The mechanisms and prevention of intravascular fluid loss after occlusion of the supraceliac aorta in dogs

The mechanisms and prevention of intravascular fluid loss after occlusion of the: supraceliac aorta in dogs Russell G. Bourchier, MB, ChB, Peter Gloviczki, M D , M a r k V. Larson, M D , Qing-hua W u , M D , J o h n W. Hallett, Jr., M D , David A. Ahlquist, M D , and Peter C. Pairolero, M D Rochester, Minn. Mechanisms of intravascular fluid depletion after temporary occlusion of the supraceliac aorta were investigated in a canine model. During ischemia and reperfiasion, hemodynamic parameters, superior mesenteric artery flow, intestinal mucosal perfusion, and mucosal permeability were monitored. After 12 hours of reperfusion, the volumes of intravenous electrolyte fluid required to maintain hemodynamic stability and fluid lost into the gastrointestinal tract and peritoneal cavity were measured. The distribution of total body water was analyzed by use of radionuclide dilution techniques. Group A animals underwent laparotomy only, group B had the supraceliac aorta occluded for 45 minutes, group C had superoxide dismutase administered after 45 minutes of aortic occlusion, and group D animals were exposed to mild hypothermia during a similar ischemia and reperfusion period. No significant difference was found in mean superior mesenteric artery flow or mucosal perfusion during ischemia among groups B, C, and D. During reperfusion superior mesenteric artery flow returned to values similar to control in all groups. Aortic occlusion increased mucosal permeability most significantly in group B (p < 0.01). Mean intravenous fluid requirements (ml/mg) were the following: group A, 80 -+ 5; group B, 201 -+ 9 (p < 0.01); group C, 116 +- 7 (p < 0.05); group D, 245 +- 24 (p < 0.05). Mean gastrointestinal fluid loss was highest in the hypothermic group and smallest if superoxide dismutase was given. Mean intracellnlar fluid volume was increased in groups B and D compared with group A (p < 0.01). We conclude that aortic occlusion and reperfusion results in increased requirement for intravenous fluid as fluid shifts into the intracellnlar space and is lost into the gastrointestinal tract. These fluid shifts are increased by hypothermia but can be minimized by superoxide dismutase. (J V^sc SURG 1991; 13:637-45.)

The postoperative course of patients undergoing thoracoabdominal aortic reconstruction or mesenteric revascularization is often marked by a requirement for large volumes of intravenous fluid to maintain hemodynamic stability. Although some continuing loss into the operative field may occur, other factors may also play more important roles in producing intravascular volume depletion. These factors may include the effects of ischemia and reperfusion to visceral and somatic regions distal to the site of aortic occlusion) -3 Hypothermia may also

MATERIAL AND METHOD

From the Sectionof Vascular Surgery,and Divisionof Gastroenterology and InternalMedicine,Mayo Clinicand Mayo Foundation, Rochester. Presented at the FourteenthAnnualMeeting of the Midwestern Vascular Surgical Society, Toledo, Ohio, September 14-15, 1990. Reprint requests: Peter Gloviczki,MD, MayoClinic,200 First St. SW, Rochester,MN 55905 24/6/28507

Twenty-three mongrel dogs (19.8 to 25.5 kg) were divided into four groups. Group A animals (n = 6) served as control and underwent laparotomy only. Group B animals (n = 6) had the supraceliac aorta cross-clamped for 45 minutes. Group C animals (n = 6) were similar to group B, except that superoxide dismutase (15,000 units/kg; bovine liver, Sigma Chemical Co., St. Louis, Mo.) was infused

be associated with increased fluid requirement in these patients. The goal of this study was to define the mechanisms of intravascular fluid loss after temporary occlusion of the supraceliac aorta in a canine model. An attempt was also made to determine whether the quantity and distribution of fluid loss coudd be altered by mild hypothermia or by reducing the reperfusion injury with an oxygen-derived free radical scavenger.

