Mixed Type of Total Anomalous Pulmonary Venous Connection: Outcomes and Novel Classication

Purpose To summarize the diagnosis and treatment of 13 patients with mixed-type total anomalous pulmonary venous connection (TAPVC) and propose a novel classication for mixed TAPVC. Methods A retrospective review of 13 with mixed a The diagnosis of mixed-type TAPVC was made in all patients using echocardiography combined with computed tomography angiography. According to the mixed TAPVC anatomy, there were 3 patients with type I TAPVC (2+2 veins), 10 patients with type II TAPVC (3+1 veins) and no patients with type III TAPVC. Correspondingly, there was 1 patient with the “SVC+VV” subtype, 2 patients with the “CS+C” subtype, 8 patients with the “CS+VV” subtype, 1 patient with the “CS+SVC” subtype and 1 patient with the “RA+SVC” subtype according to our classication system. All patients underwent cardiopulmonary bypass surgery.


Introduction
Total anomalous pulmonary venous connection (TAPVC) is divided into supracardiac, cardiac, infracardiac and mixed types 1 . With the accumulation of experience in surgical treatment and intensive postoperative care, the outcomes of TAPVC have improved over time, with reported mortality rates consistently < 10% 2 . However, mixed-type TAPVC, in which more than one level of pulmonary venous drainage exists, seems to be the most problematic subgroup. In some series, the mortality rate of patients with mixed TAPVC is still high 2 , but because of the rarity of this condition, few reports have focused exclusively on this subtype, and it is di cult to derive enough experience regarding surgery. A useful classi cation system for mixed type TAPVC that can indicate a suitable surgical approach is di cult to formulate due to the variability of the pulmonary venous connections, although one classi cation for subtypes of mixed TAPVC has been proposed 3,4 . The diagnosis and treatment of mixed TAPVC are still challenging and can be daunting. The aim of the present study was to retrospectively assess the results and to share our experience in diagnosing and surgically treating several cases of mixed TAPVC and propose a novel classi cation that can provide information related to the selection of the surgical method.
Patients And Methods TAPVC can be classi ed into 4 subtypes, and the mixed type is the rarest form, accounting for only 5-10% of TAPVC cases 3 . The surgical approach varies for different types of mixed TAPVC. Accurate diagnosis, especially regarding how all four pulmonary veins drain, has been emphasized as a critical factor of achieving improved results. We referenced the classi cation of mixed TAPVC subtypes proposed by Chowdhury et al 4 and propose a new classi cation based on our experience.
The classi cation we propose is based on each site of pulmonary vein drainage into the systemic venous system and how all four pulmonary veins drain, such as whether a site of pulmonary venous con uence is present, which veins drain to the con uent site, and whether pulmonary venous obstruction (PVO) is present. We analyzed the published literature 1-4 focusing on the morphology of the mixed subtypes of TAPVC and summarized the sites of the pulmonary veins. The pulmonary venous drainage sites were always as follows: (1) Coronary sinus (CS): In this group, one or more pulmonary veins are connected to the CS, and the key point of this group is the site of the drainage rather than whether the pulmonary veins form a con uent site or the number of pulmonary veins. (2) Con uence (C): In this group, two or more pulmonary veins form a con uent site, which is then connected to the vertical vein (VV), the superior vena cava (SVC) or inferior vena cava (IVC). The key point of this group is that a con uent site or common chamber is formed. (3) VV: Usually one pulmonary vein drains to the brachiocephalic vein via the VV. The