Background Significant morbidity after Fontan operation results in either takedown, heart transplantation, or death. Initial creation of a fenestration results in less morbidity and mortality; however, the role of late creation of a fenestration in aiding patients manifesting morbidity after an initial nonfenestrated Fontan operation is unclear.
Methods and Results We reviewed our experience with late creation of a surgical fenestration in 9 patients (5.2±3.1 years old) exhibiting chronic effusions (n=4) or protein-losing enteropathy (PLE) (n=5) after lateral tunnel-type Fontan operation. Patients with effusions had creation via coronary punch of two or three 3-mm defects; patients with PLE had creation of a large, 5-mm defect. One child with effusions and multisystem organ failure before fenestration died 7 weeks after surgery secondary to low cardiac output; the other 3 had resolution of effusions within 4 to 6 weeks. Of the 5 with PLE, 3 had normalization of serum proteins and resolution of symptoms at 2 to 6 weeks. The 2 failures had arterial saturations >89% after surgery. Follow-up was from 25 to 30 months. Spontaneous closure of defects occurred in all 3 with effusions. No return of symptoms was noted in 2; however, the third reaccumulated effusions and has undergone refenestration with a large defect. All 3 patients with PLE have remained asymptomatic with patency of the fenestration (4 to 5 mm on echocardiography) and arterial saturation ≤85% for >2 years.
Conclusions Late surgical creation of fenestration results in resolution of morbidity after Fontan operation. Improvement is related to the degree of right-to-left shunt created.
Despite technical advancements and overall improving survival outcome, some patients exhibit significant morbidity after the Fontan operation for functional single ventricle. Low cardiac output, chronic effusions, and debilitating PLE are conditions that plague some children after Fontan operation. It is not uncommon for these patients to die unless the “Fontan physiology” is abandoned, with either takedown of the connections or heart transplantation.1
One modification of the Fontan operation, fenestration of the systemic venous pathway, has been reported to result in reduced morbidity and mortality.2 3 4 5 6 Recently, success has been reported with catheter dilation of fenestration defects that have undergone spontaneous closure.7 In addition, successful relief of PLE has been reported after catheter-created fenestration of an interatrial septum in a patient with an atriopulmonary-type Fontan procedure.8 Little information exists on the effectiveness of late fenestration in alleviating symptoms in patients with an intact, nonfenestrated, lateral tunnel–type systemic venous pathway, a common form of Fontan connection. We report on our experience with late surgical fenestration in an attempt to alleviate morbidity after the Fontan operation.
Between September 1994 and October 1995, 9 patients underwent late surgical fenestration of their systemic venous pathways (Table⇓). Age at fenestration operation was 5.2±3.1 years (range, 1.2 to 9.6 years). Age at Fontan operation was 2.2±1.1 years (range, 1 to 4.5 years). The interval between Fontan operation and fenestration operation ranged from 3 to 81 months. The Fontan operation consisted of creation of an intra-atrial lateral tunnel made of PTFE. Two patients had partial exclusion of hepatic veins from the systemic venous pathway at the Fontan operation; 1 (patient 5) had progressive cyanosis leading to surgical reinclusion of hepatic veins, and the other (patient 2) had reinclusion of hepatic veins into the systemic venous pathway at the time of fenestration. Seven of the 9 patients underwent hemi-Fontan operation at least 6 months before Fontan operation; all had palliative procedures performed within the first 3 months of life.
Indications for fenestration included severe chronic effusions (n=4) or PLE (n=5). Patients with effusions had pleural/pericardial drainage for a duration of 3 to 16 months (mean, 9±6 months) after surgery. Effusions persisted despite frequent pleurocenteses, diuretic therapy, and administration of ACEIs. In addition, 2 patients failed to respond to intrapleural steroid administration. Patients with PLE had onset at 6 months to 5 years after Fontan operation. Symptoms included diarrhea, peripheral edema, and ascites. PLE was diagnosed by serum albumin level <3 g/dL, total protein <5 g/dL, and elevated stool α1-antitrypsin clearance. Three patients underwent small-bowel biopsy, with normal histology in 1 and lymphangiectasia noted in 2. Treatment included parenteral albumin infusions, diuretics, and afterload reduction without resolution. In addition, 3 patients (patients 5, 6, and 7) were treated with oral steroids with some success (transient rise in serum protein levels); however, none could be successfully weaned without return of PLE.
