• Open Access

Transesophageal Echocardiography Guided Patent Ductus Arteriosus Occlusion with a Duct Occluder


  • This study was done at the department of Cardiology of Clinica Veterinaria Gran Sasso, Milan, Italy.



Angiography and fluoroscopy are the standard methods to guide transcatheter occlusion of patent ductus arteriosus (PDA). The use of iodinated contrast agents and radiation exposure pose risks of animals and staff.


To assess feasibility of transesophageal echocardiography (TEE) for device size selection and procedure monitoring for PDA occlusion with a duct occluder (DO) without the use of angiography.


Eighty client-owned dogs with left-to-right PDA.


Prospective study. Dogs with left-to-right PDA undergoing transcatheter occlusion were included. Procedures were performed without angiography and device size selection was based on TEE measurements. Procedures were monitored with simultaneous TEE and fluoroscopy and both methods were compared. Visualization of the ductus and dimensions obtained by TEE and transthoracic echocardiography (TTE) were compared.


Complete PDA occlusion was achieved in 79/80 cases. TEE was consistently superior to TTE for PDA visualization and the latter showed higher values for ductal dimensions when compared to the former. TEE provided adequate procedure monitoring in 73 cases (91%). Fluoroscopy exposure time (2.77 ± 1.2 minutes (mean, SD)) was lower than previously reported for the same procedure.

Conclusions and Clinical Importance

TEE is a useful and efficient tool for device size selection and can be used for procedure monitoring in most cases. Fluoroscopy exposure time can be reduced and the use of contrast agents can be avoided. However, fluoroscopy is required in a minority of cases when TEE monitoring is not feasible or incomplete and should be available for this procedure.


distal ampulla diameter


duct occluder


minimal ductal diameter


patent ductus arteriosus


transesophageal echocardiography


transthoracic echocardiography

Catheter-based definitive treatment for patent ductus arteriosus (PDA) in dogs has become a frequent alternative to surgical treatment and several devices have been used for this purpose. The use of a commercial duct occluder (Amplatz® Canine Duct Occluder1) (DO) gained popularity in the last few years because of its straightforward technique, high efficacy and low rate of complications.[1-5] Although its use is limited by animal size, ductal morphology and dimensions, transcatheter PDA closure with a DO is currently the preferred method in most cases.[4, 5] This device is available in several sizes and selection is based on minimal ductal (MD) and ampulla diameters.[1, 6] Accurate device size selection is required for procedure success and to avoid potential complications such as device embolization.[2]

Angiography is the standard method for ductus morphology and size assessment as well as for procedure guidance[1-6] but transesophageal (TEE)[3, 5-8] and more recently transthoracic (TTE)[9, 10] echocardiography have also been used for these purposes. Both ultrasound-based methods have been shown to correlate with angiography for ductal dimension measurement[6, 8, 11, 12] with TEE most accurately reflecting angiographic measurements.[6, 12] TEE also provides comprehensive, real-time views of the ductus and adjacent structures and its use for occlusion procedure guidance has been described in humans[13-15] and dogs.[6-8] Ultrasound methods do not expose operators and animals to ionizing radiation nor require the use of iodinated contrast agents which contribute to increased animal and staff safety.

We hypothesized that definitive treatment of PDA with a DO could be performed using TEE for device selection therefore avoiding angiography and for procedure monitoring reducing radiation exposure. The objectives of this study were to describe a large number of cases treated with a DO without angiography and to assess the feasibility of TEE for procedure monitoring.

Materials and Methods

Client-owned dogs weighing 3 kg or more, with a left-to-right shunting PDA, undergoing treatment with a DO were included in this study. Dogs with concomitant congenital heart disease or pulmonary hypertension were excluded.

A written consent form was signed by the owners and complete physical examination, routine blood work, electrocardiogram, blood pressure and thorax radiographs in right lateral and ventrodorsal views were performed in all dogs.

A complete TTE examination was performed in all dogs before the procedure using MyLab 502 and MyLab Class C units3 with 2.5–10 MHz transducers selected according to body weight. The PDA was visualized from the right and left parasternal views as previously described[11] and the measurements were obtained from the view that provided the best image quality and alignment. Two measures were made: the MD and the distal ampulla diameter (AD). The latter was a line 2–4 mm proximal and parallel to the former, which was the expected position of the cupped proximal disc according to its height (Fig 1).[1] Measurements of the MD and AD were obtained at the moment of the largest diameter during the cardiac cycle.

