Right axillary thoracotomy versus median sternotomy for repair of subarterial ventricular septal defect in adolescents and adults Authors: Jiachen Li, Yalun Qu, Zhongyi Han, Lizhi Lv, Yabing Duan, Guofeng Xing, Gang Li, Yuefeng Cao, Qiang Wang Published: 2026 ABSTRACT BackgroundSubarterial ventricular septal defect (sVSD) has a relatively high incidence in the Asian population. Surgical repair remains the preferred treatment. This research aims to report on the application of a modification of the surgical approach in the treatment of sVSD in adolescents and adults; (2)MethodsThis retrospective cohort study included 74 consecutive patients aged 14 years or older with sVSD who underwent surgical repair between January 2016 and December 2024.Patients were divided into Group A (right axillary thoracotomy) and Group B (median sternotomy). Perioperative data and long-term follow-up outcomes were analyzed and compared; (3) Results: Baseline characteristics were comparable between groups. The right axillary thoracotomy group had longer cardiopulmonary bypass and aortic cross-clamp times but significantly lower postoperative drainage volumes. No significant differences were observed in mechanical ventilation time, intensive care unit stay, hospital length of stay, or pre-discharge ventricular function. During long-term follow-up, no deaths, reinterventions, residual shunts, or complete atrioventricular block occurred in either group; (4) Conclusions: In patients aged 14 years or older with sVSD, right axillary thoracotomy achieves perioperative and long-term outcomes comparable to conventional sternotomy, while offering advantages in reduced surgical trauma. Introduction Subarterial ventricular septal defect (sVSD) has a relatively high incidence in the Asian population, accounting for about 30% of all kinds of ventricular septal defect (VSD) [1]. It is adjacent to the aortic valve and pulmonary valve, often combined with aortic valve prolapse and insufficiency, with a low spontaneous closure rate and high interventional risk [2]. Therefore, surgical operation has always been the first choice. With the development of right axillary thoracotomy technology over the past three decades, some scholars [3, 4] have attempted to repair sVSD through right axillary thoracotomy and achieved good results, but most of the patients were young children with small weight [5], and few studies have focused on older children or adult patients with larger body size. This study intended to conduct a cohort study on patients over 14 years old (including adults) with sVSD in our center, compare the clinical effects of right axillary thoracotomy with common median sternotomy, including the early and mid-term follow-up results. Materials and methods General information This retrospective cohort study included consecutive patients aged ≥ 14 years (including adults) with subarterial ventricular septal defect (sVSD) who underwent surgical repair at our center between January 2016 and December 2024. Inclusion criteria were: (1) diagnosis of sVSD confirmed by echocardiography, with or without concomitant atrial septal defect, and (2) presence of mild to moderate pulmonary hypertension. Exclusion criteria included: (1) concomitant intracardiac or extracardiac malformations requiring simultaneous surgical correction (excluding atrial septal defect), (2) moderate to severe pulmonary hypertension, (3) cardiac insufficiency, and (4) prior cardiac surgery. Patients were divided into two groups according to the surgical approach: Group A (right axillary thoracotomy) and Group B (median sternotomy). All subsequent analyses and comparisons were performed between Group A and Group B. Baseline demographic and clinical variables collected included age, sex, height, weight, ventricular septal defect diameter, left ventricular ejection fraction, and presence of preoperative aortic insufficiency. Intraoperative and postoperative variables included aortic cross-clamp time, cardiopulmonary bypass (CPB) duration, postoperative drainage volume, duration of mechanical ventilation, intensive care unit (ICU) stay, length of hospital stay (LOS), and residual shunt. Surgical technique Right axillary thoracotomy approach In Group A, the patient positioning and incision planning are illustrated in Fig. 1, with a positional roll placed beneath the left axilla to elevate the operative field. The patient was stabilized using anterior and posterior supports, and the right upper limb was abducted above the