- Email: [email protected]
Non-Protein Energy Overloading Induces Bacterial Translocation During Total Parenteral Nutrition in Newborn Rabbits
BASIC NUTRITIONAL INVESTIGATION
Nutrition Vol. 14, No. 5, 1998
Non-Protein Energy Overloading Induces Bacterial Translocation During Total Parenteral Nutrition in Newborn Rabbits TAKESHI YAMANOUCHI, MD, SACHIYO SUITA, MD, AND KOUJI MASUMOTO, MD From the Department of Pediatric Surgery, Faculty of Medicine, Kyushu University, Fukuoka, Japan Date accepted: 25 October 1997 ABSTRACT
The effect of energy intake provided by total parenteral nutrition (TPN) on the incidence of bacterial translocation and the relationship between TPN-induced cholestasis and bacterial translocation were investigated in newborn animals. Forty-six Japanese white newborn rabbits were divided into three groups: TPN-H group (high energy TPN; 280 kcal z kg21 z d21), TPN-L group (low energy TPN; 180 kcal z kg21 z d21), and a control group (breast fed). On day 8, they were all killed for investigation of the presence of bacterial translocation, for blood chemistry analysis, and for histological examination of the ileum. The incidence of translocation to any of mesenteric lymph nodes and liver and spleen was significantly higher in the TPN-H group (67%) than in both the TPN-L group (13%) and the control group (10%) (P , 0.01). No difference was seen in ileum morphology between the two TPN groups. Although the mean bilirubin level of the TPN-H group tended to be higher than the TPN-L group, whether or not bacterial translocation occurred was not found to be closely related to the degree of TPN-associated cholestasis. In conclusion, parenteral nonprotein energy overloading increased the incidence of bacterial translocation in the newborn rabbit. However, bacterial translocation did not appear to be associated with the development of TPN-associated cholestasis. Nutrition 1998;14:443– 447. ©Elsevier Science Inc. 1998 Key words: bacterial translocation, total parenteral nutrition, cholestasis, newborn, rabbit
INTRODUCTION
Bacterial translocation is a term introduced by Berg and Garlington1 to describe the passage of viable bacteria from the gastrointestinal tract through the epithelial mucosa into the lamina propria and then to the mesenteric lymph nodes and possibly other organs. Alexander et al.2 later expanded the definition to include the passage of viable and nonviable microbes and their byproducts (including endotoxins) across an intestinal barrier. Bacterial translocation is promoted by several factors including a disruption of the intestinal microflora leading to bacterial overgrowth,3 radiation,4 burn,5 intestinal ischemia,6 hemorrhagic shock,7 endotoxin,8 obstructive jaundice,9 small bowel transplantation,10 and decreased intestinal motility.11 The significance of bacterial translocation in humans remains controversial,12 but in critically ill patients bacterial translocation is suspected to cause sepsis and multiple organ failure.13 Total parenteral nutrition (TPN) is believed to promote
bacterial translocation,14 but the effect of the energy content of TPN on the incidence of bacterial translocation has yet to be investigated. Bacterial translocation has been postulated to be one of the factors inducing liver dysfunction during TPN.15 It is widely known that newborn infants, especially premature infants, tend to develop cholestasis during TPN. The gut barrier and systemic immune response in newborn infants is so immature16 that neonates have been postulated to be prone to bacterial translocation from their own intestines. This phenomenon suggests bacterial translocation to be the primary cause of TPN-associated cholestasis in neonates, although the cause of TPN-associated cholestasis is generally indicated to be multifactorial.17 The aim of this study was, therefore, to investigate the effect of high caloric TPN on bacterial translocation and to also clarify the relationship between TPN-associated cholestasis and bacterial translocation in neonates.
Correspondence to: Takeshi Yamanouchi, MD, Department of Pediatric Surgery, Faculty of Medicine, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.
Nutrition 14:443– 447, 1998 ©Elsevier Science Inc. 1998 Printed in the USA. All rights reserved.
0899-9007/98/$19.00 PII S0899-9007(98)00015-X
444
NEONATAL BACTERIAL TRANSLOCATION DURING TPN TABLE I. COMPOSITION OF SOLUTIONS
Volume (mL z kg21 z d21) Carbohydrate (g z kg21 z d21) Amino acid† (g z kg21 z d21) Fat‡ (g z kg21 z d21) Energy (kcal z kg21 z d21) NP cal/N (kcal/g)
TPN-H
TPN-L
Control*
230 28 12 13.0 280 119
230 16 12 7.5 180 68
200 4 14 16.0 200 70
* Breast fed; composition of rabbit breast milk is based on Hata et al.18 † Amizet 10 (Tanabe Ph Co., Osaka, Japan). ‡ 20% Intralipos (Midorijuji Ph Co., Osaka, Japan). TPN-H, high-energy total parenteral nutrition; TPN-L, low-energy total parenteral nutrition; NP, nonprotein; N, nitrogen.
