- Email: [email protected]
Bacterial Translocation from the Biliary Tract to Blood and Lymph in Rats with Obstructive Jaundice
JOURNAL OF SURGICAL RESEARCH ARTICLE NO.
74, 125–130 (1998)
JR975192
Bacterial Translocation from the Biliary Tract to Blood and Lymph in Rats with Obstructive Jaundice1 Thomas M. Karsten, M.D.,* Thomas M. van Gulik, M.D.,* Lodewijk Spanjaard, M.D.,† Anne Bosma, M.D.,‡ Marius A. van der Bergh Weerman,‡ Koert P. Dingemans,‡ Jacob Dankert, M.D.,† and Dirk J. Gouma, M.D.* *Department of Surgery, †Department of Bacteriology, and ‡Department of Pathology, Academic Medical Centre, Amsterdam, The Netherlands Submitted for publication June 18, 1997
Background. The disruption of the hepatocyte tight junctions observed in biliary obstruction suggests altered permeability of the blood–bile barrier. In this study the role of biliary obstruction and increased biliary pressure on the translocation of bacteria from biliary tract to bloodstream and lymphatic system were evaluated. Materials and Methods. Rats underwent distal bile duct ligation (BDL, n Å 33) for two weeks or a sham celiotomy (n Å 21). Seventeen of the 33 BDL rats underwent subsequent biliary decompression by a choledochojejunostomy (CJ). Two weeks after the final operation, a laparotomy was performed again and the CBD, the thoracic duct, and the caval vein were canulated. Next, a suspension containing 108 Escherichia coli/ml was retrogradely infused in the CBD for 5 min at 5 or 20 cm H2O above the secretory biliary pressure. Results. A higher biliary infusion pressure resulted in a significant increase of cfu E.coli per milliliter of blood in all the three groups (Sham, BDL, CJ). BDL rats showed significantly more bacterial translocation to the bloodstream than the shams. After biliary decompression, translocation normalized to the control levels. At 5 cm H2O infusion pressure only one lymph culture was positive (CJ group). At 20 cm H2O overpressure, nine lymph cultures were E.coli positive (P Å 0.03). These were found mainly in groups with a nonobstructed bile duct (Sham and CJ 40% vs BDL 10%). Conclusion. Translocation of bacteria from biliary tract to bloodstream increased at higher intrabiliary pressures. Longstanding bile duct obstruction was an independent determinant for cholangiovenous reflux. Bacterial translocation to the lymphatic system did not parallel translocation to the bloodstream, although in the nonobstructed biliary tract, increased bacterial translocation to the lymphatic system was pressure related. q 1998 Academic Press
ally result in septicemia. Several studies have identified elevated intrabiliary pressure as the causative factor in the development of septicemia [1 –4]. However, the effect of bile duct obstruction on the translocation of bacteria from the biliary tract to the bloodstream and lymphatic system has only been scarcely studied. Prolonged biliary obstruction gives rise to numerous changes in the liver and the host defense system. Histological changes include dilatation and proliferation of the biliary tract [5], desintegration of the extrahepatic biliary epithelium [6], hepatic patchy necrosis [7], and disruption of tight junctions [8, 9]. Host defense alterations entail dysfunction of chemotaxis and decreased killing potential of polymorphonuclear cells [10] and Kupffer cells [11, 12]. In addition, the secretion of IgA in the bile is diminished, by which the ability to bind bacteria and maintain mucosal integrity decreases [13]. These changes resulting from prolonged biliary tract obstruction, in conjunction with the increased biliary duct pressure, may induce translocation of bacteria from the biliary tract into bloodstream and lymphatic system. Thus far, only one study has addressed this aspect and failed to identify a role for biliary obstruction in the occurrence of bacterial cholangiovenous reflux (14). However, the duration of bile duct obstruction was short, 48 h. In the present study, we investigated separately the effect of biliary pressure and biliary obstruction on bacterial translocation from bile duct to bloodstream and lymphatic system in an experimental model in the rat, applying bile duct obstruction for 2 weeks. Furthermore, whether surgical decompression after bile duct obstruction would affect the occurrence of bacterial translocation was investigated. MATERIALS AND METHODS
INTRODUCTION
It is well known that whenever the biliary tract is obstructed, cholangitis can develop, which can eventu1 Paper presented at the 2nd World Congress of the IHPBA, Bologna, Italy, June 2–6, 1996.
Animals and surgical procedures. Fifty-seven male Wistar rats, weighing 280 to 380 g, were used. They were allowed free access to rat chow and water and were not fasted before operation or death. All experiments were approved by the University of Amsterdam animal care committee. Anesthesia was induced by injecting xylazine/ketamine/atropine intramuscularly and laparotomy was performed under clean, but not sterile conditions. During the surgical 0022-4804/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
125
AID
JSR 5192
/
6n28h$$301
04-01-98 08:59:59
srga
126
JOURNAL OF SURGICAL RESEARCH: VOL. 74, NO. 2, FEBRUARY 1, 1998
FIG. 1. Experimental design.
procedures, body temperature was maintained by a heated matress at 377C. Ringer’s glucose solution was administered via the tail vein at a rate of 0.5 ml/min. The experimental design is essentially a modification of the experiments described by Raper et al. [14] (Fig. 1). The animals were randomized to undergo sham celiotomy (sham), bile duct ligation (BDL), or bile duct ligation for 2 weeks followed by a Roux-en-Y choledochojejunostomy (CJ). During the sham celiotomy the duodenum and the bile duct were mobilized, whereas in BDL rats, the distal bile duct was doubly ligated and divided. In the CJ group, the choledochojejunostomy was constructed end-toend between the dilated stump of the bile duct and the first jejunal loop. The Roux en-Y end-to-side enteroenterostomy was constructed approximately 8 cm distal of the anastomosis with the bile duct. Both anastomoses were performed using continuous prolene 7-O sutures (polypropylene, Ethicon, FRG). During the first surgical procedure blood samples for bilirubin measurement were taken from the tail vein. Two weeks after the preparatory operations in each group, a laparotomy was performed again for the actual experiment. The caval vein was canulated via the right lumbal vein with the tip of the canula being positioned at the outflow point of the hepatic veins (polyethylene, 0.58 i.d. 1 0.96 o.d. mm, Portex, England). Thus, posthepatic blood could be aspirated. Next, the thoracic duct was dissected just below the diaphragm and canulated with polyethylene tubing (0.58 1 0.96 mm, Portex, England) and secured with a ligature. Lymph flow was measured before and during the perfusion experiment. Finally, the bile duct was canulated. Sham-operated animals were canulated with nylon tubing (0.5 mm i.d. 1 0.63 mm o.d., Portex, England), whereas BDL rats and rats with a CJ were canulated with larger tubing (1 mm i.d. 1 2 mm o.d., Medica bv, The Netherlands) because of the increased diameter of the bile duct. Bacteria. A stock culture of Escherichia coli (O149 K91 AC/) cultured from fecal specimens of a pig was used for the experiments (IDDLO, Lelystad, The Netherlands) (15). To establish pathogenicity of the pig E.coli for the rat, 0.1 ml of an E.coli suspension in phosphate-buffered saline containing 107 bacteria was injected into the common bile duct in two rats after which the bile duct was ligated. Both rats developed a purulent cholangitis within 24 h, and one died. Biliary pressure measurements and bacterial infusion. The free end of the bile duct canula was connected to a three-way stop-cock connector, of which the first port led to a pressure transducer connected to an XY recorder (Fig. 2). By closing off the two other ends of the three-way stop-cock connector, a closed system was created, in which, after pressure buildup and stabilization for 5 min, the maximum biliary secretory pressure of the liver was measured. After the initial pressure measurement, the second port of the three-way connector was attached to a 50-ml syringe filled with the bacterial suspension on a variable-speed pump, and the third port was connected to an overflow catheter which drained freely in an empty vial. In this fashion an open end system was created, by which, during retrograde infusion of the bacterial suspension in the bile
AID
JSR 5192
/
6n28h$$302
04-01-98 08:59:59
duct, the pressure during infusion could be altered by raising the open end of the overflow catheter. By subtracting the volume of bacterial suspension collected in the vial from the volume injected, the volume of the bacterial suspension infused into the biliary duct was assessed. At the start of each experiment (t Å 0), a suspension of 108 colonyforming units (cfu) E.coli/ml in phosphate-buffered (pH 7,4) 0.9% NaCl was retrogradely infused for 5 min into the CBD at a pressure of 5 or 20 cm H2O higher than the maximal initial secretory biliary pressure. At t Å 0 and t Å 5 min, 0.5 ml posthepatic blood and 0.1 ml lymphatic fluid were sampled and 20-fold diluted in fresh Columbia broth. Subsequently, quantitative cultures of the blood and lymph samples were made by plating 0.1 ml of 10-fold serial dilutions of the broth. After incubation at 377C for 24 h the number of E.coli colonies was counted. At the end of the experiment the animals were killed by an overdose of intravenous pentobarbital. Statistics. Statistical analysis was carried out with the nonparametric, unpaired, two-tailed Mann–Whitney test unless reported otherwise. P values õ0.05 were accepted as significant.
