- Email: [email protected]
Hepatobiliary carcinoma in primary sclerosing cholangitis: a case control study
Journal of Hepatology 34 (2001) 792±798
www.elsevier.com/locate/jhep
Hepatobiliary carcinoma in primary sclerosing cholangitis: a case control study Marjut Leidenius 1, Krister HoÈckerstedt 1,*, Ulrika Broome 2, Bo-GoÈran Ericzon 2, StyrbjoÈrn Friman 3, Michael Olausson 3, Erik Schrumpf 4 1
Transplantation and Liver Surgery, Helsinki University Hospital, Kasarmikatu 11±13, 00130 Helsinki, Finland 2 Unit of Gastroenterology and Hepatology, Huddinge University Hospital, Huddinge, Sweden 3 Department of Transplantation and Liver Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden 4 Department of Medicine, Rikshospitalet, Oslo, Norway
Backround/Aims: Hepatobiliary carcinoma (HBC) has been considered to be a late complication of end-stage primary sclerosing cholangitis (PSC). The incidence of HBC is approximately 20% in PSC patients evaluated for liver transplantation. The diagnosis of HBC is dif®cult, at least at an early stage and the prognosis is poor even after liver transplantation. The aim of the study was to look for signs and risk factors for developing hepatobiliary carcinoma in patients with PSC. Methods: Thirty-six consecutive patients with PSC and HBC (32 with bile duct carcinoma, BDC, and four with hepatocellular carcinoma, HCC) were pair-matched to control patients referred for liver transplantation because of PSC but who did not have HBC. Gender and age at referral were used as matching factors. Clinical and biochemical data were registered. Results: PSC patients with BDC had a shorter median duration of PSC (1 year) compared with the controls (7 years) and PSC patients with HCC (8 years). There were no statistically signi®cant differences in the liver biochemistry between the patient groups. Varices were more common in patients with PSC and HCC (100%) than in controls (56%) or patients with PSC and HBC (12%) (P , 0:0005). Conclusions: The relatively short duration of PSC and the absence of varices in patients with BDC suggest that BDC, unlike HCC, is not necessarily a late complication of end-stage PSC, as was previously assumed. q 2001 European Association for the Study of the Liver. Published by Elsevier Science B.V. All rights reserved. Keywords: Primary sclerosing cholangitis; Hepatobiliary carcinoma; Ulcerative colitis
1. Introduction Patients with primary sclerosing cholangitis (PSC) have an increased risk of developing gastrointestinal cancer, in particular hepatobiliary malignancies [1±5]. The risk of bile duct cancer (BDC) ranges from 6 to 14% among all PSC patients [2±5]. The prevalence of BDC is approximately 20% in PSC patients evaluated for liver transplantation [6±9]. The prevalence of hepatocellular carcinoma (HCC) in PSC patients is less well known but it is estimated to be 2% in PSC undergoing liver transplantation [9]. PSC is also an additional risk factor for neoplastic alteration of the Received 7 June 2000; received in revised form 12 December 2000; accepted 5 January 2001 * Corresponding author. Tel.: 1358-9-4718-8271; fax: 1358-9-174975.
colorectal mucosa in ulcerative colitis (UC) [10±12]. Broome et al. also noted that the incidence of cholangiocarcinoma in PSC was higher in patients with concomitant UC and colorectal neoplasia. Hepatobiliary carcinoma (HBC) is usually considered to be a late complication of end-stage PSC [2,13,14]. The diagnosis of BDC is dif®cult, at least at an early stage and the prognosis is poor even after liver transplantation [15]. Therefore BDC is nowadays considered as a contra-indication for liver transplantation in many transplantation units. Clinical, biochemical and radiological signs for predicting cholangiocarcinoma in PSC patients have been sought for. Rapid clinical deterioration with jaundice, weight loss associated with an appreciable rise in serum bilirubin [3,16] and abdominal discomfort [3,17] are considered to be signs
0168-8278/01/$20.00 q 2001 European Association for the Study of the Liver. Published by Elsevier Science B.V. All rights reserved. PII: S01 68-8278(01)0004 2-3
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
of (advanced) cholangiocarcinoma in PSC. Cholangiographic features suggesting malignant degeneration include progressive stricture formation, ductal dilatation or polypoid masses [14]. The monitoring of tumor markers CEA and CA 19-9 in serum or bile may be useful for indicating hepatobiliary malignancy in patients with PSC [17±20]. The diagnostic speci®city of bile cytology is very high (up to 100%), but disadvantages of the method are invasiveness and the relatively low sensitivity (approximately 50%), when distinguishing between benign and malignant biliary obstruction [21]. Bile duct dysplasia has been found in over 60% of patients with PSC and cholangiocarcinoma [17]; dysplasia may precede cholangiocarcinoma in patients with PSC [22]. The diagnosis of HBC is dif®cult but important when considering the indications and timing of liver transplantation in PSC. Therefore the aim of this study was to examine clinical or biochemical differences between patients with PSC and HBC and control patients with PSC alone, to look for signs and risk factors for developing hepatobiliary carcinoma in PSC. 2. Materials and methods 2.1. Patients Data on 36 consecutive patients who were referred for liver transplantation between 1987 and 1996 because of PSC and also had HBC, were collected retrospectively from patient records at four Nordic Liver Transplantation Centers: Helsinki in Finland, Stockholm and Gothenburg in Sweden and Oslo in Norway. The study was approved by the Ethical Committees of all transplantation centers. The study was performed in accordance with the Declaration of Helsinki. The diagnosis of PSC was based on established biochemical, radiological and histopathological criteria [23,24]. The onset of PSC was de®ned as the point when cholangiography with ®ndings consistent of PSC was made. The duration of PSC was calculated in two different ways: from the time of cholangiography and also from the time of ®rst presentation of abnormal serum liver function tests. The histologically or cytologically proven diagnosis of HBC was made during laparotomy (usually liver transplantation or an attempt of liver transplantation) in 26 patients, by liver biopsy in eight patients and at autopsy in two patients. Thirty-one patients had bile duct carcinoma (in the intra- or extrahepatic bile ducts or both), one patient had carcinoma of the gallbladder and four patients had hepatocellular carcinoma. Each patient with PSC and HBC included in this study was pair-matched at each center with a control patient also referred for liver transplantation because of PSC but who did not have HBC. Hepatobiliary malignancy was excluded in the control group by careful histological examination of explanted livers after transplantation and by long term follow-up. The matching factors included gender and age at referral as close as possible.
