- Email: [email protected]
The effect of mesenteric venous hypertension on gut motility and absorption
JOURNAL
OF SURGICAL
RESEARCH
48,562-567
(1990)
The Effect of Mesenteric Venous Hypertension on Gut Motility and Absorption DONALD L. JACOBS, M.D., JOHN LOF, M.S., EAMONN M. M. QUIGI,EY, M.D., ALI D. SPANTA, M.D., AND LAYTON F. RIKKERS, M.D. Ilepartments
of Surgery and Internal
Medicine,
University
of Nebraska Medical Center, 42nd Street and Dewey Avenue, Omaha, Nebraska 68105
Presentedat the Annual Meeting of the Association
for Academic Surgery, Louisville,
November
15-18, 1989.
may also be operative. Altered gut motility, for example, could also influence intestinal absorption [8]. We have previously demonstrated alterations in both gastric emptying [9, lo] and small intestinal transit [ll] in portal hypertension. The pathogenesis of these motor changes and their contribution to impaired absorptive function in portal hypertension have not been defined. In particular, it has not been established whether these changes in small intestinal physiology are entirely consequent upon passive venous congestion or whether they reflect ‘the effects of altered levels of enteric hormones and neuropeptides as a result of portasystemic shunting [ 12, 131. Our aim, therefore, was to assess the effects of chronic mesenteric venous hypertension, in the absence of hepatic dysfunction or portasystemic shunting, on intestinal motility and absorption.
The effects of portal hypertension on gut function may be mediated by venous congestion and altered circulating levels of enteric hormones and neuropeptides. We designed this study to determine the effects of chronic intestinal venous hypertension (VHT), in isolation, on gut motility and absorption. In 10 dogs, a 20- to 26-cm loop of jejunum was isolated from the fecal stream, but myoneural continuity was maintained with the proximal bowel by a seromuscular bridge. In 5 dogs, VHT was created in the loop by a fixed stenosis of its venous drainage; a sham procedure was performed in a further 5 animals. Serosal monopolar electrodes were placed in all animals. Absorptive function and myoelectrical activity were studied over a I-week period. Venous hypertension was achieved and sustained in the VHT animals; loop vein pressures for VHT vs control in cm Hz0 (means f SEM) are: initial-29.8 + 1.8 vs 7.5 + 0.4 (P < O.Ol), and at 4 weeko17.6 f 6.88 vs 7.3 f 0.2 (P < 0.01). Absorption of Na+, Cl , glucose, and water was impaired in VHT loops. Normal patterns of fasting and postprandial myoelectrical activity were preserved in the VHT animals. We conclude that chronic VHT, in the absence of portosystemic shunting, results in impaired absorption of water, glucose, and electrolytes without any change in in& 1990 Academic Press. Inc. testinal motility.
METHODS Preparation
of Animals
Ten adult female mongrel dogs, weighing between 15 and 20 kg, underwent celiotomy under pentobarbital sodium anesthesia. Beginning at a point 30 cm distal to the ligament of Treitz, a 20- to 25-cm segment of jejunum, drained by either a single or at most two veins at the base of the mesentery, was identified. This segment was fashioned into a loop which, though isolated from the lumen of the gastrointestinal tract, maintained neuromuscular continuity with the proximal small intestine via a bridge of tunica muscularis between the proximal end of the loop and the intact jejunum [14-161. A Thomas cannula was placed in the loop near the proximal end and brought out to the abdominal wall through a separate incision. The distal end of the loop was brought out to the abdominal wall as an end jejunostomy (Fig. 1). Two weeks later all animals underwent a second celiotomy under general anesthesia. The vascular supply to the loop was identified and skeletonized at the base of the mesentery. To permit intraoperative recording of venous pressure, a needle connected to a water-filled but non-
INTRODUCTION The etiology of malnutrition associated with advanced cirrhosis and portal hypertension is multifactorial and thought to include such factors as poor nutritional intake, impaired hepatic metabolism of absorbed nutrients, and intestinal malabsorption [l-3]. With respect to the latter, it has been suggested that intestinal venous congestion consequent on portal hypertension may directly alter absorptive function [3, 41. Indeed, previous studies, in animals, have demonstrated that acute intestinal venous hypertension causes a decrease in the absorption of water, sodium, chloride, and ammonia [5, 61. Furthermore, chronic intestinal venous hypertension results in malabsorption of vitamin D:, in the rat [7]. Other mechanisms oozz-4804/90 $1.50 Ckapyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Kentucky,
562
JACOBS
ET AL.: MESENTERIC
VENOUS
FIG. 1. Experimental model: bridged loop transection. A 20- to 25 cm loop that is luminally isolated yet maintains myoneural continuity with the proximal intestine via the bridge of tunica muscularis (inset). Note the location of the serosal monopolar electrodes.
perfused catheter was inserted into a peripheral vein in the venous arcade of the loop. The catheter was, in turn, connected to a saline-filled manometer. In five dogs mesenteric venous hypertension was created in the loop (VHT group) as follows: a calibrated stenosis of the previously defined, principal draining vein
HYPERTENSION
AND
GUT
563
FUNCTION
was performed with a silk suture until peripheral venous pressure measurements demonstrated a three- to fourfold stable elevation in venous pressure. In those instances where two veins drained the loop, the smaller of the two veins was ligated and the larger stenosed. The remaining five animals (control group) also underwent exposure of the mesenteric vessels and had venous pressure measurements performed; in these animals, however, venous ligation was not performed. At this time nine silver-silver chloride monopolar serosal electrodes were placed in all animals; four electrodes were placed at lo-cm intervals along the proximal (intact) duodenum and jejunum; two at 10 cm apart on the loop and three at lo-cm intervals on the jejunum distal to the jejunojejunostomy. Leads from all electrodes came together in an abdominal cannula which was brought out onto the abdominal wall through a separate stab incision and sutured in place. This cannula served as the reference electrode (Fig. 2). Motility
and Absorption
Studies
Following a l-week recovery period studies of myoelectrical activity and absorption were performed at 5- to 6day intervals over a 4-week period. Dogs were fasted for 18 hr before each study and rested quietly in a Pavlov sling during recordings. Recordings were first performed in the fasted state until at least three complete interdigestive myoelectrical complexes (ID-
-.-
FIG. 2. Fasting myoelectrical activity-VHT animal. In this example, Phase III of the IDMEC can be seen to propagate from the duodenum through the proximal small intestine across the muscularis bridge into the loop. Note also the propagation across the jejunojejunostomy.
