Introduction
The ductus arteriosus connects the aorta and the pulmonary artery and is considered essential for intrauterine life. After birth, the ductus arteriosus typically closes spontaneously in around 48–72 hours due to muscle contraction and the decrease in placental prostaglandins. If this process does not occur, a patent ductus arteriosus is present. Reference Backes, Hill and Shelton1 Preterm infants and low birth weight are risk factors for this condition. Patent ductus arteriosus is present in approximately 50% of preterm infants. Reference Mallick, Nayak and Tripathy2
Persistence of the ductus arteriosus can lead to critical physiological changes, most notably increased pulmonary flow, which is clinically manifested by respiratory distress and an increased need for oxygen and ventilation. In the long term, this may result in chronic lung injury and bronchopulmonary dysplasia. The pathophysiology of a patent ductus arteriosus also includes a volumetric overload of the left-sided chambers, which can cause systolic dysfunction and heart failure. In cases of a wider patent ductus arteriosus, systemic blood shunting may lead to low systemic blood flow and a drop in diastolic pressure, which may be associated with comorbidities such as necrotising enterocolitis, retinopathy of prematurity, and periventricular haemorrhage. Reference Parkerson, Philip, Talati and Sathanandam3,Reference Dice and Bhatia4
Despite evolving over the years, the optimal patent ductus arteriosus treatment strategy remains controversial. Surgical closure of the patent ductus arteriosus used to be the only option when medical therapy failed; however, it could lead to severe complications, such as retinopathy of prematurity, bronchopulmonary dysplasia, and neurodevelopmental impairment. Reference Weisz, More, McNamara and Shah5,Reference Janz-Robinson, Badawi, Walker, Bajuk and Abdel-Latif6 Therefore, transcatheter percutaneous closure has been developed as an alternative to surgical patent ductus arteriosus closure.
More recently, with the development of the Amplatzer Duct Occluder II Additional Sizes and Piccolo devices in 2015, it has become possible to perform transcatheter percutaneous closure in children weighing less than 2,000 g. Reference Drighil, Jufan and Omrane7 Nevertheless, there are also concerns regarding the procedure’s safety, especially in low-birth-weight infants, who may develop left pulmonary artery stenosis and coarctation of the aorta. Reference Wang, Lin, Hsieh, Wei, Ju and Wu8 Therefore, we aimed to perform a meta-analysis on the efficacy and safety of transcatheter percutaneous closure versus surgical ligation of patent ductus arteriosus in low-birth- weight preterm infants (<2,500 g).
Methods
Enrollment and exclusion criteria
The enrollment criteria included the following: (1) randomised or nonrandomised studies; (2) including low-birth-weight preterm infants (<2.5 kg); (3) diagnosed with patent ductus arteriosus; and (4) comparing transcatheter percutaneous closure with surgical ligation of patent ductus arteriosus. We excluded reviews, meta-analyses, editorials, animal studies, and case reports. Studies on other CHDs or systemic or metabolic diseases were also excluded. Abstracts were included. There were no language restrictions. The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) on 01/17/2023 (PROSPERO ID CRD42023389847).
Endpoints
The outcomes of interest were mortality, haemodynamic instability (haemodynamic instability was defined as a post-ligation cardiac syndrome and the need for inotropic medication), bronchopulmonary dysplasia, major unexpected complications (pneumothorax, chylothorax, infection, and pulmonary artery injury), and associated morbidities (necrotising enterocolitis, retinopathy of prematurity, intraventricular haemorrhage, and periventricular leukomalacia). Haemodynamic instability was defined as a post-ligation cardiac syndrome and the need for inotropic medication.
Literature search
We systematically searched PubMed, the Cochrane Library, and Embase from inception to November 16, 2023. Literature searches were carried out using Medical Subject Headings terms and free-text terms. The search strategy included the keywords “neonates”, “preterm”, “low weight”, “low birth weight”, “patent ductus arteriosus”, “transcatheter”, and “Amplatzer”. Reference lists of included articles and previous meta-analyses were searched for other related studies.
