A wide range of microorganisms can produce septicemia, especially in an immunocompromised host. Particularly in gastroenterological surgery, the major culprits are gram-negative bacteria (1), and the most common early manifestation is undoubtedly peritonitis. The traditional therapeutic approach has been adequate surgical intervention, application of antibiotics and parenteral fluid, and nutritional support. However, failures continue to be high, which has stimulated the search for new treatment modalities. One such principle is immunotherapy. This therapeutic approach is based on the notion that the host s own protective mechanisms might be mobilized in the defense against invasive organisms.
Immune functions are stimulated by polysaccharides from a variety of edible mushrooms (2, 3) and plants (4). In particular, polysaccharides, especially β-glucans, are now widely recognized as biological response modifiers or modulators that act through their ability to activate macrophages (5). Moreover, polysaccharides like β-glucans protect against various bacterial infections (6). We have recently shown that β-glucans protect against Mycobacterium bovis (7) and Streptococcus pneumoniae 6B infections in mice (8).
Agaricus blazei Murill (AbM) is an edible Basidiomycetes species primarily found in a damp mountainous coastal region (Piedade) outside of São Paulo, Brazil. It has been extensively used in folk medicine for life-threatening diseases throughout the world (2). The protein-bound polysaccharides, "proteoglucan," isolated from AbM, have marked tumoricidal activity in mouse tumor models (9, 10). In fact, Ohno et al. (11) proposed that the 1,3-β-glucan in AbM was responsible for the antitumor activity. Biologic activities associated with in vivo use of these biopolymers include anticancer agents (12), enhanced defense against bacterial challenges (13), increased hemopoietic activity and radioprotective effects (14), and improved healing of bowel anastamoses (15). In vitro, these molecules have been reported to influence the morphology of macrophages (16), and in vivo, in the activation of the natural killer cells (17), release of cytokines (e.g., TNF-á, IL-6, and IL-1) (18), release of nitric oxide (19), lysosomal enzyme secretion (20), release of hydrogen peroxide (21), activation of arachidonic acid metabolism (22), and activation of the alternate complement pathway (23).
Beyond our initial study on pneumococcal infection and AbM (24), the effect of AbM against other bacterial infections has, to our knowledge, not been documented. Patients undergoing bowel operations are prone to development of peritonitis. The use of antibiotics against such infections may cause bacterial resistance, especially at hospitals with major turnover operations.
The aim of this study was firstly to establish a safe and secure animal experimental model resembling iatrogenic peritonitis occurring during bowel operations. Secondly, to examine the usefulness of AbM given prophylactically against such infections.
MATERIALS AND METHODS
Four- to 5-week-old female inbred BALB/cHsdOla strain mice with a body weight of 18 to 20 g were obtained from Harlan Scandinavia, Allerod, Denmark. Upon arrival, 8 mice were placed in each cage and housed for 1 week in air with 60% humidity at 25°C. All animals were fed ad libitum with rat and mouse maintenance diet (RM1[E], Special Diets Services Limited, Witham, Essex, England) and they had free access to tap water. The feed contained no antibiotic agent. The room was illuminated using a 12-h light and darkness cycle (light on at 7:00 am). The animals were not fasted before the start of the experiment. The mice were treated according to the ethical guidelines of the Norwegian Institute of Public Health. The experimental protocol was approved by the local ethical committee and performed in compliance with laws and rules regulating animal experiments in Norway.
An aqueous extract ("gold label") from AbM of the family basiodiomycetes was obtained from ACE Co., Ltd, Gifu-ken, Japan, and stored at 4°C in a dark glass bottle. The extract is sold as health food in Japan. The lipopolysaccharide (LPS) content of the extract was analyzed using the Pyrogene Recombinant Factor C Endotoxin Detection System (Bio Whittaker, Rockland, Me) and was found to be miniscule, i.e., 1.5 EU of LPS equivalents per mL of AbM (1 EU = 0.1 ng LPS). Control mice were treated with phosphate-buffered saline (PBS) of pH = 7.3.
