5"I

THE RESULT OF VAGOTOMY ANDGASTROENTEROSTOMY AFTER

THREE TO FIVE YEARSBy B. V. McEvEDY, F.R.C.S.

Senior Registrar, University College Hospital, London

and G. K. KIRKLAND, M.B., Ch.B., D.P.H.Assistant Surgeon, Oldham Royal Infirmary and Ancoats Hospital

It is only iI years since Lester Dragstedt re-introduced vagal section in the treatment of pepticulcers, yet in this short period, after attaining apopularity almost equal to that of partial gas-trectomy, the operation has fallen into disreputeand is now rarely performed. Despite the manyfailures large numbers of patients have been curedby vagotomy although the different factors pro-ducing success or failu"re have not been adequatelyelucidated. In the three years prior to I948 it wasshown that vagotomy was contra-indicated for-gastric ulcers and that for duodenal ulcers agastroenterostomy must also be performed toovercome the gastric stasis produced by paralysisof the stomach. This paper sets out the results ofone surgical team between I948 and I95I, duringwhich period all duodenal ulcers were treated byvagal section with, in all cases, the vagotomyaccompanied by a gastroenterostomy.The technique of the operation was essentially

that put forward by Dragstedt (I947). The gastro-enterostomy was placed within 2 in. of the pylorusas on barium meal, this was found to be the mostdependent part of the stomach; with the classicaltype of gastroenterostomy on an atonic stomach apouch is formed between the stoma and the pylorusin which food is retained for long periods. All theoperations were performed under local anaesthesiaas this was thought to be the safest of allanaesthetics.

The Follow-up SeriesIn this paper are considered i84 cases of peptic

ulcers operated upon by the late Peter G. McEvedyand one of the writers (G.K.K.). The operationswere performed between 1948 and July I951, sothat all the patients have been followed for overthree years and some for more than six years. Thefollow-up has been largely by questionnaire, butnearly a quarter of the cases have been seen

personally; while questionnaire follow-up hasbeen criticized as inaccurate, it is of interest thatno case seen was worse than the questionnaire hadimplied while several were better. The series con-sisted of I75 duodenal ulcer and nine jejunal ulcerpatients.Of the I84 cases operated upon two died in the

immediate post-operative period, both from in-testinal obstruction. One was a case of malrotationof the gut and death was from massive volvulus,while in the second case the obstruction was dueto old tuberculous adhesion4 around the terminalileum. This mortality of just over i per cent. issimilar to that reported in many series.

Duodenal UlcersOf the I75 duodenal ulcer patients, six are ex-

cluded from the series as their letters were returnedundelivered, while seven further patients failed toreply, so that a 92.5 per cent. follow-up wasobtained (i6z cases). Seven patients have diedsince their operations from causes unrelated totheir peptic ulcers or their operation but un-fortunately none were submitted to post-mortemexamination. Two patients are in sanatoria andso are not subjected to conditions where theirulcers might be troublesome. There were twooperative deaths. These i i cases are excludedfrom the follow-up. Thus i i cases (86 per cent.)of duodenal ulcers are reviewed.The results were classified into good, improved

and bad categories. The good results were thosein which there were no complaints and the patientswere satisfied with the operation. In the im-proved cases were included those who, althoughable to work and getting no pain of indigestion,complained ofsymptoms that prevented them fromforgetting about their stomach completely. Thebad results were where the ulcer pain had per-sisted, further operations had been required or

Protected by copyright.

on January 14, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.31.360.511 on 1 October 1955. D

ownloaded from

512 POSTGRADUATE MEDICAL JOURNAL October 955

where it was found impossible to lead a normallife as a result of abdominal symptoms.