key point is that no con uent site or common chamber is formed and the blood drains to the brachiocephalic vein. (4) Right atrium (RA): Pulmonary veins are connected directly to the RA. (5) SVC: One or more pulmonary vein drains to the SVC, not the brachiocephalic vein, without forming a con uent site or common chamber. (6) IVC: One or more pulmonary vein drains to the IVC without forming a con uent site. (7) Bizarre anatomical variants (BAVs): In this group, the anatomical variants were bizarre and required pulmonary venous rechanneling with individualized surgical techniques. The classi cation we propose is based on the various combinations of drainage sites, such as "CS + C" or "SVC + RA + CS" 5 .
In this retrospective analysis, we reviewed the clinical records of patients who underwent repair for mixed TAPVC between January 2010 and November 2019. All of these patients underwent biventricular correction. Patients with a single ventricle, atrial isomerism and other complex congenital heart diseases were excluded. Patient characteristics, such as weight and age comorbidities, pulmonary venous anatomy, demographics, and clinical, operative, postoperative, and follow-up data were recorded.
A total of 231 patients who were diagnosed with TAPVC underwent surgical repair at the Children's Hospital of Chongqing Medical University. 13 patients (5.6%) had mixed TAPVC; of them, 11 were male, and 2 were female. The median weight at surgery was 4.6 ± 1.0 kg (3.4-7.3 kg), and the median age at surgery was 96.2 ± 81.2 days (10-242 days). 10 patients were younger than 6 months, including 3 neonates. Preoperative PVO occurred in 3 patients (Table 1). The pattern distribution among the 13 patients is shown in Table 2. 3 patients (23.1%) had type I TAPVC.
In this subtype, the right pulmonary veins in 2 patients (the fourth and twelfth patients) connected to the CS and left superior and inferior veins to form a site of con uence with the brachiocephalic vein via a VV ("CS + C" subtype in our classi cation system); in the other patient (the eighth patient), the right superior and inferior veins formed a site of con uence and connected to the IVC, while the left superior and inferior veins formed a site of con uence and connected to the brachiocephalic vein via a VV ("C + C" subtype in our classi cation system). 10 (76.9%) patients had type II TAPVC. In this group, the right pulmonary vein and the left inferior pulmonary vein in 8 (80%) patients connected to the CS, and the left superior pulmonary vein connected to the brachiocephalic vein via a VV ("CS + VV" subtype in our classi cation system). In 2 patients, the left pulmonary vein and the right inferior pulmonary vein connected to the RA, and the right superior pulmonary vein connected to the SVC ("RA + SVC" subtype in our classi cation system).  Surgical correction in cases of type I TAPVC (the fourth and twelfth patients) involved complete unroo ng of the CS wall into the LA and patch closure of the atrial septal defect (ASD) (similar to treatment of the cardiac type); then, the con uent was anastomosed to the LA. Surgical correction in the eighth patient involved anastomosing the left con uent section to the LA and the right con uent section to the back of the RA, followed by ba ing to the LA to avoid the excessive tension involved in anastomosis to the LA. Surgical correction in cases of type II TAPVC (the rst, third, fth, sixth, seventh, ninth, tenth and thirteenth patients) involved complete unroo ng of the CS wall into the LA, anastomosis of the VV to the LA or left atrial appendage (LAA) after transection, and closure of the VV near the innominate vein. In the second and eleventh patients, a Warden procedure was carried out, and the left pulmonary veins were rechanneled to the LA with ba ing while closure of the ASD redirected the pulmonary veins to the LA (van Son procedure).

