Author + information
- Received December 11, 1997
- Revision received March 6, 1999
- Accepted April 14, 1999
- Published online August 1, 1999.
- Takanori Yasu, MD, PhD∗,
- Geert W Schmid-Schönbein, PhD†,* (, )
- Bruno Cotter, MD∗ and
- Anthony N DeMaria, MD, FACC∗
- ↵*Reprint requests and correspondence: Dr. Geert W. Schmid-Schönbein, Institute for Biomedical Engineering and Department of Bioengineering, University of California, at San Diego, 9500 Gilman Drive, La Jolla, California 92093-0412.
The purpose of this study was to test the hypothesis that a subgroup of QW7437 microbubbles, dodecafluoropentane-based ultrasound contrast microspheres, resides for prolonged periods in the microvasculature.
QW7437 produces echo enhancement in myocardium which may persist relatively longer than opacification in the left ventricular cavity. The mechanism for this persistent enhancement remains unknown.
The transit of fluorescently labeled erythrocytes was examined by fluorescence intravital microscopy in the microvessels in five rat mesenteries. Ten rats were used to observe the behavior of fluorescently labeled QW7437 microbubbles in the mesenteric microcirculation.
There was no significant change in erythrocyte velocity in the arterioles and venules after the administration of QW7437 microbubbles (0.05 ml/kg) preactivated by negative hydrodynamic pressure. Of 552 microbubbles observed in four arterioles and five capillaries, 549 (99.5%) passed without stoppage (≥0.1 s stoppage); only one stopped transiently in arteriole and two in capillaries, each for <0.5 s. Sixty-five of 478 microbubbles (13.6%) observed in six postcapillary venules 11 to 30 μm in diameter and 24 of 408 microbubbles (5.9%) in four venules 31 to 50 μm in diameter stopped transiently (0.1 to 180 s) with an attachment to venular endothelium; the remaining microbubbles passed through the venules without stoppage.
Prolonged survival as microbubbles in the circulation and transient stoppage of a subgroup of microbubbles in the microvasculature, particularly in venules, are potential mechanisms for the persistent tissue echo enhancement by QW7437 microbubbles during contrast echocardiography.
The development of new ultrasound contrast techniques offers an opportunity to expand the application of echocardiography in clinical cardiology (1–3). A number of new ultrasonic contrast agents have recently been developed based upon the concept that extended stability of microbubbles in blood would enable myocardial enhancement to be achieved by intravenous administration (4–7). One of these agents, a dodecafluoropentane (DDFP) emulsion which converts to a gas at body temperature (37°C), yields enhancement of myocardium, kidney and other organs, which is relatively unusual because it can persist beyond left ventricular cavity opacification (4–7). The mechanism for the persistent gray-scale enhancement of the tissue signal remains unknown. In fact, limited data exist regarding the behavior of air-filled ultrasound contrast microbubbles within the microcirculation (8–11).
We hypothesized that a subgroup of DDFP microbubbles (QW7437) in vivo might remain in the normal microvasculature longer than erythrocytes, especially at sites with lower shear stress. To investigate this hypothesis we serially observed the behavior of fluorescently labeled QW7437 microbubbles and erythrocytes in mesenteric arterioles, capillaries and venules of rats by intravital fluorescence microscopy.
All animal procedures were reviewed and approved by the University of California at San Diego Animal Subjects Committee. Wistar rats (n = 15, at 13 to 15 weeks of age, 280 to 400 g; Charles River Breeding Laboratories, Wilmington, Massachusetts) were used for the study. After general anesthesia with pentobarbital sodium (30 mg/kg IM), a catheter (PE-50 tubing, Clay Adams, Parsippany, New Jersey) was inserted into the left femoral artery and systemic blood pressure was continuously monitored. Another catheter was inserted into the left femoral vein. The animals were placed on a water heating pad maintained at 37°C. The abdomen was opened by a small midline incision and the ileocecal part of the mesentery was carefully exteriorized and draped over a thin plastic foil (Reynold’s Plastic Wrap, Richmond, Virginia) to minimize exposure to air and room temperature during intravital microscopy (12). The preparation was maintained at 37°C and continuously superfused (1.0 ml/min) with a Krebs-Henseleit bicarbonate-buffered solution saturated with a 95% N2and 5% CO2gas mixture to maintain physiologic pH.