I949 1950 I95I TotalTotal 65 53 33 I5IGood 53 8i.5% 47 88.7% 30 90.9% I30 86.i%Improved 9 13.9% 2 3.8% I 3% I2 7.9%Bad 3 4.6% 4 7.5% 2 6.I% 9 6%

In reviewing this series it is important to com-pare the results with those of duodenal ulcerstreated by gastroenterostomy alone. Clark (I951)in his follow-up of over 300 gastroenterostomycases traced for I5 to 20 years, found that thegreat majority of anastomotic ulcers occurred inthe first three years and after about ten years thosecases that were symptom free almost always -re-mained well. Classifying his cases in a similarmanner to ours, his results are: Good, 68 per cent.;improved, 9 per cent.; bad, 23 per cent.

Cooper (1,948), in a ten-year follow-up of 257duodenal ulcers treated by gastroenterostomyalone, found that the poor results were practicallyalways due to a jejunal ulcer and that an activeduodenal ulcer in the presence of a functioningstoma was very rare. In the 49 failures (I9 percent.) of his series, jejunal ulcers were definitelypresent in 37 and probably present in a further 12;47 of the failures occurred in the first five yearsand 13 of them settled with medical care and werefinally classified as good results.

In our series time will show if the steady upwardtrend in the good result percentage is due to im-proving surgery or whether the jejunal ulcers havenot yet developed in the later cases. While theselate results in the treatment of duodenal ulcers byvagotomy and gastroenterostomy can be comparedfavourably with those after other forms of treat-ment, the percentage of failures is still high.A comparison of pre-operative findings and the

post-operative results shows that some slightselection of patients suitable for vagotomy ispossible. In the pre-operative investigation of acase of duodenal ulcer, certain features can bedetermined: Age, length of .history, presence ofpyloric stenosis, sex, history of haematemeses andthe results of the fractional test meal.

Age. Surgeons have always been chary ofoperating upon the younger patient with a duo-denal ulcer with the result that many young menhave been forced to lead unnaturally quiet livesuntil, after ten years or more, an operation hasreturned them to a normal life. In this series agehad little bearing on the likelihood of a goodresult ensuing.

Age Total Good Improved BadUnder 4o 57 49 86% 4 7% 4 7%Over 40 88 75 85.2% 6 6.8% 7 8%Over 6o 6 5 83.3% I I6.7% 0

Length of history. Similarly the duration of thedisease does not affect the result. Both thesefindings are unlike those seen after gastroenter-ostomy alone, where the results improve the olderthe patient and the longer the history.

History Total Good Improved BadUnders years 4o 36 90% I 2.5% 3 7.5%5 to lo years 42 35 83.3% 7 i6.7% 0Over Ioyears 69 59 85.5% 4 5.8% 6 8.7%

Pyloric stenosis. This has long been known togive excellent results with gastroenterostomy alone,the poor results usually being in those cases wherethe obstruction is due merely to temporary oedemaaround the ulcer.' If the pyloric stenosis isdiagnosed only on the clinical history of vomitingof large quantities of food, the presence of visibleperistalsis or the operative finding of a dilatedstomach with a hypertrophied wall, the resultswith gastroenterostomy alone are excellent. Theaddition of a vagotomy does not appear to affectthe result in these cases.

Stenosis Total Good Improved BadPresent 3I 29 93.6% I 3.2% I 3.2%Absent 120 0OI 84.2% II 9.2% 8 6.6%

Sex. Only I4 women were included in theseries, an incidence of 9 per cent. All had goodresults from the operation and although thenumber is too small to be of significance this issimilar to the results after gastroenterostomyalone, where women are found to be less liable tohave severe relapses.

Previous haematemeses. Haematemeses of in-sufficient severity to require immediate operationhad occurred in i i of our cases, of whom sevenhave obtained good results, two are improved andtwo are bad results. In'none had there been anyfurther haemorrhages but the number of cases istoo small to carry any statistical significance.

