Br. J. Surg. Vol. 70 (1983) 393-397 Printed in Great Britain

Proffered review

R. EARLAM

The London Hospital. Whitechapel, London E l .

Bile reflux and the Roux en Yanastomosis

Bile reflux mto the stomach and then subsequently into theoesophagus occurs frequently after the two basic forms ofpartial gastrectomy Billroth I and Billroth II. Similarly bilereflux follows the different operations performed to com-pensate for the temporary slowing of gastric emptying after avagotomy—Hemeke-Mikulicz pyloroplasty, Finney pyloro-plasty and gastrojejunoslomy. When large amounts of bile enterthe stomach, after a Billroth II partial gastrectomy forexample, bilious vomiting occurs. In the early days of gastricsurgery there were several attempts at altering the techniques ofthe original partial gastrectomy to reduce bile reflux, butanother method was to divert bile away from the stomach by aRoux en Y anastomosis (Fig. 1), an end-to-side jejunojejuno-stomy, which was successful for gross bile vomiting.

The reflux of smaller amounts of bile either into the intactstomach or after different types of pyloroplasty or vagotomywithout actual bilious vomiting, can cause symptoms whicharise from the stomach or the lower oesophagus. The differenttechniques available for recognition of this duodenogastricreflux will be discussed. It will then be clear that bile refluxoccurs more frequently than has been previously thought andthat bilious vomiting represents the most extreme form ofreflux, with many other symptoms of lesser amounts of refluxbeing unrecognized. Not all patients with bile reflux havesymptoms, even though there may be histological changes inthe mucosa of the stomach and oesophagus. On the other handthere are some patients who have a very sensitive loweroesophageal mucosa with minimal histological changes, inwhom all bile must be removed from the stomach before theylose their symptoms.

Most patients with gastro-oesophageal reflux disease can betreated by simply repairing the anatomical defect of a hiatushernia. Over 80 per cent of the patients do well. Otheroperations are based on the reduction of gastric acid, and manypatients can be successfully treated by an antrectomy with orwithout the different types of vagotomy. However, in everysurgeon's series there are a few problem patients, and it is thesewho may have bile reflux. It is not known whether they can beidentified at the time of their first operation or even whether itwould be essential to do so. The author puts great weight on theclinical estimation of bile regurgitation by asking the patientwhether fluid comes back into the mouth that is 'yellow, greenor bile-stained'. The surgical treatment of these particularproblem patients is by the Roux en Y procedure.

Apart from reconstruction after a total gastrectomy thedecision to do a Roux en Y anastomosis must not be takenlight ly. Although some surgeons do it as a primary procedure,in the author's opinion, it is best reserved for failures after anoriginal procedure for reflux disease, ft not only prevents bilereaching the lower oesophagus but also reduces gastric aciditybecause a vagotomy and antrectomy are done as essentialcomponents in addition to the jejunojejunostomy. Subsequentcomplications consist of those that can normally follow apartial gastrectomy and those that may occur after operationson the small bowel. They are more frequent than after a partialgaslrectomy alone and are many more than after a hiatushernia repair or a vagotomy and drainage procedure. Avagotomy is usually added to avoid the complication of stomalulceration. Diarrhoea can occur after both a Billroth IIoperation and a Roux en Y anastomosis. Some surgeonsconsider that a more important factor than the type ofvagotomy is the size of the gastric exit or pyloroplasty. It mustbe emphasized that it is not gastro-oesophageal reflux that isaltered, but the corrosive nature of the gastric fluid.

1897

Fig. 1. Roux en Y anastomosis.

This review is concerned with the application of the Roux enY anastomosis for the relief of oesophageal symptoms. Em-phasis is placed on the physiology and pathology of bile reflux.No technical or surgical details about the operation areincluded, other than the fact that the anastomosis must be atleast 40cm away from the lower oesophagus.

Historical aspects of early gastric surgery relating to bilerefluxThe Roux en Y anastomosis (Fig. 1) was described by Roux in1897 (1). Cesar Roux was born in Switzerland in Mont-La-Villein 1857 and died in 1915. He obtained an MD from Berne andbecame professor of surgery, pathology and gynaecology inLausanne. Being Swiss French he wrote in French and thecorrect name for the anastomosis is Roux en Y grecque. Hisother original operation was the Roux loop (Fig. 2), describedin 1907, done to bridge the gap between the oesophagus and theduodenum (2, 3) after a total gastrectomy. This loop is asegment of jejunum with its vascular pedicle and should not beconfused with the Roux en Y anastomosis.

The Roux en Y operation was done for postoperative bilevomiting of such an extreme nature that it could cause death.Nowadays it is performed for minor degrees of bile refluxcausing less severe symptoms. Nevertheless many lessons werelearnt in the last century which should not be forgotten. It isimpossible to understand the full implications now or in thoseearly days without knowing about the first gastric operations.Billroth did his first partial gastrectomy for pyloric cancer in1881. This was not the original operation because it had beendone experimentally before in dogs, and both Pean andRydgier had operated on patients who had died soon after-wards, but Billroth's patient was the first to survive.

