treatment vagotomy, gastrectomy, vagotomysummary duodenogastric reflux of bile acids and...

Gut, 1982, 23, 569-577

Bile acid and lysolecithin concentrations in the stomachin patients with duodenal ulcer before operation andafter treatment by highly selective vagotomy, partialgastrectomy, or truncal vagotomy and drainageP DEWAR,* R KING, and D JOHNSTON

From the University Department ofSurgery, The General Infirmary, Leeds

SUMMARY Duodenogastric reflux of bile acids and lysolecithin in the course of a standard testmeal was measured in normal people and in patients with duodenal ulcer before operation andmore than one year after highly selective vagotomy, Polya partial gastrectomy, truncal vagotomyand pyloroplasty, and truncal vagotomy and gastrojejunostomy. Before operation, duodenalulcer patients had significantly higher fasting, post-prandial, and peak bile acid concentrations inthe stomach than had normal subjects. After Polya partial gastrectomy, fasting, post-prandial,and peak concentrations of bile acids and lysolecithin were significantly higher than inpreoperative duodenal ulcer patients. After highly selective vagotomy, in contrast, bile acidconcentrations in the stomach were significantly lower than in preoperative duodenal ulcerpatients and post-prandial and peak lysolecithin concentrations were less than half (NS) thoserecorded in preoperative duodenal ulcer patients. After highly selective vagotomy, bile acidconcentrations were also significantly lower than bile acid concentrations after Polya partialgastrectomy, truncal vagotomy and pyloroplasty, and truncal vagotomy and gastrojejunostomy;and post-prandial and peak lysolecithin concentrations were significantly lower than after Polyapartial gastrectomy and truncal vagotomy and gastrojejunostomy. Thus, when used in thetreatment of patients with duodenal ulcer, highly selective vagotomy keeps 'bile' out of thestomach, probably through its effect on gastric smooth muscle, combined with the preservation ofan intact antropyloroduodenal segment. In contrast, Polya partial gastrectomy, truncal vagotomyand gastrojejunostomy, and truncal vagotom' and pyloroplasty all lead to a significant increase inreflux of bile acids and lysolecithin into the stomach. The clinical importance of these findings isthat both gastritis and, in the long term, gastric carcinoma may prove to be less common afterhighly selective vagotomy than after partial gastrectomy or vagotomy with a drainage procedure.

Traditional operative procedures for duodenal ulcersuch as partial gastrectomy and vagotomy with adrainage procedure are known to increase duodeno-gastric reflux.1 Duodenal content in the stomachbreaks the gastric mucosal barrier by its detergentaction and, with or without the aid of acid, pepsin,and agents such as aspirin or alcohol, produces achronic gastric mucosal reaction or gastritis.>4 Such'bile reflux gastritis' may produce symptoms such as

' Address for correspondence: Mr E P Dewar, Department of Surgery, RoyalNaval Hospital, Haslar, Gosport, Hants P012 2AA.

Received for publication 19 October 1981

epigastric pain, nausea, vomiting, and loss ofweight5-7 and, in the long term, there is anincreased risk of gastric carcinoma. 10

Highly selective vagotomy, unlike these pro-cedures, preserves the pylorus" and might thus beexpected not to produce an increase in duodeno-gastric reflux. In this paper we show that, afterhighly selective vagotomy, reflux of bile acids 'andlysolecithin into the stomach is, in fact, significantlyless than in patients with duodenal ulcer beforeoperation. As expected, both partial gastrectomyand truncal vagotomy with a drainage procedurewere followed by significant increases in duodeno-gastric reflux.

569

on July 31, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.23.7.569 on 1 July 1982. Dow

nloaded from

http://gut.bmj.com/

Dewar, King, and Johnston

Methods

PATIENTSSix groups of patients were studied (Table 1). Onegroup consisted of 16 patients with active duodenalulceration who had not been operated upon. Asecond control group comprised nine normalsubjects with no history of disease of the gastro-intestinal tract. The other four groups consisted ofpatients who had been treated surgically forduodenal ulcer more than one year previously: 14 byhighly selective vagotomy, 11 by Polya partialgastrectomy, nine by truncal vagotomy andHeineke-Mikulicz pyloroplasty, and eight by truncalvagotomy and gastrojejunostomy. The groups didnot differ significantly with respect to age and sex.

