Liver Histology and Follow up Of68 Patients With

468 Gut, 1990, 31, 468-472

Liver histology and follow up of 68 patients with ulcerative colitis and normal liver function tests Gut: first published as 10.1136/gut.31.4.468 on 1 April 1990. Downloaded from

U Broome', H Glaumann, R Hultcrantz

Abstract the colitis,4' but so far no such correlation has Hepatobiliary disorders are well known com- been found. plications in patients with ulcerative colitis but In 1966, primary sclerosing cholangitis was it is not possible to predict those patients with recognised in patients with ulcerative colitis.9 10 ulcerative colitis who will eventually devleop Initially, it was described as an unusual finding,8- liver disease. In this study, liver biopsies from 10 but with an increased awareness of the disease 74 patients with ulcerative colitis have been and with the availability of endoscopic retro- reevaluated. None of the patients showed grade cholangiography in addition to a possible clinical or biochemical signs of liver disease at absolute increase of the disease, primary scleros- the time of biopsy. Thirty seven (50%) had a ing cholangitis is today found in 2-4% of all completely normal liver biopsy. The others patients with ulcerative colitis." 12 Except for one showed minimal portal inflammation or fatty study by Perrett et al,3 who found minor histo- infiltration. The biopsies of three patients logical abnormalities in six of 13 patients with displayed concentric, periductular fibrosis, or ulcerative colitis without clinical or biochemical so called 'onion skin' lesions. None showed evidence of liver disease, little is known about other signs of fibrosis or cirrhosis. The histo- the incidence ofalterations in the liver in patients logical findings were unrelated to either with ulcerative colitis in general. Also, aetio- activity or extent ofcolitis, except for the onion logical factors leading to liver disease in patients skin lesions which were seen exclusively in with ulcerative colitis are not clarified. biopsies of patients with involvement of the In the present study liver biopsies from 74 total colon. Sixty eight of the 74 patients were patients suffering from ulcerative colitis with reviewed after a mean interval of 18 years. The normal liver function tests have been reexamined majority had no symptoms of hepatobiliary according to the new classification system ofliver disorders and only two had developed liver biopsies.'4 The patients were followed up in disease; one suffered from cryptogenic order to assess if patients with ulcerative colitis, http://gut.bmj.com/ cirrhosis, possibly due to non-A, non-B hepa- but lacking biochemical parameters of liver titis and the other of an autoimmune liver disease, run a lower risk of developing hepato- disease and later developed a bile duct car- biliary complications compared-with all patients cinoma. Both were women with total colonic with ulcerative colitis. We also wanted to estab- involvement. At the time of the first liver lish if any specific histological abnormality biopsy one showed no histological abnormali- would be of value in predicting the future risk ties and the other only minor fatty infiltration. that a patient with ulcerative colitis will develop on October 2, 2021 by guest. Protected copyright. Thus, minor abnormalities in liver tissue are liver disease. common in patients with ulcerative colitis without biochemical evidence of liver disease. The morphological changes are oflittle help in Methods predicting the future risk of a patient with ulcerative colitis developing a clinically rele- PATI ENTS vant hepatobiliary complication. The absence Between 1958 and 1968, at St Erik's Hospital in of biochemical parameters for liver disease Stockholm, Sweden, liver biopsies as well as during the early years of ulcerative colitis biochemical liver function tests were routinely predict a favourable longterm diagnosis as performed on patients with ulcerative colitis. In regards hepatobiliary complications. 78 of the patients, the liver function was more extensively examined and was found to be completely normal. Four of the patients later In 1874 Thomas' described an association developed Crohn's disease and were therefore between ulcerative colitis and liver disease. Since excluded from our present study. many reports have confirmed this 3 Clinical data were obtained from the hospital Departments ofMedicine then, finding2 U Broome and it is today considered a fact that the inci- records at the time of the biopsy. There were 34 R Hultcrantz dence of liver disease in patients with ulcerative men and 40 women. The mean age at onset of colitis is greater than in the general population.3 ulcerative colitis was 31 years (8-69). At this time Infectious Diseases, and Pathology, Huddinge The most common histological findings are fatty the Ethics committee system was not operating University Hospital and infiltration and portal inflammatory cell infil- in Sweden but as far as we know informed Roslagstull Hospital, trates consisting of collections of lymphocytes consent was not obtained. The diagnosis of Stockholm, Sweden H Glaumann and plasma cells. Often, but not always, these ulcerative colitis was based on the following are and a of diarrhoea and/or rectal Correspondence to: Ulrika cells concentrated around the bile ducts, criteria: history Broome, MD, Huddinge eventually hyaline fibrous tissue surrounding bleeding for six weeks or more; at least one University Hospital, S-141 86 the mucosa Huddinge, Sweden. bile ducts may develop.2-7 There have been sigmoidoscopy showing friability of Accepted for publication attempts to correlate the histological changes with contact bleeding and/or petechial haemor- 10 July 1989 seen in liver tissue to the severity and extent of rhages and/or ulceration with intervening inflam- Liverhistology andfollow up of68patients with ulcerative colitis and normal liverfunction tests 469

