Use of pulmonary artery flow restrictors in neonates with congenital heart disease: Timing and technical considerations

Abstract

Objective

Neonates with congenital heart disease often require restriction of pulmonary blood flow while awaiting definitive palliation or repair. Historically, these patients were managed with pulmonary artery banding. Recently, the use of transcatheter-modified Micro Vascular Plugs as pulmonary flow restrictors has presented a promising alternative for these high-risk patients.

Methods

We report our experience with transcatheter pulmonary flow restrictors in neonates with congenital heart disease and increased pulmonary blood flow. Outcomes assessed included successful device placement, procedural complications, procedural mortality, late device migration, progression to surgery, and need for pulmonary artery repair at the time of surgical device removal.

Results

Between 2023 and 2025, 9 neonates (median age 14 days) with congenital heart disease, pulmonary overcirculation, and other preoperative risk factors received transcatheter pulmonary flow restrictors. Two patients received a third pulmonary flow restrictor as a second procedure. Twenty pulmonary flow restrictors were placed (17 Micro Vascular Plug 7Q and 3 Micro Vascular Plug 9Q devices). All patients survived the procedure. There were 2 early device migrations and no other procedural complications. In 2 patients, late pulmonary flow restrictor migration left a right and left upper lobe artery unprotected, requiring additional pulmonary flow restrictors. Six patients were successfully bridged to surgery (median interval: 29.5 days), and 1 patient had the transcatheter device removed after 2 weeks. At surgery, the removal of pulmonary flow restrictors was uncomplicated, up to 5.5 months postimplantation, and no pulmonary artery repair was required.

Conclusions

Transcatheter palliation using pulmonary flow restrictors is a safe and effective strategy for neonates with congenital heart disease and increased pulmonary blood flow. This approach offers a viable alternative to pulmonary artery banding for bridging high-risk patients to definitive surgical repair, including delayed Norwood.

Angiograms before (A) and after (B) the second flow restrictor was placed in the proximal right PA to protect the right upper lobe artery.

Central Message.

Early intervention, sizing techniques to minimize risk of device migration, and using a second device in a single vessel if needed are keys to successfully bridging to surgical repair/palliation.

Perspective.

Transcatheter implantation of PFRs is a useful alternative to surgical PA banding in neonates with complex CHD. This approach is a less-invasive yet effective bridge to definitive surgical repair or palliation, including a delayed Norwood. Early decision-making and early intervention are keys to optimal device placement and clinical stabilization.

Restriction of excessive pulmonary blood flow is crucial in the stabilization and palliation of neonates with hypoplastic left heart syndrome (HLHS) and other complex congenital heart disease (CHD) defects. Medical management is often ineffective, and surgical pulmonary artery banding (PAB) has been a traditional approach to reduce pulmonary blood flow. PAB is commonly pursued for neonates with pulmonary overcirculation and known risk factors for poor outcomes such as prematurity, low birth weight, extracardiac malformations, preoperative mechanical ventilation, and shock-related end-organ dysfunction.1, 2, 3 Recently, an innovative transcatheter method for restricting pulmonary artery (PA) blood flow with a modified endovascular pulmonary flow restrictor (PFR) has emerged as an alternative to PAB.4, 5, 6

Khan and colleagues⁴ initially described this approach in a swine model using a manually modified Medtronic Micro Vascular Plug (MVP) positioned in the proximal branch pulmonary arteries. In 2020, Shranz and colleagues⁵ described 6 neonates successfully treated using the modified MVP without mortality during implantation or subsequent surgery. In 2023, Haddad and colleagues⁷ reported on 14 patients, expanding the spectrum of infants with CHD safely treated with PFRs. Next, Sperotto and colleagues⁸ published their experience with PFRs in high-risk single-ventricle neonates (n = 17) and compared outcomes with a historical cohort palliated with a hybrid stage 1. In that study, PFRs resulted in a significantly lower all-cause 6-month mortality risk compared with the historical cohort.

Given the paucity of data, we report our experience with PFRs in neonates with CHD and increased pulmonary blood flow in reference to procedure timing, safety, technique, feasibility of transcatheter and surgical device removal, management of distal device migration, and progression to surgery. We also report on the successful use of PFRs in a 950-g infant with an interrupted aortic arch.

