing the duration of postoperative effusions and the length of hospital stay. The single prospective, randomized study has also shown that use of a fenestration improves clinical outcomes. While there are some disadvantages to a fenestration; these are of relatively limited clinical importance. Selective rather than routine use of a fenestration relies on the ability to predict outcome based on preoperative or intraoperative variables. However, none of these variables has been found to have useful predictive value in the individual patient. In conclusion, the Fontan procedure can be done without a fenestration, with good outcomes. However, the positive effects of a fenestration on duration of effusions and length of hospital stay argue that use of a fenestration should be routine during the Fontan procedure. Semin Thorac Cardiovasc Surg Pediatr Card Surg Ann 13:55-59 © 2010 Elsevier Inc. All rights reserved.
Background
T
he goal of this article is to make the case that use of a fenestration should be routine during the Fontan procedure. The crux of this issue can be seen in the chest radiograms shown in Fig. 1: pleural effusions early after a Fontan procedure. Effusions, in turn, lead to drainage tubes, and to increased length of hospital stay. Before exploring these clinical issues, I would like to first establish the physiologic effects of a fenestration in a patient who has undergone a Fontan procedure. What does a fenestration actually do? The best answer to this question is provided by looking at closure of fenestrations in the cardiac catheterization lab. Fenestration closures are generally carried out weeks or months after surgery. In this setting, there are no confounding effects of cardiopulmonary bypass and the other perioperative effects of open heart surgery. The cath lab setting also allows for accurate measurement of oxygen saturations, blood flows, and oxygen delivery. There have been several studies that have examined the effects of fenestration closure in the cath lab. Their results are summarized
Professor of Surgery, From the Section of Pediatric Cardiothoracic Surgery, Medical University of South Carolina, Charleston, SC. Address correspondence to Scott M. Bradley, MD, Pediatric Cardiothoracic Surgery, Medical University of South Carolina, CSB 424, 96 Jonathan Lucas St., Charleston, SC 29425; e-mail: [email protected]
1092-9126/10/$-see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1053/j.pcsu.2010.01.004
in Table 1. These results are striking for their consistency: fenestration closure results in a rise in central venous pressure (“Fontan pressure”) and in systemic oxygen saturation, but a fall in cardiac index. The fall in cardiac index more than offsets the rise in oxygen saturation, so that overall systemic oxygen delivery decreases with fenestration closure (Table 1). We can infer that the physiologic effects of a fenestration in the Fontan patient are to decrease “Fontan pressure” and systemic oxygen saturation, but to improve both cardiac index and systemic oxygen delivery. The question remains, do these physiologic benefits translate into improved clinical outcomes?
Clinical Outcomes Several studies have evaluated the clinical effects of a fenestration in patients undergoing a Fontan procedure.6-9 All but one of these studies9 have been nonrandomized. They have consistently found that a fenestration decreases the duration of pleural effusions and the length of hospital stay. One of the earliest studies is from Children’s Hospital Boston,6 and includes among its coauthors my opponent in this debate, Frank Hanley. The Boston study compared 56 patients undergoing a Fontan procedure with no fenestration with 91 patients with a fenestration. The fenestrated patients were judged to be at higher surgical risk by virtue of higher preoperative pulmonary artery pressure, pulmonary vascular re55
S.M. Bradley
56
Figure 1 Chest radiograms showing pleural effusions following a Fontan procedure.
sistance, ventricular filling pressure, and pulmonary artery distortion. The results are summarized in Table 2. Despite being at higher risk, the fenestrated patients had better outcomes. Their mortality was lower, although this was not statistically significant. Patients with a fenestration had an average duration of pleural effusions that was 13 days shorter, and average hospital stay that was 8 days shorter. These differences were not just statistically significant, but also clinically meaningful. Airan et al7 reported on 348 patients undergoing a Fontan procedure from 1988 to 1997. Their analysis was focused on the role of a fenestration. All patients (126) operated since 1994 underwent routine fenestration. The overall Fontan failure rate was 14%; pleural effusions lasting more than 10 days were seen in 27%. In a multivariable analysis, the only risk factor for Fontan failure was absence of a fenestration; risk factors for pleural effusions were absence of a fenestration and aortic cross-clamp time greater than 60 minutes. While this study showed a beneficial effect of fenestration use on postoperative pleural effusions, it is subject to the criti-
Table 1 Effects of Fenestration Closure in the Cardiac Catheterization Laboratory
Chicago1 1992 Yale2 1992 UCLA3 1995 Boston4 1995 Michigan5 1998
CVP
O2 Saturation
Cardiac Index
O2 Delivery
1
1
2
2
¡
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
Abbreviations: CVP, central venous pressure (“Fontan pressure”); O2, oxygen.
