Metabolic Responses to Enteral and Parenteral Nutrition H. L. A. Lickley, MD, Toronto, Ontario, Canada N. S. Track, PhD, Toronto, Ontario, Canada M. Vranic, MD, Toronto, Ontario, Canada K. D. Bury, MD, Toronto, Ontario, Canada
Although the basic materials for formulation of a complete nutritional support solution have been available for some time, problems of intravenous delivery have precluded clinical applicability. Dudrick et al [I] in 1968 developed a method whereby hypertonic solutions could be delivered directly into a central vein. Safe continuous infusion of hypertonic solutions containing glucose, amino acids, minerals, and vitamins are now carried out over long periods of time, resulting in positive caloric and nitrogen balance, weight gain, and complete nutritional support in a wide variety of clinical situations. Total parenteral nutrition (TPN) has become an important therapeutic ,advance for (a) the replenishment of nutritionally depleted patients, (b) the nutritional maintenance of patients when the oral feeding route is unavailable, or (c) when total bowel rest is required as part of the therapeutic management of problems such as inflammatory bowel disease [2,3]. A parallel development in clinical nutrition has been the perfection of chemically formulated liquid diet for oral administration [4]. Such diets are partially predigested, precisely formulated, and residue-free and have been modified to reduce osmolality. In selected clinical cases these diets have also maintained nutritional support in various disease states [5]. A rat model for the long-term continuous intravenous infusion of TPN solutions was first described by Steiger, Vars, and Dudrick [6] and adapted for use in our laboratory. The detailed composition of the solution which we found provided the best growth rate, nitrogen balance, and animal survival data has been described [7]. A previous study in which two groups of rats were
From the Departments of Surgery and Physiology, University of Toronto and McMaster University, Hamilton, and the Wellesley Hospital and Women’s College Hospital, Toronto, Ontario, Canada. This work was supported in part by the Medical Research Council of Canada, the Dorothy Frances Graham Fund, Women’s College Hospital, and the Atkinson Charitable Foundation. Reprint requests should be addressed to H. L. A. Lickley, MD, FRCS(C), C/O Women’s College Hospital, 76 Grenville Street, Toronto, Ontario, Canada M5S 182. Presented at the Eighteenth Annual Meeting of the Society for Surgery of the Alimentary Tract, Toronto, Ontario, Canada, May 24-25, 1977.
172
fed the same amount of the TPN solution orally or intravenously for three weeks revealed significantly greater weight gain in the orally fed group. It has been shown that the enteric administration of glucose gives rise to a greater release of insulin than an equivalent intravenous infusion with concomitant augmentation of the disposal of glucose [8]. There is evidence that the metabolic responses to enterically administered glucose are mediated through intestinal and hepatic factors [9]. Similar mechanisms may exist for the disposal of all ingested nutrients. Gastrointestinal augmentation of the release of pancreatic glucagon occurs after the oral administration of amino acids [IO]. Gut hormones have been implicated in gastrointestinal augmentation of insulin and glucagon release [II]. Secretin, pancreozymincholecystokinin, gastrin, glucagon-like immunoreactivity (GLI), gastric inhibitory polypeptide (GIP), and, vasoactive intestinal polypeptide have been suggested as candidates for a role in this response [12,13]. Material and Methods Seven pairs of male Biobreeding@ Wistar rats were studied. Under light Nembutale anesthesia, a Silastice catheter was inserted into the external jugular vein for intravenous feeding in seven of the animals. An identical catheter was placed into the stomach via a standard gastrostomy technic in the remaining animals. In all the animals the feeding catheters were tunneled subcutaneously and brought out in the midscapular region. The catheters were passed through a stainless, steel protective coil and attached to a swivel apparatus allowing mobility. Sterile polyethylene tubing was connected from the swivel apparatus through a #903 Halter@ pump to the sterile diet source. Paired intravenously and intragastrically fed animals were simultaneously infused through one Holter pump at an initial rate of 1 ml/hr, increasing over three days to a maximum rate of 2 ml/hr. The animals were placed in metabolic cages and allowed water ad libitum. The same diet was delivered to all animals and consisted of 12 per cent pure amino acids, 30 per cent dextrose monohydrate, and all the electrolytes, trace elements, and vitamins known to be essential for nutritional maintenance. The complete formulation of the diet has been previously
The American
Journal of Surgery
Enteral and Parenteral Nutrition
TABLE I
Fluid Balance Daily intake Diet + HP0
Daily Output Urine
52 f 1 ml 52 f 2 ml
30 f 1 ml 28 f 2 ml
Intravenously fed rats lntragastrically fed rats
Balance 22 f 24 f
1 ml 1 ml
described [7]. Fluid intake and urine output were measured daily. The initial mean weight of the intravenously fed animals was 233 f 4 gm and that of the intragastrically fed animals 231 f 5 gm. The animals were weighed at the end of the fourteen day study period. Blood samples were taken at the start of the study, at day 8, and at day 14 for the determination of hematocrit, blood glucose (BG), serum immunoreactive insulin (IRI), serum immunoreactive pancreatic glucagon (IRG), serum enteroglucagon or GLI, serum GIP, and serum gastrin levels. Initial values were obtained while the animals were receiving a normal laboratory diet ad libitum. Results were analyzed statistically using the paired t test.
0
2
4
6
8
10 12 14
Days
Figure 1. Mean changes in serum im*munoreactlve lnsulln (/RI) in seven pairs of normal rats which received Intravenous infusions of a glucose-amino acid mlxtura elthar intravenously (IV) or intragaotricafly (10). Vertical bars represent standard errors of the Mans.