Intravital fluorescence microscopy and image analysis
The mesenteric microcirculation was visualized through an intravital fluorescence microscope (× 25, water-immersion objective lens, Leitz, Wetzlar, Germany) with the use of a silicone-intensified target tube camera (Series 66, Dage-MTI, Michigan City, Indiana). To elicit fluorescent images, the sample was illuminated with a 200-W mercury lamp. The light was passed through a quartz collection, a heat filter (Model KG-2; Carl Zeiss, Thornwood, New York) and a 515- to 560-nm wavelength excitation filter (Leitz, Rockleigh, New Jersey) to epi-illuminate the sample. Fluorescent emission from the selected area was passed through a 580-nm thick bandpass filter (Leitz). Transillumination was provided by a 250-W halogen light source with heat filter. The video vertical frame rate of the camera was 30/s. The overall magnification on the TV monitor was approximately × 910. The microscope was focused to limit the observations to the median plane of the vessel of interest. The images were stored on a videocassette recorder (Model AG-6300, Panasonic, Tokyo, Japan) for off-line analysis. All images were compiled and analyzed on a Macintosh computer with public domain National Institutes of Health image software (Version 1.60). The computer was calibrated in the x-y planes using a calibration reticule (10-μm divisions). The vessel diameters refer to inner lumen measurements; no corrections were made for noncircular cross sections (13). The velocity of labeled microbubbles and erythrocytes in arterioles, capillaries and venules was measured twice with frame by frame analysis (9), and the two measurements were averaged. We defined a transient stoppage of microbubbles as stopping in the vessels persisting for 0.1 s (3 frames) or longer, and less than 3 min. The frequency of microbubbles adhering to the endothelial cells (EC) was determined by dividing the number of adhering microbubbles per 200 μm length of each vessel by the total number of microbubbles observed in that vessel.
QW7437 (Sonus Pharmaceuticals, Bothell, Washington) is a premixed liquid-in-liquid emulsion of DDFP which converts to a gas at body temperature (37°C). The emulsion contains droplets with a mean diameter of approximately 0.3 μm. On intravenous administration, this liquid emulsion becomes a dispersion of microbubbles with an average diameter of 2 to 5 μm and a mean concentration of 1012microbubbles/ml (4,6). The emulusion is negatively charged to minimize adhesion of the microbubbles to the vascular EC (5). The surface of QW7437 was labeled by a fluorescent dye PKH-26 (Zynaxis Cell Science, Malvern, Pennsylvania) (Fig. 1), a label for cell membrane lipids (14–17). To observe the microbubbles in vivo, the following fluorescent labeling technique was used: 1) a 10−3-mol/liter stock solution of the fluorescent dye PKH-26 was added to the emulsions of QW7437 for a final PKH-26 concentration of 5 × 10−6mol/liter; 2) the mixtures were gently stirred in the dark for 5 min at room temperature; 3) after drawing 1 ml of labeled or nonlabeled QW7437 into a 30-ml syringe with a stopcock, a negative hydrodynamic pressure was manually imposed on the QW7437 by pulling back the plunger abruptly to the 10-ml line, then releasing it immediately to convert some of the QW7437 emulsion from liquid to gas before intravenous administration (7,18,19); 4) to stop the staining reaction, 1 ml rat plasma was added to the PKH-26 with QW7437 mixtures and gently mixed for 1 min. The mixtures were drawn into a 1-ml syringe for injection. For all injections, the QW7437 emulsion was administered as soon as possible after application of the negative hydrodynamic pressure. In a preliminary study, myocardial contrast echocardiography (MCE) by intravenous injection of 0.05 ml/kg of PKH-labeled QW7437 preactivated by negative hydrodynamic pressure showed similar intensity and duration of myocardial enhancement in rats as 0.05ml/kg of nonlabeled QW7437 (data not shown). This finding supports the concept that PKH staining does not change the echogenicity of QW7437 microbubbles.
