Tricuspid valve detachment for ventricular septal defect closure in infants <5 kg: should we be hesitant?

Abstract

 

Open in new tabDownload slide

OBJECTIVES

We compared the clinical outcomes between tricuspid valve detachment (TVD) and non-TVD for ventricular septal defect (VSD) closure in infants <5 kg.

METHODS

From January 2004 to April 2020, 462 infants <5 kg with VSD without more complex intracardiac lesions and who had undergone VSD closure through the trans-atrial approach were enrolled. Propensity score-matching analysis was performed. Clinical outcomes were compared between the paired TVD group (group D) and paired non-TVD group (group N).

RESULTS

The median age and body weight at operation were 1.9 months [interquartile range(IQR), 1.4–2.5] and 4.2 kg (IQR, 3.7–4.6). The median follow-up duration was 83.4 months (IQR, 43.5–130.4). After matching, 44 pairs were extracted from each group. There were no significant differences in all-cause mortality (P = 0.176), reoperation (P = 0.172), postoperative morbidities, including residual VSD, aortic regurgitation, atrioventricular block and significant tricuspid regurgitation (TR) (P = 0.346) between group D and group N. However, group D showed significantly less TR progression during follow-up (P = 0.019).

CONCLUSIONS

In infants <5 kg, TVD can be a reasonable and valid option for successful VSD closure without morbidities, including TR progression if the indication exists.

INTRODUCTION

The most common form of congenital heart disease in childhood is ventricular septal defect (VSD), occurring in 50% of children with congenital heart disease and in 20% as an isolated lesion [1]. Over time, advancement in surgical and cardiopulmonary bypass techniques and devices has enabled a shift towards earlier repair of congenital heart defects. Currently, early closure of VSD is considered to reduce the time required for therapy to prevent or treat heart failure, maintain growth and minimize exposure to increased pulmonary pressure [2, 3].

In cases where it is difficult to expose the VSD margin, tricuspid regurgitation (TR), residual VSD and atrioventricular block (AVB) can occur due to improper suture or patch placement. Injury to the tricuspid valve (TV) structure may also occur while exerting traction to expose the VSD. Several studies have demonstrated that tricuspid valve detachment (TVD) is a safe technique to improve exposure for VSD closure without increasing the risk of TR, residual VSD and AVB [4–14].

As the timing of surgery is getting earlier, there are cases where it is difficult to find the VSD margin, even in patients with low body weight (LW). In these cases, most surgeons are concerned about damage to the relatively fragile and immature TV leaflet and subvalvar apparatus and, therefore, hesitate to perform TVD, due to complication such as TR progression. Furthermore, there have been only few studies on infants with LW who have undergone VSD closure, specifically with TVD; the impact of TVD on the clinical outcomes of VSD closure in patients with LW remains controversial.

We reviewed and compared the clinical outcomes between TVD and non-TVD (NTVD) for VSD closure in infants <5 kg.

MATERIALS AND METHODS

Patient selection

Since all available data were obtained from a retrospective review, the Institutional Review Board approved and waived the need for patient consent (approval no: H-2001-067-1094).

From January 2004 to April 2020, 576 patients <5 kg underwent VSD closure through trans-atrial approach in our institution. Of these patients, 114 patients who either underwent VSD closure through additional trans-pulmonary arterial approach or ventriculotomy (n = 40), had VSDs associated with a more complex intracardiac lesions that were corrected concomitantly (n = 71) or had previously undergone cardiac surgery (n = 3) were excluded. Thus, a total of 462 eligible patients were enrolled. Indications for operation included (i) failure to thrive/congestive heart failure and (ii) elevated pulmonary vascular resistance [2, 15].

Operative strategy and surgical technique

All procedures were performed under moderate hypothermic cardiopulmonary bypass. After aortic cross-clamp, the VSD was exposed through right atriotomy.

The indications of TVD were: (i) multiple TV chordal arrangement obscuring the margins of the defect, (ii) TV aneurysm that precluded easy access to the defect and (iii) high position of the defect with outlet extension requiring excessive traction on the TV leaflet for exposure [10]. With these indications, if the VSD was challenging to access, TVD was used at the surgeon's discretion for better exposure of the VSD margin.