Preparation of Standard Fecal Solution
The inoculum was made by pooled dry fecal pellets that were randomly collected from all the mice-containing cages 1 day before the challenge. Within 30 min of challenge, 10 mL of pooled pellets, weighing 4.0 ± 0.5 g, were mixed with 30 mL of PBS at room temperature. The mixture was thoroughly shaken before filtering through a double layer of surgical gauze with a mesh diameter of 1 to 2 mm. This stem solution was defined as the undiluted fecal inoculum (1/1). Further dilutions were made by adding PBS. As Tannock (25) pointed out, the numerically predominant microbes (>99%) in mice were anaerobic bacteria, of which lactobacilli, bacteroides, fusiforms, spiral-shaped bacteria, enterobacteria, and enterococci were represented. Microaerophilic and facultatively anaerobic bacteria were also presented in appreciable numbers. Although our initial attempts to cultivate anaerobic bacteria did not provide sufficient evidence under the laboratory conditions, their presence cannot be ruled out. Quantitative bacteriologic studies performed on fresh inoculum showed a major (>90%) aerobic bacterial population that was evenly distributed between coliforms, enterococci, á-hemolytic streptococci, and nonhemolytic streptococci.
Induction of Peritonitis and Treatment
The following dilutions of fecal stem solution were used: 1/8, 1/10, 1/12, 1/16, 1/32, and 1/64. The i.p. injection of 1/4 solution resulted in 100% mortality within 1 d, whereas mice receiving i.p. injections of fecal solutions 1/16, 1/32, and 1/64 considerably survived (50%, 75%, and 100%, respectively). Accordingly, we ended up using 3 degrees of peritonitis defined as severe (1/8), moderate (1/10), and mild (1/12) peritonitis. Experiments were conducted in groups of 8 mice placed in 1 cage. Two hundred microliters of fecal solution, in various dilutions, was injected i.p. via the lower half of the abdomen using a 1-mL syringe, using a cannula with a diameter of 0.05 mm. Control mice were administered i.p. with 200 μL of PBS instead of the fecal solution. The mice were first treated by orogastric intubation, given 200 μL of AbM or 200 μL of PBS as control, 1 day (24 h) before induction of fecal peritonitis. Eight mice were used in each group with AbM or PBS per experiment, which was done twice.
Detection of Microorganisms in Blood
The blood samples were drawn from the saphenous vein at various time intervals after the induction of peritonitis and diluted to 1/10, 1/100, 1/103, and 1/104 in Todd-Hewitt medium. The blood samples from moribund mice were obtained by transthoracic cardiac puncture. The diluted blood samples were cultivated on blood agar plates in atmospheric air and humidity at 37°C for 24 h. Blood agar was used both as an enriched medium for fastidious bacteria and as a differential medium. Lactose agar plates were used as a selective medium for the detection and enumeration of coliform organisms. Bacterial density was assessed by counting colony-forming units (CFU) and expressed as CFU/mL after incubation. Cultivation for growth of anaerobic bacteria was not performed.
The diluted blood samples were cultivated on blood agar plates. Bacterial density (CFU) was assessed during incubation for 24 h. Twenty-five microliters of blood was harvested from each mouse at 3 and 12 h, and 1, 2, 3, 6, and 7 days after the challenge, and the number of CFU/mL was calculated. The moribund mouse had a level of 24 × 105 CFU/mL.
Three days before the experiments, some of the mice (for details see the results) were harmlessly implanted with miniature glass-encapsulated, implantable, programmable, temperature transponders (IPTT-300, Bio Medic Data System, Inc., supplied by Plexx, Ab Elst, The Netherlands) beneath the hairy skin in the dorsal part of the upper trunk. The transponders were used to measure body temperature before and during the experiments.
After the mice were induced with peritonitis by i.p. injection with different fecal solutions, they were inspected every 2 to 3 h for the first 24 h, followed by every 8 h during day 2, and every 12 h during day 3 to evaluate serious morbidity and mortality. When the mice were considered moribund, as demonstrated by curved spine, shivering, and/or immobility, they were killed by cervical dislocation. Time of death was registered, autopsy was immediately performed, and gross pathological changes were noted.
Statistical calculations were conducted using the Statistical Package for Social Sciences (version 10.0; Chicago, Ill.). Descriptive statistics are expressed as means and SEM. The statistical significance of differences within and between groups was determined by analysis of variance (ANOVA), using the General Linear Model with repeated measurements. The α level was set at 0.01. P < 0.05 was considered statistically significant. The overall survival was analyzed using the Kaplan-Meier survival analysis with the log-rank test, which compares survival distributions between groups. The post hoc tests for more than 2 dependent categorical variables were analyzed by the Tukey-HSD test.