MATERIALS AND METHODS
Animals Adult female Japanese white rabbits with timed pregnancies of 14 to 15 d of gestation, weighing 3.5 to 4.5 kg, were obtained (Kyudo Co. Ltd., Fukuoka, Japan). The rabbits were housed in individual rabbit cages until delivery to the animal facility of Kyushu University. At 30 d of gestation, they were anesthetized with thyamiral (25 mg/kg body weight [BWT]) and their newborn rabbits were rapidly delivered by a hysterotomy. The newborn rabbits were quickly wrapped with gauze and placed in an infant incubator under conditions of 34°C temperature and 100% humidity. In the infant incubator their body weights were measured after birth and thereafter every day of the study. Thirty-six newborn rabbits, weighing from 40 to 70 g, were then used for the parenteral nutrition study. Ten newborn rabbits were breast fed for 8 d as a control, but birth weight could not be measured in order to prevent the mothers from engaging in cannibalism. The protocol of this study was approved by the Kyushu University Institutional Animal Care and Use Committee. Insertion of Central Venous Catheters Within 24 h after birth, newborn rabbits were anesthetized with ketamine (20 mg/kg BWT) subcutaneously and locally anesthetized with 1.0% lidocaine. The skin over the neck and midscapular region was then prepared with povidone iodine. With sterile instruments and using aseptic techniques, the jugular vein was isolated and ligated and a silastic catheter (0.012 in ID, 0.025 in OD, Dow-Corning Corp., Tokyo, Japan) was inserted through the jugular vein into the superior vena cava. The free end of the catheter was tunneled subcutaneously and brought out at the midscapular region and connected to the harness and the swivel. The newborn rabbits were housed in individual cages at temperatures ranging from 33 to 35°C throughout the study period. Nutritional Regimen After catheterization, the newborn rabbits were randomized into one of two different parenteral nutrition groups. The composition of the hyperalimentation solutions are summarized in Table I. The expected composition of rabbit breast milk is also shown in Table I. The animals in the high-energy total parenteral nutrition (TPN-H) group (n 5 19) received a high energy solution (280 kcal z kg21 z d21), which was thought to include a 40% higher energy intake than the breast-fed rabbits, whose glucose intake was thought to be the largest when the glucose energy accounts for
one-half of total energy. A dose of 28 g of glucose z kg21 z d21 was considered to be the upper limit for parenterally fed newborn rabbits due both to the fact that newborn rabbits are known to not tolerate glucose overloading,18 and based on our preliminary experiments. The animals in the low-energy total parenteral nutrition (TPN-L) group (n 5 17) received a lower energy hyperalimentation solution (180 kcal z kg21 z d21), which was thought to be a 10% lower energy intake than the breast-fed newborn rabbits (Table I). The amount of amino acids was equal in these two group and fat energy accounted for one-half the total nonprotein energy. On days 1 and 2, both groups were fed parenterally with the half-energy solution for acclimation, and from day 3 they were fed with full strength TPN for 6 d. The nutritional composition of the administered solution is indicated in Table I. Each solution was supplemented with electrolytes, trace minerals, and vitamins. The actual administered volume was arranged according to body weight, which was measured daily. The control group (n 5 10) was nourished by their lactating mothers for 8 d. Testing for Bacterial Translocation and Specimens At the end of the study period, each animal was anesthetized with ether and weighed. Using aseptic techniques, a laparotomy incision was made and a blood sample was obtained by an aortic puncture for serum chemistry. Next, the exposed viscera was swabbed with a sterile cotton-tipped applicator stick, which was cultured in tryptic soy broth (TSB, Eiken Co., Ltd., Tokyo, Japan) to detect any accidental bacterial contamination. The mesenteric lymph nodes complex (MLN), liver and spleen were removed and weighed, and then each was placed in grinding tubes containing a nine-fold volume of TSB. The organs were homogenized with sterile glass grinders. Then the sternum was opened aseptically and the central venous catheter tip was removed for an aerobic culture in TSB. If the catheter culture was positive, the data of this animal was excluded. Portions (0.2 mL) of organ homogenates were cultured aerobically on MacConkey’s agar (Eiken Co., Ltd.) to detect gram-negative enteric bacilli, and blood agar (Eiken Co., Ltd.) to detect gram-positive cocci. The plates were examined after 24 and 48 h of incubation at 37°C. Gram-positive bacteria was identified using standard procedures. Facultative gram-negative bacteria was identified by the API 20E System (Analytab Products, Plainview, NY, USA). The terminal ileum and the liver were excised and fixed in 10% formalin, embedded in paraffin, and stained with hematoxylin and eosin stain for the histological analysis. Both the intestinal structure and the liver histology were evaluated by light microscopy. The villous height and crypt depth were measured by a micrometer. A minimum of 10 villi were counted in each sample. Blood samples were centrifuged to obtain plasma samples. These samples were kept frozen at 220°C for additional chemical analysis. The plasma total bilirubin (T.Bil), direct bilirubin (D.Bil), glutamic-oxaloacetic transaminase (GOT), blood urea nitrogen (BUN), triacylglycerol (TG), total cholesterol (T.Chol), and blood sugar (BS) were determined by an autoanalyzer (Hitachi 736, Hitachi Co., Ltd., Tokyo, Japan). Statistics The incidence of translocation (discontinuous data) was evaluated with the Fisher’s exact test. The continuous data were expressed as the mean 6 SD and were tested by an analysis of variance with the Dunnet test. P values less than 0.05 were considered to be significant. RESULTS
No animal had any positive swab cultures in the peritoneum. There were three catheter-related infections; one in the TPN-H group and two in the TPN-L group. Therefore data were evaluated
NEONATAL BACTERIAL TRANSLOCATION DURING TPN
445
TABLE II. BODY WEIGHT AND LIVER WEIGHT
Initial body weight (g) Final body weight (g) Percent increase in body weight Liver weight/body weight (%)
TPN-H
TPN-L
Control*
57.2 6 6.6 85.3 6 8.8 49.7 6 8.7 3.44 6 0.62‡
59.1 6 6.2 84.8 6 10.4 43.2 6 7.4 2.88 6 0.52
137.2 6 16.3† 3.02 6 0.16
* Breast-fed. † P , 0.01 compared with the TPN-H and TPN-L groups. ‡ P , 0.05 compared with the TPN-L group. TPN-H, high-energy total parenteral nutrition; TPN-L, low-energy total parenteral nutrition.
in 18 newborn rabbits of the TPN-H group, 15 newborn rabbits in the TPN-L group, and 10 newborn rabbits in the control group. Body Weight Gain and Liver Weight The initial and final body weights are shown in Table II. The final body weight in the control group was higher than in the other groups. No significant difference was observed between the TPN-H group and the TPN-L group. The liver weight/body weight ratio was higher in the TPN-H group than in the TPN-L group, which, therefore, suggested hepatomegaly in the TPN-H group. The characteristic appearance of the animals in the TPN-H group was depilation, which was not recognized in the animals in the TPN-L or the control groups. Incidence of Bacterial Translocation and Species of Translocated Bacteria The results of translocation are shown in Figure 1. Translocation to the MLN occurred in 39% (7/18) of the animals in the TPN-H group, 7% (1/15) of the animals in the TPN-L group, and 10% (1/10) of the animals in the control group. This represents a significant increase in the MLN translocation in the TPN-H group compared to the TPN-L group (P , 0.05). Translocation to the liver occurred in 28% (5/18) of the animals in the TPN-H group, 7% (1/15) of the animals in the TPN-L group, and 0% (0/10) of the animals in the control group, while translocation to the spleen occurred in 33% (4/12) of the animals in the TPN-H group, 0%
(1/10) of the animals in the TPN-L group, and 0% (1/10) of the animals in the control group (not significant). Translocation to either MLN, the liver or spleen (total incidence of translocation) was 67% (12/18) of the animals in the TPN-H group, 13% (2/15) of the animals in the TPN-L group, and 10% (1/10) of the animals in the control group. This represents a significant increase in the TPN-H group compared with both the TPN-L group and the control group (P , 0.01). The translocated bacteria mainly included Staphylococcus aureus (47%), Corinebacterium (20%), and Enterococcus fecalis (13%). There was no difference in the species of translocated bacteria among the groups. Escherichia coli was not identified in any of the samples. Blood Chemistry The blood chemistry data are summarized in Table III. The plasma level of bilirubin in the TPN-H group was significantly higher than in the control group. The plasma level of GOT was not different among the three groups. The plasma level of BUN in the TPN-L group was significantly higher than in the other groups. The plasma level of triacylglycerol and blood sugar in the TPN-L group was significantly lower than in the other groups. To assess the relationship between TPN-associated cholestasis and bacterial translocation, all the TPN animals (TPN-H and TPN-L) were divided into two groups according to presence or absence of bacterial translocation. Figure 2 shows a comparison of the plasma level of direct bilirubin between the bacterial translocation-positive group and the bacterial translocation-negative group. The level of direct bilirubin was not related to the presence of bacterial translocation. Histology The histological findings of the liver samples showed cholestasis, which demonstrated bile plugs in the interlobular bile ducts and various bile pigments in the Kupffer cells, which were associated with increased levels of serum bilirubin. The villous height and crypt depth in both parenterally fed groups were lower than in the control group (P , 0.05). No difference was noted between the TPN-H and TPN-L groups (Table IV). No other pathological findings, including subepithelial edema, were observed in any of the groups. DISCUSSION 14
FIG. 1. Incidence of bacterial translocation in the mesenteric lymph nodes (MLN), liver, and spleen. The total incidence of translocation means any positive culture.
