549

BENIGN DUODENO-COLIC FISTULAEM. J. GRAYSON,* M.B., M.R.C.P., N. D. O'CONNELL,t M.B., D.M.R.D.

Addenbrooke's Hospital, Cambridge

In I857 Dr. Charles Murchison,15 in a reviewof alimentary fistulae, quoted a case in which'the duodenum communicated with the colonthrough the gall bladder, the nature of the diseasenot being stated '. However, the first fully docu-mented case of benign duodeno-colic fistulawhich we have been able to trace was reportedby Sanderson in I863.28 The majority of thereported fistulae between the duodenum andcolon have been malignant and benign duodeno-colic fistulae are rare. In a review of the literaturewe have been able to find only 25 cases (see table).There are some other cases in the literaturewhich we have omitted from the table for thefollowing reasons. We have not included the casereported by Blondeaul (quoted by Clayton andThornton2), as it was thought to be malignant.It would seem that the case reported by Krock6is the same as that of Olson.19 Garland andWyatt4 reported a case following surgery for pan-creatitis, but the site of the fistula would seem tobe in doubt. Judd and Burden5 reported twocases of internal biliary fistulae where the duo-denum connected through the gall bladder to thecolon, but we are not including these two cases, asthey were not fully documented. For a similarreason we are not including the two separate casesmentioned by Lyons9 and Crill3 in the discussionfollowing the case report of Olson.19 We describebelow two further cases of this uncommon con-dition and discuss the clinical features anddiagnosis.Case IThe patient, a 40-year-old male clerk, was admitted

on 19.9.58 from the out-patient clinic with a historyof diarrhoea and abdominal pain. Two and a halfyears previously he had had intermittent lower abdominalpain with some looseness of the stools lasting for abouta week. Six months later he had a second episode,also lasting for a week. Four months before he presentedat the clinic he had a further attack of severe colickylower abdominal pain, eased by passing flatus and associ-ated with a feeling of abdominal distension. At firsthe had a normal, formed daily motion, but later thestools became fluid, although only once per day.During this illness he lost 20 lb. (9. kg.) in weight

* At present at St. George's Hospital, London.t At present at Cardiff Royal Infirmary.

and had a poor appetite. After five weeks his symptomssubsided until three weeks before he was admittedwhen, following a meal of mussels, he had a recurrenceof the lower abdominal pain and his daily stool becameloose. On no occasion had the stool been more frequentthan once daily and he had noticed no blood. Twoweeks prior to admission he vomited some brown fluidon two occasions and once a small clot of blood. Eightdays before admission his stools became formed, threedays later his abdominal pain ceased, and for four dayshe had no bowel action. During this last attack helost a further 20 lb. (9.1 kg.) in weight. He had at notime suffered from dyspepsia.

Previous History. No serious illnesses.On examination he was an ill-looking man, who had

obviously lost weight. His tongue was furred and hismucous membranes were slightly pale. His bloodpressure was II5/65 and there were no cardiac, respira-tory or neurological abnormalities. There was noabnormal finding on abdominal examination. Rectalexamination was normal. Sigmoidoscopy on 22.9.58showed a rather dry, granular mucosa which was undulyreddened, but not friable. It was thought that theappearances were abnormal, but that they were nottypical of ulcerative colitis. A mucosal biopsy was takenfrom the sigmoid colon. On histological examinationthis revealed an excess of inflammatory cells, but noabnormality of the glandular pattern.

Laboratory Investigations. Haemoglobin I2.5 g. pero00 ml.; mean corpuscular haemoglobin concentration29.7%; blood film, normochromia and some aniso-cytosis; white cell count I3,000 per mm.3 with 78%polymorphonuclear neutrophil leucocytes; E.S.R. (Win-trobe) 45 mm. in the first hour; stool culture, nopathogens isolated.