The mean interval between operation and thesestudies was four years in the highly selectivevagotomy group, eight years in the truncalvagotomy and drainage groups, and 20 years in thePolya partial gastrectomy group. The results ofoperation were good (Visick grades 1 or 2)12 in allthese postoperative patients. The diagnosis ofduodenal ulceration had been made in each patientby barium meal examination and/or fibreoptic endo-scopy and had been confirmed at operation in eachpatient who had undergone surgery. The nature andpurpose of the investigations were carefullyexplained to each patient and his or her consentobtained.

FIBREOPTIC OESOPHAGOGASTRODUODENOSCOPYEach patient underwent this procedure, which was

performed for diagnostic purposes in the pre-operative duodenal ulcer patients and in the post-operative patients both to confirm that the ulcer hadhealed and to exclude the presence of any otherpathological condition in the stomach andduodenum. At endoscopy the presence or absenceof bile reflux was recorded and multiple gastric andduodenal biopsies were taken in each case to formthe basis of a separate report on the morphologicalstate of the gastric mucosa in relation to 'bile' reflux.

There was no evidence of gall bladder disease in any

patient, either on oral cholecystography or atoperation.

TEST MEAL AND SAMPLING TECHNIQUEAfter the patient had fasted overnight, a size 12nasogastric tube was positioned in the stomach usingthe water recovery method. 13 Resting juice was

aspirated and discarded. Three 8 ml samples ofgastric juice were collected by gentle suction with a

syringe at 20 minute intervals during the 'basal'hour. A standard Lundh test meal was then given. Itconsisted of 300 ml of water containing 13.5 g

protein, 18 g corn oil, and 53 g carbohydrate, andwas ingested over a five minute period. A furtherfour samples of gastric content were then collectedat 20 minute intervals. The patient remained seatedin a chair throughout the period of study. Theprocedure was well tolerated and none of thepatients complained of nausea during the test.The samples of gastric aspirate were placed in

tubes containing 0O002 M ethylenediaminetetraceticacid and 0.005 M calcium chloride to inhibitphospholipase A activity. This prevented in vitrohydrolysis of lecithin to lysolecithin, a processcatalysed by phospholipase A and activated by bileacids and trypsin.

MEASUREMENTS OF BILE ACIDSConcentrations of total bile acids were measuredusing a commercial kit (Sterognost 3 c R_Pho.Nyegaard).15 The method depends on the activity ofa non-specific 3oc-hydroxysteroid dehydrogenaselinked to the reduction of nicotinamide adeninedinucleotide (NAD+). All 3oc-hydroxysteroid bileacids, both free and conjugated, are converted tothe corresponding 3-ketosteroids with the formationof nicotinamide adenine dinucleotide reduced(NADH).

Determinations were made in duplicate in a

medium containing 0.0003 M NAD+, 0.082 Mhydrazine hydrate buffer, pH 9.5 and 20 gI of gastricaspirate in a final volume of 1.52 ml. The reduction

Table 1 Details ofpatients

Length offollow-up (yr)Age (yr) after operation

SexGroup n Mean Range (male) Mean Range

Control 9 46-4 33-74 6DU 16 42-0 22-58 11 -

HSV 14 50-1 30-68 11 4-4 1-10PG 11 55.2 35-80 8 20-4 3-33TV+P 9 50-5 33-70 6 7-7 3-13TV+GJ 8 57.0 44-74 6 8-4 3-13

570


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/

Bile acid and Iysolecithin concentrations in stomach

of NAD+ was followed at 340 nm using a PyeUnicam SP 1800 dual beam spectrophotometer. Thesmallest detectable difference in bile acid concentra-tion was governed by the sensitivity of the spectro-photometer. For the instrument used this was

equivalent to 10 ,mmol/l of bile acid. The betweenassay coefficient of variation using this techniquewas 5%.

Bile acid concentrations in each 20 minute periodwere used to calculate the mean bile acid concentra-tions (1) in the fasting state for one hour, and (2)after the meal for 80 minutes. These values were

termed the mean fasting concentration and themean post-prandial concentration, respectively. Themean peak concentration was calculated from thehighest concentration of bile acid recorded in eachpatient.

