British J o d of Surgery 1994,81, 259-264

Compartment syndrome in experimental chronic obstructive pancreatitis: effect of decompressing the main pancreatic duct N. D. KARANJIA, A. L. WIDDISON, F. LEUNG, C. ALVAREZ, F. J. LUTRIN and H. A. REBER

Departments of Surgery, Sepulveda Veterans’ Affairs Medical Center, Sepulveda and University of California at Los Angeles School of Medicine. Los Angeles. California, USA Correspondence to: Professor H. A . Reber, University of Cal$ornia at Los Angeles Medical Center, 10 833 L.e Conte Avenue, Los Anyeleu, California 90 024, USA

Chronic pancreatitis is characterized by persistent and severe pain, which can be relieved by decompression of the main pancreatic duct (MPD). Both ductal and interstitial pressures have been shown to be increased in chronic pancreatitis in patients. A study was carried out of pancreatic interstitial pressure and pancreatic blood flow in normal cats and those in which chronic obstructive pancreatitis had been induced 5 weeks earlier to determine the effect of decompression of the MPD. In the normal pancreas, median(interquarti1e range (i.q.r.)) basal Inter- stitial pressure was O.OS(1.2) mmHg and median(i.q.r.) basal pancreatic blood flow 583(24.3)ml per min per 100g. Secretory stimulation did not change the interstitial pressure significantly, but was associated with a 40 per cent increase in median(i.q.r.) blood flow to 81*8(458)ml per min per 1OOg. In contrast, in chronic obstructive pancreatitis, the median(i.q.r.) basal interstitial pressure was 2.0(1-5) mmHg, which was significantly higher than in the normal gland, and median(i.q.r.) pancreatic blood flow

was 38.3(9.8) ml per m h per 100 g, significantly lower than in the normal pancreas. Furthermore, secretory stimulation was assodated with a significant increase in median(i.q.r.) interstitial pressure to 3.3(1.6) mmHg and a simultaneous decrease in median(i.q.r.) blood flow to 31.5(13.7) ml per min per 100 g. After decompression of the MPD in cats with chronic obstructive pancreatitis, the median(i.q.r.) basal interstitial pressure was 2.0(1.4)mmHg and on secretory stimulation l.tS(1-5) mmHg. Decompression thus prevented the increase in interstitial pressure seen in the animals with obstruction. In contrast, ductal decompression improved the median(i.q.r.) basal pancreatic blood flow to 45.9(384) ml per min per l00g and, furthermore, this increased significantly on secretory stimulation to a median(i.q.r.) of 814(478)ml per min per 100g. Decompression thus restored the normal pattern of secretory hyperaemia. Within the conflnes of this model, these observations demonstrate that chronic obstructive pancreatitis exhibits a compartment syndrome that is relieved by duct drainage.

There is no adequate explanation for the intractable pain experienced by patients with advanced chronic pancreatitis. It is also not known why the disease often progresses, despite removal of the apparent causes. Several investigators have observed raised pancreatic ductal l v 2 and tissue3 pressures in these patients and have suggested that these are in some way responsible for the pain. Surgery has been used to relieve pain in chronic pancreatitis. Whereas some surgeons believe that resection of the parts of the pancreas most involved with disease is a more effective approach, others prefer procedures designed to drain the hypertensive ductal system. Both strategies are associated with short-term pain relief in at least 80 per cent of patients4, although the mechanism by which duct drainage alleviates pain is unclear.

In an earlier study using a model of chronic obstructive pancreatitis in cats, increases in pancreatic interstitial pressure similar to those seen clinically were observed. Under basal conditions, glands with chronic obstructive pancreatitis had normal main pancreatic duct (MPD) pressures, but parenchymal interstitial pressures were greater than in the normal pancreas’. Perfusion of the MPD, at rates mimicking secretion, caused a dramatic increase in both the MPD and parenchymal pressure’. In contrast, MPD pressure in the normal pancreas rose by only a small amount and parenchymal pressure was virtually unchanged5.