Radiology. On 26.9.58 a barium enema examinationdemonstrated a fistula between the colon and the upperalimentary tract. There was no other colonic lesionand, in particular, no radiological evidence of ulcerativecolitis. A chest radiograph was normal. On I.I0.58a barium meal examination showed no abnormality ofthe oesophagus or stomach. The duodenal cap wasdeformed and an ulcer was present in the second partof the duodenum with kinking of the duodenum andsome hold-up of barium at this site. The remainder ofthe duodenum emptied rapidly and showed a jaggedcontour. No fistula was seen at this examination. On6.10.58 a repeat barium enema examination demon-strated a fistula between the region of the ulcer in thesecond part of the duodenum and the proximal part of thetransverse colon (Fig. i).

Course. For the first three days the patient had alow-grade pyrexia of 99.6°F. (37.6°C.) and had nobowel action. Thereafter he was apyrexial and had adaily formed motion.

After preliminary phthalyl-sulphathiazole and oralstreptomycin, a laparotomy was performed on 7.I0.58

copyright. on M

ay 8, 2020 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.36.419.549 on 1 S

eptember 1960. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

FIG. i.-Case i. Barium enema examination showingfilling of the duodenal cap and the distal part of thestomach from the colon.

by Mr. P. H. R. Ghey through a right paramedianincision. The hepatic flexure of the colon was boundto the duodenum by many adhesions. Dissection dis-closed a fistulous track from the second part of theduodenum to the region of the proximal transversecolon. This was divided and the colonic opening closed.A Polya gastrectomy was then performed with closureof the duodenal stump just distal to the ulcer.The post-operative course was uneventful and the

patient is now symptom-free and has regained hisnormal weight.Case 2The patient, a farmer aged 33 years, attended the

out-patients' department on 30.6.55, complaining ofdiarrhoea for seven weeks, peri-umbilical pain on de-faecation and loss of weight. He had seven or eightpale motions at night but not by day and there was noslime or blood present in the stool. There was novomiting. He had lost a lot of weight but was uncertainof the amount. During the year preceding his attendanceat clinic he had had several attacks of abdominal painwith slight tenderness in the umbilical region and wasadmitted to another hospital during one such attack,being diagnosed as a case of colitis. At no time duringthese attacks was he pyrexial. Five months previously

FIG. 2.-Case 2. Barium enema examination outliningthe obstruc ive lesion in the proximal transversecolon.

he was treated for a lumbar disc lesion and in February,I955, he had an acute attack of dysentery when ShigellaSonne was isolated from his stools. It was subsequentlyproved that his whole family were also infected but weresymptom-free. Since then, two negative stool cultureshad been obtained from the patient.

Previous History. No serious illnesses.On examination the patient was thin, wasted and

definitely ill with some distension of the lower abdomen.Peristalsis was active. No anaemia. B.P. 2zo/80.Rectal examination was normal and sigmoidoscopy wasnot performed. The stools were clay coloured.He was admitted to hospital for investigation. A

barium enema on x5.7.55 showed an obstructive lesionin the proximal transverse colon (Fig. 2).At operation on 21.7.55 the surgeon, Mr. J. F. R.

Withycombe, found a large mass of inflammatory tissueinvolving the caecum and ascending colon with afistulous connection between the duodenum and theregion of the hepatic flexure. The fistula was closed anda right hemicolectomy was performed with anastomosisof the terminal ileum to the mid-transverse colon.Following the operation he did well until the tenth post-operative day when he developed a subcutaneous abscess

September i 96o550copyright.

on May 8, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.419.549 on 1 Septem

ber 1960. Dow

nloaded from

GRAYSON and O'CONNELL: Benign Duodeno-Colic Fistulae

Clinical Features Site of FistulaMethod of

Author Sex Age Aetiology Diar- Wt. Diagnosis Part ofPain Vom. rhoea Loss Duodenum Colon

Sanderson 1863 M 30 Duodenal ulcer + ++ - ? Autopsy First Transverse(faecal)

