- Email: [email protected]
Importance of Collateral Circulation in the Vascularly Occluded Feline Intestine
GASTROENTEROLOGY 1987;92:1215-9
Importance of Collateral Circulation in the Vascularly Occluded Feline Intestine ANDRE J. PREMEN, VIOLETA BANCHS, WILLIAM A. WOMACK, PETER R. KVIETYS, and D. NEIL GRANGER Department of Physiology, College of Medicine, University of South Alabama, Mobile, Alabama
The aim of this study was to assess the extent of collateral blood flow provided by the celiac and inferior mesenteric arteries to the intestines during total occlusion of the superior mesenteric artery (SMA). In anesthetized cats, blood flow to the pancreas, duodenum, jejunum, ileum, and colon was measured with radioactive microspheres (15 /Lm in diameter) before and during occlusion of the SMA. Superior mesenteric artery occlusion significantly decreased (by 63%) blood flow to the head of the pancreas. Flow to the neck and tail of the pancreas was not altered. Blood flow to the proximal and distal duodenum was significantly reduced by 35% and 61 %, respectively. Along the entire jejunum and ileum, SMA occlusion markedly decreased blood flow by an average of 71%. In the proximal colon, blood flow decreased by 63%, whereas flow to the middle and distal colon was not affected by SMA occlusion. Reduction in total wall blood flow to the small and large intestines was largely due to a marked reduction in mucosa/submucosa blood flow; muscularis/serosa flow was not affected. The results of this study suggest that total occlusion of the SMA does not compromise blood flow to the neck and tail of the pancreas and middle and distal colon (tissues that are normally perfused with blood from either the celiac or inferior mesenteric arteries). Perfusion through collaterals maintains flow to the head of the pancreas and gut (from duodenum to proximal colon) to within 30%-65% of control (preocclusion) flow. An important new observation of this study is that collateral blood vessels are much more effective Received July 8, 1986. Accepted November 17, 1986. Address request for reprints to: D. Neil Granger, Ph.D., Department of Physiology and Biophysics, Louisiana State University Medical Center, P.O. Box 33932, Shreveport, Louisiana 71130-3932. This work was supported by grant AM33548 from the Institute of Arthritis, Diabetes and Digestive and Kidney Diseases. The authors thank Ursula Romano for preparation of the manuscript. © 1987 by the American Gastroenterological Association 0016-5085/87/$3.50
in preventing ischemia in the muscularis/serosa than in the mucosa/submucosa after SMA occlusion. Clinical observations suggest that collateral blood flow derived from the celiac and inferior mesenteric arteries affords a significant degree of protection against small bowel ischemia when the superior mesenteric artery (SMA) is occluded (1-9). Despite the recognized importance of the collateral circulation, quantitative studies of collateral blood flow in the small bowel have only recently been undertaken (10,11). Bulkley and associates (10) have shown that collateral vessels between adjacent segments of canine small bowel can maintain blood flow in one segment at -55% of its control level when the artery to that segment is totally occluded. Virtually all of this collateral flow was carried by precapillary (arterial) channels. Two-thirds of the collateral flow was carried by extramural (marginal) channels, whereas the remaining one-third was supplied by intramural connections. Although these studies have improved our understanding of collateral blood flow between adjacent vascular arcades, it remains unclear whether these findings can be extrapolated to situations where the SMA is totally occluded. Thus, the aim of the present study was to determine the extent of collateral blood flow provided by the celiac and inferior mesenteric arteries to the small intestine during total occlusion of the SMA.
Materials and Methods Surgical Preparation Experiments were conducted on 6 cats (weighing 2.45 ± 0.15 kg) of either sex that were fasted for 18-24 h. The animals were initially anesthetized with ketamine hydrochloride (30 mg/kg, Lm.) then maintained under anesthesia with intravenous pentobarbital sodium. A tracheostomy was performed to facilitate breathing and as Abbreviation used in this paper: SMA, superior mesenteric artery.
1216
PREMEN ET AL.
a means of artificial respiration if the cats failed to breathe spontaneously. The right carotid and left femoral arteries were cannulated with PE-190 tubing. The carotid cannula was carefully advanced into the left ventricle while continuously monitoring pressure. The position of the cannula was verified by the characteristic ventricular pressure waveform and by inspection at necropsy. The femoral cannula was threaded into the abdominal aorta and used for monitoring mean arterial pressure and withdrawal of the reference blood sample (reference organ). Arterial and ventricular pressures were recorded on a Grass recorder, (model 7D, Grass Instrument Co., Quincy, Mass.). A cannula was placed in a femoral vein (PE-190) for administration of isotonic saline to compensate for evaporative water loss and as a route for pentobarbital sodium administration. The abdomen was opened through a midline incision. The SMA, near its origin on the abdominal aorta, was isolated from surrounding tissues with care taken to avoid damage to the periarterial nerves. A loose ligature of silk (2-0) was placed around the SMA. The visceral organs were carefully replaced in the abdomen and time was allowed (-45 min) for all pressures to reach a steady state. The abdominal incision was covered with saline-soaked gauze and plastic wrap. Body temperature was maintained at 38°C with a thermistor-controlled heat lamp.
