• Open Access

Long-Term Follow-Up after Transvenous Single Coil Embolization of Patent Ductus Arteriosus in Dogs


  • This study was performed at the Department of Clinical Studies, Small Animal Clinic (Internal Medicine and Surgery), Justus-Liebig-University Giessen, Giessen, Hessen, Germany. The meeting was presented at the 14th Annual Congress of the European Society of Veterinary Internal Medicine, September 9–11, 2004, Barcelona, Spain.

Corresponding author: Nicolai Hildebrandt, Klinikum Veterinärmedizin, Klinik für Kleintiere (Innere Medizin), Justus-Liebig-University Giessen, Frankfurter Street 126, 35392 Giessen, Hessen, Germany; e-mail: nicolai.b.hildebrandt@vetmed.uni-giessen.de.


Background: Long-term follow-up studies after interventional therapy of patent ductus arteriosus (PDA) in dogs are rare.

Hypothesis: Transvenous PDA embolization with a single detachable coil is a highly effective method in patients with an angiographically determined PDA ≤ 4.0 mm.

Animals: Twenty-eight dogs with an angiographic PDA ≤ 4.0 mm were included.

Methods: Prospective follow-up study after PDA coil embolization.

Results: The median follow-up time was 792 days (range, 2–3, 248 days). The rate of complete closure demonstrated by Doppler color flow was 54% at day 3 after intervention and the final cumulative rate was 71%. The rate of complete closure was significantly different between small and moderately sized PDA over the study period (P < .0001) and finally was 100 and 50%, respectively. In 16 dogs with complete closure, no recanalization was found. Disappearance of the continuous heart murmur was found in 89% after 3 days, and this increased to a final cumulative rate of 96%. Indexed left ventricular internal diameter in diastole (LVDd-I) decreased significantly (P < .0001). In the group with moderately sized PDA, a significant difference (P= .0256) was seen in LVDd-I between patients with and without residual shunt after exclusion of patients with persistent severe mitral valve regurgitation.

Conclusion and Clinical Importance: Long-term follow-up after single coil embolization showed complete closure in all small PDA but a residual shunt with mild hemodynamic consequences was present in half of the moderately sized PDA.


left ventricular diastolic diameter index


patent ductus arteriosus

Transcatheter closure has been studied widely during recent years in human medicine and has been considered to be a safe and relatively noninvasive method to close a patent ductus arteriosus (PDA). Several reports describe the technique of catheter intervention in dogs, and report the results immediately after the procedure.1–11,a Interventional closure with multiple vascular occlusion coils by transarterial access is the most commonly used technique.4,5,8,10,12,a To decrease the rate of complications, effort and costs of the procedure, a single coil technique has been evaluated in retrospective studies in human13 and veterinary4 medicine and in a prospective study in humans.14

Long-term follow-up (for at least 1 year) after placement of different catheter-based devices, especially the most commonly used systems such as free coils,15–18 detachable coils,18–20 the Duct-Occlud-Occlusion system,21 Amplatzer Duct Occluder,20,22–24 adjustable buttoned device,25 and Rashkind-occluder-system,20,26 often is described in human medicine. The main purpose of these studies was to assess the long-term efficacy and safety of these devices by evaluating the rate and time of complete occlusion, and to judge potential residual shunt flow by Doppler color flow examination. In veterinary medicine, only a limited number of long-term follow-up studies after interventional therapy2,7,10,27 or surgical ligation27–29 are available.

The aim of the present study was to evaluate the prevalence and the timing of a complete closure and recanalization after embolization of PDA with a minimal diameter ≤ 4.0 mm, as determined by angiography, in dogs, using a transvenous single coil technique. The 2nd goal was to evaluate the hemodynamic impact of residual shunting.

