• Open Access

Transvenous Coil Embolization of Patent Ductus Arteriosus in Small (≤3.0 kg) Dogs


  • This work was done at the Small Animal Clinic, University of Giessen, Germany. Previously presented at 26th ACVIM Congress, June 4–7, 2008, San Antonio, TX.

Corresponding author: E. Henrich, Frankfurter Str. 126, 35392 Giessen, Hessen, Germany; e-mail: estelle.henrich@vetmed.uni-giessen.de.


Background: Surgical and interventional therapy for occlusion of a patent ductus arteriosus (PDA) in small dogs is challenging. Interventional closure of a PDA is rarely described in small dogs.

Hypothesis: Transvenous single-coil occlusion of a PDA in small (≤3.0 kg) dogs is possible and safe.

Animals: Twenty-one client-owned dogs with a left-to-right shunting PDA.

Methods: Prospective clinical study. Inclusion criteria were a left-to-right shunting PDA and a body weight ≤3.0 kg. Dogs with additional congenital cardiac diseases were excluded. Without arterial access, a single detachable coil was implanted by a transvenous approach with a 4 Fr catheter.

Results: Twenty-one dogs were the study population with Chihuahua and Yorkshire Terrier being the commonest breeds (n = 6 and n = 5, respectively). There were 14 female and 7 male dogs. The age range was 1.9–83.5 months (median, 7.7 months), and the body weight was 1.0–2.9 kg (1.87 ± 0.45). By angiography, the minimal ductal diameter measured 1.2–2.4 mm (median, 1.8 mm) and the PDA ampulla diameter was 2.4–5.9 mm (median, 4.6 mm). Coil implantation was successful in all dogs. After detachment of the coil from the delivery cable, repositioning of the pulmonary loop of the coil became necessary in 1 dog. The prevalence of immediate closure was 76%. The prevalence of cumulative closure was 90%.

Conclusion: For an experienced cardiologist, transvenous occlusion of a PDA in small dogs is possible with a 4 Fr catheter and a commercial single detachable coil. Arterial access is not essential. The procedure is safe and successful in experienced hands.


minimal ductal diameter


patent ductus arteriosus

If closure is not undertaken in dogs with patent ductus arteriosus (PDA), case fatality rate is ∼64% in the 1st year after diagnosis.1 Closure is therefore strongly indicated. Small or toy breeds of dogs have the highest risk for developing a PDA.2 Surgical ligation of PDAs has been carried out for >40 years with a high prevalence of success (>95%) in uncomplicated cases. The prevalence of perioperative death rises dramatically (40–100%) if substantial hemorrhage occurs, which is most likely in older dogs and is mostly fatal in small dogs.1,3–5 The use of catheter-guided intravascular occlusion devices such as free or detachable 0.038-in. coils, Gianturco vascular occlusion device, Amplatzer duct occluder, Amplatzer vascular plug, and more recently the Amplatzer canine duct occluder has gained popularity. Most of these systems were implanted using an arterial approach. Transarterial coil occlusion through the femoral artery in infants has been described.6,7 However, the femoral artery in small dogs may be too small for an introducer of at least 5 Fr. Therefore, only single cases of small dogs are described. To reduce the introducer size to 3 Fr, 1 research team used free 0.025-in. coils in 10 small dogs,8 and another used an approach through the larger carotid artery for coil implantation (0.035–0.038 in.) with a 4 or 5 Fr catheter in 2 small dogs.9

Transvenous implantation of PDA occlusion devices is an alternative, and has been used successfully for detachable coils,10,11 for the Amplatzer duct occluder,11–13 and for the Amplatzer vascular plug in dogs.11,14 The specific complications of the transvenous procedure are induction of arrhythmias, difficult retrograde PDA access, and kinking of the delivery sheath.10,12,13 Nevertheless, successful cases in cats15 and in small dogs10,14 have been reported. The results of a retrospective study in small dogs will be published soon.11 The transvenous procedure is preferred in human medicine (particularly in small babies) because of lower prevalence of complications.16

The goal of the present study was to examine the feasibility and success of transvenous single-coil occlusion of a PDA in small (≤3.0 kg) dogs.

Materials and Methods


Between November 1999 and 2007, 241 dogs were diagnosed with an isolated left-to-right PDA and treated by catheter intervention at the University of Giessen (Hessen, Germany). Twenty-one dogs had a body weight ≤3.0 kg and no other congenital heart diseases. These dogs were selected for this prospective clinical study.