Measurement of Organ Blood Flow Blood flow to the pancreas, small bowel, colon, and kidneys was determined using two different radiolabeled micro spheres; scandium 46 (46 SC) and strontium 85 (85 Sr). Each microsphere (15 ± 3 /Lm diameter) (3M Medical Products, st. Paul, Minn.) was suspended in 0.9% saline containing one drop of 0.05% Tween 80. Before injection, each microsphere was dispersed in an ultrasonic bath (Bransonic 220) and thoroughly mixed with a vortex shaker (Fisher Scientific Co., Pittsburgh, Pa.).
Experimental Protocol After a steady state was achieved (control), a 1.0-ml suspension containing -750,000 microspheres, labeled with either 46SC or 85Sr, was injected into the left ventricle over a period of 15-20 s. The carotid cannula was flushed with warmed isotonic saline. Ten seconds before injection, the reference sample was withdrawn from the left femoral artery at a rate of 2.72 mllmin for a total of 75 s using a pump (Harvard Apparatus, Millis, Mass.). A volume of a warmed physiologic electrolyte solution (Plasma-Lyte, Travenol) equal to the reference sample (-3.4 ml) was reinfused through the carotid cannula during the withdrawal period. This procedure maintained volume in the animals during the withdrawal period. Arterial pressure was monitored both before and after the reference sample withdrawal period to ensure that our procedure did not affect arterial pressure. After the first microsphere injection, the ligature around the SMA was tied to totally occlude blood flow through the vessel. Arterial pressure was carefully monitored and "occluded" blood flow to the various organs was not measured until arterial pressure reached a new steady state (-15-20 min after occlusion of
GASTROENTEROLOGY Vol. 92, No.5, Part 1
the vessel). The second microsphere was injected into the left ventricle using the identical protocol as described above. The animals were killed with an intravenous injection of pentobarbital sodium. The abdominal cavity was opened and the pancreas, duodenum, jejunum, ileum, colon, and both kidneys were removed for determination of organ blood flows. The pancreas was divided into three regions: the head (adjacent to the duodenum), the neck (adjacent to the gastric antrum), and the tail (adjacent to the spleen). Ligatures were placed at the junction between (a) duodenum and pylorus, (b) duodenum and jejunum (ligament of Treitz), (c) ileum and colon (ileocecal valve), and (d) colon and rectum. The duodenum was excised and divided into two equal portions (proximal and distal). The remaining small bowel (jejunoileum) was divided into 6-cm segments. The colon was divided into three equal parts representing proximal, middle, and distal portions. Intestinal segments were opened along their antimesenteric border, rinsed in isotonic saline, and pinned flat on their mucosal surface on preweighed nonabsorbant weighing paper. The muscularis/serosa was carefully removed from the mucosa/submucosa by dissection with a spatula. The intestinal tissues, pancreas, and kidneys were placed in preweighed counting tubes. The weighing paper was added to the mucosa/submucosa compartment to avoid excessive loss of tissue and microspheres. Larger organs were dissected into smaller portions for uniform geometry within the counting tubes. Organ and reference blood sample activities of 46SC and 85Sr were measured in a LKB Compu-Gamma spectrometer (LKB Instruments, Inc., Rockville, Md.). The error in measurement of the radioactivity induced by spillover of 46SC into the 85Sr channel was corrected using 46SC and 85Sr standards.
Calculation of Blood Flow Organ blood flow in milliliters per minute was calculated using the following equation: BF =
Organ cpm x WR Reference sample cpm
,
where BF is blood flow, cpm is counts per minute, and WR the reference-blood sample withdrawal rate (2.72 mllmin). Organ blood flow was normalized to milliliters per minute per gram organ weight. In each cat, the following criteria were used to evaluate the reliability of microsphere injections: (a) absence of large changes in mean arterial pressure (± 15 mmHg) before or after microsphere injections and (b) agreement within 10% of the total counts per minute for both kidneys.
Data Analysis All values are expressed as mean ± SEM. Each cat served as its own control and SMA-occluded blood flows were compared against their respective controls using Student's paired t-test (two-tail). A probability <0.05 was considered significant.
May 1987
COLLATERAL BLOOD FLOW DURING OCCLUSION
Table 1. Pancreatic Blood Flow During Total Occlusion of the Superior Mesenteric Artery Region of pancreas
Control flow (mllmin . g)
SMA occluded flow a (mllmin . g)
Head Neck Tail
1.12 ± 0.29 1.16 ± 0.27 1.45 ± 0.30
0.41 ± 0.13 b 1.27 ± 0.27 2.01 ± 0.27c
SMA, superior mesenteric artery. Values are mean ± SEM. a Superior mesenteric artery occluded blood flows were measured after arterial pressure achieved a steady state (-15-20 min after total occlusion of the SMA). b P < 0.05 compared to respective control. c p = 0.07. n = 6.