In the same year the first gastrojejunostomy was done inBillroth's clinic by Wolfler as a palliative procedure for anotherpyloric carcinoma. In 1884 gastroenterostomy was taken a stepfurther with the removal of the pyloric tumour and closure of

394 R. Earlam

1907 1893Fig. 2. Roux loop. Fig. 3. Brciun anastomosis.

ihe duodenum to give the Billroth II type of partial gastrec-toiny. Originally these operations were done for cancers of thestomach, but Rydgier's first partial gastrectomy was done for agastric ulcer. Hemeke and Mikulicz both performed theireponymous pyloroplasty independently of each other in 1886.These were exciting days, because all the classic operationswere done for the first time within 6 years.

At the turn of the century duodenal ulceration became morecommon, and the original operations were then applied in suchan enthusiastic fashion that William Mayo could report 307gastroenterostomies by 1905. The rationale of doing a partialgastrectomy instead of a drainage procedure was originallybased on removing as much stomach as possible to reduce acidproduction. In 1905 Edkins demonstrated that the antrum wasthe source of the hormone gastrin, which increased gastricsecretion. Thereafter antrectomies could be rationalized, butthe surgeons actually had no idea of how clever they werebefore 1905.

Historically the next great advance was Lester Dragstedt'struncal vagotomy in 1943. The effect of the vagus on acidproduction had been known at the beginning of the century.Continued gastric secretion after a vagotomy was interpretedby Pavlov as failure to cut all the vagal fibres. His obsessionwith neural reflexes caused him to miss the whole world ofhormones. Dragstedt, after experience with vagotomy in dogs,persuaded a surgeon to do a simple truncal section alone onhimself for his duodenal ulcer. Later a drainage operation hadto be done, but this was the start of the vagotomy era.Variations in the method of vagotomy which have followed areirrelevant compared with this great advance made by LesterDragstedt.

The main technical problem in early gastric surgery was themortality rate. In 1884 Rydgier reported 43 partial gastrec-tomies gleaned from the literature with 28 fatalities. Themortality from gastrojejunostomy between 1881-1885 was 70per cent and between 1886-1892 it had decreased to 40 percent. The majority of these deaths were due to 'regurgitantbilious vomiting' or 'circulus vitiosus'. The injurious effect ofbile and pancreatic secretions on the stomach was discovered ata very early stage. The first Billroth II patient eventually died ofbilious vomiting due to a spur obstruction at the anastomosisand not due to a recurrence. Surgical techniques for preventingor relieving bile reflux were soon developed. The majority ofearly gastrojejunostomies were performed in an antecolicfashion on to the anterior wall of the stomach. In 1885, vonHacker, a pupil of Billroth, first described the retrocolic anasto-mosis onto the posterior wall of the stomach through thetransverse mesocolon. This procedure, together with shorten-

ing the afferent loop, first suggested in 1901 by Petersen fromCzerny's clinic, dramatically reduced the number of patientswith bilious vomiting.

An end-to-side jejunojejunostomy is a satisfactory way ofdealing with bile in the stomach after a gastrojejunostomy, withor without an antrectomy: it was first suggested by Wolflerwho demonstrated its effect in a dog in 1883. Doyen describedit again in 1893, but it was Roux who popularized it afterhis description in 1897.

The other operation for preventing bile reflux was the sidc-to-side jejunojejunostomy first described by Braun in 1893 (4)(Fig. 3). It was originally applied to the jejunum after agastrojejunostomy, when bilious vomiting had to be avoided orrelieved. Its main use nowadays is to prevent bile reflux into theoesophagus after a total gastrectomy and oesophagojejuno-stomy. A further modification has recently been described inwhich two side-to-side anastomoses are made (5).

Bile reflux in the intact stomachIt is still not proven whether some bile can reflux in normalasymptomatic people and can therefore be regarded as withinnormal limits. Intubation certainly induces nausea and bileregurgitation, but non-invasive techniques do not have thisdisadvantage. Beaumont, the Canadian doctor who lookedafter Alexis St. Martin with a gastric fistula, is always cited asthe first to describe bile reflux in a 'normal' person (6). But ifone reads the book in detail, his patient does not appearentirely normal remembering that he had a large gastric fistulaand was very irascible. The etymological derivation of the wordcholeric shows that the ancient Greeks were perfectly aware ofthe association between bile and anger.

Bile reflux was first associated clinically and experimentallywith gastric ulceration in 1914 (7, 8) and this association hasbeen observed on many subsequent occasions (9-15). There isa definite anatomical abnormality of the pyloric muscle (16)which probably explains why the reflux persists after the ulcerheals (17).