MEASUREMENT OF LYSOLECITHINPhospholipids were extracted from 2 ml of aspirateby the method of Bligh and Dyer.16 Samples of thisextract (4 ml) were evaporated under nitrogen,redissolved in a small quantity of chloroform, andapplied to thin layer chromatographic plates of silicagel H (thickness 0.5 mm) together with standards oflysolecithin. Plates were developed in chloroform/methanol/water (65:25:4 by vol)17 and the separatelipid spots demonstrated by iodine vapour. Samplespots containing lysolecithin were scraped from theplate and assayed for phosphate after degradation. 18To each sample of silica gel was added 1. 1 ml 11.7Mperchloric acid. This mixture was heated on asandbath and maintained at a temperature ofapproximately 230°C to ensure digestion of the lipid.After cooling, the original volume of perchloric acid

was restored by weight and the liberated phosphatewas assayed using a colorimetric technique.19

Standards of free phosphate were used in thecalculation of the amount of phosphate liberated byacid degradation but the efficiency of extraction oflysolecithin was determined by the addition ofknown quantities of pure lysolecithin to both lowrange and high range patient samples. In this way itwas possible to quantify the loss of lysolecithin at allstages of the extraction. The overall recovery was

found to be 45±5% (mean± 1 SD) and this includedthe losses on extraction, evaporation, transfer, thinlayer chromatography, degradation, and assay. Thefasting, post-prandial, and peak values of lyso-lecithin were calculated as for the bile acids.

STATISTICAL METHODSThe values reported represent the means of dupli-cate determinations on each sample plus or minusone standard error of the mean. The significance ofdifferences between the groups was determined bymeans of the Wilcoxon rank sum test.20

Results

BILE ACID CONCENTRATIONS IN STOMACHBile acid concentrations in the stomach in the sixgroups of patients are shown in Figs. 1 and 2 and inTable 2.

Patients with duodenal ulcer, before operation,had significantly higher fasting, post-prandial, andpeak bile acid concentrations in the stomach thandid normal subjects.

In patients who had undergone highly selectivevagotomy fasting, post-prandial, and peak bile acid

Table 2 Bile acid concentrations (,umol 1-')

Significance Post- Significance SignificanceGroup Fasting ofdifference prandial ofdifference Peak of difference

Control 11 51 112p<0-OOl p<0-002 p<0-001

DU 3811 514 J 906p.p<O-Ol01 p<O-O1 p<O-O1

HSV 45 86 196

p<O-OOOl 4 p<O-ooo1 p<O-looPG 4880< 8451 12 464 Jp<OO1 p<0-000 p<b-O

TV+P 4541 507 1298NS p<0.001 p<0-01

TV+GJ 424 3647 6986HSV vs control p<0-65 NS NSHSV vs TV+P p<0-01 p<0-05 p<0.05HSV vs TV+GJ p<0-01 p<0-0001 p<0-0001PG vs DU p<0-001 p<0-0001 p<0-0001PG vs TV+GJ p<0-01 p<0-05 p<0.05

Mean fasting, post-prandial, and peak values for each group of patients. The standard errors of the means are shown in Figs. 1 and 2.

571


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/


Eg 700D

- 600+1c

E500

E 400-C

.° 300-

C

8 200-c

0

100.

U)S

T T

20 40 60 80 1io iio 140(minutes)

Fig. 1 Bile acid concentrations in the stomach in thefasting state and after a Lundh test meal, in nine normalsubjects (control) (0), 16 preoperative duodenal ulcerpatients (0), and 14 patients after highly selective vagotomy(A).

concentrations were significantly lower than those induodenal ulcer patients (Fig. 1, Table 2). Post-prandial and peak bile acid concentrations afterhighly selective vagotomy were similar to those innormal controls (Fig. 1), but in the fasting state bileacid concentrations of highly selective vagotomypatients were higher than in controls. Bile acid

12000 Lundh meal120000

+1c

SlE 80001

E I

c 4000-

, 2000 /

(minutes)Fig. 2 Bile acid concentrations in the stomach in thefasting state and after a Lundh test meal, in 14 duodenalulcer patients treated by highly selective vagotomy (A), 11by Polya partial gastrectomy (0), nine by truncal vagotomyand pyloroplasty (E), and eight by truncal vagotomy andgastrojejunostomy (A). Note change in scalefrom Fig. 1.

concentrations in the stomach were significantly lessafter highly selective vagotomy than after truncalvagotomy and pyloroplasty or truncal vagotomy andgastrojejunostomy.

