Specific contrast ultrasound using sterically stabilized microbubbles for early diagnosis of thromboembolic disease in a rabbit model

Abstract

Specific contrast ultrasound is widely applied in diagnostic procedures on humans but remains underused in veterinary medicine. The objective of this study was to evaluate the use of microbubble-based contrast for rapid ultrasonographic diagnosis of thrombosis in small animals, using male New Zealand white rabbits (average weight about 3.5 kg) as a model. It was hypothesized that the use of microbubble-based contrast agents will result in a faster and more precise diagnosis in our model of thrombosis. A pro-coagulant environment had been previously established by combining endothelial denudation and external vessel wall damage. Visualization of thrombi was achieved by application of contrast microbubbles [sterically stabilized, phospholipid-based microbubbles filled with sulfur hexafluoride (SF₆) gas] and ultrasonography. As a result, rapid and clear diagnosis of thrombi in aorta abdominalis was achieved within 10 to 30 s (mean: 17.3 s) by applying microbubbles as an ultrasound contrast medium. In the control group, diagnosis was not possible or took 90 to 180 s. Therefore, sterically stabilized microbubbles were found to be a suitable contrast agent for the rapid diagnosis of thrombi in an experimental model in rabbits. This contrast agent could be of practical importance in small animal practice for rapid diagnosis of thrombosis.

Introduction

Systemic thromboembolism is a frequent and life-threatening complication of cardiomyopathy in cats. Stasis of blood within dilated cardiac chambers, turbulent flow, and increased platelet reactivity combine to predispose a cardiomyopathic patient to systemic thromboembolism (1). Typically, the clot lodges at the aortic furcation (saddle thrombus), which results in a severe ischemic insult to the pelvic limbs and/or tail. A smaller clot may enter 1 iliac artery and cause unilateral paresis or paralysis of the limb. Besides pelvic limbs, 1 thoracic limb and certain organs, such as the kidneys, gastrointestinal tract, and brain, can also be affected. Vasoactive agents, such as prostaglandin or serotonin, that are released by platelets at the site of the thrombus result in constriction of collateral and regional vessels, which further contributes to ischemia and reduces the blood flow to terminal spinal cord segments (1,2). A saddle thrombus results in a consistent constellation of physical abnormalities, including a pelvic limb paresis or paralysis, absence of femoral pulsation, cyanosis, and local hypothermia. Hind limb musculature is typically firm and painful. Organ dysfunction may occur depending on the location of the thrombus. Affected cats almost always have significant underlying cardiac disease and sometimes even heart failure. Overall prognosis of this complication is guarded. Approximately 50% of affected cats do not survive the congestive heart failure and systemic thromboembolic crisis and die within 6 to 36 h. Those surviving typically show steady improvement in limb function within 24 to 72 h after presentation. The prognosis is poor for cats not showing an early improvement and euthanasia should be considered (1–4).

Regardless of the underlying cardiac disease, local control of the thrombus must be achieved within a relatively short period of time. Symptoms usually begin with pelvic paresis or paralysis with marked pain. Absence of femoral pulsation is a typical sign but to make a decision, an early and proper diagnosis is essential (1,2). In general, small-animal practice angiography is typically unavailable and Doppler abdominal ultrasonography requires an experienced radiologist in order to confirm the arterial thrombosis beyond a reasonable doubt (1–4). For better visualization of the distal aorta (dorsal to the urinary bladder), we suggest using novel contrast media to make the thrombus more transparent. Although specific contrast ultrasound is widely applied in diagnostic procedures on humans, it is still underused in veterinary medicine. Early and precise diagnosis of thromboembolic disease in cats could potentially improve the treatment, thus resulting in a better prognosis. Feline patients that are treated earlier in the disease process might benefit from an improved quality of life.

