Percutaneous Closure of the Patent Ductus Arteriosus in Infants ≤2 kg: IMPACT Registry Insights

Data were obtained following strict quality assurance standards from the IMPACT Registry, as has been previously described.³⁵  The current study included data from IMPACT, version 2.1. Data elements and definitions are available online at the NCDR Web site.

Patients who were ≤2 kg at the time of percutaneous PDA closure in a contemporary cohort (April 2016-June 2021) were included. Exclusion criteria included history of cardiac catheterization or surgery, diagnosis of a congenital heart defect (with the exclusion of patent foramen ovale or atrial septal defect), and infants undergoing additional procedures besides PDA closure (i.e., atrial septal defect closure).

We assessed age at intervention (<30 and ≥30 days), weight at the time of the procedure (<1000, 1001–1500, and 1501–2000 g), sex, presence of known genetic or chromosomal syndrome (the database provided a list of preexisting genetic conditions such as Down syndrome, 22q11 deletion), and presence of nonrespiratory comorbidities at the time or up to 48 hours before arrival to the catheterization laboratory (neonatal-specific preprocedure conditions included necrotizing enterocolitis, sepsis). Center-specific data of hospital-level annual PDA catheterization case volumes in patients ≤2 kg was also analyzed and stratified in quartiles.

Mode of sedation (general anesthesia, oral/intranasal), airway status (i.e., previously intubated, elective intubation, continuous positive airway pressure) and preprocedural use of diuretics and pulmonary vasodilators were also recorded.

PDA characteristics (angiographic type based on the Krichenko classification, minimum diameter, diameter at aortic side, and length) were documented. Ductal length <4 mm was considered a “short” PDA. Device specifications were recorded when available. Type of access (venous versus arterial or both), fluoroscopy time and dose, and duration of procedure were also recorded. Hemodynamic data, including systolic pulmonary artery pressure, ratio of pulmonary to systemic blood flow, and indexed pulmonary vascular resistance were documented.

Primary outcome was defined as the composite of technical failure and/or MAE. The individual components of the primary outcome were also analyzed. Procedural failure was defined as unable to successfully place a device in the PDA. In the situation of device embolization (which was successfully retrieved) but the PDA was closed, during the same procedure with a different device, procedural success was assigned. MAEs were defined a priori according to previous publications: cardiac arrest, cardiac tamponade requiring pericardial drainage, airway event requiring escalation of care (defined as episode of apnea, hypoxia, or obstruction during the procedure requiring intubation/reintubation), device malposition or thrombus requiring surgery, device embolization requiring device retrieval, unplanned cardiac or vascular surgery attributed to the catheterization, subsequent cardiac catheterization within 72 hours, arrhythmia requiring treatment (cardioversion, antiarrhythmic medication, or pacemaker) and major bleeding event (defined as bleeding between start of the procedure and 72 hours postprocedure, bleeding at access site, or hematoma at access site and associated with 1 of the following: (1) hemoglobin drop of ≥3 g/dL; (2) transfusion of red blood cells; (3) intervention at the bleeding site to reserve, stop, or correct the bleeding).²⁰  Adverse events were included up to 30 days following the catheterization procedure unless otherwise indicated. Unplanned surgery and need for subsequent cardiac catheterization were considered until the time of hospital discharge. Because of the inability to attribute all adverse events to the cardiac catheterization, MAEs were chosen if they could be definitively linked to the procedure (i.e., device embolization), or highly likely to be attributed (i.e., cardiac tamponade). Death was not included as an MAE because in this version of the IMPACT Registry, it cannot be directly linked to cardiac catheterization given the multiple competing confounding illnesses in extremely preterm infants. We did, however, record deaths in the first 24 hours (highly likely to be attributed to the procedure), first 7 days, and 30 days after the procedure. Minor adverse events such as left pulmonary artery or arch obstruction not requiring intervention were not recorded as part of the Registry.