After Polya partial gastrectomy fasting, post-prandial, and peak bile acid concentrations in thestomach were significantly greater than those ofcontrols and duodenal ulcer patients before opera-tion, and also significantly greater than those inpatients treated by highly selective vagotomy. Bileacid concentrations in the stomach after Polyapartial gastrectomy were significantly greater thanthose after truncal vagotomy and pyloroplasty andtruncal vagotomy and gastrojejunostomy. Bile acidconcentrations in the stomach of fasting patientswere relatively low after truncal vagotomy andpyloroplasty and truncal vagotomy and gastro-jejunostomy and high after Polya partialgastrectomy. After the meal, bile acid concentra-tions were highest after Polya partial gastrectomy,very high after truncal vagotomy and gastro-jejunostomy, and high after truncal vagotomy andpyloroplasty. Duodenogastric reflux was much lessafter truncal vagotomy and pyloroplasty than aftertruncal vagotomy and gastrojejunostomy or Polyapartial gastrectomy (Fig. 2).

LYSOLECITHIN CONCENTRATIONS IN STOMACHLysolecithin concentrations in the stomach in the sixgroups of patients are shown in Figs. 3 and 4 and inTable 3.

150-ED

+1

0E

100-

E

._

* 50-p

o 40-.' 30-* 20-o5 10-J

Lundh m

I

20 40 60 80(minutes)

100 120 140

Fig. 3 Lysolecithin concentrations in the stomach in thefasting state and after a Lundh test meal, in nine normalsubjects (control) (0), 16 preoperative duodenal ulcerpatients (0), and 14 patients after highly selective vagotomy(A).

* * . . .-

572


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/


Table 3 Lysolecithin concentrations (,mol 1')

Significance Post- Significance SignificanceGroup Fasting of difference prandial ofdifference Peak ofdifference

Control 29 22 47NS NS NS

DU 25 62 102NS NS NS

HSV 22 22 42-p<0-OOl p<O-OOOl p<0-OOOl

PG 270 627 1140p<0-0l p<O-OOl p<0-OOl

TV+P 30 45 90JNS p<0-01 p<0-01

TV+GJ 22 282 605]HSV vs TV+GJ NS p<0-001 p<0-001PG vs control p<0-01 p<0001 p<0-0001PG vs DU p<0-001 p<0-0001 p<0-0001PG vs TV+GJ p<0-01 p<0-05 p<0.05

Mean fasting, post-prandial, and peak values for each group of patients. Standard errors are shown in Figs. 3 and 4.

Both in the fasting state and after the meal,lysolecithin concentrations in highly selectivevagotomy patients were very similar to those innormal subjects (Fig. 3).

In duodenal ulcer patie-ts before operation,lysolecithin concentrations in the fasting state weresimilar to those of normal subjects and highlyselective vagotomy patients, but post-prandial andpeak levels were more than double those in thesetwo groups (Fig. 3, Table 3).

(nimtes)

Fig. 4 Lysolecithin concentrations in the stomach in thefasting state and after a Lundh test meal, in 14 patients afterhighly selective vagotomy (A), 11 Polya partial gastrectomy(0), nine truncal vagotomy andpyloroplasty (a), and eighttruncal vagotomy and gastrojejunostomy (A).

When the four postoperative groups of patientsare considered (Fig. 4), lysolecithin concentrationswere low in the fasting state in the highly selectivevagotomy, truncal vagotomy and pyloroplasty, andtruncal vagotomy and gastrojejunostomy groups,and no significant differences were found betweenthese groups. The fasting lysolecithin concentrationsin these three groups of patients were all signifi-cantly lower than the fasting lysolecithin concentra-tions in the Polya partial gastrectomy group.After the meal, lysolecithin concentrations

remained relatively low after highly selectivevagotomy and truncal vagotomy and pyloroplasty,but were greatly and significantly raised after bothPolya partial gastrectomy and truncal vagotomy andgastrojejunostomy (Fig. 4).

After Polya partial gastrectomy lysolecithin con-centrations (fasting, post-prandial, and peak) wereall significantly greater than lysolecithin concentra-tions after both truncal vagotomy and pyloroplastyand truncal vagotomy and gastrojejunostomy.

