Recurrent gastrointestinal bleeding in a ITlan with Osler-Weber-Rendu syndrome and intestinal varices

A patient with Osler-Weber-Rendu disease and idiopathic intestinal l'arices who presented with recurrent gastrointestinal hemorrhage and iron deficiency anemia is reported. The unique coexistence of these two rare disorders is disw~scd and the literature reviewed. Can J Gastroenterol 1988;2(2):65-66

until that time to hemorrhoids.A hemnrrhoidectomy did not influence the al-mo~l monthly epi~odcs of bleeding.
At the time of his initial assessment, and on many occasions su bsequently, he had developed profound iron deficiency anemia, often to the range of 60 to 70 g/L, and not compensated for by chronic oral iron therapy.He had developed unstable angina on several occasions during periods of worsening anemia and, in 1986, underwen t coronary arte r y bypass graft surgery for occlusive coronary artery disease.At no time over the past 10 years had the patien t been seen and assessed while actively bleeding, bu t scools had been positive for occult blood on several occasions.The remainder of the h istor y, physical examination and laboratory investigations were normal.In particular, the re was no cl in ical or biochemical evidence of unde rl ying liver disease or por tal hypertension.T here was no suggestion o f coagulopathy at any time.T he re had been no previous abdominal surge ry.
The patient had been extensively in-vc~tigated on several occasions, particu-Gc1srro111rc,rinal D1.1ctrsc., Research UnH <1nd 1he De/Jartmcnr of Medicme, Q11een's University, larly over the past year, during which time he h ad been tra nsfused with over 20 units of packed red b lood cells.Esophagogastroduodenoscopy, small bowel enema and air con trast bariu m enema had failed to demonstrate any abnormality.In particular no varices were visualized in the esophagus, stomach or duodenum during upper endoscopy.At colonoscopy large colonic varice~ ( Figure I) were observed d iffuse ly throughout the left and righ t colon and also in the terminal ileum (Figure 2).In add ition, there were l to 2 mm diameter tclangiectatic vascular lesions scattered in the region of the splenic fl exu re and descendin g colon (Figure 3 ).Several of these lesions were coagulated w ith a bicap electrode, (ACM!, BC-60B BICAP), despite uncertainty that these were che cause of the gastrointestinal hemorrhage.
During the same admission a superior mesen teric artery angiogram, with delayed venous phase, fai led to identify any abnormality.Ultrasound o f the abdomen showed the portal vein to be pate n t and the spleen to be of normal size.The liver parenchyma was unremarkable.h owever, a 3 to 4 cm d iameter hyperchoic lesion was detected in the right lobe.A technetium 99 su lphur colloid liver-spleen scan also demonstrated a 3 cm filling defect in the right lobe of the liver.Otherwise, the hepatic uptake of technetium 99 was homogeneous and there was no bone marrow u ptake or splenomegaly to suggest portal hypertension.Subsequently a pooled techne-F igu re I ) E11Jmco/>1c 1•1eu• o/mnccs m rhc as-,;~11Jm.~col,m tium red blotid cel l scan was perfo rmed which revealed a 'delayed hot spot' in the right lobe sugges ting that the liver lesion was a hemangioma.

DISCUSSION
The associmion between Osler-Wcber-Rend u hereditary telangiectasia and colonic variccs has never been reported.
Colonic variccs arc relatively rare bcmg fo u nd in only 0.07°1, o f 2912 consecutive auwpsics reported by Feldman ct al ( I ).In the majority of cases, colonic var-icc~ arc.,1ttriburnblc to por tal hypertension (2), usually in the setting of alcoholic liver disease.Less common causes of colon ic variccs include large bowel venous anomalies (3) and selective mcscn te ric vein obstruction ( 4 ).Rarely, chronic heart failure or constrictive pericarditis may be responsible fo r impaired colonic venous return and result in colo n ic variccs (2).
The management of recu rren t colon ic variccal blceding from portal hypertension is either resection o f the affected portion of bowel, or portosystemic decompressive s hunt ( 5).Va rices arc even less likely robe found F ig ure 2) EnJo,copic 1•1cu• 0/mnce.1 in rhc wrmmal ileum in the small bowel, as was demonstrated in the present patient.Falch uk and colleagues ( 6 J reported gastrointestina l bleeding from ilea! varices, which were believed to ,irise as a result of localized mcscnteric hypertension due to ad hesions.Except for such unusual circumsta n ces, the small bowel is ge n erall y spared extensive portosyscemic collaterals, w h ich are common in the cardiacazygous system.abdomina l wall and rectum (7).A lthough ch is was nor the case in the present patient, small bowel varices can be demonstrated by enteroclysis (8).Active bleeding from intestinal varices can be d iagnosed by scintigraphy, using labelled red b lood cells, if the examination is performed during an episode of active hemorrhage (9).
The only report which links colonic variccs a nd benign vascu lar disease of the bowel is ch ar of Lieberman a nd others (3 ).These auch0rs describe a patient with severe lower gastrointestinal hlceding, rectosigmoid varices and Gtvcrn0us hemangiomas involving the skin, l0wcr extremities and distal colon.The authors suggest th:u the underlying cavernous hemangi0ma with associated venous proli(eration prod uced the colonoscopically visuali zed varices in the rectosigmoiJ .T he p resen t patient presented with two rare intesti nal conditions, and the probability of these co-existi ng due to chance alone is quite small.Thus it is tempting to speculate that the two arc interrelated.It is possible that arreriovenous shu nting a~sociated with in te~tinal tclangiectasias resu lts in increased flow and ,cc• ondary dilatation of intestinal vei n,, however, the relatively ~mall number of colonic tclangicccasias seen in the patient would argue against this hypothesis.Neverth eless, it is possible thar more extensive relangiectactic lc~ions were rrescnt deep to the mucosa and out of view of the endoscope.
Gi\Tn the uncerraintie~ regarding the exact si te and source ofhlecding, the authors in tend to manage thb patien t with hlood rra nsfusion as necessary, and u ndcnake colonic or small bowel rcsccrion only for an cxsanguinating bleed or, if they .ireable to loca lize a foca l bleeding so urce.by angiography or red blood cell bleeding ,can.