Results
The rats receiving the standard diet intravenously gained 2.0 f 0.4 gm in weight per day, whereas the intragastrically infused animals gained 2.8 f 0.3 gm per day, which was a significantly (p < 0.05) greater weight gain. Fluid balance was similar in the two groups of animals. (Table I.) The increase in serum IRI levels was significantly greater (p < 0.001) in the intragastrically infused rats at day 8, but by day 14 the serum IRI levels had returned to initial values in both groups of animals. (Figure 1.) Serum IRG levels increased in both groups of animals, but the increase was greater (p < 0.005) in the intravenously fed rats at day 8. (Figure 2.) Serum GLI levels remained stable in the intragastrically infused rats until day 8 but subsequently declined significantly (p < 0.02). However, there was a marked decline in GLI in the intravenously infused rats which was evident by day 8. (Figure 3.) Serum levels of GLI were significantly higher in the intragastrically infused rats at both day 8 and day 14 (p < 0.001). Figure 4 shows that blood glucose levels are maintained throughout the infusion period in the intragastrically fed rats. There was an initial decrease in blood glucose level in the intravenously fed rats seen at day 8 with a partial return toward initial values by day 14. Serum levels of GIP were the same at the beginning of the infusion period and at day 14 in the intravenously infused rats but had declined significantly (p < 0.01) in the intragastrically fed rats by day 14. (Figure 5.) Serum gastrin levels decreased markedly by day 14 in both the intragastrically and intravenously infused rats. (Figure 6.)
Volume 135, February 1978
0
2
4
6
8
10 12 14
Figure 2. Mean changes in serum immunoreactlve pancreatlc giucagon ( IRG) In seven palm of normal rats whkh received intravenous Infusions of a glucose-amino acid mixture e/ther intravenously (IV) or /ntragastrka//y (IG). Vertical bars represent standard errors of the means.
600
I-
500
-
400
-
300
-
200
-
IG
IV
100 !0
2
I
I
I
4
6
8
I
I
1
10 12 14
Figure 3. Mean changes In gfucagon-Ike immunoreactivtty ( GLI) in seven pairs of normal rats whkh received /n&avenous infusions of a glucose-amino actd mixture elther intravenously (IV) or intragastrkally (IG). Vertical bars represent standard errors of the means.
173
Lickley et al
Comments
It has been suggested that the greater weight gain observed in rats given the amino acid-glucose mixture orally rather than parenterally could be explained on the basis of intermittent feeding in the former group of animals. In our studies the rats received the solution intragastrically or intravenously at precisely the same rate, and greater weight gain was observed in the intragastrically infused rats. Thus, other mechanisms must be involved in the preferential weight gain in the intragastrically infused rats. Since daily fluid intake and output were similar in both groups of animals, the greater weight gain in the latter group cannot be explained on the basis of fluid retention. Serum IRI levels increased significantly in both groups of animals by day 8, but by day 14 they had returned to preinfusion levels. The increase in serum
0
2
4
6
8 10 12 14
Days Figure 4. Mean changes in blood glucose (BG) in seven pairs of normal rafs which received intravenous infusions of a glucose-amino acid mixture either intravenously (IV) or intragastricaliy (IG). Vertical bars represent standard errors of the means.
IRI levels was significantly greater in the intragastrically infused rats. It is possible that greater release of insulin during the intragastric administration of the amino acid-glucose mixture contributed to better disposal of nutrients and thus greater weight gain in the enterically fed rats. Pancreatic glucagon levels increased significantly with the intravenous but not with the intragastric infusion of the amino acid-glucose solution. Both intraduodenal and intravenous infusions of pure amino acids have been shown to produce an increase in pancreatic glucagon [IO]. Hyperglycemia suppresses pancreatic glucagon release [14], but there is no firm evidence that the enteric infusion of glucose is associated with a greater suppression of pancreatic glucagon release than is the parenteral infusion of glucose. GLI was significantly higher in the intragastrically infused rats throughout the study period. It has been shown that the enteric administration of glucose is a potent stimulus to the release of GLI [15] and that GLI can stimulate the release of insulin [16]. Therefore, it has been suggested that GLI may be a gastrointestinal hormone involved in the “enteroinsular axis.” Glucagon is known to be a potent stimulus to insulin release [16]. Insulin, in turn, may be essential for the suppression of glucagon secretion by hyperglycemia, for Unger et al [ 171 have shown that there is a marked hyperglucagonemia in diabetes mellitus despite the presence of hyperglycemia. Thus, in the present study either GLI or IRI might have contributed to the lower levels of pancreatic glucagon seen in the intragastrically infused rats. Despite lower levels of IRI and higher levels of IRG
250 700 200
-I-
- 3 02 @_
6oo 500
l?r[ **
IV
%1IG
L
1
I
1
1
0
2
4
6
8
1 1 I 10 12 14
Days Figure 5. Mean changes in serum immunoreactive gastric inhibitory poiypeptide (G/P) in seven pairs of normal rats which received intravenous infusions of a glucose-amino acid mixture either intravenously (IV) or intragastricaliy (IG). Vertical bars represent standard errors of the means.
174
1 1 1 1 1 1 1 I 0 2 4 6 8 10 12 14 Figure 6. Mean changes in serum immunoreactive gastrin in seven pairs of normal rats which received intravenous infusions of a glucose-amino acid mixture either intravenously (IV) or intragastrically ( IG). Vertical bars represent standard errors of the means.
