| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institute for Surgical Research, University of Munich, 81377 Munich, Germany
Submitted 10 April 2003 ; accepted in final form 13 June 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
liver
However, there are two organs in which not only the nutritive but also the functional circulation is of relevance; i.e., the lung and the liver. Functional pulmonary blood flow can easily be measured by injecting the MS intravenously (10, 29). MS are carried by the venous blood into the right ventricle, where they mix homogeneously with the blood before entering the lung. In contrast, blood flow through the portal vein is laminar, preventing any uniform distribution of MS in the blood when directly injected into the portal vein.
However, portal blood flow is known to be influenced by a variety of physiological [e.g., nutrition, (11)] and pathophysiological processes [e.g., burns (30) and sepsis (18)]. Therefore, the knowledge of alterations of regional portal blood flow (RPBF) and its distribution in the liver under various conditions is of relevance. Up until now, MS have only been employed to estimate RPBF by relating the total arterial splanchnic blood flow to the liver weight (1, 4, 14). Other techniques, such as laser Doppler ultrasound (2, 23, 34), extraction of radioactively labeled colloids (7-9), and computer tomography (25), have been used to measure RPBF but have disadvantages, such as the low resolution of these methods or the impossibility of repeating the measurements.
Our hypothesis was that the disadvantages of the methods listed above can be overcome by a new experimental setup, allowing the repeated determinations of RPBF by means of FM, as well as the analysis of the distribution of portal blood flow within the liver.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
A mixing chamber was designed (Fig. 1) to create turbulent streaming of the blood after being inserted into the portal vein. The chamber consists of Perspex, which was carefully polished to achieve a maximum plane surface. The chamber can be opened at one side to allow cleaning and disinfecting. Closure of the opening is achieved by an aluminum plate, which is covered by a silicone layer.
|
In in vitro experiments, the inlet and outlet of the chamber were connected to a silicone tubing (inner diameter 10 mm, length 100 cm each; Ismatec, Wertheim, Germany). Tubing and chamber were perfused with saline without recirculation by means of a roller pump (Stoeckert Instrumente, Munich, Germany) at different flow rates (100, 300, 500, 800, and 1,200 ml/min). Turbulence was visualized by injecting ink into the chamber.
A second series of in vitro experiments served to investigate the uniformity of dispersion of FM in the blood at different flow velocities. Porcine blood collected at the Munich slaughterhouse was used to perfuse the tubing and the chamber at the lowest and highest flow rates given above. Four connectors were integrated at varying distances (17, 29, 50, and 78 cm apart from the chamber) into the tubing so as to allow catheters (inner diameter 1.0 mm, length 10.0 cm; Cavafix, B. Braun Melsungen, Melsungen, Germany) to be introduced. These catheters served to collect reference blood samples by means of two separate pumps (Harvard Apparatus) at a constant withdrawal rate of 3.24 ml/min. Sodium citrate (5 ml; B. Braun Melsungen) was drawn into the collecting syringes before blood withdrawal. FM (Molecular Probes, Leiden, The Netherlands, colors: blue, blue-green, yellow-green, orange, red, crimson, scarlet) of 15-µm diameter were vortex mixed (2 min), sonicated (5 min), and again vortex mixed (1 min). Approximately 1 x 106 FM were diluted with 19 ml of saline to a final volume of 20 ml. The Harvard pumps, each of them capable of collecting two reference blood samples, were started at that time point, and FM were injected into the chamber over a period of 1 min. Withdrawal of reference blood samples was continued for a further 2 min after completion of FM injection. The syringes were weighed individually before and after blood withdrawal to allow calculation of the volume collected. The procedure described above was repeated six times with FM of one color at the lowest (100 ml/min, FM yellow-green) and highest (1,200 ml/min, FM red) flow rate, respectively. FM were recovered from the reference blood samples by means of the sample processing unit (Gaiser Kunststoff und Metallprodukte, Kappel-Grafenhausen, Germany), as described by Raab et al. (22). The filters containing the FM were processed by a robot, and the extracted fluorescence was measured by a luminescence spectrometer (LS 50 B, Perkin Elmer, Rodgau, Germany) (33). Finally, the fluorescence intensities of each reference site and each FM species were related to the collected blood volume, and volume fluorescence values were compared.