564
JOURNAL
OF SURGICAL
RESEARCH:
MECs) had been recorded. Alternatively, if regular IDMEC activity was not observed, recordings were performed for at least 4 hr while the animal remained fasted. When the third interdigestive myoelectrical complex had traversed the most distal electrode, or 4 hr of fasting recordings had elapsed (in animals who did not exhibit regular IDMECs), a meal of 425 g of beef flavor dog food was administered (constituents, as a percentage of total weight, protein > 8%, fat > 1.5%, fiber > 2%, moisture > 78%, “Sir Johns Choice,” Bern Extrusion, Inc., Bern, KS) and recordings continued for a further hour. The output from all electrodes was transferred to a Beckman R612 recorder (Sensormedics Corp., Anaheim, CA) using alternating current amplifiers (low frequency cutoff, 0.16 Hz, for identification of slow waves; 5.3 Hz for delineation of spike patterns; high frequency cutoff, 30 Hz) and displayed at a paper speed of 0.5 mm/set. Blood samples were drawn from each animal at 2- and 4-week intervals and serum levels of sodium, chloride, potassium, albumin, urea, creatinine, alanine transaminase, aspratate transaminase, lactic dehydrogenase, and bilirubin assayed using standardized automated laboratory techniques. Immediately on completion of each myoelectrical recording an absorption study was performed. Loops were first gently flushed with saline to remove luminal mucus and debris. Continuous perfusion of the loop with an isoosmolar solution comprising 110 mM NaCl, 5 n&f KCl, 30 mM NaH&03, 5% polyethylene glycol (PEG), and 2.34% dextrose in water was then performed via the Thomas cannula at a rate of 70 ml/hr using an I Med infusion pump (Model 960, I Med Corp., San Diego, CA). To determine at what time steady-state conditions had been established, trace amounts of [14C]PEG were added to the perfusion fluid and radioactivity of the effluent was measured in a scintillation counter (Packard Autogamma 5650, Donners Grove, IL). Loop output from the jejunostomy was collected at 5-min intervals and each of the three successive 5-min aliquots, collected following achievement of steady state, was measured and assayed for sodium, chloride, potassium, bicarbonate, and glucose. These assays were performed using standardized, automated laboratory techniques. Terminal
VOL. 48, NO. 6, JUNE
1990
activity. Slow wave frequency at each electrode was determined over three l-min intervals both during Phase I of the interdigestive cycle and during the postprandial period on each study day. The periodicity, propagation velocity, site of proximal origin, extent of propagation, and duration of Phase III of the IDMEC at each electrode site were analyzed using methods described by Code and Marlett [ 171. With respect to postprandial myoelectrical activity, both the time to onset and the total duration of fed-type spike activity (expressed as a percentage of the total postprandial recording time) were defined at each electrode site. Net absorption or secretion of sodium, chloride, potassium, bicarbonate, glucose, and water were defined as the difference between output (calculated from concentration obtained from assay and measured volume) and known input in each 5-min period. To allow for differences in loop length, absorption data were adjusted for loop length and expressed per 20 cm intestine per 5 min. All values are expressed as means + standard deviation of the mean. Statistical comparison between control and VHT animals was performed using Student’s t test for unpaired data. RESULTS
General Condition All animals maintained their body weight throughout the study period, and serum levels of electrolytes, urea, creatinine, and liver enzymes remained within the normal range. Examination of the loop at the final laparotomy did not reveal any evidence of ischemia or infarction. Mesenteric
Vein Pressure
As expected, mesenteric venous pressure was significantly higher in the VHT animals immediately following ligation of the loop vein (mesenteric venous pressure for control vs VHT (mean + SD), 7.5 +- 0.4 vs 29.8 f 1.8, P < 0.05). As evidenced by the pressure readings obtained at the final celiotomy, mesenteric venous hypertension was sustained for the duration of the study period in the VHT group (7.3 -+ 0.2 vs 17.6 f. 6.88, P < 0.01).
Pressure Measurements
On completion of all absorption and motility studies a final celiotomy was performed under general anesthesia. A peripheral vein within the arcade of veins draining the loop and a mesenteric vein remote from the loop were cannulated and venous pressure was recorded. Each animal was then sacrificed by intravenous administration of T61. Data Analysis Recordings of myoelectrical activity were analyzed visually for slow waves (pacemaker potentials) and spike
Motility
Studies
Recordings of fasting and postprandial myoelectrical activity were obtained from all 10 animals. In all animals and at all electrode sites, myoelectrical activity featured omnipresent slow wave activity with intermittent superimposed spike activity. During fasting, spike activity was organized into recurring cycles of the IDMEC. Each phase of the IDMEC could be identified at each electrode site in both control and VHT animals. Postprandial recordings featured intermittent, apparently random spike activity.
JACOBS
ET AL.: MESENTERIC
VENOUS
HYPERTENSION
TABLE Parameters
of Fasting
and Postprandial
Myoelectrical
Duodenum proximal anastamosis
Slow wave frequency (cycles/min) IDMEC incidence (cycles/hr) Phase III duration (min) Time to onset fed pattern (min) Duration fed pattern (W total fed recording time)
AND
GUT
565
FUNCTION
1
Activity
in Control
and Mesenteric-Vein-Ligated
Animals
to Loop
Jejunum distal to anastomosis
Control
VHT”
Control
VHT
Control
VHT
18.8 f 0.7
18.8 f 1.2
17.6 f 1.1
17.8 1?I1.5
13.9 * 0.7
14.3 f 0.4
0.66 + 0.36 5.9 2~2.4
0.58 f 0.31 5.9 3~0.8
0.65 f 0.26 5.7 f 1.4
0.62 iz 0.30 5.4 f 1.0
0.73 zk 0.29 7.5 f 6.7
0.61 f 0.21 7.6 + 2.8
6.5 f 6.5
7.0 f 6.9
14 + 12.7
8.4 f 5.8*
11.6 f 10.9
8.3 + 7.2
89.8 f 9.7
88.3 f 11.2
70.5 + 27.4
86.6 f 9.0*
80.5 f 22.8
86.9 2 11.3
a Mesenteric venous hypertension group. * P < 0.05 control vs mesenteric-vein-ligated.