Results
Hospital mortality was de ned as mortality occurring less than 30 days after surgery or before discharge. All other deaths were de ned as late mortality. ASD closure was performed in all patients. There was 1 case (the third patient) of late mortality because of PVO 7 months after surgery. The patient developed metabolic acidosis and respiratory distress that required intubation. The parents refused surgical treatment when echocardiography showed pulmonary venous ow at the anastomotic of 2.9 m/s. No other reoperations for postoperative PVO were required. Follow-up data were obtained for all survivors except the rst patient at an average of 3.4 ± 2.2 years (5 months to 8 years) after surgery, and the remaining patients were asymptomatic at the last follow-up. The survival rates at 3 and 5 years were both 90.9% ± 8.7% (95% CI: 73.8%-108%) (Fig. 1).

Discussion
Mixed TAPVC is the least frequent subtype, accounting for 5-10% of patients with TAPVC in large series; the corresponding proportion in our series is 6.4%. Because of the unpredictability of the pulmonary venous connections, mixed TAPVC is also the most problematic type and requires a combination of techniques, including the individualized anastomosis of pulmonary veins. An individually tailored operation should be formulated to successfully manage the highly variable anatomical pulmonary vein connections. Therefore, a correct preoperative understanding of the anatomy and an accurate description of all anomalously draining pulmonary veins are the rst issues to be addressed 6 .
Echocardiography is a valuable, safe, cost-effective and easily reproducible imaging tool that should always be the primary mode of diagnosis. It can provide anatomical details regarding the position of anomalous pulmonary venous connections and the presence of venous obstruction in the majority of patients. In most cases, echocardiography is su cient and preferable. Because of the highly variable anatomy, as noted above, echocardiography can sometimes identify only two or three veins, which can lead to misdiagnosis. Some reports 7 have shown that the mixed type is the most di cult type to diagnose, with 4 (44.4%) patients who were diagnosed only after surgical examination 6 . In order to identify all four veins, computed tomography angiography should be applied, which can provide not only an accurate description of the pulmonary venous drainage pattern but also delineate the pulmonary vein anatomy in patients with suspected asymmetrical PVO, despite the high radiation exposure 8,9 . A merit of computed tomography angiography is three-dimensional reconstruction, which can provide a precise noninvasive visual description of the pulmonary vein connections 8,9 . Cardiac catheterization is the third diagnostic method that can be considered. Due to the invasiveness, which will worsen the nal outcome, it is not recommended as a routine examination 4  the superior vena cava was performed using the Warden procedure, in which the SVC was transected above the highest anomalous pulmonary vein, and the cardiac end of the SVC is oversewn. A ba e is then constructed in such a fashion that the right atrial ori ce of the SVC is connected to the LA and closure of the ASD is achieved; then, the cephalad end of the SVC is sewn to the right atrial appendage.
Many other surgical techniques have been described and proven to be useful and safe. 6 (6) ICV: In these patients, the veins were anastomosed to the back of the RA and then ba ed to the LA; sometimes, the veins can be anastomosed to the LA directly 10 . (7) BAVs: In this group, the anatomical variants were bizarre and required pulmonary venous rechanneling on an individualized basis. The operative details for each morphological subtype should apply to the surgical approach and even allow two or more approaches to be combined 3 . The classi cation we propose is based on the various combinations of drainage sites combined with the corresponding approaches. The "CS + VV" pattern was most frequent in our series (61.5%), and the same tendency was found in most other studies 3,4 .
Rarely, patients may require the management of an isolated pulmonary vein or the drainage of a solitary pulmonary vein from one lung at different levels, which must be individually tailored. An obstructed isolated pulmonary vein mandates rechanneling into the LA, whereas unobstructed isolated anomalous pulmonary venous drainage may be left uncorrected, can be corrected during a second operation if needed, or may require removal of the anomalously connected lobe, which is well tolerated even in infants 11,12 . Patients with a single uncorrected vein usually have an excellent long-term outcome; after all, a single anomalously draining vein account for approximately 20% of the total pulmonary blood ow 12 .
Unlike the surgical results of mixed TAPVC patients described in other studies, in our series, the statistical analyses showed a good outcome regarding early mortality, with an occurrence rate of 7.7%. With satisfactory primary repair, the pulmonary venous ow at the site of anastomosis was less than 1.2 m/s, as shown by postoperative echocardiography, which is considered adequate 13 . To achieve good outcomes, the following tips can be applied during anastomosis: (1) Sutureless repair, which has been reported to result in no mortality or reintervention, can be applied. Patients treated with sutureless repair or in whom TAPVC was partially unrepaired revealed reasonable early and medium-term physiological tolerance, without the need for reintervention 14 .
(2) Ensure precise geometric alignment and avoid tension, torsion, and rotation, making as large an anastomotic site as possible. (3) Maintain patency of the foramen ovale in patients with moderate-to-severe pulmonary hypertension for decompression of the right-sided chambers in the event of a pulmonary hypertensive crisis.
There are many limitations to our research. This was a retrospective single-center study, and the sample size was small. Additionally, the types of mixed TAPVC in our series are the more common types, unlike those in some other studies, which presented di culties during the operation. The classi cation also has limitations; it cannot account for all potential sites of pulmonary vein drainage, and the more uncommon types are classi ed as BAVs in our classi cation.

Conclusion
In summary, mixed TAPVC is a comparatively rare variant of TAPVC. The diagnosis and accurate description of the pulmonary venous anatomy has important implications in the surgical treatment, and the variable geometry of multiple sites of pulmonary venous con uence is challenging. Echocardiography combined with computed tomography angiography is su cient in most cases. The surgical scheme is determined by the drainage sites of the pulmonary veins. Therefore, we propose a novel classi cation for the rst time on the basis of the drainage sites to guide the surgical approach. Operative survivors can achieve good long-term survival similar to that of patients with other forms of TAPVC if a suitable approach is applied to allow the anastomosis to be unobstructed. Declarations Ethics approval and consent to participate:This research received ethical approval and consent.And it didn't require the approval of an ethics committee.
Consent for publication All patients' information contained in this manuscript has been published with the consent of their parents.
Availability of data and material The data and materials in the manuscript are available, and the original data for the relevant results is owned by myself and can be contacted if it's needed. Figure 1 Kaplan-Meier survival curve for all patients.