Bile reflux also occurs in duodenal ulceration (18, 19). Inpentagastrin studies bile has been found in either the basalsamples or the stimulated juice in the majority of patients (20).Reflux also occurs in atrophic gastritis (21, 22), alcoholicgastritis (10), gall bladder dyspepsia (23, 24), chronic non-specific lung disease (25) and stress ulcers (26, 27).

Bile in the postoperative stomachThe sensitive tests to measure bile reflux into the stomachdeveloped during the last 15 years have confirmed what wasknown at the beginning of the century about bilious vomiting.

Roux en Y anastomosis

They arc important because the results have established thateven a small amount of bile is able to cause symptoms andgastritis. They help to quantify the degree of pyloric regurgi-Uition thai can occur postoperatively.

As expected bile reflux is common after a Billroth II type ofpartial gastrectomy (13, 27). Over 50 per cent of the restingjuice samples will be bile-stained and 75 per cent of the nightsamples (28). Bile is also present after a Billroth I partialgastrectomy but to a lesser extent (27). The fact that bile refluxis a cause of gastric ulceration, yet the operation commonlyperformed for its cure causes even more bile to enter thestomach must sound illogical. Perhaps the abnormal anatomi-cal arrangements of the smooth muscle of the lesser curve (29),which arc said to be responsible for gastric ulceration, may bethe most important factor. A Billroth I partial gastrectomy willremove this abnormal muscle and allow more bile reflux, butsymptoms and the ulcer usually disappear.

Bile reflux can also occur after a vagotomy of any type (19,27, 30, 31). It must be realized however that reflux also occursin duodenal ulceration, the original disease, so all of thesepatients may originally have an abnormal pylorus. Highlyselective vagotomy without drainage procedure is associatedwith the least bile reflux (18, 32), No studies have shownwhether there is more reflux after a partial gastrectomy when avagotomy is added.

Much more reflux of bile occurs after a gastrojejunostomythan a pyloroplasty (19. 27, 33), and reflux after a gastro-jcjunostomy is higher than after excision of both the pylorusand antrum.

A Roux en Y anastomosis is followed by the least amount ofbile reflux into the stomach (19, 20, 27, 34). The anastomosismust be at least 40cm and possibly 50cm from the loweroesophagus (34). When a partial gastrectomy is converted to aRoux en Y the amount of bile in the stomach is reduced and thegastritis can be reversed experimentally (14, 35). Side-to-sidejejunojejunostomy is not as effective as a Roux en Y anasto-mosis in removing bile from the lower oesophagus, as demon-strated by the radioactive technetium HIDA technique (34). Ajejunal loop has been advocated for intractable symptomscomplicating peptic ulcer surgery, but no bile reflux studieswere done to compare its effect with other procedures (36).

Bile in the oesophagusBile in the stomach is a prerequisite for bile in the oesophagusexcept after total gastrectomy. Most stomachs secrete acid, sothat the bile is mixed with hydrochloric acid. Much confusionhas occurred because of the term 'alkaline oesophagitis', whichhas been favoured by surgeons ignoring the true use of theword alkaline. If pH 7-0 is accepted as the border between acidand alkali then true alkaline oesophagitis should only bepresent with the intact stomach in pernicious anaemia andatrophic gastritis, and, in the postoperative state, after a totalgastrectomy. Alkaline oesophagitis and heartburn only occurin pernicious anaemia (37) and after a total gastrectomy.Experimentally trypsin has been shown to damage oesophagealmucosa (38). After total gastrectomy most surgeons do notfavour either an oesophagoduodenal anastomosis or an inter-posing jejunal Roux loop, because of the complication of bilereflux and oesophagitis. There have been no quantitativestudies using modern techniques of measuring bile after thesetwo particular operations where the refluxing material isalkaline and a total gastrectomy has been performed.

On the other hand bile mixed with gastric acid juice to a pHbelow 4-0 is corrosive in the oesophagus, as first demonstratedin 1951 (39) and confirmed on numerous occasions (40-43).Experimentally it has been shown that the addition of bile toacid at any given pH level makes it more corrosive to theoesophageal mucosa (43-45). Moreover, in pain reproductiontests adding bile to the perfusing acid fluid causes pain at alower threshold (45).

Histological changes in the lower oesophagus have neverbeen shown to correlate well with symptoms. Recently, with therealization that pyloric regurgitation of bile can occur throughan abnormal pylorus in gastrooesophageal reflux disease(46-48), studies have been made to combine the appearances in

the stomach mucosa with those in the oesophagus. Gastritis isfrequently present and is associated with abnormal histologicalchanges in the oesophageal mucosa (49, 50). The presence ofbile in the stomach refluxing through the gastro-oesophagealsphincter would appear to be the common factor.