Table 1 compares a number of parameters of fasting and postprandial myoelectrical activity from electrodes situated on the intact duodenum, the loop, and the jejunum distal to the jejunojejunostomy in the control and VHT animals. Values for slow wave frequency, IDMEC incidence, and Phase III duration were similar in both groups; in particular, values within the loop were almost identical. Both the time to onset and duration of the fed pattern differed within the loop in the VHT animals; the fed pattern began earlier and lasted longer in the VHT animals (time to onset fed pattern (min) mean + SD, control vs VHT, 14.0 + 12.7 vs 8.4 + 5.8, P < 0.05; and duration fed pattern, percentage total fed recording time, 70.5 + 27.4 vs 86.6 + 9.0, P < 0.05). Induction of venous hypertension did not affect propagation of myoelectrical signals, either into or through the loop (velocity of propagation (cm/min) across the bridge, 5.1 + 6.7 vs 4.3 + 2.0, NS; and within the loop, 6.9 f 5.5 vs 5.7 k 3.4, NS). Overall 98.2 +- 7.1% of all IDMECs originating in the duodenum propagated to the loop in the control animals; 100% did so in the VHT animals. Ectopic initiation of IDMECs within the loop was similarly unusual in both groups: 0 and 7.1% of all IDMECs originating in either of the two loop electrodes in control and VHT animals, respectively. Absorption
and bicarbonate absorption were also less in VHT animals, but this difference did not reach statistical significance. DISCUSSION
The principal finding of this study was impairment of absorption of water and a number of electrolytes from jejunal segments exposed to chronic mesenteric venous hypertension. Motor activity, in contrast, was relatively unaffected by induction of venous hypertension, suggesting that the changes in absorption reflected a primary alteration in intestinal mucosal transport. In this model, sustained venous hypertension was achieved and its effects, in isolation from the other common accompaniments of portal hypertension, namely, hepatic dysfunction and portasystemic shunting, could be defined. Our findings in the control animals confirm that retention of a bridge of tunica muscularis will ensure normal propagation of myoelectrical signals to a luminally isolated loop [ 14-161. Previous studies of intestinal absorption in patients with cirrhosis and portal hypertension have provided
Studies
Absorption studies were completed in 9 of the 10 animals (4 controls, 5 VHT). The Thomas cannula became dislodged in the fifth control animal, preventing completion of absorption studies. Table 2 compares results of absorption studies in the two groups. Net absorption of sodium, chloride, glucose, and water was significantly reduced in the VHT animals; values in the VHT animals were between 13 and 22% lower than those in controls. Average values for notassium
TABLE Control
Na bed K (meg) Cl bed CO2 beq) glucose (mg) &O (4
423 8.3 336 78 8547 2.6
+ f + f f f
VHT
109” 4.9 88 37 2139 1.0
’ Mean f SEM, absorption/20
2
515 10.2 412 92 9827 3.3
cm loop/5 min.
2 155 f 7.8 31 122 2 33 f 2046 f 1.3
P
566
JOURNAL
OF SURGICAL
RESEARCH:
variable results [2,18-211. Thus, while some studies demonstrated decreased D-xylose absorption in cirrhosis and portal hypertension 14, 19-211 and D-xylose absorption has been shown to increase in portal hypertensive patients following a nonselective shunt operation [ 41, Norman and colleagues [18] failed to demonstrate any deficit in absorption of D-xylose, water, or electrolytes in their patients with chronic liver disease and elevated hepatic vein wedge pressures. Interpretation of these human studies has been complicated by the presence of factors such as alcoholism, malnutrition, and multiorgan system dysfunction. Available data from animal studies do not permit firm conclusions either. Thus, while Grandison and colleagues [6] reported a reduction in ammonia, water, sodium, and chloride absorption from the dog colon following acute elevation of portal pressure, the relevance of these results to chronic intestinal VHT is uncertain. Sarfeh and colleagues induced portal hypertension in the rat by performing both side-to-side mesenteric arteriovenous anastomosis and portal vein banding [ 71. While resulting in significant portal hypertension, this model maintains portal vein flow such that hepatic architecture and excretory function are preserved. Absorptive studies in this model demonstrated normal absorption of the amino acids valine and tryptophan but impaired absorption of vitamin Ds. In concluding that venous hypertension alone resulted in the selective impairment of intestinal absorption of some nutrients, these authors fail to take account of the fact that the portasystemic shunting which is a basic feature of their model may also have an impact on gut absorptive function. Diversion of enteric hormones and peptides from the liver may result in significant changes in their metabolism with consequent alterations in their concentration in the portal and systemic circulations. Thus, for example, increased leveis of glucagon in the portal system have been proposed to play a major role in the maintenance of splanchnic hyperemia in portal hypertension [12, 13, 22, 231 and may also explain the delayed transit observed in the portal hypertensive rat [ll]. Our model permits isolation of the effects of chronic intestinal venous hypertension from those of either hepatic dysfunction or portasystemic shunting. The observed differences in intestinal function between the VHT and control loops can be attributed entirely, therefore, to the effects of venous congestion. In the jejunum, glucose and bicarbonate are absorbed primarily by active transport mechanisms; whereas, sodium, potassium, chloride, and water transport takes place largely by passive diffusion 1241. Therefore, the results in this study demonstrating significantly less absorption of glucose but not of bicarbonate and less absorption of sodium, chloride, and water but not of potassium seem to indicate a lack of selectivity of the effect of VHT on passive or active transport processes. If VHT was exerting its effect through increased hydrostatic pressure alone, one would have ex-
VOL. 48, NO. 6, JUNE
1990
petted a greater impairment of passive absorption [5,25]. Since such selectivity was absent in this study, we suggest that increased hydrostatic pressure alone cannot explain the effect of VHT on absorptive function. Previous studies in the dog colon showed that total colonic blood flow decreased in response to acute increases in venous pressure and led to the proposal that this decreased flow could result in impaired absorption [6]. Increased mesenteric venous pressure also affects small intestinal blood flow. In the rat, acute VHT results in a diversion of flow from mucosal to muscularis capillaries [26]. Such redistribution could well alter mucosal absorptive function. Similar diversion of blood away from the mucosa has been demonstrated in the stomach of the portal hypertensive rat [27] and has been associated with impaired mucosal function as evidenced by decreased hydrogen ion secretion and impaired mucosal resistance to various ulcerogenic agents [28-301. These changes in intestinal absorption do not appear to be due to altered motor activity. Thus, the VHT animals demonstrated normal patterns of fasting and postprandial myoelectrical activity at all sites including those within the loop. Control and VHT animals differed only in the time to onset and duration of fed myoelectrical activity. Given that postprandial motor activity is thought to promote mixing and to increase mucosal contact time [31], these differences would be expected to promote rather than to inhibit absorption. The findings in this study also suggest that venous congestion alone cannot explain the transit delay we described previously in the portal hypertensive rat [ll] and that some other feature of that model such as portasystemic shunting with resultant altered metabolism of hormones such as glucagon, which can inhibit motility, [32] may be responsible. REFERENCES 1.