Alverdy et al. first showed that 2 wk of TPN increased the incidence of bacterial translocation, from 0% to 66%, in rat models. They, therefore, considered bacterial translocation to have contributed to bacterial overgrowth, while also impairing the intestinal defense and including lower levels of secretary IgA in
446
NEONATAL BACTERIAL TRANSLOCATION DURING TPN TABLE III. BLOOD CHEMISTRY
TPN-H TPN-L Control
Total bilirubin (mg/dL)
Direct bilirubin (mg/dL)
GOT (IU/L)
BUN (mg/dL)
Triacylglycerol (mg/dL)
Blood glucose (mg/dL)
1.34 6 0.92* 1.08 6 0.94 0.57 6 0.52*
1.11 6 0.90* 0.84 6 0.78 0.40 6 0.46*
36.6 6 21.2 35.2 6 31.5 43.3 6 43.3
23.5 6 7.1* 39.9 6 13.6*,† 19.9 6 10.8†
189.5 6 180.6* 55.0 6 44.0*,† 122.7 6 78.1†
123.0 6 41.5* 58.1 6 36.3*,† 208.8 6 24.6†
*,† Group means with the same superscripts are significantly different. GOT, glutamic-oxaloacetic transaminase; BUN, blood urea nitrogen; TPN-H, high-energy total parenteral nutrition, TPN-L, low-energy total parenteral nutrition.
bile. Thereafter, additional studies showed that enteral nutrition19 and dietary fiber20 prevented TPN-associated bacterial translocation. However, no studies have so far focused on the effect of the amount of caloric energy administered by TPN on bacterial translocation. This study was a first experiment to suggest that TPN with nonprotein energy overloading increased the incidence of bacterial translocation in newborn animals. Factors promoting bacterial translocation were generally divided into three groups: a physical disturbance of mucosal barrier, an impaired host immune defense mechanism, and an altered gut microbial ecology.8 In this study, histologically both the villous height and crypt depth in TPN-fed groups were significantly lower than those in the breast-fed control group, but no difference was observed between the TPN-fed groups in the intestinal structure. These findings suggest that a mechanical disturbance of the mucosal barrier might not play an important role in the bacterial translocation in this model. TPN impairs the immune function in many ways,21 including a reduction of the secretion of IgA,14 a decreased number of T lymphocytes in the intestinal mucosa,22 a depleted production of superoxide and the inhibition of phagocytosis activity by the macrophages19 in animal models. In clinical studies, Moore et al.23 and the Veterans Administration Cooperative study24 showed that patients receiving TPN demonstrated a significant increase in the number of infectious complications. However, an impairment of the immune function in these experimental and clinical studies might be primarily attributed to the absence of intraluminal nutrients rather than to TPN itself. Because these studies did not refer
to the effect of the amount of macronutrient, they could not explain why nonprotein energy overloading promoted bacterial translocation. We observed an elevated plasma level of triacylglycerol in the TPN-H group. It was reported that intravenously administered long-chain fatty acid emulsions impaired the clearance of bacteria in the reticuloendothelial system.25 We, therefore, suspected that translocated bacteria could not be killed by an impaired reticuloendothelial system, therefore, resulting in an increased incidence of bacterial translocation in the TPN-H group. Studies of bacterial translocation have used radioisotope-labeled bacteria to detect all translocated bacteria, whether they were viable or nonviable.26 Therefore, by using this technique, we could count the number of killed bacteria in the lymph nodes and other organs and evaluate immune function altered by TPN with nonprotein energy overloading. An impaired host immune defense was suspected to contribute the most to the increased incidence of bacterial translocation by parenteral energy overloading. However, as we did not evaluate the gut microbial ecology in this study, the true mechanism remains unclear. We could not clarify whether glucose or fat might increase the incidence of bacterial translocation. It has been reported that newborn rabbits could not tolerate glucose overloading, while in our preliminary study they could not tolerate fat overloading without a sufficient glucose intake. We, therefore, fixed the energy balance of glucose and fat to elucidate the difference in the energy intake between the two TPN groups. Additional studies, in which the total energy and the energy balance of glucose and fat are modified are, therefore, still required to clarify the mechanism of increased incidence of bacterial translocation. No study has yet shown any direct evidence that bacterial translocation during TPN causes liver dysfunction. According to
TABLE IV. MORPHOLOGIC PARAMETERS OF ILEAL MUCOSA
Villous height (mm) Crypt depth (mm)
FIG. 2. The relationship between bacterial translocation and the plasma levels of direct bilirubin. D.Bil, direct bilirubin; BT, bacterial translocation.
TPN-H
TPN-L
Control*
413.3 6 78.5
417.9 6 63.3
480.5 6 39.5†
111.0 6 21.5
104.2 6 17.9
134.4 6 19.7†
* Breast fed. † P , 0.05 compared with the TPN-H and TPN-L groups. TPN-H, high-energy total parenteral nutrition, TPN-L, low-energy total parenteral nutrition.