Microbubble size in vitro
An in vitro study was carried out to assess the effect of the body temperature (37°C) on the size distribution of QW7437 microbubbles. Fresh human venous blood was drawn by a clean venipuncture of an anticubital vein of a healthy volunteer into a plastic syringe containing ethylenediaminetetraacetic acid–2Na (1 mg/ml) with a pressure-resistant stopcock just before the experiment. Special attention was taken not to expose the sampled blood to the air. Labeled QW7437 (0.05 ml) after the application of negative pressure was mixed with 2 ml of the fresh human blood at 37°C in a 3-ml syringe with a stopcock. The syringe was gently and continuously rotated in a water bath at 37°C for 60 s. Slide glasses were warmed on a heating pad at 37°C until microbubbles were placed on them. Then one drop of the mixture was immediately placed on a glass slide with a coverslip. Microbubble sizes were observed using microscopy (× 40, water-immersion objective lens, Leitz, Wetzlar, Germany). The images were recorded with a color CCD (coupled charge device) camera (DEI-470, Optonics, Goleta, California) and stored on VHS tape using the videocassette recorder (Panasonic, Tokyo, Japan) within 60 s after placing the microbubbles on the glass slide. The diameters of 500 microbubbles were measured by off-line analysis using National Institutes of Health image software as described above. The overall magnification on the TV monitor for these measurements was approximately × 2,000. To get control values before mixing with blood, one drop of labeled QW7437 solution after the application of negative pressure was placed on a glass slide with a coverslip.
In vivo protocol 1
Five Wistar rats were used for protocol 1 to serially measure the velocity of fluorescently labeled erythrocytes in the mesenteric microcirculation before and for 10 min after the intravenous administration of QW7437. Erythrocytes from separate donor rats were labeled with PKH-26 according to the manufacturer’s instructions. PKH-labeled erythrocytes are stable in vivo (14–16). This enhanced stability allowed us to determine the erythrocyte velocity in vivo throughout the 10-min course of the experiment after a single bolus injection. Injection of 0.1 ml of PKH-labeled erythrocytes was carried out via the left femoral vein. After recording a control image, 0.05 ml/kg of nonlabeled QW7437 treated with the negative hydrodynamic pressure (5,20)was immediately mixed with 0.2 ml whole blood drawn from the femoral vein, and injected over a 15-s period via the femoral vein. The catheter was flushed with 0.25 ml saline after each injection.
A selected tissue area was recorded under fluorescent microscopy for 1 min after the onset of the QW7437 bolus injection, and for periods of 50 s each at 2, 5 and 10 min after the injection. Five arterioles with diameters between 11 and 30 μm and four venules with diameters between 25 and 50 μm were recorded for analysis. After the recording of each fluorescent image, transillumination images were obtained for off-line measurements of vessel diameters.
In vivo protocol 2
Ten Wistar rats were used for protocol 2 to determine whether vascular obstruction or transient stoppage of QW7437 microbubbles with attachment to microvascular EC occurred or not. Of the mixtures containing 0.05 ml/kg of labeled QW7437 after application of the negative pressure, 0.1 ml/kg was injected over a 15-s period via the left femoral vein in the same way as described in protocol 1. Only one bolus of QW7437 was administered to each rat. Fluorescent images of selected tissue areas were recorded continuously for 10 min after the injections. Four arterioles examined ranged in diameter between 13 μm and 29 μm, five capillaries in diameter between 5 μm and 8 μm and 10 venules in diameter between 11 μm and 50 μm. When microbubbles stopped in vessels for more than 2 s, we simultaneously switched to semibright field for a few seconds, to determine whether vascular obstruction occurred or not. We skipped this semibright field observation when the other microbubbles passed by retained microbubbles in venules, because this passing phenomenon clearly demonstrates QW7437 microbubble attachment to EC without obstruction. Fifteen minutes after QW7437 injection, 0.1-ml PKH-labeled erythrocytes were injected in two rats into the left femoral vein to compare the velocities of microbubbles and erythrocytes in individual venules with diameters of 12 μm, 13 μm and 44 μm.