Dogs ranged in weight from 1.0 to 2.9 kg (1.87 ± 0.45 kg) and in age from 1.9 to 83.5 months (median, 7.7 months). There were 7 males and 14 females. Breeds represented included Chihuahua (n = 6), Yorkshire Terrier (n = 5), Maltese (n = 2), mixed-breed dogs (n = 2), and 1 each of other breeds.

Clinical signs were present in 5/21 dogs (exercise intolerance [5/5], cough [4/5], and collapse [1/5]). Continuous heart murmur was detected in all dogs (grade IV/VI: n = 5, grade V/VI: n = 16). None of the dogs showed rhythm disturbances under electrocardiography. Under plain radiography, 16/20 dogs showed signs of congestion. Under echocardiography, 10 dogs showed regurgitation of the mitral valve (mild n = 6; moderate n = 3; severe n = 1). At the time of catheter intervention, 14/21 of the dogs were on heart medication (β-methyldigoxin n = 4; furosemide n = 11; benazepril n = 10).

Catheter Intervention

General anesthesia was induced by levomethadonhydrochloride with fenpipramidhydrochloride (0.5 mg/kg) and diazepam (0.5 mg/kg) and maintained with isoflurane (1.7–2.0%). Anesthetic monitoring included pulse oximetry, electrocardiography, measurement of arterial blood pressure by the Doppler method, capnography, and measurement of esophageal temperature. The position of the right femoral vein was analyzed with a Doppler device.a The femoral vein was punctured with a 21-G butterfly needle.b After cutting off the tube, a guidewirec was placed into the vein. A 4 Fr introducer sheathd was inserted and a hemostatic valve connected. A 4 Fr multipurpose cathetere was passed through the right side of the heart into the main pulmonary artery with the help of the guidewire. The PDA was crossed in a retrograde manner with a bigger guidewire.f If this was not possible, ∼1.0 mL/kg of contrast agent was injected into the main pulmonary artery. The levophase of the angiogram was used to localize the PDA, which permitted successful retrograde passage of the wire. After switching to an angiographic catheter,g angiography was undertaken by injecting an iodinated contrast agent (1.0–1.2 mL/kg) into the descending aorta and/or ductal ampulla with an automated injector (Fig 1A). Angiography was recorded at 25 frames/s on 30 mm film or digitally. PDA morphology was classified.17 The minimal ductal diameter (MDD) and the diameter and length of the ampulla were measured to the nearest 0.1 mm by comparison with a radiodense scale.

Figure 1.

 Fluoroscopy of a dog with uncomplicated transvenous patent ductus arteriosus (PDA) coil occlusion. (A) Lateral angiography. A pigtail catheter was placed retrograde through the PDA into the descending aorta. Contrast medium is visible in the descending aorta, the long conical ductal ampulla, and the main pulmonary artery. The ampulla diameter (white arrow) and minimal ductal diameter (black arrow) were measured to be 3.8 and 2.0 mm, respectively. (B) Fluoroscopic image immediately before coil detachment: A detachable 0.038-in. coil with a diameter of 5 mm and 5 loops has been positioned within the ductal ampulla, leaving 1 loop in the pulmonary artery.

A detachable 0.038-in. coilh with a loop diameter of at least twice the MDD was chosen.18 The number of loops was selected to fit into the ampulla. After reverting to the multipurpose catheter, the coil was advanced to the tip of the catheter. The first 2–3 loops of the occlusion coil were extruded into the descending aorta, and the catheter pulled back into the mid-portion of the ductal ampulla. Additional loops were then placed in the ductal ampulla, leaving 0.5–1 loop within the catheter. The catheter was pulled back into the pulmonary artery. The last 0.5–1 loop was placed into the pulmonary artery (Fig 1B) and the coil detached. A pressure bandage was applied immediately before removal of the sheath and persisted for 6 hours to control hemorrhage. Dogs were examined over 3–5 days for complications such as bleeding, or hematuria. Clavulanic-acid potentiated amoxicillin (15.0–20.0 mg/kg IV or PO q12h) was administered for 7 days.