Materials and Methods

Between June 1996 and January 1999, embolization of a PDA with an angiographic minimal diameter ≤4.0 mm was performed in 28 dogs with a single detachable coil and transvenous access, in accordance with a previously published technique.6 Without exclusion, all patients were selected for the prospective follow-up study and the data were analyzed up to June 2005. The dogs were distributed among 14 different breeds. The most common breeds were German Shepherd dogs (n = 5), Miniature Schnauzer (n = 4), and Polski Owczarek Nizinny (n = 4). Eighteen dogs were female and 10 were male. The median body weight was 5.5 kg (range, 1.5–35.0 kg) and the median age was 5.8 months (range, 2.6–65.5).

Clinical signs were found on initial presentation in 12/28 (43%) dogs (New York Heart Association class II [n = 2], class III [n = 6], and class IV [n = 4]). Atrial fibrillation also was identified in 1 dog. In all patients, a continuous heart murmur (≥IV/VI) was found at initial presentation. A full echocardiographic examination, including 2-dimensional, M-mode and Doppler color flow evaluation, was performed in all patients before intervention. The PDA shunts were graded using Doppler color flow examination in 3 groups according to a publication in human medicine30 (grade 1 = only a minimal flow through the PDA at entrance into the main pulmonic artery [MPA]; grade 2 = small jet into the MPA, which does not reach the pulmonic valve; grade 3 = broad shunt reaching the pulmonic valve). Grades 2 or 3 PDA flow was found in 4 and 24 dogs, respectively. The left ventricular internal diameter in diastole index (LVDd-I) was measured from the right parasternal long-axis view in M-mode according to guidelines from human medicine31 and the related index LVDd-I was calculated according to the allometric scaling method.32 Mitral valve regurgitation was found and graded33 in 13 dogs (mild, n = 6; moderate, n = 3; severe, n = 4). Angiography of the descending aorta indicated a minimal median PDA diameter of 2.6 ± 0.7 mm. According to classification in humans,34 20 dogs were classified with PDA type E (elongated conical), 7 with type A (conical), and 1 with type D (multiple constrictions). Each PDA was classified as small, up to a minimal diameter of 2.5 mm, or as moderately sized, between 2.6 and 4.0 mm. Twelve dogs with a small PDA (1.9 ± 0.3 mm) were treated with 5 mm coils and 3 (n = 2) or 5 (n = 10) loops. Sixteen dogs with a moderately sized PDA (3.2 ± 0.3 mm) received an 8 mm coil with 5 loops. The dog with atrial fibrillation (n = 1) was treated successfully by electrical cardioversion.

In all patients, a 1st reexamination was done between 1 and 3 days after intervention. Additional follow-up examinations were proposed to the owners at 3, 12, and 24 months as well as at each 2nd year after PDA embolization. After taking a history, a physical examination with auscultation and a complete echocardiographic examination, including color flow Doppler, were performed by 1 of the 2 authors (N.H., M.S.) with a standard stethoscope and 1 of 2 ultrasound machinesb at each follow-up.

Complete closure was defined as the absence of a residual shunt across the occluded PDA as judged by Doppler color flow mapping.17 Complete closure found in the first examination after intervention was defined as immediate closure, and in one of the following examinations as delayed closure.10,16 Recanalization was diagnosed when Doppler color flow showed a reappearance of the shunt in patients with a previously documented complete closure.15 Auscultatory success was defined as disappearance of the continuous heart murmur after catheter intervention. Analysis of the LVDd-I was performed in each patient with at least 3 months follow-up.

Statistical Analysis

The data were analyzed for normal distribution by visual inspection and the omnibus normality test of D'Agostino and Pearson. A Kaplan-Meier product-limit analysis technique was used to estimate the rate of complete closure and auscultatory success, respectively. The log rank test was used to compare the results between patients with small and moderately sized PDA. The difference in LVDd-I between the initial and final examinations was compared by a two-sided paired Student's t-test. In the dogs with moderately sized PDA and without severe mitral valve regurgitation an unpaired t-test was used to compare LVDd-I values between patients with and without residual shunt at the final examination. All statistical calculations and illustrations were performed using a statistical software package.c A P-value of ≤.05 was considered significant.