2.2. Clinical examination Presence or absence of various clinical symptoms and signs were recorded at the time of evaluation for liver transplantation. These included jaundice, itching, presence of varices, variceal bleeding, febrile cholangitis, epigastric pain and weight loss during 12 months. Patients in both groups had undergone colonoscopy with multiple biopsies to diagnose or to exclude concomitant in¯ammatory bowel disease (IBD), except for those who had undergone proctocolectomy. The diagnosis of IBD was based on
793
clinical features and the radiological, endoscopic and histological appearance of the large bowel and rectum [25,26]. All patients with longstanding colitis were included in colonoscopy surveillance in order to detect colorectal carcinoma or dysplasia.
2.3. Biochemical methods The biochemical tests for liver function used as part of the medical evaluation prior to transplantation included measurements of serum concentrations of alanine transferases (S-ALT), alkaline phosphatases (SALP), total bilirubin and albumin, as well as prothrombin complex assay (P-TT-SPA). The tumor marker CEA was determined in 16 patients and 16 controls and CA 19-9 in 13 patients and 11 controls.
2.4. Statistical methods Fisher's exact test was used to compare percentages. Medians were compared using the Mann-Whitney U-test.
3. Results The median age at referral was 45 years in both patients (range 18±67 years) and controls (range 27±64 years). PSC patients with HBC were somewhat older at onset of PSC (not signi®cant (NS)) and had a 5 years shorter median duration of PSC (P , 0:01) than control patients without HBC when the onset of PSC was de®ned as the time point when cholangiography with ®ndings consistent of PSC was made. The median duration of PSC was 6 years (range 0±24 years) in patients with HBC and 8 years (range 0±23 years) in controls (NS) calculated from the time of ®rst presentation of abnormal serum liver function tests. The proportion of patients with very short duration of PSC (0±1 years) was 31% (11/36) in the HBC group and 6% (2/36) in the control group (P , 0:02) using the time of ®rst presentation of abnormal serum liver function tests as the onset of PSC. (Table 1). Altogether 31 (86%) PSC patients with HBC also had IBD: 27 had UC, three had unde®ned colitis and one patient had Crohn's disease. Twenty-nine (81%) control patients Table 1 Characteristics of PSC patients with HBC and control patients without HBC at referral for liver transplantation a
Age (years) Men/women Age at onset of PSC (years) Duration of PSC (years) b Duration of PSC (years) d Patients with 0±1 yearsc duration of PSC a
HBC (n 36)
Control (n 36)
45 (18±70) 26/10 41 (14±66) 2 (0±18) c 6 (0±24) 11/36 e
45 (27±70) 26/10 35 (13±63) 7 (0±20) 8 (0±23) 2/36
Values are median (range). The onset of PSC calculated from the time of cholangiography c The onset of PSC calculated from the time of ®rst presentation of abnormal serum liver function tests. d P , 0:01. e P , 0:02. b
794
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
Table 2 In¯ammatory bowel disease (IBD) in PSC patients with HBC and control patients without HBC a
IBD (%) Age at onset of IBD (years) Duration of IBD to the time of colectomy or to the referral (years) Duration of IBD to the referral (years) Duration of IBD to onset of PSC (years) Colorectal dysplasia or carcinoma (%) Colectomy rate (%)
HBC (n 36)
Control (n 36)
86 20 (8±50) 18 (2±35)
81 22 (13±69) 14 (0±33)
22 (4±44)
20 (0±40)
16 (9±43)
11 (19±33)
39
28
45
41
a Values are median (range). The differences between the groups were not statistically signi®cant.
without HBC had IBD: 28 had UC and one had unde®ned colitis. The median age at onset of IBD was 20 years in the HBC patients and 22 years in the controls. Median duration of IBD to colectomy or to the date of referral was 18 years and 22 years to the date of referral in patients with HBC, and 14 years and 20 years, respectively, in the control patients. The median interval between the onset of IBD and the onset of PSC was 16 years and 11 years in the controls (NS). Six patients with HBC had undergone colectomy because of
active colitis, six because of colorectal dysplasia or carcinoma and two patients had undergone prophylactic proctocolectomy because of long-standing UC. Five control patients were operated on because of active UC and seven because of colorectal neoplasia. Twelve patients (39%) with HBC and eight controls (28%) exhibited colorectal dysplasia or carcinoma (NS) (Table 2). There were no statistically signi®cant differences in the liver biochemistry at referral between the patients with HBC and the controls without HBC. Serum CEA was measured in 16 HBC patients and in 16 controls. Serum CA 19-9 was determined in 13 HBC patients and in 11 control patients. Increased levels of both CEA (median 9.45 mg/l) and CA 19-9 (median 120 units/l) were observed in the HBC patients compared with the control group (median CEA 2 mg/l and median CA 19-9 37 units/l) (P , 0:05, both for CEA and CA 19-9) (Table 3). Both tumor markers were measured in eight patients and eight controls and the combined tumor marker index was calculated [19]. The index exceeded 400 units in ®ve patients with BDC, whereas the index was under 400 in all patients with HCC and in all controls (P , 0:05 between BDC patients and controls). Jaundice, itching or febrile cholangitis were as common in patients with PSC and HBC as in the control patients with PSC alone. Esophageal (or peristomal) varices were more common in control patients with PSC alone (56%) than in patients with PSC and HBC (12%) (P , 0:0005). Control
Table 3 Characteristics of PSC patients with BDC or HCC and control patients without HBC at referral for liver transplantation a
Age (years) Men/women Age at onset of PSC (years) b Duration of PSC (years) b S-ALT (,40 units/l) S-ALP (,280 units/l) S-Bilirubin (,20 mmol/l) P-TT-SPA (.70%) S-Albumin (.36 g/l) CEA (,5 mg/l) CA 19±9 (,37 units/l) Varices (%) Bleeding from varices (%) Jaundice (%) Itching (%) Febrile cholangitis (%) Epigastric pain (%) Weight loss in 12 months (%) a b c d e f g h
BDC (n 32)
HCC (n 4)
Control (n 36)
45 (18±70) 22/10 41 (14±66) 1 (0±17) c 96 (26±282) 1380 (378±3446) 180 (10±658) 91 (26±150) 31 (16±40) 10.4 (1±120) d 120 (35±8100) e 0f 0g 85 68 31 67 h 75
49 (18±67) 4/0 33 (18±59) 8 (0±18) 101 (28±240) 620 (228±2460) 63 (31±85) 67 (41±130) 28 (24±43) 4.4 (1.5±7.2) 6 (2±10) 100 100 50 0 0 33 0
45 (27±70) 26/10 35 (13±63) 7 (0±20) 117 (29±406) 1155 (379±3480) 106 (8±500) 68 (25±130) 27 (16±44) 2 (1±7) 37 (12±82) 56 31 83 55 28 7 39
Values are median (range). The onset of PSC calculated from the time of cholangiography. P , 0:005 compared with the controls. P , 0:05 compared with the controls. P , 0:0005 compared with the controls, P , 0:05 compared with the HCC patients. P , 0:00005 compared with the controls and to the HCC patients. P , 0:001 compared with the controls, P , 0:00005 compared with the HCC patients. P , 0:0005 compared with the controls.
