Off-pump pulmonary valvotomy for pulmonary atresia with intact ventricular septum: a safe and effective alternative with comparable outcomes

Abstract

Background

Right ventricle (RV) decompression for pulmonary atresia with intact ventricular septum (PA/IVS) could be achieved surgically or percutaneously, yet transcatheter valvotomy is not widely accessible in remote regions and classic on-pump pulmonary valvotomy procedure carries risks of by pass in neonatal patients. We aim to summarize our 15-year’s surgical experience treating PA/IVS with on-pump valvulotomy, off-pump valvulotomy and hybrid techniques.

Methods

From January 2010 to March 2025, 104 PA/IVS children who underwent off-pump, on-pump pulmonary valvulotomy or hybrid therapy in our hospital were retrospectively reviewed, including off-pump group (n = 56), on-pump group (n = 34) and hybrid group (n = 14). The postoperative mortality, operation time, early and mid-term follow-up of the three groups were analyzed.

Results

Both the off-pump and hybrid groups exhibited significantly shorter operative times, reduced vasoactive drug support, shorter intubation times, and shorter intensive care unit (ICU) and postoperative hospital stays compared to the on-pump group (P < 0.001). Cerebral, intestinal and renal oxygen monitor suggested that no obvious hypoxia was caused by clamping in off-pump procedure (P > 0.05). At final follow-up, RV hypoplasia, tricuspid Z score and pulmonary regurgitation were not significant among groups (P > 0.05). However, the hybrid group demonstrated a significantly higher rate of restenosis.

Conclusions

The off-pump pulmonary valvotomy is an effective, safe and technical-friendly operation. It provides superior early outcomes compared to on-pump surgery and a lower restenosis rate relative to hybrid techniques, highlighting its value for broader adoption across medical centers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-026-05721-z.

Introduction

Pulmonary atresia with intact ventricular septum (PA/IVS) is an uncommon type of cyanotic congenital heart disease, affecting 3 to 8 of every 100,000 living births, comprising fewer than 1% of all heart defects [1, 2]. PA/IVS exhibits a spectrum of morphologic abnormalities, such as right ventricle (RV) hypoplasia, tricuspid valve (TV) hypoplasia, and fistulae connecting the RV to the coronary circulation [3, 4]. Due to its high natural mortality in the neonatal stage, it is one of the few types of congenital defects that require early interventions soon after birth. However, both neonatal operation and the anatomical heterogeneity of PA/IVS itself necessitate great demands to its surgical techniques and strategies.

In patients with mild to moderate hypoplastic right ventricle, adequate tricuspid valve and without right ventricular-dependent coronary circulation (RVDCC), RV decompression is generally adopted as primary strategy [2]. It aims to create antegrade flow through pulmonary valves, which further promote the growth of RV and optimize the chance of biventricular correction. RV decompression now could be achieved surgically or percutaneously while the choice is various in different centers. The most classic and well-accepted way for RV decompression is direct pulmonary valvotomy under cardiopulmonary bypass (CPB). However, the 1-year survival of this on-pump procedure is merely 70–75% as reported, likely due to the intolerance of sternotomy and CPB in neonatal patients [5–7]. With the advances of catheterization in recent decades, PA/IVS could also be treated by transcatheter valvotomy and balloon dilation. Although this minimal-invasive procedure avoids CPB, it seems to be associated with considerable rate of procedural failure, complications and surgical reintervention in several studies [8]. Recently, a hybrid approach combining surgical access with a transcatheter intervention to open the RVOT was described. Potential advantages of the hybrid approach include avoidance of neonatal CPB, mitigation of the technical failure rate of the percutaneous approach, due to direct access to the RVOT, and a lower complication rate (i.e., no risk of RVOT perforation and associated sequelae) [9]. However, hybrid therapy’s limitations include the substantial need for specialized equipment and multidisciplinary collaboration, along with uncertainty in achieving complete stenosis relief. These constraints confine its application to large medical centers and preclude its use in remote areas.

To avoid the limitation and risks of above strategies, we carried out surgical pulmonary valvotomy without CPB as an alternative for PA/IVS patients in our center. This off-pump technique repairs pulmonary atresia under direct surgical vision and largely prevent the damage of CPB. However, it has not yet been rigorously compared with a contemporaneous cohort undergoing the standard surgical approach. Thus, we describe and summarize our experience performing pulmonary valvotomy for PA/IVS using both on-pump and off-pump techniques, as well as a hybrid approach, as the primary procedure. The off-pump approach shows considerable promise, given its commendable early and mid-term outcomes.