The median number of reexaminations per patient was 3 (range, 1–8), and median follow-up time was 792.5 days (range, 2–3,248 days). In the symptomatic dogs, resolution of clinical signs occurred in 11/12 after embolization of the PDA. In the remaining dog, dyspnea persisted despite complete occlusion of the shunt because of severe mitral valve regurgitation, and this led to death 306 days after embolization. There were 3 cases of noncardiac death during the study period.

Complete Closure


Immediate closure was seen in 15/28 dogs (54%). During the follow-up period, 5 delayed closures were identified after variable time frames (105, 111, 300, 422, 1,114 days), leading to a cumulative rate of complete closure in 20/28 patients (71%). The median time to complete closure was 2 days (range, 1–1,114 days; n = 20).

Small Versus Moderately Sized PDA

Immediate closure was found in 11/12 dogs (91%) with a small PDA and in 4/16 dogs (25%) with moderately sized PDA. Owing to 1 and 4 delayed closures, the cumulative rate of complete closure at the end of the study increased to 100 and 50%, respectively. The median time to complete closure for patients with a small PDA (n = 12) was 2 days (range, 1–105 days) compared with 59 days (range, 1–1114 days) for patients with moderately sized PDA (n = 8). The course of complete closure was significantly different between the groups (P < .0001; Fig 1). The residual shunt in the 8 dogs with moderately sized PDA was classified as grade 1 in 2 dogs, grade 2 in 4 dogs, and grade 3 in 2 dogs at their final examination.

Figure 1.

 Kaplan-Meier product-limit analysis of complete closure judged by Doppler color flow, comparing patients with small and moderately sized PDA.


In 16 dogs, a median of 2 (range, 1–7) examinations were performed after documentation of complete closure (median time, 849.5 days; range, 103–3,137 days) and did not show any recanalization.

Within the follow-up period, 2 dogs with moderately sized PDA showed an increase of Doppler color flow shunt grade (grades 1–2 within 665 days and grades 1–3 within 1,018 days). In both of these dogs, no later follow-up information was available.

Auscultatory Success


25/28 (89%) dogs showed disappearance of the continuous heart murmur 1–3 days after coil implantation. Over time, the cumulative rate of auscultatory success increased to 27/28 (96%).

Small Versus Moderately Sized PDA

Auscultatory success was found in 12/12 (100%) dogs with a small PDA 1–3 days after the intervention. In the group with moderately sized PDA, auscultatory success increased from 13/16 (81%) to 15/16 (94%) during the follow-up period. The time to auscultatory success was a median of 1.5 days (range, 1–3 days) in the small PDA group (n = 12) and a median of 3 days (range, 1–100 days) in the moderately sized PDA group (n = 15). The course of auscultatory success was significantly different between dogs with small and moderately sized PDA (P= .0002; Fig 2).

Figure 2.

 Kaplan-Meier product-limit analysis of auscultatory success, comparing patients with small and moderately sized PDA.

Reappearance of Continuous Murmur

In 25 dogs, at least 1 reexamination was performed after documentation of auscultatory success. In a median of 2 (range, 1–7) examinations during a median period of 791 days (range, 87–3,247 days) reappearance of the continuous heart murmur was found in 4 dogs with moderately sized PDA (at 90, 170, 372, 999 days). Two of these dogs were presented for additional reexamination and in both the murmur again had disappeared.

Of the 8 dogs with residual shunting demonstrated by Doppler color flow, 3 (38%) had continuous heart murmurs; the shunt grade was grade 2 (n = 2) or grade 3 (n = 1). In contrast, the 5 patients with inaudible residual shunt had grade 1 (n = 2), grade 2 (n = 2), or grade 3 (n = 1) shunts.


One dog with a small PDA had complete PDA closure and an LVDd-I of 1.663 (within the 95% prediction interval) the day after the intervention, but, because of an absence of follow-up, the patient was excluded from further analyses. In the remaining 27 dogs, the LVDd-I decreased significantly during the follow-up period (P < .0001) from 2.105 ± 0.418 to 1.693 ± 0.446 (Fig 3). All 11 dogs with small PDA that were analyzed showed complete closure at the end of the follow-up period (median, 718; range, 105–2,362 days). The LVDd-I was 1.943 ± 0.395 before intervention (6/11 over the upper 95% prediction interval) and decreased significantly (P= .0010) to 1.527 ± 0.171. At the final examination, none of the 11 dogs showed LVDd-I above the upper 95% prediction interval. Among the 11 dogs, 1 case of mild and 1 of moderate mitral valve regurgitation were found.