The TV leaflet was detached with a parallel, circumferential incision in the annulus of TV, and the incision was made in either the annulus of the septal leaflet, the anterior leaflet, or in both. The incision was made ∼2 mm from the margin of the TV annulus to prevent damage to the aortic valve.

VSD closure was mainly performed using glutaraldehyde-fixed autologous pericardial patch using 6–0 or 7–0 polypropylene interrupted pledgetted suture technique. After VSD closure, the detached TV leaflet was repaired with 7–0 or 8–0 polypropylene continuous suture technique as appropriate. After VSD closure and TV repair, cold saline was directly injected into the right ventricle to evaluate for TR. Based on the saline injection test and postoperative trans-oesophageal echocardiography, if needed, additional TV repair including simple coaptation sutures for partial obliteration of the anteroseptal commissure or Kay-annuloplasty was performed to reduce the TR degree. The geometry and function of TV should be maintained and restored during the TV leaflet detachment and repair.

Echocardiographic evaluation

The perioperative and follow-up TR grade was evaluated by transthoracic echocardiography based on the American Society of Echocardiography criteria: 0, none to trivial; 1, mild; 2, mild to moderate; 3, moderate; 4, severe [16].

TR aggravation was defined as TR that was worse at discharge compared to preoperative TR, and TR progression was defined as that was worse at the last follow-up compared to TR at discharge. Significant TR was defined as moderate or more than moderate TR (TR grade ≥3).

Evaluation of clinical outcomes

In-hospital mortality was defined as death during the same hospitalization. Late mortality was defined as any death after discharge. The end-point was all-cause mortality after surgery, reoperation and significant TR. The last echocardiographic findings at or before the end-point were used for statistical analysis.

Regular (3–6-month intervals) echocardiographic and clinical follow-up was performed at the outpatient clinic. The median follow-up duration was 83.4 months [interquartile range (IQR), 43.3–130.5].

Statistical analysis

Statistical analyses were performed using IBM SPSS statistical software (version 25.0, IBM Inc., Armonk, NY, USA) and SAS (version 9.4, SAS Institute, Cary, NC, USA). All P-values were two-tailed and a P-value of <0.05 was considered statistically significant.

Continuous variables were expressed as median with IQR and categorical variables were presented as frequencies with percentages. Differences in baseline characteristics and clinical factors between the groups were evaluated using Student’s t-test or Wilcoxon rank sum test for continuous variables and chi-square test or fisher’s exact test for categorical variables. If there are >20% of cells with an expected value of <5 in a contingency table, Fisher’s exact test was used.

We performed a propensity score matching (PSM) between the TVD and NTVD groups to reduce the effect of selection bias. First, multiple logistic regression analysis was applied to get propensity score of the following covariates: age and body weight at operation, sex, gestational age <37 weeks, low birth weight (<2.5 kg), down syndrome, TR grade, concomitant TV repair and VSD type. The discrimination and calibration ability of the propensity score model was assessed by means of the C-statistic and the Hosmer–Lemeshow statistic (C-statistic 0.689 and Hosmer–Lemeshow goodness-of-fit test P-value 0.081); 1:1 matched samples adjusted for confounders were obtained from all patients using the nearest neighbour method. Standardized difference was used for comparing the covariates balance after matching and the standardized mean difference was within ±0.1 [17]. There was no significant difference between the groups for all variables (P-value <0.05). Comparisons between matched groups were performed with the McNemar and paired t-test for categorical and continuous variables, respectively. Survival rates related with clinical outcomes were estimated using the Kaplan–Meier method. Log-rank test was performed to compare the survival curves between the groups.

RESULTS

Patient demographic and perioperative data

Forty-six and 416 patients underwent VSD closure with TVD and VSD closure without TVD, respectively. Demographic and perioperative data are summarized in Tables 1 and 2. Overall, the median age and body weight at the time of operation were 1.9 months (IQR, 1.4–2.5) and 4.2 kg (IQR, 3.7–4.6). After PSM, 44 pairs were extracted from each group, paired TVD (group D) versus paired NTVD (group N). There were no significant differences in demographic and perioperative data, except for TV aneurysmal tissue between group D and group N (Tables 1 and 2). Group D had more TV aneurysmal tissue than group N (20 vs 7, P = 0.011). There were no significant differences in the median follow-up duration and echocardiographic follow-up duration between the 2 groups (Table 1).

