Complications in Obstructive Jaundice: Role of Endotoxins J. W. M. GREVE, D. J. G O U M A & W. A. BUURMAN Dept. of Surgery, University of Limburg. Maastricht, The Netherlands
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
Greve JWM, Gouma DJ. Buurman WA. Complications in obstructive jaundice: role of endotoxins. Scand J Gastroenterol 1992;27 Suppl 194%-12. Surgical treatment of patients with obstructive jaundice is associated with a high postoperative morbidity and mortality. A correlation was suggested between endotoxins and the observed complications. The mechanism by which endotoxins affect the negative outcome in operated jaundiced patients was, however, not clear. nor was the mechanism of clinically used preventive treatments. Several experiments were therefore performed in rats with biliary obstruction, to investigate whether and how endotoxins are active. The role of endotoxins was studied in a model in which endotoxins were absent. In germfree rats (free of bacteria and thus of endotoxin) the effect of biliary obstruction was studied and compared with biliary obstruction in conventional rats. To substantiate further the role of endotoxin, anti-endotoxin treatments (oral lactulose or internal drainage) were tested in rats with obstructive jaundice undergoing a severe surgical trauma. It is shown that endotoxins are responsible for complications (suppression of cellular immunity, kidney function. mortality) and that these complications can be prevented with an anti-endotoxin treatment. These results may have implications for preoperative treatment of jaundiced patients. Key words: Cellular immunity; endotoxin; internal biliary drainage; lactulose; mortality; obstructive jaundice: renal failure; tumor necrosis factor Jan Willem M. Greve, M . D . , Dept. of Surgery, University of Limburg, P.O. Box 5800, NL-6202 AZ Maastricht, The Netherlands
It is well known that patients with obstructive jaundice are more prone to postoperative complications than nonjaundiced patients (1,2). Most frequent are postoperative infections and renal failure (2,3). High plasma bilirubin levels, high plasma bile acid levels, and impairment of the nutritional status were suggested as a possible cause, but were found not to be involved in the pathogenesis of these complications. Another factor that was suggested to play an important role in the pathogenesis of the enhanced morbidity and mortality in obstructive jaundice is the systemic and/or portal endotoxemia frequently observed in these patients (4,5). The endotoxemia in obstructive jaundice is considered to result from an increased absorption of gut-derived endotoxins. The mechanism by which endotoxins negatively influence the postoperative outcome of jaundiced patients is, however, not known. The immune system is suggested to play an important role in the pathogenisis of the infectious complications. Experimental bile duct ligation in animals resulted in a suppression of cellular immunity (6,7). In patients with obstructive jaundice a similar reduction of cellular immunity was found (8). Since endotoxemia disappeared after internal biliary drainage in rats and since it was found in another study that cellular immunity recovered after internal drainage, it has been suggested that endotoxins affect the immune system (7,9). On the other hand, several clinical studies have suggested that endotoxins are related to the renal impairment frequently observed in jaundiced patients (2,4,10). These
observations were confirmed by experimental studies (11,12). OBJECTIVES A first series of experiments was performed to study the role of endotoxins with regard to the impaired immune system. It is difficult, however, to study selectively the role of endotoxins in a model with experimental biliary obstruction. Therefore germfree rats, in which endotoxins are virtually absent, were used to study the effect of biliary obstruction on the immune system without the presence of endotoxins. The results were compared with the effect of biliary obstruction in conventional rats. In a second series of experiments the effect of anti-endotoxin treatments (for example, oral lactulose or internal biliary drainage) was tested in jaundiced rats undergoing a severe surgical trauma, to further substantiate the role of endotoxins in biliary obstruction, with emphasis on kidney function and mortality. EXPERIMENTS Endotoxins and cellular immunity in obstructive jaundice The effect of endotoxins on cellular immunity was studied in a rat model with experimental biliary obstruction. In germfree rats (carrying no gram-negative bacteria and thus virtually no endotoxin) a biliary obstruction was obtained by ligation and transection of the common bile duct (13).