cism that fenestration use was limited to the patients operated in the later part of the study (historical controls). One of the most recent studies is from the National Heart, Lung and Blood Institute (NHLBI)-sponsored Pediatric Heart Network Fontan cross-sectional study.8 This study includes seven centers and 546 patients who have undergone a Fontan procedure, 361 (66%) with a fenestration. As in the Boston study, the fenestrated patients were judged to be at higher surgical risk, by virtue of more preoperative atrioventricular valve regurgitation. Despite this, the fenestrated patients had lower incidences of postoperative pericardial effusions as well as pleural effusions, and shorter average hospital stay (Table 3). In summary, several nonrandomized studies have indicated that use of a fenestration in a Fontan procedure is associated with improved clinical outcomes, in particular shorter duration of postoperative effusions and shorter hospital stay, despite the fenestrated patients generally being at higher surgical risk. These studies are subject to all of the limitations of nonrandomized studies with historical controls. For most of the topics discussed and debated at our national meetings, nonrandomized studies comprise the best evidence we have. We do not have the gold standard: a prospective, randomized study.
Table 2 Effect of a Fenestration on the Outcome of the Fontan Operation Fenestration
Yes
No
P Value
Fontan failure Duration of effusions (mean; days) Effusions > 14 days Hospital stay (mean; days)
7% 11
11% 24
NS <.01
13% 13
38% 21
<.01 <.01
Data from Bridges et al.6
Fenestration should be routine during the fontan procedure
57
Table 3 Effect of a Fenestration on the Outcome of the Fontan Operation Fenestration
Yes
No
P Value
Pericardial effusions Pleural effusions Hospital stay (mean; days)
7% 21% 16
21% 35% 22
<.01 <.01 <.01
Data from Atz et al.8
The Gold Standard In the case of the current topic, the gold standard study has actually been performed. It was published in 2002 by Lemler et al9 from Dallas Children’s Hospital. Standard-risk patients undergoing a Fontan procedure were prospectively randomized to receive a fenestration (25 patients) or no fenestration (24 patients). The results are summarized in Table 4. Patients with a fenestration had a shorter duration of postoperative effusions and shorter hospital stay, as well as a lower incidence of additional procedures. Once again, these differences were both statistically significant and clinically important. Figure 2, reproduced from their paper, shows the effect of a fenestration on the percentage of patients with a chest tube in place over the postoperative period. This prospective, randomized study indicates that use of a fenestration improves the clinical outcome of a Fontan procedure, even in patients considered at standard surgical risk.9
Other Inferences from the Literature The discussion thus far has focused on comparisons within institutions. There are several reasons that it is difficult to reach valid conclusions regarding the use of a fenestration by comparing results between institutions. Operative mortality for the Fontan procedure is now low, despite a wide variety of surgical approaches.10,11 This makes mortality an insufficiently sensitive outcome to assess the effect of any variation in approach, including the use of a fenestration. Duration of effusions and hospital length of stay are outcomes with more variation than mortality. However, both of these outcomes are affected by many factors that vary from one institution to the next. For example, different institutions (and even surgeons) have different protocols dictating postoperative chest tube removal. This makes it very difficult to make meaningful Table 4 Effect of a Fenestration on the Outcome of the Fontan Operation Fenestration Chest tube drainage (median; days) Hospital stay (median; days) Additional procedures (no. of patients) Additional procedures (% of patients) Data from Lemler et al.9
P Value
Yes
No
10
18
<.01
12 4
23 56
.03 <.01
13%
69%
<.01
Figure 2 Duration of chest tube drainage in fenestrated versus nonfenestrated patients. The Kaplan-Meier curve demonstrates that the duration of chest tube drainage is shorter in fenestrated patients (solid line) than in nonfenestrated patients (broken line). (Reprinted with permission.9)
comparisons of the duration of effusions and length of hospital stay from one institution to another.
Short-Term and Long-Term Considerations It can be questioned whether the duration of effusions and length of hospital stay are important considerations. What is wrong with spending some extra time in a hospital? Time in the hospital clearly costs money. It also carries social and psychological costs for the patient and the family. Time in the hospital increases the risk of other complications, such as infection. Finally, time in the hospital may even be bad for the brain. The most recent analysis of the Children’s Hospital of Philadelphia apolipoprotein E (APOE) study looked at predictors of impaired neurodevelopmental outcomes at 1 year of age after infant cardiac surgery.12 Longer length of hospital stay was an independent, multivariable predictor of lower psychomotor development index. Thus, efforts to decrease hospital length of stay, such as the use of a fenestration, seem worthwhile. The use of a fenestration may have implications not just for early, but also for late outcomes. There is some data that early pleural effusions predict late outcome. In a study from the University of Michigan looking at 636 patients who had undergone a Fontan procedure from 1992 to 2007, late proteinlosing enteropathy was seen in 6%.13 Postoperative pleural effusions lasting more than 14 days were a significant risk factor both for worse long-term survival and for the development of protein-losing enteropathy. These data suggest that decreasing early effusions may have a positive effect on longterm outcomes.