head and secured to a head frame. A curvilinear skin incision (7–8 cm) was made between the intersection of the posterior axillary line with the third intercostal space and the intersection of the midaxillary line with the fifth intercostal space. The muscle layers were dissected along the direction of muscle fibers to avoid transection, and subcutaneous dissection was extended along the inferior border of the pectoralis major muscle to the fourth intercostal space. The thoracic cavity was entered through the fourth intercostal space along the superior border of the fifth rib. The right lung was gently retracted posteriorly using moist gauze to expose the pericardium and the right phrenic nerve. The pericardium was incised longitudinally approximately 2 cm anterior to the phrenic nerve, extending from the pericardial reflection of the ascending aorta to that of the inferior vena cava. The operative field and pericardial exposure are shown in Fig. 2. Multiple traction sutures were placed along the right pericardial margin as well as near the pulmonary artery, ascending aorta, and superior vena cava to optimize exposure of the operative field. Cardiopulmonary bypass establishment Cardiopulmonary bypass (CPB) was established under direct vision via central cannulation of the ascending aorta, superior vena cava, and inferior vena cava. Central cannulation was adopted as the standard strategy in this study because it is reproducible and provides sufficient exposure in most adolescent and adult patients, whereas peripheral cannulation was not routinely used but may be considered selectively in cases with suboptimal exposure. In patients with a wider thoracic cavity, exposure of the ascending aorta can be technically challenging; therefore, gentle inferior (caudal) traction of the ascending aorta was applied using vascular forceps to lower its relative position and improve visualization of the cannulation site. A 90-degree angled venous cannula was preferentially used for the superior vena cava, while a 180-degree cannula was used for the inferior vena cava, with an insertion depth of approximately 4–5 cm depending on patient height. After aortic cross-clamping, cardioplegia was administered via the aortic root. Intracardiac repair via trans-tricuspid approach Following cardioplegic arrest, a right atriotomy was performed, and the subarterial ventricular septal defect (sVSD) was exposed through the tricuspid valve. In adolescent and adult patients, exposure through the pulmonary artery is often limited due to the relatively deeper position of the main pulmonary artery in the right axillary operative field; therefore, a trans-tricuspid approach was preferentially adopted in this cohort. A suction device was inserted into the defect and directed toward the caudal (foot) side, providing inferior traction of the interventricular septum and facilitating exposure of the superior margin of the defect at the level of the pulmonary valve annulus. The trans-tricuspid exposure of the defect and surgical orientation are illustrated in Fig. 3. Adequate exposure was achieved through leaflet retraction and traction techniques, and routine detachment of the tricuspid anterior leaflet was not required. The defect was repaired using a bovine pericardial patch of appropriate size, and continuous suturing was initiated from the 1 o’clock position from the surgeon’s perspective. At the juxta-arterial superior margin, sutures were placed with shallow, precise bites along the fibrous rim beneath the semilunar valves, with careful attention to avoid distortion of the aortic and pulmonary valve annuli, and reinforcement within the pulmonary sinus was not routinely required. Completion After completion of the repair, rewarming was initiated and the aortic cross-clamp was removed. The patient was weaned from CPB once satisfactory results were confirmed by intraoperative transesophageal echocardiography. Standard adult cardiac surgical instruments were used in all cases, and no specialized minimally invasive instruments were routinely required. Median sternotomy In Group B, the patient was placed in the supine position with the chest elevated. A longitudinal incision was made along the midline of the sternum, the sternum was sawed open, and the sternum was retracted with a retractor. The pericardium was incised and suspended to fully expose the aorta and superior and inferior vena cava. After heparinization, intubation was performed in the aorta, superior vena cava, and