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
Complications in Obstructive Jaundice
The control group underwent a sham operation. Endotoxins in portal and systemic plasma were determined with a chromogenic limulus amebocyte lysate test (lower detection limit, 1-10 pg/ml) (14). Two weeks after the biliary obstruction or sham operation, cellular immunity was examined by culturing spleen lymphocytes taken from the rats. After mitogen stimulation lymphocyte reactivity was determined by [3H]thymidine incorporation (15). As expected, no plasma endotoxin could be detected in the germfree rats. After bile duct ligation in the germfree rats no alteration of the cellular immunity was observed. Bile duct ligation in conventional (Wistar) rats resulted in a significant suppression of cellular immunity as compared with sham (7223 2 5494 cpm versus 39,570 19,530 cpm; p < 0.01) (15). This suppression of the cellular immunity could not be related directly to endotoxins, since neither systemic nor portal endotoxin levels were significantly increased in jaundiced rats. However, the main difference between germfree and conventional rats is the absence of bacteria and thus of endotoxins, suggesting a correlation of endotoxins and suppression of cellular immunity in obstructive jaundice. However, other bacterial products could also be responsible for the observed effect. Therefore the effect of administering a chronic low dose of endotoxins on cellular immunity was studied. Rats in the treatment group received 250 kg endotoxin/100 g body weight every 8 h during 3 days; controls received a saline solution. Thereafter the rats were killed to study immunity. Suppression of cellular immunity in the endotoxin-treated Wistar rats was comparable to that in jaundiced conventional rats (3404 ? 1119 cpm versus 14,670 c 4551 cpm;p < 0.01). Moreover, in contrast to biliary obstruction chronic low-dose endotoxin did suppress cellular immunity in germfree rats (47,849 ? 7211 cpm versus 65,143 2 7221 cpm; p < 0.01). To substantiate the role of endotoxins, the production of spontaneous tumor necrosis factor (TNF) by peritoneal macrophages from all the rats in the experiment were studied (TNF is an important mediator of endotoxin toxicity, which is mainly produced by mononuclear phagocytes). TNF production by the peritoneal macrophages was measured with a cytotoxicity assay with L929 mouse cells, as described by Spofford et a1 (16). Peritoneal macrophages from conventional jaundiced rats released significantly more TNF than macrophages from control rats (1254 487 pg/ml versus 220 +- 259 pg/ml; p < 0.001). Peritoneal macrophages from bile duct-ligated and sham-treated germfree, endotoxin-'free' rats did not spontaneously release TNF (25 2 20.2 pg/ml versus 26.5 20.7 pg/ml; NS). Finally, the spontaneous TNF production by the peritoneal macrophages of each individual rat was correlated to the cellular immunity. A significant correlation was found between TNF levels and suppression of cellular immunity (r = -0.717, p < 0.004) in the conventional rats. These
*
*
9
results suggest a relation between endotoxins (inducing TNF release) and suppression of cellular immunity. Effects of preoperative treatment on survival and kidney function in jaundiced rats Throughout this experiment male Wistar rats were used. The study was designed to evaluate the effect of preoperative treatment in jaundiced rats on the outcome after a severe surgical trauma. Since the kidney is at risk in jaundiced subjects bilateral renal ischemia was chosen as the surgical trauma. Preoperative treatment consisted of either the antiendotoxin lactulose (17) (Duphar, Weesp, The Netherlands) or internal biliary drainage. Five experimental groups were formed. These groups were sham-operated rats (Sham); bile duct-ligated rats (BDL); bile duct ligation followed by internal biliary drainage (BDL-Drain); bile duct ligation treated with oral lactulose (BDL-Lact), and, finally, a BDL group undergoing a sham renal ischemia (BDL-SI). Variables were blood urea nitrogen (BUN) and survival (18). Oral lactulose was given by gavage three times a day during 3 days before renal ischemia; all other rats were given similar amounts of saline by gavage. Internal biliary drainage was performed by inserting a silastic tube in the dilated common bile duct, connecting it with the duodenum. In all rats this resulted in complete normalization of bilirubin levels within 2 weeks. Bilateral renal ischemia in Sham rats resulted in a sharp rise in BUN levels, indicative of kidney function, which recovered after 1-3 days. In BDL rats the rise in BUN was significantly higher during the first 24 h compared with Sham (change in BUN in mmol/l: Sham, 41.6 5 6.7 versus BDL, 54.7 5 4.2; p < 0.001). BDL rats developed complete renal failure without recovery. Since most BDL rats died within 48 h, further statistical analysis was not possible. Sham ischemia in BDL-SI rats did not affect BUN. Anti-endotoxin treatment (oral lactulose) did not significantly reduce deterioration of renal function after renal ischemia (change in BUN in mmol/l: BDL-Lact, 47.8 ? 12.6 versus BDL, 54.7? 4.2; NS). However, a tendency to improve was noted, since changes in BUN in BDL-Lact rats were not significantly different from those in Sham rats, whereas renal impairment in BDL rats was (change in BUN in mmol/l: BDL-Lact, 47.8 ? 12.6 versus sham, 41.6 2 4.7; NS). Internal biliary drainage resulted in a significantly reduced renal impairment compared with both BDL and BDL-Lact rats (change in BUN in mmol/l: BDL-Drain, 34.0 2 15.6 mmol/l versus BDL; p < 0.001; BDL-Drain versus BDLLact; p < 0.05). Change in BUN in BDL-Drain was not significantly different from that in Sham rats. Mortality after renal ischemia was recorded during 7 days, and thereafter the remaining rats were killed. After renal ischemia in Sham rats 4 of 10 rats died, resulting in a survival time of 134 C 49 h. No BDL rat survived after renal
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
10
J . W.M. Greve et al.