It was hypothesized that, in this model of chronic obstructive pancreatitis, the restrictive fibrotic capsule that envelopes the gland prevented dissipation of the MPD pressures generated during secretion. Thus, increased interstitial pressures developed.

Paper accepted 4 May 1993

This situation would be analogous to a compartment syndrome. Pancreatic secretion in this setting could lead to a corresponding reduction in pancreatic blood flow. Decompres- sion of the MPD might be expected to prevent the increase in main duct and interstitial pressure during secretion, and so restore blood flow. The present study was designed to test these hypotheses.

Materials and methods Adult mongrel cats of either sex and weight 2 4 kg were used in all experiments. Cats were cared for according to the guidelines established for animal care and use by the Sepulveda Veterans’ Affairs Medical Center Office of Research and Development. Before surgery, cats were fasted for 24 h and anaesthetized with intramuscular xylazine 1 mg/kg and intraperitoneal sodium pentobarbitone 25 m a g . During blood flow determinations, anaesthesia was maintained with continuous intravenous infusion of sodium pentobarbitone 2.5 mg per kg per h.

Model of chronic pancreatitis

Chronic obstructive pancreatitis was induced in cats over 5 weeks by partial occlusion of the MPD6. Anaesthetized adult cats underwent laparotomy. The main duct was isolated in the neck of the pancreas at a constant position anterior to the portal vein. The lumen of the duct was partially occluded using a 3/0 silk ligature tied around it and two strands of 310 polypropylene which were removed on completion of the tie. This consistently produced a greater than 75 per cent reduction in the normal MPD diameter. The abdomen was then closed and the animals allowed to recover. The model is shown schematically in Fig. 1. Analgesics were regularly administered immediately after surgery and later as required by an independent veterinary surgeon. After 5 weeks, histological features in the body and tail of the pancreas

259

260 N. D. KARANJIA, A. L. WIDDISON, F. LEUNG, C. ALVAREZ, F. J. LUTRINand H. A. REBER

included dense fibrosis, a pronounced chronic inflammatory cell infiltrate, some acinar destruction (Fig. 20) and the development of a thick capsule (Fig. 2h) typical of chronic pancreatitish. The unobstructed accessory lobe remained histologically normal, which probably prcvcnted the development of pancreatic insufficiency. Earlier work had shown that these histological changes are irreversible despite delayed relief of duct obstruction. Preliminary unpublished studies showed that normal pancreatic tissue had a maximal secretory rate of I ml per h per g, whereas in chronic obstructive pancreatitis rates of 0.5-07 ml per h per g were seen.

Ptrticrcriric. hlood,/korc. mcusuromen!

Pancreatic blood flow was measured using the technique of hydrogen gas clearance’ ‘ I with an intraductal electrode. This method has previously been validated in the authors’ laboratory and results correlated well (r=O.88. Pc0.001) with blood flow measurements using plabelled microspheres’.

Briefly, the concentration of hydrogen in the tissues is determined from thc current generated when molecular hydrogen is oxidized to hydrogen ions and electrons. A platinum wire catalyses the reaction and serves as an electrode to conduct the current. The magnitude of the current is directly proportional to the local concentration of hydrogen and the rate of change of hydrogen concentration in the tissues is thus proportional to local blood flow’.

Normal fasted mongrel cats were anaesthetized and placed under a heat lamp. A tracheostomy was performed and an endotracheal tube of external diameter 4.5 mm inserted. Animals were allowed to breathe

/ Main lobe

Head

Accessory lobe

spontaneously so that 3 per cent hydrogen in air could be administered when necessary. An external jugular vein and a femoral vein were cannulated for the administration of fluids and drugs. A femoral artery was cannulated for continuous blood pressure readings. A laparotomy was performed, and the MPD identified in the tail of the pancreas and cannulated with a platinum electrode (Fig. 3). The electrode was fashioned from 0127 mm platinum wire insulated by polyethylene tubing (external diameter @61 mm), with the tip of the wire bared and bent back on itself to create a loop. This loop was sited in the mid-body of the pancreas to achieve maximal stability. The cannula was tied in place by a suture at the point of entry into the duct. and the abdominal viscera were returned to the peritoneal cavity. A Ag/AgCl reference electrode was also placed in the abdomen above the dome of the liver.