RESULTS
Bilirubin Levels, Body Weights, Maximal Biliary Pressure, and Lymph Flow In either group of rats, bile duct ligation for 2 weeks resulted in a significant increase of bilirubin levels in plasma and lymph. Sham celiotomy did not modify the bilirubin levels. Restoration of bile flow with a CJ normalized the elevated bilirubin levels to preoperative
FIG. 2. Experimental setup during bacterial infusion into the biliary tract. The caval vein is canulated through a lumbal vein (A), and the tip of the canula directed at the hepatic vein confluence (arrow). The canulated CBD (B) is retrogradely infused by a pump (E) connected to a pressure recorder (D) and an outflow canula (C). The height of the overflow canula determines the infusion pressure. The thoracic duct is canulated (F) below the diafragm downstream of the inflow of the hepatic lymph.
srga
127
KARSTEN ET AL.: BACTERIAL TRANSLOCATION TO BLOOD AND LYMPH
TABLE 2 Mean ({SD) Infused Volumes (ml) of E. coli Suspension (108/ml) into the Biliary Tract for 5 min
FIG. 3. Bilirubin levels in lymph and blood.
levels (Fig. 3). Body weight of sham-operated controls increased significantly during the 2 weeks between celiotomy and the final experiment (Table 1). In contrast, BDL rats had an average weight loss of 15 g during the 2 weeks of cholestasis (P õ 0.05). Only the CJ group with 5 cm H2O overpressure gained weight (mean) after the bile flow was restored, whereas in the CJ group with 20 cm overpressure no significant weight gain was noted (average 10 g in 2 weeks) after decompression, although serum bilirubin levels returned to normal. The maximum secretory biliary pressure of the liver was not affected by the surgical procedures, since no significant difference in pressure measurement results between the groups of rats was noted. Thoracic duct lymph flow in either group of rats increased significantly after biliary obstruction (P õ 0.05) (Table 1). Restoration of bile flow resulted in a normalization of the lymph flow. During infusion of the bacterial suspension into the common bile duct, the lymph flow was related to the intrabiliary pressure: at 20 cm water overpressure lymph flow was significantly higher than at 5 cm water overpressure in all rats. Retrograde Biliary Infusion of Bacterial Suspension Retrograde infusion of the bacterial suspension containing 108 E.coli/ml into the bile duct was performed
Procedure
Overpressure at 5 cm H2O
Overpressure at 20 cm H2O
Sham BDL CJ
0.28 { 0.06 0.91 { 0.33 0.71 { 0.7
1.31 { 0.09 4.9 { 1.1 3.0 { 0.95
for 5 min at a constant pressure of either 5 or 20 cm H2O above the maximum secretory biliary pressure. In all three groups, the volume of infused suspension into the common bile duct at 20 cm H2O was significantly higher than that at 5 cm H2O overpressure (Table 2). The infused volumes also differed considerably between the three groups of rats, since the dimensions of the dilated and proliferated bile duct system of the BDL rats and, to a lesser degree, of the CJ rats, were significantly greater than the sham-operated animals. Translocation of E.Coli to the Blood after Retrograde Biliary Infusion for 5 min All preinfusion blood cultures (t Å 0) were sterile. In all 3 groups of rats, biliary infusion pressure at 5 and 20 cm H2O resulted in positive bloodcultures in all rats. The number of cfu of E.coli per milliliter of blood was significantly higher at 20 cm H2O (sham, 1.9 { 2.7 vs 446 { 963 1 104; BDL, 31 { 36 vs 4424 { 8246 1 104; CJ, 2.5 { 4.3 vs 81 { 76 1 104 (data in mean { SD)) (Table 3). Since all groups were infused with various volumes to maintain a constant biliary pressure, the absolute bacterial load for each group differed proportionally. However, after standardization for the number of cfu in the blood per milliliter of infused bacterial suspension, the bacterial translocation from bile duct to bloodstream at the highest intrabiliary pressure was still significantly higher than at lower pressures. BDL rats showed, at both infusion pressures, significantly more translocation of E.coli from the biliary
TABLE 1 Mean ({SD) of Body Weights (g), Maximum Biliary Secretory Pressure (cm H2O), and Lymph Flow (ml/min) Weight
Group
Initial
Sham, 5 cm H2O (n Å 11) Sham, 20 cm H2O (n Å 10) BDL, 5 cm H2O (n Å 10) BDL, 20 cm H2O (n Å 9) CJ, 5 cm H2O (n Å 8) CJ, 20 cm H2O (n Å 9)
310 360 380 356 340 381
{ { { { { {
15 35 23 8 38 34
After sham or BDL 321 375 364 342 321 365
{ { { { { {
35* 43* 23* 31* 28* 30*
After CJ — — — — 335 { 26† 355 { 32‡
Maximum biliary pressure 18.5 20.7 18 18.9 19.6 19.1
{ { { { { {
2 1.6 3.5 3.6 1.3 1.2
Initial lymph flow 0.05 0.07 0.15 0.17 0.09 0.08
{ { { { { {
0.02 0.05 0.06 0.06 0.06 0.05
Lymph flow during biliary infusion 0.08 0.12 0.18 0.29 0.14 0.18
{ { { { { {
0.04† 0.08† 0.04† 0.07† 0.03† 0.06†
Note. Body weights: *P õ 0.05 for initial weight vs BDL weight or sham weight. †P Å 0.015 for BDL weight vs CJ weight. ‡P Å 0.26 (ns) for BDL weight vs CJ weight. Lymph flow: †P 0.05 for baseline lymph flow vs lymph flow during infusion. P by one-tailed paired Mann– Whitney test.