637

638 Bourchier et al.

over 15 minutes into the aorta just distal to the cross-clamp. The infusion was started 5 minutes before release of the cross-clamp. Group D animals (n = 5) were similar to group B, except that mild hypothermia was maintained during the ischemia and the reperfusion period. Core temperature was allowed to fall to a mean of 33.5 +_ 1.3 ° C at aortic occlusion, 32.1 + 1.5 ° C at the time of cross-clamp release, and 30.9 + 1.6 ° C after the first hour of reperfusion. Animal care complied with the National Institute of Health guidelines as given in "Principles of Laboratory Animal Care" and "Guide for the Care and Use of Laboratory Animals" (NIH Publication No. 80-23, revised 1985). Operative procedure Anesthesia was induced with methohexital (10 mg/kg), endotracheal intubation was performed, and ventilation was controlled with a Mark IV (Bird Products Corp., Palm Springs, Calif.) ventilator to maintain arterial Po2 and Pco2 within normal limits. Anesthesia was maintained with halothane (1% to 1.5%). Core temperature was held constant above 35.5 ° C by controlling ambient room temperature and with heating blankets in all groups except group D. A thermistor-tipped pulmonary artery catheter was placed via the right external jugular vein to monitor pulmonary wedge pressure, cardiac output, and core temperature. Systemic arterial blood pressure was monitored via a cannula placed in the right carotid artery, and a urinary catheter was placed to monitor urinary output. After midline laparotomy the aorta was mobilized above the origin of the celiac trunk. A 22-gauge cannula was placed in the portal vein for blood sampling, and the origin of the superior mesenteric artery (SMA) was dissected to allow placement of a 4 mm probe connected to an ultrasonic blood flow meter (Transonic Systems Inc. Ithaca, N.Y.). The coaxial probe of a reflectance spectrophotometer (Tissue Spectrum Analyzer TS 200; Sumitomo Electric Industries, Ltd., Osaka, Japan) was introduced into the lumen of the midjejunum through a small enterotomy to measure changes in mucosal hemoglobin concentration and changes in mucosal oxygen saturation. 4"s Both these parameters were expressed as calculated indexes dependent on differences in tissue absorption of light, as discussed in detail by Su et al.s Changes in indexes ofmucosal hemoglobin concentration and in indexes of mucosal oxygen saturation have been shown to correlate directly with alterations in gastrointestinal blood flow? "s These measurements allowed repeated, rapid assessment of intestinal mucosal perfusion.

Journal of VASCULAR SURGERY

After baseline measurements of hemodynamic and mucosal perfusion parameters, heparin (100 units/kg given intravenously) was administered to all groups, and the supraceliac aorta was occluded with a single cross-clamp for 45 minutes in groups B, C, and D. A cannula was introduced into the aorta just distal to the cross-clamp via a branch of the splenic artery, and 200 ml of cold normal saline solution was infused over the 45-minute occlusion period in group D to further augment temperature reduction. In groups A, B, and C 200 ml of 37 ° C normal saline solution was infused over the same time period. Immediately before reperfusion, 10 ml of fluorescein isothiocyanate (FITC)-dextran 4000 (FITC-dextran 4000, Sigma Chemical Co., St. Louis, Mo.) was injected into the lumen of a 10 cm loop of proximal jejunum isolated between soft occluding clamps to measure changes in mucosal permeability, as described herein. Blood flow at this time was also measured in the SMA as were indexes ofmucosal perfusion. After release of the aortic clamp, SMA flow and mucosal perfusion indexes were measured for a period of 1 hour. Portal vein blood samples were taken at timed intervals to measure FITC-dextran 4000 concentration. All intraabdominal cannulas and probes were then removed and the abdominal incision closed. On recovery from anesthesia the animals were extubated, sedated with morphine (0.5 to 1.0 mg/kg intravenously) and placed in a restraining sling for 12 hours. Intravenous electrolyte solution (PlasmaLyte A, p H 7.4, 294 mosm/L, Na + 140 mEq/L, K ÷ 5 mEq/L, Mg 2+ 3 mEq/L, CI- 98 mEq/L, acetate 27 mEq/L, gluconate 23 mEq/L, Baxter Laboratories, Longwood, Fla.) was administered to maintain hemodynamic parameters at baseline levels. Systemic blood pressure, pulmonary capillary wedge pressure, and cardiac output were monitored, but the pulmonary artery wedge pressure was used as the principal indicator of intravascular volume and the need for volume replacement. Urine and gastrointestinal losses were collected and measured. At the end of the 12-hour observation period, the volumes of total body water, extracellular water, and plasma water were determined by use of radionuclide dilution techniques as described herein. At the time the animals were killed, intraperitoneal and intraluminal gastrointestinal fluid volumes were also measured. The animals were killed with an overdose of sodium pentobarbital. Measurement o f intestinal permeability A fluorescein-labeled low molecular weight dextran, fluorescein-isothiocyanate-conjugated dextran, average molecular weight 3900; was used to quantify