Data are shown as median (interquartile range [IQR]) for continuous variables and n (%) for categorical. Infants experiencing the composite outcome of technical failure and/or MAE were compared with those who did not experience it by the Wilcoxon test for continuous variables and χ² or Fisher exact tests for categorical variables. Modified hierarchical logistic regression was used to estimate the odds ratio of covariates predicting MAE or composite success endpoints with a random intercept to account for the clustering of patients in hospitals. A priori variables for the multivariable models included interaction of age groups (<30 and ≥30 days) and weight groups (<1 and ≥1 kg)³⁶  at the time of procedure into their 4 distinct pairings (with the largest and older infants as the group of refence), any arterial access and annual hospital vol of PDA percutaneous closure in ≤2 kg (hospitals with higher than the median case volumes versus the others). These thresholds were chosen based on expert opinion of perceived higher risk profile and based on historical studies of interventional PDA closure.^36,37  Results were considered significant if P < .05. SAS version 9.2 (SAS Institute, Cary, North Carolina) was used for all analyses. Given the US Food and Drug Administration approval of the use of the Amplatzer Piccolo Occluder for infants weighing >700 g, we analyzed outcomes in a high-risk subpopulation of <700 g.

The IMPACT Registry recorded an attempted 1594 percutaneous PDA closures in patients ≤2 kg between April 2016 and June 2021. Procedures among infants with prior cardiac surgery or catheterization (n = 3) or undergoing additional procedures beyond PDA closure (n = 4) were excluded from analysis. The final cohort for analysis included data from 87 hospitals in 1587 infants ≤2 kg undergoing attempted percutaneous PDA closure. Figure 1 shows the volume of cases performed in each quarter. A total of 628 (39.5%) of the procedures were performed in the final 4 quarters (last year) included in the analysis. Among 1587 attempted PDA closures, 1549 (97.6%) had a device successfully implanted, MAE occurred in 61 (3.8%) and the composite outcome occurred in 85 (5.3%) of the patients.

Median (IQR) gestational age at birth was 25 (24, 27) weeks and approximately 37% (n = 565) of the infants were born <25 weeks. Approximately half of the population were male infants (n = 804, 50.7%). Median (IQR) age and weight at intervention were 37 (27, 49) days and 1.2 (1.0, 1.6) kg, respectively. There were 16 patients who were <700 g at the time of percutaneous PDA closure. Incidence of major nonrespiratory comorbidity at the time of catheterization was 8% (n = 52 [5.8%] with necrotizing enterocolitis and 19 [2.1%] with sepsis, of 886). Genetic/chromosomal syndrome was described in 16 (1%). Table 1 compares demographics and procedural characteristics of infants experiencing the composite outcome of technical failure and/or MAE to characteristics of those not experiencing the composite outcome. Figure 2 depicts the distribution of the composite outcome according to weight (panel A) and age (panel B) at the time of the procedure. A biphasic curve was noted when assessing age at the time of procedure with an initial decrease in the incidence of the composite outcome between the first postnatal week and approximately 30 days and a secondary mild increase in the incidence around 40 to 50 days.

The median (IQR) hospital-level annual PDA catheterization case volumes in patients ≤2 kg was 2.7 (1.1, 5.6). PDA catheterization volume among infants ≤2 kg (number of hospitals) across each quartile is as follows: first quartile, 0.2 to <1.1 (n = 22); second quartile, 1.1 to <2.7 (n = 22); third quartile, 2.7 to <5.7 (n = 22); fourth quartile, 5.7 to 21.9 (n = 21).

Many infants (n = 530, 33.3%) were receiving diuretics before intervention and a small number (n = 40, 2.5%) were receiving vasodilators. Most patients were intubated before the procedure (n = 1121, 70.6%) and nearly all interventions were performed under general anesthesia (n = 1565, 98.6%).

The median (IQR) minimal PDA diameter, diameter on the aortic side, and length were 2.6 (2, 3.2), 3.9 (3.3, 4.5), and 9.8 (8, 11.3) mm, respectively. A total of 32 (2%) were considered short PDAs. Exclusive venous access was used in 1537 (96.8%) of cases. Among 1549 infants with successful device closure, 685 (44.2%) had the Amplatzer Piccolo Occluder (Piccolo, Abbott, Plymouth, Minnesota), 378 (24.4%) had the microvascular plug (Medtronic, Minneapolis, Minnesota), 249 (16%) had the Amplatzer Ductal Occluder II Additional Sizes (ADO-II-AS, Abbott), and 67 (4.3%) had the Amplatzer Vascular Plug - II (AVP-II, Abbott). Other various devices were used in 170 patients (10.9%). Additional procedural characteristics are available in the Supplemental Information.