Patients and Methods

Nine consecutive patients who received PFRs from March 2023 to May 2025 at Nemours Children's Hospital, Florida were reviewed. The Institutional Review Board approved this study with a waiver of informed consent (IRB-1790747, 4/25/25).

Definitions

Demographics, clinical data, and patient outcomes were abstracted from the electronic medical record. A PFR procedure was considered successful when a PFR was implanted in each PA (Figure 1) with echocardiographic evidence of pulmonary blood flow restriction (Figure 2). Procedural complications included early PFR distal migration, vascular injury, and unanticipated cardiac interventions during PFR procedure. PFR migration was categorized as early if this occurred during the procedure or late if recognized after the patient left the catheterization laboratory (eg, follow-up x-ray). Distal PFR migration was considered significant (thus recorded) if 1 or more lung segments had resultant unrestricted pulmonary blood flow. Procedural mortality was defined as death due to the PFR procedure or within 30 days.

Figure 1.

Main PA angiogram showing placement of PFRs after deployment.

Figure 2.

A, Color Doppler imaging showing flow through flow restrictor. B, Continuous waveform Doppler interrogation showing diastolic runoff suggesting appropriate restriction.

Primary outcomes included procedural success, procedural complications, procedure mortality, and progression to device removal or planned surgical intervention. Secondary outcomes included late device migration and need for PA repair/reconstruction at the time of surgical device removal.

Procedure

The PFR procedure was performed with systemic heparinization and biplane fluoroscopy. Femoral venous access was obtained with a 4F sheath introducer (upsized if concomitant ductal stenting or atrial septostomy was planned). The tricuspid valve was crossed with a 0.018-inch Choice guide wire and standard 2.7F Progreat or a steerable Swift NINJA (Merit Medical Systems, Inc) microcatheter. A 4F multipurpose Radiofocus Glide Cath (Terumo Corp) was used for selective hand angiograms in the left pulmonary artery (LPA) in 30° left anterior oblique/lateral 90° projections and in the right pulmonary artery (RPA) in 30° right anterior oblique/lateral 90° projections. A 40° to 45° caudal with 15° to 20° left anterior oblique/lateral 90° projection was also used to evaluate PA bifurcation and PFR positions before deployment. A Y-connector was placed at the end of the catheter to prevent blood loss and allow angiograms. Angiographic measurements of proximal and distal diameters of both branch PAs were taken for device selection (Video 1). Concomitant procedures and PFRs were performed in the order deemed best by the interventionalist.

After size selection and preparation of the MVP device to create the PFR, the system was inserted in the glide catheter through the Y-connector. At the time of deployment, we placed the distal radiopaque marker at the target area distally and then rapidly uncovered the PFR. If needed, the device was recaptured back into the catheter and repositioned. Before release, an angiogram checked the patency of the upper lobe branches and the relationship of the PFR's proximal end to the pulmonary valve. Echocardiography then confirmed the device position and measured the pressure gradient with continuous-wave Doppler tracing. Bilateral pulmonary vein wedge pressures were also obtained to evaluate distal PA pressures after device deployment. Postprocedure, patients received a 24-hour heparin infusion to maintain an activated partial thromboplastin time of 60 to 80 seconds, and aspirin (5 mg/kg) and clopidogrel (0.2 mg/kg) were initiated. For patients with unprotected flow in an upper lobe PA branch after initial PFR placement, an additional device was placed more proximally when feasible (Video 1).

Device Selection and Preparation

MVP devices were originally designed for occlusion of peripheral vasculature. For the same vessel size, a larger device is needed in high-flow, low-resistance branch PAs compared with less dynamic peripheral vessels to prevent migration. Our size selection algorithm was MVP-5Q for vessel diameters less than 3.5 to 4 mm, MVP-7Q for diameters less than 5.5 to 6 mm, and MVP-9Q for less than 8 mm. We also considered the branch PA length before bifurcation given the different device lengths.