In Vivo Experiments
Measurement of RPBF. The experiments were authorized by the local Bavarian government (no. 3-29/96) and approved by the institutional animal care and use committee.
Pigs (n = 6, 23.5 ± 2.9 kg body wt) of the German Land-race of either sex were anesthetized by an intramuscular injection of midazolam (0.75-1.5 mg/kg body wt; Dormicum, Hoffmann La Roche, Grenzach-Whylen, Germany), azaperon (25 mg/kg body wt; Stresnil, Janssen-Cilag, Neusss, Germany), ketamine (25 mg/kg body wt; Ketavet, Pharmacia & Upjohn, Erlangen, Germany), and 0.5 mg of atropine (B. Braun Melsungen). Next, a catheter (Insyte-W, 20 GA, Becton Dickinson, Madrid, Spain) was introduced into an ear vein for administration of drugs. The animals were fixed in the supine position on a heated (37°C) table, relaxed by an intravenous injection of 4 mg of pancuronium bromide (Curamed, Karlsruhe, Germany), intubated, and mechanically ventilated (Servo 900, Draeger, Luebeck, Germany). Ventilator settings were adjusted to maintain normocapnia. Anesthesia was continued by continuous infusion of midazolam (30-60 mg/h) and fentanyl (0.47-0.94 mg/h; Janssen-Cilag) via a catheter (16 gauge) placed in the right jugular vein. Arterial blood pressure was monitored via a catheter (5 Fr) inserted into the right carotid artery. The catheter was connected to a pressure transducer (DTX, Becton Dickinson) linked to a computer.
The abdomen of the animals was opened by a midline incision. To minimize irritation of the intestine during surgical preparation, drapes were fixed to the wound edges by clamps. The drapes were then tensed upward and sideways to give enough space for the surgical procedure with minimal manipulation of the intestine. If the stomach was overfilled, it was punctured, emptied by suction, and closed again by means of a circular suture. The common bile duct was then cannulated for the collection of bile over an initial period of 10 min. The volume of bile collected at this time point was regarded as a parameter of liver function and served as a control value for the comparison of bile flow values obtained at later time points.
Next the portal vein was freed from connective tissue, taking care to harm the surrounding tissue as little as possible. To evaluate the influence of the surgical preparation on liver function, bile was collected for a further period of 10 min.
The animals were anticoagulated by an intravenous injection of 15,000 units of heparin (Braun Melsungen) 5 min before occlusion of the portal vein. The vessel was completely occluded close to the liver by two vascular clamps and incised in between the clamps. For implantation of the mixing chamber described above, two silicone tubes (inner diameter of 10 mm, length of 20 cm each; Ismatec) were introduced (one retrograde, the other antegrade) through the incision and fixed by ligatures. The tube placed close to the liver had a connector integrated (Fig. 1), allowing the operator to introduce and advance (10-15 cm) a catheter (inner diameter 1.0 mm, length 10.0 cm; Cavafix, B. Braun Melsungen) for reference blood sampling. The open ends of the tubes were then connected to the inlet and outlet of the mixing chamber, respectively. After the clamp was opened close to the liver, the tubing and chamber were emptied via the connector. Portal reperfusion was established by opening the clamp proximal to the direction of flow. During the time of portal occlusion, 6% HAES solution (200-300 ml; Fresenius Kabi, Bad Homburg, Germany) was infused via the jugular vein to stabilize the systemic hemodynamics. The duration of portal occlusion as well as arterial blood pressures before, during, and after occlusion of the portal vein were recorded. After reestablishment of portal flow, another bile collection was performed to assess the influence of the maneuver on the liver. After completion of the surgical exposure, the abdominal organs were covered with a plastic film and a heating pouch so as to maintain the species-specific intraabdominal temperature of 39.0°C.
A period of 45 min of stabilization was allowed before the first injection of FM.
