| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
 
|
|
|
||
|
|
Anesth Analg 2003;97:1789-1794 © 2003 International Anesthesia Research Society
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Cerebral lipid microemboli (LME) may cause postoperative cognitive dysfunction after orthopedic and cardiovascular surgery. In 13 anesthetized rats, we created a cranial window to study LME using orthogonal polarization spectral imaging videomicroscopy. All rats received 0.2 mL of human marrow fat, obtained from surgical waste during arthroplasty, injected into the superior vena cava. Five rats died within seconds of this injection, despite resuscitation efforts. Seven minutes later, we injected an additional 0.1 mL in 6 of the 8 surviving rats. We observed the videomicroscopy for 1 h in all 8 rats. Arterial blood pressure (BP) was continuously measured. No LME were observed in the first 7 min (n = 8); however, within seconds of the additional 0.1 mL injection, mean BP decreased from 79 ± 31 mm Hg to 28 ± 12 mm Hg (n = 6; P < 0.02). Epinephrine and crystalloid infusion increased BP to 161 ± 9 mm Hg and 20–100 LME were seen within 5 min. LME changed shape and fragmentation, erosion, and streaming patterns were noted, with transient arteriolar occlusion (10–220 s). Increasing BP resulted in reperfusion of occluded arterioles. No venous LME were noted. Postmortem, brain and lung LME were found with no patent foramen ovale. This model may be useful in studying cerebral LME.
IMPLICATIONS: Marrow lipid may pass through the lung during orthopedic surgery, creating cerebral lipid microemboli (LME). We created a cranial window in rats to study LME flowing through pial-cortical vessels. Cerebral LME appeared after resuscitation from hypotension and vessel occlusion was transient. This model may be useful in studying cerebral LME.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Postoperative neurological dysfunction occurs after orthopedic surgery (1,2). Symptoms range from subtle, subclinical changes in neurocognitive tests to stroke, confusion, or coma (1–3). The causes of neurological dysfunction are multifactorial, with hypoxia, hypoperfusion, and drugs proposed as etiologic factors (2). Lipid microemboli (LME) have also been implicated as a cause of neurological changes after orthopedic surgery and after cardiopulmonary bypass (CPB) (4,5). In both clinical situations, the source of LME is bone marrow, either from pressurization of long bones during total hip arthroplasty (THA) and intramedullary rod fixation of a fractured femur or from reinfusion of "shed" blood containing fat from the sternum during CPB.
Clinical reports of cerebral LME often rely on indirect evidence, such as transesophageal echocardiography (TEE) or transcranial Doppler ultrasound (6,7). Using TEE, Christie et al. (6) demonstrated echogenic material in the right ventricle in more than 90% of patients during medullary reaming for fresh fractures of the femur. A clinical report of postmortem brain histology in a patient who died after THA showed evidence of LME (8). Animal studies clearly showed that marrow-derived fat emboli can pass through the pulmonary vasculature within minutes to hours of cemented arthroplasty and are found in the brain (9). In this observational study, we demonstrate the dynamic characteristics of cerebral LME using videomicroscopy of rat pial-cortical vessels after injection of marrow fat into the superior vena cava.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This study was approved by the Institutional Animal Care Committee and conformed to the standards of the Canadian Council on Animal Care. The experiments were performed on 13 male Sprague-Dawley rats weighing 300–400 g. They were anesthetized using intraperitoneal ketamine hydrochloride (100 mg/kg), intubated, and the lungs were mechanically ventilated with oxygen (FIO2 = 1.0). Anesthesia was maintained using inhaled isoflurane (0.75%–1.5%). Ventilation was adjusted to maintain the partial pressure of carbon dioxide in arterial blood (PaCO2) between 35 mm Hg and 45 mm Hg. Body temperature was recorded using a rectal probe and maintained with a heating pad. Lidocaine 2% was injected subcutaneously on the ventral surface of the tail and a 22-gauge cannula was inserted into the tail artery for sampling of blood and continuous monitoring of mean arterial blood pressure (BP). The right internal jugular vein was cannulated using a 20-gauge catheter. This catheter was later used for the injection of drugs and marrow fat into the superior vena cava.