All results are expressed as means ± SD. Time course of labeled erythrocyte velocities (maximum and mean) and vessel diameter in protocol 1 and systolic blood pressure data in protocol 2 were assessed by one-factor analysis of variance for repeated measures. Velocities of the labeled microbubbles was compared with those of labeled erythrocytes in individual venules by unpaired Student ttest. Frequency of transiently adhering microbubbles was compared between arterioles, capillaries and venules by chi-square or Fisher exact probability test for group data. Probability levels of less than 0.05 were considered to be statistically significant. The statistical analysis was carried out with Stat View IV software (Abacus Concepts, Berkeley, California).
Microbubble size in vitro
The in vitro QW7437 diameter of 500 microbubbles immediately after the application of negative pressure was 5.4 ± 2.6 μm (mean ± SD), 4.8 μm (median), 2.5 μm (10th, percentile), 8.0 μm (90th, percentile) with a range between 1 and 22 μm and 4.4% of microbubbles larger than 10 μm in diameter. After mixture with human blood at 37°C for 1 min, the QW7437 diameter was increased to 5.9 ± 5.2 μm (mean ± SD, p < 0.05 vs. before mixing with human blood at 37°C(n = 335), 3.5 μm (median), 2.0 μm (10th, percentile), 12.0 μm (90th percentile) with a range between 1 and 34 μm and 12.4% of microbubbles larger than 10 μm in diameter.
In vivo protocol 1
A summary of the velocity data of labeled erythrocytes and measurement of vessel diameter is presented in Table 1. More than 50 labeled erythrocytes were visible and assessed in every vessel of interest during each examination period. There was no significant change in erythrocyte velocity in each vessel after the administration of QW7437 compared with the erythrocyte velocity before the administration. No labeled erythrocytes adhered to the vessels. The diameters of the arterioles and venules remained unchanged.
In vivo protocol 2
Neither tachypnea (not quantified) nor significant change in systolic blood pressure or heart rate over time was observed during protocol 2 (before administration, 130 ± 13 mm Hg, 342 ± 39 beats/min; 30 s after administration, 128 ± 13 mm Hg, 343 ± 29 beats/min; at 1 min, 131 ± 15 mm Hg, 341 ± 33 beats/min; at 5 min, 128 ± 13 mm Hg, 335 ± 32 beats/min, and at 10 min, 126 ± 17 mm Hg, 337 ± 35 beats/min).
More than 50 labeled microbubbles (about 3 × 10−7% of intravenously injected microbubbles) were observed within each observation field after each injection. Several microbubbles could be visualized in the microvasculatures for periods of 5 min after administration. In four arterioles, 399 out of 400 microbubbles passed rapidly through arterioles without transient stoppage during a period of 10 min after the injection (Fig. 2A and 3), ⇓⇓and only one microbubble (0.25%) transiently stopped in an arteriole for 0.27 s. The velocity of QW7437 microbubbles in arterioles (n = 400, 3.59 ± 1.93 mm/s, p = 0.41) was not significantly different from that of labeled erythrocytes in arterioles after administration of the agent in protocol 1 (n = 257, 3.70 ± 1.16 mm/s). Among a total of 152 microbubbles observed in capillaries, 150 microbubbles flowed freely through five capillaries without transient stoppage (Fig. 2B and 3). Only two microbubbles (1.3%) were observed to be transiently stopped on capillary EC for less than 0.5 s. Unfortunately we could not observe the passing phenomenon or persistent blood flow in semibright field when the three microbubbles showed transient stoppage in arterioles and a capillary.