Evaluation of clinical outcomes

In the overall cohort, all-cause mortalities were 9 (1.9%). In-hospital mortality occurred in 7 (1.5%) patients (TVD 3/46, 6.5% vs NTVD 4/416, 1.0%). The causes of in-hospital mortalities in the TVD group were septic shock (n = 1), respiratory arrest associated with tracheoesophageal fistula (n = 1) and ischaemic encephalopathy (n = 1), all of which occurred >3 months after surgery (postoperative 4.2, 11.8 and 19.8 months). The causes of in-hospital mortalities in the NTVD group were septic shock (n = 2) and cardiogenic shock (n = 2). Of these patients, postoperative venoarterial extracorporeal membrane oxygenation was required in 3 patients. Among 455 early survivors, late mortality occurred in 2 patients. There was 1 mortality in the TVD group at postoperative 86.8 months. The patient underwent heart transplantation for dilated cardiomyopathy and died due to acute rejection. The other late mortality in the NTVD group occurred 18.3 months after the operation, and involved a patient with acute myeloid leukaemia who developed, acute respiratory distress syndrome.

The overall survival rate at 1, 5 and 10 years were 98.7%, 98.2% and 97.7%, respectively. There was a significant difference in overall survival rate between the TVD group and the NTVD group (90.9% vs 98.5% at 10-year, P = 0.001). In the matched cohort, all-cause mortalities were 5 (5.7%). In-hospital mortality occurred in 3 (group D 3/44, 6.8% vs group N 0/44, 0.0%), and late mortality occurred in 2 of 85 early survivors. The Kaplan–Meier curve showed there was no significant difference in overall survival rate between group D and group N (P = 0.176) (Fig. 1).

Eight (1.7%) patients, from the NTVD group underwent reoperation. The causes of reoperation were residual VSD (n = 3), greater than grade 2 TR (n = 3), and left ventricular outflow tract obstruction (n = 2). The freedom from reoperation rate at 1, 5 and 10 years were 99.6%, 98.7% and 97.3%, respectively. There was no significant difference in the freedom from reoperation rate between the TVD group and the NTVD group (100.0% vs 97.1% at 10 years, P = 0.347). In the matched cohort, 2 patients required reoperation (group D 0/44, 0.0% vs group N 2/44, 4.5%). In the Kaplan–Meier curve, there was no significant difference in the freedom from reoperation rate between group D and group N (P = 0.172) (Fig. 1).

In the overall cohort, there were no significant differences in postoperative morbidities between the 2 groups (Table 3). There was residual VSD in 25 patients (TVD 0 versus NTVD 25). Of these patients, 3 patients underwent reoperation. Aortic regurgitation (AR) was newly developed in 8 patients (TVD 1 versus NTVD 7). In all these patients, the AR grade was mild without further progression. Three patients in the NTVD group required permanent pacemaker implantation due to irreversible complete AVB. In the matched cohort, there were no significant differences in morbidities such as residual VSD, AR, AVB and junctional ectopic tachycardia between group D and group N (Table 3).

Changes in TR

In the overall cohort, the NTVD group showed more TR aggravation than the TVD group at discharge (P = 0.030). However, in the matched cohort, there was no significant difference in TR aggravation between group D and group N (P = 0.763).

During the follow-up period, the freedom from TR progression rate at 1, 5 and 10 years were 99.2%, 97.0% and 87.4%, respectively. There was no significant difference in the freedom from TR progression rate between the TVD group and the NTVD group (100.0% vs 86.3% at 10 years, P = 0.380). However, in the matched cohort, the Kaplan–Meier curve showed the freedom from TR progression rate was higher in group D compared to group N (P = 0.019) (Fig. 2). Figure 3 shows the distribution of TR over time in the matched cohort.

There was no significant TR in the TVD group at the last follow-up (TVD 0/46, 0.0% vs NTVD 8/416, 1.9%). The freedom from significant TR rate at 1, 5 and 10 years were 99.3%, 98.7% and 96.8%, respectively. There was no significant difference in the freedom from significant TR rate between the TVD group and the NTVD group (100.0% vs 96.5% at 10 years, P = 0.375). In the matched cohort, significant TR occurred in 1 patient from group N. In the Kaplan–Meier curve, there was no significant difference in freedom from significant TR rate between group D and group N (P = 0.346) (Fig. 2).