Figure 3.

 Index of the left ventricular internal diameter in diastole before intervention and at the final examination in 27 dogs. 11/11 dogs with a small PDA showed finally a complete closure with an LVDd-I within the 95% prediction interval. 8/16 dogs with a moderately sized PDA had finally a complete closure; the LVDd-I of all 5 dogs without severe mitral valve regurgitation were within the 95% prediction interval. 8/16 dogs with a moderately sized PDA had a residual shunt and maximal moderate mitral valve regurgitation at the last follow-up; the LVDd-I was increased in 1 dog. Dotted lines, upper and lower limits of the 95% prediction interval.

Of the 16 dogs with a moderately sized PDA, 8 developed a complete occlusion of the shunt over time. In these dogs, the LVDd-I was 2.256 ± 0.475 before intervention, and decreased significantly (P= .0365) to the last follow-up examination (median, 1199; range, 110–3,248 days) with a value of 1.899 ± 0.765. Three of the 8 dogs had a left ventricular diameter above the upper 95% prediction interval. All 3 showed persistent severe mitral valve regurgitation, in contrast to the other 5 dogs with mild (n = 2) or no (n = 3) regurgitation at their last examination.

In 8 of 16 dogs with a moderately sized PDA, incomplete shunt occlusion was documented at the last follow-up examination (median, 891; range, 90–1,098 days). The LVDd-I decreased significantly (P= .0066) from 2.178 ± 0.356 before intervention to 1.716 ± 0.135 at the end of the follow-up period. One of the 8 dogs had a left ventricular diameter above the upper 95% prediction interval in the absence of mitral valve regurgitation. Of the other 7 dogs, 1 showed mild mitral valve regurgitation.

To evaluate the impact of residual shunting at the final examination in the group with moderately sized PDA, we excluded the 3 patients with severe mitral valve insufficiency. In the remaining 5 dogs with complete closure, the LVDd-I (1.521 ± 0.129) was significantly lower (P= .0256) than in the 8 dogs with residual shunting (1.716 ± 0.135).


Coil embolization of PDA is performed in human15,16,35–37 as well as veterinary medicine,5,10,38–40,a most of the time using a multiple coil technique to achieve immediate complete closure. In contrast to the application of a single coil, the risk of coil embolization into the aorta and other systemic arteries10,41,42 or the pulmonary artery10,42–44 increases when several coils are implanted. Other potential risks of this technique are mechanical hemolysis10,45,46 or stenosis of the left pulmonary artery.18,47 As performed previously in human13,14 and veterinary medicine,4 in the present study application of a single coil was used to decrease these risks. The single coil technique leads to a lower rate of immediate closure in humans13,14 and in dogs,4,6 but long-term complete closure rates are good in humans13,14 and have not been defined in veterinary medicine.

Doppler examination, especially Doppler color flow mapping, was used in the current study to demonstrate the presence or absence of residual ductal flow. This diagnostic technique shows the highest sensitivity in human patients,26 and in particular it is more sensitive than angiographic detection of flow in the PDA.15

The residual shunt flow through the PDA was classified, as described by Swensson et al,30 into 3 grades, because at the beginning of this prospective study, a scale adapted for veterinary medicine was not available. In the meanwhile, a similar system of echocardiographic shunt quantification in dogs has been published.48

