Hypoplastic left heart syndrome is a fatal disease if left untreated, with a mortality rate of 90–95% in the first month. Reference Sano, Huang, Kasahara, Yoshizumi, Kotani and Ishino1 The classical Norwood stage 1 procedure, Reference Norwood, Lang and Hansen2 Sano modification, Reference Sano, Ishino and Kawada3 and hybrid approach Reference Gibbs, Wren, Watterson, Hunter and Hamilton4 can be considered as the three options for palliation in stage 1 surgery for hypoplastic left heart syndrome. Early mortality of Norwood Stage I surgery decreased from initial 80% to 10–16% in specific centers recently. Reference Iannettoni, Bove and Mosca5–Reference Mascio, Irons and Ittenbach8 It is well known that high-volume centres with more experience in hypoplast surgery present better results. Reference Pontailler, Gaudin and Lenoir9,Reference Pasquali, Jacobs and He10 Early mortality of Norwood Stage I surgery in our country is around 40–60% (authors personal communication). Reference Özdemir, Korun and Dedemoğlu11 Our cent has been approved four years ago, therefore with limited experience and with the burden of no patient selection, this apparently high early mortality rate is actually comparable to results of similar centres in the country. We presented our method earlier as a case report, in which we used a part of the pulmonary artery wall as the new systemic pathway. Pulmonary flow was maintained with the aortopulmonary window that was placed in the systemic pathway. Reference Çelik and Özkan12 Here we present the short-term results of our limited patient series for whom we performed our own method of palliation.
Material and methods
Patient selection
The introduced technique was used in all patients with hypoplastic left heart syndrome and its variants between February 2019 and March 2021. Twelve of them were diagnosed as hypoplastic left heart syndrome, and four had hypoplastic left heart syndrome variants. All patients underwent detailed physical examination, laboratory tests, and transthoracic echocardiography and most of them underwent angiographic imaging. Necessary information was provided to all patients’ relatives, and their written informed consents for surgery were obtained. This study was approved by Baskent University Institutional Review Board (Project no: KA21/534).
Operative style and technique
The thymus was removed following a median sternotomy in all patients. The arterial cannula was placed in the ascending aorta, innominate artery, or ductus arteriosus, depending on anatomical variations. A “U”-shaped arteriotomy was performed on the anterior aspect of the main pulmonary artery, with its opening facing cephalad, and a flap of the native pulmonary artery was created. The borders of the incision were limited to at least 1 mm distal to the pulmonary valve commissures and proximal to the left and right pulmonary artery orifices. A fenestrated bovine pericardial patch (Edwards bovine pericardial patch, Edwards Lifesciences, Irvine, CA, USA) was placed at the distal pulmonary artery to isolate the branch pulmonary arteries and create a new pulmonary outflow pathway. This patch was fenestrated using a 2.5 or 3 mm aortic punch according to patient weight and was held down with 6/0 propylene suture (Propilen®,Dogsan, Trabzon, Turkey). Thus, pulmonary flow was limited by the size of the fenestration on the patch (Fig 1). A new systemic pathway was created by suturing the rectangular bovine pericardium with 6/0 propylene, which was prepared to fit the pulmonary artery flap (Fig 2).
Figure 1. A 3-mm aortopulmonary window was created between the new systemic flow path and the pulmonary artery.
Figure 2. New systemic flow path.
The aforementioned procedures were performed after the ductus arteriosus was divided in patients for whom aortic cannulation was performed. A blood-free area was provided by snaring the arterial cannula in patients for whom ductus arteriosus cannulation was done. Both systemic and pulmonary blood flow pathway repairs were performed on cardiopulmonary bypass with a perfused heart.