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
patients also seemed to have bleeding from varices more often (31 vs. 12% in HBC patients) but the difference was not statistically signi®cant. Epigastric pain was more common in HBC patients (63%) than in control PSC patients (7%) (P , 0:00005). Seventy-one percent of HBC patients had lost weight during last the 12 months, compared with 39% of the controls (NS) (Table 3). Four patients with PSC and HBC had HCC, and 32 had BDC. The median age at referral was 45 years (range 18±70 years) in BDC patients and 49 years (range 18±67 years) in HCC patients (NS). The age at onset of PSC was 41 years (range 14±66 years) in the BDC patients and 33 years (range 18±59 years) in the HCC patients (NS). The median duration of PSC was 1 year (range 0±17 years) in patients with BDC and 8 years (range 0±18 years) in patients with HCC (NS) when the onset of PSC was de®ned as the point when cholangiography with ®ndings consistent of PSC was made. The median duration of PSC calculated from the time of ®rst presentation of abnormal serum liver function tests was 5 years (range 0±24 years) in the BDC group and 15 years (range 0±18 years) in the HCC group (NS). Both the median S-ALP (1380 vs. 620 units/l) and S-bil (180 vs. 63 mmol/l) were somewhat higher in patients with BDC compared with HCC patients, but the differences were not statistically signi®cant. Varices and variceal bleeding were encountered in all of the HCC patients but in none of the BDC patients (P , 0:00005). The median serum CEA was 10.4 mg/l (range 1±120 mg/l) in the BDC patients and 4.4 mg/l (1.5± 7.2 mg/l) in the HCC patients (NS). The median serum CA 19-9 was higher in the BDC patients (120 units/l, range 35± 8100 units/l) than in the HCC patients (6 units/l, range 2±10 units/l) (P , 0:05) (Table 3). 4. Discussion The major ®nding in this case-control study is that HBC, especially BDC, may develop also in patients with an early phase PSC and thus is not necessarily a late complication of end-stage PSC as has been previously assumed [2,13,14]. Patients with HBC had a ®ve year shorter median duration of PSC than the control patients without HBC. The HBC patients had a tendency towards onset of PSC at a later age. They also seemed to have a longer duration of IBD before the onset of PSC. In a recent study by Bergqvist et al., there were no differences in the duration of PSC between the PSC patients with and without HBC [17]. They de®ned the onset of PSC as the time of ®rst presentation of abnormal serum liver function tests consistent with the diagnosis of PSC. It should be mentioned that some of the patients in the study by Bergqvist et al. (12 HBC patients and four control patients) are actually also included in the present study. In the present study the onset of PSC was de®ned as the time point when cholangiography with ®ndings consistent of PSC was made. In addition, the duration of PSC was
795
calculated in two different ways: from the time of cholangiography and from the time of ®rst presentation of abnormal serum liver function tests. Abnormal liver function tests do not necessarily stand for PSC. Majority of PSC patients have IBD (over 80% in the present study), usually extensive UC. Transient elevations in liver function tests are encountered in patients with UC in up to 50% [27]. Usually these transient abnormalities in liver function tests are associated with active colitis or surgery [28,29]. At least mild, persistent liver involvement has been found in 21% of UC patients at the time of colectomy although the criteria of PSC was ful®lled only in one third of them [30]. These milder forms of hepatobiliary dysfunction did not show any progression in the follow-up [30]. The golden standard in the diagnosis of PSC is endoscopic retrograde cholangiography (ERC). ERC is an invasive procedure with signi®cant complication risks. Therefore it has not always been carried out at the time of ®rst presentation of abnormal liver function tests. This may cause delay in the diagnosis of PSC in some patients. In addition, PSC has even been encountered without laboratory abnormalities [31]. Therefore it is impossible to determine the exact onset of PSC by serum liver function tests alone. Presence of esophageal or peristomal varices or variceal bleedings were rare in the HBC patients compared with the control patients. This ®nding supports our presumption that HBC may develop before disease progression to end-stage PSC. Correspondingly, in an earlier study by Miros et al. PSC patients with HBC seemed to have a less advanced histological stage of PSC than those without HBC [16]. Wee et al. have also detected HBC in patients with small duct PSC [2]. Unfortunately, suf®cient data regarding the histological stage of PSC patients with HBC was not available in the present study. With regard to patients with PSC and HCC the ®ndings of the present study are in close agreement with a study by Harnois et al., suggesting HCC to be a complication of end-stage PSC [9]. All HCC patients had bled from esophageal varices as a sign of chronic, end-stage liver disease while none of BDC patients had peristomal or esophageal varices. Patients with HCC had a longer median duration of PSC compared with those with BDC. Also the ®ndings at liver biochemistry were slightly different in BDC patients, showing a greater degree of cholestasis in them. Correspondingly, itching was encountered in 61% of BDC patients but in none of the HCC patients. However, the differences between HCC and BDC patients in duration of PSC or in parameters representing cholestasis were not statistically signi®cant, probably due to the small number of HCC patients. To our knowledge, these ®ndings on HCC in PSC patients have not been reported earlier. Sixteen patients had had PSC just for 1 year or less before the detection of HBC, when the onset of PSC was de®ned as the time point when cholangiography with ®ndings consistent of PSC was made. This may give rise to doubt about the diagnosis of PSC in these patients. One may argue that in
796
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