Transient stoppage of microbubbles occurred more frequently in venules than in arterioles or capillaries (p < 0.0001). When microbubbles stopped in venules for more than 2 s, semibright field showed persistent erythrocyte flow in all of these particular venules. The venules examined ranged in diameter between 11 and 50 μm, and were subdivided into those ≤30 μm and those >30 μm. Transient stoppage occurred more frequently in smaller venules with diameters between 11 μm and 30 μm (13.6%, p < 0.001, Fig. 4A and 5)⇓⇓than in larger venules with diameters between 31 μm and 50 μm (5.9%, Fig. 4B). The adhering microbubbles remained stationary at one site for periods ranging from 0.1 s to 120 s (Table 2). Among the set of stopping bubbles, 60 microbubbles (65.2%) were attached to venular EC for more than 1.0 s and only six microbubbles (6.5%) were stationary for more than 60.0 s. The diameter of the fluorescent microbubble images which stopped in venules and remained in sharp focus on still frame review was 6.4 ± 1.2 μm (n = 69). Their size remained unchanged during the period of stoppage. No microbubbles showed permanent stoppage (>3 min) or extravascular emigration in the selected observation areas.
Microbubble velocity relative to that of erythrocytes varied among individual venules. In venule #1 (vessel diameter of 12 μm) and #2 (vessel diameter of 13 μm) in Table 2, velocities of labeled QW7437 microbubbles (venule #1: 0.88 ± 0.67 mm/s, n = 73; venule #2: 0.91 ± 0.55 mm/s, n = 84) were significantly slower (p < 0.05) than those of erythrocytes (venule #1: 1.14 ± 0.55 mm/s, n = 50; venule #2: 1.13 ± 0.4 mm/s, n = 68). In venule #8 (vessel diameter of 44 μm), however, velocities of labeled QW7437 microbubbles (1.58 ± 0.72 mm/s, n = 107, p = 0.1) were not significantly different from those of erythrocytes (1.74 ± 0.60 mm/s, n = 90). When the velocity of labeled erythrocytes in 25- to 50-μm venules (n = 412, 1.82 ± 0.55 mm/s) after administration of nonlabeled QW7437 in protocol 1 was compared with the velocity of QW7437 microbubbles in similar size venules in protocol 2, the latter was significantly slower (n = 485, 1.10 ± 0.80 mm/s, p < 0.001).
The present measurements document for the first time the rheologic characteristics of DDFP ultrasound contrast microbubbles in the microcirculation using a fluorescence intravital microscopic approach. The data demonstrate that, after injection of a clinical dose of QW7437, erythrocyte velocity is unaltered in arterioles and venules in the presence of the microbubbles, and transient stoppage of microbubbles occurs rarely in arterioles and more frequently in venules.
Microbubble life span in vivo
QW7437 microbubbles were observed in the circulation, including arterioles, for periods up to 5 min after administration. This represents a longer period of stability than that seen with sonicated albumin microbubbles, which become invisible in the cheek pouch microcirculation of hamsters (9)and in the mesentery microcirculation of rats (unpublished data) within 2 min of intravenous administration. The potential mechanisms by which QW7437 microbubbles persist in blood and provide longer acoustic contrast than air or carbon dioxide microbubbles include: 1) DDFP has a much lower solubility in water than oxygen or nitrogen; 2) it has slower diffusion in blood because of its larger molecular weight (228 D); 3) it has a low concentration of saturation in blood, and 4) it contains carbon–fluorine bonds which are not metabolized in the body but are excreted through the lung (6,7).