The American Journal of Surgery
Enteral
with the intravenous infusions, there was throughout the study a significant decrease in blood glucose that was not observed with the intragastric infusions. This finding perhaps indicates a greater renal clearance of glucose with the intravenous infusions. GIP levels were maintained during the intravenous infusions but decreased markedly with the intragastric infusions. This is somewhat unexpected, for the administration of oral glucose has been shown to increase circulating levels of GIP [18]. However, recent studies have shown that physiologic levels of exogenous insulin suppress the GIP response to intraduodenal glucose [19], suggesting the existence of negative feedback regulation of GIP release by endogenous insulin. Since higher levels of IRI were observed with the intragastric infusions in the present study, perhaps this could account for the suppression of serum GIP. Serum gastrin levels decreased markedly with the intravenous administration of the amino acid-glucose solution and also, though to a less marked degree, with the intragastric infusions. It will be important to determine whether or not these changes have occurred by day 8, at which point changes in IRI, IRG, and blood sugar levels are maximal. Summary
Seven pairs of rats were simultaneously infused with a chemically formulated nutritionally complete amino acid-glucose diet which was delivered, at the same rate, into a central vein or into a feeding gastrostomy. The intragastrically infused rats showed greater weight gain than did the intravenously infused rats. This could not be explained by fluid retention since intake and output were similar in the two groups of animals. There was a greater increase in serum immunoreactive insulin (IRI) at day 8 in the intragastrically infused animals, but a smaller increment in serum immunoreactive pancreatic glucagon (IRG) at that point. Levels of enteroglucagon or glucagon-like immunoreactivity (GLI) were maintained in the intragastrically infused rats but declined markedly in the intravenously infused rats. It is possible that the greater release of IRI seen with the intragastric amino acid-glucose feeding contributes to better disposal of nutrients and greater weight gain. The presence of nutrients in the intestinal lumen may have stimulated the release of GLI, which in turn is insulinotropic. Acknowledgment:
We thank M. Grayston, J. Kanarens, M. Kovacevic, and H. Karl for their excellent assistance.
Volume 135, February 1978
and Parenteral
Nutrition
References 1. Dudrick SJ, Wilmore DW, Vars HM, et al: Long-term total parenteral nutrition with growth, development, and positive nitrogen balance. Surgery 64: 134, 1966. 2. Aguirre A, Fischer JE, Welch CE: The role of surgery and hyperalimentation in the therapy of gastrointestinal-cutaneous fistulae. Ann Surg 180: 393. 1974. 3. Fischer JE, Foster GS, Abel RM, et al: Hyperalimentation as primary therapy for inflammatory bowel disease. Am J Surg 125: 165, 1973. 4. Bury KD. Stephens RV, Randall HT: Use of a chemically defined liquid elemental diet for nutritional management of flstulas of the alimentary tract. Am J Surg 121: 174, 1971. 5. Voitk A, Echave V, Feller JH, et al: Experience with elemental diet in the Rx of inflammatory bowel disease. Arch Surg 107: 329, 1973. 6. Steiger E. Vars HM, Dudrick SJ: A technique for long-term intravenous feeding in unrestrained rats. Arch Surg 104: 330, 1972. 7. Bury KD, Grayston M, Kanarens J: Modification of methodology and diet composition for prolonged intravenous feeding In unrestrained rats. Am J C/in Nutr (In press.) 8. McIntyre N, Holdsworth CD, Turner DS: Intestinal factors in the control of insulin secretion. J C/in Endocrinol Mefab 25: 1317, 1965. 9. Lickley HLA, Chisholm D, Rabinovitch A, Wexler M, Dupre J: Effects of portacaval anastomosis on glucose tolerance In the dog: evidence of an interaction between the gut and the liver in oral glucose tolerance. Metabolism 24: 1157, 1975. 10. Ohneda A, Parada E. Eisentraut AM, Unger RH: Characterization of response of circulating glucagon In intraduodenal and intravenous administration of amino acids. &I 47: 2305, 1968. 11. Dupre J, Chisholm DJ, Lickley HLA, et al: Regulatory functions of the intestine in the metabolism of nutrients. froc 4th Inf Congress Endocrinol (Excerpta Medica International Congress Series), Washington, 1972. 12. Dupre J, Curtis JD, Unger RH, Waddell RW, Beck JC: Effects of secretin, pancreozymin or gastrln on the response of the endocrine pancreas to administration of glucose or arginlne in man. &I 48: 745, 1969. 13. Brown JC, Dryburgh JR, Ross SA, Dupre J: Identlfioation and actions of gastric inhibitory polypeptide. Recent Prog Horn, Res 31: 487, 7975. 14. Unger RH, Eisentraut AM, McCall MS, Madison LL: Glucagon antibodies and an immunoassay for glucagon. $/ 40: 1280, 1961. 15. Samols E, Tyler J, Marri G, Marks V: lmmunochemlcal glucagon in human pancreas, gut and plasma. Lancet 2: 1257, 1966. 16. Samols E, Marri G, Marks V: Promotion of Insulin secretion by glucagon. Lancer 2: 415, 1965. 17. Unger RH, Aguilar-Parada E, Miiller WA, Eisentraut AM Studies of pancreatic alpha cell function in normal and diabetic subjects. JCI 49: 837, 1970. 18. Pederson RA, Schubert HE, Brown JC: Gastric inhibitory polypeptide: its physiological release and insullnotropic action in the dog. Diabetes 24: 1050, 1975. 19. Sirinek KR, Pace WG, Crockett SE, O’Dorislo TM, Mazzaferrl EL, Cataland S: Insulin-induced attenuation of glucosestimulated gastric inhibitory polypeptide secretion. Am J Surg 135: 151.1978.