Enquist, I. F., Golding, M. R., Aiello, R. G., Fierst, S. M., and Solomon, N. A. The effect of portal hypertension on intestinal absorption. Surg. Gynecol. O&et. 120: 87, 1965.
2. Sun, D. C. H., Albacete, R. A., and Chen, J. K. Malabsorption studies in cirrhosis of the liver. Arch. Intern. Med. 119: 567,1967. 3. Losowsky, M. S., and Walker, B. Liver disease and malabsorption. Gmtroenterology 56: 589,1969. 4. Rikkers, L. F. Portal hemodynamics, intestinal absorption and post-shunt
encephalopathy.
Surgery 94: 126,1983.
5. Shields, R., and Code, C. Effect of increased portal pressure on absorption of water and sodium in the ileum of dogs. Amer. J. Physiol. 200: 775, 1961. 6. Grandison, A. S., Harrison, I. D., and Shields, R. Effects of wide variations in portal pressure on mesenteric blood flow and absorption from the canine colon. Gut 21: 475, 1980. 7. Sarfeh, I. J., Aaronson, S., Lombino, D., Rypins, E. $., Mason, G. R., Dadufalza, L., and Hollander, D. Selective impairment of nutrient absorption from intestine with chronic venous hypertension. Surgery 99: 166, 1986. role in gut a. Quigley, E. M. M. Small intestinal motor activity-Its homeostasis in health and disease. Q. J. Med. 66: 799, 1987.
JACOBS 9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19.
20.
ET AL.: MESENTERIC
VENOUS
Reilly, J. A., Forst, C. F., Nielson, E., Quigley, E. M. M., and Rikkers, L. F. The effect of portal hypertension on gastric emptying of liquids. Curr. Surg. 46: 33, 1989. Reilly, J. A., Jr., Markin, R. S., Quigley, E. M. M., and Rikkers, L. F. Accelerated gastric emptying in the portal hypertensive rat. Surg. Forum. 39: 136, 1988. Reilly, J. A., Jr., Quigley, E. M. M. and Rikkers, L. F. Small intestinal transit in the portal hypertensive rat. Submitted for publication. Sherwin, R., Joshi, P., Hendler, R., Felix, P., and Corm; H. 0. Hyperglucagonemia in Laennec’s cirrhosis-The role of portalsystemic shunting. N. Engl. J. Med. 290: 239, 1974. Benoit, J. N., Barrowman, J. A., Harper, S. L., Kvietys, P. R., and Granger, D. N. Role of humoral factors in the intestinal hyperemia associated with chronic portal hypertension. Amer. J. Physiol. 247: G486, 1984. Collin, J., Kelly, K. A., and Phillips, S. F. Enhancement of absorption from the intact and transected canine small intestine by electrical pacing. Gastroenterology 76: 1422, 1979. Sarr, M. G., and Kelly, K. A. Patterns of movement of liquids and solids through canine jejunum. Amer. J. Physiol. 239: G497,1980. Quigley, E. M. M., Phillips, S. F., Dent, J., and Taylor, B. M. Myoelectrical activity and intraluminal pressure of the canine ileocolonic sphincter. Gastroenterology 85: 1054, 1983. Code, C. F., and Marlett, J. A. The interdigestive myoelectrical complex of the stomach and small bowel of dogs. J. Physiol. 296: 289, 1978. Norman, D. A., Atkins, J. M., Seelig, L. L., Jr., Gomez-Sanchez, C., and Krejs, G. J. Water and electrolyte movement and mucosal morphology in the jejunum of patients with portal hypertension. Gastroenterology 79: 707, 1980. Baranoa, E., Orrego, H., Fernandez, O., Amenabar, E., Maldonado, E., Tag, F., and Salinas, A. Absorptive function of the small intestine in liver cirrhosis. Amer. J. Dig. Dis. 7: 318, 1962. Marin, G. A., Clark, M. L., and Senior, J. R. Distribution of Dxylose in sequestered fluid resulting in false-positive tests for malabsorption. Ann. Intern. Med. 69: 1155, 1968.
HYPERTENSION
AND
GUT
FUNCTION
567
21.
Marin, G. A., Clark, M. L.,,and Senior, J. R. Studies of malabsorption occurring in patients with Laennec’s cirrhosis. Gastroenterology 56: 727, 1969.
22.
Vorobioff, J., Bredfeldt, J. E., and Groszmann, R. H. Hyperdynamic circulation in portal-hypertensive rat model: A primary factor for maintenance of chronic portal hypertension. Amer. J. Physiol. 244: G52, 1983.
23.
Benoit, J. N., and Granger, D. M. Splanchinc hemodynamics chronic portal hypertension. Semin. Liver Dis. 6: 287, 1986.
24.
Powell, D. W. Intestinal water and electrolyte transport. In L. R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 2nd ed. New York: Raven Press, 1987. Pp. 1267-1305.
25.
Yablonski, M. E., and Lifson, N. Mechanisms of production of intestinal secretion by elevated venous pressure. J. Clin. Inuest. 57: 904, 1976.
26.
Davis, M. J., and Gore, R. W. Capillary pressure in rat intestinal muscle and mucosal villi during venous pressure elevation. Amer. J. Physiol. 244: H174-H187, 1985.
27.
Sarfeh, I. J., and Tarnawski, A. Gastric mucosal vasculopathy portal hypertension. Gastroenterology 93: 1129, 1987.
28.
Sarfeh, I. J., Tarnawski, A., Malki, A., Mason, G. R., Mach, T., and Ivey, K. J. Portal hypertension and gastric mucosal injury in rats: Effects of alcohol. Gastroenterology 84: 987, 1983.
29.
Sarfeh, I. J., Tarnawski, A., Maeda, R., Raymont, K., Mason, G. R., and Ivey, K. J. The gastric mucosa in portal hypertension: Effects of topical bile acid. Stand. J. Gastroenterol. 12(Suppl.): 189, 1984.
30.
Pique, J. M., Leung, F. W., Kitamora, T., Sarfeh, I. J., Tarnawski, A., and Guth, P. H. Gastric mucosal blood flow and acid secretion in portal hypertensive rats. Gastroenterology 95: 727, 733, 1988.
31.
Schemann, M., and Ehrlein, H-J. Postprandial patterns of canine jejunal motility and transit of luminal content. Gastroenterology 90: 991, 1986.
32.