Methods

Study population

From January 2010 to March 2025, a total of 104 PA/IVS patients who received surgical pulmonary valvotomy or hybrid approach as primary operation in Shanghai Children’s Medical Center were retrospected. The study protocol conformed to the 1975 Declaration of Helsinki. Ethical issues were approved by Shanghai Children Medical Center Ethics Committee (SCMCIRB-K2022121-1). Informed consent was obtained from legal guardians. 34 patients received pulmonary valvotomy under CPB, 56 patients received pulmonary valvotomy without CPB as off-pump group, and 14 patients received hybrid approach as hybrid group. All patients underwent follow-up every 2 months in the first year, and once a year after. All patients met the inclusion criteria as following: (1) diagnosis as PA/IVS by echocardiography and enhanced CT; (2) Membranous atresia of the pulmonary valves (PV) with mild to moderate hypoplasia of RV; (3) Z value of TV > -4; (4) without RVDCC; (5) received only surgical pulmonary valvotomy or hybrid approach as primary operation, with or without Patent ductus arteriosus (PDA) closure or Blalock-Taussig (BT) shunt. Patients received PV transannular patch repair and patients with severe RV hypoplasia, RVDCC, or other concomitant conditions such as severe extra-cardiac malformations, severe nervous system diseases, infection or immune system diseases, were excluded.

Surgical techniques

The surgery for three groups was taken under median sternotomy. For the on-pump pulmonary valvotomy, CPB was established routinely after cannulation of aorta, superior vena cava and inferior vena cava. A transverse incision was then made on the anterior wall of the main pulmonary artery at about 1 cm above the pulmonary annular parallel to the cannula of CPB. After exposure of the fused PV, pulmonary valvotomy was performed under direct vision. Sutured the incision of the main pulmonary artery to complete the operation.

As shown in Fig. 1, for the off-pump pulmonary valvotomy procedure, inflow occlusion was initiated by clamping the distal end of the main pulmonary artery following block of superior and inferior vena cava. The same transverse incision was then made on the anterior wall of the main pulmonary artery, and gently set tractions on both side of the incision. Cut open the fused pulmonary valves and expanded the valvular orifice to 4 ~ 6 mm with vascular forceps, followed by quickly clamping the incision of the main pulmonary artery with lateral forceps and restoring the blood flow of the superior and inferior vena cava and the distal end of the pulmonary artery. Lastly sutured the incision of the main pulmonary artery to complete the procedure. A standardized anticoagulation protocol was followed: approximately 5 min prior to inflow occlusion, systemic heparinization was achieved with an initial intravenous dose of unfractionated heparin at 100 IU/kg. Activated clotting time (ACT) was measured using a point-of-care device before heparin administration (baseline), 5 min after administration, with a target ACT of > 250 s; if ACT fell below this threshold, an additional bolus of 25–50 IU/kg was given. Upon completion of the suture and confirmation of hemostasis, protamine sulfate was administered slowly over 10–15 min for heparin reversal at a dose of 1 mg per 100 IU of the initial heparin dose, based on the initial dose or the most recent ACT value, to minimize the risk of hypotension or anaphylactoid reactions.

Fig. 1.

Surgical schematic of off-pump valvotomy. SVC, superior and inferior vena cava; IVC, inferior vena cava; MPA, main pulmonary artery; PV, pulmonary valve

The initial hybrid procedure combined pulmonary valvuloplasty and balloon dilatation. A purse-string suture was placed in the right ventricular outflow tract (RVOT), approximately 1.5–2.0 cm from the pulmonary trunk. Under continuous transesophageal echocardiography guidance, a intravenous catheter punctured the RVOT wall through the center of the purse string and then perforated the atretic pulmonary valve (PV). A guide wire was advanced through the needle, followed by a sheath that was passed over the guide wire into the RVOT. The balloon diameter was selected to match the PV annulus, and sequential balloon dilatations of the PV were performed until transesophageal echocardiography confirmed satisfactory pulmonary pulsatile blood flow.

Before completion, we generally examined the change of SatO₂ to decide whether to perform modified Blalock Taussig shunt (MBTS) or ligate the PDA. If SatO₂ increased to more than 85%, PDA was then temporarily blocked and SatO₂ was monitored. If SatO₂ showed no significant changes, PDA could be ligated at the same time. If SatO₂ was significantly reduced to less than 75%, PDA was kept open. If SatO₂ was further lower than 70%, MBTS was added and PDA was ligated. Patients received different procedures were showed in Fig. 2.

Fig. 2.