Test meal results. From the test meal two par-ticular points were determined; firstly the pre-sence of a hyperchlorhydria and secondly the levelof the resting acid; it was felt that the latter wasthe important reading in the test meal as it gavesome indication of the level of acidity whichbathed the duodenum in the hours between meals,while the hyperacidity at meals is at least onlytemporary and is buffered by the food. Un-fortunately in only 70 cases was a test meal per-formed pre-operatively but no significant differencecould be demonstrated in the results achieved inthe' different groups of gastric acidity.Test Meal Total Good Improved BadHyperchlor-

hydria 49 44 89.8% 3 6.I% 2 4.2%Normal acid 2I 17 8i% 2 9.5% 2 9.5%Resting acidHigh 38 3I 8i.6% 3 7.9% 4 IO.5%Normal 32 28 87.5% 3 9.4% I 3.I%

Protected by copyright.

on January 14, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.31.360.511 on 1 October 1955. D

ownloaded from

October 1955 McEVEDY and KIRKLAND: Result of Vagotomy After Three to Five. Years 5I3

A new book for the postgraduateOPHTHALMOLOGYA TEXTBOOK FOR DIPLOMA STUDENTSby P. D. TREVOR-ROPERM.B., B.Chir.(Camb.), F.R.C.S., D.O.M.S.(Eng.)Curator, Dept. of Pathology, Institute of Ophthalmology;Asst. Ophthalmic Surgeon, Westminster Hospital.This textbook aims to cover the field of ophthalmology required by the postgraduatestudent seeking an ophthalmic diploma. It deals at appropriate length, and in considerable XVl + 656 pp.detail, with the anatomy and physiology of the eye, with physical and physiological optics 449 ilustrationsincluding the prescription of spectacles, and then covers systematicaly the diseases of the 8 colour platesouter and inner eye. The emphasis falls throughout on clinical expernence. (1955) 75s. net

A new book for the specialistSTUDIES ON THE CEREBRAL CORTEXby S. RAMON Y CAJAL(Translation from the Spanish by Lisbeth M. Kraft, Yale University.)The timeliness of this translation is self-evident from the widespread and growing interest inthe limbic system. Nevertheless, a precise knowledge of anatomy is essential to a thoroughelucidation of function. When Ram6n y Cajal undertook and completed the series of xii + 179 pp.studies included in this volume he probably gave us the most precise and comprehensive 108 illustrationssingle anatomical account of the limbic system that is available. (1955) 27s. 6d. net

Descriptive leaflets available on request

LLOYD-LUKE (MEDICAL BOOKS) LTD., 49 Newman Street, W.1

Thus in the pre-operative selection of cases,while pyloric stenosis patients and women are moreliable to get a good result, youth and a short historydo not contra-indicate the operation. The onlyother pre-operative test that might be of use is thenight secretion volume and acidity but this was notperformed in our series. Vagotomy might befound to give better results in those patients with ahigh level of night secretion, where presumably thevagal tone is high, than in those with a normallevel, thus giving a useful pre-operative method ofselection of cases.

In the post-operative examination completevagal section is the main factor in deternmining theprognosis and the only means for deciding if thishas been accomplished is the insulin test. It hasbeen found that in the presence of hypoglycaemia^overactivity of the vagal centre in the hypothalamusoccurs with an increased secretion of gastric acid.'The test is performed on the resting stomach, theresting gastric juice acidity and the fasting bloodsugar are estimated, after which 15 units of solubleinsulin are injected intravenously. Any rise in thegastric acidity is taken to indicate that vagalimpulses are reaching the stomach. It has beenfound that to stimulate the hypothalamus theblood sugar must fall rapidly to below 40 mgm. percent, and even then 6 per cent. of the pre-operative

cases fail to show any rise in gastric acid. Thetest is purely qualitative and a positive result givesno indication of the number of vagal fibres thatare stimulating the stomach. A negative result isalso difficult to evaluate as the tube may not havebeen in the stomach or the patient may be one ofthe 6 per cent. who give a false negative unless apre-operative test has excluded these patients.Finally the test is not without discomfort to thepatient.

In our series the insulin test was performed upon74 cases and in these a rise in acidity occurred in25 per cent.; thus it appears that in at least aquarter of our cases vagal section was incomplete.Repeated insulin tests on a few patients during theyears of the follow-up showed that negative testsremained negative and that a test done in theimmediate post-operative period was quite re-liable; this also indicated that no vagal regenera-tion had occurred.