Rees 933 F 62 ? Typhoid ulcer + + - + Operation Second TransverseOrmandy and M 31 Ulcerative colitis + + + ++ Autopsy First Transverse

Bargen 1939 (faecal)McPeak 1940 M 8 Duodenal ulcer - + + + Barium enema First Hepatic

flexureMcPeak 1940 M 46 Duodenal ulcer + - ++ + Barium enema Second/ Transverse

ThirdMcClinton 1944 M 47 Duodenal ulcer + + + + Barium enema Second Transverse

(faecal)Lovell 1947 M 55 Cholecystitis + +4 ? + Barium enema First Transverse

(faecal)Masters 1948 M 40 Regional ileitis + + ++ Operation Second TransverseRailton 1948 F 45 Duodenal ulcer + + - ? Barium meal First TransverseOgilvie 1950 F 52 Tuberculous gland - - ++ + Barium meal Third AscendingOgilvie 1950 M 54 Tuberculous gland - + + + + + Barium enema Third AscendingRife I95I M 45 Duodenal ulcer - - ++ + Barium meal Second Transverse

and enemaRansom 1951 M 28 Ulcerative colitis - - + + Barium enema Unknown Transverse

(blood)Ransom 1951 M 12 Ulcerative colitis ? - + + Barium enema Unknown Transverse

(blood)Winfield I95I M ? ?- - + ? Barium enema Third SigmoidOlson 1951 F 68 ? Duodenal ulcer + + + + Barium meal Second Hepatic

(faecal) (blood) and enema flexureClayton and M 48 Appendicitis + - + + + Barium meal Second/ CaecumThornton 1953 ? Duodenal ulcer and enema Third

Neville 1954 M ? Cholecystitis and gall + + - - Barium enema First Hepaticstones (faecal) flexure

Rosenqvist and F 26 ? Foreign body - + ++ + Barium enema Second/ TransverseSjoberg 1955 ? Tuberculous gland Third

Rosenqvist and M 61 Cholecystitis and gall + - ++ + Barium enema First TransverseSjoberg 1955 stones

Nash and M 39 Duodenal ulcer + + - + Operation Second HepaticDaland 1956 flexure

Pautler I958 M 72 Duodenal ulcer ? + + + + Barium enema First TransverseMichell 1959 M 43 ? - - ? Barium enema Second Transverse

(blood)Own case I M 40 Duodenal ulcer + - 4++ + Barium enema Second TransverseOwn case 2 M 33 Regional ileitis + - + + Operation ? Hepatic

______ flexureTABLE I.

DETAILS OF REPORTED CASES OF DUODENO-COLIC FISTULAE

which was drained satisfactorily. He was dischargedfrom hospital a week later, apyrexial, gaining weightand feeling very well. He has had no further com-plications.

Pathology Report (Dr. R. A. Parker). ' The specimenincluded caecum, appendix, terminal ileum and theproximal x6 cm. of colon. Six cm. from the ileo-caecalvalve, in the narrowed segment, there was considerablethickening of the bowel wall apparently due to sub-mucous fibrosis, without complete destruction of themuscularis mucosae. Outside the bowel wall in thisregion a certain amount of fibrosis was present in themesocolon. Similar submucosal thickening was presentin the proximal 2 cm. of the appendix and in the distal4.5 cm. of the ileum. A number of superficial ulcerswere also present in the ileum, caecum, appendix andascending colon with shelving edges and with necrotic,haemorrhagic floors formed by the submucous layer.One such ulcer was situated in the narrowed segmentmentioned above and led to a fistula which communi-cated with the duodenum. The histological appearanceswere those of a chronic inflammatory lesion, composedmainly of follicular aggregations of lymphocytes withcentres composed of epithelial cells'. The lesion wasdiagnosed as regional ileitis.