AID
JSR 5192
/
6n28h$$302
04-01-98 08:59:59
srga
128
JOURNAL OF SURGICAL RESEARCH: VOL. 74, NO. 2, FEBRUARY 1, 1998
TABLE 3 Mean ({SD) Number of E. coli Colony-Forming Units (cfu 1 104/ml) per Milliliter in Blood Collected after 5 min of Biliary Infusion of an E. coli Suspension Containing 108/ml at Two Infusion Pressures Overpressure, 5 cm H2O
Overpressure, 20 cm H2O
1.9 { 2.7 31 { 36† 2.5 { 4.3‡
446 { 963 4424 { 8426† 81 { 76‡
Sham BDL CJ
P
Overpressure, 5 cm H2O (correcteda)
Overpressure, 20 cm H2O (correcteda)
P
0.03 0.005 0.002
7.6 { 12 33 { 39† 3.9 { 7.3‡
315 { 650 831 { 1467† 28 { 24‡
0.01 0.03 0.006
Note. For both pressure groups: †P õ 0.05, sham vs BDL; ‡P õ 0.05, BDL vs CJ.
tract to the bloodstream compared to the sham-operated controls (5 cm H2O: sham, 1.9 { 2.7 vs BDL, 31 { 36 1 104; P Å 0.01; and 20 cm H2O: sham, 446 { 963 vs BDL, 4424 { 8426 1 104; P Å 0.02). Restoration of bile flow after 2 weeks of cholestasis in CJ rats reduced translocation to the bloodstream to sham control levels in both pressure groups. Translocation of E.coli to the Thoracic Duct after Retrograde Biliary Infusion for 5 min All preinfusion lymph cultures were sterile. At 5 cm H2O overpressure, only one lymph culture among 29 rats showed growth of E.coli (Table 4). This finding was noted in the decompressed group. At 20 cm H2O overpressure, a total 9 cultures from the thoracic duct were positive for E.coli. In this pressure group the number of positive lymph cultures was significantly less in the BDL groups than in the sham and decompressed groups (1/29 vs 9/28; P Å 0.03; Fischer’s exact test). DISCUSSION
This study clearly showed that biliary pressure and a period of bile duct obstruction are determinants of bacterial translocation from the biliary tract to the bloodstream. The major importance of intrabiliary
pressure in regurgitation of bile constituents to the bloodstream is in accordance with earlier studies. In 1947, it was demonstrated by Mixter et al. that contrast media refluxed to the blood after injection into the bile duct [2]. In 1969, Huang et al. showed that the process of bacterial regurgitation, whether via cholangiovenous or lymphatic routes, could be readily initiated in the dog, once the pressure in the CBD reached 20 cm H2O [1]. A similar, pressure-related passage of bacteria from bile ducts to bloodstream has been demonstrated in a rat model [14]. In the present study, retrograde infusion of the CBD was conducted at overpressures of 5 and 20 cm H2O, yielding absolute pressures in the CBD of approximately 25 and 40 cm H2O. These high pressures were necessary to overcome the maximal secretory biliary pressure. It can be argued that such pressures far above the maximal secretory biliary pressure may create artifacts. However, it is known that in case of an infection, pressures in the duct also exceed the maximal secretory biliary pressure [16]. Furthermore, in humans, the maximal secretory biliary pressure is 40 cm H2O and during contraction of the gallbladder, intrabiliary pressure can even rise to 80 cm H2O, which is much higher than the pressures used in the present study [17]. This study also showed a significant increase in bac-
TABLE 4 Absolute Number of Colony Forming Units E. coli per Milliliter in Thoracic Duct Lymph Cultures Collected 5 min after Biliary Infusion of an E. coli Suspension of 108/ml at Two Infusion Pressures Sham, 5 cm H2O (n Å 11)
— — — — — — — — — —
BDL, 5 cm H2O (n Å 10)
CJ, 5 cm H2O (n Å 8)
Sham, 20 cm H2O (n Å 10)
BDL, 20 cm H2O (n Å 9)
CJ, 20 cm H2O (n Å 9)
— — — — — — — — — —
— — — — — 400 — —
1.7 E 7 1200 7500 — 300 — — — — —
— — 4000 — — — — — —
— — — 7500 100 5000 — 10000 —
Note. All groups combined: P Å 0.03 for positive cultures at 5 cm overpressure (1 of 29) vs 20 cm overpressure (9 of 28). Groups with a nonobstructed bile duct (sham and CJ): P Å 0.05 for positive cultures at 5 cm overpressure (1 of 19) vs 20 cm overpressure (8 of 19). Nonsignificant difference in each individual group (sham, BDL, or CJ) between 5 and 20 cm overpressure (Fisher’s exact test).
AID
JSR 5192
/
6n28h$$303
04-01-98 08:59:59
srga
KARSTEN ET AL.: BACTERIAL TRANSLOCATION TO BLOOD AND LYMPH
terial translocation to the systemic circulation after a 2-week period of BDL. The many pathophysiological changes resulting from obstructive jaundice may be involved in this phenomenon. It has been suggested that breakdown of the tight junctions between the hepatocytes damages the anatomical barrier between bile canaliculi and sinusoids and facilitates cholangiovenous reflux [18, 19]. Epithelial disruption of the biliary tract occurring after longstanding bile duct obstruction may also promote reflux [6]. Furthermore, the sinusoids are insufficiently cleared from refluxed bacteria due to dysfunction of the Kupffer cells, enhancing invasion into the systemic circulation [20]. The results of this study, however, are in contrast with other reports in which no increased cholangiovenous reflux after biliary obstruction could be demonstrated [14]. This discrepancy between the results can probably be attributed to the extended period of cholestasis (2 weeks) used in the present study, which induces more severe damage to the liver and the reticuloendothelial system. Furthermore, in the present setup, blood samples were taken from the caval vein at the outflow site of the hepatic veins, thus ensuring sampling of posthepatic blood. Raper et al. used blood samples collected from the femoral artery, which had already passed the lungs [14]. It is known that after bile duct obstruction, when Kupffer cell function is depressed, a compensatory increase of clearance by the lung macrophages and the spleen occurs [21]. Pulmonary clearance of bacteria could therefore account for low bacterial counts in femoral artery blood. After 2 weeks of decompression by choledochojejunostomy the bacterial translocation observed after bile duct obstruction returned to the normal control level. Several studies have readily shown that bile duct decompression restores most of the changes of the liver parenchyma [22, 23]. Complete recovery of liver synthesis function was estimated to be achieved between 4 and 6 weeks [24]. A report by Katz et al. showed complete reversal of hepatocyte phagocyte dysfunction and a decrease in compensatory bacterial trapping after 2 weeks of biliary decompression in rats [25]. The present results are consistent with these findings and provide evidence in favor of preoperative biliary drainage. Further studies should address the optimal duration of biliary decompression in order to reverse changes in translocation due to bile duct obstruction. Lymph flow doubled after bile duct obstruction and returned to normal after decompression. The biliary infusion experiments clearly demonstrated that lymph flow was directly related to intrabiliary pressure, which is consistent with other studies [1, 26]. The data from the lymph cultures, however, did not parallel those from the blood cultures. This is in contrast with studies by Stewart [27], Yamamoto [28] in rats, and Huang in dogs [1], which all suggested that reflux to the lymphatics occurred before and at lower pressures than reflux to the bloodstream. However, Stewart and Yamamoto used particulate matter to
AID
JSR 5192
/
6n28h$$303
04-01-98 08:59:59
129
identify paths of reflux, and only in the study by Huang were viable bacteria used. It is generally assumed that hepatic lymph originates in the interstitial space of the liver parenchyma, i.e., the space’s of Disse and Mall [29]. The lymph tract starts in the terminal lymphatics in the portal areas, which drain into collecting vessels along the veins. Hepatic lymph ultimately drains by two routes in the thoracic duct [26, 30–32]. The first route (sampled in this study) follows the hepatoduodenal ligament and ends alongside the superior mesenteric artery into the cisternae chyli which, in turn, drain into the abdominal part of the thoracic duct. The second route parallels the caval vein and enters the thoracic duct at the level of the right subclavian vein. The mechanism by which bacterial translocation predominantly takes place to the bloodstream and not to the lymphatic system is not clear. Although the second extrahepatic lymph route was not canulated in this study, it is unlikely that bacterial translocation would be directed only to the lymph drainage system alongside the caval vein. In addition, bacterial clearance of the lymph by lymph nodes along the route to the thoracic duct, or shunting of bacteria through lymphaticovenous communications remains a matter of speculation (33). This study answers no questions concerning the pathway that bacteria take in the liver parenchyma, during their passage from bile duct to either the sinusoids or the lymph tract. In theory, this can be either a transcellular or paracellular route at the level of the bile canaliculus or the larger bile ducts. Earlier studies showed, at pressures up to 50 cm H2O, that reflux occurred from bile ductules to the spaces of Disse and Mall [27, 28]. Only by blocking these low-resistance pathways can passage via biliary canaliculi occur. Another study showed that bacterial reflux occurred predominantly across the hepatocyte, by transcellular pathways [14]. The relevance of these pathways and their relation to bile duct pressure and various bacterial species require further investigation. In conclusion, this study confirms that the translocation of bacteria from biliary tract to bloodstream increases at higher intrabiliary pressures. It was also shown that longstanding bile duct obstruction is an independent determinant for cholangiovenous reflux. Bacterial migration to the lymphatic system did not parallel translocation to the bloodstream, although in a nonobstructed biliary tract bacterial translocation to lymph seemed to be pressure-related. ACKNOWLEDGMENTS The authors are grateful to Mr. H. de Wit for his skillful assistance during the surgical experiments and to Mrs. P. Brinke and Mrs. E. Reurekas and Mr. G. Shatorje, for their assistance in performing the quantitative analysis of the blood and lymph cultures. The authors express their gratitude Mr. F. van Zijderveld who kindly provided the stock culture of E.coli and the monoclonal antibodies.
