Original Article | Published:

Impact of balloon atrial septostomy in neonates with transposition of great arteries

Journal of Perinatology volume 31, pages 494499 (2011) | Download Citation



To examine the impact of balloon atrial septostomy (BAS) on cardio-respiratory status, need for prostaglandin E1 (PGE1) and postoperative outcomes in infants with transposition of great arteries (TGA).

Study Design:

Single-center retrospective review of consecutive neonates with dTGA.


BAS was performed in 42 (70%) infants and resulted in a significant increase in minimum (61 to 76%) and maximum (80 to 90%) oxygen (O2) saturations and a drop in FiO2. BAS was ‘successful,’ that is, PGE1 was discontinued in 16 (38%) infants. Three infants died; four infants developed strokes, all of whom had undergone BAS. The duration of hospitalization, ventilation and O2 need did not differ between infants without BAS, ‘successful’ BAS and unsuccessful BAS. PGE1 duration correlated with duration of hospitalization, ventilation, O2 need and peak respiratory severity score (P<0.03).


We speculate that limiting BAS for clinical hypoxemia and aggressive weaning of PGE1 following BAS would improve outcomes.


Transposition of great arteries (TGA) is the second most common congenital cardiac defect encountered in early infancy and the leading reason for transfer to a cardiac unit in the first 2 weeks of life.1, 2 The preoperative management of TGA consists of prostaglandin E1 (PGE1) infusion to maintain ductal patency and balloon atrial septostomy (BAS) to enhance mixing at the atrial level. PGE1 is effective in alleviating hypoxemia in the majority of infants with TGA.3, 4, 5, 6 The value of BAS has been demonstrated and a significant decrease in early mortality demonstrated for severely hypoxic neonates.3, 7, 8, 9, 10, 11, 12 BAS continues to be extensively used for widely varying indications in the era of early surgical repair. Lately, however, concerns about the ‘routine’ use of BAS have emerged, with some studies demonstrating an association with stroke and the recognition that the closure of the atrial septal defect may prolong the surgical repair and contribute to perioperative risk.13, 14, 15 In addition, despite successful creation of an adequate-sized anatomic atrial septal defect in most cases, a high rate of ‘failed BAS’ with suboptimal increase in systemic oxygenation has been reported.16 Apart from clinical improvement in oxygenation, another goal of the BAS procedure is discontinuation of PGE1.17 The purported benefits of discontinuation of PGE1 include elimination of associated adverse effects such as apnea and fever and according to a single study, shorter postoperative length of hospital and intensive care unit stay.18 Whether and in what proportion of infants BAS is efficacious in this particular goal and its impact, if any, on clinical outcomes remain unclear.

Our broad objective in this study was to examine the impact of BAS on cardio-respiratory status, need for PGE1 and postoperative clinical outcomes in infants with TGA, in the era of prompt PGE1 administration and early surgical repair. Our specific aims were threefold: (1) to determine the effects of BAS in neonates with TGA on cardio-respiratory parameters and postprocedure requirement of PGE1, (2) to identify factors associated with successful BAS, defined a priori as discontinuation of PGE1 and (3) to compare overall clinical outcomes between three groups of infants with TGA: no BAS (group 1), successful BAS (group 2) and unsuccessful BAS (group 3).


This was a retrospective chart review of consecutive patients.

Inclusion criteria

All neonates (<1-month-old) with a diagnosis of TGA admitted to Children's Hospital of Michigan Neonatal Intensive Care Unit between 1 January 2002 and 11 January 2008 identified by the neonatal intensive care unit discharge database.

Exclusion criteria

TGA associated with other major extra-cardiac congenital malformations, infants in whom a decision was made not to continue aggressive care and complex congenital cardiac defects (of which TGA was a component) who underwent surgical repair other than an arterial switch or who required PGE1 for coexisting lesions were excluded from further analyses.