After long-term follow-up, a complete closure was found in 71% of our dogs. This rate is lower than that reported in human patients treated with single coil technique (83%,14 94%13). This difference could be explained by relatively large PDA (2.6 ± 0.7 mm) in our dogs compared with the studies of humans (1.8 ± 0.7 mm,14 2.2 ± 0.6 mm13). In both studies of human patients, a negative influence of PDA size on complete closure rate leading to different rates of closure between small and moderately sized PDA (<2.0 mm, 92%; 2.0–4.0 mm, 64%)14 was seen. We found similar differences in complete closure between small (≤2.5 mm, 100%) and moderately sized PDA (2.5–4.0 mm, 50%) in the present study. At the time of the study, only 3 coil diameters were available (3, 5, 8 mm). The use of an additional size (6.5 mm) shows improved short-term results, but long-term data are pending.1

Multiple coil implantation leads to a high rate of long-term complete PDA closure in human medicine.19 In 2 studies in dogs, this rate was lower (46%27 and 61%10), as found in the present single coil study (71%). The authors of the 1st study of dogs27 identified an inadequate number or size of the coils used to be a potential cause for the low closure rate. On the other hand, an accurate comparison is difficult, because specific information on the minimal PDA diameter, PDA shape, and the size of the coils used was not provided. The lower rate of complete closure in the 2nd study of dogs10 may have been caused by the inclusion of patients with large PDA (>4.0 mm), leading to a relatively wide mean PDA diameter (3.1 ± 1.4 mm) when compared with our study (2.6 ± 0.7 mm).

A delayed closure was seen in the present study predominantly in the first 3–12 months after catheter intervention. This coincides with the findings in follow-up studies of human patients15,16,19 as well as in other veterinary10,27 studies. Thrombosis induced by the polyester fibers connected to the coil probably is responsible for this development, after initial mechanical obstruction by the coil material. This is an acute mechanism, and explains the changes that occur early after coil implantation. In the dogs of the present study, the documented occlusions did not show a uniform course. Most of the time, a step-by-step reduction in the shunt was seen at different time points. This indicates an ongoing thromboembolic process.

No recanalization was seen in the current study, in contrast to other studies of coil implantation in dogs10,27 as well as in human patients.15,16,49 Possible causes are the inclusion of only small and moderately sized PDA and the dominant PDA morphology, which showed a long ampulla and pulmonary narrowing (Type E) in 71% of the dogs. In human medicine, a relationship was found between recanalization and ductus morphology.49 In particular, the type B (window-type) without ampulla had a higher risk of shunt development after complete occlusion. In contrast, type E, with its long ampulla, offered good conditions for development of thrombosis at the position of the implanted coil material, and for subsequent endothelization. Later recanalization is unlikely.

Although no recanalization was found in the present study, an increase in residual shunt flow was seen in some cases. In dogs, as in humans, many changes occur after coil implantation, induced by thrombosis, thrombolysis, endothelization, and fibrosis, which could result in variable outcomes. Most of the changes occurred in the first 6 months after intervention. A group of researchers in human medicine has postulated that there is a critical flow limit above which the shunt increases because the primarily induced thrombotic material is washed away.16 Below this limit the shunt should close spontaneously. Evidence from the 2 dogs with an increase in shunt flow, which were graded in their 1st Doppler color flow examination as grade 1, does not support this theory. The thesis that a small residual flow generally is a high velocity flow, which can wash thrombotic material away and could increase the residual PDA flow, is more likely in these dogs.

As an alternative to coil embolization, occluder systems are used in humans and in dogs. The Amplatz Duct Occluder showed a complete occlusion rate of 100% and no signs of recanalization in long-term follow-up studies performed in human patients.20,50 In dogs, 95% had no residual ductal flow 1 year after the implantation of an Amplatz Duct Occluder.7 Most of the delayed occlusions were seen in the first 3 months, and at the time of reexamination no recanalization was found in these dogs.7 The main limiting factor is the high price of the system in contrast to the detachable coils used currently.5 With the Amplatz Canine Duct Occluder a complete occlusion was confirmed in 94% of the canine patients.2 Long-term follow-up over 12 months did not show a delayed occlusion or recanalization. This excellent success is combined with higher cost, a large introducer, and strictly transarterial access.

