JOURNAL
OF SURGICAL
RESEARCH
51, 267-270 (19%)
Liver Bacterial Clearance following Hepatic Artery Ligation and Portacaval Shunt SCHMUEL
KATZ,
M.D., MARCUS
A. JIMENEZ,
M.D., WILLIAM
E. LEHMKUHLER,
Section of Pediatric Surgery, Department of Surgery, Indianu University and the James Whitcomb Riley Hospital for Children, Indianapolis, Submitted
for publication
M.D.
School of Medicine, Indiana 46202
July 3, 1990
80% of the entire RES function [2]. The unique feature of the hepatic circulation is the reciprocal relationship between the portal vein and the hepatic artery [3, 41. This dual blood supply to the liver is occasionally altered by pathologic, therapeutic, and surgical conditions such as cirrhosis, tumor, trauma, thrombosis, hepatic artery ligation or embolization, and portacaval shunt. The impairment of the liver blood flow may affect the RES and may predispose to infectious complications. The purpose of this study is to evaluate and compare the effect of hepatic artery ligation and portacaval shunt on blood clearance and organ localization of viable Eschrichiu coli.
The reticuloendothelial system (RES) plays an important role in removing bacteria, endotoxins, and immune complexes from the circulation. Hepatic phagocytosis accounts for more than 80% of RES function. The dual hepatic blood supply (hepatic artery/portal vein) may be altered by pathologic states and surgical procedures. This study evaluates and compares the effect of hepatic artery ligation and portacaval shunt on hepatic trapping of viable Escherichia coli. Thirty rats were placed in three groups: Group I was composed of sham operated controls; Group II underwent end-to-side portacaval shunt (PCS); and in Group III, hepatic artery ligation (HAL) was performed. At 2 weeks following the operation 10 ’ ?+radiolabeled viable E. coli were injected via the tail vein. At 10 min, bacterial distribution in the different organs was determined. Tissue samples were processed for liquid scintillation counting. The final distribution of bacteria was calculated from the input specific activity (dpm/bacteria) and expressed as the mean percentage of injected viable E. coli per gram of tissue and per organ weight. There was a significant decrease of bacterial trapping by the liver in rats following PCS (Group II), 46.0 + 10.4% vs controls 77.1 f 3.73% (P < 0.005). This was partially compensated for by a significant increase of bacterial trapping by the lung. The decreased clearance in PCS rats is due to a reduction in liver mass compared to that in controls. Bacterial localization in HAL (Group III) rats was similar to that in controls. These data show that PCS decreases hepatic clearance and increases pulmonary localization of viable E. coli. This phagocytic dysfunction may contribute to increased susceptibility to infection following portacaval shunt. o 1991 Academic Press,
M.D., AND JAY L. GROSFELD,
MATERIALS
Experimental
AND
METHODS
Animals
Thirty male Sprague-Dawley rats (Harlan SpragueDawley, Inc., Indianapolis, IN) weighing 200-225 g were housed two per cage and fed with standard Rodent Laboratory Chow No. 5001 (Purina Mills, Inc., St Louis, MO) and tap water ad libitum. All of the following procedures were approved by the IUPUI Animal Care Committee (Indianapolis, IN). Laparotomy was performed under 0.15 ml/rat im ketamine “cocktail” anesthesia (containing 100 mg ketamine, 2.2 mg promazine, 0.4 mg atropine per milliliter) and clean surgical technique. Rats were placed in three experimental groups: Group I, sham operated controls, underwent mobilization of the hepatoduodenal ligament (n = 10). Group II underwent endto-side portacaval shunt (n = 10); the portal vein was dissected, the gastroduodenal vein was ligated and divided, and the portal vein was ligated and divided in the hilum of the liver and was anastomosed in an end-toside fashion to the vena cava at the level of the right renal vein, using microsurgical technique with continuous 8-O nylon sutures. In Group III, the main hepatic artery was ligated and divided between small titanium clips (Ethicon Inc., Somerville, NJ) (n = 10). The abdomen was closed in two layers using continuous 4-O Vicry1 sutures.
Inc.
INTRODUCTION
The reticuloendothelial system (RES) plays an important role in removing bacteria, endotoxins, immune complexes, and other particulate matter from the circulation [ 11. Hepatic phagocytosis accounts for more than 267
0022-4804/91$1.50 All
Copyright 0 1991 by Academic Press, Inc. rights of reproduction in any form reserved.