mation of the mucosa, and a barium enema with 100' radiological evidence of ulceration or narrowing or shortening of the colon; and characteristic microscopical changes in the biopsy.'5 80- The extent of the colitis was determined by a'c Gut: first published as 10.1136/gut.31.4.468 on 1 April 1990. Downloaded from sigmoidoscopy and barium enema. Thirty seven ._0 patients were found to have total involvement of 60- IRemission the colon, 28 left sided involvement and nine had proctitis. The inflammatory activity ofthe colitis E Active colitis 40- had been determined. Remission was defined as CL- a state in which the patient had no bowel 0 symptoms and the sigmoidoscopy showed no 20- activity. The colitis was defined as active when the sigmoidoscopy showed oedema, excess n I * I- -11 - a mucus, signs of bleeding and/or ulceration and . . M-F7- () I .f the patient had symptoms of diarrhoea. Fifteen O ' + ++ I +++ ++++ patients were in remission, and in 59 patients the Degree of portal inflammation colitis was active. Two patients had been sub- Figure 1: Comparison between the degree ofportal inflammation in patients with active colitis and with colitis in jected to cholecystectomy, and one had pre- remission. viously suffered from hepatitis A. Three patients were classified as over consumers of alcohol while the others consumed no alcohol or only all had been autopsied. The time of death was small amounts at weekends. Sulphasalazine was then taken as the time of follow up. Among the taken by 21 of the patients, corticosteroids by 17 six patients whom it was not possible to follow and local corticosteroids by 27 patients. up, we know that two women died at the age of. Guanicil, a non-absorbable sulphonamide, was 92 and 80 respectively. The fate ofthe remaining taken by nine patients and 10 patients were four patients are unknown to us today (Table). taking no medicine at the time of the liver The mean time offollow up was 18 years (0-28). biopsy. Thus, because of the various medical treatments and the small group ofpatients, it was not meaningful to correlate the histological find- Results ings to the drugs taken. The liver function in all patients was evaluated LIVER BIOPSY INTERPRETATION by bilirubin, transaminases, alkaline phos- Among the 74 patients studied, 37 (50%) had a phatase, and Bromsulphalein retention tests. All completely normal initial liver biopsy. Twenty http://gut.bmj.com/ these tests were normal in all 74 patients. Liver three biopsies showed fatty infiltration, nine biopsy was performed on average seven years portal inflammation, and in five there were both after the onset of the ulcerative colitis. The fatty infiltration and portal inflammation. Three percutaneous technique was used and each displayed fatty infiltration sufficient to make the biopsy stained with haematoxylin-eosin and van diagnosis of fatty liver (more than 50% of the Gieson. The biopsies were reevaluated by one hepatocytes involved). One of these three liver pathologist and one hepatologist according patients was in clinical remission, whereas two on October 2, 2021 by guest. Protected copyright. to the classification system of liver diseases'4 and suffered from active colitis. In the liver biopsy without knowing the patients identity. Special from one patient in remission more extensive emphasis was placed on alterations in the portal portal inflammation was seen. tracts such as inflammation and fibrosis. Hepato- The quantitative data regarding portal inflam- cellular degeneration and fatty infiltration were mation and lobular changes in relation to the also carefully assessed. Portal inflammation and inflammatory activity of the colitis are given in fatty infiltration were subjectively quantified on Figures 1 and 2, respectively, where it can be an arbitary scale from 0 to + + + +. seen that there was no correlation between the Through The National Population Registry it was possible to trace 68 patients. The time of 100, follow up was taken for 60 patients as the last recorded hospital visit when a clinical examination was carried out with measurements of * 80

transaminases, alkaline phosphatase and bili- 0. rubin. Biochemical data of the liver function [jI Remission7 .. 60 could not be found for two patients and the last Active colitis recorded clinical examination was then taken as the time offollow up. Four patients had died and E = 2040