In comparing the results in those cases wherevagal section was incomplete with those where theinsulin test appeared to show that the vagotomywas complete, the results are definitely better inthe latter group. Many cases in which the sectionwas incomplete, however, give excellent results andthis is to be expected in a test that is qualitativerather than quantitative as in all cases a very

Protected by copyright.

on January 14, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.31.360.511 on 1 October 1955. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

careful vagal dissection had been performed andpresumably only a few small nerves had been leftin most patients.

VagotomyCompleteIncornplete

Total Good Improved Bad55 74.3% 50 90.9% 3 5.5% 2 3.6 /o19 25.6% 13 68.4% 5 26.3% I 5.3%

In the two bad results in the group where vagalsection appears to be complete, both requiredfurther operation for recurrence of their originalabdominal pain. In both their duodenal ulcershad healed and while in one no abnormality wasdiscovered at operation, in the other a hyper-trophic gastritis, noted at the time of his firstoperation, was still present and unchanged and wasthe probable cause of the symptoms. Neither ofthese patients was improved by their secondoperation.

In the bad result patient, whose vagotomy wasincomplete, a large jejunal ulcer was found at hissubsequent operation and this was treated by apartial gastrectomy; since that time he has beenvery well. This patient was operated upon by anassistant and was his only case in the series; vagalsection was probably very incomplete and this wasreally a duodenal ulcer treated by simple gastro-enterostomy. The poorer results in cases treatedby surgeons with only a small series has beennoted by Jordan (1952).The only other test that might be of use in

determining if the vagal section has been com-plete, is the post-operative night secretion estima-tion. This test was not performed in our series asit was felt that the results would be untrustworthydue to the gastroenterostomy and the reducedgastric tone and mobility; both of which wouldaffect the amount of the gastric aspirate. Howeverin a total vagotomy resting gastric section wouldbe practically negligible, while the amount ofsecretion in the other cases might give someindication of the degree of vagal section.Thus in the post-operative determination of the

prognosis, the insulin test is the main factor avail-able but if the operation has been performedthoroughly, even if the vagotomy is incomplete, ahigh proportion of good results will be achieved.

The bad results. Nine patients received badresults from their operation and in four of these asecond operation was performed. As alreadymentioned, in one a large jejunal ulcer was foundand in another a hypertrophic gastritis had failedto subside, but in all four cases the originalduodenal ulcer had healed. In the other two ofthe four operation cases no abnormality was dis-covered. In the five patients who have not hadany further operative treatment the original duo-denal ulcer appears to be active in one but in theother four cases no cause can be found to account

for their symptoms. Thus out of the ninefailures, in six there is no obvious abnormalitydespite thorough investigation. The symptomsof which they complain include nausea and vomit-ing, indigestion and typical ulcer pain. Probablythese symptoms are largely the result of faultymechanics due to the loss of tone and motility ofthe stomach with imperfect emptying and sub-sequent gastritis. The appearance of a gastritisis very difficult to be certain about and may easilybe missed at laparotomy or on examination of apartial gastrectomy specimen. The two patientswith negative laparotomies received no benefitfrom their partial gastrectomies and possibly evenafter resection stasis might still occur in the atonicstomach and prevent relief of symptoms. Bariumstudies in the six idiopathic bad results, however,failed to demonstrate any very marked gastricstasis, so as yet the cause of these failures isundetermined.

The improved cases. Most of the symptoms hereagain appeared to be due to mild mechanicaldefects as a result of the reduced gastric tone.The commonest trouble was a feeling of distensionafter meals with flatulence and belching; othercomplaints were very similar to the post-gastrec-tomy dumping syndrome, with faintness andsweating after large meals necessitating the patientlying down. In most of this group of patientssymptoms appeared to be getting less frequent,possibly as a result of recovery of gastric tone,avoidance of large meals and by gastric hyper-trophy overcoming any mechanical obstruction.