Two independent reviewers (BO and JL) examined the electronic searches and obtained full reports of all citations that were likely to meet the predefined selection criteria. Disagreements were resolved by consensus and discussion with a third reviewer (SE). If data provided by the studies were insufficient or did not meet inclusion criteria, the article was excluded.
Data extraction and quality assessment
Two authors (BO and JL) independently extracted the data following predefined search criteria and quality assessment. Information collected from studies included: author study, year of publication, number of patients, sex distribution, gestational age at birth, birth weight, age at intervention, weight at intervention, and ductus arteriosus size.
The risk of bias was assessed with the Cochrane Risk of Bias in Non-Randomized Studies of Interventions tool, following recommendations from the Cochrane Handbook for Systematic Reviews of Interventions. Two independent investigators (CM and MP) performed the risk of bias. Disagreements were resolved with the senior author (SE). Reference Wang, Lin, Hsieh, Wei, Ju and Wu8
Statistical analysis
The Cochrane Collaboration Review Manager version 5.4 software was used for statistical analyses. We calculated odds ratios for dichotomous data and weighted mean differences with 95% confidence intervals for continuous data. The Cochran Q test formally tested inter-study heterogeneity of treatment effects (p < 0.10). The I2 statistic was examined, and we considered I2 ≥ 25% to indicate significant heterogeneity between the trials. We adopted a fixed-effects model when I2 < 25%; otherwise, the origination of the heterogeneity was analysed to verify whether a random-effects model could be used.
Results
Search results
The initial search yielded 780 publications, of which 725 were excluded by review of title and abstract or due to duplicate records. The remaining 55 studies were reviewed in full. Of these, 43 studies were excluded due to overlapping populations, no transcatheter percutaneous closure or surgical ligation group, no outcome of interest, or other exclusion criteria. Ultimately, 12 retrospective cohorts were included (Fig. 1).
Figure 1. PRISMA flow diagram of study screening and selection.
Study characteristics
The included studies were published between 2017 and 2023. In total, there were 4,668 preterm infants with low birth weight and patent ductus arterious, of whom 966 (20.7%) were in the transcatheter group and 3,702 (79.3%) patients were included in the surgical group. Table 1 presents the baseline characteristics of the studies analysed in this meta-analysis, including the average gestational age and range of ductus arteriosus size.
Table 1. Characteristics of the trials included
DA = ductus arteriosus; TCPC = transcatheter percutaneous closure; SL = surgical ligation; NA = not applicable.
Pooled analysis of all studies
All-cause mortality was significantly reduced with transcatheter percutaneous closure compared to surgical ligation (OR 0.28; 95% confidence interval 0.18–0.43; p < 0.00001; I2 = 0%, Fig. 2). Similarly, haemodynamic instability was less frequent in the transcatheter group (OR 0.10; 95% confidence interval 0.05–0.21; p < 0.001; I2 = 14%; Fig. 3).
Figure 2. All-cause mortality was significantly reduced in the transcatheter closure group.
Figure 3. Haemodynamic instability was significantly reduced in the transcatheter closure group.
There was no significant difference between the transcatheter and surgical groups for the outcome of bronchopulmonary dysplasia (OR 0.93; 95% confidence interval 0.46–1.87; p = 0.83; I2 = 0%; Fig. 4) or major complications (OR 0.76, 95% confidence interval 0.34–1.69; p = 0.51; I2 = 43%; Fig. 5). Major procedural complications were defined as pneumothorax, chylothorax, infection, or injury to the pulmonary artery.
Figure 4. Bronchopulmonary dysplasia was not significantly different between transcatheter closure and surgical ligation.
Figure 5. Major unexpected complications were not significantly different between transcatheter closure and surgical ligation.