REFERENCES 1. Huang, T., Bass, J. A., and Williams, R. D. The significance of biliary pressure in cholangitis. Arch. Surg. 98: 629–632, 1969.
srga
130
JOURNAL OF SURGICAL RESEARCH: VOL. 74, NO. 2, FEBRUARY 1, 1998
2. Mixer, H., Rigler, L., and Oddone, M. Experimental studies on biliary regurgitation during cholangiography. Gastroenterology 9: 64–80, 1947. 3. Jacobsson, B., Kjellander, J., and Rosengren, B. Cholangiovenous reflux. An experimental study. Acta Chir. Scand. 123: 316– 321, 1962. 4. Yoshimoto, H., Ikeda, S., Tanaka, M., and Matssumoto, S. Relationship of biliary pressure to cholangiovenous reflux during endoscopic retrograde ballon catheter cholangiography. Dig. Dis. Sci. 34: 16–20, 1989. 5. Desmet, V. J. Cholestasis: extrahepatic obstruction and secondary biliary cirrhosis. In McSween, R. N. M., Anthony, P. P., and Scheuer, P. J. (Eds.), Pathology of the Liver, 2nd ed. Edinburgh: Churchill–Livingstone, 1987. pp. 364–423. 6. Karsten, T. M., Coene, P. P. L. O., van Gulik, T. M., Bosma, A., et al. Morphological changes of extrahepatic bile ducts during obstruction and subsequent decompression by endoprosthesis. Surgery 111: 562–568, 1992. 7. Shibayama, Y. Endotoxaemia and hepatic injury in obstructive jaundice. J. Pathol. 159: 335–339, 1989. 8. Robenek, H., Herwig, J., and Themann, H. The morphologic characteristics of intracellular junctions between normal human liver cells and cells from patients with extrahepatic cholestasis. Am. J. Pathol. 100: 93–114, 1980. 9. Landman, L. Cholestasis induced alterations of the trans- and paracellular pathways in rat hepatocytes. Histochemistry 103: 3–9, 1995. 10. Andy, O. J., Grogan, J. B., Griswold, J. A., and Scott-Conner, C. E. H. Peritoneal neutrophil chemotaxis is impaired in biliary obstruction. Am. Surg. 58: 28–32, 1992. 11. Scott-Conner, C. E. H., Grogan, J. B., Scher, K. S., and Bernstein, J. Impaired clearance of Escherichia coli bacteremia in early biliary obstruction. Am. J. Surg. 157: 210–214, 1989. 12. Ding, J. W., Andersson, R., Soltesz, V., Willen, R., and Bengmark, S. Obstructive jaundice impaires reticuloendothelial function and promotes bacterial translocation in the rat. J. Surg. Res. 57: 238–245, 1994. 13. Rank, J., and Wilson, I. D. Changes in IgA following varying degrees of biliary obstruction in the rat. Hepatology 3: 241– 247, 1983. 14. Raper, S. E., Barker, M. E., Jones, A. L., and Way, L. W. Anatomic correlates of bacterial cholangiovenous reflux. Surgery 105: 352–359, 1989. 15. Zijderveld van, F. G., Westenbrink, F., Anakotta, J., Brouwers, R. A. M., et al. Characterization of the F41 fimbrial antigen of enterotoxigenic Escherichia coli by using monoclonal antibodies. Infect. Immun. 57: 1192–1199, 1989. 16. Csendes, A., Sepulveda, A., Burdiles, A., Braghetto, I., et al. Common bile duct pressure in patients with CBD stones with or without acute suppurative cholangitis. Arch. Surg. 123: 697– 699, 1988.
AID
JSR 5192
/
6n28h$$304
04-01-98 08:59:59
17. Toresoli, A., Corraziari, E., Habib, F. C., and Cicala, M. Pressure relationships within the human bile tract. Normal and abnormal physiology. [Review]. Scan. J. Gastroenterol. Suppl. 175: 52–57, 1990. 18. Ker, C. G., Huang, T. J., and Sheen, P. C. Electron microscopic assessment of bile regurgitation of human in extrahepatic obstructive jaundice. Eur. Surg. Res. 17: 197–206, 1985. 19. Toyota, N., Miyai, K., and Hardison, W. G. M. Effect of biliary pressure versus high bile acid flux on the permeability of hepatocellular tight junction. Lab. Invest. 50: 536–543, 1984. 20. Bradfield, J. W. B. Control of spillover: The importance of Kupffer cell function in clinical medicine. Lancet 2: 883–86, 1974. 21. Katz, S., Grosfeld, J. L., and Gross, K. Impaired clearance and trapping in obstructive jaundice. Ann. Surg. 199: 14–20, 1984. 22. Abdel-Aziz, G., Lebeau, G., Rescan, P. Y., Clement, B., et al. Reversibility of hepatic fibrosis in experimentally induced cholestasis in the rat. Am. J. Pathol. 137: 1333–1342, 1990. 23. Zimmerman, H., Reichen, J., Zimmerman, A., Sagesser, H., et al. Reversibility of secondary biliary fibrosis by biliodigestive anastomosis in the rat. Gastroenterology 103: 579–89, 1992. 24. Koyama, K., Takagi, Y., Ito, K., and Sato, T. Experimental and clinical studies on the effect of experimental obstructive jaundice. Am. J. Surg. 158: 423–427, 1989. 25. Katz, S., Yang, R., Rodefeld, M. J., Folkening, W. J., et al. Impaired hepatic clearance is reversed by surgical relief of obstructive jaundice. J. Pediatr. Surg. 26: 401–406, 1991. 26. Gyure, L. A., Hall, J. G., Lapis, K., and Thimar, J. The relief of biliary obstruction in experimental rats by diverting the hepatic lymph into the duodenum. Int. J. Exp. Pathol. 72: 345–348, 1991. 27. Stewart, L., Pellegrini, C. A., and Way, L. W. Cholangiovenous reflux pathways as defined by corrosion casting and scanning electron microscopy. Am. J. Surg. 155: 23–28, 1988. 28. Yamamoto, K., and Phillips, M. J. A hitherto unrecognized bile ductular plexus in the normal rat liver. Hepatology 4: 381–385, 1984. 29. Henriksen, J., Horn, T., and Christoffersen, P. The blood– lymph barrier in the liver. A review based on morphological and functional concept of normal and cirrhotic liver. Liver 4: 221–232, 1984. 30. Tilney, N. L. Patterns of lymphatic drainage in the adult laboratory rat. J. Anat. 109: 369–383, 1971. 31. Bollman, J. L., Cain, J. C., and Grindlay, J. H. Techniques for the collection of lymph from the liver, small intestine, or thoracic duct of the rat. J. Lab. Clin. Med. 33: 1349–1352, 1948. 32. Friedman, M., Byers, S. O., and Omoto, C. Some characteristics of hepatic lymph in the intact rat. Am. J. Physiol. 184: 11–17, 1956. 33. Hidden, G., Menard, P., and Zorn, J. Y. Lymphaticovenous communications. Anat. Clin. 7: 83–91, 1985.