Mechanism of mucosal damage by bileThe simplest experiments are performed by placing bile in thestomach, or stomach in the bile (51). The former always causesgastritis. The latter can be achieved by putting a strip of gastricmucosa in the gall bladder—without subsequent change. Theaction of pure bile is minimal, but as the pH drops it becomesmore severe (52, 53). The following sequence of events has beensuggested in an excellent article (48). Bile changes the characterof the gastric mucus (54, 55) so that the underlying epithelialcells are damaged. The gastric mucosal barrier, which is aphysiological concept, is destroyed (56-59). Hydrogen ionsleak in, histamine is released as a non-specific reaction todamage and more hydrogen ions may be secreted into thelumen. Over a long term the gastric acid secretion is reducedbecause the damage leads to atrophic gastritis and a progres-sive loss of parietal cells.

An important chronic change in the gastric mucosal cells,apart from atrophic gastritis, is gastrointestinal metaplasia andneoplasia. Cancer in the proximal stump of the stomach isaccepted as a long term complication after Billroth II types ofpartial gastrectomy. If a Roux en Y diversion of bile reversesthe histological changes in a partial gastrectomy, it would seemlogical that it must be a safer long term procedure than aBillroth II type of partial gastrectomy in regard to thisparticular complication, but there are no long term studiesavailable, presumably because of the rarity of this particularoperation.

Tests for bile reflux into the stomach or oesophagus1. Symptoms alone (27).2. Bile in the gastric aspirate or in the lower oesophagus seen

macroscopically.3. Measurements of naturally occuring bile pigments (60),

salts (9) or acids (18) in the gastric juice.4. Radio-opaque contrast material placed in the duodenum

and observed radiologically (Capper's pyloric regurgi-tation test) (20).

5. Measurement of artificially introduced substances: (a)C14-labelled bile (61); (b) bromsulphalein (62); (c) indo-cyanin green and phenol red (62-64); (d} lysolecithin (18,27, 65, 66, 67); (<?) polyethylene glycol (68); (/) tech-netium-labelled HIDA (34, 69, 70).

6. Mathematical calculation based on sodium concentrationof gastric juice (19, 62, 71).

7. Indirectly by histological examination of gastric mucosa(72, 73).

8. Physiological profile of pylorus (47, 63, 74-76).

Indications for the Roux en Y anastomosis1. After a total gastrectomy.2. If bile reflux occurs after a pyloroplasty.3. Failed partial gastrectomy (77).4. Failed hiatus hernia repair.5. Failed conservative management of a benign peptic stric-

ture in the lower oesophagus (78, 79).6. Occasionally as a primary procedure for duodenogastric

reflux with or without oesophagitis.

Medical treatment of bile refluxComplete cessation of smoking is essential. It is known that thelower oesophageal sphincter is temporarily weakened after avagotomy and then returns to normal, but after a partialgastrectomy it has a permanently reduced pressure (80, 81).Many drugs alter this pressure, an important one being nicotinewhich is known not only to reduce the resting gastro-oesopha-geal sphincter pressure, but also to increase bile reflux into thestomach (75, 82-84).

Bed rest as treatment is out of fashion, although proven to bebeneficial (85), but it should be appreciated that pyloric

396 R. Earlam

regurgitation occurs more frequently in the supine position thansitting up (10). Experiments measuring the pylonc pressureprofile may vary if this factor is not recorded.

Antacids (86), of which aluminium hydroxide is the best (87,88), can not only neutralize acid but also adsorb bile salts (89).Cholestyramine is a specific agent for adsorbing bile salts. Itwas originally developed for reducing the level of cholesterol inthe blood (90), but it can be used for symptomatic treatment(91) although not pleasant to take.

Mctoclopramidc promotes gastric emptying and reducespyloric regurgitation. Its main effect on bile in the stomach isprobably due to the increased speed of gastric emptying.

Carbeiioxolone stimulates mucus secretion and prolongs thelife of the gastric epithelial cell. It does not reduce pyloric bileregurgitation, and only prevents damage to the inucosa. Analginate preparation mixed with carbenoxolone is marketed inthe United Kingdom as Pyrogastrone.

The effect of cholecystectomy on bile refluxCholecystectomy is nowadays not done for vague abdominalsymptoms because the results were unreliable. There areprobably no specific symptoms of indigestion associated withgall bladder pathology (92). One study showed that patientswith flatulent dyspepsia were equally divided as to whether theyhad gall bladder disease or not (93). Cholecystectomy isnotoriously unreliable for relieving the symptoms of flatulentdyspepsia (24, 94), but it can relieve gastro-oesophageal refluxsymptoms (24, 95).

Bile reflux through the pylorus occurs in many patients withgall bladder disease (23, 24, 96) and usually reverts to normalafter » cholecystectomy. This would fit in with the fact thatpatients with a non-functioning gall bladder are less likely tohave gastrooesophageal reflux symptoms before cholecystec-tomy than those with normal function (92, 97). The gastro-oesophageal sphincter pressures do not alter after cholecystec-tomy (98), but the oesophageal mucosa is certainly less sensitiveto stimulation by acid dripped into it after a cholecystectomy(95). The cause of the bile reflux into the stomach is unknown.There may be a reflex spasm of the pyloric muscle or there maysimply be less concentrated bile in the duodenum after chole-cystectomy (97). It is therefore likely that there are no changesin the oesophagus after a cholecystectomy other than recoveryof damage to the epithelium: there is less bile in the stomachdue to less concentrated bile being secreted into the duodenumand being available to reflux; there may be some abnormalityof the pyloric muscle which reverts to normal aftercholecystectomy.