Discussion

The method used to obtain samples of gastriccontent could be criticised on the grounds that thepresence of a nasogastric tube may produce nauseaand, in itself, influence duodenogastric reflux. Thesame conditions prevailed, however, in all subjectsand the differences observed between groups werein the main so great that it seems unlikely that thepresence of the tube had much influence on theresults. Manifestly, the nasogastric tube did notcause much reflux in the control or in the highlyselective vagotomy groups. The use of radio-isotopes and functional scintigraphy21 has recently

573


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/


been introduced to demonstrate duodenogastricreflux of bile. Such a technique, however, does notat present permit accurate measurement of refluxand its use is thus complementary to the method thatwe used, rather than a substitute for it.The findings of very low bile acid concentrations

in the stomach in the control group are in agreementwith the findings of Rhodes et a122 and Fisher andCohen.23 They again show the value of a normalantropyloroduodenal segment in preventing refluxof 'bile' into the stomach.24 The preoperativeduodenal ulcer patients had significantly higher bileacid concentrations in the stomach than thecontrols, both in the fasting state and after a meal.These results confirm, and give quantitative expres-sion to, the previous observations of Capper and hiscolleagues25 that patients with duodenal ulcer showevidence of duodenogastric reflux more often thancontrol subjects.

After highly selective vagotomy, relatively lowconcentrations of bile acid were found in thestomach, only the fasting values being significantlygreater than those of normal subjects. All threevalues of bile acid concentration (fasting, post-prandial, and peak) were significantly lower afterhighly selective vagotomy than in duodenal ulcerpatients before operation (Table 2), and, as the twogroup of patients were well matched, the differenceis almost certainly a real one. This finding is at firstsight surprising and pre and postoperative studies onthe same patients will be required to confirm thispoint.The mechanism of duodenogastric reflux is not

fully understood, but it may involve abnormalities ofmotility and pressure in the antropyloroduodenalcomplex.23 Pressure increases in the pyloricsphincter of normal people on response to hormonalstimulation by secretin and cholecystokinin, butthere is no response in patients with gastric ulcer.Similarly, duodenal acidification reduces duodeno-gastric reflux by increasing pyloric sphincterpressure in normal people, but in patients withgastric ulcer it produces little or no increase inpressure. These findings suggest that dysfunction ofthe pyloric sphincter may be responsible, in wholeor in part, for increased reflux of duodenal contentinto the stomach. In patients with duodenal ulcer,the presence of the ulcer itself just beyond thepylorus could be responsible for abnormal functionof the pylorus. Presumably the normal pressuregradient across the antropyloroduodenal complex isaltered because of loss of normal pyloric function.Entry of chyme into the duodenum not onlystimulates the secretion of bile but also increasesduodenal motility and reduces antral motility.26Thus increased pressure in the duodenum and

decreased pressure in the antrum, in the presence ofdysfunction of the pylorus, may create suitableconditions for duodenal content to enter thestomach. This may explain the similar findings forbile acid and lysolecithin concentrations in thepreoperative duodenal ulcer and truncal vagotomyand pyloroplasty groups.Though many patients with duodenal ulcer

probably have some dysfunction of the pylorus,treatment by highly selective vagotomy will at leastpreserve whatever normal function the antropyloro-duodenal segment still possesses, because thepylorus is neither destroyed nor by-passed and themotor innvervation of the antropyloroduodenalcomplex is preserved.1' In addition, highly selectivevagotomy may actually improve the function of thepylorus by curing the ulcer and thereby diminish-ing oedema, muscular spasm, and mechanicaldeformity.We think, however, that the best explanation for

the finding of diminished duodenogastric reflux afterhighly selective vagotomy is to be found in theeffects of that operation on gastric smooth muscleand on intragastric pressure. After vagal denerva-tion of the body and fundus of the stomach,vagovagal inhibitory reflexes are destroyed,27leading to impairment of receptive relaxation andaccommodation to distension by the muscular wallsof the stomach: the altered pressure/volumerelationships after ingestion of a meal thus result inhigher pressures than normal within the stomach.28This increase in intragastric pressure will tend toprevent reflux from the duodenum into the stomach,and it has also been shown to increase the rate ofgastric emptying of liquids,29 which will itselfdiscourage reflux.Thus, through its effect on gastric pressure/

volume relationships, by preservation of an intactantropyloroduodenal segment, and, perhaps, bycuring the ulcer itself, highly selective vagotomykeeps 'bile' out of the stomach.The results of the studies on lysolecithin suggest

that lysolecithin concentrations in the stomach afterhighly selective vagotomy are normal. Postprandiallevels of lysolecithin in duodenal ulcer patientsbefore operation were nearly three times those innormal subjects. Perhaps because of the wide rangeof values recorded, however, there was nostatistically significant difference between the twogroups.The effectiveness of highly selective vagotomy as

an anti-reflux operation is apparent when it iscompared with Polya partial gastrectomy, truncalvagotomy and pyloroplasty, and truncal vagotomyand gastrojejunostomy, all of which are operationsthat either remove the pylorus or interfere with its