Among procedural failures (n = 38, 2.3%), the device was implanted but not released (n = 16), or implanted, released, retrieved, and not subsequently placed successfully during the same case (n = 22). Rates of procedural failure were similar across weight (<1000 g, 1.6%; 1000–1499 g, 2.7%; 1500–2000 g, 2.4%; P = .561) and age (<30 days, 1.6%; ≥30 days, 2.7%; P = .17) thresholds. Technical failure was more common in patients with some arterial access versus those with only venous access (8.3% vs 2.1%; P = .006). Procedure duration was longer in patients with technical failure versus those who were successful (106 [69–162] vs 49 [35–71] minutes; P < .001). There were no technical failures in patients <700 g.

Composite outcome of MAE and/or technical failure occurred in 85 (5.3%) of the total cohort and 2 of 16 (12.5%) patients <700 g. Covariates associated with the composite outcome are shown in Table 2. Hospitals with an annual volume of cases above the median had lower incidence of the composite outcome. Figure 3 depicts the rate of the composite outcome and the median weight at the time of the procedure over time.

Covariates associated with the likelihood of MAEs are shown in Table 3. Hospitals with an annual volume of cases above the median had less incidence of MAEs. The most common MAE was device embolization requiring retrieval (n = 20 [1.3%], 14 of which were retrieved via catheterization, 4 requiring surgery). Other MAEs reported included: cardiac arrest (n = 11, 0.7%), cardiac tamponade (n = 5, 0.3%), device malposition or thrombus requiring surgery/intervention (n = 10, 0.6%), unplanned cardiac or vascular surgery (n = 20, 1.3%), need for subsequent catheterization (within 72 hours) (n = 9, 0.6%), arrhythmia requiring treatment (n = 10, 0.6%), major bleeding event (n = 9, 0.6%), and airway escalation of care (n = 8, 0.5%). The rate of MAE in patients less than 700 g was 12.5% (2 of 16 patients) vs 3.7% (59 of 1571) in patients 700 to 2000 g.

Secondary outcomes are reported in the Supplemental Information.

Across 87 hospitals, we observed 1587 attempted percutaneous PDA closures in infants ≤2 kg, with a 97% technical success rate and 3.8% MAEs. This is the largest report of short-term outcomes of percutaneous PDA closure in the most immature neonates. The 2017 analysis of the IMPACT Registry reported clinical outcomes of only 19 patients who were ≤2 kg at the time of the procedure.²⁰  Since then, the practice of percutaneous PDA closure has continued to increase and has expanded exponentially toward smaller infants. In the current study, almost 40% of the procedures were performed in the final year that was included. The median patient weight has decreased to 1.2 kg and the incidence of MAE and/or technical failure has also diminished. Importantly, the rates of procedure failure, MAEs and composite outcome were irrespective of age or weight at intervention. This is similar to the findings of a recent meta-analysis including infants ≤1.5 kg in which year of publication/procedure was directly associated with technical success and adverse events, although the probability of technical failure was inversely related to age.²¹  In addition, the results of the Amplatzer Piccolo Occluder trial support this trend by reporting a success rate more than 95% with a small risk of safety endpoints.³⁴ 

There has been a secular trend toward a more conservative/nonintervention approach to the management of PDA in extremely preterm infants, which has been compounded by the fact that multiple trials of medical therapy have failed to show benefit.³⁸  However, there are data from observational studies reporting an association between PDA, duration of shunt exposure, and neonatal comorbidities.^39,40  A post hoc analysis on the PDA randomized controlled trial cohort showed that early elimination of shunt may reduce respiratory morbidity,⁴¹  whereas a recent epoch study revealed that transition to a nonintervention approach to the PDA resulted in an absolute increase of 31% in bronchopulmonary dysplasia in infants born preterm at less than 26 weeks’ gestational age.⁹  Furthermore, prolonged PDA exposure has a known detrimental effect in the pulmonary vascular bed development and increased risk for pulmonary hypertension.^11,42,43  Therefore, given the lack of comparative data between conservative versus definitive intervention in high-risk populations, it is imperative to understand the safety endpoints of percutaneous PDA closure in the care of progressively smaller neonates.

According to the current study, percutaneous PDA closure is a relatively short procedure such that 92% of the cases are done in less than 2 hours and fluoroscopy time is on average 7 minutes. Most PDAs were classified as type C and the incidence of short PDAs was extremely small. Additionally, most procedures were performed exclusively through venous access. Patients with the composite outcome, procedure failure or MAE, had a higher incidence of arterial access. The absence of arterial injuries reported, and small number of major bleeding events (n = 9), suggests that the need for arterial access may reflect the technical complexity of some cases. When there are no complications and/or technical difficulties, arterial access is not indicated nor routinely performed in small infants.^32,34  This hypothesis is further corroborated by the longer duration of the procedure in those experiencing the composite outcome.