The good results. These were remarkable forthe excellence of their results, the immediatepost-operative relief of their symptoms and therapid increase in well-being. Most patients had agood appetite and enjoyed large meals and a fulllife. Many had put on weight and formed amarked contrast to the successful partial gastrec-tomy cases where often, although free from allpain and indigestion, only small meals can betaken, the appetite is poor or cannot be assuagedwithout discomfort and the patients fail to put onweight for many years. None of our cases com-plained of diarrhoea although many said that theirbowels were improved and that they resorted tolaxatives less frequently.

In comparing the results after vagotomy andgastroenterostomy with those after partial gastrec-tomy, three main features are available forcomparison; the mortality, the recurrence ofulceration and the associated effects of the opera-tion. The mortality in the best centres for partialgastrectomy is rarely under 3 per cent. and ismuch higher throughout the country; for vagotomyand gastroenterostomy a mortality of about i percent. as in this series is about that to be expected,

October 19 555I4

Protected by copyright.

on January 14, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.31.360.511 on 1 October 1955. D

ownloaded from

October I955 McEVEDY and KIRKLAND: Result of Vagotomy After Three to Five Years

while large series without any deaths are recorded(Dragstedt, 1950). The incidence of recurrentulceration after partial gastrectomy is low, prob-ably in the region of 3 per cent. in the centres wherea high resection is performed but higher after themore average resections performed elsewhere. Inour series the bad results formed 6 per cent. ofthe cases although, as explained, only one of thesewas due to a jejunal ulcer. Among the associatedeffects the ability of the patient to enjoy his life,work hard and be really fit must be taken intoaccount as well as the minor annoying featuressuch as post-prandial distention and the dumpingsyndrome. One of the most satisfactory featuresof our series was the general well-being of thepatients and this, together with the lower mor-tality and the nearly comparable percentage ofgood results makes the comparison of the twooperations more equal.

Gastrojejunal UlcersTreatment of this type of peptic ulcer remains

the one popular indication-for the operation; goodresults are reported by numerous authorities (Orrand Johnson, I949) and the effects are consideredby others to be superior to repeated resection inthose cases that follow partial gastrectomy (Walters,I955). It is this fact that makes it difficult tounderstand the poor regard for the operation inthe treatment of duodenal ulcers. If, as isgenerally admitted, the main failing of gastro-enterostomy for duodenal ulcers is the high in-cidence of jejunal ulcers and if the latter arebelieved to be cured by vagotomy, it is reasonableto assume that associating the gastroenterostomywith a vagotomy will prevent the formation ofjejunal ulcers and cure the duodenal ulcers.

In discussing jejunal ulcers two main varietiesmust be considered, those following gastro-enterostomy and those after partial gastrectomy.After partial gastrectomy they rarely occur unlessthe original lesion was a duodenal ulcer; aftergastroenterostomy they are less common in thepresence of pyloric stenosis, in elderly patients orif there is a long history of ulceration, all con-ditions where the acidity tends to be lowered.Nearly always the ulcer occurs in the efferent loopof the jejunum or on the line of the anastomosisand is usually quite small, only about 50 per cent.being demonstrable on barium studies.

In our series there were only seven jejunalulcers and all followed gastroenterostomy forduodenal ulcers; their second operations occurringone to nineteen years later, averaging six years.Six were treated by simple vagotomy and four ofthese have remained very well since their opera-tions five years ago; the seventh had his vagotomycombined with re-doing of the gastroenterostomy

and has also been very well for five years. Of theother two, one began to develop obstructive symp-toms after six months and at a further operationthe jejunal ulcer was healed but in healing hadobstructed the stoma; since partial gastrectomy hehas been very well. The other case originally hada gastroenterostomy in 1943 followed, in I947, byan enteroanastomosis and developed a jejunalulcer which was treated by vagotomy in I949;however his jejunal ulcer failed to heal and a yearlater he had a partial gastrectomy, since when hehas been very well. The association of the entero-anastomosis with a gastroenterostomy carries ahigh risk of jejunal ulceration and is almost com-parable with the Mann-Williamson operationi indogs. Maybe the presence of this enteroanasto-mosis explains the failure of the ulcer to heal, butin this one case the vagotomy was also incompleteas shown by the insulin test. No cases of jejunalulcer following partial gastrectomy were treated inthis series.Two cases of gastrojejuno-colic fistulae were