The Human Investigation Committee of Wayne State University approved the study and waiver of parental consent. Demographic, clinical and outcomes data were collected by retrospective chart review of infants who met eligibility criteria. Echocardiographic reports of all echocardiograms performed on included infants were reviewed; the initial diagnostic echocardiogram was used to evaluate size of shunts and cardiac function at presentation. The echocardiogram closest to BAS was used to collect data on ‘pre-BAS’ variables and echocardiograms performed within 24 h of the procedure were considered the post-BAS study. A ‘successful BAS’ was defined as successful discontinuation of PGE1 anytime before surgical repair, irrespective of the number of attempts involved. For assessing the effects of BAS, clinical and ventilatory variables were assessed over a 4-h period at three points in time, (1) at presentation, (2) 24 h after BAS and (3) immediately before surgical repair. Respiratory severity scores (RSSs) and modified inotropic score were calculated using standard formulae (mean airway pressure × fraction of inspired oxygen) and {[(dopamine × duration)+(dobutamine × duration)+(nitroglycerine × duration)+(nitroprusside × duration)] × 0.1}+{[(milrinone × duration)+(arginine-vasopressin × duration)] × 2}+{(epinephrine × duration) × 10}.19, 20

BAS was performed either in the catheterization laboratory, which is in close proximity to our Neonatal Intensive Care Unit or at the bedside by pediatric interventional cardiologists. During the earlier part of the study, a preoperative angiography was often performed in infants with TGA to delineate specific coronary artery anatomy in the catheterization laboratory and BAS was performed during the same procedure. At least one attempt to taper and discontinue PGE1 after BAS was made in all cases. PGE1 was restarted in case of significant desaturations, hypoxia or metabolic acidosis, according to the discretion of the clinical team.

Statistical analysis

Descriptive statistics was employed as appropriate (mean±s.d., median and range for continuous variables and number and frequencies for categorical variables). Between groups comparison of mean values was by Student's t-tests and Mann–Whitney U-test for respective parametric and nonparametric datasets. χ2 test was used to compare proportions between groups. One-way analysis of variance and Kruskal–Wallis tests were used respectively for parametric and nonparametric comparisons of continuous variables among three groups. Bivariate correlation using Spearman's ρ was used to correlate duration of PG use and other continuous outcome variables. Binary logistic regression, with Hosmer–Lemeshow goodness of fit was used to examine the association between clinical and echocardiographic variables and ‘successful BAS’ as well as a favorable postoperative outcome. Level of significance was set at P-value<0.05.


Patient characteristics

A total of 88 patients with a discharge diagnosis of TGA were identified from the neonatal intensive care unit database (Figure 1). Of these, 28 patients were excluded (25 met the exclusion criteria, 2 patients died on the way before transfer to our hospital and 1 patient underwent BAS at the referring hospital). Table 1 describes baseline demographic and clinical characteristics of the study cohort. As expected, 70% of our cohort were males and most (n=54, 90%) were term infants. PGE1 was initiated in 58 (96.6%) infants; the two patients who did not require PGE1 had large ventricular septal defects (VSD) with good mixing. TGA with intact ventricular septum was diagnosed in 40 (66.6%) cases, while the remaining 20 (33.3%) had TGA with a small (7), moderate (7) or large VSD (6). Infants who had coexisting defects such as dextrocardia or Taussig–Bing anomaly were reclassified into either TGA with intact ventricular septum or TGA with VSD for study purposes. At presentation, median (range) minimum and maximum saturations (SaO2), maximum fraction of inspired oxygen (FiO2) and maximum RSS were 64.5% (20 to 95), 85% (54 to 96), 60% (21 to 100) and 4.6% (1.5 to 20), respectively. During the preoperative period, 40 (66.7%) infants were mechanically ventilated. Before BAS, 17 (28.3%) infants were on pressors, all of whom were on dopamine, 7 (11.7%) on dopamine and dobutamine, 1 (1.7%) on dopamine, dobutamine and epinephrine drips and another on dopamine and epinephrine. A total of 42 (70%) infants underwent BAS at a median age of 15.5 h (1.5 to 168 h), 8.7 h (0 to 168 h) after admission. Documented indications for BAS were desaturations on pulse oximetry despite PGE1 in 12 (28%), echocardiographic evidence of restrictive mixing in 14 (33.3%), echocardiographic restricted mixing and desaturations in 8 (19%) and ‘routine’ in 8 (19%) cases. The median timing for arterial switch operation was 7 days; there were two infants who underwent delayed surgical repair up to 30 days because of necrotizing enterocolitis and sepsis. Four infants developed a stroke (on ultrasound/MRI) before surgical repair and three infants died, two before surgical repair. Causes of death were ‘multiple organ dysfunction syndrome’ following repair in one infant, necrotizing enterocolitis, Escherichia coli bloodstream infection and pneumonia in another and sudden cardiac arrest, thought to be secondary to myocardial ischemia in the third.