Flow chart of outcome for 109 patients following surgery. PA/IVS, pulmonary atresia with intact ventricular septum; MBTS, modified Blalock Taussig shunt; PDA, Patent ductus arteriosus

Collection of clinical data

Demographic data of the three groups of children were collected, including gender, month age, weight, body surface area. Preoperationally, all patients went through detailed examination of echocardiography, enhanced CT, and MRI when necessary. The hypoplasia of RV and tricuspid regurgitation was graded as mild, moderate and severe by echocardiography under the criteria described before [10]. Z value of TV was calculated by the diameter of TV under echocardiography and body surface area, using the normogram of the Congenital Heart Surgeon Study Group Protocol. In early post-operational period, patients were closely monitored, including vital signs, cardiac functions, hemodynamic changes, internal environmental hemostasis, and adverse events. The operation time, post-operational SatO₂%, mechanical ventilation time. hemodynamics drug maintainence, ICU stay length, in-hospital stay length and intraoperative RBC transfusion were collected. Early mortality was defined as death occurring before hospital discharge. Early reintervention was defined as the need for PDA ligation or MBTS procedure before hospital discharge, or the requirement for other surgical or catheter-based interventions.

Near-infrared spectroscopy

All patients received a standardized anesthetic protocol administered by the attending anesthesiologist. Intraoperative multi-site near-infrared spectroscopy (NIRS) monitoring was employed to assess regional oxygen saturation (rSO₂) in the brain, kidney, and intestine for comprehensive evaluation of organ perfusion. The sensors were positioned as follows: cerebral oxygen saturation sensors were placed on the forehead; renal oxygen saturation sensors were placed on the right paravertebral region (corresponding to the T10–L2 vertebral levels); and intestinal oxygen saturation sensors were placed on the anterior midline abdomen (between the umbilicus and xiphoid process). To ensure procedural safety, predefined alarm thresholds were established: during inflow occlusion, a decrease in cerebral rSO₂ of more than 10% from baseline triggered a warning; the acceptable intraoperative cerebral oxygen saturation was defined as an absolute value greater than 50%.

Neurodevelopmental testing

Neurodevelopmental outcomes were assessed by an experienced pediatrician using the Bayley Scales of Infant and Toddler Development III (Bayley III). This standardized test, designed for children aged 1 to 42 months, is widely recognized for its good inter-rater reliability and serves as an established measure for evaluating cognitive, language, and motor development [11]. The scores are standardized to a mean of 100 and a standard deviation of 15. Developmental delay was defined as a composite score below 85, representing more than one standard deviation from the normative mean.

Follow-up

After discharge, all patients underwent follow-up every 2 months in the first year, and once a year after. The median follow-up period was 80.6 months (range 3-116 months). We excluded patients who were lost of follow-up. 3 patients (8.8%) in the on-pump group and 3 (5.4%) in the off-pump group were lost to follow-up, while no losses occurred in the hybrid group. Echocardiography and enhanced CT were conducted each time to assess the growth of RV, TV size and ejection fraction. Pulmonary stenosis and pulmonary regurgitation were graded as mild, mild to moderate, moderate, moderate to severe and severe. Study endpoints included mortality prior to subsequent surgery and the achievement of either biventricular or 1.5 ventricular circulation among survivors. Patients were classified as having achieved biventricular circulation without further intervention if they exhibited no clinically significant cyanosis, along with echocardiographic evidence of a tripartite right ventricle, a RVOT pressure gradient ≤ 40 mmHg, no or trivial PDA flow, and minimal shunt across the patent foramen ovale. Significant associated issues—such as RVOT obstruction from subvalvular muscular hypertrophy, pulmonary valve annular hypoplasia, PDA shunt, or tricuspid regurgitation—were addressed before final classification as biventricular circulation. Patients with persistent cyanosis despite adequate forward flow and ongoing right ventricular hypoplasia between 36 and 60 months of age underwent 1.5 ventricular circulation via a bidirectional cavopulmonary shunt. Patients who did not meet either endpoint were categorized as “wait”.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 9.5(GraphPad Software, La Jolla CA, USA) and IBM SPSS Statistics version 22.0 (IBM-SPSS Inc., Armonk, NY, USA). Continuous variables with a normal distribution are reported as mean ± standard deviation and were compared using one-way analysis of variance (ANOVA), followed by Tukey’s test for post-hoc pairwise comparisons. Continuous variables under skewed distribution were illustrated by median with range. Categorical variables are expressed as frequency (percentage). Group differences in categorical data were assessed with the chi-square test or Fisher’s exact test. Time-to-event outcomes were analyzed using the Kaplan–Meier method, and differences between groups were compared with the log-rank test. A P-value < 0.05 was considered statistically significant.

Results

Baseline characteristics

Among them, 34 patients received pulmonary valvotomy under CPB, among these patients there were 11 cases underwent PDA ligation and 2 cases underwent MBTS + PDA ligation simultaneously, as termed on-pump group, 56 patients received pulmonary valvotomy without CPB as off-pump group, among these patients there were 20 cases underwent PDA ligation and 5 cases underwent MBTS + PDA ligation simultaneously. And 14 patients received hybrid approach as hybrid group, among these patients there were 12 cases underwent PDA ligation and 2 case underwent MBTS + PDA ligation simultaneously. No difference was observed among the groups in gender, age, weight, RV hypoplasia, tricuspid Z score, LVEF and preoperational SatO2 (Table 1).