Clinical FeaturesAetiology

It will be seen from Table i that 20 of thereported cases of benign duodeno-colic fistulaeoccurred in men, and five in women. Such fistulae

can occur at any age ranging in the above reportedcases from I2 to 72 years.As might be expected, duodenal ulceration is

the commonest cause of the fistulae, being thecause in nine of the above reported cases and apossible cause in two others. Cholecystitis andulcerative colitis have each been reported in threecases, and regional ileitis in a further two. Tuber-culous glands, typhoid ulceration, appendicitis,duodenal diverticulum and foreign bodies havebeen postulated as other causes.

SymptomatologyThe typical features of duodeno-colic fistulae

include diarrhoea, weight loss, vomiting andabdominal pain. Diarrhoea may be gross and wasprominent in 19 out of the 25 reported cases.Understandably, blood was present with the stoolsin two of the cases of ulcerative colitis, but it wasalso noted in two other cases where the aetiologywas not due to colonic disease. This point isdiscussed later. In such fistulae it might beexpected that the stool would contain undigestedfood, but this has not been reported.Weight loss may be severe, ranging from i lb.

September 960 55Icopyright.

on May 8, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.419.549 on 1 Septem

ber 1960. Dow

nloaded from

552 POSTGRADUATE MEDICAL JOURNAL September 1960

(5 kg.) to 84 lb. (38.2 kg.) in 20 out of the 25above cases.Vomiting was not so frequent, being reported in

only 12 cases, but when the vomitus was faecal,as it was in half of these cases, it was a usefuldiagnostic point. Pain was rather non-specific andseemed in most cases to depend on the cause ofthe fistula.

DiagnosisIn our first case the provisional diagnosis was

that of ulcerative colitis, although the weight lossseemed out of proportion to the severity of thediarrhoea. Although the sigmoidoscopic appear-ances were abnormal, they were not quite typicalof this condition. A mucosal biopsy was takenwhich showed inflammation, but no diagnosticfeatures of ulcerative colitis. Barium enema showeda duodeno-colic fistula, although this was notdemonstrated by a barium meal examination.

In our second case an inflammatory intestinallesion was suspected on clinical grounds. Anobstructive lesion in the ascending colon wasshown on barium enema, but the duodeno-colicfistula was not revealed. In this case the diagnosiswas made at operation.

In the 25 reported cases, as shown in the table,I9 were diagnosed by barium examination. Ofthe others, four of the fistulae were demonstratedfirst at operation and the remaining two at autopsy.The fistula was demonstrated by barium enema

examination in 17 (85%) of the 20 cases where itwas performed, whereas i6 barium meal examina-tions showed the fistulae in only six (37.5%) cases.On no occasion did a barium meal examinationreveal a fistula which a barium enema had failedto show. These results confirm the contention ofMedhurstl1 that the most useful diagnostic methodis barium enema examination. This is probablydue to the higher pressure gradient that obtainsbetween the two viscera on barium enema ex-amination.

TreatmentOperative treatment was performed in 22 out of

25 cases with only one post-operative death, dueto a pelvic abscess. Operative procedures includedsimple ligation or excision of the fistula, sometimescombined with partial or total colectomy or withgastro-enterostomy or partial gastrectomy. At thetime of the original reports the 21 patients were ingood health. Death ensued in the three patientswhere operation was not performed.Discussion

Diarrhoea and loss of weight are often thepresenting complaints in patients with fistulaebetween the upper alimentary tract and the colon.

Until recent years the explanation of these featureswas assumed to be the short-circuiting of food andgastric secretions into the colon. Convincingevidence, however, has been presented by severalauthors that the diarrhoea is due to retrogradepassage of colonic contents through the fistulawith resultant contamination of the small bowel.Lowden8 reports an inflammatory jejunitis andsteatorrhoea in cases of gastro-jejunocolic fistulae.Rosenqvist and Sjoberg26 also believe that theclinical features of diarrhoea and malabsorptionin such fistulae are due to jejunitis with resultantrapid passage of food through the small bowel.