Results Mean arterial pressure averaged 123 ± 6 mmHg during the control (preinjection) period. Mean arterial pressure was not altered during or after injection of the first microsphere. After the SMA was totally occluded, mean arterial pressure rose rapidly to a pressure of 171 ± 8 mmHg and then gradually declined toward the baseline pressure, plateauing at a steady state pressure of 137 ± 4 mmHg. Approximately 15-20 min were required to achieve the new steady state. After injection of the second microsphere, mean arterial pressure decreased slightly to 128 ± 7 mmHg. Renal blood flow was not affected by occlusion of the SMA. Left kidney blood flow changed from 2.08 ± 0.23 to 2.15 ± 0.19 and right kidney blood flow from 2.04 ± 0.22 to 2.09 ± 0.18 mllmin . g, respectively. The equality between left and right kidney blood flows during the control and occlusion period corroborated a homogeneous distribution of the injected microspheres. Table 1 summarizes the changes in blood flow to
1217
the pancreas (head, neck, and tail) after occlusion of the SMA. Occlusion produced a differential effect on the various regions of the pancreas. Blood flow to the head decreased significantly by an average of 63%, whereas neck blood flow was unaltered. Tail blood flow demonstrated a hyperemia (39%) that was nearly significant (p = 0.07). The changes in total wall blood flow to the small intestine after occlusion of the SMA are summarized in Table 2. Occlusion of the vessel significantly reduced duodenal, jejunal, and ileal blood flow. Proximal duodenal blood flow fell by 35%, whereas distal duodenal blood flow fell by 61 %. Blood flow to the remainder of the small bowel (jejunoileum) was reduced by an average of 71% (range 59%-76%). When the small bowel was partitioned into five equal segments (expressd as a percentage of total jejunoileallength), a remarkably consistent decrease in blood flow (mean of 71 %) was observed (Table 2). Estimates of intramural blood flow distribution (mucosa/submucosa and muscularis/serosa) revealed that the reductions in total blood flow to the intestines were due exclusively to a marked decrease in mucosa/submucosa flow (Table 2). There were no significant changes in blood flow to the muscularis/ serosa layer of the intestines. Control blood flow to the muscularis/serosa ranged from 0.043 ± 0.013 to 0.063 ± 0.020 mllmin' g in the duodenum, from 0.075 ± 0.041 to 0.170 ± 0.054 mllmin' g in the jejunoileum, and from 0.044 ± 0.011 to 0.135 ± 0.057 mllmin' g in the colon.
Discussion The results from the present study quantitate the importance of collateral blood flow to the pan-
Table 2. Intestinal Blood Flow During Total Occlusion of the Superior Mesenteric Artery Total wall blood flow
Mucosa/submucosa blood flow
Region of intestine
Control (mllmin . g)
SMA occluded a (mllmin . g)
Control (mllmin . g)
SMA occluded a (mllmin . g)
Duodenum Proximal Distal
1.10 ± 0.22 0.92 ± 0.15
0.72 ± 0.17 b 0.36 ± 0.08 b
1.50 ± 0.36 1.57 ± 0.32
1.05 ± 0.26 b 0.58 ± 0.08 b
Jejunoileum 20% total length 40% total length 60% total length 80% total length 100% total length
0.78 0.69 0.62 0.60 0.63
Colon Proximal Middle Distal
1.22 ± 0.39 1.41 ± 0.48 1.45 ± 0.43
± ± ± ± ±
0.08 0.07 0.05 0.05 0.04
0.22 0.18 0.18 0.18 0.19
± ± ± ± ±
0.02b O.Olb O.Olb 0.02b 0.02b
0.45 ± 0.15 b 0.93 ± 0.40 1.27 ± 0.42
1.19 1.13 1.11 0.89 0.95
± ± ± ± ±
0.25 0.24 0.18 0.16 0.20
1.51 ± 0.63 2.23 ± 0.78 2.33 ± 0.73
0.32 0.31 0.30 0.24 0.33
± ± ± ± ±
0.05 b 0.04b 0.05 b 0.05 b 0.07 b
0.67 ± 0.23 b 1.98 ± 0.80 2.03 ± 0.72
SMA, superior mesenteric artery. Values are mean ± SEM. a Superior mesenteric artery occluded blood flows were measured after arterial pressure achieved a steady state (-15-20 min after total occlusion of the SMA). b P < 0.05 compared to respective control. n = 6.