Both electrodes were connected by shielded copper wire to a polarographic unit. The platinum electrode was positive and the reference electrode negative. Output from the polarographic unit was amplified and recorded graphically on a Gilson chart recorder (Gilson Medical Electronics, Middleton, Wisconsin, USA). Simultaneously, the preamplifier current was fed to the input of an analogue-digital converter. Digital output was sampled every 6 s by a microcomputer and stored on floppy disks.

A stable baseline was achieved within 10-15 min before blood flow measurement was performed Hydrogen (3 per cent) in air was administered by a rubber tube placed over the tracheostomy cannula.

-

inurn electrode

W Fig. 3 Schematic diagram showing the experimental model for the measurement of blood flow in normal cats and those with chronic obstructive pancreatitis using hydrogen gas clearance and an

Fig. 1 Schematic diagram showing placement of a constricting tie around the neck of the main pancreatic duct for the induction of chronic pancreatitis intraductal electrode

a b

Fig. 2 Photomicrographs of cat pancrcas 5 weeks after partial ligation of the main pancreatic duct. a Characteristic histological changes of chronic pancreatitis with fibrosis, chronic inflammatory infiltrate and acinar destruction. (Original magnification x 200.) b Development of a thick capsule (arrows) characteristic of chronic pancreatitis. (Original magnification x 300)

Brirish Journul of Surgery 1994, 81. 259-264

EXPERIMENTAL CHRONIC OBSTRUCTIVE PANCREATITIS- 261

The current tracing gradually rose and reached a plateau, usually within 10-15min. Hydrogen was administered for l5min and then discontinued. The current tracing then gradually fell and returned to baseline. The digitalized desaturation clearance curves were analysed by a computerized direct curve-fitting software program utilizing a Gauss-Newton non-linear least-squares regression routinelo. The half-life (tt) of the decay curve was calculated and tissue blood flow computed using monoexponential kinetics, which predict* l o that tissue blood flow is given by the relationship:

Flow (ml per s per g tissue)=ln2/tt

where rt is in seconds and In is the natural logarithm. This value was small and was multiplied by 60 and then 100 to calculate the flow in millilitres per minute per 100 g.

Pancreatic bloodflow at rest and during secretion

Three group of cats were fasted, anaesthetized and prepared for the measurement of pancreatic blood flow as described. Group 1 animals (n = 9) were normal; group 2 (n = 8) had chronic obstructive pancreatitis; group 3 (n=7) had chronic obstructive pancreatitis (induced 5 weeks previously) but immediately after cannulation of the MPD in the tail with the platinum electrode, and before any measurements were taken, the MPD was decompressed through an incision just proximal to the site of the previous silk tie by incising the anterior surface longitudinally for 5 mm.

Pancreatic blood flow was measured under basal conditions and then during maximal secretory stimulation by infusion and bolus administration of hormones. Secretin 2 units/kg and cholecystokinin octapeptide 0.08 pglkg dissolved together in 101111 0.9 mol/l NaCl were given as an intravenous infusion starting 5 min before the desaturation phase of blood flow measurement and continued for 20 min. At 2 min before the desaturation phase of blood flow measurement, an identical amount of the same hormones was administered as a bolus. Good contact was reestablished between the electrode and the duct, and the immediate fall in blood pressure was always allowed to recover before the desaturation phase, which was analysed for pancreatic blood flow. On conclusion of the blood flow studies, the body and tail of the

pancreas were resected from the line of the portal vein and weighed. The product of this weight (in grams) and tissue blood flow (millilitres per minute per 100 g), divided by 100, yielded the average total blood flow to the body and tail (millilitres per minute). The glands were examined histologically to confirm changes typical of chronic pancreatitis.