In our first case, however, there was histologicalinflammation of the recto-sigmoid mucosa. Whilstaccepting that the main cause of the diarrhoea induodeno-colic fistulae is retrograde passage offaeces through the fistulac with resultant jejunitis,we feel that sometimes the pressure in the duo-denum may exceed that of the colon, causing inter-mittent flow of acid gastric juice into the colon withresultant colitis and that this may contribute to thediarrhoea. In two other reported cases, withoutcolonic disease, where blood was present in thestool, it may be that there was also some acid-induced colitis.

SummaryThe literature of benign duodeno-colic fistulae is

reviewed and two further cases are added. Thecondition is rare, the commonest cause beingduodenal ulceration. The patient usually com-plains of diarrhoea with associated weight loss,which may be severe. The method of choice ofdiagnosis is by barium enema. The condition isusually curable by surgery. The mechanism of thediarrhoea in this condition is discussed.

AcknowledgmentsWe should like to thank Dr. A. P. Dick, Dr.

L. B. Cole, Mr. P. H. R. Ghey and Mr. J. F. R.Withycombe for their kind permission to publishthe case reports, and Dr. R. A. Parker for thepathological details. We are grateful to Drs. A. P.Dick and F. R. Berridge for their criticism andadvice.

REFERENCESI. BLONDEAU, A., DERRIEU and DE LAROQUETTEMIRAMOND (I935), cited by Clayton and Thornton.2. CLAYTON, R. S., and THORNTON, W. L. (1953), Radiology,60, 832.3. CRILL, G. (i95i), quoted by Olson, personal communication.4. GARLAND, J. G., and WYATT, K. (I9SI), Rocky Mtn med.J., 48 426.5. JUDD, E. S., and BURDEN, V. G. (1925), Ann. Surg., 81, 305.6. KROCK, F. (I95s), in discussion with Ogilvie.7. LOVELL, D. L. (1947), Sth. Surg., 13, IS2.8. LOWDEN, A. G. R. (I953), Brit.J. Surg., 4x, xI3.9. LYONS, C. (I951), quoted by Olson, personal communication.

xo. MASTERS, H. (1948), J. Mt Sinai Hosp., IS, 264.i . McCLINTON, J. B. (I944), Canad. med. Ass. J., 5S, 434.Continued on page 556.

copyright. on M

ay 8, 2020 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.36.419.549 on 1 S

eptember 1960. D

ownloaded from

556 POSTGRADUATE MEDICAL JOURNAL September I960

normal, and this symptom cannot be attributed tocontinuing glandular fever. Cases 8 and :9 havebeen included to show the dramatic effect ofprednisone subjectively and on the pyrexia, butthey are too recent for continued improvement tobe reported. We can vouch for the good healthof Case io who is a medical colleague.DosageOur patients usually started with between

20-60 mg. prednisone daily in divided dosage, tobe reduced slowly over the next few weeks.Case 3 might have responded to continued pred-nisone but we felt bound to change to ACTH forthe reasons stated. The apparent superiority ofACTH over oral steroids here, and in otherdiseases such as ulcerative colitis, may well lie inthe matter of effective dosage rather than anyspecific therapeutic effect. We consider that someauthors have not used a sufficient initial dose, norhave they continued the therapy for long enough,and for this reason we would not agree with Masonand Adams7 that the speed of recovery is similarto those cases treated with bed rest only. We alsofeel that the dosage should be lessened slowly, andCases 2 and 7 demonstrate the relapse whichoccurs if the drug is withdrawn too quickly.