inferior vena cava respectively to establish CPB. After aortic cross-clamping, cardioplegic solution was perfused through the aortic root. The main pulmonary artery was incised, and the sVSD was explored through the pulmonary artery incision. A suction device was inserted into the ventricular septal defect and pulled to expose the ventricular septal defect. A bovine pericardium of the same size was cut and used to continuously repair the ventricular septal defect clockwise from the 1 o’clock position of the surgeon’s side, taking care to avoid damaging the valves. After the operation, rewarming was performed and the aorta was unclamped. Cardiopulmonary bypass was stopped and withdrawn after the transesophageal echocardiography result was satisfactory. Ethical statement This single-center, retrospective cohort study obtained approval from the Ethics Committee of Beijing Anzhen Hospital, Capital Medical University, on 6 August 2024 (protocol number 2025334#). The study data were collected and stored in accordance with the ethical guidelines outlined in the WMA Declaration. The establishment and ongoing management of the study database were reviewed and approved by the Ethics Committee of Beijing Anzhen Hospital. Evaluation standard There are some evaluation criteria that can help us understand the preoperative condition and postoperative recovery of patients: Early follow-up results mainly refer to death or reoperation within 30 days after surgery or before hospital discharge. Long-term follow-up results mainly referred to death or reoperation within one to ten years after surgery. Statistical analysis All data were statistically analyzed using SPSS Statistics 29.0. First, the K-S test was used to test the normality of the measurement data. Normally distributed data were described as mean and standard deviation (x ± s), and intergroup comparison was performed using the t-test; nonnormally distributed ones were described as median (interquartile range) [M (P25, P75)], and intergroup comparison was performed using the Mann-Whitney U test. Categorical variables were described as case number and percentage, and intergroup comparison was performed using the chi-square test. The intergroup comparison of follow-up results was represented by Kaplan-Meier curve. P ≤ 0.05 was considered statistically significant. Results Seventy-four patients were included and divided into two groups according to the surgical approach: Group A (n = 38) and Group B (n = 36). There were no significant differences between Group A and Group B in baseline characteristics, including age (20.63 ± 6.64 vs. 25.11 ± 7.48 years, p = 0.062), height (167.63 ± 7.63 vs. 163.44 ± 20.72 cm, p = 0.170), and weight (55.47 ± 8.13 vs. 56.78 ± 9.96 kg, p = 0.664). Similarly, no significant differences were observed in preoperative ventricular septal defect diameter (12.42 ± 5.70 vs. 10.89 ± 5.48 mm, p = 0.410), left ventricular ejection fraction (66.29 ± 5.34% vs. 67.00 ± 5.87%, p = 0.720), or the incidence of preoperative aortic insufficiency (8/38 vs. 2/36, p = 0.350). Detailed baseline characteristics are summarized in Table 1. Table 1Baseline demographic and clinical characteristics of the study populationVariablesGroup A (n=38)Group B (n=36)P valueAge (years)20.63 ± 6.6425.11 ± 7.480.062Male, n (%)8 (21.05%)16 (44.44%)0.170Height (cm)167.63 ± 6.73163.44 ± 20.720.195Weight (kg)55.47 ± 8.1356.78 ± 9.960.664VSD diameter (mm)12.42 ± 5.7010.89 ± 5.480.410LVEF (%)66.29 ± 5.3467.00 ± 5.870.720Preoperative AI, n (%)8 (21.05%)2 (5.56%)0.350Data are presented as mean ± standard deviation or number (percentage)VSD: ventricular septal defect; LVEF: left ventricular ejection fraction; AI: aortic insufficiency Table 1Baseline demographic and clinical characteristics of the study populationVariablesGroup A (n=38)Group B (n=36)P valueAge (years)20.63 ± 6.6425.11 ± 7.480.062Male, n (%)8 (21.05%)16 (44.44%)0.170Height (cm)167.63 ± 6.73163.44 ± 20.720.195Weight (kg)55.47 ± 8.1356.78 ± 9.960.664VSD diameter (mm)12.42 ± 5.7010.89 ± 5.480.410LVEF (%)66.29 ± 5.3467.00 ± 5.870.720Preoperative AI, n (%)8 (21.05%)2 (5.56%)0.350Data are presented as mean ± standard deviation or number (percentage)VSD: ventricular septal defect; LVEF: left ventricular ejection fraction; AI: aortic insufficiency Baseline demographic and clinical characteristics of the study population Data are presented as mean ± standard deviation or number (percentage) VSD: ventricular septal defect; LVEF: left ventricular ejection fraction; AI: aortic insufficiency Intraoperatively, Group A had significantly longer cardiopulmonary bypass (CPB) time (71.17 ± 12.84 vs. 51.56 ± 7.55 min, p < 0.001) and aortic cross-clamp time (41.94 ± 11.26 vs. 25.67 ± 5.00 min, p < 0.001) compared with Group B. Postoperatively, Group A demonstrated significantly lower drainage volumes within the first 24 h (360.00 vs. 620.00 mL, p = 0.004) and the second 24 h (140.00 vs. 200.00 mL, p = 0.003). No significant differences were observed between the two groups in mechanical ventilation time (5.11 ± 1.78 vs. 4.61 ± 2.31 h, p = 0.470), ICU stay (1.00 vs. 1.00 days, p = 0.258), length of hospital stay (10.00 vs. 10.00 days, p = 0.878), or pre-discharge left ventricular ejection fraction (65.59 ± 6.16% vs. 64.71 ± 3.12%, p = 0.614). Detailed perioperative and postoperative outcomes are presented in Table 2. Table 2Perioperative and postoperative outcomes of the two groupsVariablesGroup A (n=38)Group B (n=36)P valueCPB time (min)71.17 ± 12.8451.56 ± 7.55<0.001Surgical approachThrough tricuspid valveThrough pulmonary valve38 (100.00%)0 (0.00)0 (0.00)36 (100.00%)--Aortic cross-clamp time (min)41.94 ± 11.2625.67 ± 5.00<0.001Mechanical ventilation (h)5.11 ± 1.784.61 ± 2.310.470ICU stay (days)1.00 (1.00–1.00)1.00 (1.00–1.25)0.258Hospital stay (days)10.00 (8.00–15.00)10.00 (9.75–13.00)0.878Drainage 0–24 h (mL)360 (250–753)620 (535–725)0.004Drainage 24–48 h (mL)140 (63–170)200 (165–215)0.003Early death, n (%)0 (0.00)0 (0.00)-Pre-discharge reintervention, n (%)0 (0.00)0 (0.00)-Pre-discharge LVEF (%)65.59 ± 6.1664.71 ± 3.120.614Residual shunt, n (%)0 (0.00)0 (0.00)-Third-degree AV block, n (%)0 (0.00)0 (0.00)-Postoperative AI, n (%)0 (0.00)1 (2.78%)0.911Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage)CPB: cardiopulmonary bypass; LVEF: left ventricular ejection fraction; ICU: intensive care unit; AI: aortic insufficiency Table 2Perioperative and postoperative outcomes of the two groupsVariablesGroup A (n=38)Group B (n=36)P valueCPB time (min)71.17 ± 12.8451.56 ± 7.55<0.001Surgical approachThrough tricuspid valveThrough pulmonary valve38 (100.00%)0 (0.00)0 (0.00)36 (100.00%)--Aortic cross-clamp time (min)41.94 ± 11.2625.67 ± 5.00<0.001Mechanical ventilation (h)5.11 ± 1.784.61 ± 2.310.470ICU stay (days)1.00 (1.00–1.00)1.00 (1.00–1.25)0.258Hospital stay (days)10.00 (8.00–15.00)10.00 (9.75–13.00)0.878Drainage 0–24 h (mL)360 (250–753)620 (535–725)0.004Drainage 24–48 h (mL)140 (63–170)200 (165–215)0.003Early death, n (%)0 (0.00)0 (0.00)-Pre-discharge reintervention, n (%)0 (0.00)0 (0.00)-Pre-discharge LVEF (%)65.59 ± 6.1664.71 ± 3.120.614Residual shunt, n (%)0 (0.00)0 (0.00)-Third-degree AV block, n (%)0 (0.00)0 (0.00)-Postoperative AI, n (%)0 (0.00)1 (2.78%)0.911Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage)CPB: cardiopulmonary bypass; LVEF: left ventricular ejection fraction; ICU: intensive care unit; AI: aortic insufficiency Perioperative and postoperative outcomes of the two groups Surgical approach Through tricuspid valve Through pulmonary valve 38 (100.00%) 0 (0.00) 0 (0.00) 36 (100.00%) - - Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage) CPB: cardiopulmonary bypass; LVEF: left ventricular ejection fraction; ICU: intensive care unit; AI: aortic insufficiency During follow-up, no deaths, reinterventions, residual shunts, or third-degree atrioventricular block occurred in either group. Follow-up duration differed between the groups (76.00 vs. 93.50 months, p = 0.014). One patient in Group B with preoperative moderate aortic insufficiency had persistent mild-to-moderate aortic insufficiency during approximately nine years of follow-up. Kaplan–Meier analysis demonstrated no significant differences in overall survival, freedom from reintervention, residual shunt, or aortic regurgitation between the two groups (Fig. 4). Detailed follow-up outcomes are summarized in Table 3. Table 3Mid-term follow-up outcomes of the two groupsVariablesGroup A (n=38)Group B (n=36)P valueFollow-up time (months)76.0 (63.0, 90.0)93.5 (83.5, 100.5)0.014Follow-up rate (n, %)38 (100.00%)36 (100.00%)-LVEF (%)65.59±6.1664.71±3.120.614Death (n, %)0 (0.00)0 (0.00)-Reintervention (n, %)0 (0.00)0 (0.00)-Residual shunt (n, %)0 (0.00)0 (0.00)-III°atrioventricular block (n, %)0 (0.00)0 (0.00)-AI (n, %)0 (0.00)1 (2.78%)0.911TI (n, %)0 (0.00)0 (0.00)-Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage)LVEF: left ventricular ejection fraction; AV: atrioventricular “-” indicates not applicable Table 3Mid-term follow-up outcomes