Discussion Roger Keith cepted that the mones has been teroinsular axis,
(Toronto, Ontario, Canada): It is well acsecretin family of gastrointestinal horimplicated as the candidate for the enand the authors are suggesting that GLI
175
Lickley et al
may be that hormone. Brown and others have suggested that GIP, wit& its biphasic release after meals and especially with high carbohydrate, may be the hormone. They not only recommend we not rule out the old hormone secretin (the radioimmunoassays are not well accepted at the present time), but the data might suggest that secretin may be that candidate hormone as well. You have shown in the face of continued gastric infusion of nutrients a suppressed and continually suppressed gastrin level which is in keeping with secretin effect, and the significant elevation of insulin with a rather minimal elevation of GLI in the face of nutrient infusion may tend to put a little doubt on the GLI hypothesis. We have previously shown in humans that the constant gastric infusion of an elemental diet will tend to produce a secretin-like stimulating effect on the exocrine pancreatic function, albeit a lesser one than with regular feedings. Regarding the weight gain, do you have any data on the liver weight of your paired models, particularly data relative to any fatty change in the liver? Joel B. Freeman (Iowa City, IA): I am not sure that we can talk about trophic effects without knowing gastrin levels. Oral intake of any diet will maintain the small intestinal mucosal mass more efficiently than an intravenous infusion of the same diet. This effect may be mediated through gastrin. This comment is not intended to minify the beneficial effects of “intestinal rest” in the appropriate patient. Food given enterally should be superior, and that is what your study seems to indicate. As you know, we gave a typical total parenteral nutrition solution both enterally and intravenously to a group of normal volunteers using a crossover design. Nitrogen balance and weight gain were identical when the enteral and intravenous periods were compared. This leads me to believe that our subjects metabolized these nutrients equally well, regardless of the route of administration. However, on an individual basis some of the subjects did significantly worse when the diet was given through a nasogastric tube. This was probably due to osmotic diarrhea resulting in increased fecal nitrogen losses. Your study implied that the enteral route was superior, although it was in rats rather than human volunteers. I suspect there may be a hormonal cause for this. Do you have more data in this regard? Why did the rat diet consist of 50 per cent essential amino acids? I was slightly confused to observe that insulin levels decreased with feedings in your experiments. Generally insulin levels are elevated during the infusion of hypertonic dextrose. This is certainly true during total parenteral nutrition. One minor criticism: I would like to see insulin levels plotted against glucose levels. This makes ‘the data easier to interpret.
Alexander C. Charters, III (San Diego, CA): Have the authors compared oral administration with portal infusion? In a simplistic way, perhaps, if the liver gets first shot they might do almost as well.
176
Lawrence W. Way (San Francisco, CA): Were these differences maintained over time? It appeared that the differences were rapidly changing over the two week period of observation. What happened afterwards? H. L. A. Lickley (closing): Doctor Way, we have not tried longer infusion periods. In our study, the maximum changes were seen at day 8, and by day 14 most parameters measured seemed to have returned toward their original values. Also, if one does some mental gymnastics, although I realize that extrapolations can be dangerous, this two week period of TPN delivery in our rats would be roughly equivalent to two and a half years of TPN in a human. One final point with respect to your question, it is extremely difficult technically to maintain the animals on TPN for longer than two weeks. In answer to Doctor Charters’ question, the idea of comparing the responses to intraportal and intravenous TPN infusions is an intriguing one which came to our minds recently when someone raised the point that in our present study a different amino acid mixture reached the general circulation when delivered intragastrically as opposed to intravenously. It would be reasonable to assume that the amino acid composition of the nutrient solution would change upon initial passage through the liver, and certainly when an amino acid mixture is delivered peripherally there is uptake of branched chain amino acids by muscle. We plan to undertake a study such as you suggest comparing responses to intragastric, intraportal, and intravenous infusions in the near future. Preliminary gastin data did suggest that serum gastrin levels fell slightly less markedly with the enteric feeding. Doctor Freeman, nitrogen balance studies were carried out in our animals. A positive riitrogen balance of about 0.3 gm per day was achieved. As you noted, nitrogen balance was equivalent in the two groups of animals. Fifty per cent of the amino acid solution, not 50 per cent of the diet, was in the form of essential amino acids. The formula had been shown to give maximal weight gain in the rat. Serum insulin levels did increase during both external and parenteral feeding above baseline values, which were obtained in the fed state. In answer to Doctor Keith’s suggestion concerning the possible role of secretin in contributing to the higher insulin levels and greater weight gain seen in the intragastrically fed animals, studies carried out many years ago by Sum and Preshaw (Lancet 2: 340,1967) do shed some light on this. With the intraduodenal administration of glucose they were unable to demonstrate increased pancreatic output of fluid and electrolytes, suggesting that secretin is not released in response to the ingestion of glucose and therefore secretin is not likely to be the hormone involved in the enteroinsular axis. Confirmation of their findings using a secretin radioimmunoassay would be valuable, but most secretin assays to date have been fraught with problems and the stimulatory and inhibitory factors that determine secretin release have not yet been completely characterized. We do not yet have data on liver weights or fatty changes in the liver in our animals.