Diagnostic cardiac catheterization was performed before consideration for surgical fenestration. The range of aortic saturation was slightly lower in the patients with effusions (88% to 93%) than in those with PLE (90% to 95%). No patient had angiographic evidence of significant right-to-left shunt. Arterial desaturation in the effusion group was attributed to lung disease and chronic atelectasis. Pulmonary artery pressures were higher in the effusion group (16 to 20 mm Hg) than in the PLE group (9 to 15 mm Hg). One patient in the effusion group had severely diminished right ventricular systolic function (patient 1); 1 patient in the PLE group had an atretic left pulmonary artery after previous multiple attempts at surgical reconstruction (patient 5). Catheter interventions performed before surgery included coil embolization of prominent aortopulmonary collaterals in 3, dilation of mild left pulmonary artery stenosis in 2, and baffle fenestration in 1, all without significant abatement of symptoms.
A right atriotomy was performed exposing the pulmonary venous side of the lateral tunnel. The presence of a small fenestration (2 to 3 mm) previously created via catheter intervention was noted in one patient (patient 4). A thin layer of fibrous neointima was peeled off the pulmonary venous side of the baffle, exposing a portion of the PTFE patch, which was incised, and a coronary punch was inserted. Two or three small (3-mm) punches were inserted in patients with effusions; those with PLE had creation of a single, large (5-mm) defect.
All patients survived surgery. One child with effusions and low cardiac output (patient 4) was on long-term ventilation and had severe multisystem organ failure before surgical fenestration. No symptomatic relief occurred, and death ensued 7 weeks after surgery. The other 8 patients were discharged home, on ACEIs, diuretics, and aspirin, 5 days to 6 weeks after surgery.
Resolution of effusions occurred in patients 1, 2, and 3 (small, multiple fenestrations) within 4 to 6 weeks after surgery; resolution of PLE (single, large fenestration) occurred in patients 5, 6, and 7 with normalization of serum proteins and disappearance of edema and ascites within 2 to 6 weeks after surgery. Discharge room air arterial saturations in these 6 successfully treated patients were ≤86%. PLE persisted in patients 8 and 9; discharge room air arterial saturations after fenestration in these two patients were relatively high, 90% and 89%, respectively. Patient 8 had evidence on echocardiography of an intimal flap partially occluding the fenestration at discharge, and patient 9, the oldest in the series, had a large, 5-mm unobstructed defect.
Follow-up for the 3 patients with effusions who responded to fenestration was 25 to 30 months at the time of this writing. Patients 1 and 2 had spontaneous closure of their multiple small defects as detected by echocardiography within the first year after surgery, with room air arterial saturations rising to 90% to 92%. Both patients have remained effusion free for 26 and 25 months after surgery, respectively. Patient 3 had spontaneous closure of the multiple small defects 11 months after surgery, with a rise in arterial saturation to 90%, and large bilateral pleural effusions returned. Twelve months after initial fenestration, the patient was taken back to the operating room for refenestration. Inspection of the PTFE patch revealed a glistening, smooth layer of fibrous neointima on the pulmonary venous side covering the previous fenestration sites. No thrombus was noted. A large, single 5-mm fenestration was placed. Resolution of effusions occurred 4 weeks thereafter. At follow-up 18 months after the second fenestration, the patient remains effusion free, with a large, 4- to 5-mm defect seen on echocardiography and room air arterial saturation of 84%.
Follow-up for the 3 patients with PLE who responded to fenestration is 24 to 28 months (Figure⇓). All are maintained on ACEIs and aspirin. Patient 5 exhibited lability in serum protein levels 18 months after surgery, with return of mild edema and total protein level of 4.1 g/dL and albumin of 2.1 g/dL. Echocardiography showed a large fenestration defect. The patient was placed on a small dose of diuretic, responded with an increase in serum proteins, and at 28 months after surgery is symptom free, with normal serum proteins and arterial saturation of 84%. Patients 6 and 7 are both symptom free without hypoproteinemia >2 years after surgery; on echocardiography, large defects (4 to 5 mm) are present in both, with arterial saturations of 82% to 86%.
Of the 2 patients whose PLE persisted despite fenestration, patient 8 continues to exhibit abnormally low levels of serum protein and is being treated conservatively with diuretics at the referring institution. Patient 9 underwent balloon dilation of the defect, with consideration given to the possibility that the 5-mm defect was not large enough relative to the patient’s body size to promote adequate right-to-left shunting. Room air arterial saturation before dilation was 90%. After dilation with a 12-mm-diameter balloon, arterial saturation was 80%, with angiography confirming right-to-left shunting at the defect level. Within 2 weeks, however, arterial saturation once again rose to 90%. The child subsequently underwent further surgery, with removal of the systemic venous baffle and placement of a new baffle with creation of a larger 6-mm fenestration. The postoperative course was complicated by stroke, recurrent atrial arrhythmia, ventricular dysfunction, continued hypoproteinemia, and large pleural effusions. The patient subsequently underwent heart transplantation and has had resolution of PLE.
