• Open Access

Minimally Invasive Per-Catheter Patent Ductus Arteriosus Occlusion in Dogs Using a Prototype Duct Occluder



This article is corrected by:

  1. Errata: Erratum Volume 22, Issue 3, xxviii, Article first published online: May 2008

  • Presented in part as an abstract at the 26th Annual American College of Veterinary Internal Medicine Forum, Louisville, KY, June 1, 2006.

Corresponding author: Anthony H. Tobias, BVSc, PhD, DACVIM (Cardiology), Veterinary Clinical Sciences Department, College of Veterinary Medicine, University of Minnesota, 1365 Gortner Avenue, Saint Paul, MN 55108; e-mail: tobia004@umn.edu.


Background: Per-catheter patent ductus arteriosus (PDA) occlusion in dogs with devices intended for humans is associated with technical difficulties, high rates of procedure abandonment, device migration, and residual ductal flow.

Hypothesis: Use of a custom-made canine duct occluder (Amplatz Canine Duct Occluder, ACDO) would be feasible in dogs of varying weights and somatotypes and effective in occluding a wide range of PDA shapes and sizes.

Animals: Eighteen client-owned dogs of various breeds with PDA. Weights ranged from 3.8 to 32.3 kg (median, 17.8 kg), and angiographic minimal ductal diameters ranged from 1.1 to 6.9 mm (median, 3.7 mm). Ductal morphologies included types IIA, IIB, and III.

Methods: Per-catheter PDA occlusion with the ACDO was performed in all dogs. Persistent or recurrent ductal flow was assessed at the end of the procedure by angiography and at 1 day, 3 months, and ≥12 months after the procedure by echocardiography.

Results: Successful ACDO placement was achieved in all 18 dogs. One dog required a 2nd procedure with a larger ACDO after the 1st device migrated to the pulmonary vasculature. Complete occlusion was confirmed in 17 of 18 dogs during the procedure, as well as at 1 day and 3 months after the procedure, and in 12 of 13 dogs evaluated at ≥12 months after the procedure.

Conclusions and Clinical Importance: Per-catheter PDA occlusion in dogs with the ACDO is feasible and effective in dogs of a wide range of weights and somatotypes and with PDAs of varying shapes and sizes.

Patent ductus arteriosus (PDA) is a common congenital cardiovascular abnormality in dogs that, if left uncorrected, typically leads to complications that can include mitral valve regurgitation, left-sided cardiac chamber dilation, arrhythmias, left-sided congestive heart failure (CHF), and death. Surgical ligation via thoracotomy is an established method for PDA closure in dogs. More recently, minimally invasive per-catheter techniques for PDA occlusion in dogs using a variety of devices intended for cardiovascular interventions in humans have been investigated.a,b,1–12 Of these devices, embolization coils have been studied most extensively. However, the veterinary experience with embolization coils has been associated with technical difficulties, procedure abandonment, and a relatively high incidence of device migration and residual ductal flow.c,3,5,10,13 Self-expanding devices such as the Amplatzer Duct Occluder and the Amplatzer Vascular Plug, which are designed for PDA or peripheral vascular occlusion in human patients, respectively, have also been used for PDA occlusion in dogs.d,6–8,11,12 However, the PDA in humans typically differs from that in dogs with respect to shape and minimal ductal diameter (MDD).4,10 Consequently, we, in collaboration with a biomedical device company,e designed and developed a device, the Amplatz Canine Duct Occluder (ACDO), specifically for the canine ductus, as well as a per-catheter procedure for ACDO deployment. The objectives of this study were to (1) evaluate the deployment procedure in dogs of various weights and somatotypes and (2) investigate the efficacy of the prototype ACDO in occluding PDAs in dogs over a wide range of ductal shapes and sizes.

Materials and Methods


Between March 2005 and September 2006, 18 dogs with left-to-right shunting PDA at the University of Minnesota Veterinary Medical Center (n=15), North Carolina State University Veterinary Teaching Hospital (n=2), and Purdue University Veterinary Teaching Hospital (n=1) were enrolled in the study. To facilitate methodological consistency among cases, the authors either performed or proctored all evaluations and procedures at the 3 institutions. The study was approved by the Institutional Animal Care and Use Committee of the University of Minnesota.