Preparation and injection of MS. To analyze the uniformity of MS deposition in the liver tissue, FM of two different colors were injected simultaneously at three different time points. FM stock solutions were prepared as described in In Vitro Experiments. FM of two different colors (5 x 105 each = 0.5 ml of stock solution) were withdrawn from the glass vials by means of a syringe, diluted with 9.0 ml of saline, and mixed by mechanically shaking the syringe. The syringe was connected to the injection port of the mixing chamber, and the Harvard pump (withdrawal rate 3.24 ml/min), used for reference blood sampling, was started.
The injection of FM was begun when blood appeared in the reference catheter (
30 s after starting the pump). FM were injected continuously over a period of 1 min by the same person in all of the experiments to avoid interindividual variations. Thirty minutes were allowed to elapse before the subsequent injection.
Arterial blood pressure was recorded before and after each injection of FM.
After each FM injection, bile was collected for a period of 10 min, and a sample of systemic arterial blood was taken for laboratory investigation.
Organ dissection and recovery of MS. The animals were killed by an intravenous overdose of potassium chloride (Braun Melsungen) at the end of the experiments, and the liver was removed. The liver was placed into 10% formalin for fixation for a period of at least 1 wk.
The porcine liver's anatomy and its vascularization differ from that of the human liver. In contrast to the human liver, the porcine liver consists of three lobes: right, left, and median lobe, with the median lobe being separable into two sections along the ligamentum teres (19). Each of these subunits (left and right as well as left median and right median lobe) are entered by an individual branch of the portal vein. These branches are highly variable in their distribution. Dissection of the porcine liver according to its anatomic structures, therefore, appears to be more reasonable than dissection into segments.
Superficial connective tissue as well as the gall bladder were removed. The livers were then dissected into the left, median, and right lobe. In addition, the median lobe was separated along the ligamentum teres into the left and right median lobe. The individual sections were then dissected horizontally into the diaphragmatic layer (close to the diaphragm), a central layer (from the middle of the organ), and the visceral layer (close to the hepatoduodenal ligament). Recovery of FM from liver samples heavier than 4 g is very time consuming, as the digestion period has to be prolonged with increasing sample weight. Therefore, care was taken to keep the weight of the specimens <4 g during dissection of the layers mentioned above. An illustration of the dissection method is given in Fig. 2. Immediately after dissection, the weight of the individual tissue samples was determined and recorded.
|
Shunting of FM through the portal microcirculation would lead to an underestimation of RPBF. However, the lung has shown itself to be an organ capable of retaining almost 100% of recirculating spheres (5). FM shunted through the liver would, consequently, be trapped in the lungs. To verify the amount of FM shunted through the liver, we removed the lungs from three of the animals, dissected them into cubes of 2-3 cm3 (n = 285), and analyzed them for the presence of FM.
FM from liver and lung samples were recovered by a method described elsewhere (22, 33).
Parameters evaluated. Besides regional blood flow values, the following parameters were assessed during the experiments.
Mean arterial blood pressure (MAP) was recorded after cannulation of the carotid artery; after isolation; before, during, and after occlusion of the portal vein; and before and after each injection of FM.
Bile was collected for a period of 10 min before and after isolation of the portal vein and after the individual injections of FM.
Arterial blood samples were taken after cannulation of the carotid artery, after isolation and after occlusion of the portal vein, as well as after each injection of FM. The blood samples served to determine the following parameters: hemoglobin, hematocrit, 
-glutamyl transferase, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, lactate dehydrogenase, cholinesterase (CHE), urea, total bilirubin and albumin, as well as white and red blood cells and platelets counts.
Statistical Analysis
Data were analyzed by means of one-way ANOVA. To isolate values significantly differing from control, Dunnett's method was applied. A P value < 0.05 was considered to indicate a statistically significant difference.
Blood flow values obtained by simultaneously injected FM were compared by means of the Whitney rank-sum test and by the method of Bland and Altman (3). In addition, blood flow values of the individual injections were correlated, and a regression analysis was performed.
