Abstracts
1997 Digestive Disease Week

Endotoxin and bacterial translocation during elective surgery: studies on mechanisms and prevention of gut barrier failure.

D Berger, E Bolke, M Seidelmann, HG Beger. Department of General Surgery, University of Ulm, Germany.

Objective: In a prospective, observational, clinical study the time sequence of endotoxin release from the GI-tract and bacterial translocation to mesenteric lymph nodes during and after gastric, pancreatic or colonic resection were correlated with the systemic release of hemodynamically active mediators (thromboxane B_{2}, 6-keto-PGF1lpha, leucotriene B_{4}), interleukin-6 and with the GI microcirculation as determined by mucosal tonometry. 56 pts. aging from 33 to 77 years were followed up perioperatively until the 5th po. day. In a further controlled randomized study 40 patients before gastric or pancreatic resection were enrolled and pretreated with an immunogloblin-enriched bovine colostral milk preparation for 3 days after randomization to the verum group. The time course of endotoxin and Il-6 plasma levels was followed up until the 5th po. day. Statistical analysis was performed by using the Kruskal-Wallis test, chi2-test and Mann-Whitney-U test. Median values and quartiles are given for all measurements except the pHi values which were demonstrated as mean±S.E.M. due to gaussian distribution. Measurements and main results: Endotoxin plasma levels started to increase significantly after induction of anesthesia from 0.04 to 0.13 EU/ml. Peak values of 0.24 EU/ml were observed before the closure of the abdominal wall. Mesenteric lymph nodes obtained from 34 pts. after mobilization of the organ which should be resected were found positive for bacterial growth in 12 cases. At the same time point 50 out of 56 pts. had increased endotoxin plasma levels (0.19 EU/ml). Intramucosal pH started to decrease after skin incision from 7.34±0.11 to 7.29±0.09 and reached a minimum 6h po. of 7.19±0.07. The prostanoid release was only very slight reaching a significant maximum after operation (299 pg/ml). Enteral pretreatment with the colostral preparation significantly reduced perioperative endotoxemia as checked by comparing the "area under curve" (p<0.05) in the randomized study. Endotoxin plasma levels increased at later time points and decreased earlier in the verum group whereas the control group revealed a comparable course as found in the previous observational study. Maximal Il-6 levels were reduced from 384 to 288 pg/ml by the colostral preparation. Due to the wide variance that effect was nonsignificant. Conclusion: Endotoxin translocation is much more frequent than bacterial translocation (p<0.001) and represents a very sensitive marker of gut barrier failure. The described arachidonic acid derived mediators may not be the cause of gut barrier damage. Microcirculatory disturbances as observed by a decrease of pHi was found after the onset of endotoxin leakage. The hypothesis of gut derived endotoxemia during abdominal surgery is strongly supported by the observation of reduced perioperative endotoxemia after enteral pretreatment with a bovine colostral milk preparation.