268
JOURNAL
Radiolabeling
OF SURGICAL
RESEARCH:
of Bacteria
E. coli (clinical isolate, Indiana University Hospital, Indianapolis) was cultured to the late exponential phase at 37°C on a shaker bath in 10 ml of chemically defined liquid medium (NH&l, 2 g; Na,HPO,, 6 g; KH,PO,, 3 g; NaCl, 5 g; MgCl, * 6H,O, 0.3 g; glucose, 4 g; Na,SO,, 0.005 g; and distilled water, 11 g) containing 5 &i/ml of [35S]Na,S0, (New England Nuclear, Boston, MA). 35Slabeled bacteria were harvested and washed three times by centrifugation (27,00Og/min) in normal saline. Bacterial concentration was adjusted spectrophotometrically to an absorbance at 540 nm that corresponded to approximately 10’ viable E. coli per milliliter as determined by a standard spread plate technique. Samples of the bacterial suspension were taken for radioactivity determination prior to the injection of the different groups of rats. Typically, the specific activity of 35Swas 106-lo7 disintegrations per minute (dpm) per 10’ viable bacteria. Bacterial
Clearance and Organ Distribution
For bacterial clearance and organ distribution determination, the animals were sedated with an intramuscular injection of a ketamine cocktail and kept in a restraining cage. The tail was soaked in warm water (50°C). One milliliter of 35S-labeled viable E. coli (10’ bacteria/ml) was injected via a lateral tail vein. To evaluate bacterial clearance, the tail was transected, 50 ~1 of tail blood was obtained from actively bleeding proximal tail vessels using heparinized micropipets at 1,2,5, and 10 min following bacterial challenge, and samples were transferred to glass scintillation vials. Ten minutes following injection of viable E. coli, at laparotomy, the patency of the portacaval shunt was verified. Bacterial distribution in the different organs was determined; liver, spleen, kidneys, and lungs were immediately removed and weighed. Small tissue samples (50-100 mg) were processed for liquid scintillation counting. The final distribution of bacteria in the different organs was calculated from the input specific activity (dpm/bacteria) and expressed as percentage of the injected bacteria per gram of tissue and per organ weight. All experimental groups were studied 2 weeks following the surgical procedures. Radioactivity
VOL.
Determination
Samples of the blood and excised tissues were dissolved by treatment of 1.0 ml of 100% Soluene 350 (Packard Camberra Co., Downers Grove, IL) at 50°C overnight. The dissolved tissues were dried by 100% isopropyl alcohol and decolorized by treatment with 30% H,O, for 30 min. Samples were cooled to room temperature and 15 ml of Insta-Gel X-F scintillation fluid (Packard Instrument Co., Downers Grove, IL) was added. Radioactivity was determined using a 2200 CA Tricarb liquid scintillation counter (Packard Instrument Co., Downers Grove, IL) interfaced with an IBM
51, NO. 3, SEPTEMBER
1991
% dl)m/orqan SO
60
40
20
0 lung
liver
m
control
m
portacaval
shunt l
kidney
0
hepatlc
art
ligation
p ( 0 005
FIG. 1. Organ localization of intravenously injected radiolabeled viable E. coli in the experimental groups 2 weeks following operation. Values represent means f SD expressed as percentage of the injected bacteria (dpm) per total organ. Bacterial distribution was determined 10 min following injection. The changes in liver and lung distribution in rats following portacaval shunt are significant (P < 0.005).
microcomputer. Data were corrected for quench using a calibration curve determined by external standardization for 35S. Radioactivity was expressed as 35S dpm. Statistical
Analysis
Results were expressed as the mean + standard deviation. Data were analyzed using the unpaired Student t test. RESULTS
Intravenous injection of viable radiolabeled E. coli was followed by immediate clearance of bacteria from the circulation (more than 95% in the first 5 min) and subsequent localization of bacteria in the different organs in all experimental animals. Bacterial distribution in the sham operated controls in the liver, spleen, kidneys, and lungs was 77.1 + 3.73, 8.43 f 0.54, 0.26 f 0.04, and 1.13 f 0.69%, respectively. There was no significant change in bacterial localization in these four organs following hepatic artery ligation (HAL) (Group III) (Fig. 1). There was, however, a significant decrease of bacterial trapping by the liver in rats following portacaval shunt (PCS) (Group II) 45.0 f 10.4% vs controls 77.1 & 3.37% (P > 0.005). This decrease in the ability of the liver to trap bacteria was partially compensated by an increase of lung localization of bacteria following portacaval shunt, 6.7 + 3.22% compared to controls, 1.13 f 0.69% (P > 0.005) (Fig. 1). There was no significant difference in bacterial distribution in the spleen and the kidneys in the three experimental groups. The decrease in hepatic bacterial clearance in the PCS rats may be due to the decrease of the percentage liver mass, 2.4 + 0.6% of the
KATZ % dpm/g
tissue
LIVER
m Control
ET AL.:
BACTERIAL
CLEARANCE
T
SPLEEN
KIDNEY
LUNG
portacaval
shunt
*p
0
hepatlc
art
ligatfon
c
(0005
FIG. 2.