Tissue samples containing fewer than 400 FM were regarded as nonvalid specimens (5) and were, therefore, excluded from the analysis.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Injection of ink into the mixing chamber perfused with saline revealed three different regions of turbulence inside the mixing chamber (Fig. 1). The shape of the turbulences did not show any flow rate-dependent alterations. Comparison of mean fluorescence intensities of the samples collected at the four different withdrawal sites revealed no differences. Mean fluorescence intensity per milliliter of blood was 29.4 ± 8.0, 26.8 ± 8.5, 28.8 ± 8.3, and 28.6 ± 7.1 (n = 6 each) at the different withdrawal sites when the tubing was perfused with blood at a rate of 100 ml/min and 11.4 ± 1.7, 9.6 ± 1.9, 11.9 ± 2.4, and 10.2 ± 2.7 (n = 6 each) at a perfusion rate of 1,200 ml/min. These data indicate homogeneous dispersion of the FM in the blood by the mixing chamber.
In Vivo Experiments
The median duration of portal occlusion was 9.0 ± 3.1 min. A period of 45.5 ± 6.0 min elapsed between portal reperfusion and the first injection of MS. Before the second and the third injection of FM, 26.0 ± 6.1 and 23 ± 7.5 min of stabilization were allowed, respectively.
The mean weight of the formalin-fixed livers was 513 ± 124 g. Relation of the liver weight to the body wt yielded a highly constant ratio of 2.1 ± 0.4%.
Hematological and Hemodynamic Parameters
Cell counts revealed no significant decrease of platelets and white and red blood cells at all of the time points evaluated. MAP decreased from 68 ± 15 mmHg (control) to 61 ± 8 mmHg during portal occlusion. This decline was not statistically significant (Table 1). After establishment of the bypass and portal reperfusion, MAP increased significantly to a value of 76 ± 5 mmHg (P = 0.024). This increase was not significant compared with control values. No significant changes were noted before and after the individual injections of FM (Table 1).
|
No bleeding from the portal vein or any entering blood vessels occurred during the experiments. Consequently, the red blood cell counts, hemoglobin concentration, and hematocrit were unchanged (Table 1).
Liver Enzymes and Function
The serum concentration of the enzymes investigated (Table 2) remained within the normal range and showed no changes compared with control values at any time points of measurement. Only CHE activity decreased significantly (P = 0.009) after the second injection of FM, indicating some impairment of liver function. This finding is paralleled by a significant decrease (P = 0.004) of the serum concentration of albumin at the same time point of measurement. However, the serum concentrations of CHE and albumin both remained within the normal range throughout the experiments, thus indicating that the surgical preparation as well as the injection of FM did not affect the integrity of the livers. This observation is confirmed by the finding that the measurement of neither serum bilirubin nor urea showed major alterations at any time. Additionally, production of bile remained unchanged until the end of the experiments. There was a minor decrease in the production of bile from 0.4 ± 0.05 ml/min (control) to 0.35 ± 0.07 ml/min after isolation of the portal vein. From then on bile flow remained almost unaltered.
|
Comparison of Regional Blood Flow Values
Dissection of the six livers yielded 1,081 tissue samples with a mean weight of 2.85 ± 0.8 g. Blood flow values of 40 tissue samples were lost, owing to a technical problem with the luminescence spectrometer. In addition, blood flow values obtained from samples containing <400 FM were excluded from the analysis, as suggested by Buckberg et al. (5). Thus 5,458 blood flow values remained for the analysis from the three simultaneous injections of FM.
Comparison of all blood flow values (Table 3) revealed close correlation (r = 0.97 ± 0.01, range from 0.96 to 1.00), indicating homogeneous distribution of FM in the tissue.
|
In only one experiment did we use FM of the color scarlet, a color known to have higher variation of fluorescence (28) than the others (experiment 4, injection 1). The correlation of the blood flow values from this injection was the lowest of all of the experiments (r = 0.96). Analysis of the blood flow values (n = 5,156) in the absence of the results of injection 1 from experiment 4 gave better results (r = 0.99, range 0.97-1.00). The correlation plots are shown in Figs. 3 and 4.
|
|
The Mann-Whitney rank sum test revealed no significant differences between the blood flow values (n = 5,458) determined by simultaneously injected FM of two different colors.