ConclusionsBile reflux into the stomach through the pylorus is an im-portant factor in many upper abdominal disorders. If there isbile in the stomach it may mix with the gastric juices and refluxinto the oesophagus. If it occurs after a total gastrectomy theeffect will be due only to bile and pancreatic secretions (99).This subject has been reviewed here and emphasis placed on thedifferent components such as the particular bile salts (100)which may cause damage.

Originally the Roux en Y anastomosis was used for grossbilious vomiting after various failed gastric operations such asPolya partial gastrectomy or gastrojejunostomy. It was redis-covered about 70 years later for the lesser evils of 'alkalinegastritis' and bile reflux oesophagitis (101, 102). It is usuallyperformed in conjunction with a partial gastrectomy andvagotomy. The latter enables less stomach to be removed and itcan in effect be an antrcctomy. After a total gastrectomy theRoux en Y anastomosis is one of the methods available fordiverting bile and pancreatic secretions away from the loweroesophagus.

No techniques concerning the anastomosis have been dis-cussed other than emphasis on the fact that bile must bediverted at least 40cm away from the oesophageal mucosa. Theresults are not discussed here because it is difficult to assessthem when the original patient populations vary so much andthere are no controlled trials. In a recent article it was suggestedthat a new reliable non-invasive test for bile reflux was needed

(103). Technetium-labelled HIDA might be the ideal substance.It is emphasized that the success of a Roux en Y proceduredepends on reducing the acid and bile in the stomach andeliminating pyloric obstruction; reflux into the oesophagus stilloccurs but is asymptomatic. It is the only procedure availableto ensure that no bile refluxes into the stomach.

References1. Roux C.: De la gastro-enterostomie. Rev. Gvnecol. Chir. Abdom.

1897; 1:95.2. Roux C.: L'oesophago-jejuno-gastrostomosc. nouvellc operation

pour retrecissement infranchisable de foesophage. Sem. Mem.(Paris) 1907; 27: 37-40.

3. Allison P. R. and da Silva L. T.: The Roux loop. Br. J. Surg. 1953;41: 173-80.

4. Braun H.: Uber Gastro-enterostomie und gleichzeitig ausgefuhrteEntero-anastomose. Arch. Klin. Chir. 1893; 45: 361.

5. Lygidakis N. J.: Total gastrectomy for gastric carcinoma: aretrospective study of different procedures and assessment of anew technique of gastric resection. Br. J. Surg. 1981; 68: 649 55.

6. Beaumont W.: Experiments und Observations on the Gastric Juiceand Physiology of Digestion. Pittsburg: Allen. 1833.

7. Smith G. M.: An experimental study of the relation of bile toulceration of the mucous membrane of the stomach. /. Med. Res.1914; 30: 147-83.

8. Stuber B.: Zur Aetiologie des Ulcus Ventriculi. Eine neue Theorieauf experimenteller Grundlage. Miinch. Med. Wochenschr. 1914;61: 1265.

9. Rhodes J., Barnardo D. E.. Phillips S. F. et al.: Increased reflux ofbile into the stomach in patients with gastric ulcer. Gastro-enterology 1969; 57: 241-52.

10. Flint F. J. and Grech P.: Pyloric regurgitation and gastric ulcer.Gut 1970; 11: 735-7.

11 . Cocking B. and Grech P.: Pyloric reflux and the healing of gastriculcers. Gut 1973; 14: 555-7.

12. DuPlessis D. J.: The importance of the pyloric antrum in pepticulceration. 5. Afr. J. Surg. 1963; 1: 3-11.

13. DuPlessis D. J.: Pathogcnesis of gastric ulccration. Lancet 1965;1: 974-8.

14. Lawson H. H.: Effect of duodenal contents on the gastric mucosaunder experimental conditions. Lancet 1964; 1: 469-72.

15. Lawson H. H.: Gastritis and gastric ulceration. Br. J. Surg. 1966;53: 493-6.

16. Liebermann-Meffert D. and Allgower M.: Struktur der Ver-schluss Systeme des Magens. Anatomie und Pathologic derPyloroantralwand. In: Schumpelik V., Begemann F. and WernerB. (ed.): Reflukrankheit des Magen. Stuttgart: Enke, 1979.