Why Not Fenestrate? Given all of this, the question should be asked: Why not use a fenestration? One issue is the need for later closure. How-
S.M. Bradley
58 ever, spontaneous closure of Fontan fenestrations occurs in at least 20% of patients.7,14 There is then some debate whether a fenestration that remains open needs to be closed. One reason to close a fenestration is cyanosis. However, systemic oxygen saturation in a patient with a fenestration is generally 90% to 92%, which compares favorably with the lower saturations considered acceptable in patients awaiting a Fontan procedure. Another reason to close a fenestration is the risk of stroke, due to thromboembolism and right-to-left shunting. While this is a logical concern, no clear risk has been documented. For example, the Pediatric Heart Network (PHN) Fontan cross-sectional study examined 546 patients, 361 (66%) with a fenestration.8 At a median follow-up of 8.1 years, the incidence of stroke was 2% in patients with a fenestration versus 1% in those without (P ⫽ .5).8 Nonetheless, if it is decided to close a fenestration, this can be carried out as a straightforward, outpatient procedure in the cardiac catheterization laboratory. The procedure can be done by placement of a device, or by use of a snare left at the time of the Fontan operation, which avoids placement of an intravascular foreign body.
Why Not Fenestrate Selectively? A remaining question is: Why not use fenestration selectively? It has been suggested that the decision to use a fenestration in a given patient could be based on either preoperative or intraoperative factors.10,15 For example, the Fontan operation can be performed, and if the central venous pressure or transpulmonary gradient at the end of the operation is above a certain value, a fenestration can be placed at that time.10 This approach raises the “crystal ball” issue: are there any variables that accurately predict outcome after a Fontan procedure in an individual patient? Many studies have examined risk factors for pleural effusions following a Fontan procedure. A variety of factors have been identified, such as patient age (either younger or older), respiratory viral season, lower preoperative oxygen saturation, and longer cardiopulmonary bypass and cross-clamp times.7,16-18 However, every one of these variables has poor predictive value, both positive and negative. The association of each of these variables with pleural effusions is true only on average; in an individual patient, none of them is an accurate predictor. To illustrate, Fig. 3A shows the relationship between transpulmonary gradient measured at the end of a Fontan operation and subsequent duration of pleural effusions. Figure 3B shows the same relationship for central venous pressure. These data are from 50 patients undergoing a Fontan procedure at the Medical University of South Carolina from January 2006 to September 2009. Median duration of effusions for the entire group was 7 days. However, there were some patients with a low transpulmonary gradient or central venous pressure who drained for more than 2 weeks, and some with a high transpulmonary gradient or central venous pressure who drained for less than 1 week. In short, there is no crystal ball to predict postoperative course following a Fontan procedure. This argues in favor of placing a fenestration in all patients at the time of surgery.
Figure 3 A, Duration of chest tube drainage (weeks) versus transpulmonary gradient (TPG ⫽ pulmonary artery – common atrial pressure; mmHg) measured at arrival in the intensive care unit. Data are from 50 patients undergoing a Fontan procedure at the Medical University of South Carolina from January 2006 to September 2009. Median duration of chest tube drainage (dotted line) was 7 days (range, 3 days to 5.3 weeks). There is poor correlation between the two variables (R value ⫽ 0.02). B, Duration of chest tube drainage (weeks) versus central venous pressure (CVP, “Fontan pressure,” or pulmonary artery pressure; mmHg) measured at arrival in the intensive care unit. Data are from 50 patients undergoing a Fontan procedure at the Medical University of South Carolina from January 2006 to September 2009. Median duration of chest tube drainage (dotted line) was 7 days (range, 3 days to 5.3 weeks). There is poor correlation between the 2 variables (R value ⫽ 0.06).
The Off-Pump Fontan The proposal has been made that the details of the surgical approach to the Fontan procedure might affect the need for a fenestration.10 In particular, minimizing or avoiding cardiopulmonary bypass and aortic cross-clamping could preserve myocardial and pulmonary function, thus eliminating the necessity of a routine fenestration. The Stanford/University of California-San Francisco experience with an extracardiac conduit Fontan operation with minimal or no cardiopulmonary bypass has had excellent results.10 However, even with this approach, the median duration of pleural effusions was 8 days (range. 2 to 91 days), with a duration of effusions longer than 2 weeks in 15% of patients.10 The effect of cardiopulmonary bypass on postoperative effusions was specifically examined in a report from Shikata et al.19 This report included 27 patients undergoing a Fontan procedure with the use of cardiopulmonary bypass, and 47 patients undergoing an “off-pump” Fontan. Propensity matching was used to select 14 patients from each of the two groups for direct comparison. Figure 4 shows that there was no difference in the
Fenestration should be routine during the fontan procedure
59
References
Figure 4 Duration of pleural and peritoneal drainage in the cardiopulmonary bypass and off-pump patients (propensity score matched). The Kaplan-Meier curve demonstrates that there was no significant difference between the cardiopulmonary bypass and off-pump groups. (Reprinted with permission.19)
percentage of patients with postoperative effusions in the cardiopulmonary bypass and off-pump groups. Perhaps the ultimate “off-pump” approach is Fontan completion in the cardiac catheterization lab. This is a setting that avoids any side effects, not only of cardiopulmonary bypass but also of surgery, such as mediastinal dissection, tissue trauma, and inflammation. Nonetheless, an abstract reporting the results of this approach noted pleural effusions in 10 out of 11 patients, with a duration ranging from 4 days to 6 months.20 This experience argues that a propensity for effusions may be intrinsic to the transition to Fontan physiology, independent of the side effects of cardiopulmonary bypass and surgery.