some patients the cholangiographic ®ndings which have led to the diagnosis of PSC, may actually have been due to the HBC itself. The diagnosis of PSC seems to be justi®ed also in these HBC patients with a very short duration of preexisting PSC. Twelve of these 16 patients also had IBD. HBC is assumed to be very rare among patients with IBD without any co-existing large or small duct PSC [2]. Three of the four patients without IBD had undergone liver transplantation with thorough histological examination of the removed liver, with ®ndings typical of PSC and complicating HBC [2], and not to a sclerosing cholangiocarcinoma without pre-existing PSC [32] The number of the HBCpatients with very short duration of PSC was 11, when the duration of PSC was calculated from the time of ®rst presentation of abnormal serum liver function tests. We found that epigastric pain was more common in HBC patients. There was also a tendency towards a greater weight loss in this group. Both symptoms have also been related to HBC in previous studies [3,16]. However, epigastric pain as well as weight loss are signs of advanced malignancy and do not help in the early diagnosis of HBC. In addition, the presence of ascites and the use of diuretics render it dif®cult to evaluate the actual wasting of a patient merely by measuring the weight. Jaundice, itching or abnormalities in liver function tests do not distinguish between the HBC and the control group, because these could be either due to HBC or to the PSC, itself. The ®ndings of earlier studies have been controversial in regard to sudden clinical deterioration or rise in serum bilirubin as signs of HBC in PSC patients [3,16,17]. The rapid clinical or biochemical deterioration can be due to HBC as well as liver decompensation or a dominant benign stricture [17,33]. MacCarty et al. have found that markedly dilated ducts or ductal segments, large (1 cm or greater) polypoid masses and progressive stricture formation or dilatation suggest cholangiocarcinoma in PSC [14]. Van Laethem et al. considered dominant intrahepatic duct strictures as malignant and extrahepatic strictures (with benign brush cytology) to be due to the PSC itself [33]. On the contrary, Bergqvist et al. found it impossible to differentiate malignant strictures from benign by cholangiographic appearances [17]. PSC patients with dominant strictures may bene®t from endoscopic stenting [33]. However, liver transplantation must be considered in all patients with dominant strictures even in those with less advanced PSC, because it is impossible to rule out HBC. Elevated serum levels of tumor markers CEA and Ca19-9 have been detected in PSC patients with BDC compared with patients with stable or progressive PSC without BDC [17±19]. Our own results are in close agreement with these previous studies. Tumor markers CEA and CA 19-9 are widely used in follow-up of PSC patients in order to detect BDC. However, at least moderately elevated serum tumor marker levels have also been found in PSC patients without malignancies [18,19], probably due to obstructive jaundice [34]. Normal serum tumor marker levels do not exclude
BDC [18,19] and neither is monitoring CEA and CA 19-9 helpful in the diagnosis of HCC. Anyway, by combined use of both markers it may be possible to detect most occult bile duct malignancies and exclude inappropriate cases for liver transplantation more effectively [19]. Unfortunately, malignancies detected by the measurement of CEA or CA 19-9 serum levels or by combined use of these markers, are usually at an incurable stage and little can be done to help these patients [19]. The optimal timing of liver transplantation has been addressed by Wiesner et al. who developed the Mayo PSC natural history model [35]. An important limitation to applying this risk score model concerns the risk of HBC among PSC patients. The survival of patients with PSC and HBC has not been signi®cantly improved by liver transplantation, because it has been impossible to include such components as `undiagnosed HBC' in the Mayo model [15]. In Hannover, Nashan et al. compared Mayo risk scores with the incidence of HBC in PSC patients [36]. They found that all HBC patients had a risk score of 4.3 or higher. Correspondingly, they assumed that the risk of HBC is very low in PSC patients with a risk score of 4.0 or lower. Unfortunately they had only very few patients with a risk score under 4.3 in their study to support their assumption. They also noticed that most HBC patients were included in the medium and not in the high-risk score group. The most probable explanation to this con¯icting ®nding is that the components included in the Mayo model (age, serum bilirubin concentration, splenomegaly, histological stage) can hardly to be expected to distinguish between PSC patients with and without HBC. Broome et al. have noted that the risk of bile duct carcinoma in PSC was concentrated on patients with concomitant UC and colorectal neoplasia [11]. A similar tendency was observed by Leidenius et al., but the difference in incidence of hepatobiliary malignancy between PSC patients with and without colorectal dysplasia or carcinoma was not so striking [12]. The same tendency was seen in the present study (39% in patients with PSC, IBD and HBC vs. 28% in control PSC patients with IBD, NS). The difference between the incidences of colorectal dysplasia and cancer was increased (46% in patients with HBC vs. 26% in controls, NS) when patients and controls with Mb Crohn or indeterminate colitis were excluded (data not shown). However, we also found that the duration of IBD was somewhat longer (18 vs. 14 years in controls) in the HBC group and may have in¯uenced these ®ndings. The results are also biased by the fact that PSC patients with UC advanced colorectal carcinoma were not included in the present study, because such patients were not referred for liver transplantation. The detection of HBC among patients with PSC using established clinical, biochemical, radiological or cytological methods is inaccurate or even impossible at least at an early stage. The elevated serum levels in tumor markers as well as weight loss and epigastric pain are indicators of manifest carcinoma and do not provide any clues on how to identify
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
PSC patients prone to develop hepatobiliary carcinoma. New, more precise methods for detecting HBC are urgently needed to optimize the timing of liver transplantation and to improve the survival of patients with PSC after transplantation. Positron emission tomography may turn out to be a promising method for the detection even smaller foci of HBC in PSC patients [37]. The need for an early diagnosis of BDC is emphasized, because liver transplantation combined with preoperative chemotherapy has been found to be associated with prolonged survival in highly selected patients with early-stage PSC [38]. In conclusion, the ®ndings of the present study suggest that disease progression in PSC patients with BDC is different, leading to a malignancy instead of an end-stage liver disease. The relatively short duration of PSC and the absence of varices in patients with BDC suggest that BDC is not necessarily a late complication of end-stage PSC as has been assumed previously. On the contrary, HCC seems to be a complication of chronic end-stage liver disease caused by PSC. Neither clinical symptoms, signs nor biochemical ®ndings are helpful in detecting occult HBC among PSC patients.