The day after intervention, radiographs were taken (Fig 2) and urinalysis done in addition to clinical and echocardiographic examination. Follow-up clinical and echocardiographic examinations were also made after 3 months and, if the left ventricle remained dilated, again at 12 months after coil placement. In the remaining dogs, a history was taken by a phone call to the owners.

Figure 2.

 Survey radiograph from the dog (Fig 1) 1 day after coil occlusion. The coil is unchanged, with 1 loop anchored in the pulmonary artery.

Complete closure was defined as absence of a residual shunt across the occluded PDA as judged by Doppler color flow mapping.19 Absence of ductal flow in echocardiography within 24 hours of coil placement was defined as “immediate closure,” and absence of ductal flow at a subsequent Doppler examination was defined as “delayed closure.”20 Residual shunts were graded by Doppler color flow in 3 groups as described previously21 (grade 1 = only minimal flow through the PDA at the entrance into the MPA; grade 2 = small jet into the MPA which did not reach the pulmonary valve; grade 3 = broad residual shunt reaching the pulmonary valve).

The left ventricular diastolic diameter was indexed based on body weight as described previously.22

Statistical Analyses

Data were tested for normal distribution by visual inspection and the D'Agostino and Pearson omnibus normality test. Changes in the index of left ventricular diastolic diameter22 before and after coil implantation were compared by a 2-tailed Wilcoxon matched pairs test. Statistical calculations were performed by statistical software.iP≤ .05 was considered significant.


Retrograde catheterization of PDAs was possible without the need for angiography in 18 of 21 dogs. In 3 dogs, assistance of angiography to locate the duct was necessary. By angiography, 20 dogs had a PDA with an aortic ampulla and a pulmonary constriction (long conical type E, n = 16; conical type A, n = 4).17 The remaining dog showed a PDA with 2 constrictions (1 on the aortic and 1 on the pulmonary side, type D). The MDD was measured to be 1.2–2.4 mm (1.77 ± 0.37 mm), the ampulla diameter was 2.4–5.9 mm (4.45 ± 1.06 mm), and the length of the ampulla was 5.0–15.9 mm (9.98 ± 3.03 mm).23

The primary selected coil could be successfully implanted in all but 1 dog. In the latter dog, the selected coil (loop diameter, 5 mm) could not be introduced into the ampulla (ampulla diameter, 4.0 mm; MDD, 1.6 mm). After switching to the next smaller coil (loop diameter, 3 mm), this coil was placed without any problems. Finally, 7 dogs received a coil with a diameter of 3 mm and 5 loops, whereas a coil with 5 mm loop diameter and 5 loops was implanted in the other 14 dogs.

In 1 dog (2.0 kg), the pulmonary end of the coil became trapped after detachment against the wall of the main pulmonary artery in an undesired position (Fig 3A) and was repositioned with a 4 Fr wedge catheterj (Fig 3B). One dog developed bradycardia after coil implantation that required a single injection of atropine (0.01 mg/kg IV).

Figure 3.

 Fluoroscopy in a dog with an undesirable position of the coil loop. (A) Fluoroscopic image after coil detachment. The pulmonary end of the coil (white arrow) is positioned against the wall of the main pulmonary artery in a straight position. (B) Fluoroscopic image after repositioning of the coil loop. The pulmonary loop was repositioned (white arrow) with a 4 Fr wedge catheter (black arrow).

The day after the intervention, none of the dogs showed coil movement on radiography or a macroscopic hematuria. In all dogs, blood flow into the left pulmonary artery was laminar according to color Doppler assessment.

Of 21 dogs, 16 (76%) had immediate closure of the PDA. Four dogs had a residual shunt grade 3 and 1 had a grade-2 residual shunt. Of these 5 dogs, 1 (grade 3) was alive 3 months after the procedure, but was not presented for reexamination. In another dog, the grade-3 shunt stayed the same at his last follow-up (3 months). In the 3 other dogs, there was a delayed closure at 3-month follow-up. This led to a cumulative closure value of 19/21 (90%).

The index of left ventricular diastolic diameter was initially markedly increased to 1.57–2.98 (median, 2.02; n = 21) with 13 dogs above the 95% confidence interval (1.85). The day after occlusion, the index decreased significantly (P < .0001) to 1.28–2.56 (median, 1.54, n = 21), with only 2 dogs above the 95% confidence interval (Fig 4); both of these dogs had delayed ductal closure by 3 months. At this time-point, only 1 of these 2 dogs showed a slightly increased index of left ventricular diastolic diameter, which returned into the 95% confidence interval at 1-year follow-up.