Figure 1
Figure 1

Flowchart of study cohort.

Table 1: Clinical and outcome characteristics of study cohort (n=60)

Effect of balloon atrial septostomy

BAS resulted in the creation of a moderate to large-sized atrial septal defect in all except one infant with a significant decrease in the proportion of infants with a small shunt at the patent foramen ovale (73.8% before and 2.4% after BAS). A moderate-sized atrial shunt was noted after BAS in 87% of procedures performed in the catheterization laboratory and 80% of bedside BAS procedures, whereas a large shunt was documented in 13 and 20%, respectively, which were not significantly different. The single infant with a persistent small shunt underwent BAS at bedside. There were no complications directly related to the procedure. BAS resulted in a significant increase in the minimum and maximum SaO2 and drop in FiO2 requirement 24 h after the procedure (Figure 2). The peak RSSs, fractional shortening and ejection fraction did not significantly change after BAS. The significant improvement of oxygenation following BAS appeared sustained; minimum and maximum SaO2 and maximum FiO2 24 h after BAS were comparable to values just before surgical repair. Only 16 (38%) infants who underwent BAS were weaned off PGE1 successfully (meeting our a priori definition of ‘successful BAS’). Rates of successful discontinuation of PGE1 among infants who underwent BAS in the catheterization laboratory (11/26, 42.3%) and at the bedside (5/16, 31.2%) were not significantly different. The median (range) duration at PGE1 discontinuation was 53.5 h (0.5 to 398 h) following BAS. The initial attempt to discontinue PGE1 was within 24 h of BAS in 18 infants and succeeded in 10 (55.5%) whereas the initial attempt was beyond 24 h in 24 infants and was successful in 6 (25%). Table 2 compares the effects of BAS on oxygenation, ventilatory support and cardiac function before and 24 h after BAS in the two groups of infants with and without successful BAS. Although both groups showed improvements in minimum and maximum SaO2, FiO2 requirement and RSSs, the mean FiO2 requirement and RSSs were significantly higher after the procedure in the group of infants with ‘unsuccessful’ BAS. The one infant with a small shunt at patent foramen ovale even after BAS did not wean off PGE1.

Figure 2
Figure 2

Bar graph depicting cardio-respiratory variables before and 24 h after BAS. *P-value<0.05.

Table 2: Comparison of effects of BAS in the groups of infants with successful and failed discontinuation of PGE1 after the procedure

Factors associated with successful discontinuation of PGE1 after BAS

Table 3 compares characteristics of groups of infants who underwent BAS and subsequent discontinuation of PGE1 and ‘failed’ attempted discontinuation of PGE1. The only significant differences were that the successful BAS group had a higher proportion of males (93.8 vs 61.5%, P=0.021), TGA with VSD (50 vs 7.7%, P=0.014) and a moderate to large shunt at VSD (25.1 vs 3.8%, P=0.002). There were no differences in the shunts at patent foramen ovale, mean minimum or maximum SaO2, FiO2 requirements, ventilatory support (maximal peak inspiratory pressure or RSS), fractional shortening or ejection fraction at presentation among the two groups. On binary logistic regression, using gender, type of TGA (simple vs TGA with VSD), time of BAS and time of initiation of PGE1 as covariates, male gender and TGA with VSD tended to be associated with successful discontinuation of PGE1 (gender odds ratio 0.13 (95% confidence interval 0.01 to 1.23, P=0.07) and TGA with VSD odds ratio 4.05 (95% confidence interval 0.88 to 18.62, P=0.07)).

Table 3: Clinical and echocardiographic variables associated with successful BAS

Comparison of clinical outcomes of three groups of infants: no BAS, successful BAS and unsuccessful BAS: Table 4 depicts a comparison of postoperative clinical outcomes of the three groups. There were no statistically significant differences between groups, although there was a trend towards a shorter duration of ventilation in the successful BAS group. As the groups differed significantly in PGE1 duration, as expected, we correlated duration of PGE1 and clinical outcomes. A significant positive correlation was noted between the total duration of PGE1 and length of hospital stay, duration of postoperative ventilation and oxygen requirement and maximal RSS after surgical repair (all P<0.05).