574


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/


functions. The fasting, post-prandial, and peak bileacid concentrations in the stomach after highlyselective vagotomy were found to be highly signifi-cantly less than after each of these three procedures.Not only was the degree of reflux much less afterhighly selective vagotomy, but the duration ofexposure of the gastric mucosa to increasedconcentrations of 'bile' was much longer after theseother operations than after highly selectivevagotomy. Eighty minutes after the ingestion of themeal, concentrations of bile in the stomachremained high after both partial gastrectomy andtruncal vagotomy and gastrojejunostomy. Theresults of the lysolecithin estimations showed apattern similar to the results of the bile acidestimations. The highest concentrations of lyso-lecithin in the stomach were found after partialgastrectomy and truncal vagotomy and gastro-jejunostomy, especially after the meal. After truncalvagotomy and pyloroplasty the post-prandial andpeak lysolecithin concentrations, though not signifi-cantly higher than after highly selective vagotomy,were double those found after highly selectivevagotomy and remained so for at least 80 minutesafter the meal.These findings suggest that removal or bypass of

the pylorus, with gastrojejunal anastomosis, as inPolya partial gastrectomy and truncal vagotomy andgastrojejunostomy, allows much more duodenalcontent to enter the stomach than occurs afterpyloroplasty, which produces less alteration of thenormal anatomy. After partial gastrectomy,concentrations of bile acids and lysolecithin in thestomach were significantly higher than after truncalvagotomy and gastrojejunostomy. This could be dueto the inhibitory effect of truncal vagotomy on

biliary and pancreatic secretion. Again,refluxing duodenal content might be diluted to a

greater extent by gastric secretions after truncalvagotomy and gastrojejunostomy than after partialgastrectomy. Finally, the degree of reflux may berelated to some extent to the size of the stoma,which may be somewhat larger, on average, afterpartial gastrectomy than after truncal vagotomy andgastrojejunostomy.Whatever factor or combination of factors may be

involved in the increased duodenogastric reflux afterpartial gastrectomy and truncal vagotomy with a

drainage procedure, it is unlikely that the greaterlength of follow-up of the patients who underwentthese procedures was responsible for the increasedreflux. If reflux is going to occur it will probably bemanifest soon after the operation, and all thepatients studied had been operated upon more thanone year before.

Reports from many centres have highlighted the

importance of keeping 'bile' out of the stomach. Thepresence of bile acids and lysolecithin in the stomachhas been shown to increase the movement ofhydrogen ions out of the lumen of the stomach intothe mucosa,4 3335 where they cause histamine to bereleased and gastritis to develop. Lysolecithin isformed in the duodenum when phospholipase A ofpancreatic juice hydrolyses the lecithin in bile, thereaction being activated by bile acids and trypsin.The gastritis that is produced by exposure of thegastric mucosa to the combined contents of theduodenum has been shown in dogs to be con-siderably worse than the gastritis that is caused bypure pancreatic juice alone or by pure bilealone.3 36 37 It was for that reason that we measuredboth bile acids and lysolecithin. In a separate reportthe relationship between gastritis and the reflux ofbile acids and lysolecithin is being examined.

In man, regurgitation of 'bile' into the stoniachmay lead to nausea and bilious vomiting, but ofgreater importance is the prolonged irritation of thegastric mucosa, which produces chronic superficialgastritis and eventually chronic atrophic gastritis.38These changes in the gastric mucosa produce asyndrome of 'bile reflux gastritis' or 'alkaline refluxgastritis' in a proportion of patients. The syndromeis characterised by persistent epigastric pain,vomiting, anorexia, and loss of weight.