Volume 13 Number 5 May 1991

Fluid loss after aortic occlusion

639

600

[ ] Group Group i k~

._o E:: 500 $ ._ ~ 400

[ ] Group

• Group

1

Base-

line

xc on

xc off

5

10

15

20

30

40

50

60

Reperfusion, min CG134648×

Fig. 1. Superior mesenteric artery blood flow (ml/min) at baseline, at the end of 45 minutes occlusion of the supraceliac aorta (xc on), immediately after cross-clamp release (xc off), and during 60 minutes of reperfusion (mean value + SEM). changes in intestinal mucosal permeability by means of a technique described by Tagesson et al.6 and Otamiri et al. 7 Ten ml of 150 mmol/L phosphatebuffered saline (pH 7.3) containing 100 mg FITCdextran 4000 was injected into the intestinal lumen immediately before release of the cross-clamp. Portal vein blood samples (1 ml) were mixed with 9 ml Tris (50 mmol/L)-Na C1 (150 mmol/L) pH 10.3, and centrifuged at 2800g for 5 minutes. The supernatant was analyzed for FITC-dextran 4000 concentration by use of an Aminco-Bowman spectrophotofluorometer (American Instrument Company Inc., Silver Spring, Md.).

Measurement of body water distribution Volumes of plasma, extracellular, and total body water were determined by measuring the distribution volumes of radioactive iodinated I125 serum albumin (125RISA; New England Nuclear, Boston, Mass.), radiosulfate-35 (s~SO4; American Radiolabeled Chemicals Inc., St. Louis, Mo.), and tritiated w a t e r ( S H 2 0 ; New England Nuclear), respectively as described by Bauer et al.s and modified by Fowler et al.9 After the collection of baseline blood samples, each animal received injections of 12SRISA, ( = 10 ~Ci); ssSO4, ( =20 IxCi); and SHe0, (10 ~Ci). Blood samples were obtained at 15, 30, 45, and 60 minutes after injection of the isotopes and were centrifuged to obtain plasma. Two milliliter aliquots were then counted in a well-type gamma counter for ~2sI. Similar aliquots were counted from previously prepared standard solutions of ~2SRISA. Four milliliter of 10% trichloroacetic acid was added to a second 2 ml plasma sample to precipitate plasma proteins. This sample was centrifuged, and 1 ml of

the supernatant was mixed with 12 ml of scintillation fluid (Ultima Gold, Packard Instrument Company Inc., Downers Grove, Ill.) in a liquid scintillation vial and was analyzed for 35SO4 and 3H20 in a dual window automated scintillation counter. Standard samples of ssSO4 and SH20 were prepared by adding similar amounts of trichloroacetic acid and liquid scintillation fluid to provide reference dilutions of 1 in 14,000 for SH20 and 1 in 4000 for 3sSO4. These were analyzed with the experimental samples. Counts corrected for quench and cross-over were used in subsequent calculations. Linear regression analysis of 12sI and 3sSO4 counts was performed to determine the theoretic equilibrated sample counts at the time of injection. These values were used to determine the distribution volumes of 12SRISA and 3sSO4, which were taken as estimates of plasma volume and extraceUular fluid volume, respectively. Interstitial fluid volume was determined by subtracting plasma volume from extracellular fluid volume. The counts per minute of ~H20 in the plasma samples were averaged, and this value was used to calculate the distribution volume of 3H20, and hence total body water. Intracellular fluid volume was determined by subtracting extracellular fluid volume from total body water. Statistical analysis The data were analyzed by means of one way analysis of variance (ANOVA). The Newman-Keuls multiple range test was applied to detect differences between groups in the mean values of SMA flow, indexes of mucosal perfusion and permeability, sources of fluid loss, and body fluid compartment volumes. Results are reported as mean __ standard error.