NEONATAL BACTERIAL TRANSLOCATION DURING TPN
447
the ‘‘gut origin sepsis’’ hypothesis, it is postulated that either translocated bacteria or endotoxin stimulate the macrophages to release inflammatory mediators and, therefore, results in multiple organ failure.13 At the same time these translocated bacteria and endotoxins could also induce cholestasis in liver. If bacterial translocation is the cause of TPN-associated cholestasis, especially in newborn infants, the fact that newborn infants have such an immature gut barrier system that they suffer from bacterial translocation16 would, therefore, correspond with the fact that they tend to develop cholestasis during TPN. But our study, which that first investigated the TPN model in newborn animals, did not show any close relationship between bacterial translocation and TPNassociated cholestasis. The results of our study were different from previous works on bacterial translocation in species of detected bacteria. In this study, most of the translocated bacteria were gram-positive species, a small number were gram-negative enteric bacilli, but no E. coli were found. Therefore, endotoxin was suspected to be produced only to a small degree in this experiment. Endotoxin is a well-known substance that induces inflammation in various tissues and also plays an important role in TPN-associated cholestasis.27
The reason why this study failed to show that bacterial translocation-induced cholestasis might partly be due to the fact that the animals had few gram-negative bacilli, which produced endotoxin, in their flora. Although newborn rabbits are used in experimental studies on neonatal bacterial translocation,28 the development of intestinal flora in newborn rabbits has not yet been precisely determined. Additional investigations on the regulation of intestinal flora, including inoculation of E. coli, are, therefore, called for to determine the effect of bacterial translocation on TPN-associated cholestasis. In conclusion, nonprotein energy overloading was found to increase the incidence of bacterial translocation in the neonatal rabbit TPN model. Immunosuppression induced by TPN might contribute to this result, but the true mechanism remains unclear. Bacterial translocation did not appear to be associated with the development of TPN-associated cholestasis. However, few translocated bacteria found in this study were gram-negative enteric bacilli producing endotoxin. Additional studies on the intestinal flora are, therefore, still called for to clarify the relationship between bacterial translocation and TPN-induced cholestasis.
REFERENCES 1. Berg RD, Garlington AW. Translocation of certain indigenous bacteria from the gastrointestinal tract to the mesenteric lymph nodes and other organs in a gnotobiotic mouse model. Infect Immun 1979;23:403 2. Alexander JW, Boyce ST, Babcock GF, et al. The process of microbial translocation. Ann Surg 1990;212:496 3. Berg RD. Inhibition of Escherichia coli translocation from the gastrointestinal tract by normal cecal flora in gnotobiotic or antibioticdecontaminated mice. Infect Immun 1980;29:1073 4. Brook I, MacVittie TJ, Walker RI. Recovery of aerobic and anaerobic bacteria from irradiated mice. Infect Immun 1984;46:270 5. Maejima K, Deitch EA, Berg RD. Bacterial translocation from the gastrointestinal tracts of rats receiving thermal injury. Infect Immun 1984;43:6 6. Bennion RS, Wilson SE, Williams RA. Early portal anaerobic bacteremia in mesenteric ischemia. Arch Surg 1984;119:151 7. Baker JW, Deitch EA, Li M, et al. Hemorrhagic shock induces bacterial translocation from the gut. J Trauma 1988;28:896 8. Deitch EA, Berg R, Specian R. Endotoxin promotes the translocation of bacteria from the gut. Arch Surg 1987;122:185 9. Deitch EA, Sittig K, Berg R, et al. Obstructive jaundice promotes bacterial translocation from the gut. Am J Surg 1990;159:79 10. Grant D, Hurlbut D, Zhong R, et al. Intestinal permeability and bacterial translocation following small bowel transplantation in the rat. Transplantation 1991;52:221 11. Kueppers PM, Miller TA, Chung-Ying K, et al. Effect of total parenteral nutrition plus morphine on bacterial translocation in rats. Ann Surg 1993;217:286 12. Lipman TO. Bacterial translocation and enteral nutrition in humans: an outsider looks in. JPEN 1995;19:156 13. Deitch EA. Multiple organ failure: pathophysiology and potential future therapy. Ann Surg 1992;216:117 14. Alverdy JC, Aoys E, Moss GS. Total parenteral nutrition promotes bacterial translocation from the gut. Surgery 1988;104:185 15. Pappo I, Bercovier H, Berry E, et al. Antitumor necrosis factor antibodies reduce hepatic steatosis during total parenteral nutrition and bowel rest in the rat. JPEN 1995;19:80
16. Van Camp JM, Tomaselli V, Coran AG. Bacterial translocation in the neonate. Curr Opin Pediatr 1994;6:327 17. Merritt RJ. Cholestasis associated with total parenteral nutrition. J Pediatr Gastroenterol Nutr 1986;5:9 18. Hata S, Kamata S, Nezu R, et al. A newborn rabbit model for total parenteral nutrition: Effects of nutritional components on cholestasis. JPEN 1989;13:265 19. Shou J, Lappin J, Minnard EA, Daly JM. Total parenteral nutrition, bacterial translocation, and host immune function. Am J Surg 1994; 167:145 20. Spaeth G, Specian RD, Berg RD, Deitch EA. Bulk prevents bacterial translocation induced by the oral administration of total parenteral solution. JPEN 1990;14:442 21. Alverdy JC, Burke D. Total parenteral nutrition: iatrogenic immunosuppression. Nutrition 1992;8:359 22. Alverdy JC, Aoys E, Weiss-Carrington P, Burke D. The effect of glutamine-enriched TPN on gut immune cellularity. J Surg Res 1992; 52:34 23. Moore