ischemia; the survival time of 54 f 35 h was significantly lower than that of Sham rats (p c 0.001). BDL-SI rats all survived the 7-day observation period after sham renal ischemia. In contrast to renal function, lactulose treatment of BDL rats significantly improved survival time after renal ischemia (BDL-Lact, 125 2 53 h versus BDL 54 2 35 h; p < 0.01). Moreover, survival time of BDL-Lact rats was not significantly different from that of Sham rats, Survival time in rats with internal biliary drainage was increased significantly compared with BDL rats (BDL-Drain 156 2 35 h versus 54 ? 35 h; p < 0.001) and was not significantly different from that of BDL-Lact and Sham rats. The experimental model did allow daily monitoring of plasma endotoxins; therefore TNF plasma levels were measured. In jaundiced rats significant amounts of TNF were
found. These TNF levels did not correlate with mortality. An explanation could be the very short half-life of TNF and the intermittent appearance in plasma, making single daily samples unreliable. As was recently shown, TNF stimulates a de novo synthesis of triglycerides besides inhibiting lipoprotein lipase, both resulting in increased plasma triglyceride levels (19,20). We therefore also measured plasma triglyceride levels as indicator of TNF release and indirectly of endotoxins. In jaundiced rats a significant relation was found between serum triglycerides and survival (triglycerides, 2.8 2 0.9 mmol/l versus 5.9 2 1.3 mmol/l for surviving versus decreased jaundiced rats, respectively; p < 0.001). DISCUSSION It was known from previous experiments that biliary obstruc-
OBSTRUCTIVE JAUNDICE
I
HISTORY LABORATORY ULTRA SOUND ' CT-SCAN
*
DIAGNOSIS
* *
1
1
ERCP
I + NO CANULATION or CANULATION NO @RAIN
I I
1
INTERNAL DRAIN
EXTERNAL DRAIN or NO DRAIN INOPERABLE
-4
I
ANTIBIOTICS
1
I I
I
I
c 3d
I
(RADICAL) RESECTION
I
*3w
I
PALLIATION (DRAIN)
1
Fig. 1. Proposed decision tree for supportive preoperative therapy in jaundiced patients. The preventive treatments are shown in the relation to the means to obtain them. Shadow boxes show the desired supportive therapy. CT = computed tomography; ERCP = endoscopic retrograde cholangiopancreaticography; PTC-D = percutaneous transhepatic cholangiography (drainage).
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
Complications in Obstructive Jaundice
tion affects the immune system. In rats and mice with experimental biliary obstruction suppression of the specific cellular immunity was found (6,7). In these studies the hyperbilirubinemia o r the high bile acid levels were suggested to be responsible . In the present experiments cellular immunity was significantly suppressed in conventional rats with obstructive jaundice, supporting the mentioned experiments. In germfree (endotoxin-'free') jaundiced rats, however, cellular immunity was not impaired. Despite the absence of endotoxemia in conventional BDL rats, strong evidence was found for a crucial role of endotoxins with regard to the impaired immune system. In germfree rats endotoxin levels in feces were at least a thousand times lower than in conventional rats, being one of the major differences between the two rat strains. Moreover, chronic low-dose endotoxin administration induced similar defects of the cellular immunity as bile duct ligation in conventional rats. Finally, a correlation was shown between spontaneous TNF production by peritoneal macrophages, as indicator of endotoxin toxicity, and suppression of cellular immunity in jaundiced rats. A possible explanation of the unexpected absence of endotoxins is the difficult endotoxin assay. However, a recent study, in which a similar sensitive endotoxin assay was used, confirmed our results (21). In summary, evidence was given by the first experiment that suppression of cellular immunity is caused by endotoxins from the gastrointestinal tract. The results of this study suggested that one of the aims of preoperative supportive therapy in jaundiced patients should be prevention of the toxic effect of endotoxins from the bowel lumen. The clinical and experimental observation that endotoxins are related to impairment of kidney function and death could not be directly confirmed in our experiments (2, 4, 10-12). As mentioned above, no substantial endotoxemia was found in our jaundiced rats. However, the second experiment provided indirect evidence that endotoxins d o play an important role. In two separate studies lactulose was shown to inhibit the toxic effects of endotoxin (17,22). In the present study preventive treatment with oral lactulose significantly reduced mortality after renal ischemia in jaundiced rats, and there was a clear but not statistically significant tendency for it to reduce the impairment of kidney function. In the rats treated with lactulose no other variables improved, which suggests that the observed effect must be due to inhibition of the toxic effect of endotoxin. The superior effect of internal biliary drainage on kidney function and mortality after renal ischemia. compared with lactulose, was to be expected. Internal biliary drainage not only restores the bile flow to the gut, thus inactivating endotoxins in the bowel lumen with bile acids, but also restores normal liver function, reestablishing the effective clearance of endotoxins present in portal blood, providing an optimal protection against the toxic effect of endotoxins. In conclusion, the two presented experiments provide