srga
74, 125–130 (1998)
JR975192
Bacterial Translocation from the Biliary Tract to Blood and Lymph in Rats with Obstructive Jaundice1 Thomas M. Karsten, M.D.,* Thomas M. van Gulik, M.D.,* Lodewijk Spanjaard, M.D.,† Anne Bosma, M.D.,‡ Marius A. van der Bergh Weerman,‡ Koert P. Dingemans,‡ Jacob Dankert, M.D.,† and Dirk J. Gouma, M.D.* *Department of Surgery, †Department of Bacteriology, and ‡Department of Pathology, Academic Medical Centre, Amsterdam, The Netherlands Submitted for publication June 18, 1997
Background. The disruption of the hepatocyte tight junctions observed in biliary obstruction suggests altered permeability of the blood–bile barrier. In this study the role of biliary obstruction and increased biliary pressure on the translocation of bacteria from biliary tract to bloodstream and lymphatic system were evaluated. Materials and Methods. Rats underwent distal bile duct ligation (BDL, n Å 33) for two weeks or a sham celiotomy (n Å 21). Seventeen of the 33 BDL rats underwent subsequent biliary decompression by a choledochojejunostomy (CJ). Two weeks after the final operation, a laparotomy was performed again and the CBD, the thoracic duct, and the caval vein were canulated. Next, a suspension containing 108 Escherichia coli/ml was retrogradely infused in the CBD for 5 min at 5 or 20 cm H2O above the secretory biliary pressure. Results. A higher biliary infusion pressure resulted in a significant increase of cfu E.coli per milliliter of blood in all the three groups (Sham, BDL, CJ). BDL rats showed significantly more bacterial translocation to the bloodstream than the shams. After biliary decompression, translocation normalized to the control levels. At 5 cm H2O infusion pressure only one lymph culture was positive (CJ group). At 20 cm H2O overpressure, nine lymph cultures were E.coli positive (P Å 0.03). These were found mainly in groups with a nonobstructed bile duct (Sham and CJ 40% vs BDL 10%). Conclusion. Translocation of bacteria from biliary tract to bloodstream increased at higher intrabiliary pressures. Longstanding bile duct obstruction was an independent determinant for cholangiovenous reflux. Bacterial translocation to the lymphatic system did not parallel translocation to the bloodstream, although in the nonobstructed biliary tract, increased bacterial translocation to the lymphatic system was pressure related. q 1998 Academic Press
ally result in septicemia. Several studies have identified elevated intrabiliary pressure as the causative factor in the development of septicemia [1 –4]. However, the effect of bile duct obstruction on the translocation of bacteria from the biliary tract to the bloodstream and lymphatic system has only been scarcely studied. Prolonged biliary obstruction gives rise to numerous changes in the liver and the host defense system. Histological changes include dilatation and proliferation of the biliary tract [5], desintegration of the extrahepatic biliary epithelium [6], hepatic patchy necrosis [7], and disruption of tight junctions [8, 9]. Host defense alterations entail dysfunction of chemotaxis and decreased killing potential of polymorphonuclear cells [10] and Kupffer cells [11, 12]. In addition, the secretion of IgA in the bile is diminished, by which the ability to bind bacteria and maintain mucosal integrity decreases [13]. These changes resulting from prolonged biliary tract obstruction, in conjunction with the increased biliary duct pressure, may induce translocation of bacteria from the biliary tract into bloodstream and lymphatic system. Thus far, only one study has addressed this aspect and failed to identify a role for biliary obstruction in the occurrence of bacterial cholangiovenous reflux (14). However, the duration of bile duct obstruction was short, 48 h. In the present study, we investigated separately the effect of biliary pressure and biliary obstruction on bacterial translocation from bile duct to bloodstream and lymphatic system in an experimental model in the rat, applying bile duct obstruction for 2 weeks. Furthermore, whether surgical decompression after bile duct obstruction would affect the occurrence of bacterial translocation was investigated. MATERIALS AND METHODS
INTRODUCTION
It is well known that whenever the biliary tract is obstructed, cholangitis can develop, which can eventu1 Paper presented at the 2nd World Congress of the IHPBA, Bologna, Italy, June 2–6, 1996.
Animals and surgical procedures. Fifty-seven male Wistar rats, weighing 280 to 380 g, were used. They were allowed free access to rat chow and water and were not fasted before operation or death. All experiments were approved by the University of Amsterdam animal care committee. Anesthesia was induced by injecting xylazine/ketamine/atropine intramuscularly and laparotomy was performed under clean, but not sterile conditions. During the surgical 0022-4804/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
125
AID
JSR 5192
/
6n28h$$301
04-01-98 08:59:59
srga
126
JOURNAL OF SURGICAL RESEARCH: VOL. 74, NO. 2, FEBRUARY 1, 1998
FIG. 1. Experimental design.
procedures, body temperature was maintained by a heated matress at 377C. Ringer’s glucose solution was administered via the tail vein at a rate of 0.5 ml/min. The experimental design is essentially a modification of the experiments described by Raper et al. [14] (Fig. 1). The animals were randomized to undergo sham celiotomy (sham), bile duct ligation (BDL), or bile duct ligation for 2 weeks followed by a Roux-en-Y choledochojejunostomy (CJ). During the sham celiotomy the duodenum and the bile duct were mobilized, whereas in BDL rats, the distal bile duct was doubly ligated and divided. In the CJ group, the choledochojejunostomy was constructed end-toend between the dilated stump of the bile duct and the first jejunal loop. The Roux en-Y end-to-side enteroenterostomy was constructed approximately 8 cm distal of the anastomosis with the bile duct. Both anastomoses were performed using continuous prolene 7-O sutures (polypropylene, Ethicon, FRG). During the first surgical procedure blood samples for bilirubin measurement were taken from the tail vein. Two weeks after the preparatory operations in each group, a laparotomy was performed again for the actual experiment. The caval vein was canulated via the right lumbal vein with the tip of the canula being positioned at the outflow point of the hepatic veins (polyethylene, 0.58 i.d. 1 0.96 o.d. mm, Portex, England). Thus, posthepatic blood could be aspirated. Next, the thoracic duct was dissected just below the diaphragm and canulated with polyethylene tubing (0.58 1 0.96 mm, Portex, England) and secured with a ligature. Lymph flow was measured before and during the perfusion experiment. Finally, the bile duct was canulated. Sham-operated animals were canulated with nylon tubing (0.5 mm i.d. 1 0.63 mm o.d., Portex, England), whereas BDL rats and rats with a CJ were canulated with larger tubing (1 mm i.d. 1 2 mm o.d., Medica bv, The Netherlands) because of the increased diameter of the bile duct. Bacteria. A stock culture of Escherichia coli (O149 K91 AC/) cultured from fecal specimens of a pig was used for the experiments (IDDLO, Lelystad, The Netherlands) (15). To establish pathogenicity of the pig E.coli for the rat, 0.1 ml of an E.coli suspension in phosphate-buffered saline containing 107 bacteria was injected into the common bile duct in two rats after which the bile duct was ligated. Both rats developed a purulent cholangitis within 24 h, and one died. Biliary pressure measurements and bacterial infusion. The free end of the bile duct canula was connected to a three-way stop-cock connector, of which the first port led to a pressure transducer connected to an XY recorder (Fig. 2). By closing off the two other ends of the three-way stop-cock connector, a closed system was created, in which, after pressure buildup and stabilization for 5 min, the maximum biliary secretory pressure of the liver was measured. After the initial pressure measurement, the second port of the three-way connector was attached to a 50-ml syringe filled with the bacterial suspension on a variable-speed pump, and the third port was connected to an overflow catheter which drained freely in an empty vial. In this fashion an open end system was created, by which, during retrograde infusion of the bacterial suspension in the bile
AID
JSR 5192
/
6n28h$$302
04-01-98 08:59:59
duct, the pressure during infusion could be altered by raising the open end of the overflow catheter. By subtracting the volume of bacterial suspension collected in the vial from the volume injected, the volume of the bacterial suspension infused into the biliary duct was assessed. At the start of each experiment (t Å 0), a suspension of 108 colonyforming units (cfu) E.coli/ml in phosphate-buffered (pH 7,4) 0.9% NaCl was retrogradely infused for 5 min into the CBD at a pressure of 5 or 20 cm H2O higher than the maximal initial secretory biliary pressure. At t Å 0 and t Å 5 min, 0.5 ml posthepatic blood and 0.1 ml lymphatic fluid were sampled and 20-fold diluted in fresh Columbia broth. Subsequently, quantitative cultures of the blood and lymph samples were made by plating 0.1 ml of 10-fold serial dilutions of the broth. After incubation at 377C for 24 h the number of E.coli colonies was counted. At the end of the experiment the animals were killed by an overdose of intravenous pentobarbital. Statistics. Statistical analysis was carried out with the nonparametric, unpaired, two-tailed Mann–Whitney test unless reported otherwise. P values õ0.05 were accepted as significant.