Each animal was placed prone, and the head was fixed in a stereotactic frame. The scalp was incised, and connective tissues were removed bilaterally. Over the right hemisphere, the temporal muscle was reflected, exposing the lateral aspect of the skull. Using a variable speed drill with a spherical burr (diameter, 2.5 mm), a groove was made in the skull approximately 10 mm long and 8 mm wide. Drilling continued until the bone was sufficiently thin to remove a cranial cap with fine serrated forceps. The dura mater was carefully incised, lifted, and reflected, exposing the cortical surface of brain and the pial-cortical vessels. The plastic optical probe of the videomicroscopy unit was inserted into the cranial window.
To visualize microvascular flow, we used an orthogonal polarization spectral (OPS) video microscope (CytoscanTM, Cytometrics, Philadelphia, PA). The OPS video microscope uses high-intensity light that is passed through a spectral filter to isolate light at a wavelength of 550 ± 35 nm (10). The light is polarized and focused on the target area. Reflected light passes through an orthogonal polarizer placed in front of a charge-coupled camera such that only the depolarized light reaches the camera. This results in an image that is similar to that seen in typical transmission intravital videomicroscopy. We used the 5x probe that has a x167 magnification. The probe was attached to a clamp and adjusted until it just made contact with the cerebral cortex. The probe was focused using the built-in stepper motor until we obtained a field in which both arterial and venous vessels were clearly visualized and normal flow was maintained. The video out signal from the CytoscanTM represents an area of 1 mm diameter and was displayed on a 12-in monitor and continuously recorded with a S-VHS videocassette recorder for subsequent offline analysis of images.
Human fat was salvaged from surgical waste during femoral reaming in patients undergoing THA procedures in the operating room. The liquid surgical waste was heparinized and centrifuged to remove particulate matter. The supernatant lipid was removed from the blood and stored at -18°C and then warmed to 37°C before injection.
We injected 0.2 mL of the liquid marrow fat into the superior vena cava of 13 rats. Only rats surviving 7 min after the initial injection of fat with no hemodynamic instability continued in the study. In 2 rats, we continued monitoring for 1 h with no further interventions. In the other surviving rats that did not suffer a cardiac arrest within 7 min, a second dose of 0.1 mL fat was given into the superior vena cava. Both injections were followed by a flush of 0.5 mL saline into the superior vena cava. Whenever BP decreased to less than 50 mm Hg, epinephrine diluted in saline (10 µg/mL) was given IV either as a bolus (maximum of 5 µg/kg) or as an infusion at a rate of 0.35 µg · kg-1 · min-1. When given as a bolus, the epinephrine was flushed through the superior vena cava cannula with 0.5 mL saline. Images of the cerebral microcirculation were recorded continuously in each experiment.
At the end of the study, the heart and lungs were removed "en bloc," fixed in inflation, and stained for fat with osmium tetroxide as previously described (9). The brain was also removed, fixed, and stained with osmium tetroxide to detect intravascular fat.
A specimen of the marrow fat was examined using electron microscopy to determine the nature of the injected material. The specimen was placed in agar cups, then processed using routine methods, including fixation in glutaraldehyde and osmium tetroxide followed by embedding in Spurr resin. Five one-micron plastic sections were stained with toluidine blue. Sections were stained with uranyl acetate and lead citrate and then examined using FEI CM100 and Tecnai 20 transmission electron microscopes (FEI, Matwah, NJ). Further examination was performed with energy filtered transmission electron microscopy using a GIF2001 energy filter (Gatan, Pleasanton, CA) in an attempt to reveal some detail of structure. This technique can filter the electrons in the beam according to how much energy they have lost as they pass through the specimen. Every element causes a specific energy loss, so this allows each element to be imaged separately.