Dotevall, G., and Kock, N. G. The effect of glucagon on intestinal motility in man. Gastroenterology 45: 364, 1963.
in
in
OF SURGICAL
RESEARCH
48,562-567
(1990)
The Effect of Mesenteric Venous Hypertension on Gut Motility and Absorption DONALD L. JACOBS, M.D., JOHN LOF, M.S., EAMONN M. M. QUIGI,EY, M.D., ALI D. SPANTA, M.D., AND LAYTON F. RIKKERS, M.D. Ilepartments
of Surgery and Internal
Medicine,
University
of Nebraska Medical Center, 42nd Street and Dewey Avenue, Omaha, Nebraska 68105
Presentedat the Annual Meeting of the Association
for Academic Surgery, Louisville,
November
15-18, 1989.
may also be operative. Altered gut motility, for example, could also influence intestinal absorption [8]. We have previously demonstrated alterations in both gastric emptying [9, lo] and small intestinal transit [ll] in portal hypertension. The pathogenesis of these motor changes and their contribution to impaired absorptive function in portal hypertension have not been defined. In particular, it has not been established whether these changes in small intestinal physiology are entirely consequent upon passive venous congestion or whether they reflect ‘the effects of altered levels of enteric hormones and neuropeptides as a result of portasystemic shunting [ 12, 131. Our aim, therefore, was to assess the effects of chronic mesenteric venous hypertension, in the absence of hepatic dysfunction or portasystemic shunting, on intestinal motility and absorption.
The effects of portal hypertension on gut function may be mediated by venous congestion and altered circulating levels of enteric hormones and neuropeptides. We designed this study to determine the effects of chronic intestinal venous hypertension (VHT), in isolation, on gut motility and absorption. In 10 dogs, a 20- to 26-cm loop of jejunum was isolated from the fecal stream, but myoneural continuity was maintained with the proximal bowel by a seromuscular bridge. In 5 dogs, VHT was created in the loop by a fixed stenosis of its venous drainage; a sham procedure was performed in a further 5 animals. Serosal monopolar electrodes were placed in all animals. Absorptive function and myoelectrical activity were studied over a I-week period. Venous hypertension was achieved and sustained in the VHT animals; loop vein pressures for VHT vs control in cm Hz0 (means f SEM) are: initial-29.8 + 1.8 vs 7.5 + 0.4 (P < O.Ol), and at 4 weeko17.6 f 6.88 vs 7.3 f 0.2 (P < 0.01). Absorption of Na+, Cl , glucose, and water was impaired in VHT loops. Normal patterns of fasting and postprandial myoelectrical activity were preserved in the VHT animals. We conclude that chronic VHT, in the absence of portosystemic shunting, results in impaired absorption of water, glucose, and electrolytes without any change in in& 1990 Academic Press. Inc. testinal motility.
METHODS Preparation
of Animals
Ten adult female mongrel dogs, weighing between 15 and 20 kg, underwent celiotomy under pentobarbital sodium anesthesia. Beginning at a point 30 cm distal to the ligament of Treitz, a 20- to 25-cm segment of jejunum, drained by either a single or at most two veins at the base of the mesentery, was identified. This segment was fashioned into a loop which, though isolated from the lumen of the gastrointestinal tract, maintained neuromuscular continuity with the proximal small intestine via a bridge of tunica muscularis between the proximal end of the loop and the intact jejunum [14-161. A Thomas cannula was placed in the loop near the proximal end and brought out to the abdominal wall through a separate incision. The distal end of the loop was brought out to the abdominal wall as an end jejunostomy (Fig. 1). Two weeks later all animals underwent a second celiotomy under general anesthesia. The vascular supply to the loop was identified and skeletonized at the base of the mesentery. To permit intraoperative recording of venous pressure, a needle connected to a water-filled but non-
INTRODUCTION The etiology of malnutrition associated with advanced cirrhosis and portal hypertension is multifactorial and thought to include such factors as poor nutritional intake, impaired hepatic metabolism of absorbed nutrients, and intestinal malabsorption [l-3]. With respect to the latter, it has been suggested that intestinal venous congestion consequent on portal hypertension may directly alter absorptive function [3, 41. Indeed, previous studies, in animals, have demonstrated that acute intestinal venous hypertension causes a decrease in the absorption of water, sodium, chloride, and ammonia [5, 61. Furthermore, chronic intestinal venous hypertension results in malabsorption of vitamin D:, in the rat [7]. Other mechanisms oozz-4804/90 $1.50 Ckapyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Kentucky,
562
JACOBS
ET AL.: MESENTERIC
VENOUS
FIG. 1. Experimental model: bridged loop transection. A 20- to 25 cm loop that is luminally isolated yet maintains myoneural continuity with the proximal intestine via the bridge of tunica muscularis (inset). Note the location of the serosal monopolar electrodes.
perfused catheter was inserted into a peripheral vein in the venous arcade of the loop. The catheter was, in turn, connected to a saline-filled manometer. In five dogs mesenteric venous hypertension was created in the loop (VHT group) as follows: a calibrated stenosis of the previously defined, principal draining vein
HYPERTENSION
AND
GUT
563
FUNCTION
was performed with a silk suture until peripheral venous pressure measurements demonstrated a three- to fourfold stable elevation in venous pressure. In those instances where two veins drained the loop, the smaller of the two veins was ligated and the larger stenosed. The remaining five animals (control group) also underwent exposure of the mesenteric vessels and had venous pressure measurements performed; in these animals, however, venous ligation was not performed. At this time nine silver-silver chloride monopolar serosal electrodes were placed in all animals; four electrodes were placed at lo-cm intervals along the proximal (intact) duodenum and jejunum; two at 10 cm apart on the loop and three at lo-cm intervals on the jejunum distal to the jejunojejunostomy. Leads from all electrodes came together in an abdominal cannula which was brought out onto the abdominal wall through a separate stab incision and sutured in place. This cannula served as the reference electrode (Fig. 2). Motility
and Absorption
Studies
Following a l-week recovery period studies of myoelectrical activity and absorption were performed at 5- to 6day intervals over a 4-week period. Dogs were fasted for 18 hr before each study and rested quietly in a Pavlov sling during recordings. Recordings were first performed in the fasted state until at least three complete interdigestive myoelectrical complexes (ID-
-.-
FIG. 2. Fasting myoelectrical activity-VHT animal. In this example, Phase III of the IDMEC can be seen to propagate from the duodenum through the proximal small intestine across the muscularis bridge into the loop. Note also the propagation across the jejunojejunostomy.