of the two groupsVariablesGroup A (n=38)Group B (n=36)P valueFollow-up time (months)76.0 (63.0, 90.0)93.5 (83.5, 100.5)0.014Follow-up rate (n, %)38 (100.00%)36 (100.00%)-LVEF (%)65.59±6.1664.71±3.120.614Death (n, %)0 (0.00)0 (0.00)-Reintervention (n, %)0 (0.00)0 (0.00)-Residual shunt (n, %)0 (0.00)0 (0.00)-III°atrioventricular block (n, %)0 (0.00)0 (0.00)-AI (n, %)0 (0.00)1 (2.78%)0.911TI (n, %)0 (0.00)0 (0.00)-Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage)LVEF: left ventricular ejection fraction; AV: atrioventricular “-” indicates not applicable Mid-term follow-up outcomes of the two groups Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage) LVEF: left ventricular ejection fraction; AV: atrioventricular “-” indicates not applicable Discussion In this study, we demonstrated that right axillary thoracotomy repair for subarterial ventricular septal defect in adolescents and adults achieved perioperative and mid-term outcomes comparable to median sternotomy. Although cardiopulmonary bypass and aortic cross-clamp times were longer in Group A, no significant differences were observed in major clinical outcomes, including mechanical ventilation time, ICU stay, hospital length of stay, and postoperative ventricular function. In addition, postoperative drainage was significantly reduced in Group A, suggesting a potential advantage of reduced surgical trauma. Baseline demographic and clinical characteristics were comparable between the two groups, indicating that the observed differences in outcomes were primarily related to the surgical approach rather than patient-related factors. Variables such as age, body size, and defect diameter, which may influence surgical exposure and procedural complexity, were well balanced between groups, supporting the validity of the comparative analysis. Surgical repair remains the preferred treatment for subarterial ventricular septal defect due to its anatomical proximity to the aortic and pulmonary valves and the associated risk of aortic valve prolapse and insufficiency [6–7]. With the development of minimally invasive techniques, various alternative approaches have been explored to avoid the disadvantages of median sternotomy. Previous studies have reported the use of right axillary approaches for sVSD repair, but these have largely been limited to infants and small children [4–8]. Concerns have been raised regarding inadequate exposure in larger patients, and some authors have suggested that minimally invasive approaches may not be suitable for adolescents and adults [9–11]. In contrast, our findings suggest that right axillary thoracotomy can be safely extended to older patients with larger body size when appropriate surgical strategies are applied. In adolescent and adult patients, surgical exposure differs significantly from that in children due to a wider thoracic cavity and a deeper anatomical position of the great vessels. In the right axillary approach, exposure through the pulmonary artery may be limited, particularly in patients with larger body size. In this context, the trans-tricuspid approach provides a more reliable route for defect visualization, especially for juxta-arterial lesions, and helps avoid injury to the pulmonary valve. This approach has been increasingly adopted in our center since 2013 [12] and has become a preferred strategy in selected patients. Similar applications of the trans-tricuspid approach have also been reported in pediatric populations with favorable outcomes [13–14]. Furthermore, as experienced surgeons have noted, closing subpulmonic VSDs is inherently challenging, and direct aortic cannulation can be particularly difficult in patients with a high body mass index (BMI) due to the depth of the operative field. In such cases, the right axillary thoracotomy requires meticulous exposure techniques, and our trans-tricuspid strategy helps mitigate some of these visualization difficulties. Although longer CPB and aortic cross-clamp times were observed in Group A, likely reflecting the increased technical complexity of exposure, these did not translate into worse clinical outcomes, which is consistent with previous reports on minimally invasive cardiac surgery. These findings support the feasibility of extending minimally invasive approaches to adolescent and adult populations. Conclusions Right axillary thoracotomy is a safe and effective alternative to median sternotomy in patients aged 14 years or older with subarterial ventricular septal defects, providing