Similarly, the incidence of patent ductus arterious-related morbidities was not significantly different between groups. These included intraventricular haemorrhage (OR 0.87, 95% confidence interval 0.43–1.76; p = 0.69; I2 = 39%, Fig. 6), necrotising enterocolitis (OR 0.85, 95% confidence interval 0.43–1.67; p = 0.63; I2 = 0%, Fig. 7), periventricular leukomalacia (OR 0.49, 95% confidence interval 0.31–0.77; p = 0.002; I2 = 0%, Fig. 8), and retinopathy of prematurity (OR 0.70, 95% confidence interval 0.53–0.93; p = 0.01; I2 = 0%, Fig. 9).
Figure 6. Intraventricular haemorrhage was not significantly different between transcatheter closure and surgical ligation.
Figure 7. Necrotising enterocolitis was not significantly different between transcatheter closure and surgical ligation.
Figure 8. Periventricular leukomalacia was significantly reduced in the transcatheter closure group.
Figure 9. Retinopathy of prematurity was significantly reduced in the transcatheter closure group.
Risk of bias assessment
The risk of bias was assessed with the ROBINS-I tool.Reference Sterne, Hernán and Reeves 9 Seven studies were classified as serious risk of bias, whereas four were deemed at moderate risk of bias. A comprehensive assessment is outlined in Table 2. The increased risk of bias was mostly related to the risk of confounding. Due to the non-randomised nature of the studies, they were subject to confounding factors, particularly related to clinical severity. Six studies reported a tendency of sicker infants to be allocated into the transcatheter closure group, due to a perception of high surgical risk. Reference Lenoir, Wanert and Bonnet10–Reference Tabb, Aggarwal, Bajaj and Natarajan16
Table 2. Risk of bias summary for non-randomised studies (ROBINS-I)
Discussion
To our knowledge, this is the first meta-analysis comparing percutaneous transcatheter for patent ductus arteriosus closure with surgical ligation. Transcatheter percutaneous closure was associated with reduced all-cause mortality and lower haemodynamic instability. In addition, we discovered that transcatheter percutaneous closure protected against retinopathy of prematurity and periventricular leukomalacia.
There was no significant difference in bronchopulmonary dysplasia, severe complications, and related morbidities such as necrotising enterocolitis and intraventricular haemorrhage between groups. Despite the differences in group size, the more stringent inclusion criteria (preterm birth, weighing less than 2,500 g) allowed both groups to be reasonably homogeneous and comparable (see Table 1).
Different publications define post-ligation syndrome differently based on clinical evaluations, therapies, and echocardiographic findings. For example, the Tennessee group considers post-ligation syndrome clinically significant if four of seven criteria are met during the first 24 hours after the procedure: a) a peak inotrope score of >15, b) a ten mmHg decrease in mean blood pressure from baseline, c) evidence for new-onset pulmonary venous congestion chest X-ray, d) a 30% increase in respiratory severity score, e) a 20% absolute decrease in ejection fraction, f) left ventricular output 200 ml/kg/min, and g) a >20% decrease in tissue Doppler-derived lateral and medial E’). Reference Philip, Waller and Chilakala17 To simplify the definition and analysis process, we describe post-ligation syndrome as hypotension requiring inotropic support and failure of oxygenation and ventilation, which may occur 6–12 hours following ligation due to left ventricular systolic and diastolic failure, respectively. This decompensation is primarily driven by increased afterload. Reference Giesinger, Bischoff and McNamara18
The definition of a haemodynamically unstable patent ductus arteriosus needs to be clarified, and there is no consensus. Reference Lee19 Currently, most doctors decide to intervene based on clinical evidence and echocardiographic criteria to define a haemodynamically significant ductus arteriosus better. Reference McNamara and Sehgal20 However, there is variation between groups regarding the specific criteria. In this review, most articles consider a ductus arteriosus to be haemodynamically unstable when it is larger than 2 mm, with criteria of haemodynamic instability, such as using vasoactive drugs.