Tissue distribution of viable E. coli. Histogram demonstrates the specific ability of the different organs to trap viable bacteria. Values represent means f SD expressed as percentage of the injected bacteria (dpm) per g tissue. The changes in splenic and lung bacterial trapping are significant (P < 0.005).
whole body weight compared to controls 4.9 f 0.5% (P > 0.005). A slight decrease in the spleen weight was also observed in PCS rats. There was no change in the organ weight of rats following hepatic artery ligation. The specific ability of the organ to trap bacteria, namely, the amount of bacteria found in 1 g of tissue, is shown in Fig. 2. An increase in splenic and lung-specific bacterial clearance was noted but the ability of the liver to trap viable E. coli was similar to that of controls. DISCUSSION
The relative contribution made by the hepatic artery and the portal vein to the liver blood flow may be altered by a variety of pathologic, therapeutic, and surgical conditions. The structural changes observed in the cirrhotic liver (e.g., fibrosis) may increase the portal vascular resistance, resulting in portal hypertension. In addition, vasoactive drugs employed in the management of bleeding esophageal and gastric varices may adversely affect portal vein and hepatic artery flow with subsequent diminished total liver perfusion. Furthermore, the creation of an end-to-side portacaval shunt diverts the entire portal circulation, impairs total liver blood flow, and may result in a reduced liver mass [5]. Hepatic artery occlusion has become an accepted mode of treatment for both primary and secondary liver tumors [6]. Hepatic artery embolization or ligation is occasionally employed in instances of uncontrollable hemorrhage from extensive liver trauma or in cases of diffuse unresectable hepatic vascular malformation (e.g., hemangiomatosis). Hepatic dearterialization may lead to a temporary and transient impairment of some metabolic liver functions such as glycogen depletion, synthesis of bile, proteins,
AND
DEVASCULARIZATION
OF LIVER
269
triglycerides, and lipoproteins [7]. In addition, the release of lysosomal enzymes following hepatic dearterialization may reflect an ischemic injury to the liver cells [8]. This study evaluates the ability of the reticuloendothelial system of the liver to remove blood-borne bacteria following hepatic artery ligation and portacaval shunt. Traditionally, RES function has been assessed on the basis of blood clearance and liver concentration of nonbacterial particles such as gelatinized colloid carbon, radiolabeled human serum albumin, lipid emulsion, technetium and sulfur-colloid. As the rate of clearance of these compounds is dependent on the physicochemical properties as well as on the concentration of the injected particles, the specificity and the applicability of the foreign particle clearance as an indication of the cellular malfunction of the RES are controversial [9]. We have elected to study the RES function of recovering rats 2 weeks following sham operation, portacaval shunt, and hepatic artery ligation, using viable E. coli in a challenging dose adjusted to an initial blood concentration in the upper level of clinical bacteremia. Intravenous injection of viable E. coli is followed by immediate clearance from the blood stream and a typical pattern of bacterial distribution in the different organs: more than 70% of the injected bacteria are found in the liver, 5% in the spleen, and smaller amounts in the lungs and kidneys. The normal bacterial distribution following hepatic artery ligation in rats may be explained by recovery and early regeneration of the injured liver [lo]. Furthermore, the increase in portal blood flow and the development of collaterals, observed in experimental animals following hepatic dearterialization, may sufficiently perfuse the liver sinusoids. Blood-borne bacteria are therefore efficiently trapped by the liver cells. Buchholtz et al. [ll] found a significant decrease in hepatic uptake of ggmTcsulfur colloid following portacaval shunt. This was related to the reduction of portal flow with inadequate compensatory increase in hepatic artery flow. Our data show that the specific hepatic bacterial clearance (namely, dpm/g of tissue) was unchanged following portacaval shunt. This suggests that the decrease in total liver bacterial trapping may be due to the