564
JOURNAL
OF SURGICAL
RESEARCH:
MECs) had been recorded. Alternatively, if regular IDMEC activity was not observed, recordings were performed for at least 4 hr while the animal remained fasted. When the third interdigestive myoelectrical complex had traversed the most distal electrode, or 4 hr of fasting recordings had elapsed (in animals who did not exhibit regular IDMECs), a meal of 425 g of beef flavor dog food was administered (constituents, as a percentage of total weight, protein > 8%, fat > 1.5%, fiber > 2%, moisture > 78%, “Sir Johns Choice,” Bern Extrusion, Inc., Bern, KS) and recordings continued for a further hour. The output from all electrodes was transferred to a Beckman R612 recorder (Sensormedics Corp., Anaheim, CA) using alternating current amplifiers (low frequency cutoff, 0.16 Hz, for identification of slow waves; 5.3 Hz for delineation of spike patterns; high frequency cutoff, 30 Hz) and displayed at a paper speed of 0.5 mm/set. Blood samples were drawn from each animal at 2- and 4-week intervals and serum levels of sodium, chloride, potassium, albumin, urea, creatinine, alanine transaminase, aspratate transaminase, lactic dehydrogenase, and bilirubin assayed using standardized automated laboratory techniques. Immediately on completion of each myoelectrical recording an absorption study was performed. Loops were first gently flushed with saline to remove luminal mucus and debris. Continuous perfusion of the loop with an isoosmolar solution comprising 110 mM NaCl, 5 n&f KCl, 30 mM NaH&03, 5% polyethylene glycol (PEG), and 2.34% dextrose in water was then performed via the Thomas cannula at a rate of 70 ml/hr using an I Med infusion pump (Model 960, I Med Corp., San Diego, CA). To determine at what time steady-state conditions had been established, trace amounts of [14C]PEG were added to the perfusion fluid and radioactivity of the effluent was measured in a scintillation counter (Packard Autogamma 5650, Donners Grove, IL). Loop output from the jejunostomy was collected at 5-min intervals and each of the three successive 5-min aliquots, collected following achievement of steady state, was measured and assayed for sodium, chloride, potassium, bicarbonate, and glucose. These assays were performed using standardized, automated laboratory techniques. Terminal
VOL. 48, NO. 6, JUNE
1990
activity. Slow wave frequency at each electrode was determined over three l-min intervals both during Phase I of the interdigestive cycle and during the postprandial period on each study day. The periodicity, propagation velocity, site of proximal origin, extent of propagation, and duration of Phase III of the IDMEC at each electrode site were analyzed using methods described by Code and Marlett [ 171. With respect to postprandial myoelectrical activity, both the time to onset and the total duration of fed-type spike activity (expressed as a percentage of the total postprandial recording time) were defined at each electrode site. Net absorption or secretion of sodium, chloride, potassium, bicarbonate, glucose, and water were defined as the difference between output (calculated from concentration obtained from assay and measured volume) and known input in each 5-min period. To allow for differences in loop length, absorption data were adjusted for loop length and expressed per 20 cm intestine per 5 min. All values are expressed as means + standard deviation of the mean. Statistical comparison between control and VHT animals was performed using Student’s t test for unpaired data. RESULTS
General Condition All animals maintained their body weight throughout the study period, and serum levels of electrolytes, urea, creatinine, and liver enzymes remained within the normal range. Examination of the loop at the final laparotomy did not reveal any evidence of ischemia or infarction. Mesenteric
Vein Pressure
As expected, mesenteric venous pressure was significantly higher in the VHT animals immediately following ligation of the loop vein (mesenteric venous pressure for control vs VHT (mean + SD), 7.5 +- 0.4 vs 29.8 f 1.8, P < 0.05). As evidenced by the pressure readings obtained at the final celiotomy, mesenteric venous hypertension was sustained for the duration of the study period in the VHT group (7.3 -+ 0.2 vs 17.6 f. 6.88, P < 0.01).
Pressure Measurements
On completion of all absorption and motility studies a final celiotomy was performed under general anesthesia. A peripheral vein within the arcade of veins draining the loop and a mesenteric vein remote from the loop were cannulated and venous pressure was recorded. Each animal was then sacrificed by intravenous administration of T61. Data Analysis Recordings of myoelectrical activity were analyzed visually for slow waves (pacemaker potentials) and spike
Motility
Studies
Recordings of fasting and postprandial myoelectrical activity were obtained from all 10 animals. In all animals and at all electrode sites, myoelectrical activity featured omnipresent slow wave activity with intermittent superimposed spike activity. During fasting, spike activity was organized into recurring cycles of the IDMEC. Each phase of the IDMEC could be identified at each electrode site in both control and VHT animals. Postprandial recordings featured intermittent, apparently random spike activity.
JACOBS
ET AL.: MESENTERIC
VENOUS
HYPERTENSION
TABLE Parameters
of Fasting
and Postprandial
Myoelectrical
Duodenum proximal anastamosis
Slow wave frequency (cycles/min) IDMEC incidence (cycles/hr) Phase III duration (min) Time to onset fed pattern (min) Duration fed pattern (W total fed recording time)
AND
GUT
565
FUNCTION
1
Activity
in Control
and Mesenteric-Vein-Ligated
Animals
to Loop
Jejunum distal to anastomosis
Control
VHT”
Control
VHT
Control
VHT
18.8 f 0.7
18.8 f 1.2
17.6 f 1.1
17.8 1?I1.5
13.9 * 0.7
14.3 f 0.4
0.66 + 0.36 5.9 2~2.4
0.58 f 0.31 5.9 3~0.8
0.65 f 0.26 5.7 f 1.4
0.62 iz 0.30 5.4 f 1.0
0.73 zk 0.29 7.5 f 6.7
0.61 f 0.21 7.6 + 2.8
6.5 f 6.5
7.0 f 6.9
14 + 12.7
8.4 f 5.8*
11.6 f 10.9
8.3 + 7.2
89.8 f 9.7
88.3 f 11.2
70.5 + 27.4
86.6 f 9.0*
80.5 f 22.8
86.9 2 11.3
a Mesenteric venous hypertension group. * P < 0.05 control vs mesenteric-vein-ligated.