TABLE Survey offollow up of74 patients with ulcerative H colitis Type ofinvestigation Patients (n) Physical examination and liver function tests 60 n .--- .11-. -,-- 7kIzz Physical examination 2 0 + ++ +++ ++++ Autopsy 4 Liver biopsy, physical examination and liver Degree of fatty infiltration function tests 2 Lost at follow up 6 Figure 2: Comparison between the degree offatty infiltration in patients with active colitis and with colitis in remission. 470 Broome, Glaumann, Hultcrantz

Figure 3: Degree ofportal Twenty three years after the biopsy the patient inflammation in relation to showed no clinical or biochemical signs of liver the extent ofinvolvement of Proctitis the colon. disease. 0- * Left side colitis

C Gut: first published as 10.1136/gut.31.4.468 on 1 April 1990. Downloaded from 41 g Total colitis ._ PATIENT 2 Q A woman born in 1927. Onset of ulcerative 0 a' colitis was in 1949 involving the total colon. .0D E Liver biopsy was performed in 1967, at which c time the colitis was in remission. The patient was treated for many years with sulphasalazine. 0 There was no other ailment. The biopsy showed portal tracts with typical fibrosis around the bile ducts but no inflammatory infiltrates around these. In some other portal tracts, however, there were signs of chronic inflammation with Degree of portal inflammation small lymphoid aggregates. In 1987, 20 years after the biopsy, the patient was free from any histological findings in the liver and the activity clinical or biochemical parameters of liver of the colitis defined as remission or active. disease. Furthermore, only one patient had more than slight portal inflammation, the rest had very minor or no inflammation. Fatty infiltration was PATIENT 3 sparse and evidently not correlated to the portal A man born in 1949 with total ulcerative colitis inflammatory activity. starting in 1957. A liver biopsy was performed in The extent of the involvement of the colitis 1967 when the colitis was active. He was treated was correlated to the histological findings, with sulphasalazine and corticosteroids at the namely portal inflammation (Fig 3) and fatty time ofthe biopsy. He had no history ofexcessive infiltration (Fig 4). There was, however, no clear alcohol ingestion. The liver biopsy showed con- cut correlation between these two histological centric periductular fibrosis without any signs of abnormalities and the extent of the colitis. inflammation. As in patient 2, however, there Fibrosis was only seen in three patients. In the was inflammation graded to + + in some portal biopsies of these patients the interlobular bile tracts and fatty infiltration graded to +. Thirteen years later the patient showed no clinical or

ducts were surrounded by concentric fibrosis, or http://gut.bmj.com/ 'onion-skin' lesions, previously described in biochemical evidence ofliver disease. patients with sclerosing cholangitis. These patients who all displayed total colonic involvement are described below. FOLLOW UP Of the 68 patients traced two had developed evidence of liver abnormalities at the time of follow up. The first was a woman aged 52 years

PATIENT I on October 2, 2021 by guest. Protected copyright. A man born in 1926. The onset of ulcerative with onset of ulcerative colitis at an age of 27. colitis was in 1949 and involved the total colon. A The entire colon was involved. Salazopyrin was liver biopsy was performed in 1965, while the taken for many years. The first liver biopsy was patient was in remission showing two portal performed in 1965 during a fulminant attack of tracts with concentric periductular fibrosis but the colitis, two years after the onset of the no signs of inflammatory activity around the bile ulcerative colitis. The biopsy showed no histo- ducts or elsewhere in the portal tracts. Fatty logical abnormalities. In 1967 a colectomy was infiltration was graded to +. The patient was performed because of chronic persistent colitis. otherwise healthy and did not abuse alcohol. No The liver was normal at the time of surgery. In treatment for ulcerative colitis or any other the years before colectomy she received several disease was given at the time of the biopsy. units of blood during severe attacks of ulcerative colitis with concomitant anaemia and in 1970 the patient was hospitalised because of jaundice. 100- The aetiology ofthe jaundice has not been able to be established. In 1973 the transaminases were 8 80 F Proctitis7 slightly raised and there was a third liver biopsy which on reexamination showed sparse chronic c * Left side colitisI a, inflammation and some piecemeal necrosis in the a. 60- n Total colitis portal tracts. A radionucleotide scan showed a 0 normal sized liver but an enlarged spleen. Smooth muscle antibodies, mitochondrial antibodies and antinuclear factor were negative and E the patient has no symptoms or signs of liver 0 disease. The transaminases remained slightly 20 raised but less than twice the upper normal level, the alkaline phosphatase and bilirubin values Figure 4: Degree offatty were normal. The prothrombin time was some- infiltration in relation to the n4-- *-\~~~~~~~~~~~~~~~~ -s-_E- extent ofthe involvement of 0 + + +++ what prolonged. A new biopsy in 1985, 12 years the colon. Degree of fatty infiltration later, showed cirrhosis with moderate chronic Liverhistologyandfollow up of68patients with ulcerative colitis and normal liverfunction tests 471