[11] [12]
[13] [14] [15]
[16] [17]
[18]
Acknowledgements This study was in part supported by grants from Finnish Cancer Foundation and Suomen Astra.
[19] [20]
References [21] [1] Bansal P, Sonnenberg A. Sclerosing cholangitis as a risk factor for gastrointestinal cancer. Gastroenterology 1995;4:A405. [2] Wee A, Ludwig J, Coffey R, LaRusso NF, Wiesner RH. Hepatobiliary carcinoma associated with primary sclerosing cholangitis and chronic ulcerative colitis. Hum Pathol 1985;16:719±726. [3] Rosen CB, Nagorney DM, Wiesner RH, Coffey RJ, LaRusso NF. Cholangiocarcinoma complicating primary sclerosing cholangitis. Ann Surg 1991;213:21±25. [4] Schrumpf E, Abdelnoor M, Fausa O, Elgjo K, Jenssen E, Kolmannskog F. Risk factors in primary sclerosing cholangitis. J Hepatol 1994;21:1061±1066. [5] Broome U, Olsson R, LoÈoÈf L, Bodemar G, Hultcrantz R, Danielsson Ê , et al. Natural history and prognostic factors in 305 Swedish A patients with primary sclerosing cholangitis. Gut 1996;38:610±615. [6] Farrant MJ, Hayllar KM, Wilkinson ML, Karani J, Portmann BC, Westaby D, et al. Natural history and prognostic variables in primary sclerosing cholangitis. Gastroenterology 1991;100:1710±1717. [7] Farges O, Malassagne B, Sebagh M, Bismuth H. Primary sclerosing cholangitis: liver transplantation or biliary surgery. Surgery 1995;117:146±155. [8] Knechtle S, D'Alessandro A, Harms B, Pirsch J, Belzer F, Kalayoglu M. Relationship between sclerosing cholangitis, in¯ammatory bowel disease, and cancer in patients undergoing liver transplantation. Surgery 1995;118:615±620. [9] Harnois D, Gores G, Ludwig J, Steers J, LaRusso N, Wiesner R. Are patients with cirrhotic stage primary sclerosing cholangitis at risk for the development of hepatocellular carcinoma? J Hepatol 1997;27:512±516. [10] Brentnall T, Haggit R, Rabinovitch P, Kimmey M, Bronner M, Levine D, et al. Risk and natural history of colonic neoplasia in patients with
[22] [23]
[24] [25]
[26]
[27] [28] [29] [30]
797
primary sclerosing cholangitis and ulcerative colitis. Gastroenterology 1996;110:331±338. Broome U, LoÈfberg, Veress B, Eriksson LS. Primary sclerosing cholangitis and ulcerative colitis: evidence for increased neoplastic potential. Hepatology 1995;5:1404±1408. Leidenius MHK, FaÈrkkilaÈ MA, KaÈrkkaÈinen P, Taskinen EI, Kellokumpu IH, HoÈckerstedt KAV. Colorectal dysplasia and carcinoma in patients with ulcerative colitis and primary sclerosing cholangitis. Scand J Gastroenterol 1997;32:1521±1525. Coffey R, Wiesner R, Beaver S, Grif®n M, Ludwig J, MacCarty R, et al. Bile duct carcinoma: a late complication of end stage primary sclerosing cholangitis. Hepatology 1984;4:1056. MacCarty R, LaRusso N, May G, Bender C, Wiesner R, King J, et al. Cholangiocarcinoma complicating primary sclerosing cholangitis: cholangiographic appearances. Radiology 1985;156:43±46. Abu-Elmagd K, Malinchoc M, Dickson R, Fung J, Murtaugh P, Langworthy A, et al. Ef®cacy of hepatic transplantation in patients with primary sclerosing cholangitis. Surg Gynecol Obstet 1993;177:335± 344. Miros M, Kerlin P, Walker N, Harper J, Lynch S, Strong R. Predicting cholangiocarcinoma in patients with primary sclerosing cholangitis before transplantation. Gut 1991;33:1369±1373. Bergqvist A, Glaumann H, Persson B, Broome U. Risk factors and clinical presentation of hepatobiliary carcinoma in patients with primary sclerosing cholangitis: a case-control study. Hepatology 1998;27:311±316. Nichols J, Gores G, LaRusso N, Wiesner R, Nagorney D, Ritts R. Diagnostic role of serum CA 19-9 for cholangiocarcinoma in patients with primary sclerosing cholangitis. Mayo Clin Proc 1993;68:874± 879. Ramage J, Donaghy A, Farrant J, Iorns R, Willliams R. Serum tumor markers for the diagnosis of cholangiocarcinoma in primary sclerosing cholangitis. Gastroenterology 1995;108:865±869. Nakeeb A, Lipsett P, Lillemoe K, Fox-Talbot M, Coleman J, Cameron J, et al. Biliary carcinoembryonic antigen levels are a marker for cholangiocarcinoma. Am J Surg 1996;171:147±153. Cobb C, Floyd W. Usefulness of bile cytology in the diagnostic management of patients with biliary tract obstruction. Acta Cytol 1983;29:93±100. Martins E, Fleming K, Garrido M, Hine K, Chapman R. Super®cial thrombophlebitis, dysplasia and cholangiocarcinoma in primary sclerosing cholangitis. Gastroenterology 1994;107:537±542. Ê M, Rhodes JM, Summer®eld JA, Dick Chapman RWG, Arborgh BA R, Scheuer PJ, et al. Primary sclerosing cholangitis: a review of its clinical features, cholangiography, and hepatic histology. Gut 1980;21:870±877. Wiesner RH, LaRusso NF. Clinicopathologic features of the syndrome of primary sclerosing cholangitis. Gastroenterology 1980;79:2000±2006. Truelove SC. Natural history of ulcerative colitis. In: Allan RN, Keighley MRB, Alexander-Williams J, Hawkins C, editors. In¯ammatory bowel disease, 1st ed. Birmingham, UK: Churchill Livingstone, 1984. pp. 94±100. Nixon JB, Riddel RH. Histopathology of ulcerative colitis. In: Allan RN, Keighley MRB, Alexander-Williams J, Hawkins C, editors. In¯ammatory bowel disease, 1st ed. Birmingham, UK: Churchill Livingstone, 1984. pp. 194±209. Lupinetti M, Mehigan D, Cameron J. Hepatobiliary complications of ulcerative colitis. Am J Surg 1980;139:113±118. Dew M, Thompson H, Allan R. The spectrum of hepatic dysfunction in in¯ammatory bowel disease. Q J Med 1979;48:113±135. Broome U, Glaummann H, Hellers G, Nilsson B, SoÈrstad J, Hultcrantz R. Liver disease in ulcerative colitis: an epidemiological and follow up study in the county of Stockholm. Gut 1994;35:84±89. Mikkola K, Kiviluoto T, RiihelaÈ M, Taavitsainen M, JaÈrvinen H. Liver involvement and its course in patients operated on for ulcerative colitis. Hepato-Gastroenterology 1995;42:68±72.