treated, both following gastroenterostomy forduodenal ulcer. The first had his original opera-tion in 1929 and reported in a poor general con-dition in 1950 after a year's diarrhoea; at operationhe had a very large jejunal ulcer with a largefistula into the colon, vagotomy was performed, thegastroenterostomy undone and a new one made.A catheter was placed in the fistula into the colonand brought out of the lower end of the wound;this fistula discharged very little and healedspontaneously. Since that time he has been verywell.The second case had his gastroenterostomy in

1944 and developed a jejunal ulcer in 1948; thiswas followed in two months by loss of his pain andonset of diarrhoea and foul eructations. He wastreated by vagotomy and division of the fistula intothe colon, which was drained by a catheter; thegastroenterostomy was left. Post-operatively hewas very well but unfortunately died four yearslater of carcinoma of the larynx.

These results for jejunal ulcers seem to confirmthe general opinion that vagotomy will heal theulcer, but fibrous obstruction in a small number ofcases suggests that the larger ulcers should beassociated with undoing of the original anasto-mosis and formation of a new stoma; undoubtedlyclosure of the stoma is inadequate as the resultantgastric stasis will produce the undesirable effectsof simple vagotomy. The excellent results of thissimple treatment of gastrojejuno-colic fistula isalso of importance, as usually these patients are ina parlous condition and the major operation ofgastric and colonic resection often advised carriesa high mortality.Bibliography on page 537

S51

Protected by copyright.

on January 14, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.31.360.511 on 1 October 1955. D

ownloaded from

October I955 PINKERTON and ANNAMUNTHODO: Post Eclamptic Anuria 537

possibility of necrosis and autodigestion in anaberrant pancreas producing a cyst in which notrace of the original lesion could be found.

(d) Cyst of Brunner's glands. Robertson (I94I)mentioned the possibility of a retention cyst ofBrunner's glands but stated that these cysts wereusually only a few millimetres in diameter and weresymptomless. Booher and Pack (1946) reported acase of cystic tumour of duodenum of this naturein a female aged 52.Adenoma of Brunner's glands are rare. They

present with haemorrhage or obstruction. Feyrter(1934) found three in 2,800 duodena examined.Cases have been reported by Willis and Lasersohn(1925) and Balfour and Henderson (I929). Theseadenomata may undergo cystic degeneration.There was no histological evidence of this in ourspecimen.

(e) Lymphatic cyst arising in the submucosallymphatics. Lymphatic cysts are not uncommonin the mesentery of the small intestine. We havenot been able to trace any case of submucosallymphatic cyst in the literature.The only similar case of duodenal cyst in the

literature was reported by Riker (I95I). In dis-cussing the probable aetiology he mentioned thepossibility of a duodenal cyst being the result of atiny mucosal perforation with abscess formationand organization of haematoma from spontaneoussubmucosal haemorrhage.From the histology it is evident that this cyst

had been present for some time. Haemorrhageoccurred into it probably at the time of theeclamptic fits. This would account for the rathermarked fall in blood pressure which succeeded theconvulsions, the early appearance of anaemia, thepain and tenderness in the upper abdomen whichwe had regarded as hepatic in origin, the slightjaundice and especially the incomplete duodenalobstruction which was such a dominant feature ofthe case.

Treatment of duodenal cysts is on the wholeextremely unsatisfactory. One reason for this isthat they usually occur in infants who, because ofprolonged vomiting, are poor operative risks.Mortality in infants is as high as 70 per cent.(Gordimer and Bluestone, I950), but it is lower inadults. In no case in the literature has duodenec-tomy been performed for this benign condition.

SummaryA case of post eclamptic anuria complicated by

duodenal obstruction due to haemorrhage into aduodenal cyst is described. The difficulties en-countered in the management of the case arediscussed. A review of the literature on duodenalcysts is given.