Effect of secretagogues on pancreatic interstitial pressure

This experiment studied the effect of pancreatic secretagogues on pancreatic interstitial pressure in three groups of cats. Group 4 animals (n = 6) were normal; group 5 (n = 7) had chronic obstructive pancreatitis; group 6 (n=7) had chronic obstructive pancreatitis (induced 5 weeks previously) but, immediately before any measurements were taken, the MPD was decompressed as described. At the conclusion of the study in group 5, the pancreas was allowed to recover for 1 h; the MPD was then drained and interstitial pressure measured as in group 6.

All cats were anaesthetized as described above and venous access achieved through an external jugular vein. A tracheostomy was fashioned and the animals allowed to breathe spontaneously. An upper midline laparotomy was performed and the spleen, pancreas and duodenum were delivered into the wound.

Pancreatic interstitial pressure was measured using a direct- puncture technique. A 2 2 4 hypodermic needle stripped of its plastic Luer hub was connected through a short length of plastic tubing (external diameter 0-97mm) to a pressure transducer and the entire system filled with heparinized saline. The transducer was zeroed with the needle open to air in the mid-axillary line. The saline-primed needle was inserted into the body of the pancreas almost parallel to the surface of the gland so that its tip was surrounded by parenchyma In this position the pressure reading stabilized over 15-30 s; at the end of this time the average of several values was recorded as the basal interstitial pressure. The pancreas was then maximally stimulated with an intravenous bolus of secretin 2 units/kg and cholecystokinin 0.08 pghg and at the same time an infusion of secretin 2 units/kg and cholecystokinin 0.08 pg/kg mixed and dissolved in lOml 0 9 mol/l

NaCl was started and run over 20 min. Readings of interstitial pressure were recorded every 5 min over the next 15 min. To achieve uniformity the average 10-min value was compared with the basal value. On conclusion of these studies the pancreas was resected for histological examination.

Statisrical analysis

Pancreatic blood flow was expressed as tissue blood flow in millilitres per minute per 100s These data were not normally distributed and were therefore analysed by non-parametric methods. Results are expressed as median(interquart1e range (i.q.r.)). Original data were analysed by the Wilcoxon test for matched pairs for changes within a group and the Mann-Whitney U test for differences between groups after preliminary analysis of variance. Nominal data were analysed by Fisher’s exact test. Significance was assumed at P<0.05.

Results Pancreatic bloodflow at rest and during secretion In group 1, the median(i.q.r.) pancreatic blood flow to the body of the pancreas in nine normal cats was 58.3(24.3)ml per min per 100 g (Fig. 4). Stimulation with secretagogues significantly increased the value to 81-8(45.8)ml per min per lOOg (an increase was seen in eight of nine cats). The median(i.q.r.) weight of the body and tail of the pancreas in group 1 was 4.7(0*4)g. The median(i.q.r.) basal blood pressure, 1065(18-8) mmHg, was not significantly changed by secretory stimulation.

In group 2, the median(i.q.r.) pancreatic blood flow to the body of the pancreas in eight cats with chronic obstructive pancreatitis was 38.3(9.8) ml per min per 100 g. This basal flow was significantly lower than that seen in group 1. Stimulation with secretagogues significantly decreased the value further to 31q13-7) ml per min per 100 g ( P = 0 0 2 ) (a decrease was seen in seven of eight cats). The median(i.q.r.) weight of the diseased body and tail of the pancreas in animals from group 2 was 25(Q9)g and this was significantly less than that Seen in

t

T i

Normal Obstructive Acute pancreatitis decompression

Fig. 4 Median (interquartile range) pancreatic blood flow at rest (0) and after secretagogue stimulation (m) in the normal pancreas, chronic obstructive pancreatitis and after acute decompression of the main pancreatic duct in chronic obstructive pancreatitis. *P<0.05 (versus normal, Mann-Whitney U test); tP<O-OS (versus basal, Wilcoxon test)