17. Black R. B.. Roberts G. and Rhodes J.: The eflect of healing onbile reflux in gastric ulcer. Gut 1971; 12: 552-8.

18. Dewar P.. King R. and Johnston D.: Bile acid and lysolecithinconcentrations in the stomach in patients with duodenal ulcerbefore operation and after treatment by highly selective vago-tomy. partial gastrectomy, or truncal vagotomy and drainage.Gut 1982; 23: 569-77.

19. Hobsley M.: Tests of gastric secretory function. In: Sircus W. andSmith A. N. (ed.): Scientific Foundations of Gastroenterology.London: Heinemann, 1980.

20. Capper W. M., Airth G. R. and Kilby J. O.: A test for pyloricregurgitation. Lancet 1966; 2: 621-3.

21. Cole G. J.: The implications of bile in the stomach. Gut 1969; 10:864-7.

22. Capper W. M., Butler T. J. and Buckler K. G.: New alkalineareas in gastric mucosa after gastric surgery. Gut 1966; 7: 220-2.

23. Capper W. M., Butler T. J. and Kilby J. O.: Gallstones, gastricsecretion and flatulent dyspepsia. Lancet 1967; 1: 413 -15.

24. Johnson A. G.: Cholecystectomy and gallstone dyspepsia. Ann. R.Coll. Surg. 1975; 56: 69-79.

25. Beeley M. and Grech P.: Pyloric incompetence and chronic non-specific lung disease. Gut 1971; 12: 102-6.

26. Meeroff J. C., Paulsen G. and Guth P. H.: The role of bile refluxin the development of cold-restraint gastric lesions. Am. J. Dig.Dis. 1975; 20: 365-9.

27. Schumpelik V., Begemann F. and Werner B.: Refluxkrankheit desMagens. Stuttgart: Enke, 1979.

28. Borg I.: Bile admixture in gastric juice in health and in peptic ulcerbefore and after operation according to Billroth I and Billroth II.Ada Chir. Scand. 1959; 251: 97-112.

29. Oi M., Oshida K. and Sugimura S.: The location of gastric ulcers.Gastroenterology 1959; 36: 45-56.

30. Rothmund M.: Duodenogastraler Reflux. 2. Gastroenterologie1977; 15: 192-201.

31. Rothmund M., Deisler G., Kaufmann A. et al.: Duodeno-gastrischer Reflux nach Vagotomie und Pyloroplastik.Langenbecks Arch. Chir. 1976: 340: 167-78.

32. Keighley M. R. B., Asquith P., Edwards J. A. C. et al.: Theimportance of an innervated and intact antrum and pylorus inpreventing postoperative duodenogastric reflux and gastritis. Br.J. Surg. 1975; 62: 845-9.

Roux en V anastomosis 397

33. Kilby J, O,: Duodenogasiric reflux and pyloric surgery. Gastro-enterology 1970; 58: 594-5.

34. Donovan I. A., Fielding J. W. L., Bradby H. et al.: Bile diversionafter total gastrectomy. Br. J. Surg. 1982; 69: 389-90.

35. Lawson H. H.: The reversibility of postgastrectomy alkalinereflux gastritis by a Roux en Y loop. Br. J. Surg. 1972; 59: 13-15.

36. Ramus N. I., Williamson R. C. N. and Johnston D.: The use ofjejunal interposition for intractable symptoms complicatingpeptic ulcer surgery, Br. J. Surg. 1982; 69; 265-8.

37. Orlando R. C. and Bozymski E. M.: Heartburn in perniciousanemia; a consequence of bile reflux. New Engl. J. Med. 1973; 289:522-3.

38. Mud H. J., Kranendonk S. E., Obertop H. et al.: Active trypsinand reflux oesophagitis: an experimental study in rats. Br. J.Surg. 1982: 69: 269-72.

39. Cross F. S. and Wangensteen O. H.: Role of bile and pancreaticjuice in production of esophageal erosions and anemia. Proc. Soc.Exp, Biol. 1951; 77: 862-6.

40. Moffat R. C. and Berkas E. M.: Bile esophagitis. Arch. Surg.1965; 91: 963-6.

41. Crurnplin M. K. H., Stol D. W., Murphy G. M. et al.: The patternof bile salt reflux and acid secretion in sliding hiatus hernia. Br. J.Surg. 1974; 61: 611-16.

42. Gtlhson E. W., Capper W. M., Airth G, R. et al.: Hiatus herniaand heart burn. Gut 1969; 10: 609-13.

43. Gillison E. W., Nyhus L. M. and Bombeck C. T.: The significanceof bile in reflux esophagitis. Surg. Gvnecol. Obstet. 1972; 134:419-24.

44. Safaie-Shirazi S., Denbesten L. and Zike W. L.: Effect of bile saltson the tonic permeability of the esophageal mucosa and their rolein the production of esophagitis. Gastroenterologv 1975: 68:728-33.

45. Henderson R. D.. Mugashe F. and Jeejeebhoy K. N.: The role ofbile and acid in the production of esophagitis and the major defectot csophagitis. Ann. Thome. Surg. 1972; 14: 465-73.