A standard Leitz microscope scale of 1 mm divided into 100 divisions was scanned with the OPS video microscope. Areas of microvasculature demonstrating both arterioles and venules that were clearly visible were identified and selected before injection of fat. The scale and the recorded images were played back, and images of individual frames were captured using ATI software for sequential measurements and stored as TIFF images. The images were then opened in Photoshop (Adobe Systems, San Jose, CA), and the measuring tool, calibrated to the standard microscope scale, was used to measure vessel diameter and length without image enhancement. To determine velocity of the LME flowing through large vessels, we used the ATI software. Emboli were tracked from frame to frame permitting the calculation of the time required for LME to move from one location to another in subsequent frames. These frames were captured, and the distance the embolus traveled was measured and the diameter of vessel and velocity of embolus was calculated using Photoshop. Data are reported as mean ± SD. We used the SAS (Cary, NC) general linear model repeated-measures analysis of variance procedure to analyze sequential measurements. When a significant F-ratio was present (P < 0.05) between baseline and other measurements, we applied Dunnett’s and Tukey’s tests as appropriate.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Good visualization of the arteriolar and venous vessels of the pial-cortical microcirculation was obtained in all rats. The arterioles were distinguished from venules by the shape of vessel wall and pattern of flow (Fig. 1). The diameter of arteriolar vessels studied ranged from 14 µm to 104.5 µm (mean 65 ± 8 µm). Rectal temperature was maintained within the range 33.9°–37.8°C. Routine transmission electron microscopy was unable to detect any structure in the material injected. No detail or membrane structure was found in the amorphous lipid material examined using energy filtered transmission electron microscopy.
|
In all 13 rats, 0.2 mL of fat was injected. Five rats died within 7 min of injection and the resuscitation protocol did not re-establish stable circulation. These experiments were terminated, and no video images of LME were obtained. Of the 8 rats that survived the initial injection, none required the resuscitation protocol and all maintained BP > 40 mm Hg for 7 min. Two of these eight rats, who did not require any epinephrine or crystalloid for resuscitation, were followed for 1 h and maintained BP >100 mm Hg. Neither required any epinephrine or crystalloid for resuscitation (Table 1). In these two rats, only one circular LME was noted on videomicroscopy. This was noted 23 min after the injection of fat (0.2 mL) into the superior vena cava, and it rapidly passed through the field of vision without obstructing any vessel.
|
In 6 of the 8 rats that maintained BP > 40 mm Hg for 7 min after the 0.2-mL injection, we injected a further 0.1 mL of marrow fat into the superior vena cava. There was a decrease in BP from 120 ± 17 mm Hg to 79 ± 31 mm Hg (P < 0.02) after the first 0.2-mL injection, then a decrease to 28 ± 12 mm Hg within 2 min (Table 1). The decrease in BP was accompanied by a visible slowing of flow in the pial-cortical arterioles; however, no LME were noted in any rat at this phase of the study. The resuscitation protocol (epinephrine bolus and 0.5-mL crystalloid flush) was begun immediately when BP decreased to 50 mm Hg. The resuscitation protocol resulted in an increase in BP from 28 ± 12 to 161 ± 9 mm Hg (P < 0.01). Arterial PaO2 decreased from 339 ± 60 mm Hg to 75 ± 20 mm Hg (P < 0.02) within 2 min of the second fat injection.
More than 5 min after institution of the resuscitation protocol, between 20 and 100 LME per experiment were observed in the pial-cortical arterioles on videomicroscopy. The velocity of LME flow at these high perfusion pressures (BP > 110 mm Hg) was highly variable; however, in one example (Fig. 1), it was 0.63 mm/s. Within minutes the epinephrine-induced hypertension subsided, and we observed LMEs occluding arterioles in all 6 rats. Complete occlusion of flow by LMEs was recorded for periods of time ranging from 10 s to 220 s. In each experiment, epinephrine was given to treat hypotension (BP < 50 mm Hg). This resulted in increased BP, and the LMEs (viewed on OPS videomicroscopy) moved further along the vessel and out of view. Blood flow through the previously obstructed vessels was restored.