Ultrasound contrast agents are currently used to improve visualization of the microvasculature within organs and vital structures. These substances usually consist of small gas microbubbles (MBs) stabilized by a surfactant (5,6). Surfactants used include serum albumin, polymers, and phospholipids (7). Lipid-coated, gas-filled MBs represent a new class of drug delivery system with both diagnostic and therapeutic application (8). Lipid-based carrier systems represent drug vehicles composed of physiological lipids such as phospholipids, cholesterol, cholesterol esters, and triglycerides, as well as synthetic auxiliary lipids that provide the lipid particles, especially their surface, with a special function. The main advantage of MB application as drug delivery systems is a reduction of undesired side effects such as toxicity owing to drug targeting (9). Furthermore, diagnostic application of ultrasound imaging using MBs has become very popular because ultrasound is a noninvasive and relatively low-cost diagnostic tool. It uses portable, real-time imaging equipment and also avoids hazardous ionizing radiation (9,10). Microbubbles are small microspheres (typically 1 to 8 μm in diameter) filled with high-molecular-weight gases such as per-fluorocarbons or sulfur hexafluoride (SF6) that result in decreased solubility and prolonged lifespan of MBs within the circulation (9). The objective of the present study was to evaluate microbubble-based contrast for rapid ultrasonographic diagnosis of thrombosis in small animals using male New Zealand white rabbits (11,12). We hypothesize that the use of microbubble-based contrast agents will result in a faster and more precise diagnosis in this model.

Materials and methods

Preparation of liposomes and microbubbles

Liposomes were prepared by the lipid film hydration method. In brief, liposomes were composed of lipids (Avanti Polar Lipids, Alabaster, Alabama, USA) consisting of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[poly(ethyleneglycol)2000] (PEG₂₀₀₀DSPE), and Tween 80 (Sigma-Aldrich, St. Louis, Missouri, USA). Lipids were dissolved in chloroform and mixed properly. The mixture was subsequently transferred to a round-bottom flask and the solvent was removed by rotary evaporation at 37°C (Laborota 4000, Heidolph, Germany). The lipid film was hydrated by adding phosphate buffer saline (PBS) solution up to the final lipid concentration of 3 mg/mL above the transition temperature of the used lipids. Afterwards, DPPC liposomes were rapidly frozen in liquid nitrogen and thawed 5 times in a water bath at 55°C. The resulting liposome suspension was extruded through 400-nm polycarbonate membrane filters at 55°C using a Mini-Extruder (Avanti Polar Lipids). Thereafter, the size distribution of liposome suspension was determined by Zetasizer NanoZS (Malvern, Worchester, UK).

The liposome suspension (1 mL) was transferred to a hermetic 1.5-mL vial. The vial was filled with sulfur hexafluoride (SF₆) gas (Messer, Gumpoldskirchen, Austria) and mixed intensively for 30 s using the capsule mixing device 3M ESPE CapMix (3M ESPE, Germany) (13). The final composition of the MB sample was 2 mg/mL DPPC, 1 mg/mL PEG₂₀₀₀DSPE, and 0.03 mg/mL Tween 80.

Microbubble characterization by optical microscopy

The size and concentration of MBs were determined by optical microscopy (14). The MB samples were taken directly from the vial, diluted 50 times with PBS, and the final volume of 10 μL was transferred into the Bürker Counting Chamber. An Eclipse TE200 microscope (Nikon, Japan) was employed and the magnification of the objective Nikon LWD 20× was used to observe and capture the images of MBs. The images were analyzed by LUCIA software (Laboratory Imaging, Czech Republic) to determine the absolute count and size distribution of MBs.

Static light scattering (SLS)

HORIBA’s LA-950 Laser Diffraction Particle Size Distribution Analyzer (Horiba, France) was used to determine the MB size distribution by static light scattering (SLS) (14). At first, the method was optimized by Megabead NIST Traceable Particle Size Standards from 1, 3, 6, 10, to 15 μm (Polysciences, Warrington, Pennsylvania, USA). The 1.0% suspension of polystyrene microspheres in water was diluted by degassed and filtered PBS and measured. An MB sample (5 to 30 μL, 3 mg phospholipid/mL) was injected into a cuvette-type fraction cell (filled with 10 mL of PBS), equipped with a magnetic stirrer to prevent non-homogenous distribution owing to flotation of MBs. All samples were measured immediately after the application into the cuvette and were analyzed both for number- and volume-weighted size distribution.