Diastolic arrest was achieved by administering single-dose antegrade cold crystalloid cardioplegia through the usual aortic root cannula or arterial cannula under deep hypothermic circulatory arrest. Atrial septectomy was performed using the standard method. An aortotomy was performed on the aortic arch, starting at least 2 mm distal to the innominate artery and extended to the descending aorta to match the new systemic pathway. All ductal tissue was completely resected and in cases with associated aortic coarctation aortotomy was carried further distally. The new systemic pathway was anastomosed to the aortotomy with 6/0 propylene (Fig 3). In patients with aortic interruption, the medial half of the descending aorta was first sutured with 6/0 propylene to the adjacent half of the new systemic pathway. After appropriate aortotomy was performed on the ascending aorta, it was anastomosed to the new systemic pathway and descending aorta system with 6/0 propylene. After de-airing, total body perfusion was restarted. Cardiopulmonary bypass support was terminated with the usual dose of inotropes (video 1). After hemostasis, the sternum was left open in all patients, and the skin was closed with running 3/0 propylene. Sternal closure was performed within 24 h after the operation in six patients, and after 24 h in intensive care under sterile conditions in three patients. Acetylsalicylic acid 5 mg/kg was given to all patients in the postoperative period.
Figure 3. New systemic flow path–aorta anastomosis.
For the patient who underwent stage 2 bidirectional cavopulmonary shunt, aorta, and superior caval vein were cannulated, and cavopulmonary anastomosis was completed under partial cardiopulmonary bypass, sparing the antegrade pulmonary blood flow to be closed at stage 3 Fontan procedure.
Postoperative care and measurements
A standard hourly follow-up was performed in the ICU. Postoperative vancomycin was started in all patients, with the sternum left open. Mediastinal specimens for culture were collected from all patients during sternal closure. Antibiotic treatment was updated based on culture results. Urine output was checked every hour, and peritoneal dialysis was performed if needed. Enteral feeding was started and gradually increased.
Statistical analysis
Statistical analyses were performed using SPSS Statistical Software. Descriptive and frequency data were analysed separately. The minimum and maximum values and other results after statistical analyses are presented in the article.
Results
All repairs were performed using the previously described method on 16 patients (eight female and eight male patients) between February 2019 and March 2021. Reference Çelik and Özkan12 The mean age of the patients was 7.56 ± 8.41 days, and mean body weight was 3.07 ± 0.5 kg. Three patients were premature babies, and three patients weighed less than 2.5 kg. Surgery was performed on the day of hospitalisation in one patient, and in six patients, it was performed the day after admission. Nine patients underwent surgery within five days. None of the patients had genetic anomalies. In eight patients, preoperative mechanical respiratory support was required. Lactate levels of 14 patients were above normal limits (mean lactate levels 3.88 ± 1.9 mmol/L).
Ten patients had low preoperative pH levels. Only one of the operated patients was born in our hospital, while eight were referred from other centres in the province and seven were referred from other provinces.
Twelve patients were diagnosed with hypoplastic left heart syndrome, and four of them were hypoplastic left heart syndrome variants. Eight of the 12 patients who were diagnosed with hypoplastic left heart syndrome had an anatomical structure of aortic atresia and mitral atresia, while four had aortic atresia and mitral stenosis. In all patients with hypoplastic left heart syndrome, retrograde flow in the ascending aorta was noted. The mean diameter of the ascending aorta was 3.12 ± 1.14 mm. Preoperative tricuspid valve insufficiency of the patients was as follows: minimal in two patients, mild in five, moderate in four, and severe in one. Two patients had a ventricular septal defect and cor triatriatum sinistra. Demographic data and detailed diagnoses of the patients are presented in Table 1.
Table 1. Preoperative data
Abbreviations : aa: aortic atresia, as: aortic stenosis, HLHS: hypoplastic left heart syndrome, IAA: interrupted aortic arch, LA: left atria, LV: left ventricle, ma: mitral atresia, ms: mitral stenosis, m. outlet: membranous outlet, p. outlet: perimembranous outlet, RV: right venticle, VA: ventriculo arterial, VSD: ventricular septal defect.
1 Patient 2 has HLHS variant, and her systemic atrioventricular valve is tricuspid valve.
2 Patient 10 has HLHS variant, and she has two patent atrioventricular valve and both of them have mild regurgitation.
3 Patients 15 and 16 have HLHS variant and their systemic atrioventricular valve is mitral valve. Given information is about mitral valve.