Conclusion The title of this debate includes the word “routine.” Webster’s dictionary tells us that this word is from the middle French, meaning route, traveled way, a regular course of procedure. There is no doubt that the Fontan procedure can be done without a fenestration, and with good outcomes. However, the positive effects of a fenestration on duration of effusions and length of hospital stay argue that use of a fenestration should be routine during the Fontan procedure.
1. Mavroudis C, Zales VR, Backer CL, et al. Fenestrated Fontan with delayed catheter closure. Circulation 1992;II-85–II-92 2. Hijazi ZM, Fahey JT, Kleinman CS, et al. Hemodynamic evaluation before and after closure of fenestrated Fontan. Circulation 1992;86: 196-202 3. Kuhn MA, Jarmakani JM, Laks H, et al. Effect of late postoperative atrial septal defect closure on hemodynamic function in patients with a lateral tunnel Fontan procedure. J Am Coll Cardiol 1995;26:259-265 4. Bridges ND, Lock JE, Mayer JE Jr, et al. Cardiac catheterization and test occlusion of the interatrial communication after the fenestrated Fontan operation. J Am Coll Cardiol 1995;25:1712-1717 5. Lloyd TR, Rydberg A, Ludomirsky A, et al. Late fenestration closure in the hypoplastic left heart syndrome: comparison of hemodynamic changes. Am Heart J 1998;136:302-306 6. Bridges ND, Mayer JE Jr, Lock JE, et al. Effect of baffle fenestration on outcome of the modified Fontan operation. Circulation 1992;86:17621769 7. Airan B, Sharma R, Choudhary SK, et al. Univentricular repair: is routine fenestration justified? Ann Thorac Surg 2000;69:1900-1906 8. Atz AM, Travison TG, McCrindle BW, et al. Impact of surgical fenestration in a cohort of Fontan patients [abstract]. J Am Coll Cardiol 2009;53(suppl A):361A 9. Lemler MS, Scott WA, Leonard SR, et al. Fenestration improves clinical outcome of the Fontan procedure: a prospective, randomized study. Circulation 2002;105:207-212 10. Petrossian E, Reddy VM, Collins KK, et al. The extracardiac conduit Fontan operation using minimal approach extracorporeal circulation: early and midterm outcomes. J Thorac Cardiovasc Surg 2006;132: 1054-1063 11. Meyer DB, Zamora G, Wernovsky G, et al. Outcomes of the Fontan procedure using cardiopulmonary bypass with aortic cross-clamping. Ann Thorac Surg 2006;82:1611-1618 12. Fuller S, Gerdes M, Wernovsky G, et al. Predictors of impaired neurodevelopmental outcomes at one year of age after infant cardiac surgery. Eur J Cardiothorac Surg 2009;36:40-47 13. Hirsch JC, Goldberg C, Bove EL, et al. Fontan operation in the current era: a 15-year single institution experience. Ann Surg 2008;248:402410 14. Yang S, Spray TL, Gaynor JW, et al. Rate of spontaneous closure of fenestration in the modified Fontan operation [abstract]. Circulation 1999;100:I-399 15. Hsu DT, Quaegebeur JM, Ing FF, et al. Outcome after the single-stage, nonfenestrated Fontan procedure. Circulation 1997;96:II-335–II-40 16. Gupta A, Daggett C, Behera S, et al. Risk factors for persistent pleural effusion after the extracardiac Fontan procedure. J Thorac Cardiovasc Surg 2004;127:1664-1669 17. Fedderly RT, Whitstone BN, Frisbee SJ, et al. Factors related to pleural effusions after Fontan procedure in the era of fenestration. Circulation 2001;104:I-148 –I-51 18. Schreiber C, Horer J, Vogt M, et al. Nonfenestrated extracardiac total cavopulmonary connection in 132 consecutive patients. Ann Thorac Surg 2007;84:894-899 19. Shikata F, Yagihara T, Kagisaki K, et al. Does the off-pump Fontan procedure ameliorate the volume and duration of pleural and peritoneal effusions? Eur J Cardiothorac Surg 2008;34:570-575 20. Murphy JD, Murdison KA, Nehgme RA, et al. Catheter facilitated completion of Fontan procedure [abstract]. Circulation 2003;108:IV-704
Background
T
he goal of this article is to make the case that use of a fenestration should be routine during the Fontan procedure. The crux of this issue can be seen in the chest radiograms shown in Fig. 1: pleural effusions early after a Fontan procedure. Effusions, in turn, lead to drainage tubes, and to increased length of hospital stay. Before exploring these clinical issues, I would like to first establish the physiologic effects of a fenestration in a patient who has undergone a Fontan procedure. What does a fenestration actually do? The best answer to this question is provided by looking at closure of fenestrations in the cardiac catheterization lab. Fenestration closures are generally carried out weeks or months after surgery. In this setting, there are no confounding effects of cardiopulmonary bypass and the other perioperative effects of open heart surgery. The cath lab setting also allows for accurate measurement of oxygen saturations, blood flows, and oxygen delivery. There have been several studies that have examined the effects of fenestration closure in the cath lab. Their results are summarized