1218 PREMEN ET AL.
creas and intestines after occlusion of the SMA. Superior mesenteric artery occlusion exerted different qualitative and quantitative responses in the head, neck, and tail regions of the pancreas (Table 1). Blood flow to the head of the pancreas was markedly reduced, whereas blood flow in the neck region was unaffected. A hyperemia was observed in the tail region during SMA occlusion. These findings suggest that occlusion of the SMA does not compromise blood flow to the neck and tail of the pancreas (tissues that are normally perfused with blood from the celiac artery). Perfusion through collaterals maintains blood flow to the head of the pancreas to within 37% of control. Occlusion of the SMA significantly reduced duodenal blood flow. Blood flow fell by only 35% in the proximal duodenum, whereas distal duodenal blood flow fell by 61% (Table 2). It is evident that blood flow derived from the celiac artery is important in minimizing the ischemia and tissue hypoxia in the duodenum produced by SMA occlusion. It remains unclear, however, whether the blood supplied by the celiac artery is actually carried by collateral channels or represents direct nutrient flow from superior pancreaticoduodenal vessels. A similar axial gradient in tissue blood flow was observed in the colon during SMA occlusion. Occlusion significantly decreased proximal colon blood flow by 63%. Blood flow to the middle and distal colon, however, was not compromised during SMA occlusion, indicating that perfusion of these areas was largely the responsibility of the inferior mesenteric artery. As was the case in the duodenum, we are unable to discern whether the blood supplied by the inferior mesenteric artery to the proximal colon is indeed carried by collaterals. Occlusion of the SMA markedly reduced jejunal and ileal blood flow to the same extent, i.e., by -71 % (Table 2). Of particular interest is the observation that blood flow was not reduced to zero in the midsection of the jejunoileum. As this region (-50% of total length) was able to maintain its blood flow at -30% of pre occlusion values, perfusion through collaterals was, no doubt, most important. Moreover, in an isolated jejunoileum preparation (perfused by only the SMA), we observed that occlusion of the vessel reduced blood flow to zero in the small bowel. Thus, the collateral channels are apparently long enough to reach the middle of the jejunoileum. The quantitative importance of the intestinal collateral circulation in exteriorized canine jejunal loops was recently assessed in our laboratory (10,11). We found that collateral blood vessels between adjacent segments of jejunum can maintain perfusion in one segment at -55% of its control level when the single artery supplying that segment is
GASTROENTEROLOGY Vol. 92, No.5, Part 1
totally occluded. Furthermore, we (10) demonstrated that (a) virtually all of the collateral blood flow is carried by arterial channels, (b) two-thirds of the collateral flow is carried by extramural (marginal) vessels and one-third of the flow is supplied by intramural connections (e.g., the extensive submucosal vascular plexus), and (c) the major determinants of collateral blood flow in the intestine appear to be passive in nature (i.e., creation of a substantial pressure gradient between nonoccluded and occluded segments that produces a driving force for collateral flow). Extension of these observations from adjacent jejunal segments to the entire jejunoileum would suggest that occlusion of the SMA creates a large pressure gradient between vessels normally perfused by the celiac and inferior mesenteric arteries and those perfused by the SMA. The fact that we did not observe a gradient in blood flow between the ends and center of the jejunoileal segment suggests that the collateral channels are large, low-resistance vessels along which there is no substantial pressure drop. Although we are unable to define the major collateral channel from our data, the previous work by Bulkley et al. (10) suggests that most of the collateral blood flow is carried by extramural (marginal) vessels. Estimates of intramural distribution of blood flow indicate that the reductions in intestinal blood flow observed during SMA occlusion are due solely to a fall in mucosa/submucosa flow (Table 2); flow to the muscularis/serosa is not altered. This finding suggests that the collaterals supplying the mucosa offer greater resistance to blood flow than the collaterals supplying the muscularis. Alternatively, the resistance to blood flow offered by collateral channels to the mucosa and muscularis may be similar and the maintenance of muscle blood flow during SMA occlusion may result from an enhancement of visceral smooth muscle activity. We observed marked increases in intestinal motility after SMA occlusion. Chou and Grassmick (12) have suggested that an active hyperemia, similar to exercise hyperemia in skeletal muscle, occurs in the muscularis of the gut wall during tonic intestinal contractions. Thus, it is possible that an active hyperemia in the muscularis layer may have offset an occlusion-mediated decrease in blood flow as observed in the mucosa/submucosa layer.
References 1. Bonte FJ, Curry GC, Parkley RW. Experimental studies of the splanchnic circulation of the rabbit after ligation of the superior mesenteric artery. Am J RoentgenoI1969;106:691-9. 2. Meyers MH. Griffith's point: critical anastomosis at the splenic flexure. Am J Roentgenol 1976;126:77-94. 3. Michels NA, Padmanabhan S, Kornblith PL. Routes of collat-
COLLATERAL BLOOD FLOW DURING OCCLUSION
May 1987
4. 5. 6. 7. 8.
eral circulation of the gastrointestinal tract as ascertained in a dissection of 500 bodies. Int Surg 1968;49:8-28. Kittredge RD. Ischemia of the bowel. Am J Roentgenol Radium Ther Nucl Med 1968;103:400-4. Novelline RA. Waltman AC. Athanasoulis CA. Baum S. Recent advances in abdominal angiography. Adv Intern Med 1976;21:417-49. Robinson JWL. Antonio Ii JA. The intestinal response to ischemia. Pharmacology 1966;255:178-91. Todd TRJ. Weinman G. McIntyre D. Selective superior mesenteric embolization for small intestinal hemorrh'l-ge. Can J Surg 1979;22:283-6. Kyung JC. Schmidt RW. Lenz J. Effects of experimental
9. 10. 11. 12.
1219
embolization of superior mesenteric artery branch of the intestine. Invest Radiol 1979;14:207-12. Schwilden ED. Secondary circulatory disorders following rectosigmoid resection. Neth J Surg 1983;35:167-72. Bulkley GB. Womack WA. Downey JM. Kvietys PRo Granger ON. Characterization of segmental collateral blood flow in the small intestine. Am J PhysioI1985;249:G228-35. Bulkley GP. Womack WA. Downey JM. Kvietys PRo Granger ON. Collateral blood flow in segmental intestinal ischemia: effects of vasoactive agents. Surgery 1986;100:157-66. Chou CC. Grassmick B. Motility and blood flow distribution within the wall of the gastrointestinal tract. Am J Physiol 1978;235:H34-9.