798
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
[31] Aadland E, Schrumpf E, Fausa O, Elgjo K, Heilo A, Askhus T, Gjone E. Primary sclerosing cholangitis: a long-term follow-up study. Scand J Gastroenterol 1987;22:655±664. [32] Weinbren K, Mutum SS. Pathological aspects of cholangiocarcinoma. J Pathol 1983;139:217±238. [33] Van Laethem J, Deviere J, Love J, Gelin M, Cremer M, Adler M. Cholangiographic ®ndings in deteriorating primary sclerosing cholangitis. Endoscopy 1995;27:223±228. [34] Jalanko H, Kuusela P, Roberts P, Sipponen P, Haglund C, MaÈkelaÈ O. comparison of a new tumour marker, CA 19-9, with a-fetoprotein and carcinoembryonic antigen in patients with upper gastrointestinal diseases. J Clin Pathol 1984;37:218±222. [35] Wiesner R, Porayko M, Dickson R, Gores G, LaRusso N, Hay E, et al.
Selection and timing of liver transplantation in primary biliary cirrhosis and primary sclerosing cholangitis. Hepatology 1992;16:1290± 1299. [36] Nashan B, Schlitt H, Tusch G, Oldhafer K, Ringe B, Wagner S, et al. Biliary malignancies in primary sclerosing cholangitis: timing for liver transplantation. Hepatology 1996;23:1105±1111. [37] Keiding S, Hansen S, Rasmussen H, Gee A, Kruse A, Roelsgaard A, et al. Detection of cholangiocarcinoma in primary sclerosing cholangitis by positron emission tomography. Hepatology 1998;28:700±706. [38] De Vreede I, Steers J, Burch P, Rosen C, Gunderson L, Haddock M, Burgart L, Gores G. Prolonged disease-free survival after orthotopic liver transplantation plus adjuvant chemoirradiation for cholangiocarcinoma. Liver Transpl 2000;6:309±316.
www.elsevier.com/locate/jhep
Hepatobiliary carcinoma in primary sclerosing cholangitis: a case control study Marjut Leidenius 1, Krister HoÈckerstedt 1,*, Ulrika Broome 2, Bo-GoÈran Ericzon 2, StyrbjoÈrn Friman 3, Michael Olausson 3, Erik Schrumpf 4 1
Transplantation and Liver Surgery, Helsinki University Hospital, Kasarmikatu 11±13, 00130 Helsinki, Finland 2 Unit of Gastroenterology and Hepatology, Huddinge University Hospital, Huddinge, Sweden 3 Department of Transplantation and Liver Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden 4 Department of Medicine, Rikshospitalet, Oslo, Norway
Backround/Aims: Hepatobiliary carcinoma (HBC) has been considered to be a late complication of end-stage primary sclerosing cholangitis (PSC). The incidence of HBC is approximately 20% in PSC patients evaluated for liver transplantation. The diagnosis of HBC is dif®cult, at least at an early stage and the prognosis is poor even after liver transplantation. The aim of the study was to look for signs and risk factors for developing hepatobiliary carcinoma in patients with PSC. Methods: Thirty-six consecutive patients with PSC and HBC (32 with bile duct carcinoma, BDC, and four with hepatocellular carcinoma, HCC) were pair-matched to control patients referred for liver transplantation because of PSC but who did not have HBC. Gender and age at referral were used as matching factors. Clinical and biochemical data were registered. Results: PSC patients with BDC had a shorter median duration of PSC (1 year) compared with the controls (7 years) and PSC patients with HCC (8 years). There were no statistically signi®cant differences in the liver biochemistry between the patient groups. Varices were more common in patients with PSC and HCC (100%) than in controls (56%) or patients with PSC and HBC (12%) (P , 0:0005). Conclusions: The relatively short duration of PSC and the absence of varices in patients with BDC suggest that BDC, unlike HCC, is not necessarily a late complication of end-stage PSC, as was previously assumed. q 2001 European Association for the Study of the Liver. Published by Elsevier Science B.V. All rights reserved. Keywords: Primary sclerosing cholangitis; Hepatobiliary carcinoma; Ulcerative colitis
1. Introduction Patients with primary sclerosing cholangitis (PSC) have an increased risk of developing gastrointestinal cancer, in particular hepatobiliary malignancies [1±5]. The risk of bile duct cancer (BDC) ranges from 6 to 14% among all PSC patients [2±5]. The prevalence of BDC is approximately 20% in PSC patients evaluated for liver transplantation [6±9]. The prevalence of hepatocellular carcinoma (HCC) in PSC patients is less well known but it is estimated to be 2% in PSC undergoing liver transplantation [9]. PSC is also an additional risk factor for neoplastic alteration of the Received 7 June 2000; received in revised form 12 December 2000; accepted 5 January 2001 * Corresponding author. Tel.: 1358-9-4718-8271; fax: 1358-9-174975.