DISCUSSION

Successful trans-atrial repair of VSD requires adequate exposure of the VSD margin to avoid residual VSD, AVB caused from injury of the conduction system and significant TR caused by TV leaflet distortion [5, 6, 9, 10]. However, multiple chordal attachments crossing over VSD, TV aneurysmal tissue or high position of VSD may present, obstacles to identification of the VSD margin. In these cases, it may be difficult to expose the VSD margin without excessive traction of the TV leaflets, which can cause major morbidities of VSD closure. As a result, the morbidities caused by inadequate VSD exposure can potentially lead to several complications. For example, residual VSD can cause infective endocarditis, cardiac and pulmonary volume overload and increased TR [14]. In addition, TR development from damage to TV can lead to atrial dilatation and arrhythmias and right heart failure [11].

Since TVD was described in 1962 by Hudspeth et al., several studies have demonstrated that TVD, or chordal detachment, is a safe surgical option and does not affect TV function [4–14, 18, 19]. Recently, Fraser et al. [13], using PSM analysis, demonstrated that TVD is a safe and effective technique that can be undertaken without compromising valve durability and imposing increased morbidity. Although TVD was mostly performed in the septal leaflet, Maile et al. [8] revealed that TVD in the anterior leaflet can be performed safely. In our study, when the exposure of VSD was difficult, we performed TVD in the septal leaflet or anterior leaflet as appropriate to prevent VSD closure-related morbidities.

Although current surgical outcomes of isolated VSD are excellent with low mortality and morbidity, young age and LW at the time of operation remain significant risk factors for poor patient outcomes [15, 20]. Nonetheless, Bang et al. revealed that TVD could be used safely for better exposure of the VSD without an increased risk of TR, even in infants younger than 3 months [12]. However, since there are patients born prematurely, the uncorrected age cannot reflect the actual state of growth. Concerning body weight, although the debate is ongoing, several studies have reported the clinical association between body weight at operation and VSD closure outcomes [2, 3, 9, 21]. Recently, Inohara et al. [3] reported that lower body weight (<4.5 kg) was significantly associated with increased risks of early mortality and major complications in infants who underwent VSD closure.

The impact of TVD on the clinical outcomes of VSD closure in infants with LW remains controversial. We considered that body weight at operation could be a more significant and influential factor and, thus, analysed the impact of TVD for VSD closure on clinical outcomes in patients <5 kg. We performed PSM analysis to also minimize bias from prematurity; there were no significant differences in patient characteristics and perioperative data except for TV aneurysmal tissue, between the matched groups. The significant difference in TV aneurysmal tissue was naturally affected by our indications of TVD. Based on the aforementioned indications of TVD, we usually performed TVD for VSD closure at the surgeon’s discretion [10].

The present study demonstrated 2 main findings. First, surgical outcomes of TVD for VSD closure in infants <5 kg were excellent without significant TR. Second, the rate of TR progression was significantly lower in group D compared to group N.

Considering the body weight of the study population, many surgeons might be hesitant to perform TVD due to increased risk of morbidity such as TV impairment, AVB, residual VSD and AR from the immature and weak valvular and subvalvular tissues.

However, regarding TV impairment, our study revealed that TVD for VSD closure in infants <5 kg did not affect TR progression. Rather, we found that there was significant TR progression in the NTVD group. Judging from the result, we can presume that there could be a distortion of TV leaflets or subvalvular apparatus due to problems related to incorrect surgical procedure from insufficient exposure of the VSD margin. This can cause leaflet traction, inappropriate suture placement and improper patch tailoring and placement. We already know that it is not easy to identify the VSD margin accurately in small infants, even though many patients in the NTVD group did not have corresponding TVD indications. Besides, an incorrect procedure during VSD closure can affect TV deformity such as TV movement restriction and TV growth limitation in the long term, and of course, TR progression as well. Meanwhile, there has been no tricuspid stenosis in our study during follow-up, which suggests that TV growth was not limited. Based on our results, adequate VSD exposure is important and TVD can be flexibly considered if the trans-atrial approach for VSD closure is not adequate to expose the defect in infants <5 kg.