11
evidence that endotoxins play an important role in the etiology of infections and in renal complications that occur after surgery in patients with obstructive jaundice. The experimental data suggest that these complications can be prevented by a preoperative treatment, such as oral lactulose or internal biliary drainage, which reduces endotoxin toxicity. Extrapolating these experimental data to the clinical situation could result in a flow diagram as presented in Fig. 1. After the diagnosis of obstructive jaundice, which in most cases will be due to a malignancy, a sequence of diagnostic and therapeutic interventions will occur. Depending on the degree of jaundice, the cause, and the operability of the patient, an attempt is made to cannulate the obstructed biliary tract by endoscopic retrograde cholangiopancreatography , internal biliary drainage being the preoperative treatment of choice. After successful internal biliary drainage at least 10-14 days' recovery of liver functions and immune system is allowed, followed by a planned surgical procedure. In the event of an unsuccessful or partial biliary drainage, infectious complications often occur. In these patients antibiotic treatment (given prophylactically) is continued with the addition of an anti-endotoxin treatment such as oral lactulose. The patient should then be operated on within several days. It is evident that prospective randomized trials are necessary to validate such a treatment protocol. REFERENCES 1. Pain JA, Cahill CJ, Bailey ME. Perioperative complications in obstructive jaundice: therapeutic considerations. Br J Surg
1985;72:942-5. 2. Wait RB, Kahng KU. Renal failure complicating obstructive jaundice. Am J Surg 1989;157:256-63. 3. Armstrong CP, Dixon JM, Taylor TV, Davies GC. Surgical experience of deeply jaundiced patients with bile duct obstruction. Br J Surg 1984;71:234-8. 4. Wardle EN, Wright NA. Endotoxin and acute renal failure associated with obstructive jaundice. Br Med J 1970;4:472-4. 5. Pain JA, Bailey ME. Measurement of operative plasma endotoxin levels in jaundiced and non-jaundiced patients. Eur Surg Res 1987;19:207-16. 6. Pinto M, Kaplun A. Immune status in mice with experimental biliary obstruction. Clin Immunol lmmunopathol 1980;16:396405. 7. Roughneen PT, Gouma DJ, Kulkarni A D , Fanslow WF, Rowlands BJ. Impaired specific cell-mediated immunity in experimental biliary obstruction and its reversibility by internal biliary drainage. J Surg Res 1986;41:113-25. 8. Fargion SR, Cappellini MD, Podda M, Fiorelli G . Immunita cellulare nelle colestasi intra ed extra-epatiche. Min Gastr 1976;22:261-5. 9. Gouma DJ, Coelho JCU, Fisher JD, Schlegel JF, Li YF, Moody FG. Endotoxemia after relief of biliary obstruction by internal and external drainage in rats. Am J Surg 1986;151:476-9. 10. Wilkinson SP, Moody H, Stamatakis JD, Kakkar VV, Williams R. Endotoxaemia and renal failure in cirrhosis and obstructive jaundice. Br Med J 1976;2:1415-8. 1 1 . Bailey ME. Endotoxin, bile salts and renal function in obstructive jaundice. Br J Surg 1976;63:774-8. 12. Cahill CJ, Pain JA, Bailey ME. Bile salts, endotoxin and renal function in obstructive jaundice. Surg Gynecol Obstet 1987;165: 519-22. 13. Lee E. The effect of obstructive jaundice on the migration of
12
J . W.M. Greve et al.
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
reticulo-endo-thelial cells and fibroblasts into early experimental granulomata. Br J Surg 1972;59:875-7. 14. Olofsson P. Application of a quantitative spectrophotometric endotoxin assay on lymph and plasma from the rat. Eur Surg Res 1986;18:112-21. 15. Greve JW, Gouma DJ, Soeters PB, Buurman WA. Suppression of cellular immunity in obstructive jaundice is caused by endotoxins. A study with germfree rats. Gastroenterology 1990; 98~478-85. 16. Spofford BT, Daynes RA, Granger GA. Cell-mediated immunity in vitro: a highly sensitive assay for human lymphotoxin. J Immunol 1974;112:2111-6. 17. Greve JW, Gouma DJ, v Leeuwen PAM, Buurman WA. Lactulose inhibits endotoxin-induced tumour necrosis factor production by monocytes. An in vitro study. Gut 1990;31:198-203.
18. Greve JW, Maessen JG, Tiebasch T, Buurman W A , Gouma DJ. Prevention of postoperative complications in jaundiced rats. Internal biliary drainage versus oral lactulose. Ann Surg 1990; 212~221-7. 19. Beutler B, Mahony J , Le Trang N, Pekala P, Cerami A. Purification of cachectin, a lipoprotein lipase-suppressing hormone secreted by endotoxin induced rat 264.7 cells. J Exp Med 1985; 161~984-95. 20. Feingold KR, Serio MK, Adi S, Moser AH, Grunfeld. Tumor necrosis factor stimulates hepatic lipid synthesis and secretion. Endocrinology 1989;124:2336-42. 21. Roughneen PT, Kumar SC, Pellis NR, Rowlands BJ. Endotoxemia and cholestasis. Surg Gynecol Obstet 1988;167:205-10. 22. Pain JA, Bailey ME. Experimental andclinical study of lactulose in obstructive jaundice. Br J Surg 1986;73:775-8.