Figure 1.

Transthoracic echocardiography images obtained from a right parasternal short-axis view of a 7 kg Border Collie. (A) Two-dimensional image showing the ductus and measurements of the minimal ductal (MD) and distal ampulla (AD) diameters. (B) Color Doppler image showing left-to-right turbulent flow within the pulmonary artery. Ao, aorta; PA, pulmonary artery; PDA, patent ductus arteriosus.

Dogs were anesthetized and positioned in right lateral recumbency with the left hind limb abducted and extended to expose the left femoral artery.

TEE was performed using an omniplane 3–8 MHz transesophageal probe with a length of 110 cm and a tip diameter of 14 × 10 mm.3 A cranial position, longitudinal view was used to identify the PDA as previously described[8] and adjustments were made to obtain a longitudinal cross-section image of the ductus. MD and AD measurements were obtained as described for TTE (Fig 2). TTE measurements were not disclosed to the team member performing the TEE. A transverse view of the distal ampulla and pulmonic ostium of the ductus was obtained when possible. This view was obtained with retroflexion of the probe until the ductus appeared as close to horizontal as possible (Fig 3A), then the ultrasonic beam angle was increased by 90° approximately, the probe was then advanced until the aortic orifice of the ductus was seen and then slowly withdrawn until the pulmonic ostium was visible (Fig 3B).

Figure 2.

Transesophageal echocardiography images obtained from a cranial position, longitudinal view of a 32 kg Doberman Pinscher. (A) Two-dimensional image showing the ductus and measurements of the minimal ductal (MD) and distal ampulla (AD) diameters. (B) Color Doppler image showing a jet of turbulent flow from the ductus into the pulmonary artery. Ao, aorta; PA, pulmonary artery; PDA, patent ductus arteriosus.

Figure 3.

Transesophageal echocardiographic longitudinal (A) and transverse (B) views of the patent ductus arteriosus. The transverse view shows an elliptical shape of the pulmonic ostium (PO) of the ductus. Ao, aorta; PA, pulmonary artery; PDA, patent ductus arteriosus.

Device size selection was performed following 3 steps: (1) the MD in mm was multiplied by 1.8 to select the closest sized device (device waist diameter); (2) cupped proximal disc to AD ratio was calculated and if it was <1 the next (larger) device was selected; (3) a further increase in the device size was made in cases where the ductus presented a horizontal ellipse shape at the level of the distal ampulla and pulmonic ostium.

Ductus morphology was classified as type I, IIa, IIb or III (adapted from Miller et al).[16]

Meanwhile, access to the left femoral artery was achieved using a modified Seldinger technique and an appropriately sized introducer (Avanti®+4 ), able to accommodate the required guiding catheter (Vista Brite Tip® MPA6), was placed and secured in place with a suture.

Transcatheter occlusion was performed as previously described.[1, 2]

TEE, performed by a trained team member throughout the procedure, and X-ray fluoroscopy were used in combination for procedure monitoring. Two screens, one displaying TEE images and another displaying fluoroscopy, were positioned in front of the team member performing the procedure. TEE and fluoroscopy images were displayed with the same orientation (cranial structures to the left) (Fig 4).

Figure 4.

Fluoroscopy and transesophageal echocardiography side-by-side view of the different phases of the procedure: (A) placement of the guiding catheter across the patent ductus arteriosus and into the main pulmonary artery; in the fluoroscopic image, the catheter is across the ductus and in the pulmonary artery; in the transesophageal image, the catheter tip is seen within the ductus just before being advanced through the pulmonic ostium; (B) deployment of the distal disc in the main pulmonary artery; (C) apposition of the distal disc to the pulmonic ostium of the ductus; (D) deployment of the cupped proximal disc within the distal ampulla; (E) device still attached to the delivery cable assuming its native shape after complete deployment; (F) device release from delivery cable.