FA, Moore EE, Jones TN, et al. TEN versus TPN following major abdominal trauma-Reduced septic morbidity. J Trauma 1989; 29:916 24. The Veterans Affairs Total Parenteral Nutrition Cooperative Study Group. Perioperative total parenteral nutrition in surgical patients. N Engl J Med 1991;325:525 25. Fischer GW, Hunter KW, Wilson SR, Mease AD. Diminished bacterial defenses with intralipid. Lancet 1980;2:819 26. Alexander JW, Gianotti L, Pyles T, Carey MA, Babcock GF. Distribution and survival of Escherichia coli translocating from the intestine after thermal injury. Ann Surg 1991;213:558 27. Latham PS, Menkes E, Phillips MJ, Jeejeebhoy KN. Hyperalimentation-associated jaundice: an example of a serum factor inducing cholestasis in rats. Am J Clin Nutr 1985;41:61 28. Go LL, Albanese CT, Watkins SC, Simmons RL, Rowe MI. Breast milk protects the neonate from bacterial translocation. J Pediatr Surg 1994;29:1059
Nutrition Vol. 14, No. 5, 1998
Non-Protein Energy Overloading Induces Bacterial Translocation During Total Parenteral Nutrition in Newborn Rabbits TAKESHI YAMANOUCHI, MD, SACHIYO SUITA, MD, AND KOUJI MASUMOTO, MD From the Department of Pediatric Surgery, Faculty of Medicine, Kyushu University, Fukuoka, Japan Date accepted: 25 October 1997 ABSTRACT
The effect of energy intake provided by total parenteral nutrition (TPN) on the incidence of bacterial translocation and the relationship between TPN-induced cholestasis and bacterial translocation were investigated in newborn animals. Forty-six Japanese white newborn rabbits were divided into three groups: TPN-H group (high energy TPN; 280 kcal z kg21 z d21), TPN-L group (low energy TPN; 180 kcal z kg21 z d21), and a control group (breast fed). On day 8, they were all killed for investigation of the presence of bacterial translocation, for blood chemistry analysis, and for histological examination of the ileum. The incidence of translocation to any of mesenteric lymph nodes and liver and spleen was significantly higher in the TPN-H group (67%) than in both the TPN-L group (13%) and the control group (10%) (P , 0.01). No difference was seen in ileum morphology between the two TPN groups. Although the mean bilirubin level of the TPN-H group tended to be higher than the TPN-L group, whether or not bacterial translocation occurred was not found to be closely related to the degree of TPN-associated cholestasis. In conclusion, parenteral nonprotein energy overloading increased the incidence of bacterial translocation in the newborn rabbit. However, bacterial translocation did not appear to be associated with the development of TPN-associated cholestasis. Nutrition 1998;14:443– 447. ©Elsevier Science Inc. 1998 Key words: bacterial translocation, total parenteral nutrition, cholestasis, newborn, rabbit
INTRODUCTION
Bacterial translocation is a term introduced by Berg and Garlington1 to describe the passage of viable bacteria from the gastrointestinal tract through the epithelial mucosa into the lamina propria and then to the mesenteric lymph nodes and possibly other organs. Alexander et al.2 later expanded the definition to include the passage of viable and nonviable microbes and their byproducts (including endotoxins) across an intestinal barrier. Bacterial translocation is promoted by several factors including a disruption of the intestinal microflora leading to bacterial overgrowth,3 radiation,4 burn,5 intestinal ischemia,6 hemorrhagic shock,7 endotoxin,8 obstructive jaundice,9 small bowel transplantation,10 and decreased intestinal motility.11 The significance of bacterial translocation in humans remains controversial,12 but in critically ill patients bacterial translocation is suspected to cause sepsis and multiple organ failure.13 Total parenteral nutrition (TPN) is believed to promote
bacterial translocation,14 but the effect of the energy content of TPN on the incidence of bacterial translocation has yet to be investigated. Bacterial translocation has been postulated to be one of the factors inducing liver dysfunction during TPN.15 It is widely known that newborn infants, especially premature infants, tend to develop cholestasis during TPN. The gut barrier and systemic immune response in newborn infants is so immature16 that neonates have been postulated to be prone to bacterial translocation from their own intestines. This phenomenon suggests bacterial translocation to be the primary cause of TPN-associated cholestasis in neonates, although the cause of TPN-associated cholestasis is generally indicated to be multifactorial.17 The aim of this study was, therefore, to investigate the effect of high caloric TPN on bacterial translocation and to also clarify the relationship between TPN-associated cholestasis and bacterial translocation in neonates.
Correspondence to: Takeshi Yamanouchi, MD, Department of Pediatric Surgery, Faculty of Medicine, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.
Nutrition 14:443– 447, 1998 ©Elsevier Science Inc. 1998 Printed in the USA. All rights reserved.
0899-9007/98/$19.00 PII S0899-9007(98)00015-X
444
NEONATAL BACTERIAL TRANSLOCATION DURING TPN TABLE I. COMPOSITION OF SOLUTIONS
Volume (mL z kg21 z d21) Carbohydrate (g z kg21 z d21) Amino acid† (g z kg21 z d21) Fat‡ (g z kg21 z d21) Energy (kcal z kg21 z d21) NP cal/N (kcal/g)
TPN-H
TPN-L
Control*
230 28 12 13.0 280 119
230 16 12 7.5 180 68
200 4 14 16.0 200 70
* Breast fed; composition of rabbit breast milk is based on Hata et al.18 † Amizet 10 (Tanabe Ph Co., Osaka, Japan). ‡ 20% Intralipos (Midorijuji Ph Co., Osaka, Japan). TPN-H, high-energy total parenteral nutrition; TPN-L, low-energy total parenteral nutrition; NP, nonprotein; N, nitrogen.