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
Greve JWM, Gouma DJ. Buurman WA. Complications in obstructive jaundice: role of endotoxins. Scand J Gastroenterol 1992;27 Suppl 194%-12. Surgical treatment of patients with obstructive jaundice is associated with a high postoperative morbidity and mortality. A correlation was suggested between endotoxins and the observed complications. The mechanism by which endotoxins affect the negative outcome in operated jaundiced patients was, however, not clear. nor was the mechanism of clinically used preventive treatments. Several experiments were therefore performed in rats with biliary obstruction, to investigate whether and how endotoxins are active. The role of endotoxins was studied in a model in which endotoxins were absent. In germfree rats (free of bacteria and thus of endotoxin) the effect of biliary obstruction was studied and compared with biliary obstruction in conventional rats. To substantiate further the role of endotoxin, anti-endotoxin treatments (oral lactulose or internal drainage) were tested in rats with obstructive jaundice undergoing a severe surgical trauma. It is shown that endotoxins are responsible for complications (suppression of cellular immunity, kidney function. mortality) and that these complications can be prevented with an anti-endotoxin treatment. These results may have implications for preoperative treatment of jaundiced patients. Key words: Cellular immunity; endotoxin; internal biliary drainage; lactulose; mortality; obstructive jaundice: renal failure; tumor necrosis factor Jan Willem M. Greve, M . D . , Dept. of Surgery, University of Limburg, P.O. Box 5800, NL-6202 AZ Maastricht, The Netherlands
It is well known that patients with obstructive jaundice are more prone to postoperative complications than nonjaundiced patients (1,2). Most frequent are postoperative infections and renal failure (2,3). High plasma bilirubin levels, high plasma bile acid levels, and impairment of the nutritional status were suggested as a possible cause, but were found not to be involved in the pathogenesis of these complications. Another factor that was suggested to play an important role in the pathogenesis of the enhanced morbidity and mortality in obstructive jaundice is the systemic and/or portal endotoxemia frequently observed in these patients (4,5). The endotoxemia in obstructive jaundice is considered to result from an increased absorption of gut-derived endotoxins. The mechanism by which endotoxins negatively influence the postoperative outcome of jaundiced patients is, however, not known. The immune system is suggested to play an important role in the pathogenisis of the infectious complications. Experimental bile duct ligation in animals resulted in a suppression of cellular immunity (6,7). In patients with obstructive jaundice a similar reduction of cellular immunity was found (8). Since endotoxemia disappeared after internal biliary drainage in rats and since it was found in another study that cellular immunity recovered after internal drainage, it has been suggested that endotoxins affect the immune system (7,9). On the other hand, several clinical studies have suggested that endotoxins are related to the renal impairment frequently observed in jaundiced patients (2,4,10). These
observations were confirmed by experimental studies (11,12). OBJECTIVES A first series of experiments was performed to study the role of endotoxins with regard to the impaired immune system. It is difficult, however, to study selectively the role of endotoxins in a model with experimental biliary obstruction. Therefore germfree rats, in which endotoxins are virtually absent, were used to study the effect of biliary obstruction on the immune system without the presence of endotoxins. The results were compared with the effect of biliary obstruction in conventional rats. In a second series of experiments the effect of anti-endotoxin treatments (for example, oral lactulose or internal biliary drainage) was tested in jaundiced rats undergoing a severe surgical trauma, to further substantiate the role of endotoxins in biliary obstruction, with emphasis on kidney function and mortality. EXPERIMENTS Endotoxins and cellular immunity in obstructive jaundice The effect of endotoxins on cellular immunity was studied in a rat model with experimental biliary obstruction. In germfree rats (carrying no gram-negative bacteria and thus virtually no endotoxin) a biliary obstruction was obtained by ligation and transection of the common bile duct (13).