comparable perioperative and mid-term outcomes while reducing surgical trauma. Limitations This study has certain limitations. Firstly, the diameter of the VSD in the patients is about 10 mm. Limited experience is available for patients with small ventricular septal defects, particularly those < 5 mm. In fact, a larger diameter of sVSD is more conducive to surgical operation. We also have only collected data on patients with an average weight of less than 55 kg and a maximum weight of no more than 70 kg. For patients with larger weights, further investigation is required. Another limitation is that the follow-up time for right axillary thoracotomy is about two years shorter compared to median sternotomy, because the right axillary thoracotomy approach was adopted relatively late, and the number of such cases gradually increased only as the technique became more mature. The third one, the sample size is relatively small, which requires us to continue to pay attention to this disease in the future. Finally, as discussed, direct aortic cannulation can be technically demanding in patients with a high body mass index (BMI). The current study did not perform a subgroup analysis based on BMI. Future studies should further evaluate the impact of BMI on surgical exposure and optimize cannulation strategies for these specific patients. Fig. 1Patient positioning and incision planning for the right axillary thoracotomy approach. The patient is placed in the left lateral decubitus position with the right upper limb abducted above the head. Preoperative surface marking identifies the optimal intercostal space and incision location along the right axillary line. The planned incision allows direct access to the fourth intercostal space while minimizing muscle injury Patient positioning and incision planning for the right axillary thoracotomy approach. The patient is placed in the left lateral decubitus position with the right upper limb abducted above the head. Preoperative surface marking identifies the optimal intercostal space and incision location along the right axillary line. The planned incision allows direct access to the fourth intercostal space while minimizing muscle injury Fig. 2Surgical field exposure through right axillary thoracotomy Intraoperative view following right axillary thoracotomy. The pericardium is incised longitudinally anterior to the phrenic nerve and retracted with traction sutures to optimize exposure. The ascending aorta, superior vena cava, and inferior vena cava are visualized for central cannulation under direct vision Surgical field exposure through right axillary thoracotomy Intraoperative view following right axillary thoracotomy. The pericardium is incised longitudinally anterior to the phrenic nerve and retracted with traction sutures to optimize exposure. The ascending aorta, superior vena cava, and inferior vena cava are visualized for central cannulation under direct vision Fig. 3Trans-tricuspid exposure and repair of subarterial ventricular septal defect Schematic illustration of subarterial ventricular septal defect (sVSD) repair through a trans-tricuspid approach. The right atrium is opened, and the tricuspid valve leaflets are retracted to expose the defect. A suction device is inserted into the ventricular septal defect and directed toward the caudal side, allowing inferior traction of the septum and improved visualization of the superior margin adjacent to the pulmonary valve. The patch is sutured along the defect margin while carefully avoiding distortion of the semilunar valves Trans-tricuspid exposure and repair of subarterial ventricular septal defect Schematic illustration of subarterial ventricular septal defect (sVSD) repair through a trans-tricuspid approach. The right atrium is opened, and the tricuspid valve leaflets are retracted to expose the defect. A suction device is inserted into the ventricular septal defect and directed toward the caudal side, allowing inferior traction of the septum and improved visualization of the superior margin adjacent to the pulmonary valve. The patch is sutured along the defect margin while carefully avoiding distortion of the semilunar valves Fig. 4Kaplan–Meier analysis demonstrated no significant differences in overall survival, reintervention, residual shunt, or aortic regurgitation between the two groups Kaplan–Meier analysis demonstrated no significant differences in overall survival, reintervention, residual shunt, or aortic regurgitation between the two groups