reduction in liver mass observed in rats following PCS rather than the decrease in total blood flow. It should be emphasized that in spite of remarkable liver phagocytic dysfunction, there was no significant difference in bacterial clearance from the circulation. This is a result of an increased bacterial trapping by the lung which efficiently compensates for liver dysfunction. The mechanism by which liver RES impairment results in an increase in lung localization of bacteria has not been clarified. Lanser and Saba found that neutrophi1 margination in the lung occurred immediately following RES blockade and concluded that this contributed to the increase of lung localization of heat-killed Pseudomonas aeruginosa [ 121. Katz et al. found a thick-
270
JOURNAL
OF SURGICAL
RESEARCH:
ening of lung interstitium, consistent with infiltration of macrophages which are actively involved in bacterial phagocytosis, in rats with RES dysfunction following common duct ligation [13]. When the hepatic RES fails to clear blood-borne bacteria, the lung becomes a very efficient trapping organ but apparently with a relatively poor capacity to kill the sequestrated bacteria. Therefore, the persistence of viable bacteria in the lungs of rats following PCS may result in a local inflammatory reaction and may allow for a reemergence of bacteria and thus may be a potential mechanism predisposing to sepsis, endotoxemia, and pulmonary complications. Performance of a nonselective portacaval shunt in the cirrhotic patient reduces hepatic blood flow and may have a deleterious effect on blood clearance and organ distribution of viable bacteria. This phagocytic dysfunction may contribute to increased susceptibility to infection [ 14, 151 encountered in cirrhotic patients following portacaval shunt. REFERENCES Rogers, D. E. Host mechanisms which act to remove bacteria from the blood stream. Bact. Reu. 24: 50, 1960. Saba, T. M. Physiology and physiopathology of the reticuloendothelial system. Arch. Intern. Med. 126: 1031, 1970. Condon, R. E., Nyhus, I,. M., Chapman, N. O., and Harkins, H. N. Portal vein and hepatic artery interactions: Studies in isolated perfused liver. Gastroenterology 43: 547, 1962. Burton-Opitz, R. The vascularity of the liver. II. The influence of the portal blood flow, upon the flow in the hepatic artery. Q. J. Exp. Physiol. 4: 93, 1911.
VOL. 5. 6. 7.
8.
9.
10.
11.
12.
13.
14.
15.
51, NO. 3, SEPTEMBER
1991
Starzl, T. E., Porter, K. A., and Francavilla, A. The Eck fistula in animals and humans. Curr. Probl. Surg. 20: 688, 1983. Bengmark, S., and Jeppson, B. Status of ischemic therapy of hepatic tumors. Surg. Clin. North Am. 69: 411, 1989. Hasselgren, P. O., Fornander, J., Jagenburg, R., and Sundstrom, E. Effect of liver ischemia on hepatic protein synthesis in vitro and in viuo. Acta Physiol. &and. 114: 143, 1982. Almersjo, O., Bengmark, S., Emgevik, L., Haftstrom, L. O., Loughridge, B. P., and Nilsson, L. A. V. Serum enzyme changes after hepatic dearterialization in man. Ann. Surg. 167: 9, 1968. Hickman, R., Jones, G., Tyler, M., and Engelbrecht, G. H. C. The effect of portal diversion upon function of the reticuloendothelial system and upon plasma fibronectin levels. J. Hepatol. 8: 67,1987. Gross, K., Katz, S., Dunn, S. P., Cikrit, D., Rosenthal, R. S., and Grosfeld, J. L. Bacterial clearance in the intact and regenerating liver. J. Pediatr. Surg. 20: 320, 1985. Buchholtz, B., Bergquist, L., Ryden, S., and Holmin, T. Hepatic reticuloendothelial function in rats with various portasystemic shunts and total liver arterialization. J. Hepatol. 4: 80, 1987. Lanser, M. E., and Saba, T. M. Neutrophil mediated lung localization of bacteria: A mechanism for pulmonary injury. Surgery 90: 433, 1981. Katz, S., Grosfeld, J. L., Gross, K., Plager, D. A., Ross, D., Rosenthal, R. S., Hull, M., and Weber, T. R. Impaired bacterial clearance and trapping in obstructive jaundice. Ann. Surg. 199: 14, 1984. Jackson, F. C., Perrin, E. B., Felix, W. R., and Smith, A. G. A clinical investigation of the portacaval shunt. V. Survival analysis of the therapeutic operation. Ann. Surg. 174: 672, 1971. Lier, H., Grun, M. In E. Wisse and D. L. Knook (Eds.), Kupfer Cells and Other Liver Sinusoidal Cells. (Proceedings of the International Kupffer Cell Symposium, Noordwijkerhout, The Netherlands, September, 1977). Amsterdam: Elsevier/North-Holland Biomedical Press, 1978.