Characterization of the Model Used for Peritonitis
Types of aerobic bacteria
Bacteriological studies based on blood samples from mice with induced peritonitis from feces diluted 1/16 showed the growth of a facultative flora of aerobic gram-positive and gram-negative species. When we performed quantitative and qualitative characterization of the bacteremia, we found that both gram-positive streptococci and gram-negative coliform bacteria dominated.
The mice were induced with fecal peritonitis by i.p. injection of 200 μL of a diluted fecal solution. Secondary systemic infection was determined by measuring CFU in repeated blood cultures. When 1/10 or 1/12 of the fecal solution was injected, there was a sharp increase in CFU levels between 3 h and 12 h, after which a plateau phase was reached (Fig. 1A). In contrast, 1/8 of fecal solution gave a continued sharp rise in bacteremia, killing nearly all animals within 24 h (Fig. 1A). There was a highly significant (P < 0.005, 2-way between-groups ANOVA) difference in the CFU levels between groups of mice challenged with 1/8, 1/10, and 1/12 of fecal solution, resulting in severe, moderate, and mild peritonitis, respectively.
Temperature and CFU
There was a significant difference in body temperature between the 3 peritonitis groups (P = 0.002, 2-way between-groups ANOVA) (Fig. 1B). The relationship between temperature and CFU levels was investigated using Pearson product-moment coefficient (r), and a strong negative correlation was found between the 2 variables, i.e., the more the bacteria, the lower the temperature (Table 1).
As shown in Fig. 1C, 88% of the animals with severe peritonitis died within 1 day, and all were dead by the third day. In contrast, 25% of those with moderate and 60% of those with mild peritonitis survived until after 7 days. The differences were highly significant (P < 0.005, log-rank test).
Effect of AbM on Experimental Peritonitis in Mice
In the next series of experiments, either AbM or PBS was orogastrically instilled, 1 day before the fecal challenge and induction of peritonitis. The animals that had experienced severe or moderate peritonitis showed significantly lower bacteremia levels (P ≤ 0.04, 2-way between-groups ANOVA) when they had been pretreated with AbM compared with PBS (Fig. 2A). The mild peritonitis group showed a similar trend in favor of AbM, but the difference was not statistically significant.
Temperature and disease symptoms
Figure 2B shows the overall temperature for the 3 severities of peritonitis in groups pretreated with AbM versus PBS. For both groups, the temperature declined with time and increasing illness. Mice that were given AbM tended to have higher temperatures than the controls, but the difference was not statistically significant.
The mice in the severe peritonitis group which were pretreated with AbM showed a higher survival rate than the mice pretreated with PBS (Fig. 3A). The survival rate decreased from 100% to 33% after 2 days and remained at 25% after 4 days (P = 0.005, log-rank test). For the less severe peritonitis groups, a similar trend was observed, with nonsignificant differences (Fig. 3B and C). However, the overall survival rate for all peritonitis groups reached a statistically significant level in favor of AbM (P = 0.048, log-rank test) (Fig. 4). To evaluate the antibiotic effect, 8 mice were induced with severe peritonitis and simultaneous orogastric instillation of antibiotics (metronidazole [5 mg/mL], 30 μg/g and doxycyline [20 mg/mL], 10 μg/g). The survival persisted at 25% from day 3 which was comparable to the AbM treatment (Fig. 3A).
Autopsy of mice that died of fecal peritonitis revealed no sign of intra-abdominal abscess. Accordingly, the cause of death was probably caused by the bacterial septicemia.
In this study, we show that prechallenge oral administration of a mushroom extract (AbM) protects mice against lethal septicemia after fecal peritonitis, as demonstrated by a reduction in bacteremia (Fig. 2) and an increase in survival rate (Figs. 3 and 4). Recently, we reported a similar effect by oral administration of AbM prechallenge in mice subsequently given i.p. infection with the virulent S. pneumoniae serotype 6B (24). Accordingly, the immunostimulatory principle of AbM was absorbed from the gastrointestinal tract to the blood, thereby stimulating the leukocytes to combat bacteremia, which was the cause of death in these mice. In as much as AbM is particularly rich in β-glucans and practically devoid of LPS, the former substance is likely to be involved in creating the effect. β-glucans exert their action by stimulation of natural killer cells (26, 27) and monocyte-macrophage release of proinflammatory cytokines (28), which has also been demonstrated using different types of β-glucans (29, 30).