The composite outcome decreased linearly during the first postnatal month. We observed a small secondary increase at approximately 40 to 50 days (Fig 1). This may reflect procedural learning curve as the procedure was performed in older and bigger neonates in the earlier years of the cohort. With increased procedural volumes and experience, the composite outcome decreased from approximately 10% to 5%, despite decreasing weight at intervention (from median of 1.5 to 1.2 kg) (Fig 2). Alternatively, morphologic aspects of the PDA at a later age must be taken into consideration. It is possible, for example, that exposure to more courses of medical therapy could affect the shape of the ductus and make the procedure technically more challenging and prone to MAEs and/or technical failure.

The multivariable models also suggest that annual hospital volumes above the median affect the incidence of the composite outcome and MAEs. The median annual number of percutaneous PDA closures in infants ≤2 kg was very small, reflecting the average of procedures over a 5-year timespan, even though almost half of the procedures were performed in the final 1 to 2 years of the cohort. Nevertheless, it is still important to reflect that centers with higher volumes and expertise are more likely to have better outcomes. These findings align with the observations of another study which showed that centers who reported at least 10 cases of percutaneous PDA closure in patients ≤1.5 kg had lower rates of adverse events.²¹ 

The current cohort included 16 infants who were less than 700 g at the time of PDA closure, all of which were technically successful. This subset of patients had 2 MAEs and 1 infant death from cardiac arrest. Although the Amplatzer Piccolo Occluder trial only included patients who were at least 700 g,³⁴  other studies have reported outcomes below this thresold.^19,24,34  A small single-center experience of 18 infants <700 g reported 100% technical success with 1 MAE (cardiac arrest successfully resuscitated) and no deaths,⁴⁴  whereas another larger single-center experience of 100 infants ≤1000 g reported 6 adverse events: left pulmonary artery stenosis requiring device retrieval (n = 2), inferior vena cava laceration with death, pericardial effusion requiring drain placement and aortic arch obstructions requiring stenting (n = 2).⁴⁵  Specific recommendations regarding the technical aspects of percutaneous PDA closure in smaller infants have been published.³²  The absence of technical failures in this subset may relate to intervention in the smallest of neonates only in centers with high annual volumes and more experience.

The main limitations of this study are related to the nature of information that can be gathered from a registry. Detailed information regarding individual patient outcomes and its relationship to the procedure itself cannot be obtained. The IMPACT Registry allows for a multicenter analysis of important metrics without the risks for publication and selection bias. However, other possible major and minor adverse events may not be fully captured under this study design. For instance, the version of the Registry that was used in this study does not include data on tricuspid regurgitation/valve injury, which has been previously described to occur in 5% of infants ≤2 kg.³⁴  Similarly, it is not possible to adequately determine the attributability of deaths to the procedure itself or the risk profile of infants who had catastrophic outcomes. Furthermore, with the current iteration of the Registry, we are unable to make comparisons between the different devices and/or techniques that are currently available. Finally, this study did not evaluate other important metrics such as the effect on short term cardiopulmonary health, the risk for postclosure cardiorespiratory instability, possible long-term unanticipated adverse events, and the influence on neonatal morbidities. The data in this study do not provide insights regarding patient-individual benefits of the intervention.

Percutaneous PDA closure appears to be a safe and feasible procedure in infants ≤2 kg, although evidence for infants <700 g remains limited and requires careful monitoring. The incidence of MAEs has continued to decline and is strongly influenced by individual center case volumes and expertise. The establishment of regional high-volume centers of excellence for procedural intervention may help mitigate risk for technical failure and MAEs. Additionally, collaboration between neonatologists, hemodynamic specialists, and interventional cardiologists is imperative to ensure ongoing quality of studies assessing clinically relevant outcomes.

IMPACT Registry

Improving Pediatric and Adult Congenital Treatments

IQR

interquartile range

MAE

major adverse event

NCDR

National Cardiovascular Data Registry

PDA

patent ductus arteriosus

Qp:Qs

ratio of pulmonary to systemic blood flow