Professor of Surgery, From the Section of Pediatric Cardiothoracic Surgery, Medical University of South Carolina, Charleston, SC. Address correspondence to Scott M. Bradley, MD, Pediatric Cardiothoracic Surgery, Medical University of South Carolina, CSB 424, 96 Jonathan Lucas St., Charleston, SC 29425; e-mail: [email protected]
1092-9126/10/$-see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1053/j.pcsu.2010.01.004
in Table 1. These results are striking for their consistency: fenestration closure results in a rise in central venous pressure (“Fontan pressure”) and in systemic oxygen saturation, but a fall in cardiac index. The fall in cardiac index more than offsets the rise in oxygen saturation, so that overall systemic oxygen delivery decreases with fenestration closure (Table 1). We can infer that the physiologic effects of a fenestration in the Fontan patient are to decrease “Fontan pressure” and systemic oxygen saturation, but to improve both cardiac index and systemic oxygen delivery. The question remains, do these physiologic benefits translate into improved clinical outcomes?
Clinical Outcomes Several studies have evaluated the clinical effects of a fenestration in patients undergoing a Fontan procedure.6-9 All but one of these studies9 have been nonrandomized. They have consistently found that a fenestration decreases the duration of pleural effusions and the length of hospital stay. One of the earliest studies is from Children’s Hospital Boston,6 and includes among its coauthors my opponent in this debate, Frank Hanley. The Boston study compared 56 patients undergoing a Fontan procedure with no fenestration with 91 patients with a fenestration. The fenestrated patients were judged to be at higher surgical risk by virtue of higher preoperative pulmonary artery pressure, pulmonary vascular re55
S.M. Bradley
56
Figure 1 Chest radiograms showing pleural effusions following a Fontan procedure.
sistance, ventricular filling pressure, and pulmonary artery distortion. The results are summarized in Table 2. Despite being at higher risk, the fenestrated patients had better outcomes. Their mortality was lower, although this was not statistically significant. Patients with a fenestration had an average duration of pleural effusions that was 13 days shorter, and average hospital stay that was 8 days shorter. These differences were not just statistically significant, but also clinically meaningful. Airan et al7 reported on 348 patients undergoing a Fontan procedure from 1988 to 1997. Their analysis was focused on the role of a fenestration. All patients (126) operated since 1994 underwent routine fenestration. The overall Fontan failure rate was 14%; pleural effusions lasting more than 10 days were seen in 27%. In a multivariable analysis, the only risk factor for Fontan failure was absence of a fenestration; risk factors for pleural effusions were absence of a fenestration and aortic cross-clamp time greater than 60 minutes. While this study showed a beneficial effect of fenestration use on postoperative pleural effusions, it is subject to the criti-
Table 1 Effects of Fenestration Closure in the Cardiac Catheterization Laboratory
Chicago1 1992 Yale2 1992 UCLA3 1995 Boston4 1995 Michigan5 1998
CVP
O2 Saturation
Cardiac Index
O2 Delivery
1
1
2
2
¡
1
2
2
1
1
2
2
1
1
2
2
1
1
2
2
Abbreviations: CVP, central venous pressure (“Fontan pressure”); O2, oxygen.
cism that fenestration use was limited to the patients operated in the later part of the study (historical controls). One of the most recent studies is from the National Heart, Lung and Blood Institute (NHLBI)-sponsored Pediatric Heart Network Fontan cross-sectional study.8 This study includes seven centers and 546 patients who have undergone a Fontan procedure, 361 (66%) with a fenestration. As in the Boston study, the fenestrated patients were judged to be at higher surgical risk, by virtue of more preoperative atrioventricular valve regurgitation. Despite this, the fenestrated patients had lower incidences of postoperative pericardial effusions as well as pleural effusions, and shorter average hospital stay (Table 3). In summary, several nonrandomized studies have indicated that use of a fenestration in a Fontan procedure is associated with improved clinical outcomes, in particular shorter duration of postoperative effusions and shorter hospital stay, despite the fenestrated patients generally being at higher surgical risk. These studies are subject to all of the limitations of nonrandomized studies with historical controls. For most of the topics discussed and debated at our national meetings, nonrandomized studies comprise the best evidence we have. We do not have the gold standard: a prospective, randomized study.