MATERIALS AND METHODS
Animals Adult female Japanese white rabbits with timed pregnancies of 14 to 15 d of gestation, weighing 3.5 to 4.5 kg, were obtained (Kyudo Co. Ltd., Fukuoka, Japan). The rabbits were housed in individual rabbit cages until delivery to the animal facility of Kyushu University. At 30 d of gestation, they were anesthetized with thyamiral (25 mg/kg body weight [BWT]) and their newborn rabbits were rapidly delivered by a hysterotomy. The newborn rabbits were quickly wrapped with gauze and placed in an infant incubator under conditions of 34°C temperature and 100% humidity. In the infant incubator their body weights were measured after birth and thereafter every day of the study. Thirty-six newborn rabbits, weighing from 40 to 70 g, were then used for the parenteral nutrition study. Ten newborn rabbits were breast fed for 8 d as a control, but birth weight could not be measured in order to prevent the mothers from engaging in cannibalism. The protocol of this study was approved by the Kyushu University Institutional Animal Care and Use Committee. Insertion of Central Venous Catheters Within 24 h after birth, newborn rabbits were anesthetized with ketamine (20 mg/kg BWT) subcutaneously and locally anesthetized with 1.0% lidocaine. The skin over the neck and midscapular region was then prepared with povidone iodine. With sterile instruments and using aseptic techniques, the jugular vein was isolated and ligated and a silastic catheter (0.012 in ID, 0.025 in OD, Dow-Corning Corp., Tokyo, Japan) was inserted through the jugular vein into the superior vena cava. The free end of the catheter was tunneled subcutaneously and brought out at the midscapular region and connected to the harness and the swivel. The newborn rabbits were housed in individual cages at temperatures ranging from 33 to 35°C throughout the study period. Nutritional Regimen After catheterization, the newborn rabbits were randomized into one of two different parenteral nutrition groups. The composition of the hyperalimentation solutions are summarized in Table I. The expected composition of rabbit breast milk is also shown in Table I. The animals in the high-energy total parenteral nutrition (TPN-H) group (n 5 19) received a high energy solution (280 kcal z kg21 z d21), which was thought to include a 40% higher energy intake than the breast-fed rabbits, whose glucose intake was thought to be the largest when the glucose energy accounts for
one-half of total energy. A dose of 28 g of glucose z kg21 z d21 was considered to be the upper limit for parenterally fed newborn rabbits due both to the fact that newborn rabbits are known to not tolerate glucose overloading,18 and based on our preliminary experiments. The animals in the low-energy total parenteral nutrition (TPN-L) group (n 5 17) received a lower energy hyperalimentation solution (180 kcal z kg21 z d21), which was thought to be a 10% lower energy intake than the breast-fed newborn rabbits (Table I). The amount of amino acids was equal in these two group and fat energy accounted for one-half the total nonprotein energy. On days 1 and 2, both groups were fed parenterally with the half-energy solution for acclimation, and from day 3 they were fed with full strength TPN for 6 d. The nutritional composition of the administered solution is indicated in Table I. Each solution was supplemented with electrolytes, trace minerals, and vitamins. The actual administered volume was arranged according to body weight, which was measured daily. The control group (n 5 10) was nourished by their lactating mothers for 8 d. Testing for Bacterial Translocation and Specimens At the end of the study period, each animal was anesthetized with ether and weighed. Using aseptic techniques, a laparotomy incision was made and a blood sample was obtained by an aortic puncture for serum chemistry. Next, the exposed viscera was swabbed with a sterile cotton-tipped applicator stick, which was cultured in tryptic soy broth (TSB, Eiken Co., Ltd., Tokyo, Japan) to detect any accidental bacterial contamination. The mesenteric lymph nodes complex (MLN), liver and spleen were removed and weighed, and then each was placed in grinding tubes containing a nine-fold volume of TSB. The organs were homogenized with sterile glass grinders. Then the sternum was opened aseptically and the central venous catheter tip was removed for an aerobic culture in TSB. If the catheter culture was positive, the data of this animal was excluded. Portions (0.2 mL) of organ homogenates were cultured aerobically on MacConkey’s agar (Eiken Co., Ltd.) to detect gram-negative enteric bacilli, and blood agar (Eiken Co., Ltd.) to detect gram-positive cocci. The plates were examined after 24 and 48 h of incubation at 37°C. Gram-positive bacteria was identified using standard procedures. Facultative gram-negative bacteria was identified by the API 20E System (Analytab Products, Plainview, NY, USA). The terminal ileum and the liver were excised and fixed in 10% formalin, embedded in paraffin, and stained with hematoxylin and eosin stain for the histological analysis. Both the intestinal structure and the liver histology were evaluated by light microscopy. The villous height and crypt depth were measured by a micrometer. A minimum of 10 villi were counted in each sample. Blood samples were centrifuged to obtain plasma samples. These samples were kept frozen at 220°C for additional chemical analysis. The plasma total bilirubin (T.Bil), direct bilirubin (D.Bil), glutamic-oxaloacetic transaminase (GOT), blood urea nitrogen (BUN), triacylglycerol (TG), total cholesterol (T.Chol), and blood sugar (BS) were determined by an autoanalyzer (Hitachi 736, Hitachi Co., Ltd., Tokyo, Japan). Statistics The incidence of translocation (discontinuous data) was evaluated with the Fisher’s exact test. The continuous data were expressed as the mean 6 SD and were tested by an analysis of variance with the Dunnet test. P values less than 0.05 were considered to be significant. RESULTS
No animal had any positive swab cultures in the peritoneum. There were three catheter-related infections; one in the TPN-H group and two in the TPN-L group. Therefore data were evaluated
NEONATAL BACTERIAL TRANSLOCATION DURING TPN
445
TABLE II. BODY WEIGHT AND LIVER WEIGHT
Initial body weight (g) Final body weight (g) Percent increase in body weight Liver weight/body weight (%)
TPN-H
TPN-L
Control*
57.2 6 6.6 85.3 6 8.8 49.7 6 8.7 3.44 6 0.62‡
59.1 6 6.2 84.8 6 10.4 43.2 6 7.4 2.88 6 0.52
137.2 6 16.3† 3.02 6 0.16
* Breast-fed. † P , 0.01 compared with the TPN-H and TPN-L groups. ‡ P , 0.05 compared with the TPN-L group. TPN-H, high-energy total parenteral nutrition; TPN-L, low-energy total parenteral nutrition.
in 18 newborn rabbits of the TPN-H group, 15 newborn rabbits in the TPN-L group, and 10 newborn rabbits in the control group. Body Weight Gain and Liver Weight The initial and final body weights are shown in Table II. The final body weight in the control group was higher than in the other groups. No significant difference was observed between the TPN-H group and the TPN-L group. The liver weight/body weight ratio was higher in the TPN-H group than in the TPN-L group, which, therefore, suggested hepatomegaly in the TPN-H group. The characteristic appearance of the animals in the TPN-H group was depilation, which was not recognized in the animals in the TPN-L or the control groups. Incidence of Bacterial Translocation and Species of Translocated Bacteria The results of translocation are shown in Figure 1. Translocation to the MLN occurred in 39% (7/18) of the animals in the TPN-H group, 7% (1/15) of the animals in the TPN-L group, and 10% (1/10) of the animals in the control group. This represents a significant increase in the MLN translocation in the TPN-H group compared to the TPN-L group (P , 0.05). Translocation to the liver occurred in 28% (5/18) of the animals in the TPN-H group, 7% (1/15) of the animals in the TPN-L group, and 0% (0/10) of the animals in the control group, while translocation to the spleen occurred in 33% (4/12) of the animals in the TPN-H group, 0%
(1/10) of the animals in the TPN-L group, and 0% (1/10) of the animals in the control group (not significant). Translocation to either MLN, the liver or spleen (total incidence of translocation) was 67% (12/18) of the animals in the TPN-H group, 13% (2/15) of the animals in the TPN-L group, and 10% (1/10) of the animals in the control group. This represents a significant increase in the TPN-H group compared with both the TPN-L group and the control group (P , 0.01). The translocated bacteria mainly included Staphylococcus aureus (47%), Corinebacterium (20%), and Enterococcus fecalis (13%). There was no difference in the species of translocated bacteria among the groups. Escherichia coli was not identified in any of the samples. Blood Chemistry The blood chemistry data are summarized in Table III. The plasma level of bilirubin in the TPN-H group was significantly higher than in the control group. The plasma level of GOT was not different among the three groups. The plasma level of BUN in the TPN-L group was significantly higher than in the other groups. The plasma level of triacylglycerol and blood sugar in the TPN-L group was significantly lower than in the other groups. To assess the relationship between TPN-associated cholestasis and bacterial translocation, all the TPN animals (TPN-H and TPN-L) were divided into two groups according to presence or absence of bacterial translocation. Figure 2 shows a comparison of the plasma level of direct bilirubin between the bacterial translocation-positive group and the bacterial translocation-negative group. The level of direct bilirubin was not related to the presence of bacterial translocation. Histology The histological findings of the liver samples showed cholestasis, which demonstrated bile plugs in the interlobular bile ducts and various bile pigments in the Kupffer cells, which were associated with increased levels of serum bilirubin. The villous height and crypt depth in both parenterally fed groups were lower than in the control group (P , 0.05). No difference was noted between the TPN-H and TPN-L groups (Table IV). No other pathological findings, including subepithelial edema, were observed in any of the groups. DISCUSSION 14
FIG. 1. Incidence of bacterial translocation in the mesenteric lymph nodes (MLN), liver, and spleen. The total incidence of translocation means any positive culture.