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
Complications in Obstructive Jaundice
The control group underwent a sham operation. Endotoxins in portal and systemic plasma were determined with a chromogenic limulus amebocyte lysate test (lower detection limit, 1-10 pg/ml) (14). Two weeks after the biliary obstruction or sham operation, cellular immunity was examined by culturing spleen lymphocytes taken from the rats. After mitogen stimulation lymphocyte reactivity was determined by [3H]thymidine incorporation (15). As expected, no plasma endotoxin could be detected in the germfree rats. After bile duct ligation in the germfree rats no alteration of the cellular immunity was observed. Bile duct ligation in conventional (Wistar) rats resulted in a significant suppression of cellular immunity as compared with sham (7223 2 5494 cpm versus 39,570 19,530 cpm; p < 0.01) (15). This suppression of the cellular immunity could not be related directly to endotoxins, since neither systemic nor portal endotoxin levels were significantly increased in jaundiced rats. However, the main difference between germfree and conventional rats is the absence of bacteria and thus of endotoxins, suggesting a correlation of endotoxins and suppression of cellular immunity in obstructive jaundice. However, other bacterial products could also be responsible for the observed effect. Therefore the effect of administering a chronic low dose of endotoxins on cellular immunity was studied. Rats in the treatment group received 250 kg endotoxin/100 g body weight every 8 h during 3 days; controls received a saline solution. Thereafter the rats were killed to study immunity. Suppression of cellular immunity in the endotoxin-treated Wistar rats was comparable to that in jaundiced conventional rats (3404 ? 1119 cpm versus 14,670 c 4551 cpm;p < 0.01). Moreover, in contrast to biliary obstruction chronic low-dose endotoxin did suppress cellular immunity in germfree rats (47,849 ? 7211 cpm versus 65,143 2 7221 cpm; p < 0.01). To substantiate the role of endotoxins, the production of spontaneous tumor necrosis factor (TNF) by peritoneal macrophages from all the rats in the experiment were studied (TNF is an important mediator of endotoxin toxicity, which is mainly produced by mononuclear phagocytes). TNF production by the peritoneal macrophages was measured with a cytotoxicity assay with L929 mouse cells, as described by Spofford et a1 (16). Peritoneal macrophages from conventional jaundiced rats released significantly more TNF than macrophages from control rats (1254 487 pg/ml versus 220 +- 259 pg/ml; p < 0.001). Peritoneal macrophages from bile duct-ligated and sham-treated germfree, endotoxin-'free' rats did not spontaneously release TNF (25 2 20.2 pg/ml versus 26.5 20.7 pg/ml; NS). Finally, the spontaneous TNF production by the peritoneal macrophages of each individual rat was correlated to the cellular immunity. A significant correlation was found between TNF levels and suppression of cellular immunity (r = -0.717, p < 0.004) in the conventional rats. These
*
*
9
results suggest a relation between endotoxins (inducing TNF release) and suppression of cellular immunity. Effects of preoperative treatment on survival and kidney function in jaundiced rats Throughout this experiment male Wistar rats were used. The study was designed to evaluate the effect of preoperative treatment in jaundiced rats on the outcome after a severe surgical trauma. Since the kidney is at risk in jaundiced subjects bilateral renal ischemia was chosen as the surgical trauma. Preoperative treatment consisted of either the antiendotoxin lactulose (17) (Duphar, Weesp, The Netherlands) or internal biliary drainage. Five experimental groups were formed. These groups were sham-operated rats (Sham); bile duct-ligated rats (BDL); bile duct ligation followed by internal biliary drainage (BDL-Drain); bile duct ligation treated with oral lactulose (BDL-Lact), and, finally, a BDL group undergoing a sham renal ischemia (BDL-SI). Variables were blood urea nitrogen (BUN) and survival (18). Oral lactulose was given by gavage three times a day during 3 days before renal ischemia; all other rats were given similar amounts of saline by gavage. Internal biliary drainage was performed by inserting a silastic tube in the dilated common bile duct, connecting it with the duodenum. In all rats this resulted in complete normalization of bilirubin levels within 2 weeks. Bilateral renal ischemia in Sham rats resulted in a sharp rise in BUN levels, indicative of kidney function, which recovered after 1-3 days. In BDL rats the rise in BUN was significantly higher during the first 24 h compared with Sham (change in BUN in mmol/l: Sham, 41.6 5 6.7 versus BDL, 54.7 5 4.2; p < 0.001). BDL rats developed complete renal failure without recovery. Since most BDL rats died within 48 h, further statistical analysis was not possible. Sham ischemia in BDL-SI rats did not affect BUN. Anti-endotoxin treatment (oral lactulose) did not significantly reduce deterioration of renal function after renal ischemia (change in BUN in mmol/l: BDL-Lact, 47.8 ? 12.6 versus BDL, 54.7? 4.2; NS). However, a tendency to improve was noted, since changes in BUN in BDL-Lact rats were not significantly different from those in Sham rats, whereas renal impairment in BDL rats was (change in BUN in mmol/l: BDL-Lact, 47.8 ? 12.6 versus sham, 41.6 2 4.7; NS). Internal biliary drainage resulted in a significantly reduced renal impairment compared with both BDL and BDL-Lact rats (change in BUN in mmol/l: BDL-Drain, 34.0 2 15.6 mmol/l versus BDL; p < 0.001; BDL-Drain versus BDLLact; p < 0.05). Change in BUN in BDL-Drain was not significantly different from that in Sham rats. Mortality after renal ischemia was recorded during 7 days, and thereafter the remaining rats were killed. After renal ischemia in Sham rats 4 of 10 rats died, resulting in a survival time of 134 C 49 h. No BDL rat survived after renal
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
10
J . W.M. Greve et al.