British Journal of Surgery 1994.81, 259-264

262 N . D. KARANJIA, A. L. W I D D I S O N , F. L E U N G , C. ALVAREZ. F. J . L U T R I N and H. A. REBER

animals from group 1. The median(i.q.r.) basal blood pressure, 953( 1 1.2) mmHg, was not significantly different from that in normal cats and was not changed by secretagogue stimulation.

Whereas administration of secretagogues caused an increaw in pancreatic blood flow in eight of nine normal glands, it caused a decrease in seven of eight cats with chronic obstructive pancreatitis. These differences in flow after stimulation were significant (P< 005, Fisher’s exact test).

The median(i.q.r.) basal pancreatic blood flow to the body of the pancreas in seven cats with chronic obstructive pancreatitis and MPD decompression was 45.9(38.4) ml per min per 1OOg. This was not significantly different from that in the cats with chronic pancreatitis and obstructed ducts. Stimulation with secretagogues significantly increased the median(i.q.r.) blood flow to 814(47.8)ml per min per lOOg (P=003) (an increase was seen in six of seven cats). The median(i.q.r.) basal blood pressure. 9&3( 19.2) mmHg, was not significantly different from that found for animals in groups 1 and 2, and this remained unchanged during secretory stimulation.

It is of interest that secretagogues caused an increase in pancreatic blood flow in six of seven cats with chronic obstructive pancreatitis whose MPD had been decompressed, whereas they caused a decrease in seven of eight cats with undrained chronic obstructive pancreatitis. These differences were significant (P <005, Fisher’s exact test).

Eflcct of secrttagogues on pancreatic interstitial pressure

The median(i.q.r.) basal interstitial pressure in the body of the pancreas of normal cats (group 4) was 0.05(1-2)mmHg (Fig. 5). After secretory stimulation it was 0.3q1.3) mmHg, which was not significantly different. Of six cats, a small increase in interstitial pressure was registered in four while two had a decrease.

In group 5 cats with chronic obstructive pancreatitis and an obstructed MPD. median(i.q.r.) basal interstitial pressure was 2.q1.5)mmHg in the body of the gland. Secretory stimulation increased interstitial pressure in all cats. A

E 4 E s! a z 3 -

g - Q .- .- c 2 ’

t I,

Q C - .- .- 0 ’

e = o L

c Q

0

t

T

Normal Acute panamitis decompression

Fig. 5 Median (interquartile range) pancreatic interstitial pressure at rest (0) and after secretagogue stimulation (I) in the normal pancreas. chronic obstructive pancreatitis and after acute decompression or the main pancreatic duct in chronic obstructive pancreatitis. *P<0.05 (versus normal, Mann-Whitney U test); +P<0.05 (r-~rsus basal, Wilcoxon test)

median(i.q.r.) value of 3.3(1.6) mmHg was measured (P=002 versus basal); an increase was registered in all seven cats in this group.

In group 6 the median(i.q.r.) basal interstitial pressure after ductal decompression was 2.q1.4) mmHg, which was the same as the basal value in the body of the pancreas in animals without decompression. Secretory stimulation of the decompressed gland was not associated with a significant change in interstitial pressure in the body of the pancreas. Three of seven cats showed a small decrease in pressure and in four there was no change when the gland was stimulated to secrete. Similar results were seen when the group 5 cats had their ducts drained and interstitial pressure measured as described above.