Table 4: Comparison of three groups of infants: group 1 (no BAS), group 2 (successful discontinuation of PGE1 after BAS) and group 3 (BAS with failed discontinuation of PGE1)


Our findings reiterate the fact that BAS is effective in relieving hypoxemia, with an increase in oxygen saturations and decrease in oxygen requirement at 24 h. A nonrestrictive atrial communication is known to optimize mixing and improve systemic oxygen content and cardiac output as well as reduce left atrial pressures. A continuing need for PGE1 after septostomy in a sizable proportion of infants with TGA has been recently recognized by other investigators.17, 18 Finan et al. and Beattie et al. reported that 44% and 57% of their cohorts could not be weaned off PGE1, respectively.17, 18 Finan's study examined the effects of time of discontinuation of PGE1 following BAS on oxygenation and need for PGE1 reinstitution. Early (<2 h) discontinuation was associated with rebound hypoxemia and a threefold increase in the need for reinstitution of PGE1. In contrast, in our cohort, despite a fairly delayed attempt at PGE1 discontinuation (median 53 h), PGE1 needed to be reinstituted in the majority.

To our knowledge, this is the first study to attempt to identify factors associated with successful discontinuation of PGE1 after BAS. A shunt at the ventricular level, in addition to an adequate atrial communication appears to be a biologically plausible factor for no longer needing PGE1 for ductal patency. There is lack of data on gender or genomic variations in therapeutic response in this population; therefore, the ‘success’ observed in male infants needs further validation. The reasons for the continuing need for PGE1 in infants with TGA despite successful BAS remain unclear. One mechanism could be that PGE1 improves pulmonary blood flow by enhancing ductal patency and by a direct pulmonary vasodilator effect. Direct pulmonary vasodilator effects of PGE1 on pulmonary veins have been previously demonstrated, as it has an association between pulmonary hypertension and dTGA.21, 22, 23, 24

We noted a significant positive correlation with the total duration of PG use and duration of postoperative oxygen and ventilatory requirement and length of hospital stay. Similar data have been reported in another study.18 We speculate that the correlation between duration of PGE1 and adverse outcomes may be related to the effect of PGs on soft-tissue edema and respiratory drive.25, 26 The tissue edema and ‘friability’ might possibly result in delayed wound healing and prolonged ventilatory requirement, resulting in longer hospital stay.

Four infants had strokes confirmed by MRI, a rate in accord with recent data. We did not find a statistically significant difference in stroke rate between groups of infants who underwent BAS and those who did not, although this was probably due to insufficient power. The relationship between stroke, TGA and BAS in neonates has emerged as an intense focus of investigation. McQuillen et al.13, 14 found that about 40% of infants with congenital heart defects had preoperative brain injury and implicated BAS as contributory to preoperative stroke in neonates with TGA. These findings were confounded by lower Apgar scores, lower arterial oxygen saturations and hemodynamic instability. Beca et al. 27 described preoperative brain injury in 30% infants with TGA, hypoplastic left heart or pulmonary atresia with a predominant pattern of white matter injury. Stroke was diagnosed in 5% and BAS did not increase the risk of preoperative brain injury. In another study, Petit et al., in a cohort of 26 infants with TGA found periventricular leucomalacia in 38%, which was not associated with BAS but with lower preoperative oxygenation and longer time to surgery.28

The limitations of our study include its retrospective nature and the inherent associated biases. Our sample size was small, although in line with similar studies. The indications for BAS and reasons for continuing or restarting PGE1 while generally similar varied according to individual physician practices. Despite these caveats, our data from a recent cohort are representative of current practices and outcomes and examine the impact of preoperative management modalities among neonates with TGA, a relatively under-investigated area.


  1. 1.

    , . Transposition of great arteries. In: Keane JF, Flyer DC, Lock JE (eds). Nada's Pediatric Cardiology. Saunders Elsevier: St Louis, 2006, pp 645–659.

  2. 2.

    , , . Racial and temporal differences in the prevalence of heart defects. Pediatrics 2001; 107: e32.

  3. 3.

    , , , , . Role of prostaglandin E1 infusion in the management of transposition of the great arteries. Am J Cardiol 1979; 44: 691–696.

  4. 4.

    , , , , . Use of prostaglandin E1 in infants with d-transposition of the great arteries and intact ventricular septum. Am J Cardiol 1979; 44: 76–81.

  5. 5.

    , . Transposition of the great arteries. Orphanet J Rare Dis 2008; 13: 27.

  6. 6.

    , , , , , . Prostaglandin E1 treatment in patent ductus arteriosus dependent congenital heart defects. J Perinat Med 2004; 32: 368–374.