A comparison between the results of the present study and the results of long-term follow-up examinations after surgical therapy of PDA is very difficult, because data on the size and shape of the PDA and shunt quantification in the case of a residual flow are rare. The rate of complete occlusion in dogs after surgery ranges between 5527 and 95%.51 Possible reasons for these differences are, first, the different surgical techniques used, and second, the influence of the size and shape of the ductus. Obvious differences have been found among different surgical techniques, especially between the standard dissection technique (79% complete occlusion rate52) and the Jackson and Henderson dissection procedure (47% complete occlusion rate52). Delayed occlusion seems to be rare (8%) after surgery,27 but if it happens, as a rule it occurs in the 1st month after the procedure.c The most important difference between coils and other catheter-based devices and surgery is the thrombogenic activity of the implanted material, which is not present in the case of surgically treated patients. In dogs, the frequency of PDA recanalization after surgical treatment varies among studies, being between 1 and 2%53,54 or up to 19%.27

Clinical follow-up examinations by auscultation were performed over a long period of time. This is the only noninvasive diagnostic method to determine success rate after surgical occlusion of PDA. As seen in this study, the sensitivity of this method is lower than that of Doppler color flow examination. At the end of the present study, in only 3 of the 8 patients (38%) with a residual ductal flow on Doppler color examination could continuous heart murmurs be found. This correlates with the results of other studies in dogs, in which fewer than 10 or 16%27 of the dogs with documented residual flow had continuous heart murmurs. In 2 studies of human patients, a typical continuous heart murmur was not found in any patient after PDA intervention, although in 917 and 45%21 continuous flow could be documented by color Doppler. Possible explanations for the low sensitivity of the auscultation technique are the grade and velocity of the residual shunt. In the present study, differences between ultrasound and auscultation were greatest in patients with a grade 1 shunt, and only occurred in rare cases with higher residual flow. In a study of humans, patients with audible PDA had a flow velocity >4.0 m/s and patients without continuous murmur had a shunt velocity between 2.0 and 3.0 m/s.

Echocardiographic measurement of the left ventricular diastolic diameter or its index has been used to document the acute volume reduction after surgical and interventional PDA occlusion with different catheter-based devices in dogs.4,10,11,27 Only 1 study described follow-up over >12 months and showed a further reduction of left ventricular diameter with most patients being within the normal range at the end of the study.10 The same circumstances were found in all subgroups in the present study. In the study of dogs described previously,10 no differences in LVDd-I were seen between dogs with and without residual ductal shunting on long-term follow-up. In contrast, the present study showed a small but significant (P= .0266) difference in the LVIDd-I in the group of dogs with moderately sized PDA with and without residual ductal flow at the final examination, without consideration of shunt grade. Patients with severe mitral valve insufficiency were excluded from this analysis because a hemodynamic influence of this condition on LVDd-I in dogs after PDA occlusion has been postulated previously.27 However, reduced systolic function after PDA closure also should be considered as a possible reason for increased LVDd-I.

The main limitation of the present study was the loss of patients from follow-up examinations, especially in the case of a completely closed PDA with a normal-sized left ventricle. Despite the proposed reexaminations, the willingness of the owners to drive a long distance for the examination of a healthy dog was relatively low.

In conclusion, long-term follow-up examinations show that the application of a single detachable coil in dogs with a small PDA is superior to other devices and to surgical ligation, owing to the high success rate and the low costs. In the case of a moderately sized PDA, the proportion of patients with a residual shunt after coil implantation is high. Mild hemodynamic alterations are seen in the dogs with residual PDA flow. Coil modifications (length, shape, or both) or embolization with other catheter-based devices seem to be necessary for these patients to increase the rate of complete occlusions.


a Miller MW, Meurs KM, Gordon SG, Spangler EA. Transarterial ductal occlusion using Gianturco vascular occlusion coils: 43 cases (1994–1998). J Vet Intern Med 1999; 13:247 (abstract)

b Toshiba SHA 140 and Toshiba Powervision 7000, Toshiba Deutschland GmbH, Neuss, Germany

c GrapPad Prism Version 4.0 and 5.0, GrapPad Software Inc, San Diego, CA


Funding: No Support.