46. Kaye M. D. and Showalter J. P.: Pyloric incompetence in patientswith symptomatic gastro-ocsophagcal reflux. /. Lab. Clin. Med.1974; 83: 198-206.

47. Kaye M. D., Mehta S. .T. and Showalter J. P.: Manometric studieson the human pylorus. Gastroenterology 1976; 70: 477-80.

48. Rees W. and Rhodes J.: Bile reflux in gastro-oesophageal disease.Clin. Gastroenterol. 1977; 6: 179-200.

49. Moraes-Filho J. P. P. and Bettarello J. P.: Lack of specificity ofthe acid perfusion test in duodenal ulcer patients. Am. J. Dig. Dis.1974; 19: 785-90.

50. Volpicelli N. A., Yardley J. H. and Hendrix T. R.: The associ-ation of heartburn with gastritis. Am. J. Dig. Dis. 1977; 22: 333-9.

51. Kirk R. M.: Experimental gastric ulcers in the rat: the separateand combined effects of vagotomy and bileduct implantation intothe stomach. Br. J. Surg. 1970; 57: 521-4.

52. Black R. B., Hole D. and Rhodes J.: Bile damage to the gastricmucosal barrier; the influence of pH and bile acid concentration.Gaaroenterohgy 1971; 61: 178-84.

53. Eastwood G. L.: Effect of pH on bile salt injury to gastric mucosa.Qastraenterologf 1975; 68: 1456-65.

54. Grant R., Grossman M. L. and Ivy A. C.: Effect of bile on gastricmucus. Am. J. Physio!. 1948; 155: 440.

55. Rhodes J.: The aetiology of gastric ulcer. Gastroenterology 1972;63: 171-82.

56. Ivy K. J.: Gastric mucosal barrier. Gastroenterology 1971; 61:247-57.

57. Teorell T.: Untersuchungen liber die Magensaftsekretion. Scand.Arch. Physio!. 1933; 66: 225-30.

58. Davenport H. W.: Destruction of the gastric mucosal barrier bydetergents and urea. Gastroenterology 1968; 54: 175-81.

59. Davenport H. W., Warner H. A. and Code G. F.: Functionalsignificance of gastric mucosal barrier to sodium. Gastro-enterologv 1964; 47: 142-52.

60. Keighley M. R. B., Asquith P. and Alexander-Williams J.:Duodenogastric reflux: a cause of gastric mucosal hyperaemiaand symptoms after operations for peptic ulceration. Gut 1975;16: 28-32.

61. Rokkjaer M., Marqversen J., Kraglund K. et al.: Quantitativedetermination of pyloric regurgitation in response to intra-duodenal bolus injection. Scand. J- Gastroenterol. 1977; 12:827-32.

62. Faber R. G.. Russell R. C. G., Royston C. M. S. et al.: Duodenalreflux during insulin-stimulated secretion. Gut 1974; 15: 880.

63. Fisher R. and Cohen S.: Physiological characteristics of thehuman pyloric sphincter. Gastroenterology 1973; 64: 67-75.

64. Fiddian-Green R. G., Whitfield P., Russell R. C. G. et al.:Indocyanine green as a marker of duodenal reflux in aspiratedgastric juice. Br. J. Surg. 1974; 61: 323-4.

65. Johnson A. G. and McDermott S. J.: Lysolecithin: a factor in thepathogenesis of gastric ulceration? Gut 1974; 15: 710-13.

66. Borgstrom B.: Studies of the phospholipids of human bile andsmall intestinal contents. Acta Chem. Scand. 1957; 11: 749.

67. Schmidt H. and Creutzfeldt W.: The possible role of phospholi-pase A in the pathogenesis of acute pancreatitis. Scand. J.Gastroenterol. 1969; 4: 39-48.

68. Wormsley K. G.: Aspects of duodeno-gastric reflux in man. Gut1972; 13: 243-50.

69. Tonlin R. D., Malmud L. S., Stelzer F. et al.: Enterogastric refluxin normal patients and patients with Billroth II gastro-enterostomy. Gastroenterology 1979; 77: 1027-33.

70. Mackie C. R., Wisbey M. and Cuschieri A.: Milk ^Tcm EH1DAtest for enterogastric bile reflux. Br. J. Surg. 1982: 69: 101-4.

71. Hobsley M.: 'Pyloric reflux'. A modification of the two-com-ponent hypothesis of gastric secretion. Clin. Sci. Mol. Med. 1974;47: 131-41.

72. Siurala M. and Tawast M.: Duodenal regurgitation and the stateof the gastric mucosa. Acta Med. Scand. 1956; 153: 451-8.

73. Siurala M. and Varis K.: Gastritis. In: Sircus W. and Smith A. N.(ed.): Scientific Foundations of Gastroenterology. London: Heine-mann, 1980.