OF SURGICAL
RESEARCH
51, 267-270 (19%)
Liver Bacterial Clearance following Hepatic Artery Ligation and Portacaval Shunt SCHMUEL
KATZ,
M.D., MARCUS
A. JIMENEZ,
M.D., WILLIAM
E. LEHMKUHLER,
Section of Pediatric Surgery, Department of Surgery, Indianu University and the James Whitcomb Riley Hospital for Children, Indianapolis, Submitted
for publication
M.D.
School of Medicine, Indiana 46202
July 3, 1990
80% of the entire RES function [2]. The unique feature of the hepatic circulation is the reciprocal relationship between the portal vein and the hepatic artery [3, 41. This dual blood supply to the liver is occasionally altered by pathologic, therapeutic, and surgical conditions such as cirrhosis, tumor, trauma, thrombosis, hepatic artery ligation or embolization, and portacaval shunt. The impairment of the liver blood flow may affect the RES and may predispose to infectious complications. The purpose of this study is to evaluate and compare the effect of hepatic artery ligation and portacaval shunt on blood clearance and organ localization of viable Eschrichiu coli.
The reticuloendothelial system (RES) plays an important role in removing bacteria, endotoxins, and immune complexes from the circulation. Hepatic phagocytosis accounts for more than 80% of RES function. The dual hepatic blood supply (hepatic artery/portal vein) may be altered by pathologic states and surgical procedures. This study evaluates and compares the effect of hepatic artery ligation and portacaval shunt on hepatic trapping of viable Escherichia coli. Thirty rats were placed in three groups: Group I was composed of sham operated controls; Group II underwent end-to-side portacaval shunt (PCS); and in Group III, hepatic artery ligation (HAL) was performed. At 2 weeks following the operation 10 ’ ?+radiolabeled viable E. coli were injected via the tail vein. At 10 min, bacterial distribution in the different organs was determined. Tissue samples were processed for liquid scintillation counting. The final distribution of bacteria was calculated from the input specific activity (dpm/bacteria) and expressed as the mean percentage of injected viable E. coli per gram of tissue and per organ weight. There was a significant decrease of bacterial trapping by the liver in rats following PCS (Group II), 46.0 + 10.4% vs controls 77.1 f 3.73% (P < 0.005). This was partially compensated for by a significant increase of bacterial trapping by the lung. The decreased clearance in PCS rats is due to a reduction in liver mass compared to that in controls. Bacterial localization in HAL (Group III) rats was similar to that in controls. These data show that PCS decreases hepatic clearance and increases pulmonary localization of viable E. coli. This phagocytic dysfunction may contribute to increased susceptibility to infection following portacaval shunt. o 1991 Academic Press,
M.D., AND JAY L. GROSFELD,
MATERIALS
Experimental
AND
METHODS
Animals
Thirty male Sprague-Dawley rats (Harlan SpragueDawley, Inc., Indianapolis, IN) weighing 200-225 g were housed two per cage and fed with standard Rodent Laboratory Chow No. 5001 (Purina Mills, Inc., St Louis, MO) and tap water ad libitum. All of the following procedures were approved by the IUPUI Animal Care Committee (Indianapolis, IN). Laparotomy was performed under 0.15 ml/rat im ketamine “cocktail” anesthesia (containing 100 mg ketamine, 2.2 mg promazine, 0.4 mg atropine per milliliter) and clean surgical technique. Rats were placed in three experimental groups: Group I, sham operated controls, underwent mobilization of the hepatoduodenal ligament (n = 10). Group II underwent endto-side portacaval shunt (n = 10); the portal vein was dissected, the gastroduodenal vein was ligated and divided, and the portal vein was ligated and divided in the hilum of the liver and was anastomosed in an end-toside fashion to the vena cava at the level of the right renal vein, using microsurgical technique with continuous 8-O nylon sutures. In Group III, the main hepatic artery was ligated and divided between small titanium clips (Ethicon Inc., Somerville, NJ) (n = 10). The abdomen was closed in two layers using continuous 4-O Vicry1 sutures.
Inc.
INTRODUCTION
The reticuloendothelial system (RES) plays an important role in removing bacteria, endotoxins, immune complexes, and other particulate matter from the circulation [ 11. Hepatic phagocytosis accounts for more than 267
0022-4804/91$1.50 All
Copyright 0 1991 by Academic Press, Inc. rights of reproduction in any form reserved.