640

Journal of VASCULAR SURGERY

Bourchier et al.

~) 200

[] Group A k~ Group B] • Group C • Group D

I ~ 150 o O ¢-

o

m

Base- xc line on

xc off

5

10

15

20

30

40

50

60

Reperfusion, min CG 134648X 2

Fig. 2. Small intestine mucosal index of hemoglobin concentration at baseline, at the end of 45 minutes occlusion of the supraceliac aorta (xc on), immediately after cross-clamp release (xc off), and during 60 minutes of reperfusion (mean value + SEM).

l i Group AB Group C Group D

50 '°

30

~ 20 "6

'i

Base- xc line on

xc off

5

10

15

20

30

40

50

60

Reperfusion, min CG134648× 3

Fig. 3. Small intestine mucosal index of oxygen saturation at baseline, at the end of 45 minutes occlusion of the supraceliac aorta (xc on), immediately after cross-clamp release (xc off), and during 60 minutes of repeffusion (mean value _+ SEM). RESULTS Intraoperative m e a s u r e m e n t s Superior mesenteric artery flow. Mean SMA flows were similar in aLl groups at baseline measurement and fell to similar levels after aortic occlusion in groups B, C, and D (Fig. 1). After release o f the aortic cross-clamp, there was a brief hyperemic response in groups B, C, and D, but mean SMA flow returned to levels similar to group A within 5 minutes in groups B and D. The hyperemic response was slightly prolonged to 15 minutes in group C but returned to

control levels afterwards. There was a delayed increase in mean SMA flow in group D compared with group A at 30 and 40 minutes after reperfusion

(p < 0.05). Indexes o f intestinal mucosal perfusion. The mcan index o f mucosal hemoglobin conccntration was similar in all groups at basclinc, and fell with aortic occlusion in groups B, C, and D (Fig. 2). This reduction in mean index o f mucosal hcmoglobin conccntration was slightly less in group C. After cross-clamp rcleasc, mcan index o f mucosal hemo-

Volume 13 Number 5 May 1991

Fluid loss after aortic occlusion 641

........"e'......

] ~

group B

2

group C group D

--'----"

,-r

group A

1

0

i

i

i

i

i

i

10

20

30

40

50

60

70

minutes alter reperfusion

Fig. 4. Intestinal mucosal permeability (ug/ml) changes after reperfusion, measured by increases in portal vein FITC-dextran 4000 concentration (mean value - SEN[).

300

200

E 100

group h

group B

group C

group D

Fig. 5. Fluid balance (ml/kg) at 12 hours after aortic reperfusion. Total intravenous fluid administered is compared to sources of fluid loss (mean value -+ SEM). globin concentration was higher in group B than group A for the first 10 minutes of reperfusion (p < 0.05). In groups C and D, mean index of mucosal hemoglobin concentration returned to values similar to group A within the first 5 minutes of reperfusion. All groups had similar mean indexes of mucosal oxygen saturation at baseline (Fig. 3). Aortic occlusion for 45 minutes produced similar reductions in groups B, C, and D. After cross-clamp release, mean index of mucosal oxygen saturation in groups B, C, and D returned to values similar to group A, with no significant change observed over the 60-minute period of reperfusion. Mucosal permeability. Mean portal vein FITCdextran 4000 concentration was similar in all groups at 5 minutes after release of the aortic cross-clamp (Fig. 4). With continued reperfusion, group B

showed a marked increase that was greater than all other groups after 60 minutes (p < 0.01 compared to Group A). This increase in permeability was reduced by both the administration of superoxide dismutase (p < 0.01) and by hypothermia (p < 0.05), with similar mean FITC-dextran 4000 concentrations being seen in these two groups over the period of reperfusion. Mucosal permeability, however, was still increased in groups C and D when compared with group A (p < 0.01).