  7. 7.

    , . Creation of an atrial septal defect without thoracotomy: a palliative approach to complete transposition of the great arteries. JAMA 1966; 196: 991–992.

  8. 8.

    , , , . Effectiveness of the Rashkind procedure in transposition of the great arteries in infants. Circulation 1971; 43(5 Suppl): I1–I6.

  9. 9.

    , . Transposition of the great arteries. Results of palliation by balloon atrioseptostomy in thirty-one infants. Circulation 1968; 38: 453–462.

  10. 10.

    , , . Balloon atrial septostomy in infants with transposition of the great arteries (d-TGA): PGIMER experience. Indian Heart J 1990; 42: 51–54.

  11. 11.

    , , , , . Treatment of d-transposition of the great arteries: management of hypoxemia after balloon atrial septostomy. Am J Cardiol 1981; 47: 299–306.

  12. 12.

    , , . Arterial switch. Pediatr Cardiol 1998; 19: 297–307.

  13. 13.

    , , , , , et al. Balloon atrial septostomy is associated with preoperative stroke in neonates with transposition of the great arteries. Circulation 2006; 113: 183–185.

  14. 14.

    , , , , , et al. Temporal and anatomic risk profile of brain injury with neonatal repair of congenital heart defects. Stroke 2007; 38(2 Suppl): 736–741.

  15. 15.

    , , , , , et al. Role of balloon atrial septostomy before early arterial switch repair of transposition of the great arteries. J Am Coll Cardiol 1992; 19: 1025–1031.

  16. 16.

    , . Transposition of the great arteries in the neonate: failed balloon atrial septostomy. J Cardiovasc Surg 1985; 26(6): 564–567.

  17. 17.

    , , , . Early discontinuation of intravenous prostaglandin E1 after balloon atrial septostomy is associated with an increased risk of rebound hypoxemia. J Perinatol 2008; 28: 341–346.

  18. 18.

    , . Prostaglandin E2 after septostomy for simple transposition. Pedatr Cardiol 2009; 30: 447–451.

  19. 19.

    , , , , , et al. Postoperative course and hemodynamic profile after the arterial switch operation in neonates and infants. A comparison of low-flow cardiopulmonary bypass and circulatory arrest. Circulation 1995; 92: 2226–2235.

  20. 20.

    , , , , , et al. Intravenous arginine–vasopressin in children with vasodilatory shock after cardiac surgery. Circulation 1999; 100(19 Suppl): II182–II186.

  21. 21.

    , , , , , et al. Relaxant actions of nonprostanoid prostacyclin mimetics on human pulmonary artery. J Cardiovasc Pharmacol 1997; 29: 525–535.

  22. 22.

    , , , , , . Prostanoid receptors involved in the relaxation of human pulmonary vessels. Br J Pharmacol 1999; 126: 859–866.

  23. 23.

    , , , , , . Mortality in potential arterial switch candidates with transposition of the great arteries. J Am Coll Cardiol 1998; 32(3): 753–757.

  24. 24.

    , , . Nitric oxide in neonatal transposition of the great arteries. Acta Paediatr 2005; 94: 912–916.

  25. 25.

    , , , , . Soft-tissue swelling in two neonates during prostaglandin E1 therapy. Pediatr Cardiol 1986; 7: 157–160.

  26. 26.

    , . Bone and tissue changes following prostaglandin therapy in neonates. Ann Pharmacother 1996; 30: 269–274.

  27. 27.

    , , , , , et al. Pre-operative brain injury in newborn infants with transposition of the great arteries occurs at rates similar to other complex congenital heart disease and is not related to balloon atrial septostomy. J Am Coll Cardiol 2009; 53: 1807–1811.

  28. 28.

    , , , , , et al. Preoperative brain injury in transposition of the great arteries is associated with oxygenation and time to surgery, not balloon atrial septostomy. Circulation 2009; 119(5): 709–716.

Download references

Author information


  1. Department of Pediatrics, Children's Hospital of Michigan, Wayne State University, Detroit, MI, USA

    • G Hiremath
    • , G Natarajan
    • , D Math
    •  & S Aggarwal


  1. Search for G Hiremath in:

  2. Search for G Natarajan in:

  3. Search for D Math in:

  4. Search for S Aggarwal in:

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to G Natarajan.

About this article

Publication history







Further reading