268
JOURNAL
Radiolabeling
OF SURGICAL
RESEARCH:
of Bacteria
E. coli (clinical isolate, Indiana University Hospital, Indianapolis) was cultured to the late exponential phase at 37°C on a shaker bath in 10 ml of chemically defined liquid medium (NH&l, 2 g; Na,HPO,, 6 g; KH,PO,, 3 g; NaCl, 5 g; MgCl, * 6H,O, 0.3 g; glucose, 4 g; Na,SO,, 0.005 g; and distilled water, 11 g) containing 5 &i/ml of [35S]Na,S0, (New England Nuclear, Boston, MA). 35Slabeled bacteria were harvested and washed three times by centrifugation (27,00Og/min) in normal saline. Bacterial concentration was adjusted spectrophotometrically to an absorbance at 540 nm that corresponded to approximately 10’ viable E. coli per milliliter as determined by a standard spread plate technique. Samples of the bacterial suspension were taken for radioactivity determination prior to the injection of the different groups of rats. Typically, the specific activity of 35Swas 106-lo7 disintegrations per minute (dpm) per 10’ viable bacteria. Bacterial
Clearance and Organ Distribution
For bacterial clearance and organ distribution determination, the animals were sedated with an intramuscular injection of a ketamine cocktail and kept in a restraining cage. The tail was soaked in warm water (50°C). One milliliter of 35S-labeled viable E. coli (10’ bacteria/ml) was injected via a lateral tail vein. To evaluate bacterial clearance, the tail was transected, 50 ~1 of tail blood was obtained from actively bleeding proximal tail vessels using heparinized micropipets at 1,2,5, and 10 min following bacterial challenge, and samples were transferred to glass scintillation vials. Ten minutes following injection of viable E. coli, at laparotomy, the patency of the portacaval shunt was verified. Bacterial distribution in the different organs was determined; liver, spleen, kidneys, and lungs were immediately removed and weighed. Small tissue samples (50-100 mg) were processed for liquid scintillation counting. The final distribution of bacteria in the different organs was calculated from the input specific activity (dpm/bacteria) and expressed as percentage of the injected bacteria per gram of tissue and per organ weight. All experimental groups were studied 2 weeks following the surgical procedures. Radioactivity
VOL.
Determination
Samples of the blood and excised tissues were dissolved by treatment of 1.0 ml of 100% Soluene 350 (Packard Camberra Co., Downers Grove, IL) at 50°C overnight. The dissolved tissues were dried by 100% isopropyl alcohol and decolorized by treatment with 30% H,O, for 30 min. Samples were cooled to room temperature and 15 ml of Insta-Gel X-F scintillation fluid (Packard Instrument Co., Downers Grove, IL) was added. Radioactivity was determined using a 2200 CA Tricarb liquid scintillation counter (Packard Instrument Co., Downers Grove, IL) interfaced with an IBM
51, NO. 3, SEPTEMBER
1991
% dl)m/orqan SO
60
40
20
0 lung
liver
m
control
m
portacaval
shunt l
kidney
0
hepatlc
art
ligation
p ( 0 005
FIG. 1. Organ localization of intravenously injected radiolabeled viable E. coli in the experimental groups 2 weeks following operation. Values represent means f SD expressed as percentage of the injected bacteria (dpm) per total organ. Bacterial distribution was determined 10 min following injection. The changes in liver and lung distribution in rats following portacaval shunt are significant (P < 0.005).
microcomputer. Data were corrected for quench using a calibration curve determined by external standardization for 35S. Radioactivity was expressed as 35S dpm. Statistical
Analysis
Results were expressed as the mean + standard deviation. Data were analyzed using the unpaired Student t test. RESULTS
Intravenous injection of viable radiolabeled E. coli was followed by immediate clearance of bacteria from the circulation (more than 95% in the first 5 min) and subsequent localization of bacteria in the different organs in all experimental animals. Bacterial distribution in the sham operated controls in the liver, spleen, kidneys, and lungs was 77.1 + 3.73, 8.43 f 0.54, 0.26 f 0.04, and 1.13 f 0.69%, respectively. There was no significant change in bacterial localization in these four organs following hepatic artery ligation (HAL) (Group III) (Fig. 1). There was, however, a significant decrease of bacterial trapping by the liver in rats following portacaval shunt (PCS) (Group II) 45.0 f 10.4% vs controls 77.1 & 3.37% (P > 0.005). This decrease in the ability of the liver to trap bacteria was partially compensated by an increase of lung localization of bacteria following portacaval shunt, 6.7 + 3.22% compared to controls, 1.13 f 0.69% (P > 0.005) (Fig. 1). There was no significant difference in bacterial distribution in the spleen and the kidneys in the three experimental groups. The decrease in hepatic bacterial clearance in the PCS rats may be due to the decrease of the percentage liver mass, 2.4 + 0.6% of the
KATZ % dpm/g
tissue
LIVER
m Control
ET AL.:
BACTERIAL
CLEARANCE
T
SPLEEN
KIDNEY
LUNG
portacaval
shunt
*p
0
hepatlc
art
ligatfon
c
(0005
FIG. 2.