Table 1.

Baseline characteristics of the three groups

	On-pump n = 34	Off-pump n = 56	Hybrid n = 14	P-valve
Gender (male/female)	20/14	34/22	8/6	0.964
Neonates (age≤28days)	22 (64.7)	36 (64.3)	7 (50.0)	0.583
Age (days)	28.2 ± 19.9	31.1 ± 22.7	30.5 ± 19.3	0.821
Weight (kg)	3.6 ± 0.8	3.8 ± 1.2	3.9 ± 0.5	0.554
RV hypoplasia				0.236
Mild	20 (58.8)	30 (53.6)	11 (78.6)
Moderate	14 (41.2)	26 (46.4)	3 (21.4)
Tricuspid Z score	-1.55 ± 1.63	-1.58 ± 1.60	-1.65 ± 1.27	0.980
LVEF%	58.2 ± 6.5	58.3 ± 7.5	59.6 ± 6.3	0.802
Preoperational SatO2%	71.2 ± 4.6	69.4 ± 5.4	71.8 ± 6.2	0.130

Data presented as mean ± SD and n (%)

P < 0.05 was considered statistically significant

RV Right ventricle, LVEF Left ventricular ejection fraction

Hospital outcomes

As shown in Table 2, the overall acute postoperative mortality was 5.5% (6/109). In the on-pump group, 2 (5.9%) patients died, 1 from severe hypoxemia and 1 from severe right ventricular arrhythmia and heart failure. The off-pump group had 3 (5.4%) deaths, 2 of which resulted from hypoxemia. 1 patient in this group underwent reoperation with MBTS and PDA ligation but died from multiorgan dysfunction 34 h later. In the hybrid group, 1 (7.1%) patient died from severe hypoxemia and respiratory failure. No statistically significant difference in mortality was observed among the three groups (P = 1.000)In the on-pump group, early complications included hypoxemia in 2 patients, pulmonary infection in 1, and renal dysfunction in 1. 1 patient in this group required reoperation due to severe hypoxemia caused by an obstructed artificial tunnel and subsequently underwent an MBTS procedure, yielding an early reoperation rate of 2.9%. The off-pump group experienced hypoxemia in 2 cases, pulmonary infection in 2, arrhythmia in 1, and wound infection in 1. Among these, 1 patient required MBTS for hypoxemia and another underwent PDA ligation for pulmonary congestion, resulting in an early reoperation rate of 3.6%. In the hybrid group, early complications comprised hypoxemia in 3 patients, pulmonary infection in 1, and renal dysfunction in 1. 2 of these patients required MBTS for hypoxemia, leading to an early reoperation rate of 14.3%. No statistically significant difference in reintervention was observed among the three groups (P = 0.264).

Table 2.

Early outcome of the three groups

	On-pump n = 34	Off-pump n = 56	Hybrid n = 14	P-value
Early mortality	2 (5.9)	3 (5.4)	1 (7.1)	1.000
Early reintervention	1 (2.9)	2 (3.6)	2 (14.3)	0.264
Postoperational SatO2%	82.2 ± 8.7	80.6 ± 10.5	81.2 ± 10.7	0.762
Operation time (min)	107.6 ± 22.7	52.3 ± 16.8	60.4 ± 10.3	< 0.001
Mechanical ventilation time (h)	92.5 ± 23.6	57.3 ± 25.4	62.0 ± 28.9	< 0.001
Hemodynamics drug maintainence (h)	218.2 ± 32.5	122.5 ± 24.8	131.4 ± 25.8	< 0.001
ICU stay length (days)	7.6 ± 2.6	4.5 ± 2.3	5.5 ± 2.2	< 0.001
In-hospital stay length (days)	14.3 ± 5.7	8.5 ± 3.2	9.1 ± 3.8	< 0.001

Data presented as mean ± SD and n (%)

P < 0.05 was considered statistically significant

ICU Intensive care unit

All groups had significant increase in SaO₂ after operation and no difference was found in SaO₂ level among groups (P = 0.345). The mean SaO₂ increased from preoperational 71.2 ± 4.6% to 80.6 ± 10.5% (P < 0.001) in on-pump group. Likewise, the primary off-pump operation promoted SaO₂ from 69.4 ± 5.4% before operation to 82.2 ± 8.7% (P < 0.001), and the mean SaO₂ increased from preoperational 71.8 ± 6.2% to 81.2 ± 10.7% (P = 0.009) in hybrid group. Although no difference was detected among the three groups in mortality and reintervention, the off-pump group and hybrid group presented with less operation time, less tracheal incubation time, less use of vasoactive drugs, less ICU stay length and in-hospital length (P < 0.001).