inflammation in the portal tracts and piecemeal In our study, no such correlation was found necrosis, but no lobular changes and no bile duct between the activity and the extent of the proliferation. An endoscopic retrograde cholan- involvement of the large intestine on one hand giography has not yet been performed. The and the histological changes in the liver on the second patient to develop liver disease is a other. In 4% of our patients, periductal concen- Gut: first published as 10.1136/gut.31.4.468 on 1 April 1990. Downloaded from woman, born in 1931. Ulcerative colitis started tric fibrosis or 'onion skin lesions' were found. in 1949 with total colonic involvement and a liver These lesions were exclusively observed in biopsy was performed in 1959 when the colitis patients with total colonic involvement, but was active, showing minor fatty infiltration but otherwise without relation to the inflammatory no portal inflammation. She had been treated activity of the colon. with salazopyrin for many years. There was no In a study of patients with ulcerative colitis, history of alcohol abuse. In 1971 the patient who had been subjected to liver biopsy because became jaundiced and tired, and lost weight. of biochemical evidence of liver disease, the The transaminases were raised maximally, nine biopsies disclosed a high frequency of histo- times the upper normal concentration whereas logical alterations, often coinciding with findings the alkaline phosphatases activity never reached indicating primary sclerosing cholangitis."6 In more than slightly raised levels. The antinuclear this study many patients had biopsy findings factor was positive in dilution 1/1600 and anti- indicating primary sclerosing cholangitis - for bodies against smooth muscle positive 1/25. example, periductal fibrosis and inflammation in Hepatitis B surface antigen was negative. A liver the portal zones. This group of patients appar- biopsy showed bile duct proliferation, periportal ently constituted a different population from inflammation and a portal cirrhosis was sus- ours, they had clinical or biochemical signs of pected. Chronic active hepatitis was then con- liver disease, and consequently, also had a much sidered possible and corticosteroids were given. poorer prognosis than our patients. One third of Clinical and biochemical improvement occurred the patients had died 10 to 15 years after the with a decrease in the transaminases to twice the biopsy compared with our study in which only upper normal level and normalisation of the two patients (3%) developed clinically significant alkaline phosphatase. The antinuclear factor liver disease. We have no evidence that any of became negative. The corticosteroid treatment our patients died of liver disease although the was gradually withdrawn. A new episode with exact fate of four of the patients is unknown. jaundice and fatigue occurred in 1976. The Primary sclerosing cholangitis is thought to be transaminases were again high. Antibodies the major hepatobiliary complication of ulcera- against smooth muscle were positive in 1/100. tive colitis." A liver biopsy is usually not The patient was again successfully treated with diagnostic in primary sclerosing cholangitis.'7'8 corticosteroids, and the biochemical and clinical Concentric periductal fibrosis is considered as http://gut.bmj.com/ signs of liver disease slowly disappeared. In more or less pathognomonic for primary scleros- 1977, a repeat liver biopsy was performed, ing cholangitis, if other biliary disorders can be showing some fatty infiltration and some slightly excluded but is only found in about one third of enlarged portal tracts, but no signs ofcirrhosis or the liver biopsies from patients with primary chronic hepatitis. From 1978 to 1985 transa- sclerosing cholangitis,'8 probably because of minases and alkaline phosphatase were essen- sporadic distribution. The combination of tially within the normal limits. Prothrombin primary sclerosing cholangitis and ulcerative on October 2, 2021 by guest. Protected copyright. time and albumin were normal. A cholecystec- colitis is found in 50-70% of all cases with tomy was performed in 1982 and in 1987 an primary sclerosing cholangitis, in particular in ileorectal anastomosis was carried out for a patients with involvement of the entire colon. 8 9 benign stenosis between the descending and In our study, 7% ofthe patients with total colonic sigmoid colon. involvement showed some periductal fibrosis but In early 1988 the transaminases again increased none developed any clinical or biochemical to maximally 10 and the alkaline phosphatase to evidence of liver disease. The aetiology of peri- five times the upper normal limit. A computed ductal fibrosis is not known but repeated episodes tomography ofthe liver was performed revealing of cholangitis, may be one cause. None of the an enlarged liver with the intrahepatic bile ducts patients had such episodes, however. Based on being irregular and dilated. A new liver biopsy the present findings we therefore conclude that showed moderate fatty infiltration, fibrotic isolated periductal fibrosis without inflammation bridging, general bile duct proliferation, and seems to be of little clinical significance in malignant transformation of some bile ducts. No patients with ulcerative colitis. endoscopic retrograde cholangiography has been Other reports have shown that cholestasis and performed. bile ductal changes are not related. Several patients have for example, been found to have radiological and histological signs indicating Discussion primary sclerosing cholangitis even in the This study of a comparatively large group of absence of chemical parameters of bile duct patients suffering from ulcerative colitis initially disease.202' It is also known that asymptomatic without any clinical or biochemical evidence of patients with primary sclerosing cholangitis may liver disease shows that minor histological develop symptomatic liver disease several years changes are common in liver biopsies from these after the diagnosis has been established.22 To our patients. According to De Dombal et al4 and knowledge, however, it is not established Eade,8 it should be possible to relate histological whether patients, as those we describe here, alterations such as fatty infiltration and portal without both clinical and biochemical evidence inflammation to the clinical course of the colitis. of liver disease with periductal fibrosis, have a 472 Broome, Glaumann, Hultcrantz