In the longer term, prolonged contact of 'bile'with the gastric mucosa may well be a factor in thedevelopment of gastric carcinoma. Gastritisprogresses, with the development of intestinalmetaplasia, which some authors believe to be apre-cancerous lesion,38 though others doubtwhether the presence of such metaplasia is of valueas a predictor of cancer.39 Many reports have shownthat the incidence of gastric 'stump' carcinoma issignificantly increased more than 15 years afterBillroth 1 partial gastrectomy, Polya partial gastrec-tomy, or even gastrojejunostomy alone.81 40 Thecarcinomatous change usually develops close to thestoma, where the damage to the gastric mucosa isgreatest, a fact that strongly suggests that 'bile'reflux may be an important factor in the genesis ofthe condition. When Schrumpf and his colleaguesl'endoscoped 108 asymptomatic patients who hadundergone partial gastrectomy 20 to 25 years before,they found that four of them had infiltrating gastriccarcinoma and another three had carcinoma-in-situ:all but one of the remaining patients had eithersevere chronic gastritis, intestinal metaplasia, ordysplasia. Thus it seems likely that loss of thepylorus and the excessive duodenogastric reflux thatfollows exposes the patient to a considerable risk ofdeveloping gastric carcinoma in the long term.The severity of the gastritis that develops is

575


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/

576 Dewar, King, and Johnston

proportional to the amount of duodenal content thatrefluxes and to the duration of the exposure of thegastric mucosa to the irritant.36 The results of thisstudy suggest that when highly selective vagotomy isused in the treatment of duodenal ulcer, both thequantity of bile in the stomach and the length oftime that it is in contact with the gastric mucosa aresignficantly less than after other operations thatdestroy or bypass the pylorus. Meticulous, long-term studies are now needed to test our hypothesisthat the reduced duodenogastric reflux seen afterhighly selective vagotomy benefits the patient interms both of fewer symptoms of gastritis and alower risk of gastric carcinoma.

References1 Lawson HH. The role of the pyloric antrum andHofmeister valve in the development of the chronicgastric mucosal reaction. Br J Surg 1976; 63: 592-5.

2 du Plessis DJ. Gastric mucosal changes after resectionof the stomach. S Afr Med J 1962; 36: 471-8.

3 Lawson HH. Effect of duodenal contents on the gastricmucosa under experimental conditions. Lancet 1964; 1:469-72.

4 Orchard R, Reynolds K, Fox B, Andrews R, ParkinsRA, Johnson AG. Effect of lysolecithin on gastricmucosal structure and potential difference. Gut 1977;18: 457-61.

5 Scudamore HH, Eckstam EE, Fencil WJ, JaramilloCA. Bile reflux gastritis. Am J Gastroenterol 1973; 60:9-22.

6 Drapanas T, Bethea M. Reflux gastritis followinggastric surgery. Ann Surg 1974; 179: 618-27.

7 Keighley MRB, Asquith P, Alexander-Williams J.Duodenogastric reflux: A cause of gastric mucosalhyperaemia and symptoms after operations for pepticulceration. Gut 1975; 16: 28-32.

8 Stalsberg H, Taksdal S. Stomach cancer followinggastric surgery for benign conditions. Lancet 1971; 2:1175-7.

9 Domellof L, Eriksson S, Janunger K-G. Late pre-cancerous changes and carcinoma of the gastric stumpafter Billroth 1 resection. Am J Surg 1976; 137: 26-31.

10 Schrumpf E, Stadaas J, Myren J, Serck-Hanssen A,Aune S, Osnes M. Mucosal changes in the gastricstump 20-25 years after partial gastrectomy. Lancet1977; 2: 467-9.

11 Johnston D, Wilkinson AR. Highly selective vagotomywithout a drainage procedure in the treatment ofduodenal ulcer. Br J Surg 1970; 57: 289-96.

12 Goligher JC, Pulvertaft CN, de Dombal FT, ConyersJH et al. Five to eight year results of Leeds/Yorkcontrolled trial of elective surgery for duodenal ulcer.Br Med J 1968; 2: 781-7.

13 Findlay JM, Prescott RJ, Sircus W. Comparative

evaluation of water recovery test and fluoroscopicscreening in positioning a nasogastric tube duringgastric secretory studies. Br Med J 1972; 4: 458-61.

14 Magee WL, Gallai-Hatchard J, Sanders H. ThompsonRHS. The purification and properties of phospholipaseA from human pancreas. Biochem J 1962; 83: 17-25.

15 Fausa 0, Skalhegg BA. Quantitative determination ofbile acids and their conjugates using thin layerchromatography and a purified 3 oc-hydroxysteroiddehydrogenase. Scand J Gastroenterol 1974; 9: 249-54.