To further investigate the impact of transient blood flow occlusion during off-pump surgery, we paid special attention to oxygen perfusion of distal organs. During off-pump valvulotomy, the mean durations of the lowest rSO₂ in the brain, kidney, and intestine were 1.6 ± 0.3 min, 1.8 ± 0.5 min, and 1.9 ± 0.2 min, respectively. No significant differences were observed in the above indicators between pre-clamp and post-clamp 10 min or among the three groups, indicating that temporary blockage of the blood flow in off-pump procedure induced no significant more ischemic damage (Fig. 3).

Fig. 3.

Regional oxygen saturation (rSO₂) in the cerebral (A), intestinal (B), and renal regions (C). On-pump group (green), Off-pump group (blue) and Hybrid group (orange)

In the on-pump group, 14 patients (41.2%) required intraoperative RBC transfusion, with a mean volume of 1.3 ± 0.6 units. The off-pump group had 24 patients (42.9%) receiving transfusion, at a mean of 1.2 ± 0.6 units. Neither the transfusion rate (P = 0.876) nor the volume (P = 0.623) differed significantly between these two groups. Within the hybrid group, which was excluded from formal statistical comparisons due to its small sample size, only one patient (6.7%) received an intraoperative RBC transfusion of 2 units.

Neurodevelopmental outcomes

Neurodevelopmental assessments based on the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) were completed for 26 patients at a mean age of 27.3 ± 5.1 months (11 in the on-pump group and 15 in the off-pump group). The hybrid group was excluded from this analysis due to an insufficient number of assessed patients (n = 1). The mean composite scores for the on-pump and off-pump groups were 94.4 ± 7.2 versus 92.5 ± 7.4 (P = 0.516) for cognition, 91.3 ± 10.5 versus 90.8 ± 6.5 (P = 0.882) for language, and 95.4 ± 3.3 versus 93.2 ± 4.0 (P = 0.150) for motor. No statistically significant differences were observed between the two groups in any domain. All mean scores fell within the normal range (≥ 85), though slightly below the normative mean of 100. In the ex vivo circulation group, 1 case of cognitive delay and 2 cases of language delay were identified, compared with 2 cases of cognitive delay and 2 cases of language delay in the off-pump group.

Mid-term outcomes

Over a median follow-up of 80.6 months, 1 death occurred, and a total of 3 patients in the on-pump group and 3 in the off-pump group were lost to follow-up. 1 patient in the off-pump group died from severe arrhythmia after terminal surgery. The mid-term outcomes for the three groups are presented in Table 3. The vast majority of the patients in three groups had significant growth of RV, whereas 11 patients remained to be moderate hypoplasia (4 patients in on-pump group, 6 patients in off-pump group and 1 patient in hybrid group). Moreover, in comparison to TV size before operation, three groups showed less negative tricuspid Z value at last follow-up, indicating satisfactory effect could be achieve by both techniques to pursue biventricular circulation (Fig. 4). No difference was detected in aspects of RV (P = 0.060) and TV status (P = 0.988) among groups.

Table 3.

Mid-term outcome of the three groups

	On-pump n = 30	Off-pump n = 51	Hybrid n = 13	P-value.
RV hypoplasia				0.060
Mild	22 (73.3)	42 (82.4)	8 (61.5)
Moderate	4 (13.3)	6 (11.8)	1 (7.7)
Tricuspid Z score	-0.67 ± 1.44	-0.64 ± 1.63	-0.59 ± 1.38	0.988
Pulmonary stenosis				< 0.001
Mild	6 (20.0)	12 (23.5)	9 (69.2)
Moderate	1 (3.3)	2 (3.9)	3 (23.1)
Pulmonary regurgitation				0.978
None to mild/mild	25 (83.3)	40 (78.4)	11 (84.6)
Mild to moderate	4 (13.3)	9 (17.6)	2 (15.4)
Moderate	1 (3.3)	2 (3.9)	0 (0.0)
LVEF%	60.3 ± 5.4	61.5 ± 4.6	62.7 ± 5.8	0.326

Data presented as mean ± SD and n (%)

P < 0.05 was considered statistically significant

RV Right ventricle, LVEF Left ventricular ejection fraction

Fig. 4.