Tatebe et al. [22] proposed that TVD should be avoided in patients with Down syndrome and small infants because of the risk of persistent TR and AVB. However, Bol-Raap et al. [7] reported contrary results, which revealed that TVD could be performed safely in patients with Down syndrome and NTVD had a higher rate of postoperative TR than TVD in small children with relatively LW under the 10th percentile. In 44 patients with Down syndrome in our study, TR progression was observed in 3 (6.8%) patients and significant TR occurred in 1 (2.3%) patient. Even in patients with Down syndrome in the TVD group, TR progression and significant TR did not occur.

Inadequate VSD margin exposure can cause AVB, residual VSD and AR. In terms of AVB, Siehr et al. [23] suggested that patients <4 kg and the presence of an inlet VSDs might be significant risk factors for AVB. Of course, there can be a significant difference in the distribution of the study population’s body weight between the studies. However, considering the range width of body weight in our study, our results showed excellent outcomes, irrespective of body weight and VSD type. AVB occurred in 3 (0.6%) patients in all study patients. Furthermore, there was no AVB in the TVD group. In addition, recently, Giordano et al. [14] suggested that TVD should be used whenever the exposure of VSD is difficult to avoid a residual VSD and AVB without TV impairment. With regard to residual VSD, Bol-Raap et al. reported that TVD resulted in comparable residual VSD as NTVD [7]. In our study, 25 patients had residual VSD only in the NTVD group, though there was no statistically significant difference. Concerning AR, to our knowledge, there are few studies comparing AR in TVD and NTVD after VSD closure. Inadequate exposure of the VSD margin can lead to aortic valve damage by inappropriate suture placement around the annulus or aortic valve, or struggling to exert traction on the TV leaflets. On the other hand, when performing TVD, incision too close to TV annulus can cause aortic valve injury. Therefore, care must be exercised when performing an incision on the TV leaflet. In our study, AR occurred in 8 (1.7%) patients without significant difference between the groups (TVD 1 versus NTVD 7). Our study suggests that proper TVD for VSD closure would not affect aortic valve injury in infants <5 kg.

Limitations

This study had several limitations. First, this study was a non-randomized, retrospective study at a single centre. Second, the operations were performed by several surgeons. Although the indication and operative technique of TVD were pursued in the same strategy, the use of TVD was determined at the surgeon’s discretion based on the intraoperative findings. Third, although we performed PSM analysis, the results might still be biased. Given the fact that patients who underwent VSD closure with TVD present with a more difficult surgical anatomy than those who underwent VSD closure without TVD, it can be assumed that the 2 groups cannot be completely identical. In addition, the relatively small sample size should be considered. Therefore, generalization of the results should be made with caution and will require long-term studies with large sample sizes.

CONCLUSION

TVD can be safely performed, even in infants <5 kg. With the appropriate indication, our results suggest that TVD may be a reasonable and valid option that has the advantage of providing better exposure to the VSD margin for successful VSD closure without causing morbidities, including TR progression. In addition, we believe that TVD might be considered more flexibly in infants with LW if VSD margin exposure is difficult to achieve.

Conflict of interest: none declared.

Author contributions

Jae Hong Lee: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Visualization; Writing—original draft. Sungkyu Cho: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Resources; Supervision; Writing—review & editing. Jae Gun Kwak: Resources; Writing—review & editing. Hye Won Kwon: Conceptualization; Resources. Yujin Kwak: Resources. Jooncheol Min: Resources. Woong-Han Kim: Resources; Supervision. Jeong Ryul Lee: Resources.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Christoph Schmitz and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

Presented at the 34th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 8–10 October 2020.

REFERENCES

ABBREVIATIONS

 
• AR

  Aortic regurgitation

 
• AVB

  Atrioventricular block

 
• IQR

  Interquartile range

 
• LW

  Low body weight

 
• NTVD

  Non-tricuspid valve detachment

 
• PSM

  Propensity score matching

 
• TR

  Tricuspid regurgitation

 
• TV

  Tricuspid valve

 
• TVD

  Tricuspid valve detachment

 
• VSD

  Ventricular septal defect