Table 2 Effect of a Fenestration on the Outcome of the Fontan Operation Fenestration
Yes
No
P Value
Fontan failure Duration of effusions (mean; days) Effusions > 14 days Hospital stay (mean; days)
7% 11
11% 24
NS <.01
13% 13
38% 21
<.01 <.01
Data from Bridges et al.6
Fenestration should be routine during the fontan procedure
57
Table 3 Effect of a Fenestration on the Outcome of the Fontan Operation Fenestration
Yes
No
P Value
Pericardial effusions Pleural effusions Hospital stay (mean; days)
7% 21% 16
21% 35% 22
<.01 <.01 <.01
Data from Atz et al.8
The Gold Standard In the case of the current topic, the gold standard study has actually been performed. It was published in 2002 by Lemler et al9 from Dallas Children’s Hospital. Standard-risk patients undergoing a Fontan procedure were prospectively randomized to receive a fenestration (25 patients) or no fenestration (24 patients). The results are summarized in Table 4. Patients with a fenestration had a shorter duration of postoperative effusions and shorter hospital stay, as well as a lower incidence of additional procedures. Once again, these differences were both statistically significant and clinically important. Figure 2, reproduced from their paper, shows the effect of a fenestration on the percentage of patients with a chest tube in place over the postoperative period. This prospective, randomized study indicates that use of a fenestration improves the clinical outcome of a Fontan procedure, even in patients considered at standard surgical risk.9
Other Inferences from the Literature The discussion thus far has focused on comparisons within institutions. There are several reasons that it is difficult to reach valid conclusions regarding the use of a fenestration by comparing results between institutions. Operative mortality for the Fontan procedure is now low, despite a wide variety of surgical approaches.10,11 This makes mortality an insufficiently sensitive outcome to assess the effect of any variation in approach, including the use of a fenestration. Duration of effusions and hospital length of stay are outcomes with more variation than mortality. However, both of these outcomes are affected by many factors that vary from one institution to the next. For example, different institutions (and even surgeons) have different protocols dictating postoperative chest tube removal. This makes it very difficult to make meaningful Table 4 Effect of a Fenestration on the Outcome of the Fontan Operation Fenestration Chest tube drainage (median; days) Hospital stay (median; days) Additional procedures (no. of patients) Additional procedures (% of patients) Data from Lemler et al.9
P Value
Yes
No
10
18
<.01
12 4
23 56
.03 <.01
13%
69%
<.01
Figure 2 Duration of chest tube drainage in fenestrated versus nonfenestrated patients. The Kaplan-Meier curve demonstrates that the duration of chest tube drainage is shorter in fenestrated patients (solid line) than in nonfenestrated patients (broken line). (Reprinted with permission.9)
comparisons of the duration of effusions and length of hospital stay from one institution to another.
Short-Term and Long-Term Considerations It can be questioned whether the duration of effusions and length of hospital stay are important considerations. What is wrong with spending some extra time in a hospital? Time in the hospital clearly costs money. It also carries social and psychological costs for the patient and the family. Time in the hospital increases the risk of other complications, such as infection. Finally, time in the hospital may even be bad for the brain. The most recent analysis of the Children’s Hospital of Philadelphia apolipoprotein E (APOE) study looked at predictors of impaired neurodevelopmental outcomes at 1 year of age after infant cardiac surgery.12 Longer length of hospital stay was an independent, multivariable predictor of lower psychomotor development index. Thus, efforts to decrease hospital length of stay, such as the use of a fenestration, seem worthwhile. The use of a fenestration may have implications not just for early, but also for late outcomes. There is some data that early pleural effusions predict late outcome. In a study from the University of Michigan looking at 636 patients who had undergone a Fontan procedure from 1992 to 2007, late proteinlosing enteropathy was seen in 6%.13 Postoperative pleural effusions lasting more than 14 days were a significant risk factor both for worse long-term survival and for the development of protein-losing enteropathy. These data suggest that decreasing early effusions may have a positive effect on longterm outcomes.