16 Bligh EG, Dyer WJ. A rapid method of total lipidextraction and purification. Can J Biochem Physiol1959; 37: 911-7.

17 Nelson GJ. Quantitation, composition and meta-bolism. In: Blood lipids and lipoproteins. New York:Wiley Interscience, 1972: 55.

18 Marinetti GV. Chromatographic separation, identifica-tion and analysis of phosphatides. J Lipid Res 1962; 3:1-20.

19 Fiske CH, Subbarrow Y. Colorimetric determinationof phosphorus. J Biol Chem 1935; 66: 375-400.

20 Snedecor GW, Cochran WG. Statistical methods Ames,Iowa: Iowa State University Press, 1967.

21 Mackie CR, Wisbey M, Cuschieri A. Milk'TcMEHIDA test for enterogastric bile reflux. Br JSurg 1981; 68: 351.

22 Rhodes J, Barnardo DE, Phillips SF, Rovelstad RA,Hofmann AF. Increased reflux of bile into the stomachin patients with gastric ulcer. Gastroenterology 1969;57: 241-52.

23 Fisher RS, Cohen S. Pyloric sphincter dysfunction inpatients with gastric ulcer. N EngI J Med 1973; 288:273-6.

24 Keighley MRB, Asquith P, Edwards JAC, Alexander-Williams J. The importance of an innervated and intactantrum and pylorus in preventing postoperativeduodenogastric reflux and gastritis. Br J Surg 1975; 62:845-9.

25 Capper WM, Airth GR, Kilby JO. A test for pyloricregurgitation. Lancet 1966; 2: 621-3.

26 Brink BM, Schlegel JF, Code CF. The pressure profileof the gastroduodenal junction zone in dogs. Gut 1965;6: 163-71.

27 Martinson J. Studies on the efferent vagal control ofthe stomach. Acta Physiol Scand 165; 65: suppl 255:5-24.

28 Stadass J, Aune S. Intragastric pressure/volume rela-tionship before and after vagotomy. Acta Chir Scand1970; 136: 611-5.

29 Wilbur BG, Kelly KA. Effect of proximal gastric,complete gastric, and truncal vagotomy on caninegastric electric activity, motility and emptying. AnnSurg 1973; 3: 295-303.

30 Malagelada JR, GO VLW, Summerskill WHJ. Alteredpancreatic and biliary function after vagotomy andpyloroplasty. Gastroenterology 1974; 66: 22-7.

31 MacGregor IL, Parent J, Meyer JH. Gastric emptyingof liquid meals and pancreatic and biliary secretionafter subtotal gastrectomy or truncal vagotomy andpyloroplasty in man. Gastroenterology 1977; 72: 195-205.

32 Lavigne ME, Wiley ZD, Martin P et al. A study of


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/

Bile acid and lysolecithin concentrations in stomach 577

gastric, pancreatic and biliary secretion, and the rate ofgastric emptying following parietal cell vagotomy. Am JSurg 1979; 138: 644-51.

33 Ivey KJ, DenBesten L, Clifton JA. Effect of bile saltson ionic movement across the human gastric mucosa.Gastroenterology 1970; 59: 683-90.

34 Davenport HW. Destruction of the gastric mucosalbarrier by detergents and urea. Gastroenterology 1968;54: 175-81.

35 Davenport HW. Effect of lysolecithin, digitonin andphospholipase A upon the dog's gastric mucosalbarrier. Gastroenterology 1970; 59: 505-9.

36 Lawson HH. The effect of the duodenal contents onthe gastric mucosa. S Afr J Surg 1965; 3: 79-92.

37 Dragstedt LR, Woodward ER, Seito T, Isaza J,Rodriguez JR, Samiian R. The question of bileregurgitation as a cause of gastric ulcer. Ann Surg 1971;174: 548-59.

38 Morson BC, Dawson IMP. Gastrointestinal pathology.Oxford: Blackwell, 1972.

39 Domellof L, Eriksson S, Janunger K-G. Carcinomaand possible precancerous changes of the gastric stumpafter Billroth II resection. Gastroenterology 1977; 73:462-8.

40 Siurala M, Lehtola J, Ihamaki T. Atrophic gastritis andits sequelae. Scand J Gastroenterol 1974; 9: 441-6.

41 Leading Article. After gastrectomy. Br Med J 1981;282: 1096-7.


http://gut.bmj.com

/G


nloaded from

http://gut.bmj.com/

treatment vagotomy, gastrectomy, vagotomysummary duodenogastric reflux of bile acids and...

Documents