Microbubble velocity and transient adhesion to the endothelium
Keller et al. (9)have reported that sonicated albumin microbubbles showed a velocity profile similar to that of erythrocytes in the cheek pouch microcirculation of hamsters. They simultaneously measured the velocities in arterioles of sonicated albumin microbubbles labeled with dichlorotriazinylaminofluorescein-1-dihydrochloride and calcein-labeled erythrocytes using intravital fluorescence microscopy equipped with a strobe illuminator. We measured the velocity of labeled erythrocytes and microbubbles in separate protocols, since preliminary experiments had shown that it was difficult to distinguish PKH-labeled erythrocytes from PKH-labeled microbubbles in vivo.
The slow intravenous injection of 0.05 ml/kg preactivated QW7437 microbubbles over 15 s did not produce a change in systemic blood pressure and heart rate. In addition, this clinical dose of QW7437 neither changed the vessel diameters of arterioles or venules, nor altered erythrocyte velocities in arterioles and venules. These results are consistent with the previous reports (18,19).
In vitro results showed a heterogeneous growth of QW7437 microbubbles after 1 min mixture with blood at 37°C. However, no obstruction was seen with QW7437 microbubbles in arterioles and capillaries in the mesenteric microcirculation. Obviously, microbubbles are subjected to different pressures in vitro than in vivo. In addition, after intravenous injection, microbubbles, (which are significantly larger than the pulmonary capillary vessel diameters) may be filtered by the pulmonary microcirculation, despite the fact that some small bubbles have a potential to grow after passing through the pulmonary microcirculation. The diameters of the transiently stopped QW7437 microbubbles’ images remained unchanged during stoppage and were slightly larger than the mean bubble size measured in vitro, although QW7437 bubbles which stopped in venules for a longer duration exhibited fluorescent quenching (loss in fluorescent brightness). Beppu et al. (6)have reported that an alternate DDFP-based agent, QW3600, can result in microbubbles as large as three to eight times the size of an erythrocyte in both an arteriole and a venule in rat mesentery, and can induce a gradual slowing and stoppage of blood flow. However, these observations were made after intravenous administration of QW3600 in doses that were much greater (2 to 4 ml/kg) than those used for QW7437, and were given without the application of negative pressure as used in this study. In recent clinical trials (7)in which QW3600 has been given in lesser doses (0.1 ml/kg or less) after hydrodynamic preactivation, the duration of contrast opacification has been substantially shorter than that reported in the animal studies without hydrodynamic preactivation technique (6). In addition, the negative electrical charge on QW7437 to some extent inhibits microbubble attachment to the vascular EC (19). Beppu et al. (6)and Main et al. (19)have suggested that the long residence of some microbubbles in microvessels with obstructed blood flow represented a possible explanation for the prolonged myocardial opacification by DDFP microbubbles. Our results suggest that long residence of some microbubbles in venules without obstructed blood flow also contributes to prolonged myocardial opacification.
The velocities of QW7437 microbubbles were significantly slower and manifested a wider spectrum of values than labeled erythrocytes in the venules with diameters of 12 μm and 13 μm. This difference in axial velocities was likely caused by the transient stoppage of some microbubbles at venular EC. A previous study (9)also reported that a few sonicated albumin microbubbles were observed to adhere to the EC of collecting venules. The interaction of erythrocytes with the relatively large microbubbles in capillaries and venules appears to resemble that of erythrocytes with leukocytes (21,22). Adhesion molecules on EC may have a role in the overall process of attachment (22–25). Iigo et al. (25)have reported that postcapillary venular EC are the major portion expressing intercellular adhesion molecule-1 among arterioles, capillaries and venules in the rat mesentery. Further research is needed to explore the molecular mechanisms for the phenomenon of microbubble adhesion to the vessel EC.