Why Not Fenestrate? Given all of this, the question should be asked: Why not use a fenestration? One issue is the need for later closure. How-
S.M. Bradley
58 ever, spontaneous closure of Fontan fenestrations occurs in at least 20% of patients.7,14 There is then some debate whether a fenestration that remains open needs to be closed. One reason to close a fenestration is cyanosis. However, systemic oxygen saturation in a patient with a fenestration is generally 90% to 92%, which compares favorably with the lower saturations considered acceptable in patients awaiting a Fontan procedure. Another reason to close a fenestration is the risk of stroke, due to thromboembolism and right-to-left shunting. While this is a logical concern, no clear risk has been documented. For example, the Pediatric Heart Network (PHN) Fontan cross-sectional study examined 546 patients, 361 (66%) with a fenestration.8 At a median follow-up of 8.1 years, the incidence of stroke was 2% in patients with a fenestration versus 1% in those without (P ⫽ .5).8 Nonetheless, if it is decided to close a fenestration, this can be carried out as a straightforward, outpatient procedure in the cardiac catheterization laboratory. The procedure can be done by placement of a device, or by use of a snare left at the time of the Fontan operation, which avoids placement of an intravascular foreign body.
Why Not Fenestrate Selectively? A remaining question is: Why not use fenestration selectively? It has been suggested that the decision to use a fenestration in a given patient could be based on either preoperative or intraoperative factors.10,15 For example, the Fontan operation can be performed, and if the central venous pressure or transpulmonary gradient at the end of the operation is above a certain value, a fenestration can be placed at that time.10 This approach raises the “crystal ball” issue: are there any variables that accurately predict outcome after a Fontan procedure in an individual patient? Many studies have examined risk factors for pleural effusions following a Fontan procedure. A variety of factors have been identified, such as patient age (either younger or older), respiratory viral season, lower preoperative oxygen saturation, and longer cardiopulmonary bypass and cross-clamp times.7,16-18 However, every one of these variables has poor predictive value, both positive and negative. The association of each of these variables with pleural effusions is true only on average; in an individual patient, none of them is an accurate predictor. To illustrate, Fig. 3A shows the relationship between transpulmonary gradient measured at the end of a Fontan operation and subsequent duration of pleural effusions. Figure 3B shows the same relationship for central venous pressure. These data are from 50 patients undergoing a Fontan procedure at the Medical University of South Carolina from January 2006 to September 2009. Median duration of effusions for the entire group was 7 days. However, there were some patients with a low transpulmonary gradient or central venous pressure who drained for more than 2 weeks, and some with a high transpulmonary gradient or central venous pressure who drained for less than 1 week. In short, there is no crystal ball to predict postoperative course following a Fontan procedure. This argues in favor of placing a fenestration in all patients at the time of surgery.
Figure 3 A, Duration of chest tube drainage (weeks) versus transpulmonary gradient (TPG ⫽ pulmonary artery – common atrial pressure; mmHg) measured at arrival in the intensive care unit. Data are from 50 patients undergoing a Fontan procedure at the Medical University of South Carolina from January 2006 to September 2009. Median duration of chest tube drainage (dotted line) was 7 days (range, 3 days to 5.3 weeks). There is poor correlation between the two variables (R value ⫽ 0.02). B, Duration of chest tube drainage (weeks) versus central venous pressure (CVP, “Fontan pressure,” or pulmonary artery pressure; mmHg) measured at arrival in the intensive care unit. Data are from 50 patients undergoing a Fontan procedure at the Medical University of South Carolina from January 2006 to September 2009. Median duration of chest tube drainage (dotted line) was 7 days (range, 3 days to 5.3 weeks). There is poor correlation between the 2 variables (R value ⫽ 0.06).
The Off-Pump Fontan The proposal has been made that the details of the surgical approach to the Fontan procedure might affect the need for a fenestration.10 In particular, minimizing or avoiding cardiopulmonary bypass and aortic cross-clamping could preserve myocardial and pulmonary function, thus eliminating the necessity of a routine fenestration. The Stanford/University of California-San Francisco experience with an extracardiac conduit Fontan operation with minimal or no cardiopulmonary bypass has had excellent results.10 However, even with this approach, the median duration of pleural effusions was 8 days (range. 2 to 91 days), with a duration of effusions longer than 2 weeks in 15% of patients.10 The effect of cardiopulmonary bypass on postoperative effusions was specifically examined in a report from Shikata et al.19 This report included 27 patients undergoing a Fontan procedure with the use of cardiopulmonary bypass, and 47 patients undergoing an “off-pump” Fontan. Propensity matching was used to select 14 patients from each of the two groups for direct comparison. Figure 4 shows that there was no difference in the
Fenestration should be routine during the fontan procedure
59
References
Figure 4 Duration of pleural and peritoneal drainage in the cardiopulmonary bypass and off-pump patients (propensity score matched). The Kaplan-Meier curve demonstrates that there was no significant difference between the cardiopulmonary bypass and off-pump groups. (Reprinted with permission.19)
percentage of patients with postoperative effusions in the cardiopulmonary bypass and off-pump groups. Perhaps the ultimate “off-pump” approach is Fontan completion in the cardiac catheterization lab. This is a setting that avoids any side effects, not only of cardiopulmonary bypass but also of surgery, such as mediastinal dissection, tissue trauma, and inflammation. Nonetheless, an abstract reporting the results of this approach noted pleural effusions in 10 out of 11 patients, with a duration ranging from 4 days to 6 months.20 This experience argues that a propensity for effusions may be intrinsic to the transition to Fontan physiology, independent of the side effects of cardiopulmonary bypass and surgery.