Ductus arteriosus cannulation was performed in eight patients, ascending aortic cannulation in six, innominate artery cannulation in one, and double arterial cannulation (innominate artery + ductus arteriosus) in one. Mean cardiopulmonary bypass time was 158.5 ± 23.6 mins, and mean cross clamping time was 40.5 ± 15.49 mins. Deep hypothermic circulatory arrest guided repair was performed in all patients except two. The mean deep hypothermic circulatory arrest time was 33.2 ± 7.2 mins. In two patients, repair was performed with antegrade selective cerebral perfusion and moderate hypothermia. In one of the patients who underwent repair with antegrade selective cerebral perfusion, a 5-minute total circulatory arrest was performed while providing cardioplegia through the aortic cannula. Atrial septectomy was performed in 14 patients during deep hypothermic circulatory arrest. Conventional ultrafiltration was performed in 12 patients (mean, 600 ± 361.9 ml).The operative data are presented in Table 2.
Table 2. Operative data
APW: aortopulmonary window; ASCP: antegrade selective cerebral perfusion; CPB: cardiopulmonary bypass; DHTCA: deep hypothermic total circulatory arrest; X clamp: cross clamp.
Seven patients (43.7%) died during the early postoperative period. Four of them were lost due to low cardiac output and three died because of sepsis. Nine patients (56.3%) were discharged. The mean length of ICU stay and hospital stay were 42.6 ± 24.5 days and 47.6 ± 23.3 days, respectively. One patient died at home waiting stage 2 surgery. Three patients underwent stage 2 surgery. Five patients are waiting for stage 2. Follow-up duration for all survivors was 23 months.
Discussion
Hypoplastic left heart syndrome, with a mortality rate of 90% in the first month without treatment, is the most fatal CHD. Reference Sano, Huang, Kasahara, Yoshizumi, Kotani and Ishino1,Reference Pontailler, Gaudin and Lenoir9,Reference Pasquali, Jacobs and He10,Reference Morris, Outcalt and Menashe13 Although the mortality rate of surgery has decreased considerably during the recent era, it remains high. Reference Sano, Huang, Kasahara, Yoshizumi, Kotani and Ishino1 Highly experienced institutions have a mortality rate of 15% after stage 1 surgery, while mortality rates reach 25–33% in most centres. Reference Pontailler, Gaudin and Lenoir9 In centres with low patient density (<16 patients/year), mortality rate can be as high as 75%. Reference Pasquali, Jacobs and He10
The ideal approach for hypoplastic left heart syndrome is planning the treatment after making a prenatal diagnosis and providing detailed information to the family. Among the treatment options, a decision should be made whether to terminate the pregnancy, provide postpartum comfort care, or perform surgery. If surgical treatment is considered, delivery should be conducted at the centre where hypoplastic left heart syndrome stage 1 surgery will be performed. Although different techniques are defined as surgical treatment options, there are three techniques most commonly used nowadays: the classical Norwood stage 1 surgery, Reference Norwood, Lang and Hansen2 the Sano modification, Reference Sano, Ishino and Kawada3 and the hybrid approach. Reference Gibbs, Wren, Watterson, Hunter and Hamilton4
As Norwood stage 1 surgery and the Sano modification have high mortality rates, hybrid procedures are preferred in some centres. However, patient selection remains debatable. All three methods have disadvantages and advantages. In the classical Norwood stage 1 surgery, there are serious problems such as being an extensive surgery, long cardiopulmonary bypass time, difficulty shaping the patch for the systemic pathway, the existence of a modified Blalock–Taussig shunt, which may potentially cause low diastolic pressure, bleeding, thrombosis of the shunt, and pulmonary artery distortion as a late complication. Moreover, it is also known that the smaller the diameter of the graft used for the modified Blalock–Taussig shunt, the greater the risk of developing stenosis and reintervention. Reference Myers, Ghanayem and Cao14,Reference O’Connor, Ravishankar and Ballweg15 The difference between the Sano modification and the classical Norwood stage 1 procedure is the preference of the shunt. Placing a shunt to the pulmonary artery from the anterior aspect of the right ventricle prevents low diastolic pressure but makes it necessary to perform a ventriculotomy on the right ventricle, which is the only systemic ventricle; it can thus be speculated that a Sano shunt may cause right ventricular dysfunction, right ventricular volume load, and arrhythmia. The aim of the hybrid procedure is to choose a less invasive method for the first stage and perform comprehensive surgical repair in the future. However, the success of this procedure, which requires a multidisciplinary approach and experienced staff, depends