The American Journal of Surgery
Although the basic materials for formulation of a complete nutritional support solution have been available for some time, problems of intravenous delivery have precluded clinical applicability. Dudrick et al [I] in 1968 developed a method whereby hypertonic solutions could be delivered directly into a central vein. Safe continuous infusion of hypertonic solutions containing glucose, amino acids, minerals, and vitamins are now carried out over long periods of time, resulting in positive caloric and nitrogen balance, weight gain, and complete nutritional support in a wide variety of clinical situations. Total parenteral nutrition (TPN) has become an important therapeutic ,advance for (a) the replenishment of nutritionally depleted patients, (b) the nutritional maintenance of patients when the oral feeding route is unavailable, or (c) when total bowel rest is required as part of the therapeutic management of problems such as inflammatory bowel disease [2,3]. A parallel development in clinical nutrition has been the perfection of chemically formulated liquid diet for oral administration [4]. Such diets are partially predigested, precisely formulated, and residue-free and have been modified to reduce osmolality. In selected clinical cases these diets have also maintained nutritional support in various disease states [5]. A rat model for the long-term continuous intravenous infusion of TPN solutions was first described by Steiger, Vars, and Dudrick [6] and adapted for use in our laboratory. The detailed composition of the solution which we found provided the best growth rate, nitrogen balance, and animal survival data has been described [7]. A previous study in which two groups of rats were
From the Departments of Surgery and Physiology, University of Toronto and McMaster University, Hamilton, and the Wellesley Hospital and Women’s College Hospital, Toronto, Ontario, Canada. This work was supported in part by the Medical Research Council of Canada, the Dorothy Frances Graham Fund, Women’s College Hospital, and the Atkinson Charitable Foundation. Reprint requests should be addressed to H. L. A. Lickley, MD, FRCS(C), C/O Women’s College Hospital, 76 Grenville Street, Toronto, Ontario, Canada M5S 182. Presented at the Eighteenth Annual Meeting of the Society for Surgery of the Alimentary Tract, Toronto, Ontario, Canada, May 24-25, 1977.
172
fed the same amount of the TPN solution orally or intravenously for three weeks revealed significantly greater weight gain in the orally fed group. It has been shown that the enteric administration of glucose gives rise to a greater release of insulin than an equivalent intravenous infusion with concomitant augmentation of the disposal of glucose [8]. There is evidence that the metabolic responses to enterically administered glucose are mediated through intestinal and hepatic factors [9]. Similar mechanisms may exist for the disposal of all ingested nutrients. Gastrointestinal augmentation of the release of pancreatic glucagon occurs after the oral administration of amino acids [IO]. Gut hormones have been implicated in gastrointestinal augmentation of insulin and glucagon release [II]. Secretin, pancreozymincholecystokinin, gastrin, glucagon-like immunoreactivity (GLI), gastric inhibitory polypeptide (GIP), and, vasoactive intestinal polypeptide have been suggested as candidates for a role in this response [12,13]. Material and Methods Seven pairs of male Biobreeding@ Wistar rats were studied. Under light Nembutale anesthesia, a Silastice catheter was inserted into the external jugular vein for intravenous feeding in seven of the animals. An identical catheter was placed into the stomach via a standard gastrostomy technic in the remaining animals. In all the animals the feeding catheters were tunneled subcutaneously and brought out in the midscapular region. The catheters were passed through a stainless, steel protective coil and attached to a swivel apparatus allowing mobility. Sterile polyethylene tubing was connected from the swivel apparatus through a #903 Halter@ pump to the sterile diet source. Paired intravenously and intragastrically fed animals were simultaneously infused through one Holter pump at an initial rate of 1 ml/hr, increasing over three days to a maximum rate of 2 ml/hr. The animals were placed in metabolic cages and allowed water ad libitum. The same diet was delivered to all animals and consisted of 12 per cent pure amino acids, 30 per cent dextrose monohydrate, and all the electrolytes, trace elements, and vitamins known to be essential for nutritional maintenance. The complete formulation of the diet has been previously
The American
Journal of Surgery
Enteral and Parenteral Nutrition
TABLE I
Fluid Balance Daily intake Diet + HP0
Daily Output Urine
52 f 1 ml 52 f 2 ml
30 f 1 ml 28 f 2 ml
Intravenously fed rats lntragastrically fed rats
Balance 22 f 24 f
1 ml 1 ml
described [7]. Fluid intake and urine output were measured daily. The initial mean weight of the intravenously fed animals was 233 f 4 gm and that of the intragastrically fed animals 231 f 5 gm. The animals were weighed at the end of the fourteen day study period. Blood samples were taken at the start of the study, at day 8, and at day 14 for the determination of hematocrit, blood glucose (BG), serum immunoreactive insulin (IRI), serum immunoreactive pancreatic glucagon (IRG), serum enteroglucagon or GLI, serum GIP, and serum gastrin levels. Initial values were obtained while the animals were receiving a normal laboratory diet ad libitum. Results were analyzed statistically using the paired t test.
0
2
4
6
8
10 12 14
Days
Figure 1. Mean changes in serum im*munoreactlve lnsulln (/RI) in seven pairs of normal rats which received Intravenous infusions of a glucose-amino acid mlxtura elthar intravenously (IV) or intragaotricafly (10). Vertical bars represent standard errors of the Mans.