Figure 4.

 Left ventricular end-diastolic diameter index (LVDd-I) at different times. Dogs have been grouped according to patent ductus arteriosus (PDA) closure as immediate closure (IC; PDA closed the day after coil implantation), delayed closure (DC; PDA not closed the day after coil implantation but closed at any further follow-up), and residual shunt (RS; incomplete closure at last follow-up). Before coil implantation (baseline) the LVDd-I is markedly increased (13/21 values above the confidence interval). Within 24 hours after coil occlusion, the index is significantly reduced (P < .0001), with only 2 values above the 95% confidence interval. Both dogs with LVDd-I values greater than the predicted interval at the evaluation at 24 hours had delayed closure and normal values at the final follow-up. Dotted lines represent the 95% confidence interval.

The day after the procedure, mitral regurgitation decreased in 1 dog from moderate to mild. In the remaining 9 dogs, the grading of mitral valve regurgitation stayed the same. The severe insufficiency of the mitral valve in 1 dog before and 1 day after the procedure decreased markedly in severity to moderate and then to mild at 3- and 12-month follow-up, respectively. This dog also had the highest value of left ventricular diastolic diameter index at each time-point in the study (Fig 4).

According to the dog history provided by the owners, none of the 21 dogs had signs at 3- and 12-month follow-up.


The present study showed that transvenous implantation of a single detachable coil was safely achievable in all dogs and that the small residual shunts were without any hemodynamic relevance.

Dogs with a body weight ≤3.0 kg were selected for several reasons. First, percutaneous arterial access in dogs >3.0 kg is reliable.10 Second, a surgical study of small dogs established nearly the same cut-off for body weight (≤3.2 kg) and therefore provided a comparison with the current study.24

Overall, dogs of the current study had a similar body weight (1.0–2.8; 1.87 ± 0.45 kg) to those of the earlier surgical study (0.5–3.2; median, 2.5 kg),24 to the study with 0.025-in. transarterial coils (0.9–1.7; 1.38 ± 0.22 kg),8 and to the study with 0.038-in. transvenous coils in humans (1.6–2.7; median 2.2 kg).25

Percutaneous vascular puncture of either artery or vein itself can be challenging. A thin-walled short needle and a soft-tip guidewire are helpful for both. If the guidewire cannot be advanced into the vessel, venipuncture can be repeated several times. In contrast, a 2nd arterial puncture might be precluded by hematoma, arterial smooth muscle contraction (spasm), or both. The placement of our 4 Fr introducer was not a problem in our dogs because the changeover of dilator and introducer was extremely smooth.

Transvenous PDA catheterization has potential problems, including kinking of the catheter, induction of arrhythmias, and difficult retrograde PDA approach.10,12,13 Kinking of the catheter did not occur in the present study despite the use of small catheters. This could be because we used a braided catheter (metal mesh in the wall) and not a thin-walled sheath without wall stabilization.13 Carrying out retrograde PDA catheterization without previous angiography was relatively easy in most (18/21) of our dogs. This result can be influenced by (i) using the femoral vein and not the jugular vein; (ii) having a catheter with a multipurpose tip configuration; and (iii) the experience of our research group with this procedure.

The minimal diameter of the PDA in the present study (1.2–2.4; 1.77 ± 0.37 mm; angiography) was comparable with values found in studies using small dogs8 (1.72 ± 0.81 mm; angiography) and children25 (2.0–3.6 mm; echocardiography). There is a relationship between body weight and MDD in dogs.23 Most complications that have been described for coil occlusion of the PDA in an unselected population are implantation failure, surgical death, pulmonary or aortic coil embolism, left pulmonary artery stenosis, hemodynamic significant residual shunt, and mechanical hemolysis.20,26–28 Most of these complications are caused by a relatively large PDA and associated factors, such as presence of congestive heart failure, use of multiple coils, and big residual shunts after coil implantation. None of the complications described above were noted in the present study. This could be attributable to 1 or more factors including a small PDA in small dogs, a transvenous approach, a single-coil procedure, detachable coils, relatively stiff coils (0.038 in.), and operator experience.