on the harmony of different teams and timing. This may be difficult to achieve in centres where the patient volume is low and mortality rates can be close to those of classical Norwood stage 1 surgery. Reference Pontailler, Gaudin and Lenoir9 Furthermore, stage 2 surgeries of patients who underwent the hybrid procedure are extremely difficult and complicated. Creating a systemic pathway is similar to that of the classical Norwood stage 1. Pulmonary artery reconstruction is also frequently required. Additionally, performing pulmonary artery plasty increases the mortality and morbidity for the bidirectional Glenn operation, which is performed during hypoplastic left heart syndrome stage 2 surgery. Reference Cleveland, Tran, Takao, Wells, Starnes and Kumar16
In our technique, we used part of the main pulmonary artery as a flap to create a neosystemic pathway and complete the rest of the systemic pathway with bovine pericardium during stage 1 surgery. Thus, at least one-fourth of the new systemic pathway is created with native tissue, which has a potential for growth. In addition, because the new systemic pathway is anastomosed to the aortic arch and the proximal descending aorta at a right angle without a curve, difficulties, such as fractures or kinking in the new systemic pathway and issues in tailoring the patch as in the classical Norwood stage 1 procedure, are avoided. We also believe that since the ascending aorta is left untouched during anastomosis, problems such as coronary flow disruption after classical repair will not be encountered, even in patients with a very small ascending aorta. Considering that patients with an aortic atresia anatomy and a very small ascending aorta have the highest mortality, we believe that our technique may be advantageous in this regard. Reference Tabbutt, Ghanayem and Ravishankar7
We used bovine pericardium for the pulmonary artery reconstruction; this was obliquely placed in the main pulmonary artery with a fenestration in the middle, and the size of which was determined according to the patient’s weight. We place the aortopulmonary window patch in such a way that the right and left pulmonary artery orifices remain distal to the patch and separate the pulmonary arteries from the new systemic pathway. As a result, a controlled aortopulmonary window was created.
Instead of using the aortopulmonary window as the source of pulmonary flow, a modified Blalock–Taussig shunt in the classical Norwood stage 1 procedure, a Sano shunt in the Sano modification, or bilateral pulmonary artery band in the hybrid procedure can be used. Although creating an aortopulmonary window causes low diastolic pressure similar to the Norwood stage 1 procedure, it was preferred over the modified Blalock–Taussig shunt in order to avoid bleeding or shunt thrombosis and distortion in the pulmonary arteries. The aortopulmonary window was likewise preferred over the Sano modification so as to avoid a right ventriculotomy, increased right ventricular volume load, right ventricular dysfunction, or arrhythmia. Finally, we decided against pulmonary artery banding as in the hybrid procedure in order to avoid the need for pulmonary artery reconstruction.
The aortopulmonary window diameter was decided based on the patient’s weight and was computed to be 0.5 mm less than the appropriate modified Blalock–Taussig shunt diameter. Reference Jonas18 In the modified Blalock–Taussig shunt, the inflow artery is the subclavian artery, which has a smaller diameter and flow than the aorta; furthermore, there is a certain length of the graft that limits pulmonary blood flow. In contrast, central shunts may result in unrestricted pulmonary blood flow. Classical central shunts such as the Waterson shunt, and the Potts shunt has many disadvantages, including difficulty in adjusting the anastomotic diameter, enlargement of the anastomosis over time, pulmonary overflow, high pressure in the pulmonary vascular bed, development of pulmonary vascular disease, flow preference to a single lung, distortions in the pulmonary artery, effects on the right ventricular outflow tract, and surgical difficulties during takedown, resulting in almost complete abandonment of these shunts over the years. Reference Cole, Muster, Fixler and Paul19,Reference Idriss, Cavallo, Nikaidoh, Paul, Koopot and Muster20 In our method, appropriately sized fenestrations can be created on a bovine pericardial patch with an aortic punch; since this tissue does not have growing potential, the diameter of the aortopulmonary window is guaranteed to remain stable. As a result, we believe that pulmonary overflow, pulmonary hypertension, and pulmonary artery distortion can be avoided. In addition, contrary to the difficulties in closure of classical aortopulmonary shunts, existing aortopulmonary windows can be easily closed with an arteriotomy to the systemic pathway.