Postoperative measurements Temporary occlusion of the supraceliac aorta resulted in a significant increase (p < 0.01, group B compared with group A) in postoperative intravenous fluid volume requirement (Fig. 5). Hypothermia during ischemia and reperfusion produced a still greater requirement for postoperative fluids (p <

642

lournal of VASCULAR SURGERY

Bourchier et al.

800 total body water

intracellular

600

extracellular

== E

400

200

0

group A

group B

group C

group D

Fig. 6. Total body water (ml/kg) and distribution into intracellular and extraceUular compartments 12 hours after reperfusion (mean value _+ SEM).

300

extracellular interstitial 200

plasma

E 100

group A

group B

group C

group D

Fig. 7. Distribution of extracellular fluid (ml/kg) as interstitial water and plasma water 12 hours after reperfusion (mean value - SEM).

0.05, group D compared with group B). The administration of superoxide dismutase (group C) significantly reduced this demand (2 < 0.01), although the mean volume administered in this group was still greater than that for laparotomy alone

(p < 0.05). There was no significant difference in mean urine output volume between any of the groups. Mean fluid loss into the intestine was greatest in group D (p < 0.01 compared with group A). Mean fluid loss in group C was not different from that seen in the control group. Mean intraperitoneal fluid loss was greater in group D (p < 0.05) than in any other group, whereas group B showed an increased loss compared with group A (p < 0.05). In group C the mean intraperitoneal fluid volume was similar to group A.

Distribution o f total body water Mean total body water volume 12 hours after reperfusion was greater in group D (20 < 0.05) than in any other group, whereas in group B it was greater (p < 0.05) than group A and group C (Fig. 6). No significant difference was found between groups in mean extracellular fluid volume. The mean intracellular volume was expanded above that measured in group A in both groups B and D (p < 0.01), with no difference found between these latter groups. If superoxide dismutase was given, the mean intracellular fluid volume did not differ from that found in group A undergoing laparotomy alone. No significant differences were found between groups in mean interstitial or plasma volumes 12 hours after reperfusion (Fig. 7).

Volume 13 Number 5 May 1991

DISCUSSION In patients undergoing major surgical procedures, fluid shifts between extracellular and intracellular compartments may be associated with significant morbidity. Patients with impairment in cardiac, renal, or pulmonary functions are especially prone to complications. These fluid shifts are most pronounced in procedures involving reconstruction of the thoracoabdominal aorta or of the mesenteric arteries if the aorta is cross-clamped above the origin of the celiac trunk. Some previous studies have suggested that surgical trauma may be associated with a shift of fluid into the intracellular space, l°'H There are several other investigators, however, who could not confirm a significant increase in intracellular volume after surgical trauma. ~ l s Considerable debate still exists concerning intracellular volume expansion after trauma, although it seems that in the immediate posttrauma phase this does occur. The seemingly conflicting results found in the literature are also due to the timing of the measurements after trauma. B6ck et al. ~s studied body fluid distribution 2 to 3 days after the trauma occurred and found expansion of the interstitial space with no intracellular volume changes. We have examined a very early phase after surgical trauma associated with severe ischemia. We found that temporary occlusion of the supraceliac aorta produced expansion of the intracellular space above that seen with laparotomy alone. This was associated with an increased fluid requirement necessary to maintain hemodynamic stability. The most likely mechanism producing this fluid shift is ischemia and reperfusion injury, which is the resuk of hypoperfusion distal to the site of aortic occlusion. Cellular swelling has been found to occur at the moment of reperfusion ofischemic tissue, 16implying an alteration in cell membrane permeability. Recent work has shown that ischemia produces a fall in cell membrane electrical potential, and this continues during reperfusion despite adequate levels of intracellular adenosine triphosphate necessary to drive the adenosine triphosphate-dependent sodiumpotassium pump? 7 This finding suggests that cell membrane damage occurs during reperfusion as a result of lipid peroxidation by oxygen-derived free radicals/8-z° In our study we found indirect confirmarion of this mechanism. The administration of a free radical scavenger, superoxide dismutase, at the time of aortic cross-clamp release significantly reduced the perioperative fluid requirement and decreased expansion of the intracellular volume. The changes in body fluid compartments after