Age and weight ranged from 5 to 104 months (median, 13 months) and from 3.8 to 32.3 kg (median, 17.8 kg), respectively, and 12 of the dogs were female. Breeds represented were 2 each of Bichon Frise, English Springer Spaniel, and mixed; and 1 each of Pembroke Corgi, Coton De Tulear, Doberman Pinscher, German Shepherd, Golden Retriever, Greyhound, Irish Setter, Labrador Retriever, Miniature Poodle, Miniature Schnauzer, Newfoundland, and Shetland Sheepdog.

Sixteen dogs were asymptomatic and were evaluated after murmurs had been detected either incidentally (n=15) or, in the case of 1 dog, after a left basilar continuous murmur was ausculted 6 months following PDA ligation with absorbable suture material. One dog had a history of coughing and exercise intolerance, and another had a history of exercise intolerance alone.

Left-to-right shunting PDA was confirmed in all cases by physical examination, thoracic radiography, ECG, and transthoracic echocardiography. Radiographs in the 2 symptomatic dogs disclosed the presence of left-sided CHF. ECG revealed atrial fibrillation with a slow ventricular response rate in 1 asymptomatic dog and atrial fibrillation with a rapid ventricular response rate in one of the symptomatic dogs. The rapid ventricular response rate was controlled with digoxin, and CHF was treated to resolution before the PDA occlusion procedures.

Echocardiographic Ductal Shapes and Sizes

Ductal shapes were assessed by transthoracic echocardiography from both right parasternal short axis and left cranial parasternal views. In all cases, the MDD was situated at the point of ductal attachment to the main pulmonary artery, the pulmonic ostium. Echocardiographic MDD was measured via 2-dimensional images recorded from the left cranial parasternal view (Fig 1A). Between 5 and 17 measurements were averaged for each dog, and echocardiographic MDD ranged from 1.6 to 6.4 mm (median, 2.7 mm). In 1 dog, a ridge of redundant, mobile ductal tissue extended 5.7 mm beyond the pulmonic ostium of the PDA into the main pulmonary artery (Fig 1B). The PDA in one of the dogs with CHF was the most tube-like, with little narrowing of the ductus through the ampulla and MDD (Fig 1C).

Figure 1.

 Two-dimensional echocardiograms recorded from the left cranial location of 3 study dogs with patent ductus arteriosus. (A) Typical ductus arteriosus (DA) with a wide ampulla and narrow minimal ductal diameter at the pulmonic ostium (arrowhead). (B) Atypical DA with ridge of excessive ductal tissue (arrowheads) extending beyond the minimal ductal diameter. (C) Tube-like DA with minimal narrowing of the ductal width at the pulmonic ostium. Ao, aorta; MPA, main pulmonary artery.

Angiographic Ductal Shapes and Sizes

With each dog under general anesthesia and in right lateral recumbency, the right femoral artery was surgically isolated and cannulated with a vascular access sheath.f Under fluoroscopic guidance, an angiographic catheterg,h was advanced into the aorta immediately cranial to the junction between the aorta and PDA (n=17) or into the ductal ampulla (n=1). Angiography was performed by injecting approximately 1 mL/kg contrast solution (iothalamate meglumine)i at 20 mL/s with a pressure injector.j

Using the system described by Miller et al14 to classify PDAs in dogs based on their angiographic morphology, type IIA (n=14), IIB (n=3), and III (n=1) ducti were identified. The dog with the type III PDA also demonstrated the most tube-like ductus during echocardiography, and in all dogs, the angiographic MDD was situated at the pulmonic ostium of the PDA. Angiographic MDD, which was acquired by comparing the width of the initial portion of the jet of contrast across the MDD with intra-aortic reference markers,g,k ranged from 1.1 to 6.9 mm (median, 3.7 mm). Ductal length and width were measured as described previously15 and ranged from 8.7 to 46.3 mm (median, 19.8 mm) and from 5.2 to 22.6 mm (median, 9.6 mm), respectively.