At no time were any formed LMEs noted in venous vessels. The shape of LME varied according to the diameter of the arteriole. Smaller LME, with a diameter less than the internal diameter of the vessel, appeared as circular intravascular white objects (Fig. 1A) and traveled rapidly across the field of vision (Fig. 1B–D). In contrast, larger LME often appeared as elliptical objects moving at variable, slower speeds (Fig. 2). These elliptical LMEs occluded the lumen of the arteriole (Fig. 2A).
|
We examined the shape of LMEs as they passed from larger arterioles into smaller vessels, and three distinctive patterns were identified. Elliptical LMEs (Fig. 2A) were often noted to divide into smaller fragments at bifurcations in the arteriolar tree (Fig. 2B). When associated with hypotension, this fragmentation resulted in occlusion of smaller vessels with cessation of flow. We also noted the appearance of intravascular streaming in elliptical LMEs. The "streaming" is identified as a distinct "tail" protruding in the direction of flow (Fig. 2A) "in front" of the embolic LME. In contrast, when a round LME was caught in a vortex of flow, usually at a bifurcation of an obstructed arteriole, we noted continuous erosion of the LME, with decreasing diameter until it disappeared. Such physical changes in LME shape were noted in all 6 rats in which LME were identified in large numbers.
On histologic examination, a large amount of intravascular fat was found in the lungs of each rat (Fig. 3). Similarly, there was a large amount of postmortem intravascular fat (stained black) in the pial-cortical vessels of the brain (Fig. 4). In the two rats in which only a total of 0.2 mL of fat was injected, there was visibly less fat noted in the brain at postmortem examination than in those given 0.3 mL of fat. No patent foramen ovale or intracardiac defect was found at postmortem in any rat.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This was an observational study describing the visualization of cerebral LME using OPS videomicroscopic imaging in a rat model. We observed LME passing through the pial-cortical vessels (Figs. 1, 2), confirming that human marrow fat, after injection into the superior vena cava, passes through the lung and can be detected in the brain in rats (Fig. 4). LME were visualized in the pial-cortical vessels only after the injection of epinephrine and crystalloid to treat systemic hypotension. This suggests that one effect of the resuscitation protocol was enhanced transpulmonary passage of fat. Whether this is attributable to increased pulmonary blood flow (cardiac output), opening of pulmonary vessels (vascular recruitment), simply increased pulmonary artery pressure, or a direct effect of epinephrine remains conjectural. Once the epinephrine-induced hypertension subsided, the LME created transient cessation of flow in pial-cortical arterioles. Flow was restored by further increasing BP. Although this vascular occlusion was transient, decreased perfusion pressure and hypoxemia are often associated with cardiopulmonary instability during intraoperative fat embolism. Under specific conditions, this study suggests that it is possible that LME may contribute to postoperative neural injury.
Drew et al. (11) studied cerebral fat emboli by using triolein injected into the carotid artery. The triolein creates a large area of infarction, a stroke model. In our model, the lungs filter and disperse the amorphous globule of marrow fat into microemboli (Fig. 4), more closely simulating the clinical findings after orthopedic and cardiovascular procedures (5,8). Colonna et al. (8) reported a patient who died approximately 4 hours after insertion of a femoral prosthesis, despite cardiopulmonary resuscitation and epinephrine administration. At autopsy, there were approximately 5000 fat emboli/cm3 of brain tissue. As in our model, a patent foramen ovale was not present, suggesting that large numbers of LME traversed the pulmonary circulation. The volume of fat injected in our study was intended to produce cardiopulmonary instability, requiring inotropic support (8,12). This suggests that the amount of fat given to these rats would be representative of a large LME load in clinical practice.