higher risk to eventually develop a significant plications. The present study also suggests that liver disease. The three patients we describe have fibrotic changes, especially periductal fibrosis, now been observed for more than 20 years and previously described in patients with primary have still not developed any biochemical para- sclerosing cholangitis" 18 do not serve as an meters or symptoms related to liver disease. unfavourable prognostic marker. Gut: first published as 10.1136/gut.31.4.468 on 1 April 1990. Downloaded from Accordingly, it seems as if fibrotic changes in 1 Thomas GH. Ulceration of the colon with enlarged fatty liver. liver biopsies from these patients do not indicate Trans Path Soc Phila 1874; 4: 87-8. 2 Kimmelstein P, Large L, Verner H. Liver damage in ulcera- a bad prognosis. The exact frequency ofprimary tive colitis. AmJ Pathol 1952; 28: 259-74. sclerosing cholangitis as seen with cholangi- 3 Kleckner M, Stauffner M, Bargen A, Dockerty M. Hepatic lesions in the living patient with chronic colitis as demon- ography among patients with ulcerative colitis is strated by needle biopsy. Gastroenterology 1952; 22: 13-33. not known. Despite a high frequency of minor 4 De Dombal FT, Goldie W, McK Watts J, Goligher JC. Hepatic histological changes in ulcerative colitis. Scand histological changes in the liver, only two of our J Gastroenterol 1966; 1: 220-7. 68 patients developed liver disease symptoms. 5 Dordal E, Glagov S, Kirsner J. Hepatic lesions in chronic inflammatory bowel disease. Gastroenterology 1%6; 52: Six patients were lost at follow up. Four of these 239-53. may have developed liver disease although it 6 Mistilis SP, Skyring AP, Goulston SJ. Pericholangitis and ulcerative colitis. Ann Intern Med 1965; 63: 1-26. seems unlikely. The two patients, who even- 7 Vinnik I, Kern F. Liver diseases in ulcerative colitis. Arch tually manifested cirrhosis, had no histological Intern Med 1963; 112: 41-9. 8 Eade MN. Liver disease in ulcerative colitis. Ann Intern Med or biochemical signs ofliver disease at the time of 1970; 72: 475-87. the first liver biopsy. In the first patient the cause 9 Warren KW, Athanassides S, Monge JI. Primary sclerosing cholangitis. Amj Surg 1966; 111: 23-38. of the cirrhosis is unknown. Even though an 10 Thorpe MEC, Scheuer PJ, Sherlock S. Primary sclerosing endoscopic retrograde cholangiography was not cholangitis, the biliary tree and ulcerative colitis. Gut 1967; 8:435-48. performed, an association with primary scleros- 11 Sheperd HA, Selby WS, Chapman RWG, et al. Ulcerative ing cholangitis is unlikely, as there were no colitis and persistent liver dysfunction. Q J Med 1983; 52: 503-13. histological changes typical for primary scleros- 12 Schrumpf E, Elgio K, Fausa 0, Gjone E, Kolmannskog F, ing cholangitis and serum alkaline phosphatase Ritland S. Sclerosing cholangitis in ulcerative colitis. Scand J Gastroenterol 1980; 15: 689-97. activities were normal. Because of frequent 13 Perrett AD, Higgins G, Johnston H, Massarella GR, True- blood transfusions non-A, non-B hepatitis seems love SC, Wright R. The liver in ulcerative colitis. QJ Med 1971; 40:211-38. to be a plausible aetiology for the cirrhotic out- 14 Scheuer PJ. Liver biopsy interpretation. London: Bailliere & come.23 The second patient had evidence of an Tindall, 1988. 15 Nordenvall B, Brostrom 0, Berglund M, et al. Incidence of autoimmune liver disease which previously has ulcerative colitis in Stockholm county. ScandJ7 Gastroenterol been described in combination with ulcerative 1985; 20: 783-90. 16 Wee A, Ludwig J. Pericholangitis in chronic ulcerative colitis: colitis.24 The patient finally developed malignant Primary sclerosing cholangitis of the small bile ducts? Ann transformation of the bile ducts, a carcinoma Intern Med 1985; 102: 581-7. known to be associated with ulcerative colitis.25 17 La Russo NF, Wiesner RH, Ludwig J, Mac Carty RL.