Animals

Male New Zealand white rabbits (approximately 6 mo old, average weight 3.5 kg) were purchased from Biotest Konarovice (Konarovice, Czech Republic) and housed for 14 d before the experiments. The animals were treated in accordance with the “Guide for Care and Use of Laboratory Animals” (DHHS publ. No. NIH 85-23, revised 1996, Office of Science and Health Reports, Bethesda, Maryland, USA). The experimental protocol was approved by the Ethical Committee at the University of Veterinary and Pharmaceutical Sciences in Brno.

Animal groups

The rabbits were randomized into 2 groups (n = 6), as we intended to precisely establish the time up to clot identification.

Anesthesia and monitoring

Anesthesia was induced with a mixture of 0.2 mg/kg body weight (BW) of medetomidine and 15 mg/kg BW of ketamine mixed in 1 syringe and administered as a single intramuscular injection into the dorsal lumbar muscles. A temporary tracheostomy was done using a number-4 tracheal tube without cuff and anesthesia was maintained by a mixture of oxygen and isoflurane ranging from 1% to 1.5% of isoflurane. A 22G catheter was inserted percutaneously into the marginal auricular vein and constant rate infusion of Ringer’s solution was administered using a perfusor at the rate of 10 mL/kg per hour. The left common carotid artery was then cannulated and the arterial line was connected to a pre-calibrated arterial blood pressure transducer for continuous recording of blood pressure and heart rate. The circulatory and respiratory variables evaluated included the systolic (SAP), mean (MAP), and diastolic (DAP) arterial blood pressure, heart rate (HR), saturation of hemoglobin by oxygen (SpO₂), and respiratory rate (RR). Systolic arterial blood pressure (SAP), MAP, and DAP were measured directly through an arterial blood pressure transducer, HR was measured by electrocar- by a sensor diogram (ECG) electrodes applied on the thorax, SpO₂ applied on the tongue, and RR was measured electronically based on thorax-impedance changes. All parameters were acquired and saved by a Mindray PM-9000Vet vital monitor (Mindray, Shenzhen, China).

Surgery

We modified a previously validated and published rabbit model system (11,12) that simulated the conditions of acute arterial thrombosis. In brief, thrombosis was induced by injecting 10 μL of whole blood obtained from the carotid arterial line into the distal segment of the abdominal aorta via the left iliac artery and kept in place by 2 tourniquets occluding the vascular segment, placed approximately 10 μL apart. A pro-coagulant environment had been previously established by combining endothelial denudation and external vessel wall damage. The clot was left in place for 30 min of maturation (Figure 1) and the tourniquets were then removed and an external constrictor (over the needle loose 4-0 silk ligation) was applied distally to prevent the clot from moving. After the tourniquet was released, the flow was partially re-established. With the use of flow probes, the blood flow distal to the thrombus was monitored for 10 min. The flow probes were then removed and organs were flushed using warm sterile saline solution. The abdominal wall was then closed in a simple continuous suture pattern (3-0 Ethibond, Ethicon) using a single layer.

Figure 1.

Arterial thrombosis rabbit model. Complete isolation of distal aorta with mural clot (arrow) that can be seen after 30 min of maturation.

Endothelial denudation was achieved by passing an inflated #4 Fogarty embolectomy catheter 30 times through an isolated aortic segment via cannulation of the left external iliac artery. Aortic stenosis was then achieved by use of an external constrictor and the lumen reduction was maintained at −95% (as assessed by angiography). External damage of the aortic wall was induced by 16 clamps proximal to the stenosis using straight Mosquito hemostat. Blood pressure and aortic blood flow were recorded by use of a Single Channel Flow Meter.