RESULTS
Bilirubin Levels, Body Weights, Maximal Biliary Pressure, and Lymph Flow In either group of rats, bile duct ligation for 2 weeks resulted in a significant increase of bilirubin levels in plasma and lymph. Sham celiotomy did not modify the bilirubin levels. Restoration of bile flow with a CJ normalized the elevated bilirubin levels to preoperative
FIG. 2. Experimental setup during bacterial infusion into the biliary tract. The caval vein is canulated through a lumbal vein (A), and the tip of the canula directed at the hepatic vein confluence (arrow). The canulated CBD (B) is retrogradely infused by a pump (E) connected to a pressure recorder (D) and an outflow canula (C). The height of the overflow canula determines the infusion pressure. The thoracic duct is canulated (F) below the diafragm downstream of the inflow of the hepatic lymph.
srga
127
KARSTEN ET AL.: BACTERIAL TRANSLOCATION TO BLOOD AND LYMPH
TABLE 2 Mean ({SD) Infused Volumes (ml) of E. coli Suspension (108/ml) into the Biliary Tract for 5 min
FIG. 3. Bilirubin levels in lymph and blood.
levels (Fig. 3). Body weight of sham-operated controls increased significantly during the 2 weeks between celiotomy and the final experiment (Table 1). In contrast, BDL rats had an average weight loss of 15 g during the 2 weeks of cholestasis (P õ 0.05). Only the CJ group with 5 cm H2O overpressure gained weight (mean) after the bile flow was restored, whereas in the CJ group with 20 cm overpressure no significant weight gain was noted (average 10 g in 2 weeks) after decompression, although serum bilirubin levels returned to normal. The maximum secretory biliary pressure of the liver was not affected by the surgical procedures, since no significant difference in pressure measurement results between the groups of rats was noted. Thoracic duct lymph flow in either group of rats increased significantly after biliary obstruction (P õ 0.05) (Table 1). Restoration of bile flow resulted in a normalization of the lymph flow. During infusion of the bacterial suspension into the common bile duct, the lymph flow was related to the intrabiliary pressure: at 20 cm water overpressure lymph flow was significantly higher than at 5 cm water overpressure in all rats. Retrograde Biliary Infusion of Bacterial Suspension Retrograde infusion of the bacterial suspension containing 108 E.coli/ml into the bile duct was performed
Procedure
Overpressure at 5 cm H2O
Overpressure at 20 cm H2O
Sham BDL CJ
0.28 { 0.06 0.91 { 0.33 0.71 { 0.7
1.31 { 0.09 4.9 { 1.1 3.0 { 0.95
for 5 min at a constant pressure of either 5 or 20 cm H2O above the maximum secretory biliary pressure. In all three groups, the volume of infused suspension into the common bile duct at 20 cm H2O was significantly higher than that at 5 cm H2O overpressure (Table 2). The infused volumes also differed considerably between the three groups of rats, since the dimensions of the dilated and proliferated bile duct system of the BDL rats and, to a lesser degree, of the CJ rats, were significantly greater than the sham-operated animals. Translocation of E.Coli to the Blood after Retrograde Biliary Infusion for 5 min All preinfusion blood cultures (t Å 0) were sterile. In all 3 groups of rats, biliary infusion pressure at 5 and 20 cm H2O resulted in positive bloodcultures in all rats. The number of cfu of E.coli per milliliter of blood was significantly higher at 20 cm H2O (sham, 1.9 { 2.7 vs 446 { 963 1 104; BDL, 31 { 36 vs 4424 { 8246 1 104; CJ, 2.5 { 4.3 vs 81 { 76 1 104 (data in mean { SD)) (Table 3). Since all groups were infused with various volumes to maintain a constant biliary pressure, the absolute bacterial load for each group differed proportionally. However, after standardization for the number of cfu in the blood per milliliter of infused bacterial suspension, the bacterial translocation from bile duct to bloodstream at the highest intrabiliary pressure was still significantly higher than at lower pressures. BDL rats showed, at both infusion pressures, significantly more translocation of E.coli from the biliary
TABLE 1 Mean ({SD) of Body Weights (g), Maximum Biliary Secretory Pressure (cm H2O), and Lymph Flow (ml/min) Weight
Group
Initial
Sham, 5 cm H2O (n Å 11) Sham, 20 cm H2O (n Å 10) BDL, 5 cm H2O (n Å 10) BDL, 20 cm H2O (n Å 9) CJ, 5 cm H2O (n Å 8) CJ, 20 cm H2O (n Å 9)
310 360 380 356 340 381
{ { { { { {
15 35 23 8 38 34
After sham or BDL 321 375 364 342 321 365
{ { { { { {
35* 43* 23* 31* 28* 30*
After CJ — — — — 335 { 26† 355 { 32‡
Maximum biliary pressure 18.5 20.7 18 18.9 19.6 19.1
{ { { { { {
2 1.6 3.5 3.6 1.3 1.2
Initial lymph flow 0.05 0.07 0.15 0.17 0.09 0.08
{ { { { { {
0.02 0.05 0.06 0.06 0.06 0.05
Lymph flow during biliary infusion 0.08 0.12 0.18 0.29 0.14 0.18
{ { { { { {
0.04† 0.08† 0.04† 0.07† 0.03† 0.06†
Note. Body weights: *P õ 0.05 for initial weight vs BDL weight or sham weight. †P Å 0.015 for BDL weight vs CJ weight. ‡P Å 0.26 (ns) for BDL weight vs CJ weight. Lymph flow: †P 0.05 for baseline lymph flow vs lymph flow during infusion. P by one-tailed paired Mann– Whitney test.
AID
JSR 5192
/
6n28h$$302
04-01-98 08:59:59
srga
128
JOURNAL OF SURGICAL RESEARCH: VOL. 74, NO. 2, FEBRUARY 1, 1998
TABLE 3 Mean ({SD) Number of E. coli Colony-Forming Units (cfu 1 104/ml) per Milliliter in Blood Collected after 5 min of Biliary Infusion of an E. coli Suspension Containing 108/ml at Two Infusion Pressures Overpressure, 5 cm H2O
Overpressure, 20 cm H2O
1.9 { 2.7 31 { 36† 2.5 { 4.3‡
446 { 963 4424 { 8426† 81 { 76‡
Sham BDL CJ
P
Overpressure, 5 cm H2O (correcteda)
Overpressure, 20 cm H2O (correcteda)
P
0.03 0.005 0.002
7.6 { 12 33 { 39† 3.9 { 7.3‡
315 { 650 831 { 1467† 28 { 24‡
0.01 0.03 0.006
Note. For both pressure groups: †P õ 0.05, sham vs BDL; ‡P õ 0.05, BDL vs CJ.