Tissue distribution of viable E. coli. Histogram demonstrates the specific ability of the different organs to trap viable bacteria. Values represent means f SD expressed as percentage of the injected bacteria (dpm) per g tissue. The changes in splenic and lung bacterial trapping are significant (P < 0.005).
whole body weight compared to controls 4.9 f 0.5% (P > 0.005). A slight decrease in the spleen weight was also observed in PCS rats. There was no change in the organ weight of rats following hepatic artery ligation. The specific ability of the organ to trap bacteria, namely, the amount of bacteria found in 1 g of tissue, is shown in Fig. 2. An increase in splenic and lung-specific bacterial clearance was noted but the ability of the liver to trap viable E. coli was similar to that of controls. DISCUSSION
The relative contribution made by the hepatic artery and the portal vein to the liver blood flow may be altered by a variety of pathologic, therapeutic, and surgical conditions. The structural changes observed in the cirrhotic liver (e.g., fibrosis) may increase the portal vascular resistance, resulting in portal hypertension. In addition, vasoactive drugs employed in the management of bleeding esophageal and gastric varices may adversely affect portal vein and hepatic artery flow with subsequent diminished total liver perfusion. Furthermore, the creation of an end-to-side portacaval shunt diverts the entire portal circulation, impairs total liver blood flow, and may result in a reduced liver mass [5]. Hepatic artery occlusion has become an accepted mode of treatment for both primary and secondary liver tumors [6]. Hepatic artery embolization or ligation is occasionally employed in instances of uncontrollable hemorrhage from extensive liver trauma or in cases of diffuse unresectable hepatic vascular malformation (e.g., hemangiomatosis). Hepatic dearterialization may lead to a temporary and transient impairment of some metabolic liver functions such as glycogen depletion, synthesis of bile, proteins,
AND
DEVASCULARIZATION
OF LIVER
269
triglycerides, and lipoproteins [7]. In addition, the release of lysosomal enzymes following hepatic dearterialization may reflect an ischemic injury to the liver cells [8]. This study evaluates the ability of the reticuloendothelial system of the liver to remove blood-borne bacteria following hepatic artery ligation and portacaval shunt. Traditionally, RES function has been assessed on the basis of blood clearance and liver concentration of nonbacterial particles such as gelatinized colloid carbon, radiolabeled human serum albumin, lipid emulsion, technetium and sulfur-colloid. As the rate of clearance of these compounds is dependent on the physicochemical properties as well as on the concentration of the injected particles, the specificity and the applicability of the foreign particle clearance as an indication of the cellular malfunction of the RES are controversial [9]. We have elected to study the RES function of recovering rats 2 weeks following sham operation, portacaval shunt, and hepatic artery ligation, using viable E. coli in a challenging dose adjusted to an initial blood concentration in the upper level of clinical bacteremia. Intravenous injection of viable E. coli is followed by immediate clearance from the blood stream and a typical pattern of bacterial distribution in the different organs: more than 70% of the injected bacteria are found in the liver, 5% in the spleen, and smaller amounts in the lungs and kidneys. The normal bacterial distribution following hepatic artery ligation in rats may be explained by recovery and early regeneration of the injured liver [lo]. Furthermore, the increase in portal blood flow and the development of collaterals, observed in experimental animals following hepatic dearterialization, may sufficiently perfuse the liver sinusoids. Blood-borne bacteria are therefore efficiently trapped by the liver cells. Buchholtz et al. [ll] found a significant decrease in hepatic uptake of ggmTcsulfur colloid following portacaval shunt. This was related to the reduction of portal flow with inadequate compensatory increase in hepatic artery flow. Our data show that the specific hepatic bacterial clearance (namely, dpm/g of tissue) was unchanged following portacaval shunt. This suggests that the decrease in total liver bacterial trapping may be due to the