Varying techniques for MVP modification have been described.^6,8, 9, 10 For the first 2 patients, we used a Number 11 blade to incise the polytetrafluoroethylene. For the subsequent 6 patients, we used a 168-mm length, 0.5-mm point strut pick to create the appropriately sized hole in the polytetrafluoroethylene membrane (Figure 3 and Video 1).

Figure 3.

PFR fenestration technique. A, 7Q MVP system. B, House strut pick (0.5 mm, 90° 6 3/4) used to create the fenestration in the MVP. C, Careful manipulation of instruments to pierce the cover. A twisting motion is used. D, Final result of fenestration in 7Q MVP.

Results

Primary Outcomes

Between 2023 and 2025, 20 devices (17 MVP-7Q and 3 MVP-9Q) were successfully placed in 9 neonates (median age: 14 days [range, 8-51]; median weight: 2.9 kg [range, 0.95-4.5]) with CHD (Table 1). Patients selected for PFR commonly had complex CHD with significant preoperative risk factors such as prematurity and end-organ dysfunction.

Table 1.

Characteristics of study population^∗

Patient	GA (wk + d)	Birth weight (kg)	Cardiac diagnosis	Comorbidities	Weight at catheterization (kg)	Age at catheterization (d)	Pulmonary flow restrictors used	Weight at surgery (kg)	Age at surgery (d)	Surgical procedure	Days with PFR
1	39 + 0	2.995	Interrupted aortic arch (type B), VSD	VACTERL syndrome, rectovaginal fistula repair, imperforate anus, gastrostomy tube	3	9	LPA: 7 mm RPA: 7 mm	3.5	45	Arch reconstruction, patch closure of VSD, PFR removal	36
2	39 + 2	3.9	HLHS (MS/AA)	Seizures	4.5	9	LPA: 7 mm RPA: 9 mm	4.7	43	Delayed Norwood with RV to PA shunt, atrial septectomy, PDA ligation, PFR removal	34
3	39	3.07	HLHS (MS/AS)	Low cardiac output syndrome	3.1	8	LPA: 7 mm RPA: 7 mm	3.9	35	Delayed Norwood with RV to PA shunt, PFR removal	27
4	37 + 2	3.185	DORV, TGA, hypoplastic transverse arch	Left vocal cord paresis	3	9	LPA: 7 mm RPA: 9 mm	4.65	22	Arch reconstruction, PFR removal, atrial septectomy, MPA banding	13
5	36 + 3	2.84	HLHS (MS/AS)	Hyperbilirubinemia, coagulopathy, gastrojejunostomy tube	2.6	17	LPA: 7 mm RPA: 7 mm	3.1	43	Delayed Norwood with RV to PA shunt, DKS, atrial septectomy, PFR removal	26
6 – Procedure 1	35 + 4	2.31	uAVC, RV hypoplasia, DORV, mildly hypoplastic arch, cor triatriatum sinister	Heterotaxy, asplenia, direct hyperbilirubinemia, malrotation without obstruction (had Ladd's procedure), hypoxic-ischemic encephalopathy, osteopenia, plagiocephaly, paroxysmal sympathetic hyperactivity	2.7	14	LPA: 7 mm RPA: 7 mm	N/A	N/A	N/A	154
6 – Procedure 2				As above + progressive left pulmonary vein stenosis	4.2	144	RPA: additional 9 mm PFR	4.7	168	AV valve repair, PFR removal, MPA band placement, resection of cor triatriatum	24
7	38 + 6	3.1	Shone's syndrome, moderate transverse arch hypoplasia, severe coarctation, VSD, moderately hypoplastic, parachute mitral valve, dysplastic bicuspid aortic valve, PDA with PDA stent, aortic arch stent	Solitary dysplastic right kidney with hydroureteronephrosis with vesicoureteral reflux, nephrostomy tube, VCRL3 mutation, gastrojejunostomy tube	3.5	51	LPA: 7 mm RPA: 7 mm	4.88 (at time of transfer)	116 (from procedure to transfer)	Transferred (parents moved) Deemed not a candidate for surgical palliation or transplant and died	N/A
8 – Procedure 1	28 + 2	0.87	VSD, interrupted arch (type B), anomalous origin right subclavian artery from RPA)	Severe BPD, umbilical hernia, 22q.11.2 deletion syndrome, metabolic bone disease of prematurity	0.955	19	LPA: 7 mm PFR RPA: 7 mm PFR	N/A	N/A	N/A	N/A
8 – Procedure 2					4	145	LPA: additional 7 mm PFR	N/A	N/A	N/A
9 – Procedure 1	34 + 4	2.17	Bicuspid aortic valve, hypoplastic left ventricle, mildly hypoplastic aortic arch, ASD, left SVC to coronary sinus	Prematurity, ambiguous genitalia, pathogenic variant in MYRF	2.4	18	LPA: 7 mm PFR RPA: 7 mm PFR	N/A	N/A	N/A	14
9 – Procedure 2					2.6	32	Transcatheter removal of bilateral PFRs	N/A	N/A	N/A