Pulmonary congestion and low systemic circulation, in the context of a patent ductus arteriosus, can lead to compromised perfusion to the bowel, kidney, and brain. This pathophysiology justifies prolonged assisted ventilation, higher mortality rates, bronchopulmonary dysplasia, necrotising enterocolitis, impaired renal function, intraventricular haemorrhage, periventricular leukomalacia, and cerebral palsy. Reference Mitra, Scrivens, Von Kursell and Disher21
We found no difference in the outcome of major complications. However, there was less retinopathy and leukomalacia in the transcatheter percutaneous closure group. This result is likely due to the need for greater haemodynamic changes associated with surgery. Importantly, with technological advancements, complications of device transcatheter percutaneous closure implantation can now be minimised, particularly in extremely low-birth-weight infants. New implant techniques have been developed to reduce device malposition or protrusion in this population. Reference Sathanandam, Gutfinger and Morray22 However, it is important to interpret our findings cautiously as evaluating comorbidities was not our primary objective.
The optimal timing for closing the ductus arteriosus remains a subject of ongoing debate. Acting promptly correlates with a reduced likelihood of associated health complications. Reference Bischoff, Kennedy and Backes23 Historically, catheterisation for ductal closure was reserved for more severe clinical scenarios, often arising from delayed closure (beyond 15 days). With the FDA’s approval of the innovative “Amplatzer Piccolo Occluder” device, extending its use to premature infants weighing >700 g, recent investigations have suggested that early closure may be a viable option, given the risks of comorbidities associated with delayed closure. Reference Regan, Benbrik and Sharma24
Recent studies indicate that early closure is considered safe and may be a preferred option for preterm babies with low birth weight. Reference Barry, Gudausky and Balzer25 These contemporary insights propose that initiating treatment before the duct becomes clinically significant may prevent damage to surrounding tissues. Furthermore, this approach may reduce reliance on nonsteroidal anti-inflammatory drugs, minimise associated side effects, and contribute to shorter hospital stays – a proposition consistent with earlier findings highlighted in a Cochrane review. Reference Mitra, Scrivens, Von Kursell and Disher21
This study has limitations. We faced challenges evaluating respiratory impairment due to a lack of standardisation. Reference Sathanandam, Gutfinger and Morray22,Reference Wei, Chen and Lin26–Reference Philip, Waller and Chilakala30 It was not possible to assess ventilation time as an outcome due to insufficient data. We only used observational studies, which are prone to confounding factors. Multivariable adjusted analyses in the individual studies to correct for severity of clinical presentation were not performed. Of note, confounding by illness severity would most likely favour the surgical group, if present. In other words, if more severe patients are excluded from the surgical group, this would typically lead to worsened outcomes among the sicker population of the transcatheter group. Nevertheless, we observed the opposite result, increasing the confidence in the findings. Finally, although most papers used clinical and echocardiographic parameters to determine which ductus arteriosus was haemodynamically significant, no definitive and uniform diagnostic criteria existed. Similarly, the criteria for respiratory impairment and the definition of post-ligation syndrome varied between different studies. Future research may benefit from the standardisation of outcome definitions.
Conclusion
In conclusion, our findings indicate that transcatheter patent ductus arteriosus closure in low-birth-weight premature infants is associated with lower all-cause mortality and haemodynamic instability than surgical treatment, with no significant difference in safety outcomes. We also found that transcatheter closure protects against retinopathy of prematurity and periventricular leukomalacia compared with surgical treatment.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1047951123004353.
Financial support
There were no external funding sources for this study.
Competing interests
None of the authors have any conflict of interest to disclose. All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.
Registration
Registered in PROSPERO Database (CRD42023389847).
Conception and design of the research and statistical analysis
Conception and design of the research: SE; acquisition, analysis, and interpretation of the data: MP, SE, CM, GN, BO, ACT, JL; writing manuscript: MP, SE, CM, GN, BO, ACT, JL; critical revision of the manuscript for intellectual content: CM, MP, SE.