This report demonstrates that late surgical fenestration is useful in alleviating morbidity after a Fontan operation. The cause of chronic effusions and PLE after Fontan is unknown, and the mechanism responsible for resolving morbidity after fenestration is similarly uncertain. Improved hemodynamics may result from creation of a baffle defect that augments systemic blood flow and lowers central venous pressures. However, as noted in our own patients with PLE, post-Fontan morbidity may be manifested despite the presence of relatively low central venous pressures. It has been demonstrated that although arterial saturation diminishes, cardiac index increases after fenestration, resulting in improved blood flow and increased oxygen delivery.5 9 By reducing resistance to venous return and increasing arterial flow, fenestration may result in an improved “perfusion profile” to end organs, resulting in resolution of the as yet undefined pathophysiological process.
Though limited, our data support the view that a critical level of right-to-left shunt must be achieved before the salutory effects of fenestration are exhibited. Improvement in our patients was noted only when arterial saturations were ≤86%, consistent with diversion of approximately one third of the systemic venous return across the fenestration.10 When mixed venous saturation is very low secondary to markedly diminished cardiac output (ie, ventricular dysfunction), extremely low arterial saturation values would be necessary for this degree of right-to-left shunt to be manifested. This may explain why patient 4, with multisystem organ failure, and patient 9, after second fenestration with atrial flutter and ventricular dysfunction, did not respond favorably to fenestration. Successfully achieving very low arterial saturations in such patients may be both difficult and clinically prohibitive.
The mechanism for spontaneous fenestration closure appears to be fibrous neointimal proliferation and not thrombus formation.11 This was evident in patient 3, in whom a smooth, thin layer of neointima was visualized completely covering the previous defect sites at second surgical fenestration. Small fenestration defects appear to undergo spontaneous closure at faster rates than large defects because of the relatively smaller gap necessary for the proliferating cells to cross. When placed de novo at the time of initial Fontan operation, large defects have been demonstrated to close spontaneously7 ; in our patients with large defects, however, very little diminution in size on echocardiography has been identified up to 2 years after surgery. It is conceivable that on the basis of the pathophysiological process at hand, these patients are unique in that there is a relatively high level of flow across the fenestration, which may in itself promote a delay in the spontaneous closure process.11
The patients described in this study underwent surgical fenestration; however, it is likely that transcatheter fenestration can accomplish similar goals. In our experience, balloon dilation of catheter or surgically created baffle defects resulted in only transiently adequate fenestrations, as judged by arterial saturation and echocardiography. Inspection of the baffle in one of these at reoperation (patient 9) revealed linear tears, with little increase in effective orifice. Recently, we have modified the technique in a patient with PLE. After Brokenbrough needle puncture of the Fontan baffle under transesophageal echocardiographic guidance, we created a flap by cutting the baffle two times with a Park blade, creating a V-shaped incision. Two months after the procedure, the arterial saturation remains 81% and serum proteins are normal.
In conclusion, late fenestration is an effective means of alleviating serious morbidity from effusions or PLE after Fontan operation and should be considered in some cases instead of Fontan connection takedown or heart transplantation. Resolution of symptoms appears to be related to the degree of right-to-left shunt created. The precise mechanism for relief of symptoms can only be speculated upon at this time. Whether spontaneous closure of the large fenestrations will result in return of morbidity is unknown.
Selected Abbreviations and Acronyms
We would like to acknowledge Dr Thomas L. Spray for contributing to the care of one of the patients in this report.
- Received February 18, 1997.
- Revision received April 28, 1997.
- Accepted April 30, 1997.
- Copyright © 1997 by American Heart Association
Bridges ND, Lock JE, Castaneda AR. Baffle fenestration with subsequent catheter closure: modification of the Fontan operation for patients at increased risk. Circulation. 1990;82:1681-1689.
Bridges ND, Mayer JE Jr, Lock JE, Jonas RA, Hanley FL, Keane JF, Perry SB, Castaneda AR. Effect of baffle fenestration on outcome of the modified Fontan operation. Circulation. 1992;86:1762-1769.
Mavroudis C, Zales VR, Backer CL, Muster AJ, Latson LA. Fenestrated Fontan with delayed catheter closure: effects of volume loading and baffle fenestration on cardiac index and oxygen delivery. Circulation. 1992;86(suppl II):II-85-II-92.
Mertens L, Dumoulin M, Gewillig M. Effect of percutaneous fenestration of the atrial septum on protein-losing enteropathy after the Fontan operation. Br Heart J. 1994;729:591-592.
Hijazi ZM, Fahey JT, Kleinman CS, Kopf GS, Hellenbrand WE. Hemodynamic evaluation before and after closure of fenestrated Fontan: an acute study of changes in oxygen delivery. Circulation. 1992;86:196-202.
Laks H. The partial Fontan procedure: a new concept and its clinical application. Circulation. 1990;82:1866-1867. Editorial.