Presence of residual flow was assessed by color Doppler imaging and classified as trivial (random color pixels with no discrete jet), mild (discrete jet which did not fill the pulmonary artery), moderate (discrete jet which partially filled the pulmonary artery), and severe (discrete jet which filled the pulmonary artery)[17] before device release from the delivery cable and 5–10 minutes after device release. Presence of residual flow was also assessed by color Doppler imaging with TTE the following day.

TEE ability to monitor the procedure was classified as “impossible” if the image was poor or was impossible to obtain and maintain an adequate view, “incomplete” if adequate view was lost during some phases of the procedure causing interruptions to the monitoring, and “complete” if optimal monitoring was possible through all phases of the procedure (ductus catheterization, distal disc deployment in the pulmonary artery and apposition to the pulmonic ostium, cupped proximal disc deployment, device stability, and residual flow assessment before release, device release, and assessment of residual flow after release). Fluoroscopy exposure time was recorded.

TTE and TEE images were rated according to the image quality as “poor” if identification of the ductus was possible, but image quality or alignment was inadequate for measurement, “sub-optimal” if morphology assessment and measurement were possible but image quality or alignment were not optimal potentially affecting measurements and “optimal” if there was good image quality and alignment.

The final device waist diameter to MD ratio and device cupped proximal disc size to AD ratio were calculated.

For comparison purposes, measurements obtained with both ultrasound methods were analyzed and compared. Additionally, after the first and second steps of the device selection criteria described above, a hypothetical device size was calculated using the TTE measurements for each case and compared to device size selected using TEE measurements.

Statistical Analysis

Results are reported as mean ± SD when normally distributed, or as median and range when not normally distributed. Kolmogorov–Smirnov test was used to assess normality. TEE and TTE measurements were compared using paired samples t-test when normally distributed and Wilcoxon signed-rank test when not normally distributed, and Bland Altman plots. Significance was set at P < .05. Statistical analysis was made with a commercially available software.5


Eighty dogs were included in the study; 60 females (75%) and 20 males (25%). Age ranged from 2 to 99 months (median, 8 months).The most represented breed was German Shepherd (n = 11) followed by Doberman Pinscher (n = 6), Pembroke Welsh Corgi (n = 5), Cavalier King Charles Spaniel (n = 4), 3 each of Australian Shepherd, Dachshund, Newfoundland, Poodle, 2 each of Border Collie, Brittany Spaniel, Cocker Spaniel, Irish Setter, Labrador Retriever, and 1 each of Bavarian Mountain Dog, Belgian Shepherd, Bichon Frisé, Bolognese, Chihuahua, Griffon, Golden Retriever, Italian Hound, Italian Spitz, Jack Russell, Shetland Sheepdog, and Weimaraner. The remaining 20 dogs were mixed breeds. Body weight ranged from 3.2 to 37 kg (median, 12 kg).

MD and AD were 4.11 ± 1.29 mm and 8.22 ±2.33 mm, respectively, with TEE and 4.83 ± 1.45 mm and 9.14 ± 2.77 mm with TTE.

There was a significant statistical difference between measurements obtained by TEE and TTE for both MD (P < .0001) and AD (P = .0003) with higher TTE mean values for both. A Bland–Altman plot revealed a bias of −0.8 and −1.1 mm for MD and AD, respectively, also showing higher values for TTE in comparison to TEE. The limits of agreement were −2.9 to 1.3 and −5.9 to 3.7 for MD and AD, respectively (Fig 5).

Figure 5.

Bland–Altman plot comparing transesophageal (TEE) and transthoracic (TTE) measurements. For minimal ductal diameter (MD) measurement there is a mean bias of −0.8 and 95% limits of agreement of −2.9 to 1.3. For ampulla diameter (AD) there is a mean bias of −1.1 and 95% limits of agreement of −5.9 to 3.7.

Ductal morphology was classified based on TEE as IIA in 47 dogs (59%) and IIB in 31 dogs (39%). No dogs presented type I or III morphology.

Quality of TEE images was classified as optimal in 57 cases, suboptimal in 21 cases and poor in 2 cases. TTE images were classified as optimal in 19 cases, suboptimal in 58 cases and poor in 2 cases. TTE quality rating was not recorded in 1 case. TEE image quality was superior to TTE in 38 cases, was the same in 39 cases and inferior in 2 cases. Both techniques allowed for identification of the ductus in all cases. Measurements and morphology classification was not attempted in cases with poor image quality.