Importance of Collateral Circulation in the Vascularly Occluded Feline Intestine ANDRE J. PREMEN, VIOLETA BANCHS, WILLIAM A. WOMACK, PETER R. KVIETYS, and D. NEIL GRANGER Department of Physiology, College of Medicine, University of South Alabama, Mobile, Alabama
The aim of this study was to assess the extent of collateral blood flow provided by the celiac and inferior mesenteric arteries to the intestines during total occlusion of the superior mesenteric artery (SMA). In anesthetized cats, blood flow to the pancreas, duodenum, jejunum, ileum, and colon was measured with radioactive microspheres (15 /Lm in diameter) before and during occlusion of the SMA. Superior mesenteric artery occlusion significantly decreased (by 63%) blood flow to the head of the pancreas. Flow to the neck and tail of the pancreas was not altered. Blood flow to the proximal and distal duodenum was significantly reduced by 35% and 61 %, respectively. Along the entire jejunum and ileum, SMA occlusion markedly decreased blood flow by an average of 71%. In the proximal colon, blood flow decreased by 63%, whereas flow to the middle and distal colon was not affected by SMA occlusion. Reduction in total wall blood flow to the small and large intestines was largely due to a marked reduction in mucosa/submucosa blood flow; muscularis/serosa flow was not affected. The results of this study suggest that total occlusion of the SMA does not compromise blood flow to the neck and tail of the pancreas and middle and distal colon (tissues that are normally perfused with blood from either the celiac or inferior mesenteric arteries). Perfusion through collaterals maintains flow to the head of the pancreas and gut (from duodenum to proximal colon) to within 30%-65% of control (preocclusion) flow. An important new observation of this study is that collateral blood vessels are much more effective Received July 8, 1986. Accepted November 17, 1986. Address request for reprints to: D. Neil Granger, Ph.D., Department of Physiology and Biophysics, Louisiana State University Medical Center, P.O. Box 33932, Shreveport, Louisiana 71130-3932. This work was supported by grant AM33548 from the Institute of Arthritis, Diabetes and Digestive and Kidney Diseases. The authors thank Ursula Romano for preparation of the manuscript. © 1987 by the American Gastroenterological Association 0016-5085/87/$3.50
in preventing ischemia in the muscularis/serosa than in the mucosa/submucosa after SMA occlusion. Clinical observations suggest that collateral blood flow derived from the celiac and inferior mesenteric arteries affords a significant degree of protection against small bowel ischemia when the superior mesenteric artery (SMA) is occluded (1-9). Despite the recognized importance of the collateral circulation, quantitative studies of collateral blood flow in the small bowel have only recently been undertaken (10,11). Bulkley and associates (10) have shown that collateral vessels between adjacent segments of canine small bowel can maintain blood flow in one segment at -55% of its control level when the artery to that segment is totally occluded. Virtually all of this collateral flow was carried by precapillary (arterial) channels. Two-thirds of the collateral flow was carried by extramural (marginal) channels, whereas the remaining one-third was supplied by intramural connections. Although these studies have improved our understanding of collateral blood flow between adjacent vascular arcades, it remains unclear whether these findings can be extrapolated to situations where the SMA is totally occluded. Thus, the aim of the present study was to determine the extent of collateral blood flow provided by the celiac and inferior mesenteric arteries to the small intestine during total occlusion of the SMA.
Materials and Methods Surgical Preparation Experiments were conducted on 6 cats (weighing 2.45 ± 0.15 kg) of either sex that were fasted for 18-24 h. The animals were initially anesthetized with ketamine hydrochloride (30 mg/kg, Lm.) then maintained under anesthesia with intravenous pentobarbital sodium. A tracheostomy was performed to facilitate breathing and as Abbreviation used in this paper: SMA, superior mesenteric artery.
1216
PREMEN ET AL.
a means of artificial respiration if the cats failed to breathe spontaneously. The right carotid and left femoral arteries were cannulated with PE-190 tubing. The carotid cannula was carefully advanced into the left ventricle while continuously monitoring pressure. The position of the cannula was verified by the characteristic ventricular pressure waveform and by inspection at necropsy. The femoral cannula was threaded into the abdominal aorta and used for monitoring mean arterial pressure and withdrawal of the reference blood sample (reference organ). Arterial and ventricular pressures were recorded on a Grass recorder, (model 7D, Grass Instrument Co., Quincy, Mass.). A cannula was placed in a femoral vein (PE-190) for administration of isotonic saline to compensate for evaporative water loss and as a route for pentobarbital sodium administration. The abdomen was opened through a midline incision. The SMA, near its origin on the abdominal aorta, was isolated from surrounding tissues with care taken to avoid damage to the periarterial nerves. A loose ligature of silk (2-0) was placed around the SMA. The visceral organs were carefully replaced in the abdomen and time was allowed (-45 min) for all pressures to reach a steady state. The abdominal incision was covered with saline-soaked gauze and plastic wrap. Body temperature was maintained at 38°C with a thermistor-controlled heat lamp.