Description of Device and Deployment Procedure

The ACDO and the procedure for its deployment have been described previously.l,16 Briefly, the ACDO is a detachable multilayer nitinol mesh device with a short waist that separates a flat distal disc from a cupped proximal disc. The device, particularly its cupped proximal disc, is specifically designed to conform to ductal morphology in dogs (Fig 2). For this study, prototype devices were constructed with waist diameters of 3 mm and 4 to 14 mm in 2 mm increments.

Figure 2.

 A prototype Amplatz Canine Duct Occluder attached to a delivery cable. The detachable multilayer nitinol mesh device is constructed with a short waist that separates a flat distal disc from a cupped proximal disc.

After measuring PDA dimensions by angiography, a guiding catheterm was placed, via the femoral artery and aorta, across the PDA and into the main pulmonary artery. The ACDO was advanced via the guiding catheter using a delivery cable. The distal disc was deployed in the main pulmonary artery and apposed to the pulmonic ostium of the PDA. The device waist and proximal disc were then deployed across the pulmonic ostium and within the distal ampulla, respectively. Proper ACDO positioning and stability were evaluated by back-and-forth maneuvering of the delivery catheter or cable as well as a small manual injection of contrast solution made through the guiding catheter. The delivery cable was then detached and removed along with the guiding catheter. Aortic angiography was performed at the conclusion of the procedure, and the catheterized femoral artery was either repaired with nonabsorbable suture materialn placed in a simple continuous pattern (n=15) or ligated (n=3).


Physical examinations, thoracic radiographs, ECG, and transthoracic echocardiography were performed 1 day and 3 months after the procedure in all dogs. At the time of manuscript submission, the same examinations had been completed at ≥12 months after the procedure in 13 of the 18 study dogs. Echocardiographic examinations included color flow and spectral Doppler evaluations to assess for persistent or recurrent ductal flow as well as for any flow disturbances within the main or branch pulmonary arteries caused by the ACDO.

Definitions of Terminology

Oversize factor was defined as the ratio of ACDO waist diameter to MDD as measured by angiography. Procedural success was evaluated by thoracic radiography and echocardiography 1 day following ACDO deployment and was defined as proper expansion and location of the ACDO across the MDD. Ductal patency was assessed at the end of the procedure by angiography and during all follow-up evaluations by color flow Doppler echocardiography. Ductal occlusion was considered complete if ductal patency was absent. Ductal occlusion was considered immediate if complete occlusion was documented by angiography during the occlusion procedure. Recurrent ductal patency was defined as ductal flow in a dog in which ductal occlusion had previously been complete.

Statistical Analyses

Ductal length and width before and after PDA occlusion were compared by the Wilcoxon signed-rank test. Measurements of MDD made by echocardiography and angiography were compared by Bland–Altman analysis. Analyses were performed using statistical software,o with P≤ .05 defined as significant.


Procedure Outcome

The initial procedure was successful in 17 of 18 dogs. A single minor complication occurred during the procedure in one of these dogs whereby an imprecise measurement of the angiographic MDD led to selection of an ACDO with an oversize factor of 3.1. As a consequence, the ACDO failed to assume its uncompressed shape during deployment because the device waist and cupped proximal disc did not fully expand within the pulmonic ostium and ductal ampulla, respectively. While still attached to the delivery cable and with little resistance, the partially compressed ACDO was advanced through the pulmonic ostium of the PDA into the main pulmonary artery, where it was recaptured into the guiding catheter and removed. After remeasuring the MDD, procedural success was achieved using a smaller ACDO with an oversize factor of 2.6.