Results
The rats receiving the standard diet intravenously gained 2.0 f 0.4 gm in weight per day, whereas the intragastrically infused animals gained 2.8 f 0.3 gm per day, which was a significantly (p < 0.05) greater weight gain. Fluid balance was similar in the two groups of animals. (Table I.) The increase in serum IRI levels was significantly greater (p < 0.001) in the intragastrically infused rats at day 8, but by day 14 the serum IRI levels had returned to initial values in both groups of animals. (Figure 1.) Serum IRG levels increased in both groups of animals, but the increase was greater (p < 0.005) in the intravenously fed rats at day 8. (Figure 2.) Serum GLI levels remained stable in the intragastrically infused rats until day 8 but subsequently declined significantly (p < 0.02). However, there was a marked decline in GLI in the intravenously infused rats which was evident by day 8. (Figure 3.) Serum levels of GLI were significantly higher in the intragastrically infused rats at both day 8 and day 14 (p < 0.001). Figure 4 shows that blood glucose levels are maintained throughout the infusion period in the intragastrically fed rats. There was an initial decrease in blood glucose level in the intravenously fed rats seen at day 8 with a partial return toward initial values by day 14. Serum levels of GIP were the same at the beginning of the infusion period and at day 14 in the intravenously infused rats but had declined significantly (p < 0.01) in the intragastrically fed rats by day 14. (Figure 5.) Serum gastrin levels decreased markedly by day 14 in both the intragastrically and intravenously infused rats. (Figure 6.)
Volume 135, February 1978
0
2
4
6
8
10 12 14
Figure 2. Mean changes in serum immunoreactlve pancreatlc giucagon ( IRG) In seven palm of normal rats whkh received intravenous Infusions of a glucose-amino acid mixture e/ther intravenously (IV) or /ntragastrka//y (IG). Vertical bars represent standard errors of the means.
600
I-
500
-
400
-
300
-
200
-
IG
IV
100 !0
2
I
I
I
4
6
8
I
I
1
10 12 14
Figure 3. Mean changes In gfucagon-Ike immunoreactivtty ( GLI) in seven pairs of normal rats whkh received /n&avenous infusions of a glucose-amino actd mixture elther intravenously (IV) or intragastrkally (IG). Vertical bars represent standard errors of the means.
173
Lickley et al
Comments
It has been suggested that the greater weight gain observed in rats given the amino acid-glucose mixture orally rather than parenterally could be explained on the basis of intermittent feeding in the former group of animals. In our studies the rats received the solution intragastrically or intravenously at precisely the same rate, and greater weight gain was observed in the intragastrically infused rats. Thus, other mechanisms must be involved in the preferential weight gain in the intragastrically infused rats. Since daily fluid intake and output were similar in both groups of animals, the greater weight gain in the latter group cannot be explained on the basis of fluid retention. Serum IRI levels increased significantly in both groups of animals by day 8, but by day 14 they had returned to preinfusion levels. The increase in serum
0
2
4
6
8 10 12 14
Days Figure 4. Mean changes in blood glucose (BG) in seven pairs of normal rafs which received intravenous infusions of a glucose-amino acid mixture either intravenously (IV) or intragastricaliy (IG). Vertical bars represent standard errors of the means.
IRI levels was significantly greater in the intragastrically infused rats. It is possible that greater release of insulin during the intragastric administration of the amino acid-glucose mixture contributed to better disposal of nutrients and thus greater weight gain in the enterically fed rats. Pancreatic glucagon levels increased significantly with the intravenous but not with the intragastric infusion of the amino acid-glucose solution. Both intraduodenal and intravenous infusions of pure amino acids have been shown to produce an increase in pancreatic glucagon [IO]. Hyperglycemia suppresses pancreatic glucagon release [14], but there is no firm evidence that the enteric infusion of glucose is associated with a greater suppression of pancreatic glucagon release than is the parenteral infusion of glucose. GLI was significantly higher in the intragastrically infused rats throughout the study period. It has been shown that the enteric administration of glucose is a potent stimulus to the release of GLI [15] and that GLI can stimulate the release of insulin [16]. Therefore, it has been suggested that GLI may be a gastrointestinal hormone involved in the “enteroinsular axis.” Glucagon is known to be a potent stimulus to insulin release [16]. Insulin, in turn, may be essential for the suppression of glucagon secretion by hyperglycemia, for Unger et al [ 171 have shown that there is a marked hyperglucagonemia in diabetes mellitus despite the presence of hyperglycemia. Thus, in the present study either GLI or IRI might have contributed to the lower levels of pancreatic glucagon seen in the intragastrically infused rats. Despite lower levels of IRI and higher levels of IRG
250 700 200
-I-
- 3 02 @_
6oo 500
l?r[ **
IV
%1IG
L
1
I
1
1
0
2
4
6
8
1 1 I 10 12 14
Days Figure 5. Mean changes in serum immunoreactive gastric inhibitory poiypeptide (G/P) in seven pairs of normal rats which received intravenous infusions of a glucose-amino acid mixture either intravenously (IV) or intragastricaliy (IG). Vertical bars represent standard errors of the means.
174
1 1 1 1 1 1 1 I 0 2 4 6 8 10 12 14 Figure 6. Mean changes in serum immunoreactive gastrin in seven pairs of normal rats which received intravenous infusions of a glucose-amino acid mixture either intravenously (IV) or intragastrically ( IG). Vertical bars represent standard errors of the means.