Unique to AbM in this model was the protective effect against infection after the oral administration, whereas a similar effect of other β-glucans was exclusively observed after i.p. administration in mice (3,8,30) or i.v. administration in high-risk surgical patients (15). One may speculate that AbM contains small or easily resorbable glucan molecules, which stimulate the mononuclear phagocytes by interaction with the β-glucan binding site of CD11b/18 and possibly the dectin-1 receptor (31) and toll-like receptor 4 (32), resulting in increased cytokine (33) and phagocytic response.
The protection by AbM seemed to be stronger in mice with severe peritonitis (Fig. 3A). One explanation could be that the stimulatory effect of AbM was declining after 1 to 2 days (Fig. 1C). In this context, repetitive stimulation of the mice with AbM throughout the experiment seemed pertinent to perform. However, this was not carried out because of increased risk for complications and adverse effects from orogastric intubation of the septic mice. Inasmuch as most of the anaerobic bacteria were eliminated during the preparation of the fecal solution because of exposure to atmospheric oxygen, we assume that the lethal septicemia was mainly caused by aerobic gram-positive and gram-negative (coliform) bacteria. However, a certain contribution of the anaerobic bacteria to the septic state of these mice is conceivable.
The model of induced peritonitis by i.p. instillation of the fecal solution by a thin syringe has few confounding variables. It is accurate in the amount of bacteria given and independent of both anesthesia and surgical procedure, with risk for complications. Such parameters deteriorate interpretation of the results in models based on bowel exclusion or cecal ligation and puncture (34).
The AbM extract was orally given before the bacterial challenge to prime the native immune system and elicit ananti-infection reaction, which subsequently will involve acquired immunity. The acquired immunity response isclassified into T helper cell type 1 (Th1) (antitumor/anti-infection), Th2 (antihelminth/anti-rejection/proallergy), and Th3/T regulatory cells (suppression/anti-inflammatory) (35) based on the distinct cytokine secretion patterns. From our own and others experiments (27), results have shown that macrophages stimulated by AbM produce proinflammatory and anti-Th1 cytokines. Ono et al. (36) reported that the change in the Th1/Th2 cytokine balance in peritonitis might induce a shift toward a Th2-dominant phenotype in severe peritonitis, and the capacity to produce IFN-γ and IL-12 by the liver Kuppfer cells is reduced. Despite the fact that the BALB/c mice are more prone to a Th2-deviated response, relative to the anti-infective Th1 response (37), we demonstrated a significant protective effect of AbM against lethal peritonitis. Accordingly, in mouse strains that are less susceptible to infection and with a stronger Th1 response, the results may have been even more favorable.
Inasmuch as β-glucans both protect against infections and have adjuvant effects on allergy development (38, 39), which is Th2-driven; it is difficult to predict the outcome of β-glucan stimulation per se on the Th1/Th2 balance. On the other hand, AbM contains a wealth of other substances, like ergosterol (3), fractionized polysaccharides (10), and alkaline extracts (11), which may well also play a major role. The overall outcome of AbM stimulation, at least on monocytes/macrophages, was reported to be a proinflammatory and Th1 response (40). Modulation of host susceptibility by macrophage stimulation could be an important adjuvant therapy in cases of severe infections. Different water-insoluble and -soluble 1,3-β-glucans are known to activate the macrophage system (41). Abdominal sepsis caused by secondary fecal peritonitis after anastomotic dehiscence is a rare but life-threatening complication of bowel surgery. Our model is comparable with this clinical condition and further demonstrates the protective effect induced by AbM.
In conclusion, we have demonstrated that AbM, given prechallenge, exerts a protective effect against fecal peritonitis induced in mice. Moreover, the model for experimental and fecal peritonitis in mice is robust, accurate, and reproducible. Further studies are needed whether the AbM may in the future be considered for use as a prophylactic anti-infective principle in patients undergoing gastrointestinal resections.
The authors thank Dr. Nils Olav Hermansen from the Department of Microbiology at Ulleval University Hospital, Oslo, for the quantitative analysis of bacterial subcultures and the staff of the animal facility at The Norwegian Institute of Public Health, Oslo, for excellent technical help.