Several patterns of LME flow in the pial-cortical vessels were identified in these rats that may or may not have pathophysiological significance to patients after orthopedic or cardiovascular surgery. Visualization of LME passing through the pial-cortical vessels gives a more dynamic impression of what happens than the single "snapshot" obtained at postmortem (Fig. 4). Our study shows that these LME are not static but rather move at variable speeds through the arteriolar and capillary vessels, depending on local factors. In our model, the most important factor appeared to be the perfusion pressure. However, the effect of other variables such as PaCO2 should be assessed.
There are many limitations to this model. For example, incision of the dura and brain exposure may alter tissue pH and cause injury and swelling of brain tissue, thus influencing capillary dynamics. Pressure exerted on the surface (pial-cortical) vessels by the plastic optical probe of the videomicroscopy unit could impede local blood flow. The fact that we observed LME in 7 of 8 rats that survived for 1 hour suggests that these factors did not seriously impede flow. The effectiveness of increasing BP on reperfusion suggests that this is an important element of resuscitation after fat embolism and enhances transcapillary LME passage in the brain. We observed an area of brain tissue of only 1 mm diameter; therefore, quantification of total LME load or regional distribution was not possible. Although we observed LME in the pial-cortical vessels, we noted large numbers of LME throughout the brain at postmortem examination (Fig. 4). In the 2 rats with the smallest quantity of fat injected (0.2 mL), only one LME was observed on OPS video microscopy and few intravascular fat emboli were seen on postmortem histologic examination. This suggests that although the OPS video microscopy gave a general indication of quantity of LME, regional differences in brain circulation and susceptibility to ischemia do exist. Attempts to extrapolate from our observations on pial-cortical vessels to other areas of the brain may not be appropriate. Similarly, there may be limitations of the model because of species differences. The fat injected was collected from human surgical waste during THA. We are assuming that the rat lung filters human marrow fat in a similar fashion to the human lung. The similarity of our histologic findings to human case reports and other animal models are striking (8,9). This suggests that the small capillary arteriolar dilations previously described are LME lodged in the microcirculation during transit through the cerebral circulation (8,13).
Blood flow to the pial-cortical circulation is relatively rapid, and there are many anastomoses. Such a vascular network is not present in deeper areas of the brain. The pattern of blood flow to the inner white matter depends on perforating vessels, and the dynamic characteristics of LME in these areas of the brain may be much different. Similarly, we did not describe the ultimate fate of LME in the brain. As we did not observe any formed globules in the venous microcirculation, we do not know if the LME observed were ultimately trapped in more distal vessels or dispersed into fine fragments that are too small to visualize on OPS video microscopy. We speculate that the dynamic characteristics of LME (streaming, fragmentation and erosion) that we observed (Fig. 2) are actually mechanisms by which the LME are broken up into fine lipid droplets that traverse the microcirculation. The "streaming" (described as a tail in the direction of flow) (Fig. 2A) may be caused by the velocity gradient across the arteriole resulting from friction at the endothelial-LME interface. There may be other processes in addition to erosion, fragmentation, and streaming that ultimately destroy LME; however, this was beyond the scope of this study. Under normal circumstances, these processes may occur in the lung and brain over time. This would account for the rarity of major neurocognitive dysfunction in patients despite the fact that more than 90% have echogenic evidence of emboli (6). In this study, it was only when epinephrine was given that LME were noted in the pial-cortical vessels.