The American Journal of Surgery
Enteral
with the intravenous infusions, there was throughout the study a significant decrease in blood glucose that was not observed with the intragastric infusions. This finding perhaps indicates a greater renal clearance of glucose with the intravenous infusions. GIP levels were maintained during the intravenous infusions but decreased markedly with the intragastric infusions. This is somewhat unexpected, for the administration of oral glucose has been shown to increase circulating levels of GIP [18]. However, recent studies have shown that physiologic levels of exogenous insulin suppress the GIP response to intraduodenal glucose [19], suggesting the existence of negative feedback regulation of GIP release by endogenous insulin. Since higher levels of IRI were observed with the intragastric infusions in the present study, perhaps this could account for the suppression of serum GIP. Serum gastrin levels decreased markedly with the intravenous administration of the amino acid-glucose solution and also, though to a less marked degree, with the intragastric infusions. It will be important to determine whether or not these changes have occurred by day 8, at which point changes in IRI, IRG, and blood sugar levels are maximal. Summary
Seven pairs of rats were simultaneously infused with a chemically formulated nutritionally complete amino acid-glucose diet which was delivered, at the same rate, into a central vein or into a feeding gastrostomy. The intragastrically infused rats showed greater weight gain than did the intravenously infused rats. This could not be explained by fluid retention since intake and output were similar in the two groups of animals. There was a greater increase in serum immunoreactive insulin (IRI) at day 8 in the intragastrically infused animals, but a smaller increment in serum immunoreactive pancreatic glucagon (IRG) at that point. Levels of enteroglucagon or glucagon-like immunoreactivity (GLI) were maintained in the intragastrically infused rats but declined markedly in the intravenously infused rats. It is possible that the greater release of IRI seen with the intragastric amino acid-glucose feeding contributes to better disposal of nutrients and greater weight gain. The presence of nutrients in the intestinal lumen may have stimulated the release of GLI, which in turn is insulinotropic. Acknowledgment:
We thank M. Grayston, J. Kanarens, M. Kovacevic, and H. Karl for their excellent assistance.
Volume 135, February 1978
and Parenteral
Nutrition
References 1. Dudrick SJ, Wilmore DW, Vars HM, et al: Long-term total parenteral nutrition with growth, development, and positive nitrogen balance. Surgery 64: 134, 1966. 2. Aguirre A, Fischer JE, Welch CE: The role of surgery and hyperalimentation in the therapy of gastrointestinal-cutaneous fistulae. Ann Surg 180: 393. 1974. 3. Fischer JE, Foster GS, Abel RM, et al: Hyperalimentation as primary therapy for inflammatory bowel disease. Am J Surg 125: 165, 1973. 4. Bury KD. Stephens RV, Randall HT: Use of a chemically defined liquid elemental diet for nutritional management of flstulas of the alimentary tract. Am J Surg 121: 174, 1971. 5. Voitk A, Echave V, Feller JH, et al: Experience with elemental diet in the Rx of inflammatory bowel disease. Arch Surg 107: 329, 1973. 6. Steiger E. Vars HM, Dudrick SJ: A technique for long-term intravenous feeding in unrestrained rats. Arch Surg 104: 330, 1972. 7. Bury KD, Grayston M, Kanarens J: Modification of methodology and diet composition for prolonged intravenous feeding In unrestrained rats. Am J C/in Nutr (In press.) 8. McIntyre N, Holdsworth CD, Turner DS: Intestinal factors in the control of insulin secretion. J C/in Endocrinol Mefab 25: 1317, 1965. 9. Lickley HLA, Chisholm D, Rabinovitch A, Wexler M, Dupre J: Effects of portacaval anastomosis on glucose tolerance In the dog: evidence of an interaction between the gut and the liver in oral glucose tolerance. Metabolism 24: 1157, 1975. 10. Ohneda A, Parada E. Eisentraut AM, Unger RH: Characterization of response of circulating glucagon In intraduodenal and intravenous administration of amino acids. &I 47: 2305, 1968. 11. Dupre J, Chisholm DJ, Lickley HLA, et al: Regulatory functions of the intestine in the metabolism of nutrients. froc 4th Inf Congress Endocrinol (Excerpta Medica International Congress Series), Washington, 1972. 12. Dupre J, Curtis JD, Unger RH, Waddell RW, Beck JC: Effects of secretin, pancreozymin or gastrln on the response of the endocrine pancreas to administration of glucose or arginlne in man. &I 48: 745, 1969. 13. Brown JC, Dryburgh JR, Ross SA, Dupre J: Identlfioation and actions of gastric inhibitory polypeptide. Recent Prog Horn, Res 31: 487, 7975. 14. Unger RH, Eisentraut AM, McCall MS, Madison LL: Glucagon antibodies and an immunoassay for glucagon. $/ 40: 1280, 1961. 15. Samols E, Tyler J, Marri G, Marks V: lmmunochemlcal glucagon in human pancreas, gut and plasma. Lancet 2: 1257, 1966. 16. Samols E, Marri G, Marks V: Promotion of Insulin secretion by glucagon. Lancer 2: 415, 1965. 17. Unger RH, Aguilar-Parada E, Miiller WA, Eisentraut AM Studies of pancreatic alpha cell function in normal and diabetic subjects. JCI 49: 837, 1970. 18. Pederson RA, Schubert HE, Brown JC: Gastric inhibitory polypeptide: its physiological release and insullnotropic action in the dog. Diabetes 24: 1050, 1975. 19. Sirinek KR, Pace WG, Crockett SE, O’Dorislo TM, Mazzaferrl EL, Cataland S: Insulin-induced attenuation of glucosestimulated gastric inhibitory polypeptide secretion. Am J Surg 135: 151.1978.