Preoperative and follow-up tricuspid valve Z score of On-pump (A), Off-pump (B) and Hybrid (C) group

Moreover, pulmonary valve status was meticulously assessed by echocardiography at each follow-up visit. Only 1 patient in the on-pump group was diagnosed with moderate pulmonary stenosis, while all others exhibited less than moderate stenosis. In the off-pump group, 12 patients had mild stenosis and 2 had moderate stenosis. The hybrid group, however, demonstrated a higher proportion and greater severity of pulmonary valve stenosis, with 9 patients exhibiting mild stenosis, 3 moderate stenosis, indicating a significantly higher incidence of pulmonary stenosis (P < 0.001). The Kaplan-Meier estimates demonstrate that the hybrid group exhibited significantly poorer freedom from moderate stenosis (hybrid versus on-pump group, P = 0.035; hybrid versus off-pump. P = 0.017), with an 8-year freedom from stenosis of only 74.0% versus 90.2% in the other two groups (Fig. 5). Most patients were evaluated as having acceptable pulmonary regurgitation, no worse than mild. Nevertheless, 1 patient in the on-pump group and 2 in the off-pump group presented with moderate pulmonary regurgitation not currently requiring reintervention. No statistically significant difference was found among the groups in pulmonary regurgitation (P = 0.978).

Fig. 5.

Freedom from moderate-to-severe pulmonary stenosis. On-pump group (red), Off-pump group (green) and Hybrid group (blue). PS, pulmonary stenosis

All patients had satisfactory cardiac function that was graded into NYHA I and II grade at last follow-up. There was also no difference in left ventricular EF% among groups (P = 0.326). The end-point outcomes of all patients in three groups at last follow-up were illustrated in Fig. 2. In on-pump group patients, 18 patients needed no operation again. Biventricular circulation was achieved in 7 patients and 1.5 ventricular correction was done in 2 patients, all of whom underwent successful second-stage surgery and performed well. 3 patients were still in waitlist for further correction. In off-pump group, 28 patients were free from reoperation. 12 patients underwent successful biventricular correction and 4 patients had 1.5 ventricular correction. 6 patients awaited further observation to decided their surgical fates. In hybrid group, 3 patients were free from reoperation. 7 patients underwent successful biventricular correction and 3 patients had 1.5 ventricular correction.

Discussion

The natural history of PA/IVS is extremely poor. Without early intervention, the mortality is as high as 50% in the first 2 weeks after birth, which even reaches to 85% in 6 months. Therefore, despite of its low incidence, PA/IVS has always been a research hotspot of new techniques in field of complex congenital heart disease. Characterized by diverse degree of underdeveloped RV and TV, a large repertoire of surgical strategies is adopted to cover the heterogeneity of PA/IVS and to improve operative results. It is now widely acknowledged that a subgroup of patients with mild to moderate RV hypoplasia could be potential candidates for biventricular circulation by first-stage RV decompression and RV recruitment in neonatal age [12, 13]. However, the optimal way for RV decompression is still controversial and varies with different centers. In the present study, we intended to legitimate the off-pump pulmonary valvotomy as a primary operation for selected PA/IVS patients. By comparing the early to mid-term outcomes of the off-pump and on-pump procedure, our results supported the off-pump pulmonary valvotomy as a safe and accessible technique that avoided CPB damage in neonates.

There are multiple forms of operation for RV decompression. Our study compared the surgical outcomes and prognosis of procedures performed with and without CPB, as well as hybrid techniques, while other techniques were excluded from analysis for various reasons. Among them, PV transannular patching is the most radical procedure for relieving the obstructed RV and providing the best forward flow. However, it is also reported to be associated with higher possibility of ventricular damage and pulmonary regurgitation [14–16]. For that reason, we have largely rendered this technique in neonates with mild to moderate RV hypoplasia in recent decades. Another promising alternative is minimal invasive transcatheter valvotomy and balloon dilation, which also avoids CPB. Using a laser guidewire or RF wire to perforate the membranous atretic PV and balloon valvuloplasty to relieve the obstructed site, this technique had excellent effect reported by multiple large centers [7, 10, 17]. Nonetheless, in centers with less experience, this closed procedure seemed to have higher rates of early hypoxemia and re-stenosis that necessitate reoperation, and other catheter procedure-related complications as well [18]. Thus, the application of transcatheter valvotomy in under-developed regions is not only restrained by the considerable expense to equip a catheterization laboratory, but is also greatly limited by its demanding technical skills. The hybrid approach, combining surgical access with transcatheter valvotomy, is welcomed by some surgeons because of its shorter procedural distance, avoidance of X-ray radiation, and high success rate. Joshi described a technique involving pulmonary valvotomy combined with left subclavian artery-to-main pulmonary artery shunt via a left thoracotomy without the use of extracorporeal circulation [19]. This approach allows for both pulmonary valvotomy and a systemic-to-pulmonary shunt to be performed through a left thoracotomy, thereby avoiding cardiopulmonary bypass. The operative time for this procedure can be controlled within 2–3 h, and postoperative adhesions are minimal, significantly reducing the risk and difficulty of reoperation [20]. Sano et al. further noted that a staged treatment strategy based on this method could ultimately achieve biventricular repair in most pediatric patients [21]. When managing patients with right ventricle-to-coronary artery fistulas, they also observed that approximately half of the pediatric patients experienced spontaneous resolution of the coronary artery fistulas through the gradual reduction of right ventricular pressure, thereby avoiding coronary steal and myocardial ischemia caused by sudden decompression [12]. This concept offers a new therapeutic perspective for subgroups traditionally considered high-risk. In contrast, the median sternotomy approach employed in the present study differs significantly in terms of incision choice and surgical strategy. Although we also avoided cardiopulmonary bypass through off-pump surgery, the median sternotomy itself remains a significant factor influencing postoperative bleeding and transfusion requirements. Furthermore, due to the involvement of sternotomy and pericardial opening, postoperative adhesions resulting from a median sternotomy are theoretically more pronounced than those from a left thoracotomy, posing a potential challenge for any subsequent reoperations that may be required.