Fluid loss after aortic occlusion 643

infrarenal aortic occlusion in patients undergoing elective abdominal aortic reconstruction have been described by Nielsen et al. 2,s They found that 50% of the intravenous fluids required to maintain extracellular fluid volume had moved into the intracellular space during the first postoperative day. This was also accompanied by a shift of sodium into the cells, s As a critique of this study, however, it should be mentioned that these conclusions were reached by indirect measurement of the intracellular volume. The method used by these authors for measuring extracellular fluids detects only approximately one half of the value measured by sulfate or bromine. In our study we found that supraceliac aortic occlusion resulted in a fluid requirement 2.5 times greater than that for laparotomy alone, and that approximately 60% of this volume entered the intracellular space. Since such increase in intracellular volume has not been found in most reports after surgical trauma, 12~s in our animal model it is most probably attributed to severe ischemia. In our study superoxide dismutase administration reduced the fluid volume requirement to 1.5 times that of laparotomy alone and prevented the intraceUular volume expansion that followed supraceliac aortic occlusion. Another major site of fluid loss from the intravascular space is the gastrointestinal tract. Crossclamping the aorta above the origin of the arteries supplying the gut results in the rapid onset of ischemia. 21 The small intestine is extremely sensitive to ischemia, with changes in mucosal enzymes being found after only 10 seconds of hypoperfusion. 22 On reperfusion a marked increase occurs in mucosal vascular permeability with loss of fluid into the interstitial space and into the lumen of the gut. 22 Evidence exists that oxygen-derived free radicals may be important in producing these permeability increases. 2s The effect of hypoperfusion of isolated segments of small intestine and the stomach and the protective role of free radical scavengers such as superoxide dismutase has been examined previously. 24-26On review of the literature, however, we failed to find an evaluation of high aortic occlusion and subsequent reperfusion with and without free radical scavengers on intestinal permeability. Bulkley et al.27 demonstrated in a canine model that intestinal flow reduction below 20 ml/min/100 gm of small intestine resulted in significant reperfusion permeability changes. We measured SMA flow rates less than 15 rrd/min in all animals in this study undergoing aortic occlusion, and have confirmed the increase in mucosal permeability on reperfusion after this degree of hypopcrfusion. We were unable to dem-

644

Journalof VASCULAR SURGERY

Bourchier et al.

onstrate any consistent differences in SMA flow or intestinal mucosal perfusion during reperfusion between any of the experimental groups. It seems, therefore, unlikely that fluid loss into the gastrointestinal tract is merely a result of increased reperfusion flow as previously suggested. 1 The importance of oxygen-derived free radicals as initiating agents leading to these effects is implied by the reduction in mucosal permeability produced by superoxide dismutase administration. In these animals there was also a reduction of fluid loss into the gut lumen and into the peritoneal cavity. It is important to mention that our study did not investigate the difference between changes in vascular endothelial and cell membrane permeability. We measured changes in intestinal permeability to FITCdextran 4000, and this method did not allow us to define whether vascular or cell membrane permeability changes were primarily involved. It is likely that vascular permeability changes are governed principally by the intracellular junctions or by loss of cells from the basement membrane. In contrast, cellular permeability changes are most probably the result of damage to the cellular membrane or the sodium pump. Although increase of intracellular volume is determined mainly by alterations in cellular membrane permeability, fluid shifts into the interstitial space is mostly governed by vascular endothelial permeability. The effects of mild hypothermia appear contradictory. Although hypothermia may develop during prolonged surgical procedures, this provides some degree of protection from the effects of ischemia, mainly as a result of reduction of tissue oxygen consumption. 28'29 Examination of the effects of temperature reduction from 40 ° C to 30 ° C in the canine ileum showed no correlating changes in blood flow, but a 2.5 to 3 time reduction in oxygen uptake. 3° Hypothermia to 28 ° C before reperfusion has been shown to prevent increased capillary permeability in a feline intestinal model of ischemiareperfusion injury, sl We found a similar protective effect of mild hypothermia, with reduction of the intestinal mucosal permeability to a degree equal to that achieved by superoxide dismutase administration. There is recent evidence, however, that hypothermia during hypoperfusion may produce changes in the cell membrane that exacerbate the effects of ischemia. 32 A reduction in cell membrane electrical potential is caused by a selective increase in permeability to sodium ions, producing a significant increase in intracellular sodium concentration. The additive effect of free radical-induced cell membrane