Primary sclerosing cholangitis. N Engl J Med 1984; 310: http://gut.bmj.com/ Thus, clear evidence of liver disease associated 899-903. 18 Chapman RWG, Arborgh BAM, Rhodes JM, et al. Primary with ulcerative colitis was only found in one or sclerosing cholangitis: a review of its clinical features, possibly two (3%) of our patients. cholangiography, and hepatic histology. Gut 1980; 21: 870-7. The range of hepatobiliary complications in 19 Wiesner RH, LaRusso NF. Clinicopathologic features of the patients with ulcerative colitis is reported to be 6 syndrome ofprimary sclerosing cholangitis. Gastroenterology 1980; 79: 200-6. to 25%. 12 2628 In the present study such complica- 20 Balasubramaniam K, Wiesner RH, LaRusso NF. Primary tions seen after almost 20 years offollow up were sclerosing cholangitis with normal serum alkaline phosphatase activity. Gastroenterology 1988; 95: 1395-8. on October 2, 2021 by guest. Protected copyright. less frequent than might be expected. Accord- 21 Clements D, Rhodes JM, Elias E. Severe bile duct lesions ingly, if there is no clinical or biochemical without biochemical evidence of cholestasis in a case of sclerosing cholangitis. J Hepatol 1986; 3: 72-4. evidence of hepatobiliary disease during the first 22 LaRusso NF, Wiesner RH, Ludwig J. Is primary sclerosing years after the onset of ulcerative colitis the cholangitis a bad disease? Gastroenterology 1987; 92: 2031-33. patients apparently run a very small risk of 23 Mattsson L, Weiland 0, Glaumann H. Long-term follow-up developing a hepatobiliary complication. pf chronic posttransfusion non-A, non-B hepatitis. Liver 1988; 8: 184-8. In conclusion, this study shows that minor 24 Olsson R, Hulthen L. Concurrence of ulcerative colitis and histological abnormalities in the liver are com- chronic active hepatitis. Scand J Gastroenterol 1975; 10: 331-4. mon in patients with ulcerative colitis without 25 Mid-Madilessi SH, Farmer RG, Sivak MV. Bile duct carci- any clinical or biochemical signs of liver disease. noma in patients with ulcerative colitis. Dig Dis Sci 1987; 32: 145-54. The histological changes are of little help in 26 Lupinetti M, Mehigan H, Cameron J. Hepatobiliary compli- predicting the risk to develop hepatobiliary cations ofulcerative colitis. Am J Surg 1979; 139: 113-8. 27 Dew MJ, Thompson H, Allan RN. The spectrum of hepatic complications. The absence of indications of dysfunction in inflammatory bowel disease. Q J Med 1979; liver disease during the first years after the onset 48: 113-35. 28 Schrumpf E, Fausa 0, Kolmannskog F, Elgio K, Ritland S, of ulcerative colitis makes it unlikely that the Gjone E. Sclerosing cholangitis in ulcerative colitis, a follow- patient will later develop any hepatobiliary com- up study. Scandj Gastroenterol 1982; 17: 33-9.