Analysis of the values by means of the method of Bland and Altman showed a bias of 2.5% and a precision of 12.7% (3.6 ± 11.1% without injection 1 in experiment 4). Data are presented in Figs. 5 and 6.
|
|
The median of the blood flow values measured by means of FM of two different colors at each simultaneous injection was calculated so as to detect any changes that had occurred between the different time points of measurement. Median RPBF obtained from the first injection of MS was 1.3 ± 0.84 ml · min-1 · g liver tissue-1 and thus significantly higher (P < 0.001) than at the second and third injection of FM (1.0 ± 0.7 and 1.0 ± 0.8 ml · min-1 · g-1). No significant difference was found between the blood flow values obtained from the second and third injection.
No fluorescence could be detected in any of the tissue samples from the three lungs analyzed, indicating that the loss of FM from the liver through shunting was negligible.
Distribution of Blood Flow in the Liver
The calculation of blood flow values obtained by simultaneous injection of FM of two different colors yielded a median RPBF of 1.07 ± 0.78 ml · min-1 · g-1 (n = 5,458).
To analyze regional distribution of blood in the liver, mean blood flow values obtained by simultaneously injected FM were calculated and related to the different locations of the samples according to the dissection scheme. The dissection scheme allowed us to compare blood flow distribution between the different lobes as well as in a layer of the liver close to the diaphragm (diaphragmatic layer), a layer from the central part of the organ, and a layer close to the stomach (visceral). Median blood flow values in the right liver lobe and the right and left part of the median lobe (1.05 ± 0.79, 1.08 ± 0.86, 1.10 ± 0.67 ml · min-1 · g-1) were significantly higher (P = 0.001) compared with the values obtained from the left lobe (1.01 ± 0.68 ml · min-1 · g-1). In addition, blood flow values of the diaphragmatic layer of the liver (1.21 ± 0.90 ml · min-1 · g-1) were significantly higher (P = 0.001) compared with the central (1.05 ± 0.73 ml · min-1 · g-1) and the visceral layer (0.99 ± 0.68 ml · min-1 · g-1).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
None of the liver enzymes' serum concentration exceeded the physiological range (16, 32) at any time point of measurement, indicating that neither the surgical preparation nor the portal occlusion caused any damage to the organ. The serum concentrations of albumin and CHE differed significantly from baseline values after the second and third injection of MS, indicating an impact on the liver function. Still, the values of these markers remained within the physiological range at all time points of measurement. We, therefore, conclude that the surgical preparation and the injection of FM did not influence the liver function adversely.
Hemoglobin concentration and hematocrit appeared to be low, even at the beginning of the experiments. However, pigs of the age used in the present study pass through a physiological anemia so that these low concentrations are common (36). No bleeding occurred during the isolation of the portal vein or during implantation of the mixing chamber. Thus the declining, yet statistically not significant, values of hemoglobin and hematocrit can be explained by the hemodilution induced by the infusion of HAES solution during portal occlusion.
The fact that no loss of cells or hemolysis was observed in the blood samples collected indicates that the contact of the blood with the artificial surfaces of the tubing and the mixing chamber did not influence the blood or its individual components.
Measurement of RPBF
FM have become an attractive alternative to radioactive MS since they were first introduced by Glenny et al. (9) in 1993. They allow research workers to determine regional organ perfusion reliably as long as the spheres are delivered to the organ in proportion to the blood. Therefore, a prerequisite of the method is that the MS have to disperse in the blood homogeneously. In experiments in which regional arterial blood flow of any organ is determined, homogeneous dispersion of FM in the blood is achieved by injecting the FM into the left atrium or ventricle where they are distributed uniformly in the blood by the turbulent streaming. However, measurement of RPBF of the liver cannot be performed reliably by simply injecting the FM into the portal vein, as the laminar flow profile at this particular injection site will not ensure adequate mixing of the spheres in the blood. To our knowledge, estimation of portal blood flow by means of MS has only been performed by relating the measured arterial splanchnic blood flow to the liver weight (1, 4, 14, 35). However, estimation of portal blood flow in this manner does, of course, not allow us to allocate blood flow values to different anatomic regions of the liver. Numerous other techniques have been applied to measure portal blood flow, such as positron emission tomography (21), extraction of indocyanine green (20, 26), wash-out of xenon-133 (25, 28), and the use of ultrasonic flow probes (24). All of these methods give reliable measurements of portal or total hepatic blood flow, but they do not give any information about the distribution of the portal blood within the organ.