The UK Central Cardiac Audit Database for infants weighing 2.5 kg and undergoing surgical ligation of the arterial duct documents a 30-day case fatality rate of 8% for this group of patients, which is higher than the value for larger children.25 The case fatality rate in the study with small dogs (1/9)24 was slightly higher than in studies with unselected dogs.1,3,29,30 Possible reasons for the higher case fatality rate may be that even a small amount of bleeding leads to relatively high blood loss in smaller patients.8 However, in human medicine, other factors such as pulmonary diseases in preterm patients must also be taken into account.16

The prevalence of implantation success is equal in transarterial multiple 0.025-in. coils in dogs (80%)8 and in transvenous single 0.038-in. coils in humans (80%),25 and is slightly higher in the transvenous single 0.038-in. coil study in dogs (100%)11 and in the present study (100%). These small differences might be caused by the low number of cases and the many factors mentioned above (primarily the relationship between coil stiffness and PDA size). It was therefore unsurprising that implantation of a 0.038-in. coil was successful in all of our 21 dogs (MDD in each case was <2.5 mm), and in all 8 human patients with a MDD ≤ 2.7 mm; this was in contrast to the 2 implantation failures in the human study with a relatively large PDA (2.8 and 3.6 mm).25 Interestingly, the implantation failure with the use of 0.025-in. coils in dogs occurred with a MDD of 1 mm each; whereas the procedure was successful in 8 other dogs with much bigger MDDs.8 Possible explanations are an extremely stretchable MDD; a selected coil that was too large (5 mm on both occasions) which has a lower stiffness than the same coil in 3-mm diameter; and incorrect measurement of the MDD. Exact measurement of the highest value of the MDD was also challenging in some of our dogs despite the high frame rate (25 frames/s) and the resolution of our catheter laboratory instrument. These problems are caused by a higher heart rate and the extremely small dimensions.

The prevalence of immediate complete closure of successful implantations was increased from 63 to 100% in 1 study of dogs with more than one 0.025-in. coil transarterially.8 The prevalence of immediate closure was 75% in application of the transvenous single 0.038-in. coil in humans25 and was 76% in our dog study. The complete closure value in the follow-up examinations was 100% in the aforementioned human study25 and 90% in our dog study. It therefore appears that if a coil can be securely placed, both methods have a good prevalence of complete closure over the long-term.

Two dogs had a residual shunt 12 months after coil placement. Different methods can be used to quantify the importance of residual shunts. Calculation of the shunt ratio by Doppler methods is possible31 but can be difficult in conscious dogs.32 We therefore used left ventricular size as a marker for significant residual shunts, as carried out in other studies.10,20,28 Both dogs with residual shunting showed a normal left ventricular size the day after intervention, therefore the amount of the residual shunt was judged to be not significant. Two dogs in our group showed delayed normalization of left ventricular size. This may have been caused by late occlusion of the PDA, by a reduced systolic function33,34 or by the presence of significant insufficiency of the mitral valve.

We did not record the anesthetic or fluoroscopic time, but the procedure was sometimes prolonged. The presence of left pulmonary artery stenosis after coil placement was assessed only by Doppler color flow, which is not very sensitive for showing a mild stenosis because of preferential flow into the normal right pulmonary artery.25


a Doppler Ultrasonic Flow detector and nominal 9 MHz in standard 3/8-inch diameter pencil probe, Parks Medical Electronics Incorporated, Aloha, OR

b 0.8 mm, G.21; Microflex, Vygon, Ecouen, France

c Invatec Skipper Guidewire, 0.018″, STF-03-cm, Invatec, Brescia, Italy

d 4-Fr introducer, désilet CE 0459, Vygon

e Angiographic catheter, MP A-1, Super Torque, Cordis Europa N.V., Roden, the Netherlands

f Exchange wire with 25 cm flexible tip, 0.020″ J EX, Schneider (Europe) AG, Bülach, Switzerland

g 4-Fr angiographic catheter, pigtail, mini, super torque, 65 cm, Cordis Europa N.V.

h IMWCE-PDA, MReye Flipper PDA closure Detachable Coil, William Cook Europe, Bjaeverskov, Denmark

i Prism 5.0 for Windows, GraphPad Software Incorporated, La Jolla, CA

j Balloon Wedge Pressure Catheter, 4-Fr, 60 cm, Teleflex Medical GmbH, Kernen, Germany