Pontailler et al. presented a technique similar to ours and shared the results. For the systemic pathway, the ductus arteriosus was resected, and pulmonary homograft interposition was performed instead. They preferred bilateral pulmonary banding, as in the hybrid procedure, for the pulmonary flow source. They reported that hospital mortality and pre-stage 2 surgery mortality were significantly reduced when they compared this method with their own classical Norwood stage 1 series. Being a centre with low patient density, they stated that this may be a convenient surgical option for similar centres. Reference Pontailler, Gaudin and Lenoir9 Although we agree with this opinion, we believe that the difficulty in stage 2 surgery for patients who underwent the hybrid procedure may occur in Pontailler’s technique.
Three patients were carried to stage 2 surgery. A bidirectional Glenn operation was performed, leaving the antegrade pulmonary flow patent. There was no need for pulmonary artery reconstruction.
Although our technique has many advantages, it also has disadvantages. First, medium- and long-term results have not yet been documented. The development of recoarctation after stage 1 repair can be a problem for the classical Norwood anastomosis. Reference Carvajal, Canter, Abarbanell and Eghtesady17 In our technique, we divided the ductal tissue; there was no need to resect the ductal tissue in any of our patients because it did not prevent the performance of an aortotomy, which would fit the new systemic pathway and extend from the middle of the aortic arch to the proximal descending aorta. Although the residual ductal tissue is located on the roof of the new systemic pathway and is unlikely to cause coarctation in the future, the ductal tissue found on the pulmonary artery side may cause constriction of the pulmonary artery, requiring a surgery that will be more complicated than anticipated. We encountered this issue in one patient and had to perform a left pulmonary artery patch plasty.
The most undesirable aspect of our technique is its inability to clearly visualise pulmonary blood flow. Due to the anatomy, it is almost impossible to visualise the aortopulmonary window echocardiographically, which is in the middle of the new systemic pathway and has a posterior flow direction. Therefore, we believe that patients should be closely followed. Angiographic imaging of patients may produce similar difficulties. It seems unlikely to accurately measure the pulmonary artery pressure before stage 2, especially in patients with a small aortopulmonary window diameter. The decision for the second stage is made according to the oxygen saturation. The patients with low saturation are evaluated with pulmonary artery imaging. Distortion of pulmonary arteries is documented and managed with either interventional techniques or surgical reconstruction. Patients with well-developed branch pulmonary arteries with suitable pulmonary artery pressure, which is measured during the surgery undergo bidirectional Glenn. Patients with undersized pulmonary arteries, with or without reconstruction for distortion are further palliated with aortopulmonary window enlargement or a Blalock shunt as a second pulmonary blood source.
We have seven mortalities (43.7%) in our series, which seems to be high for early mortality. Four of them were lost due to low cardiac output and three died because of sepsis. Low birth weight, prematurity, severe tricuspid regurgitation, aortic atresia, small-sized ascending aorta, additional cardiac anomaly, non-cardiac anomalies, chromosomal anomalies, long CPB duration, long DHCA duration, and low pH values are risk factors for mortality in classical hypoplastic left heart syndrome stage 1 surgery (classical Norwood stage 1 and Sano modification). Although the survival rates in experienced centres are 77% for the classical Norwood stage 1 surgery and 92% for the Sano modification, surgical mortality increases in patients with one or more risk factors. Reference Sano, Huang, Kasahara, Yoshizumi, Kotani and Ishino1,Reference Gaynor, Mahle and Cohen21 Only one patient was born in our centre, while the others were referred from other centres and had to be transported. Every patient except two carried at least one risk factor. On the other hand, patients who were lost had at least three risk factors. All patients but one were referred from other centres and arrived in unstable condition. Therefore, patients were operated with poor general condition. Considering the learning curve process brought about by the nature of a new technique, we believe that the surgical mortality of our series is acceptable. Although long-term results are lacking, early mortality in our brief experience is comparable to the results of other centres in the country.
In conclusion, we believe that our technique, with its described advantages, is an acceptable alternative to other currently available methods used for hypoplastic left heart syndrome stage 1 surgery.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1047951122002682
Financial support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Conflict of interest
None.