Table 1 compares a number of parameters of fasting and postprandial myoelectrical activity from electrodes situated on the intact duodenum, the loop, and the jejunum distal to the jejunojejunostomy in the control and VHT animals. Values for slow wave frequency, IDMEC incidence, and Phase III duration were similar in both groups; in particular, values within the loop were almost identical. Both the time to onset and duration of the fed pattern differed within the loop in the VHT animals; the fed pattern began earlier and lasted longer in the VHT animals (time to onset fed pattern (min) mean + SD, control vs VHT, 14.0 + 12.7 vs 8.4 + 5.8, P < 0.05; and duration fed pattern, percentage total fed recording time, 70.5 + 27.4 vs 86.6 + 9.0, P < 0.05). Induction of venous hypertension did not affect propagation of myoelectrical signals, either into or through the loop (velocity of propagation (cm/min) across the bridge, 5.1 + 6.7 vs 4.3 + 2.0, NS; and within the loop, 6.9 f 5.5 vs 5.7 k 3.4, NS). Overall 98.2 +- 7.1% of all IDMECs originating in the duodenum propagated to the loop in the control animals; 100% did so in the VHT animals. Ectopic initiation of IDMECs within the loop was similarly unusual in both groups: 0 and 7.1% of all IDMECs originating in either of the two loop electrodes in control and VHT animals, respectively. Absorption
and bicarbonate absorption were also less in VHT animals, but this difference did not reach statistical significance. DISCUSSION
The principal finding of this study was impairment of absorption of water and a number of electrolytes from jejunal segments exposed to chronic mesenteric venous hypertension. Motor activity, in contrast, was relatively unaffected by induction of venous hypertension, suggesting that the changes in absorption reflected a primary alteration in intestinal mucosal transport. In this model, sustained venous hypertension was achieved and its effects, in isolation from the other common accompaniments of portal hypertension, namely, hepatic dysfunction and portasystemic shunting, could be defined. Our findings in the control animals confirm that retention of a bridge of tunica muscularis will ensure normal propagation of myoelectrical signals to a luminally isolated loop [ 14-161. Previous studies of intestinal absorption in patients with cirrhosis and portal hypertension have provided
Studies
Absorption studies were completed in 9 of the 10 animals (4 controls, 5 VHT). The Thomas cannula became dislodged in the fifth control animal, preventing completion of absorption studies. Table 2 compares results of absorption studies in the two groups. Net absorption of sodium, chloride, glucose, and water was significantly reduced in the VHT animals; values in the VHT animals were between 13 and 22% lower than those in controls. Average values for notassium
TABLE Control
Na bed K (meg) Cl bed CO2 beq) glucose (mg) &O (4
423 8.3 336 78 8547 2.6
+ f + f f f
VHT
109” 4.9 88 37 2139 1.0
’ Mean f SEM, absorption/20
2
515 10.2 412 92 9827 3.3
cm loop/5 min.
2 155 f 7.8 31 122 2 33 f 2046 f 1.3
P
566
JOURNAL
OF SURGICAL
RESEARCH:
variable results [2,18-211. Thus, while some studies demonstrated decreased D-xylose absorption in cirrhosis and portal hypertension 14, 19-211 and D-xylose absorption has been shown to increase in portal hypertensive patients following a nonselective shunt operation [ 41, Norman and colleagues [18] failed to demonstrate any deficit in absorption of D-xylose, water, or electrolytes in their patients with chronic liver disease and elevated hepatic vein wedge pressures. Interpretation of these human studies has been complicated by the presence of factors such as alcoholism, malnutrition, and multiorgan system dysfunction. Available data from animal studies do not permit firm conclusions either. Thus, while Grandison and colleagues [6] reported a reduction in ammonia, water, sodium, and chloride absorption from the dog colon following acute elevation of portal pressure, the relevance of these results to chronic intestinal VHT is uncertain. Sarfeh and colleagues induced portal hypertension in the rat by performing both side-to-side mesenteric arteriovenous anastomosis and portal vein banding [ 71. While resulting in significant portal hypertension, this model maintains portal vein flow such that hepatic architecture and excretory function are preserved. Absorptive studies in this model demonstrated normal absorption of the amino acids valine and tryptophan but impaired absorption of vitamin Ds. In concluding that venous hypertension alone resulted in the selective impairment of intestinal absorption of some nutrients, these authors fail to take account of the fact that the portasystemic shunting which is a basic feature of their model may also have an impact on gut absorptive function. Diversion of enteric hormones and peptides from the liver may result in significant changes in their metabolism with consequent alterations in their concentration in the portal and systemic circulations. Thus, for example, increased leveis of glucagon in the portal system have been proposed to play a major role in the maintenance of splanchnic hyperemia in portal hypertension [12, 13, 22, 231 and may also explain the delayed transit observed in the portal hypertensive rat [ll]. Our model permits isolation of the effects of chronic intestinal venous hypertension from those of either hepatic dysfunction or portasystemic shunting. The observed differences in intestinal function between the VHT and control loops can be attributed entirely, therefore, to the effects of venous congestion. In the jejunum, glucose and bicarbonate are absorbed primarily by active transport mechanisms; whereas, sodium, potassium, chloride, and water transport takes place largely by passive diffusion 1241. Therefore, the results in this study demonstrating significantly less absorption of glucose but not of bicarbonate and less absorption of sodium, chloride, and water but not of potassium seem to indicate a lack of selectivity of the effect of VHT on passive or active transport processes. If VHT was exerting its effect through increased hydrostatic pressure alone, one would have ex-
VOL. 48, NO. 6, JUNE
1990
petted a greater impairment of passive absorption [5,25]. Since such selectivity was absent in this study, we suggest that increased hydrostatic pressure alone cannot explain the effect of VHT on absorptive function. Previous studies in the dog colon showed that total colonic blood flow decreased in response to acute increases in venous pressure and led to the proposal that this decreased flow could result in impaired absorption [6]. Increased mesenteric venous pressure also affects small intestinal blood flow. In the rat, acute VHT results in a diversion of flow from mucosal to muscularis capillaries [26]. Such redistribution could well alter mucosal absorptive function. Similar diversion of blood away from the mucosa has been demonstrated in the stomach of the portal hypertensive rat [27] and has been associated with impaired mucosal function as evidenced by decreased hydrogen ion secretion and impaired mucosal resistance to various ulcerogenic agents [28-301. These changes in intestinal absorption do not appear to be due to altered motor activity. Thus, the VHT animals demonstrated normal patterns of fasting and postprandial myoelectrical activity at all sites including those within the loop. Control and VHT animals differed only in the time to onset and duration of fed myoelectrical activity. Given that postprandial motor activity is thought to promote mixing and to increase mucosal contact time [31], these differences would be expected to promote rather than to inhibit absorption. The findings in this study also suggest that venous congestion alone cannot explain the transit delay we described previously in the portal hypertensive rat [ll] and that some other feature of that model such as portasystemic shunting with resultant altered metabolism of hormones such as glucagon, which can inhibit motility, [32] may be responsible. REFERENCES 1.