Device selection was based solely on TEE measurements in all but 2 cases in which image quality was poor. As TTE image quality was optimal in these 2 cases, measurements were disclosed and used for device selection. In most cases, the device size calculated based on MD had a cupped proximal disk to AD ratio ≥1 and a ductal elliptical shape was not identified; therefore, no further correction was required. The initial device size was increased in 2 dogs as the cupped proximal disc to AD ratio was <1 and in 5 dogs in which a ductal elliptical shape was identified.

Correct device deployment was achieved in 75/80 cases (94%) with the first selected device size. In 4 cases the first selected device was too small and was exchanged by a larger one. This was because of the instability detected by gentle maneuvering back and forth the delivery cable that resulted in dislodgement of the device (3 cases) and because of the presence of severe residual flow despite correct deployment and stability of the device (1 case). In 1 case, the device failed to assume its native shape and a smaller device was used (Table 1). No device migration was reported immediately after device release or the following day.

Table 1. Summary of the cases requiring device size correction during the procedure.
Dog No.Age (months)SexBreedWeight (kg)MD (mm)AD (mm)MorphologyWaist Ø/MD (initial device)Waist Ø/MD (final device)Initial Device (waist Ø, mm)Final Device (waist Ø, mm)Cupped Proximal Disc/AD (final device)
  1. MD, minimal ductal diameter; AD, distal ampulla diameter; Ø, diameter.

  2. a

    Initial device unstable.

  3. b

    Initial device size increased because of elliptical pulmonic ostium but still unstable.

  4. c

    Initially undetected elliptical pulmonic ostium.

  5. d

    Correctly delivered device but severe residual flow detected before release.

  6. e

    Device failed to assume native shape.

368FAustralian Shepherd124,99,2IIB1,842,459a122,2
623FMixed Breed102,85,7IIB2,14b2,506b72,2
673FDoberman Pinscher144,17,4IIB1,712,207c92,2
7028FGerman Shepherd294,910,3IIB1,842,049d101,7
7812MGerman Shepherd375,28,9IIA1,731,549e81,6

Final device waist diameter ranged from 3 to 14 mm (median, 7 mm). Final device waist diameter to MD median ratio was 1.76 (range, 1.54–2.5), and cupped proximal disc to AD ratio was 1.69 ± 0.27.

TEE provided complete monitoring of the entire procedure in 73/80 (91%) dogs; monitoring was incomplete in 5 cases and impossible in 2 cases.

Fluoroscopy exposure time was 2.77 ± 1.2 minutes.

Hypothetical TTE-based device size was calculated and compared with TEE device selection: in 12 cases (15%), the device would have been the same, in 25 and 4 cases it would have been 1 size larger and smaller, respectively, in 18 and 3 cases it would have been 2 sizes larger and smaller, respectively, and in 13 cases the device would have been 3 or more sizes larger.

Analyzing a subgroup of dogs (n = 17) with optimal TTE and TEE image quality, the device would not differ more than 2 sizes. In 8 and 1 cases it would have been 1 size larger and smaller, respectively, and in 5 and 2 cases it would have been 2 sizes larger and smaller, respectively.

Residual flow as assessed by TEE color Doppler was present in a majority of cases (64/80, 80%) before device release (1 severe, 11 moderate, 24 mild, 28 trivial), in 1 case the residual flow before device release was severe and the initial 9 mm device was exchanged by a 10 mm device as described above. This dog had a mild residual flow with the new device before and after release and no flow was detected the next day. After device release, residual flow was present in 28 dogs (11 mild, 17 trivial) and assessment for residual flow the day after revealed moderate residual flow in one dog. No residual flow was detected in the remaining dogs.

No dog presented clinical signs ascribable to esophageal trauma caused by the TEE probe.