The results of the present study show that regional portal blood distribution can certainly be measured by means of FM, because homogeneous dispersion of FM in the blood can be achieved by using an adequate mixing chamber. This is also obvious from the high correlation and regression coefficients, as well as by the low percent difference of the blood flow values determined by simultaneously injected FM in the animal experiment.
Prevention of recirculation of spheres that were not entrapped in the capillary field is a second prerequisite of the method and is necessary to achieve reliable measurements. In experiments in which the MS are injected systemically, recirculation of spheres is prevented by the lungs, which retain almost all MS delivered by the arterial or venous blood. As we intended to use FM in models of isolated liver perfusion, we had to investigate whether FM can pass through the portal microcirculation or are released after being entrapped. Therefore, we analyzed the lungs of three of the animals for the presence of fluorescence. In none of the lung tissue samples was fluorescence detected, indicating that the portal microcirculation, like the lung, quantitatively retains all of the MS entering the liver by the portal blood.
One hundred seventy-one liver tissue samples had to be excluded from the analysis, 40 owing to a technical defect of the spectrometer used for the measurement of fluorescence, and 131 because these specimens did not contain a sufficient amount of FM. Most of these samples were located along the ventral edge of the liver lobes. However, we do not think that the location of the specimens containing <400 FM reflects a region of physiological lesser perfusion, because most of the samples excluded were very small in size (<0.5 g), thus being too small to receive a minimum of 400 FM.
FM of the color scarlet were used in one of the experiments. This color has been reported to have a higher variation in fluorescence compared with the other colors available (27). We, therefore, recommend not to use FM of the color scarlet for measurement of regional blood flow, unless really necessary.
Portal blood flow in mammals is reported to be 1.0 ml · min-1 · g liver tissue-1 (12). Gudmundsson et al. (13) measured total portal blood flow by means of transit time flow probes and related the blood flow to the liver weight. They reported portal blood flow values of 1.2 ml · min-1 · g-1. The values measured in our study (1.07 ± 0.78 ml · min-1 · g-1) are almost identical to theirs, proving the accuracy of the measurement. Total portal blood flow (summation of the individual RPBF values) from our study is in the range reported by other groups (15, 18, 30, 31) who measured total portal blood flow in pigs of the same body wt by means of ultrasonic flow probes. This again confirms the validity of our measurement.
During portal occlusion, the intestine was still arterialized, leading to a congestion of blood in the splanchnic area. The congested blood was released from the splanchnic organs on reestablishment of the portal perfusion, leading to a significantly higher median portal blood flow as determined by the first injection of FM compared with the values obtained by the second and third injection. Therefore, prolongation of the recovery period after reestablishment of the portal circulation appears to be reasonable, for up to at least 1 h, in future studies.
Allocation of the blood flow values to the different tissue samples' location within the liver revealed that the layer close to the diaphragm is perfused significantly higher, by up to 16% compared with the central and up to 22% compared with the visceral layer. These results indicate that the distribution of the portal blood within the porcine liver is heterogeneous as in other organs, such as the heart and the kidney. We cannot determine whether this heterogeneity is physiological or whether it is caused by the unphysiological supine position of the animals during the experiments. However, most experiments in anesthetized pigs are performed with the animal in the supine position. Therefore, particularly in settings in which parameters are evaluated on the liver's diaphragmatic surface, such as near-infrared spectroscopy (23) or measurement of tissue oxygenation (4), the heterogeneity of blood distribution within the liver has to be taken into account, as it may result in an overestimation of the results by up to 22%.
The difference of RPBF in the different layers of the liver is much more pronounced than the difference between the left lobe and the rest of the organ. The latter is, therefore, probably only of statistical, but not physiological, relevance.
We conclude that the model presented allows us to determine RPBF reliably by means of FM, with a minimal interference with the liver's integrity and function. In addition, the results from the study indicate that the distribution of portal blood in the porcine liver is heterogeneous.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. E. Malarkey, K. Johnson, L. Ryan, G. Boorman, and R. R. Maronpot New Insights into Functional Aspects of Liver Morphology Toxicol Pathol, January 1, 2005; 33(1): 27 - 34. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