ischemia; the survival time of 54 f 35 h was significantly lower than that of Sham rats (p c 0.001). BDL-SI rats all survived the 7-day observation period after sham renal ischemia. In contrast to renal function, lactulose treatment of BDL rats significantly improved survival time after renal ischemia (BDL-Lact, 125 2 53 h versus BDL 54 2 35 h; p < 0.01). Moreover, survival time of BDL-Lact rats was not significantly different from that of Sham rats, Survival time in rats with internal biliary drainage was increased significantly compared with BDL rats (BDL-Drain 156 2 35 h versus 54 ? 35 h; p < 0.001) and was not significantly different from that of BDL-Lact and Sham rats. The experimental model did allow daily monitoring of plasma endotoxins; therefore TNF plasma levels were measured. In jaundiced rats significant amounts of TNF were
found. These TNF levels did not correlate with mortality. An explanation could be the very short half-life of TNF and the intermittent appearance in plasma, making single daily samples unreliable. As was recently shown, TNF stimulates a de novo synthesis of triglycerides besides inhibiting lipoprotein lipase, both resulting in increased plasma triglyceride levels (19,20). We therefore also measured plasma triglyceride levels as indicator of TNF release and indirectly of endotoxins. In jaundiced rats a significant relation was found between serum triglycerides and survival (triglycerides, 2.8 2 0.9 mmol/l versus 5.9 2 1.3 mmol/l for surviving versus decreased jaundiced rats, respectively; p < 0.001). DISCUSSION It was known from previous experiments that biliary obstruc-
OBSTRUCTIVE JAUNDICE
I
HISTORY LABORATORY ULTRA SOUND ' CT-SCAN
*
DIAGNOSIS
* *
1
1
ERCP
I + NO CANULATION or CANULATION NO @RAIN
I I
1
INTERNAL DRAIN
EXTERNAL DRAIN or NO DRAIN INOPERABLE
-4
I
ANTIBIOTICS
1
I I
I
I
c 3d
I
(RADICAL) RESECTION
I
*3w
I
PALLIATION (DRAIN)
1
Fig. 1. Proposed decision tree for supportive preoperative therapy in jaundiced patients. The preventive treatments are shown in the relation to the means to obtain them. Shadow boxes show the desired supportive therapy. CT = computed tomography; ERCP = endoscopic retrograde cholangiopancreaticography; PTC-D = percutaneous transhepatic cholangiography (drainage).
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
Complications in Obstructive Jaundice
tion affects the immune system. In rats and mice with experimental biliary obstruction suppression of the specific cellular immunity was found (6,7). In these studies the hyperbilirubinemia o r the high bile acid levels were suggested to be responsible . In the present experiments cellular immunity was significantly suppressed in conventional rats with obstructive jaundice, supporting the mentioned experiments. In germfree (endotoxin-'free') jaundiced rats, however, cellular immunity was not impaired. Despite the absence of endotoxemia in conventional BDL rats, strong evidence was found for a crucial role of endotoxins with regard to the impaired immune system. In germfree rats endotoxin levels in feces were at least a thousand times lower than in conventional rats, being one of the major differences between the two rat strains. Moreover, chronic low-dose endotoxin administration induced similar defects of the cellular immunity as bile duct ligation in conventional rats. Finally, a correlation was shown between spontaneous TNF production by peritoneal macrophages, as indicator of endotoxin toxicity, and suppression of cellular immunity in jaundiced rats. A possible explanation of the unexpected absence of endotoxins is the difficult endotoxin assay. However, a recent study, in which a similar sensitive endotoxin assay was used, confirmed our results (21). In summary, evidence was given by the first experiment that suppression of cellular immunity is caused by endotoxins from the gastrointestinal tract. The results of this study suggested that one of the aims of preoperative supportive therapy in jaundiced patients should be prevention of the toxic effect of endotoxins from the bowel lumen. The clinical and experimental observation that endotoxins are related to impairment of kidney function and death could not be directly confirmed in our experiments (2, 4, 10-12). As mentioned above, no substantial endotoxemia was found in our jaundiced rats. However, the second experiment provided indirect evidence that endotoxins d o play an important role. In two separate studies lactulose was shown to inhibit the toxic effects of endotoxin (17,22). In the present study preventive treatment with oral lactulose significantly reduced mortality after renal ischemia in jaundiced rats, and there was a clear but not statistically significant tendency for it to reduce the impairment of kidney function. In the rats treated with lactulose no other variables improved, which suggests that the observed effect must be due to inhibition of the toxic effect of endotoxin. The superior effect of internal biliary drainage on kidney function and mortality after renal ischemia. compared with lactulose, was to be expected. Internal biliary drainage not only restores the bile flow to the gut, thus inactivating endotoxins in the bowel lumen with bile acids, but also restores normal liver function, reestablishing the effective clearance of endotoxins present in portal blood, providing an optimal protection against the toxic effect of endotoxins. In conclusion, the two presented experiments provide