Thereafter, distal aortic thrombus was identified by ultrasonography with Doppler system (Vivid 7; GE Healthcare, Pittsburgh, Pennsylvania, USA). To evaluate the efficacy of the MB suspension as a contrast substance, 300 μL of saline was intravenously administered in 1 group, while MB suspension was applied to the venous system of the other group. The time to clot identification was measured and the transparency of the image was assessed in each case. The MBs were administered through the venous line in 1 bolus at a dose of 300 μL (approximately 7 × 10⁸ MBs, 0.9 mg of lipid). The study was blinded so the ultrasonographer was not aware of the treatment applied to the specific animal.

After termination of each experiment, the animal was euthanized under anesthesia using a high dose of sodium pentobarbital, a sufficient aortic segment was explanted, and the mural clot was carefully removed from the lumen. It was weighed (mg) immediately and the clot size was compared with the ultrasonography (USG) image with and without the MBs.

Endpoints

As an index of efficacy, i) time to clot identification (s); ii) transparency of the image (in the range of + to ++++); and iii) the weight of the mural clot (mg) were documented. In all the experiments, the clot identification was done by the same observer, an experienced clinician, who was blinded to the treatment groups. The system for assessing the clot visibility is outlined in Table I.

Table I.

System for assessing clot visibility

Clot transparency	Description
0	No clot identified
+	Suspected clot with no detectable contours
++	Distinctive clot, yet still doubtful contours
+++	Evident clot with ill-defined contours
++++	Evident clot with well-defined contours

Statistical analysis

Statistical analysis was carried out using statistical software Statistica 9.1 (StatSoft, Tulsa, Oklahoma, USA). The number of observations in this study is on a lower margin where it is appropriate to use statistical tests. Nonparametric Mann-Whitney U-test is undertaken for explorative comparison in order to complete results of descriptive statistics. P-values lower than 0.05 were considered to be statistically significant.

Results

Characterization of microbubbles

The size distribution of MBs was determined by static light scattering (SLS) and by optical microscopy using the HORIBA-LA 950 (HORIBA Scientific, Edison, New Jersey, USA). Its main advantage over other SLS instruments is that measurements can be done in a relatively small cuvette (10 mL) with magnetic stirring to prevent changes in the distribution of MBs in the laser beam area due to flotation. The size distribution of the MBs (expressed in terms of number) ranged from approximately 2 to 10 μm with the mean 3.52 ± 0.68 μm. This data correlates well with the optical microscopy data. The average size of MBs determined by optical microscopy was 2.11 ± 1.41 μm, with the size ranging from 0.66 to 12.10 μm (Table II, Figure 2). Optical microscopy allowed prepared MBs to be directly observed. The multimodal distribution was recognized by both techniques (Figure 2).

Table II.

Parameters of microbubbles (MB) preparation

MB size (μm) by optical microscopy
Average ± SD	2.11 ± 1.41
Min	0.66
Max	12.10
MB size (μm) by SLS
Average ± SD	3.52 ± 0.68
Min	1.94
Max	10.25
MB concentration (MBs/mL)
Optical microscopy	2.33 × 109
Cell counter	2.91 × 109

SD — Standard deviation; SLS — Static light scattering. Composition of MBs was 2 mg/mL of DPPC + 1 mg/mL of PEG₂₀₀₀DSPE + 0.03 mg/mL of Tween 80.

Figure 2.

Characterization of microbubbles. A — Optical microscopy of microbubbles. Composition was 2 mg/mL DPPC, 1 mg/mL PEG₂₀₀₀DSPE, and 0.03 mg/mL TWEEN 80. B — Transmission electron microscopy of microbubbles. During desiccation of the sample in the vacuum inside the electron microscope, the microbubbles burst and the ruptures could be seen in their membranes. C — Graph of the size distribution of microbubbles measured by static light scattering (Horiba LA 950 SLS instrument). D — Histogram of microbubbles obtained by optical microscopy.