Enquist, I. F., Golding, M. R., Aiello, R. G., Fierst, S. M., and Solomon, N. A. The effect of portal hypertension on intestinal absorption. Surg. Gynecol. O&et. 120: 87, 1965.
2. Sun, D. C. H., Albacete, R. A., and Chen, J. K. Malabsorption studies in cirrhosis of the liver. Arch. Intern. Med. 119: 567,1967. 3. Losowsky, M. S., and Walker, B. Liver disease and malabsorption. Gmtroenterology 56: 589,1969. 4. Rikkers, L. F. Portal hemodynamics, intestinal absorption and post-shunt
encephalopathy.
Surgery 94: 126,1983.
5. Shields, R., and Code, C. Effect of increased portal pressure on absorption of water and sodium in the ileum of dogs. Amer. J. Physiol. 200: 775, 1961. 6. Grandison, A. S., Harrison, I. D., and Shields, R. Effects of wide variations in portal pressure on mesenteric blood flow and absorption from the canine colon. Gut 21: 475, 1980. 7. Sarfeh, I. J., Aaronson, S., Lombino, D., Rypins, E. $., Mason, G. R., Dadufalza, L., and Hollander, D. Selective impairment of nutrient absorption from intestine with chronic venous hypertension. Surgery 99: 166, 1986. role in gut a. Quigley, E. M. M. Small intestinal motor activity-Its homeostasis in health and disease. Q. J. Med. 66: 799, 1987.
JACOBS 9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19.
20.
ET AL.: MESENTERIC
VENOUS
Reilly, J. A., Forst, C. F., Nielson, E., Quigley, E. M. M., and Rikkers, L. F. The effect of portal hypertension on gastric emptying of liquids. Curr. Surg. 46: 33, 1989. Reilly, J. A., Jr., Markin, R. S., Quigley, E. M. M., and Rikkers, L. F. Accelerated gastric emptying in the portal hypertensive rat. Surg. Forum. 39: 136, 1988. Reilly, J. A., Jr., Quigley, E. M. M. and Rikkers, L. F. Small intestinal transit in the portal hypertensive rat. Submitted for publication. Sherwin, R., Joshi, P., Hendler, R., Felix, P., and Corm; H. 0. Hyperglucagonemia in Laennec’s cirrhosis-The role of portalsystemic shunting. N. Engl. J. Med. 290: 239, 1974. Benoit, J. N., Barrowman, J. A., Harper, S. L., Kvietys, P. R., and Granger, D. N. Role of humoral factors in the intestinal hyperemia associated with chronic portal hypertension. Amer. J. Physiol. 247: G486, 1984. Collin, J., Kelly, K. A., and Phillips, S. F. Enhancement of absorption from the intact and transected canine small intestine by electrical pacing. Gastroenterology 76: 1422, 1979. Sarr, M. G., and Kelly, K. A. Patterns of movement of liquids and solids through canine jejunum. Amer. J. Physiol. 239: G497,1980. Quigley, E. M. M., Phillips, S. F., Dent, J., and Taylor, B. M. Myoelectrical activity and intraluminal pressure of the canine ileocolonic sphincter. Gastroenterology 85: 1054, 1983. Code, C. F., and Marlett, J. A. The interdigestive myoelectrical complex of the stomach and small bowel of dogs. J. Physiol. 296: 289, 1978. Norman, D. A., Atkins, J. M., Seelig, L. L., Jr., Gomez-Sanchez, C., and Krejs, G. J. Water and electrolyte movement and mucosal morphology in the jejunum of patients with portal hypertension. Gastroenterology 79: 707, 1980. Baranoa, E., Orrego, H., Fernandez, O., Amenabar, E., Maldonado, E., Tag, F., and Salinas, A. Absorptive function of the small intestine in liver cirrhosis. Amer. J. Dig. Dis. 7: 318, 1962. Marin, G. A., Clark, M. L., and Senior, J. R. Distribution of Dxylose in sequestered fluid resulting in false-positive tests for malabsorption. Ann. Intern. Med. 69: 1155, 1968.
HYPERTENSION
AND
GUT
FUNCTION
567
21.
Marin, G. A., Clark, M. L.,,and Senior, J. R. Studies of malabsorption occurring in patients with Laennec’s cirrhosis. Gastroenterology 56: 727, 1969.
22.
Vorobioff, J., Bredfeldt, J. E., and Groszmann, R. H. Hyperdynamic circulation in portal-hypertensive rat model: A primary factor for maintenance of chronic portal hypertension. Amer. J. Physiol. 244: G52, 1983.
23.
Benoit, J. N., and Granger, D. M. Splanchinc hemodynamics chronic portal hypertension. Semin. Liver Dis. 6: 287, 1986.
24.
Powell, D. W. Intestinal water and electrolyte transport. In L. R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 2nd ed. New York: Raven Press, 1987. Pp. 1267-1305.
25.
Yablonski, M. E., and Lifson, N. Mechanisms of production of intestinal secretion by elevated venous pressure. J. Clin. Inuest. 57: 904, 1976.
26.
Davis, M. J., and Gore, R. W. Capillary pressure in rat intestinal muscle and mucosal villi during venous pressure elevation. Amer. J. Physiol. 244: H174-H187, 1985.
27.
Sarfeh, I. J., and Tarnawski, A. Gastric mucosal vasculopathy portal hypertension. Gastroenterology 93: 1129, 1987.
28.
Sarfeh, I. J., Tarnawski, A., Malki, A., Mason, G. R., Mach, T., and Ivey, K. J. Portal hypertension and gastric mucosal injury in rats: Effects of alcohol. Gastroenterology 84: 987, 1983.
29.
Sarfeh, I. J., Tarnawski, A., Maeda, R., Raymont, K., Mason, G. R., and Ivey, K. J. The gastric mucosa in portal hypertension: Effects of topical bile acid. Stand. J. Gastroenterol. 12(Suppl.): 189, 1984.
30.
Pique, J. M., Leung, F. W., Kitamora, T., Sarfeh, I. J., Tarnawski, A., and Guth, P. H. Gastric mucosal blood flow and acid secretion in portal hypertensive rats. Gastroenterology 95: 727, 733, 1988.
31.
Schemann, M., and Ehrlein, H-J. Postprandial patterns of canine jejunal motility and transit of luminal content. Gastroenterology 90: 991, 1986.
32.
Dotevall, G., and Kock, N. G. The effect of glucagon on intestinal motility in man. Gastroenterology 45: 364, 1963.
in
in