Magnetic resonance imaging (MRI) techniques demonstrate cerebral edema after fat emboli (1), suggesting that areas of injury are primarily located in the subcortical white matter. Importantly, our acute study was not able to demonstrate any cerebral tissue injury associated with LME. It is possible, however, that LME in low-flow regions of the brain can cause significant ischemia-reperfusion injury contributing to leakage and perivascular edema noted on MRI. In summary, we are not aware of another study demonstrating in vivo LME in pial-cortical vessels after passage through the lung. This new rat model, using OPS imaging, allows the systematic study of dynamic characteristics of cerebral LME. OPS imaging has been used in humans to measure flow in the brain and in muscle and sublingual tissue (10), although not in clinical situations associated with LME. This model, and the OPS technology, may provide a method for studying cerebral ischemia-reperfusion injury from LME, leading to a better understanding of the mechanism of cerebral dysfunction after orthopedic and cardiac surgery.
|
Acknowledgments |
|---|
Supported, in part, by a grant from St. Michael’s Hospital Research Society, Toronto, Ontario.
The authors acknowledge the funding provided by St. Michael’s Hospital Research Foundation and the assistance of Dr. Howard Nathan (University of Ottawa) and Dr. Cheryl King (Queens University) who reviewed the manuscript during preparation. We also thank Dr. James Waddell, Department of Orthopedic Surgery, who procured the marrow fat from surgical waste.
|
Footnotes |
|---|
Supplemental material available at www.anesthesia-analgesia.org.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
- Takahashi M, Suzuki R, Osakabe Y, et al. Magnetic resonance imaging findings in cerebral fat embolism: correlation with clinical manifestations. J Trauma 1999; 46: 324–7.[Web of Science][Medline]
- Berggren D, Gustafson Y, Eriksson B, et al. Postoperative confusion after anesthesia in elderly patients with femoral neck fractures. Anesth Analg 1987; 66: 497–504.
[Abstract/Free Full Text] - Jacobson DM, Terrence CF, Reinmuth OM. The neurologic manifestations of fat embolism. Neurology 1986; 36: 847–51.
[Abstract/Free Full Text] - Robinson CM. Current concepts of respiratory insufficiency syndromes after fracture. J Bone Joint Surg Br 2001; 83: 781–91.
- Brown WR, Moody DM, Challa VR. Cerebral fat embolism from cardiopulmonary bypass. J Neuropathol Exp Neurol 1999; 58: 109–19.[Web of Science][Medline]
- Christie J, Robinson CM, Pell ACH, et al. Transcardiac echocardiography during invasive intramedullary procedures. J Bone Joint Surg Br 1995; 77: 450–5.
- Edmonds CR, Barbut D, Hager D, Sharrock NE. Intraoperative cerebral arterial embolization during total hip arthroplasty. Anesthesiology 2000; 93: 315–8.[Web of Science][Medline]
- Colonna DM, Kilgus D, Brown W, et al. Acute brain fat embolization occurring after total hip arthroplasty in the absence of a patent foramen ovale. Anesthesiology 2002; 96: 1027–9.[Medline]
- Byrick RJ, Mullen JB, Mazer CD, Guest CD. Transpulmonary systemic fat embolism: studies in mongrel dogs after cemented arthroplasty. Am J Resp Crit Care Med 1994; 150: 1416–22.[Abstract]
- Groner W, Winkelman JW, Harris AG, et al. Orthogonal polarization spectral imaging: a new method for study of the microcirculation. Nature Med 1999; 5: 1209–13.[Web of Science][Medline]
- Drew PA, Smith E, Thomas PD. Fat distribution and changes in the blood brain barrier in a rat model of cerebral fat embolism. J Neurol Sci 1998; 156: 138–43.[Medline]
- Byrick RJ, Korley RE, McKee MD, Schemitsch EH. Prolonged coma after unreamed, locked nailing of femoral shaft fracture. Anesthesiology 2001; 94: 163–5.[Medline]
- Brooker RF, Brown WR, Moody DM, et al. Cardiotomy suction: a major source of brain lipid emboli during cardiopulmonary bypass. Ann Thorac Surg 1998; 65: 1651–5.
[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
 |
|
  M. Barak, M. Kabha, D. Norman, M. Soudry, Y. Kats, and S. Milo Cerebral Microemboli During Hip Fracture Fixation: A Prospective Study Anesth. Analg., July 1, 2008; 107(1): 221 - 225. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