colorectal mucosa in ulcerative colitis (UC) [10±12]. Broome et al. also noted that the incidence of cholangiocarcinoma in PSC was higher in patients with concomitant UC and colorectal neoplasia. Hepatobiliary carcinoma (HBC) is usually considered to be a late complication of end-stage PSC [2,13,14]. The diagnosis of BDC is dif®cult, at least at an early stage and the prognosis is poor even after liver transplantation [15]. Therefore BDC is nowadays considered as a contra-indication for liver transplantation in many transplantation units. Clinical, biochemical and radiological signs for predicting cholangiocarcinoma in PSC patients have been sought for. Rapid clinical deterioration with jaundice, weight loss associated with an appreciable rise in serum bilirubin [3,16] and abdominal discomfort [3,17] are considered to be signs
0168-8278/01/$20.00 q 2001 European Association for the Study of the Liver. Published by Elsevier Science B.V. All rights reserved. PII: S01 68-8278(01)0004 2-3
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
of (advanced) cholangiocarcinoma in PSC. Cholangiographic features suggesting malignant degeneration include progressive stricture formation, ductal dilatation or polypoid masses [14]. The monitoring of tumor markers CEA and CA 19-9 in serum or bile may be useful for indicating hepatobiliary malignancy in patients with PSC [17±20]. The diagnostic speci®city of bile cytology is very high (up to 100%), but disadvantages of the method are invasiveness and the relatively low sensitivity (approximately 50%), when distinguishing between benign and malignant biliary obstruction [21]. Bile duct dysplasia has been found in over 60% of patients with PSC and cholangiocarcinoma [17]; dysplasia may precede cholangiocarcinoma in patients with PSC [22]. The diagnosis of HBC is dif®cult but important when considering the indications and timing of liver transplantation in PSC. Therefore the aim of this study was to examine clinical or biochemical differences between patients with PSC and HBC and control patients with PSC alone, to look for signs and risk factors for developing hepatobiliary carcinoma in PSC. 2. Materials and methods 2.1. Patients Data on 36 consecutive patients who were referred for liver transplantation between 1987 and 1996 because of PSC and also had HBC, were collected retrospectively from patient records at four Nordic Liver Transplantation Centers: Helsinki in Finland, Stockholm and Gothenburg in Sweden and Oslo in Norway. The study was approved by the Ethical Committees of all transplantation centers. The study was performed in accordance with the Declaration of Helsinki. The diagnosis of PSC was based on established biochemical, radiological and histopathological criteria [23,24]. The onset of PSC was de®ned as the point when cholangiography with ®ndings consistent of PSC was made. The duration of PSC was calculated in two different ways: from the time of cholangiography and also from the time of ®rst presentation of abnormal serum liver function tests. The histologically or cytologically proven diagnosis of HBC was made during laparotomy (usually liver transplantation or an attempt of liver transplantation) in 26 patients, by liver biopsy in eight patients and at autopsy in two patients. Thirty-one patients had bile duct carcinoma (in the intra- or extrahepatic bile ducts or both), one patient had carcinoma of the gallbladder and four patients had hepatocellular carcinoma. Each patient with PSC and HBC included in this study was pair-matched at each center with a control patient also referred for liver transplantation because of PSC but who did not have HBC. Hepatobiliary malignancy was excluded in the control group by careful histological examination of explanted livers after transplantation and by long term follow-up. The matching factors included gender and age at referral as close as possible.
2.2. Clinical examination Presence or absence of various clinical symptoms and signs were recorded at the time of evaluation for liver transplantation. These included jaundice, itching, presence of varices, variceal bleeding, febrile cholangitis, epigastric pain and weight loss during 12 months. Patients in both groups had undergone colonoscopy with multiple biopsies to diagnose or to exclude concomitant in¯ammatory bowel disease (IBD), except for those who had undergone proctocolectomy. The diagnosis of IBD was based on
793
clinical features and the radiological, endoscopic and histological appearance of the large bowel and rectum [25,26]. All patients with longstanding colitis were included in colonoscopy surveillance in order to detect colorectal carcinoma or dysplasia.
2.3. Biochemical methods The biochemical tests for liver function used as part of the medical evaluation prior to transplantation included measurements of serum concentrations of alanine transferases (S-ALT), alkaline phosphatases (SALP), total bilirubin and albumin, as well as prothrombin complex assay (P-TT-SPA). The tumor marker CEA was determined in 16 patients and 16 controls and CA 19-9 in 13 patients and 11 controls.
2.4. Statistical methods Fisher's exact test was used to compare percentages. Medians were compared using the Mann-Whitney U-test.