74. Brink B. M., Schlegel J. E. and Code C. F,: The pressure profile ofthe gastroduodenal junctions! zone in dogs. Gut 1965; 6. 163-71.

75. Valanzuela J. E., Defllippi C. and Csendes A.: Manometricstudies on the human pyloric sphincter. Effect of cigarettesmoking, metoclopramide and atropine. Gastroenterologv 1976;70:481-3.

76. Fisher R. S., Lipshutz W. and Cohen S.: The hormonal regulationof pyloric sphincter function. J. Clin. Invest. 1973; 52: 1289-96.

77. Kennedy T. and Green R.: Roux diversion for bile refluxfollowing gastric surgery. Br. J. Surg. 1978; 65: 323-5.

78. Tanner N. and Westerholm P.: Partial gastrectomy in the treat-ment of esophageal stricture after hiatal hernia. Am. J. Surg.1968; 115: 449-53.

79. Royston C. M. S., Dowling B. L. and Spencer J.: Antrectomywith Roux en Y anastomosis in the treatment of peptic oeso-phagitis with stricture. Br. J. Surg. 1975; 63: 605-7.

80. Earlam R. J. and Thomas P. A.: The gastro-oesophageal junctionbefore and after operations for duodenal ulcer. Br. J. Surg. 1973;60: 717-19.

81. Siewert J. R., Koch A., Stuhler R. et al.: Kardiafunktion nachdistaler Magenresektion. Gastroenterologie 1974; 12: 583-90.

82. Read N. W. and Grech P.: Effect of cigarette smoking oncompetence of the pylorus: preliminary study. Br. Med. J. 1973; 2:313-16.

83. Stanciu C. and Bennett J..R.: Smoking and gastro-oesophagealreflux. Br. Med. J. 1972; 3: 793-5.

84. Begemann F., Schway M. and Schumpelik V.: Zigarettenrauchensteigert der intragastralen Lysolecithinsehalt. Z. Gastroenterol.1978; S13.

85. Doll R. and Pygott F.: Factors influencing the rate of healing ofgastric ulcers: admission to hospital, phenobarbitone and as-corbic acid. Lancet 1952; 1: 171-5.

86. Clain J. E., Malagelada J. R., Chadwick V. S. et al.: Bindingproperties in vitro of antacids for conjugated bile acids. Gastro-enterology 1977; 73: 556-9.

87. Begemann F., Bandomer G. and Schumpelik V.: Ziir Therapiedes duodenogastralen Refluxes: Adsorption von intragastralcnLysolecithin an Antacida. Z. Gastroenterol. 1978; S12.

88. Wenger J. and Heymsfield S.: Adsorption of bile by aluminiumhydroxide. J. Clin. Pharmacol. 1974; 14: 163-5.

89. Eastwood M. A. and Girdwood R. H.: Lignin: a bile-saltsequestrating agent. Lancet 1968; 2: 1170-1.

90. Bergen S. S., van Itallie T. B., Tennent D. M. et al.: Effect of ananion exchange resin on serum cholesterol in man. Circulation1959; 20: 981.

91. Kleckner F. S., Stabler E. J., Hartzell G. ct al.: Esophagitis andgastritis secondary to bile reflux. Gastroenterology 1972; 62: 890.

92. Earlam R. J. and Thomas M.: The clinical significance ofgallstones and their radiological investigation. Br. J. Surg. 1978;65: 164-7.

93. Price W. H.: Gall-bladder dyspepsia. Br. Med. J. 1963; 2: 138-41.94. Gunn A. and Keddie N.: Some clinical observations on patients

withhgallstones. Lancet 1972; 2: 239-41.95. Southam J. A.: The effects of cholecystectomy on oesophageal

symptoms. Br. J. Surg. 1969; 56: 671-2.96. MacLean A. G.: Pyloric function and gallstone dyspepsia. Br. J.

Surg. 1972; 59: 449-54.97. Earlam R. J.: Clinical Tests of Oesophageal Function. St Albans:

Crosby. Lockwood, Staples, 1976.98. Ostick D. G. and Cowley D.: The lower oesophageal sphincter in

patients with gallstones. Br. J. Surg. 1976; 63: 154.99. Bushkin F. L. and Woodward E. R.: Alkaline reflux esophagitis.

Major Probl. Clin. Surg. 1976; 20: 64-71.100. Heaton K. W.: Bile Salts in Health and Disease. Edinburgh:

Churchill Livingstone, 1972.101. Van Heerden J. A., Priestley J. H., Farrow E. M. et al.: Post

operative alkaline reflux gastritis. Surgical implications. Am. J.Surg. 1969; 118: 427-33.

102. Van Heerden J. A., Phillips S. F., Adson M. A. et al.: Post-operative reflux gastritis. Am. J. Surg. 1975; 129: 82-8.

103. Alexander-Williams J.: Duodenogastric reflux after gastricoperations. Br. J. Surg. 1981; 68: 685-7.

Paper accepted 2 February 1983.