Studies that focused on the on-pump pulmonary valvotomy have become gradually fewer in recent years, but is still widely adopted by most pediatric hospitals in China and many other developing countries. The procedures could be performed smoothly under the aid of CPB, where the hemodynamics are stably maintained. Given these premises, the on-pump valvotomy is easily accepted by most cardiac surgeons for its safety and operative simplicity. However, the largest problem related to this technique is the intolerance of CPB in neonates. Most of the PA/IVS patients are already in extremely poor conditions before surgery, plus their low weight and immature myocardium, heart surgery under CPB therefore carries inherent risks. Several studies have demonstrated that CPB duration and its related complications were independent risk factors that affected mortality of congenital heart surgery in neonates [22]. In our study, no significant difference in early mortality was observed among the three groups, which may be attributed to the relatively milder condition of our study population and our advanced surgical and perioperative care techniques. Moreover, there also existed a paradox that once MBTS is concomitantly performed, the possibility of post-operational hemorrhage(primarily mediastinal and thoracic)greatly increases due to CPB-induced coagulation dysfunction, whereas the use of hemostatic drugs increases the risks of conduits obstruction. 1 patient with MBTS in our on-pump group developed obstructed conduit because of this reason, who unfortunately had to receive another emergent operation to solve this problem.

On the contrary, the off-pump pulmonary valvotomy avoids CPB risks while having the same surgical effects. Only one mortality was observed in our study. This technique demonstrated significant advantages over CPB in terms of endotracheal intubation duration, vasoactive drug usage, ICU length of stay, and total hospital stay, while showing comparable results to the hybrid technique. Especially, it significantly reduced the operation time to nearly 50% to that of the on-pump group. Furthermore, this technique imposes lower demands on operational skills and equipment requirements compared to transcatheter interventions, rendering it more clinically applicable. Regarding long-term outcomes, the off-pump approach demonstrated satisfactory freedom from restenosis in our institutional experience. However, given the small sample size of the Hybrid group and the uneven temporal distribution of cases, these findings should be interpreted as observed trends within our institutional experience rather than as definitive evidence of superiority of one technique over another. To be noted, however, it indeed required quick completion of pulmonary valvotomy procedure soon after the blood flow was blocked, aiming to minimize the damage of ischemia and reperfusion. To guarantee this, the surgical equipment for valvotomy was prepared directly aside the surgeon, and we placed suture traction at both side of the pulmonary incision for better vision. The procedure was generally finished in 2 min, and we suggested to strictly limit the blood blockage duration in 5 min to avoid neurologic damage. To verify its safety, we closely monitored perioperative oxygen perfusion, showing no difference with those of the on-pump group. Postoperative neurodevelopmental outcomes were also assessed. Despite the limited sample size, our data indicated that average scores for cognitive, language, and motor development in the pediatric patients all fell within the normal range. Furthermore, no significant differences were observed in any neurodevelopmental outcomes between the on-pump and off-pump groups. These findings provide preliminary but important evidence supporting the neurological safety of off-pump cardiac surgery.

A systemic-to-pulmonary artery shunt remains necessary when right heart blood flow is inadequate. The initial choice between constructing a MBTS and preserving the PDA presents a clinical dilemma. In the on-pump group, the PDA was left in situ for 8 patients. 5 of these PDAs closed spontaneously during follow-up, and maintained acceptable oxygen saturation. Within the off-pump group, the PDA was retained in 11 patients, 8 of whom experienced spontaneous closure. 7 of these 8 patients maintained satisfactory oxygen saturation, while one required MBTS due to declining saturation. In the Hybrid group, the PDA was left in 5 patients, with spontaneous closure occurring in three. 2 of these 3 patients maintained acceptable oxygen saturation, and the third received MBTS. While Li advocate for MBTS in all patients younger than one month [9], we contend that preserving the PDA offers specific advantages, despite a potentially higher reintervention rate. This approach can avoid an early and possibly unnecessary MBTS in some infants and eliminates the need to take down the MBTS during the definitive corrective operation.