Conclusion The title of this debate includes the word “routine.” Webster’s dictionary tells us that this word is from the middle French, meaning route, traveled way, a regular course of procedure. There is no doubt that the Fontan procedure can be done without a fenestration, and with good outcomes. However, the positive effects of a fenestration on duration of effusions and length of hospital stay argue that use of a fenestration should be routine during the Fontan procedure.
1. Mavroudis C, Zales VR, Backer CL, et al. Fenestrated Fontan with delayed catheter closure. Circulation 1992;II-85–II-92 2. Hijazi ZM, Fahey JT, Kleinman CS, et al. Hemodynamic evaluation before and after closure of fenestrated Fontan. Circulation 1992;86: 196-202 3. Kuhn MA, Jarmakani JM, Laks H, et al. Effect of late postoperative atrial septal defect closure on hemodynamic function in patients with a lateral tunnel Fontan procedure. J Am Coll Cardiol 1995;26:259-265 4. Bridges ND, Lock JE, Mayer JE Jr, et al. Cardiac catheterization and test occlusion of the interatrial communication after the fenestrated Fontan operation. J Am Coll Cardiol 1995;25:1712-1717 5. Lloyd TR, Rydberg A, Ludomirsky A, et al. Late fenestration closure in the hypoplastic left heart syndrome: comparison of hemodynamic changes. Am Heart J 1998;136:302-306 6. Bridges ND, Mayer JE Jr, Lock JE, et al. Effect of baffle fenestration on outcome of the modified Fontan operation. Circulation 1992;86:17621769 7. Airan B, Sharma R, Choudhary SK, et al. Univentricular repair: is routine fenestration justified? Ann Thorac Surg 2000;69:1900-1906 8. Atz AM, Travison TG, McCrindle BW, et al. Impact of surgical fenestration in a cohort of Fontan patients [abstract]. J Am Coll Cardiol 2009;53(suppl A):361A 9. Lemler MS, Scott WA, Leonard SR, et al. Fenestration improves clinical outcome of the Fontan procedure: a prospective, randomized study. Circulation 2002;105:207-212 10. Petrossian E, Reddy VM, Collins KK, et al. The extracardiac conduit Fontan operation using minimal approach extracorporeal circulation: early and midterm outcomes. J Thorac Cardiovasc Surg 2006;132: 1054-1063 11. Meyer DB, Zamora G, Wernovsky G, et al. Outcomes of the Fontan procedure using cardiopulmonary bypass with aortic cross-clamping. Ann Thorac Surg 2006;82:1611-1618 12. Fuller S, Gerdes M, Wernovsky G, et al. Predictors of impaired neurodevelopmental outcomes at one year of age after infant cardiac surgery. Eur J Cardiothorac Surg 2009;36:40-47 13. Hirsch JC, Goldberg C, Bove EL, et al. Fontan operation in the current era: a 15-year single institution experience. Ann Surg 2008;248:402410 14. Yang S, Spray TL, Gaynor JW, et al. Rate of spontaneous closure of fenestration in the modified Fontan operation [abstract]. Circulation 1999;100:I-399 15. Hsu DT, Quaegebeur JM, Ing FF, et al. Outcome after the single-stage, nonfenestrated Fontan procedure. Circulation 1997;96:II-335–II-40 16. Gupta A, Daggett C, Behera S, et al. Risk factors for persistent pleural effusion after the extracardiac Fontan procedure. J Thorac Cardiovasc Surg 2004;127:1664-1669 17. Fedderly RT, Whitstone BN, Frisbee SJ, et al. Factors related to pleural effusions after Fontan procedure in the era of fenestration. Circulation 2001;104:I-148 –I-51 18. Schreiber C, Horer J, Vogt M, et al. Nonfenestrated extracardiac total cavopulmonary connection in 132 consecutive patients. Ann Thorac Surg 2007;84:894-899 19. Shikata F, Yagihara T, Kagisaki K, et al. Does the off-pump Fontan procedure ameliorate the volume and duration of pleural and peritoneal effusions? Eur J Cardiothorac Surg 2008;34:570-575 20. Murphy JD, Murdison KA, Nehgme RA, et al. Catheter facilitated completion of Fontan procedure [abstract]. Circulation 2003;108:IV-704