The initial procedure was not successful in 1 dog because the ACDO migrated to the left caudal pulmonary artery during anesthetic recovery. This major complication occurred despite the fact that the device had assumed its fully uncompressed shape during deployment and the final angiogram had shown immediate ductal occlusion. This dog was the third in the series, and the selected ACDO waist diameter and angiographic MDD were 8 and 6.6 mm, respectively, giving an oversize factor of 1.2. No attempt was made to retrieve the ACDO, no clinical signs ascribable to the presence of the device in the left caudal pulmonary artery were observed, and a 2nd procedure was performed 5 months later. Angiographic MDD was then measured at 6.8 mm, and procedural success was achieved using a 12 mm waist diameter ACDO that conferred an oversize factor of 1.8. Thus, procedural success was ultimately achieved in all 18 dogs using devices with waist diameters of 3 (n=1), 4 (n=6), 6 (n=2), 8 (n=4), 10 (n=2), and 12 mm (n=3) and oversize factors ranging from 1.5 to 2.7 (median, 2.0).

A small amount of persistent ductal flow was present in 6 of 18 dogs during manual injection of contrast solution before ACDO detachment from the delivery cable. However, final angiograms performed following device detachment disclosed immediate ductal occlusion in 17 of the 18 procedural successes and a small amount of persistent ductal flow in 1. Ductal length and width after occlusion ranged from 10.5 to 50.0 mm (median, 25.5 mm) and from 5.4 to 25.4 mm (median, 13.7 mm), respectively. Both of these dimensions were significantly increased (P <.01) when compared with their respective pre-occlusion measurements. Fluoroscopy time ranged from 5.3 to 17.0 minutes (median, 9.4 minutes).


Echocardiography performed 1 day and 3 months after the procedure demonstrated complete ductal occlusion in 17 of 18 dogs, including the 1 dog in which a small amount of persistent ductal flow had been identified by angiography. Of the 13 dogs evaluated at ≥12 months after the procedure, 12 had complete ductal occlusion, while ductal flow was identified in 1. In all dogs, Doppler echocardiography disclosed that the ACDO did not disturb normal laminar blood flow in the main or branch pulmonary arteries at any point during the study.

The dog with ductal flow at ≥12 months after the procedure had demonstrated immediate ductal occlusion during angiography, but Doppler echocardiography disclosed trivial recurrent ductal flow 1 day after the procedure. Ductal flow progressed to a moderate degree at 3 months and was unchanged at ≥12 months. Recurrent ductal patency appeared to be associated with flow through rather than around the ACDO and was present in spite of radiographic and echocardiographic evidence of proper device expansion and location.

Complete ductal occlusion was achieved in both dogs that presented with left-sided CHF. Furosemide therapy was discontinued in 1 dog 1 day after the occlusion procedure, and CHF had not recurred by ≥12 months. However, atrial fibrillation with a rapid ventricular response rate was identified at the 3-month evaluation, and rate control was achieved with digoxin. In the 2nd dog, all CHF medications were discontinued within 75 days of ductal occlusion. While there was no evidence of recurrent CHF at the 3-month evaluation, this patient's previously noted atrial fibrillation again exhibited a rapid ventricular response rate, and digoxin therapy was resumed.

Among the 15 dogs in which femoral artery repair was performed, pulsations distal to the surgical site were present in 10 dogs 1 day after the procedure and in all 15 dogs at the 3-month evaluations. In the dog in which the ACDO had migrated and a 2nd procedure was performed, the right femoral artery was isolated, cannulated, and repaired on both occasions without complications and with return of pulsations.

Angiographic Versus Echocardiographic Measurement of MDD

Bland–Altman analysis disclosed a small measurement bias (0.3 mm) between echocardiographic and angiographic MDD measurements. However, the limits of agreement were wide (− 1.3–1.9 mm; Fig 3).

Figure 3.

 Bland–Altman analysis showing that echocardiographic measurement of minimal ductal diameter (MDD) cannot substitute for angiographic measurement of MDD for selecting an appropriately sized Amplatz Canine Duct Occluder (ACDO). Whereas the mean difference or bias between the 2 methods is small, the limits of agreement (mean ± 1.96 standard deviations, SD) are unacceptably wide for ACDO selection.