Measurement of Organ Blood Flow Blood flow to the pancreas, small bowel, colon, and kidneys was determined using two different radiolabeled micro spheres; scandium 46 (46 SC) and strontium 85 (85 Sr). Each microsphere (15 ± 3 /Lm diameter) (3M Medical Products, st. Paul, Minn.) was suspended in 0.9% saline containing one drop of 0.05% Tween 80. Before injection, each microsphere was dispersed in an ultrasonic bath (Bransonic 220) and thoroughly mixed with a vortex shaker (Fisher Scientific Co., Pittsburgh, Pa.).
Experimental Protocol After a steady state was achieved (control), a 1.0-ml suspension containing -750,000 microspheres, labeled with either 46SC or 85Sr, was injected into the left ventricle over a period of 15-20 s. The carotid cannula was flushed with warmed isotonic saline. Ten seconds before injection, the reference sample was withdrawn from the left femoral artery at a rate of 2.72 mllmin for a total of 75 s using a pump (Harvard Apparatus, Millis, Mass.). A volume of a warmed physiologic electrolyte solution (Plasma-Lyte, Travenol) equal to the reference sample (-3.4 ml) was reinfused through the carotid cannula during the withdrawal period. This procedure maintained volume in the animals during the withdrawal period. Arterial pressure was monitored both before and after the reference sample withdrawal period to ensure that our procedure did not affect arterial pressure. After the first microsphere injection, the ligature around the SMA was tied to totally occlude blood flow through the vessel. Arterial pressure was carefully monitored and "occluded" blood flow to the various organs was not measured until arterial pressure reached a new steady state (-15-20 min after occlusion of
GASTROENTEROLOGY Vol. 92, No.5, Part 1
the vessel). The second microsphere was injected into the left ventricle using the identical protocol as described above. The animals were killed with an intravenous injection of pentobarbital sodium. The abdominal cavity was opened and the pancreas, duodenum, jejunum, ileum, colon, and both kidneys were removed for determination of organ blood flows. The pancreas was divided into three regions: the head (adjacent to the duodenum), the neck (adjacent to the gastric antrum), and the tail (adjacent to the spleen). Ligatures were placed at the junction between (a) duodenum and pylorus, (b) duodenum and jejunum (ligament of Treitz), (c) ileum and colon (ileocecal valve), and (d) colon and rectum. The duodenum was excised and divided into two equal portions (proximal and distal). The remaining small bowel (jejunoileum) was divided into 6-cm segments. The colon was divided into three equal parts representing proximal, middle, and distal portions. Intestinal segments were opened along their antimesenteric border, rinsed in isotonic saline, and pinned flat on their mucosal surface on preweighed nonabsorbant weighing paper. The muscularis/serosa was carefully removed from the mucosa/submucosa by dissection with a spatula. The intestinal tissues, pancreas, and kidneys were placed in preweighed counting tubes. The weighing paper was added to the mucosa/submucosa compartment to avoid excessive loss of tissue and microspheres. Larger organs were dissected into smaller portions for uniform geometry within the counting tubes. Organ and reference blood sample activities of 46SC and 85Sr were measured in a LKB Compu-Gamma spectrometer (LKB Instruments, Inc., Rockville, Md.). The error in measurement of the radioactivity induced by spillover of 46SC into the 85Sr channel was corrected using 46SC and 85Sr standards.
Calculation of Blood Flow Organ blood flow in milliliters per minute was calculated using the following equation: BF =
Organ cpm x WR Reference sample cpm
,
where BF is blood flow, cpm is counts per minute, and WR the reference-blood sample withdrawal rate (2.72 mllmin). Organ blood flow was normalized to milliliters per minute per gram organ weight. In each cat, the following criteria were used to evaluate the reliability of microsphere injections: (a) absence of large changes in mean arterial pressure (± 15 mmHg) before or after microsphere injections and (b) agreement within 10% of the total counts per minute for both kidneys.
Data Analysis All values are expressed as mean ± SEM. Each cat served as its own control and SMA-occluded blood flows were compared against their respective controls using Student's paired t-test (two-tail). A probability <0.05 was considered significant.
May 1987
COLLATERAL BLOOD FLOW DURING OCCLUSION
Table 1. Pancreatic Blood Flow During Total Occlusion of the Superior Mesenteric Artery Region of pancreas
Control flow (mllmin . g)
SMA occluded flow a (mllmin . g)
Head Neck Tail
1.12 ± 0.29 1.16 ± 0.27 1.45 ± 0.30
0.41 ± 0.13 b 1.27 ± 0.27 2.01 ± 0.27c
SMA, superior mesenteric artery. Values are mean ± SEM. a Superior mesenteric artery occluded blood flows were measured after arterial pressure achieved a steady state (-15-20 min after total occlusion of the SMA). b P < 0.05 compared to respective control. c p = 0.07. n = 6.