Alverdy et al. first showed that 2 wk of TPN increased the incidence of bacterial translocation, from 0% to 66%, in rat models. They, therefore, considered bacterial translocation to have contributed to bacterial overgrowth, while also impairing the intestinal defense and including lower levels of secretary IgA in
446
NEONATAL BACTERIAL TRANSLOCATION DURING TPN TABLE III. BLOOD CHEMISTRY
TPN-H TPN-L Control
Total bilirubin (mg/dL)
Direct bilirubin (mg/dL)
GOT (IU/L)
BUN (mg/dL)
Triacylglycerol (mg/dL)
Blood glucose (mg/dL)
1.34 6 0.92* 1.08 6 0.94 0.57 6 0.52*
1.11 6 0.90* 0.84 6 0.78 0.40 6 0.46*
36.6 6 21.2 35.2 6 31.5 43.3 6 43.3
23.5 6 7.1* 39.9 6 13.6*,† 19.9 6 10.8†
189.5 6 180.6* 55.0 6 44.0*,† 122.7 6 78.1†
123.0 6 41.5* 58.1 6 36.3*,† 208.8 6 24.6†
*,† Group means with the same superscripts are significantly different. GOT, glutamic-oxaloacetic transaminase; BUN, blood urea nitrogen; TPN-H, high-energy total parenteral nutrition, TPN-L, low-energy total parenteral nutrition.
bile. Thereafter, additional studies showed that enteral nutrition19 and dietary fiber20 prevented TPN-associated bacterial translocation. However, no studies have so far focused on the effect of the amount of caloric energy administered by TPN on bacterial translocation. This study was a first experiment to suggest that TPN with nonprotein energy overloading increased the incidence of bacterial translocation in newborn animals. Factors promoting bacterial translocation were generally divided into three groups: a physical disturbance of mucosal barrier, an impaired host immune defense mechanism, and an altered gut microbial ecology.8 In this study, histologically both the villous height and crypt depth in TPN-fed groups were significantly lower than those in the breast-fed control group, but no difference was observed between the TPN-fed groups in the intestinal structure. These findings suggest that a mechanical disturbance of the mucosal barrier might not play an important role in the bacterial translocation in this model. TPN impairs the immune function in many ways,21 including a reduction of the secretion of IgA,14 a decreased number of T lymphocytes in the intestinal mucosa,22 a depleted production of superoxide and the inhibition of phagocytosis activity by the macrophages19 in animal models. In clinical studies, Moore et al.23 and the Veterans Administration Cooperative study24 showed that patients receiving TPN demonstrated a significant increase in the number of infectious complications. However, an impairment of the immune function in these experimental and clinical studies might be primarily attributed to the absence of intraluminal nutrients rather than to TPN itself. Because these studies did not refer
to the effect of the amount of macronutrient, they could not explain why nonprotein energy overloading promoted bacterial translocation. We observed an elevated plasma level of triacylglycerol in the TPN-H group. It was reported that intravenously administered long-chain fatty acid emulsions impaired the clearance of bacteria in the reticuloendothelial system.25 We, therefore, suspected that translocated bacteria could not be killed by an impaired reticuloendothelial system, therefore, resulting in an increased incidence of bacterial translocation in the TPN-H group. Studies of bacterial translocation have used radioisotope-labeled bacteria to detect all translocated bacteria, whether they were viable or nonviable.26 Therefore, by using this technique, we could count the number of killed bacteria in the lymph nodes and other organs and evaluate immune function altered by TPN with nonprotein energy overloading. An impaired host immune defense was suspected to contribute the most to the increased incidence of bacterial translocation by parenteral energy overloading. However, as we did not evaluate the gut microbial ecology in this study, the true mechanism remains unclear. We could not clarify whether glucose or fat might increase the incidence of bacterial translocation. It has been reported that newborn rabbits could not tolerate glucose overloading, while in our preliminary study they could not tolerate fat overloading without a sufficient glucose intake. We, therefore, fixed the energy balance of glucose and fat to elucidate the difference in the energy intake between the two TPN groups. Additional studies, in which the total energy and the energy balance of glucose and fat are modified are, therefore, still required to clarify the mechanism of increased incidence of bacterial translocation. No study has yet shown any direct evidence that bacterial translocation during TPN causes liver dysfunction. According to
TABLE IV. MORPHOLOGIC PARAMETERS OF ILEAL MUCOSA
Villous height (mm) Crypt depth (mm)
FIG. 2. The relationship between bacterial translocation and the plasma levels of direct bilirubin. D.Bil, direct bilirubin; BT, bacterial translocation.
TPN-H
TPN-L
Control*
413.3 6 78.5
417.9 6 63.3
480.5 6 39.5†
111.0 6 21.5
104.2 6 17.9
134.4 6 19.7†
* Breast fed. † P , 0.05 compared with the TPN-H and TPN-L groups. TPN-H, high-energy total parenteral nutrition, TPN-L, low-energy total parenteral nutrition.