Deployment of the prototype ACDO was straightforward across the wide range of dog weights and somatotypes and PDA shapes and sizes that constituted our study population. Device position and stability could be assessed at several stages during the deployment procedure before device detachment from the delivery cable, and the ACDO was readily recaptured if initial positioning and stability were not satisfactory. Furthermore, the arterial approach avoided right heart catheterization that can be associated with difficult retrograde PDA access, kinking of the delivery sheath within the right ventricle, and arrhythmias.p,4,7,8,17

A minor procedural complication occurred in 1 dog when a device that conferred an oversize factor of 3.1 was selected. The ACDO, which was unable to expand fully when deployed, was easily advanced through the pulmonic ostium into the main pulmonary artery, where it was recaptured. A smaller ACDO was then successfully deployed. This incident demonstrated that an excessively large device could be associated with poor stability across the pulmonic ostium of the PDA. However, this minor complication was readily identified before device detachment and effectively addressed.

The single case of device migration that required a 2nd procedure occurred early in the study, and that dog had the 2nd largest angiographic MDD (6.8 mm) in our study population. More important, the ACDO selected during the initial procedure had an oversize factor of 1.2, whereas the ACDO selected during the subsequent successful procedure had an oversize factor of 1.8. Ultimately, procedural success with the prototype ACDO was achieved in all 18 study dogs with oversize factors that ranged from 1.5 to 2.7 (median, 2.0). In contrast, either incomplete ACDO deployment or device migration occurred with oversize factors of 3.1 and 1.2, respectively. We interpret our data to suggest that ACDOs with waist diameters that are approximately 2 times larger than the angiographic MDD should be selected for this PDA occlusion procedure.

There was no instance of ACDO migration when an oversize factor of >1.5 was applied in our study population. In contrast, embolization coil migration is a common complication during PDA coil occlusion procedures,b,3,5,7 and a recent review involving 125 dogs reported a 22% incidence of coil migration into either the pulmonary or systemic arteries.10

Whereas different per-catheter PDA occlusion techniques have been recommended for different MDDs in dogs,7 the ACDO was successfully deployed across a wide range of angiographic MDDs in our study population. Furthermore, the inclusion of dogs with MDDs ≥5.0 mm (n=5) increased the risk of a poor outcome because coil occlusion is discouraged entirely for MDDs >5.0 mm.7

Different devices and strategies for PDA occlusion have been recommended for different PDA morphologies.14 However, in our study, the ACDO successfully occluded a variety of ductal shapes. Echocardiography and angiography disclosed several ductal morphologies, including types IIA, IIB, and III. The 2 most notable ducti had the 2 largest angiographic MDDs: the case in which device migration occurred during the initial procedure (MDD=6.8 mm) and one of the dogs with CHF (MDD=6.9 mm). Echocardiography demonstrated a ridge of redundant, mobile ductal tissue extending 5.7 mm into the main pulmonary artery in the case in which device migration occurred. Whether or not this tissue contributed to initial device migration is unknown, but complete and immediate ductal occlusion was achieved during the 2nd procedure with a larger ACDO. The dog with CHF had the largest angiographic MDD and a type III PDA. Although device occlusion is more difficult for a type III PDA because of lack of ductal taper and large MDDs,14 complete and immediate ductal occlusion was attained with the ACDO. Despite this success, our experience using the ACDO for type III PDAs is limited, and pending further research, we urge caution with this ductal shape.

The unique features of ACDO design and deployment obviate some of the technical difficulties, device instability issues, and control problems that have plagued coil embolization procedures.c,3,5 Such limitations have contributed to procedure abandonment in 11–27% of coil embolization attemptsb,3,10 and, on occasion, aberrant coil placement.b,10 In addition, the presence of CHF confers a 3-fold risk of coil embolization failure associated with device instability.10 Neither procedure abandonment nor aberrant ACDO placement occurred in any of our study dogs despite inclusion of 2 CHF patients.

Studies with similar or larger numbers of dogs have reported angiographically determined persistent ductal flow at the conclusion of per-catheter PDA occlusion procedures in 41–50% of dogs after the deployment of embolization coilsb,4 and in 35% of dogs after the placement of Amplatzer Duct Occluders intended for PDA occlusion in humans.8 In our study, angiograms performed at the conclusion of the ACDO deployment procedures disclosed persistent flow in only 1 of 18 dogs.