1. Litton A: Gram-negative septicaemia in surgical practice. Br J Surg
2. Wasser SP, Weis AL: Medicinal properties of substances occuring in higher basiodiomycetes mushrooms. Current perspectives. Int J Med Mush
3. Takaku T, Kimura Y, Okuda H: Isolation of an antitumor compound from Agaricus blazei
Murill and its mechanism of action. J Nutr
4. Hetland G, Samuelsen AB, Løvik M, Paulsen BS, Aaberge IS, Groeng EC, Michaelsen TE: Protective effect of Plantago major
L. pectin polysaccharide against systemic Streptococcus pneumoniae
infection in mice. Scand J Immunol
5. Cleary JA, Kelly GE, Husband AJ: The effect of molecular weight and β-1,6-linkages on priming of macrophage function in mice by (1,3)-β-D-glucan. Immunol Cell Biol
6. Reynolds JA, Kastello MD, Harrington DJ, Crabbs CL, Peters CJ, Jemski JV, Scott GH, Di Luzio NR: Glucan-induced enhancement of host resistance to selected infectious diseases. Infect Immun
7. Heland G, Løvik M, Wiker HG: Protective effect of β-glucan against Mycobacterium bovis
, BCG infection in BALB/c mice. Scand J Immunol
8. Hetland G, Ohno N, Aaberge IS, Løvik M: Protective effect of β-glucan against systemic Streptococcus pneumoniae
infection in mice. FEMS Immunol Med Microbiol
9. Itoh H, Ito H, Amano H, Noda H: Inhibitory action of a (1→6)-β-D-glucan-protein complex (F III-2-b) isolated from Agaricus blazei
Murill ("himematsutake") on Meth A fibrosarcoma-bearing mice and its antitumor mechanism. Jpn J Pharmacol
10. Kawagishi H, Inagaki R, Kanao T, Mizuno T, Shimura K, Ito H, Hagiwara T, Nakamura T: Fractionation and antitumor activity of the water-insoluble residue of Agaricus blazei
fruiting bodies. Carbohydr Res
11. Ohno N, Furukawa M, Miura NN, Adachi Y, Motoi M, Yadomae T: Antitumor β-glucan from the cultured fruit body of Agaricus blazei
. Biol Pharm Bull
12. Ooi VE, Liu F: Immunomodulation and anti-cancer activity of polysaccharide-protein complexes. Curr Med Chem
13. Babineau TJ, Marcello P, Swalis W, Kenler A, Bistrian B, Forse RA: Randomised phase I/II trial of a macrophage-specific immunomodulator (PCG-glucan) in high-risk surgical patients. Ann Surg
14. Patchen ML, D Alesandro MM, Brook I, Blakely WF, MacVittie TJ: Glucan: mechanisms involved in its "radioprotective" effect. J Leukoc Biol
15. Compton R, Williams D, Browder W: The beneficial effect of enhanced macrophage function on the healing of bowel anastomoses. Am Surg
16. Burgaleta C, Territo MC, Quan SG, Golde DW: Glucan-activated macrophages: functional characteristics and surface morphology. J Reticuloendothel Soc
17. Fujimiya Y, Suzuki Y, Oshiman K, Kobori H, Moriguchi K, Nakashima H, Matumoto Y, Takahara S, Ebina T, Katakura R: Selective tumoricidal effect of soluble proteoglucan extracted from the basidiomycete, Agaricus blazei
Murill, mediated via natural killer cell activation and apoptosis. Cancer Immunol Immunother
18. Adachi Y, Okazaki M, Ohno N, Yadomae T: Enhancement of cytokine production by macrophages by stimulated with (1→3)-β-D-glucan, grifolan (GRN), isolated from Grifola frondosa
. Biol Pharm Bull
19. Ohno N, Egawa Y, Hashimoto T, Adachi Y, Yadomae T: Effect of beta-glucans on the nitric oxide synthesis by peritoneal macrophages in mice. Biol Pharm Bull
20. Tapper H, Sundler R: Glucan receptor and zymozan-induced lysosomal enzyme secretion in macrophages. Biochem J
21. Okazaki M, Chiba N, Adachi Y, Ohno N, Yadomae T: Signal transduction pathway of β-glucans-triggered hydrogen peroxide production by murine peritoneal macrophages in vitro
. Biol Pharm Bull
22. Castro M, Ralston NV, Morgenthaler TI, Rohrbach MS, Limper AH: Candida albicans
stimulates arachidonic acid liberation from alveolar macrophages through á-mannan and β-glucan cell wall components. Infect Immun
23. Suzuki T, Ohno N, Saito K, Yadomae T: Activation of the complement system by (1→3)-β-D-glucans having different degrees of branching and different ultra-structures. J Pharmacobiodyn
24. Bernardshaw S, Johnson E, Hetland G: An extract of the mushroom Agaricus blazei
Murill administered orally protects against systemic Streptococcus pneumoniae
infection in mice. Scand J Immunol
25. Tannock GW: Normal microbiota of the gastrointestinal tract of rodents. In Mackie RI, White BA, Isaacson RE (eds.): Gastrointestinal Microbiology
, Vol.2. Chapman & Hall Microbiology Series, pp 187-215, 1997.