damage on sodium permeability should produce a marked intracellular volume expansion. The fact that this cell membrane damage is not reflected in a marked increase in intestinal mucosal permeability may suggest that the rate of free radical production is dependent on temperature and is significantly reduced at this degree ofhypothermia. This suggestion is supported by Jurkovich et al. s~ who reported that rewarming from hypothermia before reperfusion increased intestinal permeability to levels greater than that seen with normothermic ischemia-reperfusion. In our study the greatest fluid shifts were seen in the hypothermic group, suggesting that rewarming after hypothermic reperfusion produces similar changes. In summary, occlusion of the aorta above the visceral arteries produces profound effects after reperfusion that lead to rapid intravascular volume depletion. An important mechanism appears to be an alteration in cell membrane permeability occurring at the time of reperfusion as the result of the action of oxygen-derived free radicals. The result is fluid migration into the intracellular space. Similar reperfusion changes lead to an increase in permeability of the gastrointestinal tract, with fluid loss into the gut lumen and into the peritoneal cavity. Hypothermia, although seeming to provide an initial protective effect, exacerbates the effect of ischemia-reperfusion during rewarming. During surgical procedures involving supraceliac aortic occlusion, these fluid shifts could be limited by the administration of oxygenderived free radical scavengers and the prevention of intraoperative hypothermia. REFERENCES

1. Haglund U, Lundgren O. Reactions within consecutive vascular sections of the small intestine of the cat during prolonged hypotension. Acta Physiol Scand 1972;84:151-7. 2. Nielsen OM, Engell HC. The importance of plasma colloid osmotic pressure for interstitial fluid volume and fluid balance after elective abdominal vascular surgery. Ann Surg 1986; 203:25-9. 3. Nielsen OM, Engell HC. Changes in extracellular sodium content after elective abdominal vascular surgery. Acta Chir Scand 1986;152:587-91. 4. Leung FW, Morishita T, Livingston EH, Reedy T, Guth PH. Reflectance spectrophotometry for the assessment of gastroduodenal mucosal perfusion. Am J Physiol 1987;252: G797-804. 5. Su KC, Leung FW, Guth PH. Assessment of mucosal hemodynamics in normal human colon and patients with inflammatory bowel disease. Gastrointest Endosc 1989;35: 22-7. 6. Tagesson C, Sjodahl R, Thoren B. Passage of molecules through the wall of the gastrointestinal tract. A simple experimental model. Scand J Gastroenterol 1978;13:51924.