Optical microscopy and the cell counter BC-2800 VET (Mindray) were used to evaluate the concentration of MBs in the prepared samples. The counting of MBs revealed that the amount of MBs was 2.33 × 10⁹ MBs/mL and 2.91 × 10⁹ MBs/mL for optical microscopy and the cell counter, respectively (Table II).

Animal study

The mean clot weight in a control group was 116.7 (± 16.33) mg with the range of 90 to 130 mg. However, the clot was identified beyond a reasonable doubt in only 4 out of 6 animals with a transparency of 1 star in all detectable cases (Figure 3). Also, the time needed for clot identification varied from 90 to 180 s (mean: 148.8 ± 41.31 s) in the control group (Table III).

Figure 3.

Image of the clot visualized by microbubble contrast media. A — The ultrasound image of the abdominal vessel with stenosis before the application of microbubbles. B — The ultrasound image of the abdominal vessel with stenosis after the application of microbubbles. Long white arrow indicates the direction of blood flow; short black thick arrows indicate the vessel wall; long black arrows indicate stenosis; broken white arrow indicates direction of blood flow.

Table III.

Parameters of thrombi identification in animal groups

	Study group			Control group
Number	TCI (s)	TI	CW (mg)	TCI (s)	TI	CW (mg)
1	15	++++	110	90	+	130
2	17	+++	90	175	+	110
3	20	+++	130	150	+	90
4	30	++++	150	N/A	0	130
5	10	++++	110	N/A	0	110
6	12	++++	120	180	+	130
N	6		6	4		6
Mean (SD)	17.3 (7.15)		118.3 (20.41)	148.8 (41.31)		116.7 (16.33)
Median	16		115	162.5		120
Range	10 to 30		90 to 150	90 to 180		90 to 130

TCI — time to clot Identification (s); TI — (0 to +++) — transparency of the image; CW — weight of the mural clot at the end of each experiment (mg) for control (given 300 μL of saline) and study group (given 300 μL of contrast media) for clot imaging; N — number of animals; SD — standard deviation.

On the other hand, we succeeded in detecting each and every clot introduced to the study group. While the mean clot weight using MB substance applied to the venous system was 118.3 (± 20.41) mg with the range of 90 to 150 mg, the mean time to clot identification was significantly (P = 0.0027) shorter (10 to 30 s with mean of 17.3 ± 7.15 s) than the control group. The transparency of the images also differed from the control, since it varied between 3+ and 4+ (Table III). The ultrasonography was carried out in a standard manner and did not require any special manipulation with the probe. The images were clear and easily recognizable, beyond suspicion (Figure 4).

Figure 4.

Image of the clot visualized by microbubble contrast media. Black arrow and black brackets indicate thrombus; white arrows indicate the vessel walls of aorta abdominalis; white bracket indicates stenosis; broken white arrow indicates direction of blood flow.

Statistical description of the results

As shown in the statistics in Table III and in the values of the observed parameters themselves, time to clot identification and transparency of the image differed between the study group and the control group. The application of microbubbles as an ultrasound contrast medium decreased the time to clot identification by 10 times. The standard deviation and the variance in the time needed for clot identification in the study group suggests that the variability of the studied method is quite low. Due to the very large difference between the groups, a small sample size was enough to demonstrate, significantly (P = 0.0027), the lower time needed to identify the clot when microbubbles are used. The weight of the clots was well balanced between the groups. The difference in median of the weights was only 5 mg and the range was also very much alike.

It is also evident from the obtained results that the transparency of the image increases when microbubbles are applied. With the microbubbles, two thirds of images are evaluated by the highest degree of transparency. In the control group; however, two thirds of images are evaluated as the least transparent and the clot was not found at all in the remaining one third of images.