Results Mean arterial pressure averaged 123 ± 6 mmHg during the control (preinjection) period. Mean arterial pressure was not altered during or after injection of the first microsphere. After the SMA was totally occluded, mean arterial pressure rose rapidly to a pressure of 171 ± 8 mmHg and then gradually declined toward the baseline pressure, plateauing at a steady state pressure of 137 ± 4 mmHg. Approximately 15-20 min were required to achieve the new steady state. After injection of the second microsphere, mean arterial pressure decreased slightly to 128 ± 7 mmHg. Renal blood flow was not affected by occlusion of the SMA. Left kidney blood flow changed from 2.08 ± 0.23 to 2.15 ± 0.19 and right kidney blood flow from 2.04 ± 0.22 to 2.09 ± 0.18 mllmin . g, respectively. The equality between left and right kidney blood flows during the control and occlusion period corroborated a homogeneous distribution of the injected microspheres. Table 1 summarizes the changes in blood flow to
1217
the pancreas (head, neck, and tail) after occlusion of the SMA. Occlusion produced a differential effect on the various regions of the pancreas. Blood flow to the head decreased significantly by an average of 63%, whereas neck blood flow was unaltered. Tail blood flow demonstrated a hyperemia (39%) that was nearly significant (p = 0.07). The changes in total wall blood flow to the small intestine after occlusion of the SMA are summarized in Table 2. Occlusion of the vessel significantly reduced duodenal, jejunal, and ileal blood flow. Proximal duodenal blood flow fell by 35%, whereas distal duodenal blood flow fell by 61 %. Blood flow to the remainder of the small bowel (jejunoileum) was reduced by an average of 71% (range 59%-76%). When the small bowel was partitioned into five equal segments (expressd as a percentage of total jejunoileallength), a remarkably consistent decrease in blood flow (mean of 71 %) was observed (Table 2). Estimates of intramural blood flow distribution (mucosa/submucosa and muscularis/serosa) revealed that the reductions in total blood flow to the intestines were due exclusively to a marked decrease in mucosa/submucosa flow (Table 2). There were no significant changes in blood flow to the muscularis/ serosa layer of the intestines. Control blood flow to the muscularis/serosa ranged from 0.043 ± 0.013 to 0.063 ± 0.020 mllmin' g in the duodenum, from 0.075 ± 0.041 to 0.170 ± 0.054 mllmin' g in the jejunoileum, and from 0.044 ± 0.011 to 0.135 ± 0.057 mllmin' g in the colon.
Discussion The results from the present study quantitate the importance of collateral blood flow to the pan-
Table 2. Intestinal Blood Flow During Total Occlusion of the Superior Mesenteric Artery Total wall blood flow
Mucosa/submucosa blood flow
Region of intestine
Control (mllmin . g)
SMA occluded a (mllmin . g)
Control (mllmin . g)
SMA occluded a (mllmin . g)
Duodenum Proximal Distal
1.10 ± 0.22 0.92 ± 0.15
0.72 ± 0.17 b 0.36 ± 0.08 b
1.50 ± 0.36 1.57 ± 0.32
1.05 ± 0.26 b 0.58 ± 0.08 b
Jejunoileum 20% total length 40% total length 60% total length 80% total length 100% total length
0.78 0.69 0.62 0.60 0.63
Colon Proximal Middle Distal
1.22 ± 0.39 1.41 ± 0.48 1.45 ± 0.43
± ± ± ± ±
0.08 0.07 0.05 0.05 0.04
0.22 0.18 0.18 0.18 0.19
± ± ± ± ±
0.02b O.Olb O.Olb 0.02b 0.02b
0.45 ± 0.15 b 0.93 ± 0.40 1.27 ± 0.42
1.19 1.13 1.11 0.89 0.95
± ± ± ± ±
0.25 0.24 0.18 0.16 0.20
1.51 ± 0.63 2.23 ± 0.78 2.33 ± 0.73
0.32 0.31 0.30 0.24 0.33
± ± ± ± ±
0.05 b 0.04b 0.05 b 0.05 b 0.07 b
0.67 ± 0.23 b 1.98 ± 0.80 2.03 ± 0.72
SMA, superior mesenteric artery. Values are mean ± SEM. a Superior mesenteric artery occluded blood flows were measured after arterial pressure achieved a steady state (-15-20 min after total occlusion of the SMA). b P < 0.05 compared to respective control. n = 6.