GA, Gestational age; PFR, pulmonary flow restrictor; VSD, ventricular septal defect; LPA, left pulmonary artery; RPA, right pulmonary artery; HLHS, hypoplastic left heart syndrome; MS, mitral stenosis; AA, aortic atresia; PDA, patent ductus arteriosus; AS, aortic stenosis; DORV, double outlet right ventricle; TGA, transposition of the great arteries; MPA, main pulmonary artery; DKS, Damus-Kaye-Stansel; uAVC, unbalanced atrioventricular canal; N/A, not available; AV, atrioventricular; BPD, bronchopulmonary dysplasia.

^∗For patients undergoing 2 PFR procedures (bold), the procedures are documented separately as “Procedure 1” and “Procedure 2.”

There was 1 early PFR migration (Table 2). There was no procedural mortality. Concomitant catheter-based interventions including patent ductus arteriosus (PDA) stenting and balloon atrial septostomy were performed in 4 patients.

Table 2.

Pulmonary flow restrictor procedural details^∗

Patient	Implanted MVP devices		Procedural complications	Concomitant procedures	Late distal pulmonary flow restrictor migration
	LPA	RPA
1	7Q	7Q	None	No	No
2	7Q	9Q	None	BAS	No
3	7Q	7Q	None	No	No
4	7Q	9Q	None	BAS	No
5	7Q	7Q	None	No	No
6 – Procedure 1	7Q	7Q	Early RPA PFR distal migration (to RLL PA leaving RUL branch unprotected)	No	Yes, left PFR Second RPA 9Q device placed; anatomy of LPA precluded second device placement in LPA
6 – Procedure 2		9Q	None	No	No
7 Deceased	7Q	7Q	None	BAS, PDA stenting	No
8 – Procedure 1	7Q	7Q	None	No	Yes, bilateral Second LPA 7Q device placed; anatomy of RPA (anomalous takeoff of right subclavian artery) precluded device placement in RPA
8 – Procedure 2	7Q		None	PDA stenting	No
9 – Procedure 1	7Q	7Q	Early RPA PFR distal migration	No	No
9 – Procedure 2	Transcatheter removal of bilateral PFRs	None	No	No

LPA, Left pulmonary artery; RPA, right pulmonary artery; BAS, balloon atrial septostomy; PDA, patent ductus arteriosus.

^∗For patients undergoing 2 PFR procedures (bold), the procedures are documented separately as “Procedure 1” and “Procedure 2.”

Six patients ultimately underwent PFR removal and subsequent surgical repair/palliation (Table 1). The median time between PFR placement and removal was 26.5 days (range, 22-168 days). In patient 6 with 2 devices in the RPA, uneventful surgical removal occurred 168 days (5.5 months) after the device placement.

One patient (patient 9) had borderline left-sided structures at birth and underwent PFR placement for pulmonary overcirculation. Two weeks later, after spontaneous ductal closure, there was no arch gradient and the left ventricle and mitral valve were sufficient to support his cardiac output. Therefore, the PFRs were removed with no further intervention.

Secondary Outcomes

Two patients (6 and 8) experienced late distal PFR migration. Two months after the initial procedure, an additional MVP-7Q device was placed in the proximal RPA for patient 6. In patient 8, an additional MVP-9Q device was placed in the proximal LPA 4 months after the initial procedure (Table 2 and Figure 4).

Figure 4.