Discussion Duct drainage procedures for chronic pancreatitis - caudal or lateral pancreaticojejunostomy - are associated with rates of immediate pain relief equivalent to re~ection’~-’~. Proponents claim that this is the result of the relief of duct obstruction, and of duct and parenchymal hypertension. However, the mechanism by which increased duct and tissue pressures produce. pain has not been defined. Earlier work has shown a direct relationship between the high duct and interstitial pressures generated in this model of chronic obstructive pancreatitis and that this is the result of a thick fibrous capsule encasing the chronically inflamed pancreas’. Nealon et all4 showed that early duct drainage in human chronic pancreatitis delayed functional impairment and prevented progression of the severity of the disease.

Results with the present model have shown that experimental chronic obstructive pancreatitis was associated with reduced basal blood flow compared with the normal pancreas. Furthermore, the basal interstitial pressure was higher in the obstructed diseased gland and rose when the pancreas was stimulated to secrete. Relief of the ductal obstruction produced an immediate improvement in blood flow and reversed the blood flow response to secretagogue stimulation. At the same time, it prevented the increase in interstitial pressure during stimulation. These features are consistent with a compartment syndrome.

The present animal model of chronic obstructive pancreatitis produces histological and functional changes after 5 weeks of duct obstruction comparable to the clinical form of the disease”. Nevertheless, it is difficult to extrapolate these results to humans for several reasons. The pancreatitis was artificially induced and a relatively early point in the course of the disease at which to make measurements was chosen. Chronic obstructive pancreatitis is relatively uncommon. Most indi- viduals develop chronic pancreatitis after long-term abuse of ethanolI6 but, unfortunately, a suitable experimental model of alcoholic pancreatitis is not available. The relevance of the obstructive model to the alcoholic form of pancreatitis is uncertain. Morphologically, the two forms have some differences in the early stagesI6 I * . However, in chronic alcoholic pancreatitis, stricture of the ducts and the plugging of ductules by protein plugs is eventually a major contributor towards disease progression.

Pancreatic blood flow depends on a number of regulatory factors. As blood flow is the mean arterial pressureportal pressure difference divided by vascular resistance, provided that the arteriovenous pressure gradient remains unchanged pancreatic blood flow is inversely related to pancreatic vascular resistance.

Under normal conditions, arteriolar tone and precapillary sphincters are the major determinants of tissue perfusion. These

British Journal of Surgery 1994.81, 259-264

EXPERIMENTAL CHRONIC OBSTRUCTIVE PANCREATITIS 263

factors establish the vascular resistance of the normal pancreas and regulate the number of capillaries p e ~ f u s e d ' ~ * ~ ~ . In the normal pancreas after feeding or the exogenous administration of secretagogues, a hyperaemic response is brought about by the release of metabolic by-products, as well as by hormonal and neural i n f l ~ e n c e s ' ~ * ~ ~ . This hyperaemia is effected by an increase in the number of perfused capillaries in association with a decrease in the vascular re~istance'~.

The present results show that basal blood flow is reduced in chronic pancreatitis compared with that in the normal pancreas. There was no difference in mean arterial pressure between normal cats and those with chronic pancreatitis. The authors have previously noted no difference in portal vein pressure between similarly treated groups of animals22. This suggests that there is an increase in the basal vascular resistance in chronic obstructive pancreatitis.

The mechanism responsible for this increased resistance is unclear. The exocrine pancreas is functionally and morpho- logically similar to the salivary glands. It has been shown that obstruction of the main duct draining the submandibular gland leads to fibrosis and atrophy similar to that observed in the pancreas in the present model. After a few weeks this atrophy was associated with a substantial reduction in the number of capillaries feeding the submandibular glandz3. A similar reduction in the capillary surface area in the obstructed pancreas could explain the increased resistance. However, such a fixed decrease in the number of vessels cannot be reconciled with the increase in flow observed after decompression and the accompanying restoration of the hyperaemic response to secretory stimulation. Other factors must be at work.