In our study, all three techniques—on-pump, off-pump, and hybrid—were applied to the same underlying patient population: neonates with pulmonary atresia with intact ventricular septum who were candidates for biventricular repair. Core inclusion criteria included membranous pulmonary atresia, mild to moderate right ventricular hypoplasia, tricuspid valve Z-value greater than − 4, and absence of right ventricle-dependent coronary circulation. Patients with TV Z-value less than − 4 or evidence of RVDCC were excluded as they were not considered suitable for primary biventricular repair. Within this homogeneous population, technique selection was primarily driven by era effect and evolving institutional experience: in the early era, on-pump surgery served as the standard approach while the hybrid technique was introduced and applied to a broader range of patients as an emerging strategy; as experience with off-pump inflow occlusion matured, it became the preferred approach for patients with favorable anatomy, while hybrid technique was used more selectively as its limitations became better understood. Additionally, patient hemodynamic stability and individual surgeon preference also collectively influenced technique selection.

In summary, for PA/IVS patients with mild to moderate RV hypoplasia and without RVDCC, the off-pump pulmonary valvotomy has distinctive advantages over the on-pump technique. Although transcatheter and hybrid approach emerged as new promising methods in recent decades, the off-pump pulmonary valvotomy has never lost its competitive because it is more accessible, technique-friendly and acceptable. Given that the long-distance transference of PA/IVS neonates to large institutions could be highly risky, it is essential for less experienced centers and underdeveloped remote regions to consider off-pump pulmonary valvotomy as an alternative.

Limitations

This study has several main limitations. First, the neurodevelopmental assessment was limited by a small sample size of only 27 cases, and the Hybrid group was excluded from this comparison due to insufficient numbers, which may introduce selection bias. The average assessment age of 27.3 months is too early to detect subtle cognitive impairments that often manifest during school age. Without preoperative baseline assessments, it is also difficult to determine whether neurodevelopmental outcomes are attributable to the surgery or the underlying congenital disease. Furthermore, potential confounders such as socioeconomic status and early rehabilitation interventions were not adjusted for in the analysis. Second, the study’s statistical power is limited by its relatively small sample size, increasing the risk of Type II error where genuine differences, particularly for low-incidence outcomes like mortality and reintervention, may remain undetected. Consequently, the lack of a statistically significant difference in mortality among the three groups should be interpreted cautiously and does not demonstrate equivalent therapeutic efficacy. Third, Confounding bias may also arise from era effects and learning curves across the 15-year study period from 2010 to 2025, as perioperative management evolved substantially. Hybrid procedures were predominantly performed in the earlier phase (2010–2015), while on-pump and off-pump cases were more evenly distributed over time. The higher restenosis rate observed in the Hybrid group may thus reflect the initial learning curve and less mature management protocols rather than an inherent limitation of the technique itself. With only three restenosis events among 15 Hybrid cases, subgroup analysis or multivariable adjustment was not feasible. In summary, while this study offers preliminary evidence for the early safety of off-pump surgery, these limitations necessitate a cautious interpretation of the findings. Larger prospective studies with longer follow-up are required for further validation.

Conclusion

Primary RV decompression for PA/IVS patients could be achieved by multiple techniques, yet the treatment effects vary greatly with experience and regions. In this study, we presented the off-pump pulmonary valvotomy as an effective, safe and technical-friendly operation, which could be accessible by more centers while providing comparable early to mid-term results to the on-pump and hybrid procedure.

Abbreviations

CPB

Cardiopulmonary bypass

ICU

Intensive care unit

MBTS

Modified Blalock Taussig shunt

PA/IVS

Pulmonary atresia with intact ventricular septum

PDA

Patent ductus arteriosus

ACT

Activated clotting time

NIRS

Near-infrared spectroscopy

rSO₂

regional oxygen saturation

RVOT

Right ventricular outflow tract

RVDCC

Right ventricular-dependent coronary circulation

RV

Right ventricle

TV

Tricuspid valve

Authors’ contributions

L.H., B.S., XJ.Z., G.Z., K.L., XY.Z. and J.Z. contributed to data curation and writing-original draft preparation. X.H. and Z.Z. conceived, instructed, reviewed, and revised the manuscript. All the authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (grant number 82172101), the Natural Foundation of Shanghai Science and Technology Committee (grant number 23Y11907000, 23ZR1440900), Shanghai Key Research Center Construction Project-Shanghai Research Center for Pediatric Cardiovascular Diseases (2023ZZ02024), National Clinical Key Specialty Construction Project (10000015Z155080000004).

Data availability

Data supporting the study conclusions can be obtained from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by IRB of Shanghai Children’s Medical Center Affiliated to Shanghai Jiao Tong University School of Medicine (SCMCIRB-K2022121-1).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.