Angiogram before (A) and after (B) second PFR placed in proximal RPA to protect uncovered right upper lobe artery 4 months after the initial procedure in patient 6.

For patients who underwent surgical treatment after PFR (n = 6), device removal was straightforward. Intraoperative visual inspection was performed to assess for subtle endothelial or intimal injury. Although the impression of the flow restrictor was noted on the intima, there was no full-thickness injury or significant intimal injury. All devices were removed intact, and no PA repair/reconstruction was required for stenosis, scarring, or vascular injury from the device or its removal. For the patient who underwent device removal 2 weeks postimplantation, the devices were snared using a 5 mm × 125 cm PFM Multisnare through a 4F angled glide catheter advanced through a 5F Braun Compass long sheath placed in the main PA. The PFRs were removed successfully without complication.

Clinical Details

After transcatheter placement of bilateral PFRs, only 2 patients (6 and 7) experienced continued clinical heart failure with failure to thrive. Patient 6 had single-ventricle physiology (heterotaxy syndrome, unbalanced atrioventricular [AV] canal defect, and severe common AV valve regurgitation) and presented in shock with acute renal failure, and she underwent emergency bilateral PFR placement at 14 days of age. After initial recovery, she experienced late migration of the PFRs resulting in an unprotected right and left upper PA bed and worsening heart failure. A second flow restrictor was placed into the proximal RPA. Because of the anatomy of the LPA upper lobe branch, it was not possible to place another flow restrictor proximally in the LPA (Figure 4). This patient subsequently underwent removal of PFRs and repair of common AV valve and main PAB at age 6 months instead of bidirectional Glenn due to elevated distal PA pressures and elevated pulmonary vascular resistance. Patient 7 presented with Shone's syndrome, moderate transverse arch hypoplasia, severe coarctation, large outlet ventricular septal defect, and moderately hypoplastic and parachute mitral valve. This patient also had dysmorphic features with VCRL3 gene mutation and solitary dysplastic kidney. After a multidisciplinary discussion, an initial transcatheter approach was selected to defer the single versus biventricular approach decision to allow growth and achieve end-organ optimization. Over 2 catheterization procedures, the patient underwent arch/coarctation stent placement, bilateral PFR placement, PDA stenting, and balloon atrial septostomy. Bilateral PFR placement was thus undertaken late, at 51 days of age. The hospital course was complicated by renal failure and sepsis. The family relocated to another state 65 days after PFR implantation. Subsequently, the patient underwent a cardiac catheterization elsewhere and was not a candidate for surgical palliation or transplant. The patient died of progressive renal and heart failure at 7 months of age.

Discussion

Placement of bilateral PFRs is a safe and effective approach to manage pulmonary overcirculation in neonates with HLHS and other complex CHD. This technique has been reproducible and advantageous across care settings.9, 10, 11, 12, 13 Here, the procedure was technically successful with no procedural mortality. Surgical removal was uneventful without vascular injury, and no PFR-related branch PA repair was required in subsequent operations. During this experience, we developed an institutional algorithm for use of PFR to facilitate decision-making and tracking of outcomes (Figure 5).

Figure 5.

Clinical decision-making algorithm: PFRs versus bilateral PA bands for neonates with CHD and pulmonary overcirculation.

The benefits of a hybrid stage 1 Norwood (surgical PAB ± PDA stenting) in patients with HLHS, low birth weight, and noncardiac/cardiac comorbidities were recently published.¹⁴ In our experience, early placement of PFRs is a viable alternative strategy for such patients, because all patients with single ventricle physiology who underwent PFR placement before 10 days of age did well and underwent a successful delayed Norwood procedure at 27, 34, and 35 days after placement of PFRs. We undertook surgical intervention when end-organs were optimized, and the baby was extubated, tolerating enteral feeds, and gaining weight.

To mitigate the risk of device migration, the decision for PFR placement should be made early in life before branch PA dilation from decreasing pulmonary vascular resistance complicates the device sizing. Dilated branch PAs pose a challenge due to size limitation of PFRs, increased risk of device migration, and potential need for multiple devices. Therefore, early placement of PFRs is advantageous. It is possible to successfully place bilateral PFRs in very low birthweight premature neonates. Patient 8 was a 28-week premature infant, with a birth weight of 870 g, who underwent PFR placement at 19 days of age and 950 g with no complications. To our knowledge, this is the smallest patient who has undergone bilateral PFRs placement.