11
evidence that endotoxins play an important role in the etiology of infections and in renal complications that occur after surgery in patients with obstructive jaundice. The experimental data suggest that these complications can be prevented by a preoperative treatment, such as oral lactulose or internal biliary drainage, which reduces endotoxin toxicity. Extrapolating these experimental data to the clinical situation could result in a flow diagram as presented in Fig. 1. After the diagnosis of obstructive jaundice, which in most cases will be due to a malignancy, a sequence of diagnostic and therapeutic interventions will occur. Depending on the degree of jaundice, the cause, and the operability of the patient, an attempt is made to cannulate the obstructed biliary tract by endoscopic retrograde cholangiopancreatography , internal biliary drainage being the preoperative treatment of choice. After successful internal biliary drainage at least 10-14 days' recovery of liver functions and immune system is allowed, followed by a planned surgical procedure. In the event of an unsuccessful or partial biliary drainage, infectious complications often occur. In these patients antibiotic treatment (given prophylactically) is continued with the addition of an anti-endotoxin treatment such as oral lactulose. The patient should then be operated on within several days. It is evident that prospective randomized trials are necessary to validate such a treatment protocol. REFERENCES 1. Pain JA, Cahill CJ, Bailey ME. Perioperative complications in obstructive jaundice: therapeutic considerations. Br J Surg
1985;72:942-5. 2. Wait RB, Kahng KU. Renal failure complicating obstructive jaundice. Am J Surg 1989;157:256-63. 3. Armstrong CP, Dixon JM, Taylor TV, Davies GC. Surgical experience of deeply jaundiced patients with bile duct obstruction. Br J Surg 1984;71:234-8. 4. Wardle EN, Wright NA. Endotoxin and acute renal failure associated with obstructive jaundice. Br Med J 1970;4:472-4. 5. Pain JA, Bailey ME. Measurement of operative plasma endotoxin levels in jaundiced and non-jaundiced patients. Eur Surg Res 1987;19:207-16. 6. Pinto M, Kaplun A. Immune status in mice with experimental biliary obstruction. Clin Immunol lmmunopathol 1980;16:396405. 7. Roughneen PT, Gouma DJ, Kulkarni A D , Fanslow WF, Rowlands BJ. Impaired specific cell-mediated immunity in experimental biliary obstruction and its reversibility by internal biliary drainage. J Surg Res 1986;41:113-25. 8. Fargion SR, Cappellini MD, Podda M, Fiorelli G . Immunita cellulare nelle colestasi intra ed extra-epatiche. Min Gastr 1976;22:261-5. 9. Gouma DJ, Coelho JCU, Fisher JD, Schlegel JF, Li YF, Moody FG. Endotoxemia after relief of biliary obstruction by internal and external drainage in rats. Am J Surg 1986;151:476-9. 10. Wilkinson SP, Moody H, Stamatakis JD, Kakkar VV, Williams R. Endotoxaemia and renal failure in cirrhosis and obstructive jaundice. Br Med J 1976;2:1415-8. 1 1 . Bailey ME. Endotoxin, bile salts and renal function in obstructive jaundice. Br J Surg 1976;63:774-8. 12. Cahill CJ, Pain JA, Bailey ME. Bile salts, endotoxin and renal function in obstructive jaundice. Surg Gynecol Obstet 1987;165: 519-22. 13. Lee E. The effect of obstructive jaundice on the migration of
12
J . W.M. Greve et al.
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/12/15 For personal use only.
reticulo-endo-thelial cells and fibroblasts into early experimental granulomata. Br J Surg 1972;59:875-7. 14. Olofsson P. Application of a quantitative spectrophotometric endotoxin assay on lymph and plasma from the rat. Eur Surg Res 1986;18:112-21. 15. Greve JW, Gouma DJ, Soeters PB, Buurman WA. Suppression of cellular immunity in obstructive jaundice is caused by endotoxins. A study with germfree rats. Gastroenterology 1990; 98~478-85. 16. Spofford BT, Daynes RA, Granger GA. Cell-mediated immunity in vitro: a highly sensitive assay for human lymphotoxin. J Immunol 1974;112:2111-6. 17. Greve JW, Gouma DJ, v Leeuwen PAM, Buurman WA. Lactulose inhibits endotoxin-induced tumour necrosis factor production by monocytes. An in vitro study. Gut 1990;31:198-203.
18. Greve JW, Maessen JG, Tiebasch T, Buurman W A , Gouma DJ. Prevention of postoperative complications in jaundiced rats. Internal biliary drainage versus oral lactulose. Ann Surg 1990; 212~221-7. 19. Beutler B, Mahony J , Le Trang N, Pekala P, Cerami A. Purification of cachectin, a lipoprotein lipase-suppressing hormone secreted by endotoxin induced rat 264.7 cells. J Exp Med 1985; 161~984-95. 20. Feingold KR, Serio MK, Adi S, Moser AH, Grunfeld. Tumor necrosis factor stimulates hepatic lipid synthesis and secretion. Endocrinology 1989;124:2336-42. 21. Roughneen PT, Kumar SC, Pellis NR, Rowlands BJ. Endotoxemia and cholestasis. Surg Gynecol Obstet 1988;167:205-10. 22. Pain JA, Bailey ME. Experimental andclinical study of lactulose in obstructive jaundice. Br J Surg 1986;73:775-8.