Color flow Doppler echocardiography performed within 1–3 days of coil deployment in dogs with PDA disclosed residual ductal flow in 50–66% of dogs in which one or more embolization coils were deployed.4,10 The same evaluation demonstrated residual flow in 25% of dogs in which Amplatzer Duct Occluders intended for use in humans were placed.8 In contrast, color flow Doppler echocardiography performed 1 day following ACDO placement demonstrated residual ductal flow in 1 of 18 dogs.

The results of longer term color flow Doppler evaluations to assess for residual ductal flow following per-catheter PDA occlusion have also been reported. Residual flow was present in 39–42% of dogs in which one or more embolization coils were deployed at evaluations performed between 3 and 63 months after the procedure.4,10 In 19 dogs that were evaluated 3 months after implantation of Amplatzer Duct Occluders intended for use in humans, 2 had residual ductal flow.8 In the same study, 1 of 9 dogs evaluated 11–25 months after the procedure demonstrated ductal flow. In our study, color flow Doppler echocardiography detected residual flow in 1 of 18 dogs 3 months after ACDO implantation, and of the 13 dogs that have been evaluated at ≥12 months after the procedure, only 1 was found to have echocardiographic evidence of residual ductal flow. Furthermore, there was no change in occlusion status after 3 months, suggesting that when using the ACDO, it is not necessary to evaluate fully occluded patients for recurrent ductal flow beyond this time frame.


aMiller MW, Bonagura JD, Meurs KM, Lehmkuhl LB. Percutaneous catheter occlusion of patent ductus arteriosus. Proceedings of the 13th Annual American College of Veterinary Internal Medicine Forum 1995:308–310

bMiller MW, Meurs KM, Gordon SG, Spangler EA. Transarterial ductal occlusion using Gianturco vascular occlusion coils: 43 cases 1994–1998. J Vet Intern Med 1995;13:247 (abstract)

cMiller MW. Transarterial coil occlusion of patent ductus arteriosus: Outcome in 120 cases. Proceedings of the 20th Annual American College of Veterinary Internal Medicine Forum 2002;102

dHogan DF, Green HW, Sanders RA, Goodwin JA. Use of a peripheral vascular occlusion device for correction of patent ductus arteriosus in dogs. J Vet Intern Med 2006;20:730 (abstract)

eAGA Medical Corporation, Plymouth, MN

fPercutaneous Sheath Introducer Set, Arrow International Inc, Reading, PA

gRoyal Flush II White Vessel Sizing Pigtail Angiographic Catheter, Cook Inc, Bloomington, IN

hRoyal Flush Plus Pediatric Pigtail Angiographic Catheter, Cook Inc

iConray 60, Mallinckrodt Inc, St. Louis, MO

jMark V ProVis Injector, Medrad Inc, Indianola, PA

kThe Graduate Measuring Wire Guide, Cook Inc

lNguyenba TP, Tobias AH. Patent ductus arteriosus occlusion with an investigational Amplatzer® Canine Ductal Occluder. J Vet Intern Med 2006;20:730 (abstract)

mNorthstar Lumax Flex Guiding Catheter, Cook Inc

nProlene (6-0, 75 cm, RB-2 needle), Ethicon Inc, Somerville, NJ

oNCSS 2007, NCSS, Kaysville, UT

pTobias A, Jacob K, Fine D, Carpenter D. Patent ductus arteriosus occlusion with Amplatzer® Duct Occluders. Proceedings of the 20th Annual American College of Veterinary Internal Medicine Forum 2002:100–101


The authors gratefully acknowledge Drs Kurt Amplatz and Xiaoping Gu as well as John Oslund of AGA Medical Corporation for innovative design and manufacture of devices for dogs; Drs Daniel Hogan, Bruce Keene, Teresa DeFrancesco, Luca Ferasin, Christopher Stauthammer, and Michelle France for contributing cases and data; and Kristin Hohnadel and Leslie Hiber for technical assistance. AGA Medical Corporation (Plymouth, MN) funded this study.