1218 PREMEN ET AL.
creas and intestines after occlusion of the SMA. Superior mesenteric artery occlusion exerted different qualitative and quantitative responses in the head, neck, and tail regions of the pancreas (Table 1). Blood flow to the head of the pancreas was markedly reduced, whereas blood flow in the neck region was unaffected. A hyperemia was observed in the tail region during SMA occlusion. These findings suggest that occlusion of the SMA does not compromise blood flow to the neck and tail of the pancreas (tissues that are normally perfused with blood from the celiac artery). Perfusion through collaterals maintains blood flow to the head of the pancreas to within 37% of control. Occlusion of the SMA significantly reduced duodenal blood flow. Blood flow fell by only 35% in the proximal duodenum, whereas distal duodenal blood flow fell by 61% (Table 2). It is evident that blood flow derived from the celiac artery is important in minimizing the ischemia and tissue hypoxia in the duodenum produced by SMA occlusion. It remains unclear, however, whether the blood supplied by the celiac artery is actually carried by collateral channels or represents direct nutrient flow from superior pancreaticoduodenal vessels. A similar axial gradient in tissue blood flow was observed in the colon during SMA occlusion. Occlusion significantly decreased proximal colon blood flow by 63%. Blood flow to the middle and distal colon, however, was not compromised during SMA occlusion, indicating that perfusion of these areas was largely the responsibility of the inferior mesenteric artery. As was the case in the duodenum, we are unable to discern whether the blood supplied by the inferior mesenteric artery to the proximal colon is indeed carried by collaterals. Occlusion of the SMA markedly reduced jejunal and ileal blood flow to the same extent, i.e., by -71 % (Table 2). Of particular interest is the observation that blood flow was not reduced to zero in the midsection of the jejunoileum. As this region (-50% of total length) was able to maintain its blood flow at -30% of pre occlusion values, perfusion through collaterals was, no doubt, most important. Moreover, in an isolated jejunoileum preparation (perfused by only the SMA), we observed that occlusion of the vessel reduced blood flow to zero in the small bowel. Thus, the collateral channels are apparently long enough to reach the middle of the jejunoileum. The quantitative importance of the intestinal collateral circulation in exteriorized canine jejunal loops was recently assessed in our laboratory (10,11). We found that collateral blood vessels between adjacent segments of jejunum can maintain perfusion in one segment at -55% of its control level when the single artery supplying that segment is
GASTROENTEROLOGY Vol. 92, No.5, Part 1
totally occluded. Furthermore, we (10) demonstrated that (a) virtually all of the collateral blood flow is carried by arterial channels, (b) two-thirds of the collateral flow is carried by extramural (marginal) vessels and one-third of the flow is supplied by intramural connections (e.g., the extensive submucosal vascular plexus), and (c) the major determinants of collateral blood flow in the intestine appear to be passive in nature (i.e., creation of a substantial pressure gradient between nonoccluded and occluded segments that produces a driving force for collateral flow). Extension of these observations from adjacent jejunal segments to the entire jejunoileum would suggest that occlusion of the SMA creates a large pressure gradient between vessels normally perfused by the celiac and inferior mesenteric arteries and those perfused by the SMA. The fact that we did not observe a gradient in blood flow between the ends and center of the jejunoileal segment suggests that the collateral channels are large, low-resistance vessels along which there is no substantial pressure drop. Although we are unable to define the major collateral channel from our data, the previous work by Bulkley et al. (10) suggests that most of the collateral blood flow is carried by extramural (marginal) vessels. Estimates of intramural distribution of blood flow indicate that the reductions in intestinal blood flow observed during SMA occlusion are due solely to a fall in mucosa/submucosa flow (Table 2); flow to the muscularis/serosa is not altered. This finding suggests that the collaterals supplying the mucosa offer greater resistance to blood flow than the collaterals supplying the muscularis. Alternatively, the resistance to blood flow offered by collateral channels to the mucosa and muscularis may be similar and the maintenance of muscle blood flow during SMA occlusion may result from an enhancement of visceral smooth muscle activity. We observed marked increases in intestinal motility after SMA occlusion. Chou and Grassmick (12) have suggested that an active hyperemia, similar to exercise hyperemia in skeletal muscle, occurs in the muscularis of the gut wall during tonic intestinal contractions. Thus, it is possible that an active hyperemia in the muscularis layer may have offset an occlusion-mediated decrease in blood flow as observed in the mucosa/submucosa layer.
References 1. Bonte FJ, Curry GC, Parkley RW. Experimental studies of the splanchnic circulation of the rabbit after ligation of the superior mesenteric artery. Am J RoentgenoI1969;106:691-9. 2. Meyers MH. Griffith's point: critical anastomosis at the splenic flexure. Am J Roentgenol 1976;126:77-94. 3. Michels NA, Padmanabhan S, Kornblith PL. Routes of collat-
COLLATERAL BLOOD FLOW DURING OCCLUSION
May 1987
4. 5. 6. 7. 8.
eral circulation of the gastrointestinal tract as ascertained in a dissection of 500 bodies. Int Surg 1968;49:8-28. Kittredge RD. Ischemia of the bowel. Am J Roentgenol Radium Ther Nucl Med 1968;103:400-4. Novelline RA. Waltman AC. Athanasoulis CA. Baum S. Recent advances in abdominal angiography. Adv Intern Med 1976;21:417-49. Robinson JWL. Antonio Ii JA. The intestinal response to ischemia. Pharmacology 1966;255:178-91. Todd TRJ. Weinman G. McIntyre D. Selective superior mesenteric embolization for small intestinal hemorrh'l-ge. Can J Surg 1979;22:283-6. Kyung JC. Schmidt RW. Lenz J. Effects of experimental
9. 10. 11. 12.
1219
embolization of superior mesenteric artery branch of the intestine. Invest Radiol 1979;14:207-12. Schwilden ED. Secondary circulatory disorders following rectosigmoid resection. Neth J Surg 1983;35:167-72. Bulkley GB. Womack WA. Downey JM. Kvietys PRo Granger ON. Characterization of segmental collateral blood flow in the small intestine. Am J PhysioI1985;249:G228-35. Bulkley GP. Womack WA. Downey JM. Kvietys PRo Granger ON. Collateral blood flow in segmental intestinal ischemia: effects of vasoactive agents. Surgery 1986;100:157-66. Chou CC. Grassmick B. Motility and blood flow distribution within the wall of the gastrointestinal tract. Am J Physiol 1978;235:H34-9.