All patients were managed with a heparin infusion for the first 48 hours and transitioned to aspirin and clopidogrel, and we did not observe any bleeding or thrombotic complications in this series.

The late sequalae of PFRs left in situ and the potential for interval PFR balloon dilation remain unclear. Here, the longest interval between placement and surgical removal was 5.5 months, and surgical removal was uneventful. For the 6 patients who underwent surgical intervention within 40 days after PFR placement, none had increased PA pressures postoperatively. Likewise, there has been no evidence of stenosis of the branch PAs during follow-up. Successful transcatheter removal of migrated or compressed devices within the first month of the initial placement has been reported.^4,13 In our experience, transcatheter device removal at 2 weeks postimplantation was accomplished without difficulty. However, transcatheter removal with intent to upsize a device 9 weeks after initial implantation (patient 6) was not successful. The device was snared without difficulty; however, it appeared incorporated within the vessel intima, and it was not possible to retrieve it.

Technical Considerations

For reliable sizing of the device fenestration, we premark a house struck pick at the desired diameter level (Video 1). The diameter is determined by the anticipated interstage time (1.5 mm for patients awaiting delayed Norwood and 2 to 2.5 mm for biventricular candidates or if a longer interval is anticipated).

We found it more difficult to engage the LPA after ductal stenting, and the left PFR may serve as a useful landmark of the pulmonary end of the arterial duct. We suggest that the left PFR be placed first, when possible, because it is easier to access the RPA if the LPA is partially occluded with the PFR. However, we find it useful to obtain a control angiogram in the RPA to define the final diameter after placement of the left PFR, especially if takes more than 5 to 10 minutes to engage. Increased flow to the RPA during that time may dilate the RPA and may result in undersizing and distal device migration. The distal part of the MVP is bare metal; therefore, placing the distal marker at the take-off point of the first upper lobe branch PA will not jail it. Pulling back the deployed device, if not satisfied with the position, should be avoided because it will likely cause proximal migration of the PFR into the MPA. It is usually possible to place a second PFR proximal to the first one if there is distal migration of the PFR leading to unprotected right upper lobe branch. However, the left upper lobe branch frequently has an early take off from the LPA, which limits available space to place a second PFR in LPA if there is acute or late migration. We were able to place a second PFR in the proximal LPA in 1 patient in whom the PFR was significantly displaced into the left lower lobe artery, but it was not possible in another patient in whom there was not enough distance within the proximal LPA for a second device.

For the 950-g patient, it was helpful to use a 0.014-inch soft choice guide wire and 4F angled glide catheter. We also suggest using deflectable microcatheters to ease entry into branch PAs in patients with large PDAs or during second PFR implantation.

As expected, PFR placement was associated with a learning curve. Although the average procedure duration was influenced by concomitant procedures, the overall median procedure/fluoroscopy time decreased from 47 minutes (first 2) to 27 minutes (last 7) over this series. Complications from vascular access were not observed. Having a consistent team was beneficial.

The use of 3-dimensional printed models, computational flow dynamics, and in vitro or in silico testing to simulate patient-specific hemodynamics before and after virtual PFR placement could enable more precise device selection and fenestration sizes, tailored to the needs of individual patients in the future.

Conclusions

Bilateral PFR placement is a feasible and safe alternative strategy for bridging neonates with CHD to surgical repair or palliation, including a delayed Norwood. In our series, transcatheter removal of devices at 2 weeks and surgical removal of the PFRs up to 5.5 months was straightforward, with no PA injury. Timely decision-making and early intervention were associated with good outcomes, even in a very low birth weight infant.

Conflict of Interest Statement

The authors reported no conflicts of interest.

The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.

Supplementary Data

Video 1

Demonstration of the aspects of pulmonary flow restrictor procedure, including creation of the fenestration, angiographic measurement for sizing, device deployment, evaluation with echocardiography, placement of a second device, and retrieval. Video available at: https://www.jtcvs.org/article/S2950-6050(25)00031-2/fulltext.