Discussion

Some criteria need to be met in order to establish a new applicable diagnostic tool for an acute thromboembolia in small animals using contrast ultrasonography. First, the procedure itself should not be technically demanding and the contrast media must be readily available, as thromboembolic disease is often an acute condition, with the patients arriving late in the afternoon or during the night, in an emergency situation. Lyophilized MB preparations sealed under the filling gas meet these criteria and are ready for rapid reconstitution and administration. Moreover, after reconstitution, they are stable for at least a working day when stored in a refrigerator under the filling gas.

Second, intravenous (IV) administration of contrast media is preferred during the examination, since femoral artery cannulation is not a routine procedure in a small animal practice, especially when the blood flow through the artery has been compromised. The contrast media must cross the pulmonary barrier through the right cardiac outflow tract in order to enter the arterial circulation. For clinical availability, IV administration has been recommended, stressing the need for quick and easy passage through pulmonary circulation as was proved in an earlier study in mice (13).

Third, the brightness of the image must be strongly superior to any other method, showing both high sensitivity and specificity. We believe that this was the case in our study.

Even though the study was blinded, i.e., the ultrasonographer was not aware of the substance applied to the venous system, either saline or MB suspension, the difference in image was apparent from the beginning. The time to clot identification was significantly shorter and 100% of embolic clots introduced into the abdominal aorta had been identified using our MB suspension. It took significantly less time to obtain enough data to clearly diagnose the study group despite the fact that the difference in the mean explanted-clot weight was not statistically significant between the groups. Since we are looking for clinical relevance, we do not feel that the addition of 1 or more observers to assess the clot transparency would benefit the study. Also, while we are aware that larger groups of animals would most likely increase the statistical significance, we kept the numbers of animals as low as possible for ethical reasons. In our opinion, the differences between the groups are evident even when the lowest comparable numbers were used.

Although selective angiography with iodinated contrast medium has been the gold standard for diagnosing aortic thrombosis, abdominal ultrasonography is the most common accessible method in identifying aortic thrombi (15). The ultrasonographic appearance of a thrombus depends on its duration. Chronic thrombi usually appear echogenic and heterogeneous, whereas immature thrombi can be difficult to distinguish from flowing blood because they can be hypoechoic to anechoic and homogeneous (16). Doppler flow evaluation is generally required to evaluate thrombi (15,17).

Lack of a Doppler signal is a diagnostic sign of a thrombus. In these cases, it is important that the settings of the ultrasound machine are correct and that optimal images of the affected vessels are obtained. Otherwise, the lack of a Doppler signal due to poor technique could be misinterpreted as a sign of thrombosis (18).

When the conventional color or power Doppler techniques are not sensitive enough, namely in cases of deeply located and/or small vessels with slow flow or in vessels with inadequate Doppler angle, ultrasound contrast media may be used to justify the presence or absence as well as the direction of blood flow in a certain vessel. Although clinical application of commercially available contrast agents in small animals may be beneficial in diagnosing aortic thromboembolism, its widespread use might be limited because of the high price of the contrast media and the ultrasound instruments that contain contrast-specific second harmonic modalities (19).

To the authors’ knowledge, there are no reports of cases of aortic thromboembolism in small animals routinely diagnosed by the use of (non-commercial) contrast agent and general ultrasonography without the use of Doppler. Therefore, a diagnosis based on applying a less expensive agent that provides a good contrast even with the use of conventional ultrasound instrument modality might be very promising due to its common accessibility.

In conclusion, acute thromboembolism is a potentially life-threatening condition that requires an immediate and efficient action. Besides general diagnosis, the early and definitive identification of a clot is absolutely crucial for future planning. In this study, we proved the successful application of microbubble-based contrast agents for rapidly diagnosing thrombosis in a rabbit model. Our plan was to develop and introduce a safe, easy-to-use, and efficient diagnostic tool to help to localize clots as well as for possible use in therapeutic targeting. In our opinion, the MBs tested meet these criteria and, among other possible uses, should be considered in the first-choice diagnostic plan for acute thromboembolic disease.