tract to the bloodstream compared to the sham-operated controls (5 cm H2O: sham, 1.9 { 2.7 vs BDL, 31 { 36 1 104; P Å 0.01; and 20 cm H2O: sham, 446 { 963 vs BDL, 4424 { 8426 1 104; P Å 0.02). Restoration of bile flow after 2 weeks of cholestasis in CJ rats reduced translocation to the bloodstream to sham control levels in both pressure groups. Translocation of E.coli to the Thoracic Duct after Retrograde Biliary Infusion for 5 min All preinfusion lymph cultures were sterile. At 5 cm H2O overpressure, only one lymph culture among 29 rats showed growth of E.coli (Table 4). This finding was noted in the decompressed group. At 20 cm H2O overpressure, a total 9 cultures from the thoracic duct were positive for E.coli. In this pressure group the number of positive lymph cultures was significantly less in the BDL groups than in the sham and decompressed groups (1/29 vs 9/28; P Å 0.03; Fischer’s exact test). DISCUSSION
This study clearly showed that biliary pressure and a period of bile duct obstruction are determinants of bacterial translocation from the biliary tract to the bloodstream. The major importance of intrabiliary
pressure in regurgitation of bile constituents to the bloodstream is in accordance with earlier studies. In 1947, it was demonstrated by Mixter et al. that contrast media refluxed to the blood after injection into the bile duct [2]. In 1969, Huang et al. showed that the process of bacterial regurgitation, whether via cholangiovenous or lymphatic routes, could be readily initiated in the dog, once the pressure in the CBD reached 20 cm H2O [1]. A similar, pressure-related passage of bacteria from bile ducts to bloodstream has been demonstrated in a rat model [14]. In the present study, retrograde infusion of the CBD was conducted at overpressures of 5 and 20 cm H2O, yielding absolute pressures in the CBD of approximately 25 and 40 cm H2O. These high pressures were necessary to overcome the maximal secretory biliary pressure. It can be argued that such pressures far above the maximal secretory biliary pressure may create artifacts. However, it is known that in case of an infection, pressures in the duct also exceed the maximal secretory biliary pressure [16]. Furthermore, in humans, the maximal secretory biliary pressure is 40 cm H2O and during contraction of the gallbladder, intrabiliary pressure can even rise to 80 cm H2O, which is much higher than the pressures used in the present study [17]. This study also showed a significant increase in bac-
TABLE 4 Absolute Number of Colony Forming Units E. coli per Milliliter in Thoracic Duct Lymph Cultures Collected 5 min after Biliary Infusion of an E. coli Suspension of 108/ml at Two Infusion Pressures Sham, 5 cm H2O (n Å 11)
— — — — — — — — — —
BDL, 5 cm H2O (n Å 10)
CJ, 5 cm H2O (n Å 8)
Sham, 20 cm H2O (n Å 10)
BDL, 20 cm H2O (n Å 9)
CJ, 20 cm H2O (n Å 9)
— — — — — — — — — —
— — — — — 400 — —
1.7 E 7 1200 7500 — 300 — — — — —
— — 4000 — — — — — —
— — — 7500 100 5000 — 10000 —
Note. All groups combined: P Å 0.03 for positive cultures at 5 cm overpressure (1 of 29) vs 20 cm overpressure (9 of 28). Groups with a nonobstructed bile duct (sham and CJ): P Å 0.05 for positive cultures at 5 cm overpressure (1 of 19) vs 20 cm overpressure (8 of 19). Nonsignificant difference in each individual group (sham, BDL, or CJ) between 5 and 20 cm overpressure (Fisher’s exact test).
AID
JSR 5192
/
6n28h$$303
04-01-98 08:59:59
srga
KARSTEN ET AL.: BACTERIAL TRANSLOCATION TO BLOOD AND LYMPH
terial translocation to the systemic circulation after a 2-week period of BDL. The many pathophysiological changes resulting from obstructive jaundice may be involved in this phenomenon. It has been suggested that breakdown of the tight junctions between the hepatocytes damages the anatomical barrier between bile canaliculi and sinusoids and facilitates cholangiovenous reflux [18, 19]. Epithelial disruption of the biliary tract occurring after longstanding bile duct obstruction may also promote reflux [6]. Furthermore, the sinusoids are insufficiently cleared from refluxed bacteria due to dysfunction of the Kupffer cells, enhancing invasion into the systemic circulation [20]. The results of this study, however, are in contrast with other reports in which no increased cholangiovenous reflux after biliary obstruction could be demonstrated [14]. This discrepancy between the results can probably be attributed to the extended period of cholestasis (2 weeks) used in the present study, which induces more severe damage to the liver and the reticuloendothelial system. Furthermore, in the present setup, blood samples were taken from the caval vein at the outflow site of the hepatic veins, thus ensuring sampling of posthepatic blood. Raper et al. used blood samples collected from the femoral artery, which had already passed the lungs [14]. It is known that after bile duct obstruction, when Kupffer cell function is depressed, a compensatory increase of clearance by the lung macrophages and the spleen occurs [21]. Pulmonary clearance of bacteria could therefore account for low bacterial counts in femoral artery blood. After 2 weeks of decompression by choledochojejunostomy the bacterial translocation observed after bile duct obstruction returned to the normal control level. Several studies have readily shown that bile duct decompression restores most of the changes of the liver parenchyma [22, 23]. Complete recovery of liver synthesis function was estimated to be achieved between 4 and 6 weeks [24]. A report by Katz et al. showed complete reversal of hepatocyte phagocyte dysfunction and a decrease in compensatory bacterial trapping after 2 weeks of biliary decompression in rats [25]. The present results are consistent with these findings and provide evidence in favor of preoperative biliary drainage. Further studies should address the optimal duration of biliary decompression in order to reverse changes in translocation due to bile duct obstruction. Lymph flow doubled after bile duct obstruction and returned to normal after decompression. The biliary infusion experiments clearly demonstrated that lymph flow was directly related to intrabiliary pressure, which is consistent with other studies [1, 26]. The data from the lymph cultures, however, did not parallel those from the blood cultures. This is in contrast with studies by Stewart [27], Yamamoto [28] in rats, and Huang in dogs [1], which all suggested that reflux to the lymphatics occurred before and at lower pressures than reflux to the bloodstream. However, Stewart and Yamamoto used particulate matter to
AID
JSR 5192
/
6n28h$$303
04-01-98 08:59:59
129
identify paths of reflux, and only in the study by Huang were viable bacteria used. It is generally assumed that hepatic lymph originates in the interstitial space of the liver parenchyma, i.e., the space’s of Disse and Mall [29]. The lymph tract starts in the terminal lymphatics in the portal areas, which drain into collecting vessels along the veins. Hepatic lymph ultimately drains by two routes in the thoracic duct [26, 30–32]. The first route (sampled in this study) follows the hepatoduodenal ligament and ends alongside the superior mesenteric artery into the cisternae chyli which, in turn, drain into the abdominal part of the thoracic duct. The second route parallels the caval vein and enters the thoracic duct at the level of the right subclavian vein. The mechanism by which bacterial translocation predominantly takes place to the bloodstream and not to the lymphatic system is not clear. Although the second extrahepatic lymph route was not canulated in this study, it is unlikely that bacterial translocation would be directed only to the lymph drainage system alongside the caval vein. In addition, bacterial clearance of the lymph by lymph nodes along the route to the thoracic duct, or shunting of bacteria through lymphaticovenous communications remains a matter of speculation (33). This study answers no questions concerning the pathway that bacteria take in the liver parenchyma, during their passage from bile duct to either the sinusoids or the lymph tract. In theory, this can be either a transcellular or paracellular route at the level of the bile canaliculus or the larger bile ducts. Earlier studies showed, at pressures up to 50 cm H2O, that reflux occurred from bile ductules to the spaces of Disse and Mall [27, 28]. Only by blocking these low-resistance pathways can passage via biliary canaliculi occur. Another study showed that bacterial reflux occurred predominantly across the hepatocyte, by transcellular pathways [14]. The relevance of these pathways and their relation to bile duct pressure and various bacterial species require further investigation. In conclusion, this study confirms that the translocation of bacteria from biliary tract to bloodstream increases at higher intrabiliary pressures. It was also shown that longstanding bile duct obstruction is an independent determinant for cholangiovenous reflux. Bacterial migration to the lymphatic system did not parallel translocation to the bloodstream, although in a nonobstructed biliary tract bacterial translocation to lymph seemed to be pressure-related. ACKNOWLEDGMENTS The authors are grateful to Mr. H. de Wit for his skillful assistance during the surgical experiments and to Mrs. P. Brinke and Mrs. E. Reurekas and Mr. G. Shatorje, for their assistance in performing the quantitative analysis of the blood and lymph cultures. The authors express their gratitude Mr. F. van Zijderveld who kindly provided the stock culture of E.coli and the monoclonal antibodies.