Volume 13 Number 5 May 1991

7. Otamiri T, Sjodahl R, Tagesson C. An experimentalmodel for studying reversible intestinal ischemia. Acta Chir Scand 1987;153:51-6. 8. Bauer JH, WillisLR, Burt RW, Grim CE. Volume studies. II. Simukaneous determination of plasma volume, red cell mass, extracellular fluid, and total body water before and after volume expansion in dog and man. J Lab Clin Med 1975;86:1009-17. 9. Fowler WL, Johnson JA, Kurz KD, Zeigler DW, Dostal DE, Payne CG. Body fluid volumes in rats with mestranol-induced hypertension. Am J Physiol 1986;250 (Heart Circ Physiol 20):H190-5. 10. Flear CTG, Bhattacharya SS, Singh CM. Solute and water exchanges between cells and extracellularfluids in health and disturbances after trauma. J Parent Ent Nutr 1980;4:98120. 11. Moore FD. The metabolic response to injury. Springfield,Ill: Charles C. Thomas, 1952:135-6. 12. Cleland J, Pluth JR, Tauxe WN, KirklJn JW. Blood volume and body fluid compartment changes soon after closed and open intracardiac surgery. J Thorac Cardiovasc Surg 1966; 52:698-724. 13. Gutelius JR, Shizgal HM, Lopez G. The effect of trauma on extracellular water volume. Arch Surg 1968;97:206-14. 14. Shizgal HM, Solomon S, Gutelius JR. Body water distribution after operation. Surg Gynecol Obstet 1977;144:3541. 15. B6ck JC, Barker BC, Clinton AG, Wilson MB, Lewis FR. Post-traumatic changes in, and effect of colloid osmotic pressure on the distribution of body water. Ann Surg 1989;210:395-405. 16. DiBona DR, Powell WJ Jr. Quantitative correlation between cell swellingand necrosisin myocardialischemiain dogs. Circ Res 1980;47:653-65. 17. Perry MO, Fantini G. Ischemia: profile of an enemy. Reperfusion injury of skeletal muscle. J VASCSURG1987;6: 231-4. 18. McCord JM. Oxygen-derived free radicals in postischemic tissue injury. N Engl J Med 1985;312:159-63. 19. Otamiri T, Tagesson C. Role of phospholipase A2 and oxygenated free radicals in mucosal damage after small intestinal ischemia and reperfusion. Am J Surg 1989;157: 562-6.

Fluid loss after aortic occlusion 645

20. Weisfeldt ML. Reperfusion and reperfusion injury. Clin Res 1987;35:13-20. 21. Rickards E, Svensson LG, Coull A, Fimmel CJ, Hinder RA. Gastrointestinal blood flow with aortic cross-clampingeffect of vascular shunt and flunarizine. S Afr J Surg 1985;23:140-1, 22. Granger DN, Sennet M, McElearney P, Taylor AE. Effect of local arterial hypotension on cat intestinal permeability. Gastroenterology 1980;79:474-80. 23. Granger DN, Parks DA. Role of oxygen radicals in the pathogenesis of intestinal ischemia. Physiologist 1983;26: 159-63. 24. Itoh M, Guth PH. Role of oxygen-derived free radicals in hemorrhagic shock-inducedgastric lesions in the rat. Gastroenterology 1985;88:1162-7. 25. Parks DA, BulkleyGB, Granger DN. Role of oxygen-derived free radicals in digestivetract diseases. Surgery 1983;94:41522. 26. Dalsing MC, Grosfeld JL, Shiftier MA, et al. Superoxide dismutase: a cellular protective enzyme in bowel ischernia. J Surg Res 1983;34:589-96. 27. BulkleyGB, Kvietys PR, Parks DA, Perry MA, Granger DN. Relationship of blood flow and oxygen consumption to ischemicinjury in the canine smallintestine. Gastroenterology 1985;89:852-7. 28. Allen FM. Physical and toxic factors in shock. Arch Surg 1939;38:155-80. 29. Bigelow WG, Lindsay WK, Harrison RC, Gordon RA, Greenwood RF. Oxygen transport and utilization in dogs at low body temperatures. Am J Physiol 1950;160:125-37. 30. KvietysPR, Harper SL, Korthuis RJ, Granger DN. Effects of temperature on ileal blood flow and oxygenation. Am J Physiol 1985;249:G246-9. 31. Jurkovich GJ, Pitt RM, Curreri PW, Granger DN. Hypothermia prevents increased capillary permeability following ischemia-reperfusioninjury. J Surg Res 1988;44:514-21. 32. Fantini GA, Zadeh BJ, Chiao J, Krieger KH, Isom OW, Shires GT. Effect of hypothermia on cellular membrane function during low-flow extracorporeal circulation. Surgery 1987;102:132-9.

Submitted Oct. 11, 1990; accepted Feb. 7, 1991.