26. Godshall CJ, Scott MJ, Burch PT, Peyton JC, Cheadle WG: Natural killer cells participate in bacterial clearance during septic peritonitis
through interactions with macrophages. Shock
27. Sorimachi K, Akimoto K, Ikehara Y, Inafuku K, Okubo A, Yamazaki S: Secretion of TNF-α, IL-8 and nitrite oxide by macrophages activated with Agaricus blazei
Murill fractions in vitro
. Cell Struct Funct
28. Chihara G: Recent progress in immunopharmacology and therapeutic effects of polysaccharides. Dev Biol Stand
29. Doita M, Rasmussen LT, Seljelid R, Lipsky PE: Effect of a soluble aminated β-1,3-D-polyglucose on human monocytes: stimulation of cytokine and prostaglandin E2
production but not antigen-presenting function. J Leukoc Biol
30. Hetland G, Sandven P: β-1,3 glucan reduces growth of Mycobacterium tuberculosis
in macrophage cultures. FEMS Immunol Med Microbiol
31. Brown GD, Taylor PR, Reid DM, Willment JA, Williams DL, Martinez-Pomares L, Gordon S, Wong SY: Dectin-1 is a major β-glucan receptor on macrophages. J Exp Med
32. Roeder A, Kirshning CJ, Rupec RA, Schaller M, Weindl G, Korting HC: Toll-like receptors as key mediators in innate fungal immunity. Med Mycol
33. Tsujimoto H, Ono S, Majima T, Kawarabayashi N, Takayama E, Kinoshita M, Seki S, Hiraide H, Moldawer LL, Mochizuki H: Neutrophil elastase, MIP-2, and TLR-4 expression during human and experimental sepsis. Shock
34. Almdahl SM, Nordstrand K, Melby K, Østerud B, Giercksky KE: Fecal peritonitis
in the rat-an experimental model for evaluation of surgical and adjuvant therapies. Acta Chir Scand
35. Infante-Duarte C, Kamradt T: Th1/Th2 balance in infection. Springer Semin Immunopathol
36. Ono S, Ueno C, Aosasa S, Tsujimoto H, Seki S, Mochizuki H: Severe sepsis induces deficient interferon-γ and interleukin-12 production, but interleukin-12 therapy improves survival
. Am J Surg
37. Watanabe H, Numata K, Ito T, Takagi K, Matsukawa A: Innate response in Th-1 and Th-2 dominant mouse strains. Shock
38. Ormstad H, Groeng EC, Lovik M, Hetland G: The fungal cell wall component β-1,3-glucan has an adjuvant effect on the allergic response to ovalbumin in mice. J Toxicol Environ Health A
39. Instanes C, Ormstad H, Rydjord B, Wiker HG, Hetland G: Mould extracts increase the allergic response to ovalbumin in mice. Clin Exp Allergy
40. Ellertsen LK, Hetland G, Johnson E, Grinde B: Effect of a medicinal extract from Agaricus blazei Murill on gene expression in a human monocyte cell line as examined by microarrays and immuno assays. Int Immunopharmacol
41. Suzuki Y, Adachi Y, Ohno N, Yadomae T: Th1/Th2-balancing immunomodulating activity of gel-forming (1→3) β glucans from fungi. Biol Pharm Bull
Agaricus blazei Murill; peritonitis; survival; septicemia; colony-forming units; gastroenterological surgery