REFERENCES 1. Huang, T., Bass, J. A., and Williams, R. D. The significance of biliary pressure in cholangitis. Arch. Surg. 98: 629–632, 1969.
srga
130
JOURNAL OF SURGICAL RESEARCH: VOL. 74, NO. 2, FEBRUARY 1, 1998
2. Mixer, H., Rigler, L., and Oddone, M. Experimental studies on biliary regurgitation during cholangiography. Gastroenterology 9: 64–80, 1947. 3. Jacobsson, B., Kjellander, J., and Rosengren, B. Cholangiovenous reflux. An experimental study. Acta Chir. Scand. 123: 316– 321, 1962. 4. Yoshimoto, H., Ikeda, S., Tanaka, M., and Matssumoto, S. Relationship of biliary pressure to cholangiovenous reflux during endoscopic retrograde ballon catheter cholangiography. Dig. Dis. Sci. 34: 16–20, 1989. 5. Desmet, V. J. Cholestasis: extrahepatic obstruction and secondary biliary cirrhosis. In McSween, R. N. M., Anthony, P. P., and Scheuer, P. J. (Eds.), Pathology of the Liver, 2nd ed. Edinburgh: Churchill–Livingstone, 1987. pp. 364–423. 6. Karsten, T. M., Coene, P. P. L. O., van Gulik, T. M., Bosma, A., et al. Morphological changes of extrahepatic bile ducts during obstruction and subsequent decompression by endoprosthesis. Surgery 111: 562–568, 1992. 7. Shibayama, Y. Endotoxaemia and hepatic injury in obstructive jaundice. J. Pathol. 159: 335–339, 1989. 8. Robenek, H., Herwig, J., and Themann, H. The morphologic characteristics of intracellular junctions between normal human liver cells and cells from patients with extrahepatic cholestasis. Am. J. Pathol. 100: 93–114, 1980. 9. Landman, L. Cholestasis induced alterations of the trans- and paracellular pathways in rat hepatocytes. Histochemistry 103: 3–9, 1995. 10. Andy, O. J., Grogan, J. B., Griswold, J. A., and Scott-Conner, C. E. H. Peritoneal neutrophil chemotaxis is impaired in biliary obstruction. Am. Surg. 58: 28–32, 1992. 11. Scott-Conner, C. E. H., Grogan, J. B., Scher, K. S., and Bernstein, J. Impaired clearance of Escherichia coli bacteremia in early biliary obstruction. Am. J. Surg. 157: 210–214, 1989. 12. Ding, J. W., Andersson, R., Soltesz, V., Willen, R., and Bengmark, S. Obstructive jaundice impaires reticuloendothelial function and promotes bacterial translocation in the rat. J. Surg. Res. 57: 238–245, 1994. 13. Rank, J., and Wilson, I. D. Changes in IgA following varying degrees of biliary obstruction in the rat. Hepatology 3: 241– 247, 1983. 14. Raper, S. E., Barker, M. E., Jones, A. L., and Way, L. W. Anatomic correlates of bacterial cholangiovenous reflux. Surgery 105: 352–359, 1989. 15. Zijderveld van, F. G., Westenbrink, F., Anakotta, J., Brouwers, R. A. M., et al. Characterization of the F41 fimbrial antigen of enterotoxigenic Escherichia coli by using monoclonal antibodies. Infect. Immun. 57: 1192–1199, 1989. 16. Csendes, A., Sepulveda, A., Burdiles, A., Braghetto, I., et al. Common bile duct pressure in patients with CBD stones with or without acute suppurative cholangitis. Arch. Surg. 123: 697– 699, 1988.
AID
JSR 5192
/
6n28h$$304
04-01-98 08:59:59
17. Toresoli, A., Corraziari, E., Habib, F. C., and Cicala, M. Pressure relationships within the human bile tract. Normal and abnormal physiology. [Review]. Scan. J. Gastroenterol. Suppl. 175: 52–57, 1990. 18. Ker, C. G., Huang, T. J., and Sheen, P. C. Electron microscopic assessment of bile regurgitation of human in extrahepatic obstructive jaundice. Eur. Surg. Res. 17: 197–206, 1985. 19. Toyota, N., Miyai, K., and Hardison, W. G. M. Effect of biliary pressure versus high bile acid flux on the permeability of hepatocellular tight junction. Lab. Invest. 50: 536–543, 1984. 20. Bradfield, J. W. B. Control of spillover: The importance of Kupffer cell function in clinical medicine. Lancet 2: 883–86, 1974. 21. Katz, S., Grosfeld, J. L., and Gross, K. Impaired clearance and trapping in obstructive jaundice. Ann. Surg. 199: 14–20, 1984. 22. Abdel-Aziz, G., Lebeau, G., Rescan, P. Y., Clement, B., et al. Reversibility of hepatic fibrosis in experimentally induced cholestasis in the rat. Am. J. Pathol. 137: 1333–1342, 1990. 23. Zimmerman, H., Reichen, J., Zimmerman, A., Sagesser, H., et al. Reversibility of secondary biliary fibrosis by biliodigestive anastomosis in the rat. Gastroenterology 103: 579–89, 1992. 24. Koyama, K., Takagi, Y., Ito, K., and Sato, T. Experimental and clinical studies on the effect of experimental obstructive jaundice. Am. J. Surg. 158: 423–427, 1989. 25. Katz, S., Yang, R., Rodefeld, M. J., Folkening, W. J., et al. Impaired hepatic clearance is reversed by surgical relief of obstructive jaundice. J. Pediatr. Surg. 26: 401–406, 1991. 26. Gyure, L. A., Hall, J. G., Lapis, K., and Thimar, J. The relief of biliary obstruction in experimental rats by diverting the hepatic lymph into the duodenum. Int. J. Exp. Pathol. 72: 345–348, 1991. 27. Stewart, L., Pellegrini, C. A., and Way, L. W. Cholangiovenous reflux pathways as defined by corrosion casting and scanning electron microscopy. Am. J. Surg. 155: 23–28, 1988. 28. Yamamoto, K., and Phillips, M. J. A hitherto unrecognized bile ductular plexus in the normal rat liver. Hepatology 4: 381–385, 1984. 29. Henriksen, J., Horn, T., and Christoffersen, P. The blood– lymph barrier in the liver. A review based on morphological and functional concept of normal and cirrhotic liver. Liver 4: 221–232, 1984. 30. Tilney, N. L. Patterns of lymphatic drainage in the adult laboratory rat. J. Anat. 109: 369–383, 1971. 31. Bollman, J. L., Cain, J. C., and Grindlay, J. H. Techniques for the collection of lymph from the liver, small intestine, or thoracic duct of the rat. J. Lab. Clin. Med. 33: 1349–1352, 1948. 32. Friedman, M., Byers, S. O., and Omoto, C. Some characteristics of hepatic lymph in the intact rat. Am. J. Physiol. 184: 11–17, 1956. 33. Hidden, G., Menard, P., and Zorn, J. Y. Lymphaticovenous communications. Anat. Clin. 7: 83–91, 1985.
srga