This study describes a large number of cases in which TEE was used instead of angiography for device size selection, avoiding the use of iodinated contrast agents. The results show that TEE is an effective method to obtain MD and AD measurements and morphology information used for device size selection. Procedure monitoring was performed using a combination of fluoroscopy and TEE and the comparison of the 2 methods shows that TEE provides adequate monitoring for the entire procedure in most cases. However, fluoroscopy is required in cases where TEE monitoring is incomplete or impossible. The use of TEE also contributes to lower radiation exposure times to animals and staff. The comparison of TEE and TTE shows that the 2 methods lead to different device size selection in most cases.

Manufacturer instructions for device size selection recommend devices with a waist diameter of at least 1.5 times the size of the pulmonic ostium of the ductus (Amplatz® Canine Duct Occluder Instruction Manual6) (upsize factor) and previous publications recommend an upsize factor close to 2 (between 1.5 and 2.7)[1-3] based on angiographic measurements. The upsize factor used in this study was based on the previous publications as well as on a previous study by our workgroup in which TEE median upsize factor was 1.8.6 Although the median upsize factor in this study is slightly lower than that reported by other authors[1, 3] it is unclear whether there is relevant difference in terms of outcome or not, the high rate of success of both methods seems to suggest that there is not.

Most cases had a correct deployment of the device at first attempt; however, in 5 cases the device size had to be corrected during the procedure (Table 1). The initial device failed to remain stable in 3 cases and a larger one was required. Better assessment of the ductus in one of these cases revealed an elliptical shape not detected during the initial evaluation. In another case, the device was unstable despite taking into account an elliptical shape of the ductus for device selection (device 1 size larger), and required a device 2 sizes larger. In the remaining case, no clear reason was identified; measurement error or failure to detect an elliptical shape are possible explanations. In 1 occasion, the device was stable after deployment, but severe residual flow was detected suggesting that the device had not conformed completely to the ductal lumen. The decision to use a larger device was made based on the assumption that the device was too small. The presence of an elliptical shape identified with TEE[6] is assumed to contribute to the underestimation of the real MD as measurements are obtained in a sagittal plane, thus only the shorter diameter is measured. However, a correctly aligned transverse view of the distal ampulla and pulmonic ostium is necessary and depends on the ultrasound beam being parallel to the ductus short axis, which is often impossible. Assessing for elliptical shape with TEE may therefore lack sensitivity. Device size was increased in the cases where this morphology was detected but other cases were possibly undetected, as suspected in the case of unexplained device undersize and possibly also in the case with severe residual flow. Initially a device 1 size larger was used as correction for elliptical shaped ductus but this was insufficient in 1 case, which suggests that using a device that is 2 sizes larger may be more appropriate. This was adopted in the following cases. Failure of the device to assume its native shape is another described complication that can contribute to device instability or residual flow.[2] This complication is usually associated with excessive device upsize. The only case in which the device failed to expand had initial ratios of 1.7 and 1.8 for device to MD ratio and device to AD ratio, respectively (Table 1). An overestimation of the ductal dimensions was suspected. In all cases of incorrect deployment, device recapture and exchange were made without complications and the second device was correctly implanted. All cases requiring device correction because of undersize had a type IIB morphology whereas the case with device oversize had a type IIA morphology. Influence of ductal morphology on device stability has been reported for Type III ducti.[2, 16] Although the number of cases with device correction is too small to draw conclusions, these data suggest that differences between type IIA and IIB may also have an effect on device stability. A conical shape of the type IIB ductus ampulla may provide less stability to the cupped proximal disc when compared to the tubular type IIA ductus, possibly requiring slightly larger devices.

TTE measurements were used in 2 cases for device selection with successful occlusion. TTE is less expensive, widely available and does not require additional specific training in comparison to TEE. In a recent study, TTE was used for device size selection and procedure guidance,[10] however, only dogs with optimal image quality were included in that study. Our study shows a statistical significant difference in the MD and AD but more importantly shows that the 2 methods lead to different device size selection in most cases. This difference was of 3 mm or more in 13 cases (16%). Differences are still present but are smaller if image quality is taken into account. Although it is impossible to draw conclusions about different outcomes with different methods, differences shown may be important, especially if image quality is suboptimal.