It seems likely that the higher basal interstitial pressures in the pancreata of cats with chronic obstructive pancreatitis could also increase the vascular resistance, thus accounting for a lower basal blood flow. During secretion the elastic parenchyma of the normal pancreas is capable of dissipating the increased intraductal volume and intraductal pressure, which develops without much change in parenchymal interstitial pressure. In contrast, in chronic obstructive pancreatitis, the parenchymal fibrosis and restrictive fibrotic capsule that envelop the pancreas as the disease progresses limit the distensibility of the gland'. This could account for the significant increase in interstitial pressure seen during secretion in chronic pancreatitis, which in turn results in a concomitant increase in vascular resistance and therefore a further decrease in blood flow.

Thin capillary walls offer little resistance to extrinsic compression. The capillary perfusion pressure in the pancreas is estimated to be only l2mmHg at normal portal pressure^^^*^^. Thus even small increments in the external pressure surrounding pancreatic capillaries, for example caused by increased interstitial pressure, would significantly decrease blood flow. This is particularly likely in clinical settings where pancreatic tissue pressures as high as 40 mmHg have been measured in patients with chronic pancreatitis'.

Decompression of the MPD had no significant effect on the basal blood flow and interstitial pressure. This suggests that some changes in the tissue had already developed that prevented a larger change. Although this was not studied, a decrease in the number of capillaries available for perfusion, analogous to the changes in salivary glands already discussed, might have occurred. When pancreatic secretion was stimulated, however, there was a marked increase in blood flow in the decompressed glands. This suggests that additional capillaries were available for perfusion and/or that the flow rate increased through capillaries already being perfused. The fact that interstitial pressure no longer rose supports the notion that the maintenance of low vascular resistance in the gland allowed the increase in flow to occur.

We conclude from interpretation of the inverse relationship between blood flow and tissue pressure that a compartment syndrome exists. Secretion in this model of chronic obstructive pancreatitis is accompanied by ischaemia, and ductal drainage prevents this ischaemia by relieving the underlying compartment syndrome. These studies may offer a new perspective on the underlying pathophysiology of human chronic pancreatitis. It is possible that the pain accompanying chronic pancreatitis could, at least in part, have an ischaemic basis. This could explain the characteristic pattern of pain often being exacerbated by eating, which stimulates pancreatic secretion and therefore metabolic demand. This could also be the reason why pain disappears spontaneously in many patients as pancreatic function deteriorates; presumably so too would the metabolic requirements of the gland. Finally, these observations may explain the pain relief experienced by patients undergoing drainage procedures, thus providing a scientific rationale for the hitherto empirical operation of lateral pancreaticojejunostomy (Puestow procedure). Further detailed studies in patients are necessary to confirm these suggestions.

Acknowledgements The authors thank Mrs Hermione Kavanagh and Mrs Paula Berger of the Sepulveda Veterans' Affairs Medical Center and Mr Kenneth Cooper for assistance with the illustrations. This work was supported by the Medical Research Service of the US Department of Veterans' Affairs.

References 1 Madsen P, Winkler K. The intraductal pancreatic pressure in

chronic obstructive pancreatitis. Scand J Gastroenterol 1982; 17: 5 5 3 4 .

2 Bradley EL. Pancreatic duct pressure in chronic pancreatitis. Am J Surg 1982; 144: 313-15.

3 Ebbehoj N, Borly L, Madsen P, Svendsen LB. Pancreatic tissue pressure and pain in chronic pancreatitis. Pancreas 1986; 1: 556-8.

4 Alvarez C, Widdison AL, Reber HA. New perspectives in the surgical management of chronic pancreatitis. Pancreas 1991; 6 S76-81.

5 Karanjia ND, Singh SM, Widdison AL, Lutrin FJ, Reber HA. Pancreatic ductal and interstitial pressures in cats with chronic pancreatitis. Dig Dis Sci 1992; 3 7 268-73.