AcknowledgmentWe wish to thank Sister M. H. Crandall,

S.R.N., S.C.M.; Mr. R. J. Edwards, Drs. G.Bras and W. R. Cole of the Departments ofMedicine, Pathology and Radiology for their co-operation; and Professors D. B. Stewart andG. H. C. Ovens for their advice and encouragement.

BIBLIOGRAPHYBALFOUR, D. C., and HENDERSON, E. F. (1929), Ann. Surg.,

89, 30.BLACK, D. A. H., and PYRAH, L. N. (194), Proc. Roy. Soc.

Med., 47, 582.BOOHER, R. J., and PACK, G. T. (I946), Arch. Surg., 53, 588.BROOKS, B., and WEINSTEIN, A. (X943), Ibid., 117, 728.BULL, G. M., JOEKES, A. M., and LOWE, K. G. (I949), Lancet,

i, 229.CHALMERS, J. A., and FAWNS, H. T. (i9SS), Ibid., i, 79.FEYRTER, F. (i934), Virchows Arch. f. Path. Anat., 293, 509,

quoted by H. E. ROBERTSON.GAMBLE, J. L. (x947) ' Extracellular Fluid,' sth edition, Harvard

University Press. Massachusetts,GARDNER, C. E., and HART, D. (1935), J. Amer. med. Ass.,

I04, I809.GORDIMER, H., and BLUESTONE, L. (I950), Ann. Surg.,

132, 11 49.

HILDES, J. A., and FERGUSON, M. H. (1952), Gastroenterology,21, 64.

HOWARD, J. E., and CAREY, R. A. (I950), J. Clin. Endocrinol.,9, 69I.

LORBER, S. H., and MACHELIA,'T. E. (1948), Gastroenterology,Io, 892.

NEWMAN, E. A., and EISENSTEIN, M. (I949), Ibid., 13, 528.NORRIS, J. C. (1946), South med. ., 39, 549.PEPLE, W. L. (I948), Ann. Surg., 127, 912.POLSON, R. A., and ISSAC, J. E. (I953), Gastroenterology, 25,

431.RIKER, W. L. (19S1), Amer. 3. Surg., 8', 464.ROBERTSON, H. E. (I941), Arch. of Path., 31, 112.ROTH, M. (i88I), Virchows Arch. f. Path. Anat., 86, 371, quoted

by C. E. GARDNER and D. HART.RUSSELL, C. S., DEWHURST, C. J., and BRACE, J. C. (I954),

Lancet, i, 902.SANGER, M., and KLOPP, A. (i88o), Arch. f. Gynak.; I6j 4r5,

quoted by C. E. GARDNER and D. HART.SHALLOW, T. A., WAGNER, F. B., and MANGES, W. B.

(I947), Surg., 21, 532.WARREN, K. W. (19s5), Surg. Clin. N. America, 31, 789.WHEELER, W. I. DE C. (1940), Brit. J. Surg., 27, 446;WHIPPLE, A. 0. (1950), ('Non-ulcerative lesiors of the duo-

denum,' by Frederick Christopher), 'Textbook of Surgery,'5th edition, W. B. Saunders Co., Philadelphia.

WILLIS, A. M., and LASERSOHN, M. (I925), Ann. Surg., 82,952-

Bibliography continuedfrom page 515-B. V. McEvedy, F.R.C.S., and G. K. Kirkland, M.B., Ch.B., D.P.H.BIBLIOGRAPHY

CLARK, D. H. (Iusi), Brit. med. J3., i, 57.COOPER, W. A. (I948), Surgery, 23, 425.DRAGSTEDT, L. R., et al. (I947), Ann. Surg., I26, 687.DRAGSTEDT, L. R. (I950), Ibid., 132, 626.JORDAN, S. M. (1952), Gastro-Ent., 22, 295.ORR, I., and JOHNSON, H. D. (I949), Brit. med. Y., ii, 13I6.WALTERS, W., et al. (I955), Surg. Gyn. and Obstet., -Ioo, I.

Protected by copyright.

on January 14, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.31.360.511 on 1 October 1955. D

ownloaded from