The primary goals of MCE are to enhance endocardial border definition and achieve myocardial opacification by intravenous injection. Dodecafluoropentane microbubbles may serve this purpose. Previous reports have shown that MCE performed with intravenous administration of DDFP microbubbles permits the definition of myocardial area at risk and infarct size in a canine model of regional myocardial ischemia (4–6, 19), and can enhance left ventricular endocardial delineation in humans (7,18). It may also be possible to apply indicator dilution techniques to quantitate coronary blood flow by MCE (26). Considering the retention of some DDFP microbubbles in venules, we could not simply use these methods with these agents to quantify coronary blood flow. Prolongation of contrast enhancement may be of clinical value in overcoming problems of signal attenuation associated with contrast echo, and in enabling sufficient time for multiple view examination. Perfluorocarbons have been applied as blood substitutes because of their strong oxygen-binding capacity and their high tolerance in mammals (27).
Several factors may have influenced the results of this study. The microvascular structure of the mesentery is not the same as in the myocardium or other organs in terms of the size and density of capillaries (28–30). Because it is still difficult to obtain continuous observations of the myocardial microcirculation in vivo (31), we selected a mesentery preparation allowing high resolution microvascular visualization that was essential for continuous and quantitative study of microbubble kinetics (28). In an in situ rat heart study (29), subepicardial capillary mean diameter has been found to be about 4 μm during systole and about 5 μm during diastole. The range of the mesentery capillary diameter examined in the present study was from 5 μm to 8 μm with a mean value of 6.2 μm. From this point of view, although no obstruction was seen with QW7437 microbubbles in mesenteric capillaries, relatively large bubbles may have a potential to transiently or permanently occlude smaller sized capillaries in the myocardium (7).
Although microbubble size in vivo is an important factor governing their kinetics in the microcirculation, such measurements were not carried out in our experiment except on adhering microbubbles. It was not possible to obtain accurate measurements of microbubble size in vivo from images of fluorescently labeled bubbles moving in vessels.
In conclusion, the velocity of the QW7437 microbubbles during their movement in postcapillary venules is different from that of the erythrocytes. About 10% of the microbubbles transiently stopped with an attachment to venular EC. The relatively long stability of microbubbles in the circulating blood and their extended residence in venules are potential mechanisms for the more persistent myocardial wall enhancement that may be seen with QW7437 microbubbles in contrast echocardiography.
We express our appreciation to Benjamin W. Zweifach, PhD for his critical advice on the manuscript. He died on October 23, 1997, shortly before this study was submitted for publication.
☆ Dr. Yasu was partly supported by grants from the Fukuda Memorial Foundation, Tokyo, Japan and from the Japan–North America Medical Exchange Foundation, Tokyo, Japan. Parts of this work was presented at the 70th Annual Scientific Sessions of the American Heart Association, Orlando, Florida, November 10, 1997.
- endothelial cells
- myocardial contrast echocardiography
- Received December 11, 1997.
- Revision received March 6, 1999.
- Accepted April 14, 1999.
- American College of Cardiology
- Ito H,
- Tomooka T,
- Sakai N,
- et al.
- Grayburn P.A,
- Erickson J.M,
- Escobar J,
- Womack L,
- Velasco C.E
- Grayburn P.A,
- Weiss J.L,
- Hack T.C,
- et al.
- Feinstein S.B,
- Shah P.M,
- Bing R.J,
- et al.
- Keller M.K,
- Segal S.S,
- Kaul S,
- Duling B
- Jayaweera A.R,
- Edwards N,
- Glasheen W.P,
- Villaneuva F.S,
- Abbott R.D,
- Kaul S
- Ismail S,
- Jayaweera A.R,
- Camarano G,
- Gimple L.W,
- Powers E.R,
- Kaul S
- Slezak S.E,
- Horan P.K
- Grayburn P
- Ley K,
- Cerrito M,
- Arfors K.E
- Villanueva F.S,
- Jankowski R.J,
- Manaugh C,
- Wagner W.R
- Iigo Y,
- Suematsu M,
- Higashida T,
- et al.
- Clark L.C Jr.,
- Gollan F
- Sobin S.S,
- Tremer H.M
- Nellis S,
- Toggart E