6 Austin JL, Roberts C, Rosenholtz MJ, Reber HA. Effects of partial duct obstruction and drainage on pancreatic function. J Surg Res 1980; 28: 426-32.

7 Reber HA, Karanjia ND, Alvarez C er d. Pancreatic b l d flow in cats with chronic pancreatitis. Gastroenterology 1992; 103: 652-9.

8 Aukland K, Bower BF, Berliner RW. Measurement of local blood flow with hydrogen gas. Circ Res 1964,14 164-87.

9 h u n g FW, Guth PH, Scremin OU, Golanska EM, Kauffman GL. Regional gastric mucosal blood flow measurements by hydrogen gas clearance in the anaesthetized rat and rabbit. Gastroenterology

10 Livingston EH, Reedy T, hung FW, Guth PH. Computerized curve fitting in the analysis of hydrogen gas clearance. A m J Physiol1989; 257 G668-75.

1 1 Sheperd AP, Kiel JW. Gastrointestinal blood flow-measuring techniques. In: Schultz SG, Wood JD, Rauner 8, eds. Handbook of Physiology. Section 6: The Gastrointestinal System. Vol. I . Motility and Circulation. Baltimore, Maryland Waverly Press, 1989 1335-70.

12 Prim RA, Greenlee HB. Pancreatic duct drainage in 100 patients with chronic pancreatitis. Ann Sure 1981; 194: 313-20..

13 Bradley EL. Long term results ofpancreaticojejunostomy in patients with chronic pancreatitis. A m J Surg 1987; 153 207-13.

14 Nealon WM, Townsend CM, Thompson JC. Operative drainage of the pancreatic duct delays functional impairment in patients with chronic pancreatitis. Ann Surg 198R.208: 321-9.

i9m,a7: 28-36.

British Journal of Surgery 1994, 81, 259-264

264 N. D. KARANJIA, A. L. WIDDISON, F. LEUNG, C. ALVAREZ, F. J. LUTRIN and H. A. REBER

I5 Singh SM. Reber HA. The pathology of chronic pancreatitis. World J Surg 1990, 14 2-10.

16 Reber H . Chronic pancreatitis: etiology, diagnosis and pathology. In: Howard JM, Jordan GL. Reber HA, eds. Surgical Diseases of the Puwreas. Philadelphia, Pennsylvania: Lea and Febiger, 1987: 475-95.

17 Sarles H. Chronic calcifying pancreatitis - chronic alcoholic pancreatitis. Gaslrfic.rrterology 1974; 66: 604-16.

18 Sarles H . Sahel J. Pathology of chronic calcifying pancreatitis. Am J Gu.stroeritcro1 1976; 66: 117-39.

19 Kvietys PR, Granger ND, Harper SL. Circulation of the pancreas and salivary glands. In: Schultz SG, Wood JD, Rauner B, eds. Hanclhook qf Physiology. Section 6: The Gastrointestinal System. Vd. 1. Motility and Circulation. Baltimore, Maryland Waverly

Press, 1989 1565-95. 20 Kvietys PR, McLendon JM, Bulkley GB, Perry MA, Granger DN.

Pancreatic circulation: intrinsic regulation. Am J Physiol1982; 242 G596-602.

21 Heisig N. Pancreatic microcirculation under the influence of adequate secretory stimulation. Bib1 Anat 1967; 9: 17680.

22 Widdison AL, Alvarez C, Reber HA. Effects ofethanol on pancreatic blood flow in chronic pancreatitis. Surgery 1992; 112 202-10.

23 Clough G, Smaje LH. Exchange area and surface properties of the microvasculature of the rabbit submandibular gland following duct ligation. J Physiol (Lon4 1984; 354: 445-56.

24 Granger ND, Barrowman JA. Microcirculation of the alimentary tract. I. Physiology of transcapillary fluid and solute exchange. Gastroenterology 1983; 84: 84668.

British Journal of Surgery 1994. 81, 259-264