TEE was efficient at monitoring the entire procedure in 73/80 cases (91%) providing a clear view of the endovascular material and vascular structures during all stages of the procedure. TEE monitoring was incomplete in 5 (6%) cases where adequate visualization of the vascular structures and endovascular material was interrupted during the procedure. This required probe position correction causing monitoring interruptions of variable duration that would have interfered with the normal course of the procedure if fluoroscopy was not available. Procedure monitoring with TEE was impossible in 2 cases and relied completely on fluoroscopy. The reasons were as follows: severe cardiomegaly with significant elevation of the heart base in 1 case and small dimensions (4.3 kg) in the second which limited probe manipulation. PDA visualization and procedure monitoring may be enhanced using TEE probes with side-to-side motion capability (not available in probe model3 used in this study) or with pediatric probes, with smaller tip diameter, that may yield better results in small dogs.

Fluoroscopy was required in cases of incomplete or impossible TEE monitoring. It was also helpful in situations where advancing the guiding catheter in the abdominal aorta was difficult with prompt identification and correction of accidental catheterization of renal artery or other main abdominal arteries. Although TEE may provide adequate procedure monitoring for the majority of cases, fluoroscopy is essential in some cases and should be available for this procedure.

Trivial, mild, or moderate residual flow before device release was commonly observed and was variable, depending on the amount of device distortion caused by the delivery cable before release. This residual flow tended to gradually decrease during the first minutes after release and appeared to occur through the device mesh. This type of residual flow was not considered significant as it is attributable to distortion of the nitinol mesh and tends to diminish after device release.[1, 6] One dog had mild residual flow after the device release based on TEE but a moderate residual flow was detected the following day by TTE. The authors believe that the residual flow was unlikely to be increased after release, as no device deformation or position change was detected on TTE or radiographic assessment. The difference is probably related to the method used to assess residual flow. Classification of residual flow based on color Doppler may be affected by the direction of the residual flow jet, differences in ultrasound transducer frequency and settings such as gain and pulse repetition frequency, which may change the color-flow Doppler jet area signal of the residual flow. Furthermore, ultrasound shadow and reverberation caused by the device between the transducer and the main pulmonary artery may also interfere with the ultrasound signal reducing TEE ability to assess residual flow.

Fluoroscopy exposure time (2.77 ± 1.2 minutes) was significantly shorter than reported in previous studies (reported median values varying between 5.1 and 9.4 minutes)[2, 3, 5, 6] for the same procedure. This difference is in part justified by the fact that angiography was not performed reducing procedure and radiation exposure time, but also by the influence of combined TEE guidance.

Being capable of providing adequate real time soft tissue imaging, TEE provided continuous information about the relative position of the endovascular material to the vascular structures during all phases of the procedure which was advantageous over fluoroscopy. In addition TEE was able to assess residual flow using color Doppler imaging or, when needed, using ultrasound contrast with agitated saline. These features complemented fluoroscopy reducing radiation exposure time and avoiding the use of contrast media.

Angiography was not performed in this study and therefore no angiographic measurements were available for comparison; this may limit our ability to compare measuring methods. Equipment cost and need for an additional trained operator are other limitations for the use of TEE.


TEE is an effective tool for device size selection and can be used instead of angiography for this purpose avoiding the use of iodinated contrast agents and radiation exposure. TEE is also efficient for procedure monitoring in most cases, suggesting that it may be accomplished with minimal use of radiation or even without it if operators are comfortable with this method. However, in a small number of cases TEE is not feasible or is insufficient for adequate procedure monitoring and fluoroscopy is required.


Grant support: This study was not supported by a grant.

Conflict of Interest Declaration: Authors disclose no conflict of interest. There is no association of the authors with the makers or marketers of the device used in this study.


  1. 1

    Infiniti Medical™, LLC, Menlo Park, CA

  2. 2

    ESAOTE S.p.A., Florence, Italy

  3. 3

    TEE022 Multiplane; ESAOTE S.p.A., Firenze, Italy

  4. 4

    Cordis, a Division of Johnson & Johnson Medical NV, Diegem, Belgium

  5. 5

    MedCalc 12; MedCalc Software, Ostend, Belgium

  6. 6

    Oliveira P, Domenech O, Silva J, et al. Percutaneous closure of patent ductus arteriosus with Amplatz® Canine Duct Occluder in 46 dogs: Outcome and prognostic survival factors. J Vet Intern Med 2009;23:1329 (abstract)