Until more is known of the cause and mechan-isms involved in glandular fever, any interpretationof the apparent beneficial effect of steroids in thetreatment of the disease, especially of its complica-tions, can be no more than conjecture. We suspect

that the benefit may arise from the known hypo-plastic effect of steroids on lymphatic tissue, onhepatitis and encephalitis, as' antistressor ' agents,and possibly by reduction of unhelpful antigen-antibody responses.SummaryTen patients with glandular fever are reported

in whom steroid therapy had a beneficial effect;three of these had severe complications.The indications for steroid treatment are

discussed together with suggestions for effectivetherapy.AcknowledgmentWe wish to thank Dr. Ronald Jones for per-

mission to describe three of the cases who wereadmitted under his care, and Miss London forsecretarial assistance.

REFERENCESI. BARRETT, A. M. (1941), J. Hyg. (Lond.), 41, 330.2. CREDITOR, M. C., and McCURDY, H. W. (I959), Ann.

intern. Med., 50, 218.3. DORAN, J. K., and WEISBERGER, A. S. (I953), Ibid., 38,

Io58.4. FIESE, M. J., CHEN, S., and RADDING, J. (1953), Arch.

intern. Med., 92, 438.5. FREEMAN, T., and WAKEFIELD, G. S. (1958), Lancet, ii,

883.6. MANDEL, W., MARILLEY, R. J., and GAINES, L. M., jun.

(i955), J. Amer. med. Ass., I58, I021.7. MASON, W. R., and ADAMS, E. K. (x958), Amer. J. nted. Sci.,

236, 447.8. NELSEN, R. S., and DARRAGH, J. H. (1956), Amsr. J. Med.

21, 26.

NOTICE OF SPECIAL INTEREST TO SUBSCRIBERS: WH NOT'WHY NOT HAVE YOUR COPIES OF THISJOURNAL BOUND INTO YEARLY VOLUMES?' HAVE

You can have your twelve monthly iuues fully bouna in dark green pin headdoth. lettered in gilt on spine with name of Journal, Volume Number and year,complete with index at front, for 22s. 6d. post free. A limited number of out ofprint journals are available to bind into volumes snd make your library complete.Price on application giving detals of issues required to complete back volumes.

THE FELLOWSHIP OF POSTGRADUATE MEDICINE BOUN60 PORTLAND PLACE, LONDON. W.I

References continued from paze ;2-M. 7. Gravson, M.B., M.R.C.P., and N. D. O'Connell, M.B., D.M.R.D.12. McPEAK, C. N. (1940), Radiology, 34, 343.I3. MEDHURST, G. A. (1956), Brit. J. Radiol., 29, 381.14. MICHELL, R. C. (I959), Ibid., 32, 55.15. MURCHISON, C. (x857), Edinb. med. J., 3, 4.x6. NASH, H. E., and DALAND, E. M. (1956), New Engl. J.

Med., 254, 1032.17. NEVILLE, W. E. (I954), Amer. J. Surg., 87, 300.x8. OGILVIE, SIR H. (1950), Ann. Surg., 131, 899.I9. OLSON, J. D. (I95s), Amer. Surg., 17, 335.2o. ORMANDY, L., and BARGEN, J. A. (1939), Proc. Mayo Clin.,

14, 550.

21. PAUTLER, E. E., WODDALL, J. C., and GAITHER, J. G.(1958), A.M.A. Arch. intern. Med., 102, 207.

22. RAILTON, S. V. (I948), Canad. med. Ass. J., 59, 267.23. RANSOM, H. K. (I951), cited by Rife.24. REES, C. E. (1933), J. Amer. med. Ass., 1oo, 496.i2. RIFE, C. S. (1951), Arch. Surg. (Chicago), 62, 876.26. ROSENQVIST, H., and SJOBERG, S. G. (I95s), Gastro-

enterologia (Basel), 82, 285.27. ROSENQVIST, H., and SJOBERG, S. G. (1955), Acta chir.

scand., 109, 293.28. SANDERSON (1863), Trans. path. Soc. Lond., 14, I73.29. WINFIELD, J. M. (i95I), cited by Rife.

copyright. on M

ay 8, 2020 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.36.419.549 on 1 S

eptember 1960. D

ownloaded from