Discussion Roger Keith cepted that the mones has been teroinsular axis,
(Toronto, Ontario, Canada): It is well acsecretin family of gastrointestinal horimplicated as the candidate for the enand the authors are suggesting that GLI
175
Lickley et al
may be that hormone. Brown and others have suggested that GIP, wit& its biphasic release after meals and especially with high carbohydrate, may be the hormone. They not only recommend we not rule out the old hormone secretin (the radioimmunoassays are not well accepted at the present time), but the data might suggest that secretin may be that candidate hormone as well. You have shown in the face of continued gastric infusion of nutrients a suppressed and continually suppressed gastrin level which is in keeping with secretin effect, and the significant elevation of insulin with a rather minimal elevation of GLI in the face of nutrient infusion may tend to put a little doubt on the GLI hypothesis. We have previously shown in humans that the constant gastric infusion of an elemental diet will tend to produce a secretin-like stimulating effect on the exocrine pancreatic function, albeit a lesser one than with regular feedings. Regarding the weight gain, do you have any data on the liver weight of your paired models, particularly data relative to any fatty change in the liver? Joel B. Freeman (Iowa City, IA): I am not sure that we can talk about trophic effects without knowing gastrin levels. Oral intake of any diet will maintain the small intestinal mucosal mass more efficiently than an intravenous infusion of the same diet. This effect may be mediated through gastrin. This comment is not intended to minify the beneficial effects of “intestinal rest” in the appropriate patient. Food given enterally should be superior, and that is what your study seems to indicate. As you know, we gave a typical total parenteral nutrition solution both enterally and intravenously to a group of normal volunteers using a crossover design. Nitrogen balance and weight gain were identical when the enteral and intravenous periods were compared. This leads me to believe that our subjects metabolized these nutrients equally well, regardless of the route of administration. However, on an individual basis some of the subjects did significantly worse when the diet was given through a nasogastric tube. This was probably due to osmotic diarrhea resulting in increased fecal nitrogen losses. Your study implied that the enteral route was superior, although it was in rats rather than human volunteers. I suspect there may be a hormonal cause for this. Do you have more data in this regard? Why did the rat diet consist of 50 per cent essential amino acids? I was slightly confused to observe that insulin levels decreased with feedings in your experiments. Generally insulin levels are elevated during the infusion of hypertonic dextrose. This is certainly true during total parenteral nutrition. One minor criticism: I would like to see insulin levels plotted against glucose levels. This makes ‘the data easier to interpret.
Alexander C. Charters, III (San Diego, CA): Have the authors compared oral administration with portal infusion? In a simplistic way, perhaps, if the liver gets first shot they might do almost as well.
176
Lawrence W. Way (San Francisco, CA): Were these differences maintained over time? It appeared that the differences were rapidly changing over the two week period of observation. What happened afterwards? H. L. A. Lickley (closing): Doctor Way, we have not tried longer infusion periods. In our study, the maximum changes were seen at day 8, and by day 14 most parameters measured seemed to have returned toward their original values. Also, if one does some mental gymnastics, although I realize that extrapolations can be dangerous, this two week period of TPN delivery in our rats would be roughly equivalent to two and a half years of TPN in a human. One final point with respect to your question, it is extremely difficult technically to maintain the animals on TPN for longer than two weeks. In answer to Doctor Charters’ question, the idea of comparing the responses to intraportal and intravenous TPN infusions is an intriguing one which came to our minds recently when someone raised the point that in our present study a different amino acid mixture reached the general circulation when delivered intragastrically as opposed to intravenously. It would be reasonable to assume that the amino acid composition of the nutrient solution would change upon initial passage through the liver, and certainly when an amino acid mixture is delivered peripherally there is uptake of branched chain amino acids by muscle. We plan to undertake a study such as you suggest comparing responses to intragastric, intraportal, and intravenous infusions in the near future. Preliminary gastin data did suggest that serum gastrin levels fell slightly less markedly with the enteric feeding. Doctor Freeman, nitrogen balance studies were carried out in our animals. A positive riitrogen balance of about 0.3 gm per day was achieved. As you noted, nitrogen balance was equivalent in the two groups of animals. Fifty per cent of the amino acid solution, not 50 per cent of the diet, was in the form of essential amino acids. The formula had been shown to give maximal weight gain in the rat. Serum insulin levels did increase during both external and parenteral feeding above baseline values, which were obtained in the fed state. In answer to Doctor Keith’s suggestion concerning the possible role of secretin in contributing to the higher insulin levels and greater weight gain seen in the intragastrically fed animals, studies carried out many years ago by Sum and Preshaw (Lancet 2: 340,1967) do shed some light on this. With the intraduodenal administration of glucose they were unable to demonstrate increased pancreatic output of fluid and electrolytes, suggesting that secretin is not released in response to the ingestion of glucose and therefore secretin is not likely to be the hormone involved in the enteroinsular axis. Confirmation of their findings using a secretin radioimmunoassay would be valuable, but most secretin assays to date have been fraught with problems and the stimulatory and inhibitory factors that determine secretin release have not yet been completely characterized. We do not yet have data on liver weights or fatty changes in the liver in our animals.
The American Journal of Surgery