reduction in liver mass observed in rats following PCS rather than the decrease in total blood flow. It should be emphasized that in spite of remarkable liver phagocytic dysfunction, there was no significant difference in bacterial clearance from the circulation. This is a result of an increased bacterial trapping by the lung which efficiently compensates for liver dysfunction. The mechanism by which liver RES impairment results in an increase in lung localization of bacteria has not been clarified. Lanser and Saba found that neutrophi1 margination in the lung occurred immediately following RES blockade and concluded that this contributed to the increase of lung localization of heat-killed Pseudomonas aeruginosa [ 121. Katz et al. found a thick-
270
JOURNAL
OF SURGICAL
RESEARCH:
ening of lung interstitium, consistent with infiltration of macrophages which are actively involved in bacterial phagocytosis, in rats with RES dysfunction following common duct ligation [13]. When the hepatic RES fails to clear blood-borne bacteria, the lung becomes a very efficient trapping organ but apparently with a relatively poor capacity to kill the sequestrated bacteria. Therefore, the persistence of viable bacteria in the lungs of rats following PCS may result in a local inflammatory reaction and may allow for a reemergence of bacteria and thus may be a potential mechanism predisposing to sepsis, endotoxemia, and pulmonary complications. Performance of a nonselective portacaval shunt in the cirrhotic patient reduces hepatic blood flow and may have a deleterious effect on blood clearance and organ distribution of viable bacteria. This phagocytic dysfunction may contribute to increased susceptibility to infection [ 14, 151 encountered in cirrhotic patients following portacaval shunt. REFERENCES Rogers, D. E. Host mechanisms which act to remove bacteria from the blood stream. Bact. Reu. 24: 50, 1960. Saba, T. M. Physiology and physiopathology of the reticuloendothelial system. Arch. Intern. Med. 126: 1031, 1970. Condon, R. E., Nyhus, I,. M., Chapman, N. O., and Harkins, H. N. Portal vein and hepatic artery interactions: Studies in isolated perfused liver. Gastroenterology 43: 547, 1962. Burton-Opitz, R. The vascularity of the liver. II. The influence of the portal blood flow, upon the flow in the hepatic artery. Q. J. Exp. Physiol. 4: 93, 1911.
VOL. 5. 6. 7.
8.
9.
10.
11.
12.
13.
14.
15.
51, NO. 3, SEPTEMBER
1991
Starzl, T. E., Porter, K. A., and Francavilla, A. The Eck fistula in animals and humans. Curr. Probl. Surg. 20: 688, 1983. Bengmark, S., and Jeppson, B. Status of ischemic therapy of hepatic tumors. Surg. Clin. North Am. 69: 411, 1989. Hasselgren, P. O., Fornander, J., Jagenburg, R., and Sundstrom, E. Effect of liver ischemia on hepatic protein synthesis in vitro and in viuo. Acta Physiol. &and. 114: 143, 1982. Almersjo, O., Bengmark, S., Emgevik, L., Haftstrom, L. O., Loughridge, B. P., and Nilsson, L. A. V. Serum enzyme changes after hepatic dearterialization in man. Ann. Surg. 167: 9, 1968. Hickman, R., Jones, G., Tyler, M., and Engelbrecht, G. H. C. The effect of portal diversion upon function of the reticuloendothelial system and upon plasma fibronectin levels. J. Hepatol. 8: 67,1987. Gross, K., Katz, S., Dunn, S. P., Cikrit, D., Rosenthal, R. S., and Grosfeld, J. L. Bacterial clearance in the intact and regenerating liver. J. Pediatr. Surg. 20: 320, 1985. Buchholtz, B., Bergquist, L., Ryden, S., and Holmin, T. Hepatic reticuloendothelial function in rats with various portasystemic shunts and total liver arterialization. J. Hepatol. 4: 80, 1987. Lanser, M. E., and Saba, T. M. Neutrophil mediated lung localization of bacteria: A mechanism for pulmonary injury. Surgery 90: 433, 1981. Katz, S., Grosfeld, J. L., Gross, K., Plager, D. A., Ross, D., Rosenthal, R. S., Hull, M., and Weber, T. R. Impaired bacterial clearance and trapping in obstructive jaundice. Ann. Surg. 199: 14, 1984. Jackson, F. C., Perrin, E. B., Felix, W. R., and Smith, A. G. A clinical investigation of the portacaval shunt. V. Survival analysis of the therapeutic operation. Ann. Surg. 174: 672, 1971. Lier, H., Grun, M. In E. Wisse and D. L. Knook (Eds.), Kupfer Cells and Other Liver Sinusoidal Cells. (Proceedings of the International Kupffer Cell Symposium, Noordwijkerhout, The Netherlands, September, 1977). Amsterdam: Elsevier/North-Holland Biomedical Press, 1978.