NEONATAL BACTERIAL TRANSLOCATION DURING TPN
447
the ‘‘gut origin sepsis’’ hypothesis, it is postulated that either translocated bacteria or endotoxin stimulate the macrophages to release inflammatory mediators and, therefore, results in multiple organ failure.13 At the same time these translocated bacteria and endotoxins could also induce cholestasis in liver. If bacterial translocation is the cause of TPN-associated cholestasis, especially in newborn infants, the fact that newborn infants have such an immature gut barrier system that they suffer from bacterial translocation16 would, therefore, correspond with the fact that they tend to develop cholestasis during TPN. But our study, which that first investigated the TPN model in newborn animals, did not show any close relationship between bacterial translocation and TPNassociated cholestasis. The results of our study were different from previous works on bacterial translocation in species of detected bacteria. In this study, most of the translocated bacteria were gram-positive species, a small number were gram-negative enteric bacilli, but no E. coli were found. Therefore, endotoxin was suspected to be produced only to a small degree in this experiment. Endotoxin is a well-known substance that induces inflammation in various tissues and also plays an important role in TPN-associated cholestasis.27
The reason why this study failed to show that bacterial translocation-induced cholestasis might partly be due to the fact that the animals had few gram-negative bacilli, which produced endotoxin, in their flora. Although newborn rabbits are used in experimental studies on neonatal bacterial translocation,28 the development of intestinal flora in newborn rabbits has not yet been precisely determined. Additional investigations on the regulation of intestinal flora, including inoculation of E. coli, are, therefore, called for to determine the effect of bacterial translocation on TPN-associated cholestasis. In conclusion, nonprotein energy overloading was found to increase the incidence of bacterial translocation in the neonatal rabbit TPN model. Immunosuppression induced by TPN might contribute to this result, but the true mechanism remains unclear. Bacterial translocation did not appear to be associated with the development of TPN-associated cholestasis. However, few translocated bacteria found in this study were gram-negative enteric bacilli producing endotoxin. Additional studies on the intestinal flora are, therefore, still called for to clarify the relationship between bacterial translocation and TPN-induced cholestasis.
REFERENCES 1. Berg RD, Garlington AW. Translocation of certain indigenous bacteria from the gastrointestinal tract to the mesenteric lymph nodes and other organs in a gnotobiotic mouse model. Infect Immun 1979;23:403 2. Alexander JW, Boyce ST, Babcock GF, et al. The process of microbial translocation. Ann Surg 1990;212:496 3. Berg RD. Inhibition of Escherichia coli translocation from the gastrointestinal tract by normal cecal flora in gnotobiotic or antibioticdecontaminated mice. Infect Immun 1980;29:1073 4. Brook I, MacVittie TJ, Walker RI. Recovery of aerobic and anaerobic bacteria from irradiated mice. Infect Immun 1984;46:270 5. Maejima K, Deitch EA, Berg RD. Bacterial translocation from the gastrointestinal tracts of rats receiving thermal injury. Infect Immun 1984;43:6 6. Bennion RS, Wilson SE, Williams RA. Early portal anaerobic bacteremia in mesenteric ischemia. Arch Surg 1984;119:151 7. Baker JW, Deitch EA, Li M, et al. Hemorrhagic shock induces bacterial translocation from the gut. J Trauma 1988;28:896 8. Deitch EA, Berg R, Specian R. Endotoxin promotes the translocation of bacteria from the gut. Arch Surg 1987;122:185 9. Deitch EA, Sittig K, Berg R, et al. Obstructive jaundice promotes bacterial translocation from the gut. Am J Surg 1990;159:79 10. Grant D, Hurlbut D, Zhong R, et al. Intestinal permeability and bacterial translocation following small bowel transplantation in the rat. Transplantation 1991;52:221 11. Kueppers PM, Miller TA, Chung-Ying K, et al. Effect of total parenteral nutrition plus morphine on bacterial translocation in rats. Ann Surg 1993;217:286 12. Lipman TO. Bacterial translocation and enteral nutrition in humans: an outsider looks in. JPEN 1995;19:156 13. Deitch EA. Multiple organ failure: pathophysiology and potential future therapy. Ann Surg 1992;216:117 14. Alverdy JC, Aoys E, Moss GS. Total parenteral nutrition promotes bacterial translocation from the gut. Surgery 1988;104:185 15. Pappo I, Bercovier H, Berry E, et al. Antitumor necrosis factor antibodies reduce hepatic steatosis during total parenteral nutrition and bowel rest in the rat. JPEN 1995;19:80
16. Van Camp JM, Tomaselli V, Coran AG. Bacterial translocation in the neonate. Curr Opin Pediatr 1994;6:327 17. Merritt RJ. Cholestasis associated with total parenteral nutrition. J Pediatr Gastroenterol Nutr 1986;5:9 18. Hata S, Kamata S, Nezu R, et al. A newborn rabbit model for total parenteral nutrition: Effects of nutritional components on cholestasis. JPEN 1989;13:265 19. Shou J, Lappin J, Minnard EA, Daly JM. Total parenteral nutrition, bacterial translocation, and host immune function. Am J Surg 1994; 167:145 20. Spaeth G, Specian RD, Berg RD, Deitch EA. Bulk prevents bacterial translocation induced by the oral administration of total parenteral solution. JPEN 1990;14:442 21. Alverdy JC, Burke D. Total parenteral nutrition: iatrogenic immunosuppression. Nutrition 1992;8:359 22. Alverdy JC, Aoys E, Weiss-Carrington P, Burke D. The effect of glutamine-enriched TPN on gut immune cellularity. J Surg Res 1992; 52:34 23. Moore FA, Moore EE, Jones TN, et al. TEN versus TPN following major abdominal trauma-Reduced septic morbidity. J Trauma 1989; 29:916 24. The Veterans Affairs Total Parenteral Nutrition Cooperative Study Group. Perioperative total parenteral nutrition in surgical patients. N Engl J Med 1991;325:525 25. Fischer GW, Hunter KW, Wilson SR, Mease AD. Diminished bacterial defenses with intralipid. Lancet 1980;2:819 26. Alexander JW, Gianotti L, Pyles T, Carey MA, Babcock GF. Distribution and survival of Escherichia coli translocating from the intestine after thermal injury. Ann Surg 1991;213:558 27. Latham PS, Menkes E, Phillips MJ, Jeejeebhoy KN. Hyperalimentation-associated jaundice: an example of a serum factor inducing cholestasis in rats. Am J Clin Nutr 1985;41:61 28. Go LL, Albanese CT, Watkins SC, Simmons RL, Rowe MI. Breast milk protects the neonate from bacterial translocation. J Pediatr Surg 1994;29:1059