3. Results The median age at referral was 45 years in both patients (range 18±67 years) and controls (range 27±64 years). PSC patients with HBC were somewhat older at onset of PSC (not signi®cant (NS)) and had a 5 years shorter median duration of PSC (P , 0:01) than control patients without HBC when the onset of PSC was de®ned as the time point when cholangiography with ®ndings consistent of PSC was made. The median duration of PSC was 6 years (range 0±24 years) in patients with HBC and 8 years (range 0±23 years) in controls (NS) calculated from the time of ®rst presentation of abnormal serum liver function tests. The proportion of patients with very short duration of PSC (0±1 years) was 31% (11/36) in the HBC group and 6% (2/36) in the control group (P , 0:02) using the time of ®rst presentation of abnormal serum liver function tests as the onset of PSC. (Table 1). Altogether 31 (86%) PSC patients with HBC also had IBD: 27 had UC, three had unde®ned colitis and one patient had Crohn's disease. Twenty-nine (81%) control patients Table 1 Characteristics of PSC patients with HBC and control patients without HBC at referral for liver transplantation a
Age (years) Men/women Age at onset of PSC (years) Duration of PSC (years) b Duration of PSC (years) d Patients with 0±1 yearsc duration of PSC a
HBC (n 36)
Control (n 36)
45 (18±70) 26/10 41 (14±66) 2 (0±18) c 6 (0±24) 11/36 e
45 (27±70) 26/10 35 (13±63) 7 (0±20) 8 (0±23) 2/36
Values are median (range). The onset of PSC calculated from the time of cholangiography c The onset of PSC calculated from the time of ®rst presentation of abnormal serum liver function tests. d P , 0:01. e P , 0:02. b
794
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
Table 2 In¯ammatory bowel disease (IBD) in PSC patients with HBC and control patients without HBC a
IBD (%) Age at onset of IBD (years) Duration of IBD to the time of colectomy or to the referral (years) Duration of IBD to the referral (years) Duration of IBD to onset of PSC (years) Colorectal dysplasia or carcinoma (%) Colectomy rate (%)
HBC (n 36)
Control (n 36)
86 20 (8±50) 18 (2±35)
81 22 (13±69) 14 (0±33)
22 (4±44)
20 (0±40)
16 (9±43)
11 (19±33)
39
28
45
41
a Values are median (range). The differences between the groups were not statistically signi®cant.
without HBC had IBD: 28 had UC and one had unde®ned colitis. The median age at onset of IBD was 20 years in the HBC patients and 22 years in the controls. Median duration of IBD to colectomy or to the date of referral was 18 years and 22 years to the date of referral in patients with HBC, and 14 years and 20 years, respectively, in the control patients. The median interval between the onset of IBD and the onset of PSC was 16 years and 11 years in the controls (NS). Six patients with HBC had undergone colectomy because of
active colitis, six because of colorectal dysplasia or carcinoma and two patients had undergone prophylactic proctocolectomy because of long-standing UC. Five control patients were operated on because of active UC and seven because of colorectal neoplasia. Twelve patients (39%) with HBC and eight controls (28%) exhibited colorectal dysplasia or carcinoma (NS) (Table 2). There were no statistically signi®cant differences in the liver biochemistry at referral between the patients with HBC and the controls without HBC. Serum CEA was measured in 16 HBC patients and in 16 controls. Serum CA 19-9 was determined in 13 HBC patients and in 11 control patients. Increased levels of both CEA (median 9.45 mg/l) and CA 19-9 (median 120 units/l) were observed in the HBC patients compared with the control group (median CEA 2 mg/l and median CA 19-9 37 units/l) (P , 0:05, both for CEA and CA 19-9) (Table 3). Both tumor markers were measured in eight patients and eight controls and the combined tumor marker index was calculated [19]. The index exceeded 400 units in ®ve patients with BDC, whereas the index was under 400 in all patients with HCC and in all controls (P , 0:05 between BDC patients and controls). Jaundice, itching or febrile cholangitis were as common in patients with PSC and HBC as in the control patients with PSC alone. Esophageal (or peristomal) varices were more common in control patients with PSC alone (56%) than in patients with PSC and HBC (12%) (P , 0:0005). Control
Table 3 Characteristics of PSC patients with BDC or HCC and control patients without HBC at referral for liver transplantation a
Age (years) Men/women Age at onset of PSC (years) b Duration of PSC (years) b S-ALT (,40 units/l) S-ALP (,280 units/l) S-Bilirubin (,20 mmol/l) P-TT-SPA (.70%) S-Albumin (.36 g/l) CEA (,5 mg/l) CA 19±9 (,37 units/l) Varices (%) Bleeding from varices (%) Jaundice (%) Itching (%) Febrile cholangitis (%) Epigastric pain (%) Weight loss in 12 months (%) a b c d e f g h
BDC (n 32)
HCC (n 4)
Control (n 36)
45 (18±70) 22/10 41 (14±66) 1 (0±17) c 96 (26±282) 1380 (378±3446) 180 (10±658) 91 (26±150) 31 (16±40) 10.4 (1±120) d 120 (35±8100) e 0f 0g 85 68 31 67 h 75
49 (18±67) 4/0 33 (18±59) 8 (0±18) 101 (28±240) 620 (228±2460) 63 (31±85) 67 (41±130) 28 (24±43) 4.4 (1.5±7.2) 6 (2±10) 100 100 50 0 0 33 0
45 (27±70) 26/10 35 (13±63) 7 (0±20) 117 (29±406) 1155 (379±3480) 106 (8±500) 68 (25±130) 27 (16±44) 2 (1±7) 37 (12±82) 56 31 83 55 28 7 39
Values are median (range). The onset of PSC calculated from the time of cholangiography. P , 0:005 compared with the controls. P , 0:05 compared with the controls. P , 0:0005 compared with the controls, P , 0:05 compared with the HCC patients. P , 0:00005 compared with the controls and to the HCC patients. P , 0:001 compared with the controls, P , 0:00005 compared with the HCC patients. P , 0:0005 compared with the controls.
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
patients also seemed to have bleeding from varices more often (31 vs. 12% in HBC patients) but the difference was not statistically signi®cant. Epigastric pain was more common in HBC patients (63%) than in control PSC patients (7%) (P , 0:00005). Seventy-one percent of HBC patients had lost weight during last the 12 months, compared with 39% of the controls (NS) (Table 3). Four patients with PSC and HBC had HCC, and 32 had BDC. The median age at referral was 45 years (range 18±70 years) in BDC patients and 49 years (range 18±67 years) in HCC patients (NS). The age at onset of PSC was 41 years (range 14±66 years) in the BDC patients and 33 years (range 18±59 years) in the HCC patients (NS). The median duration of PSC was 1 year (range 0±17 years) in patients with BDC and 8 years (range 0±18 years) in patients with HCC (NS) when the onset of PSC was de®ned as the point when cholangiography with ®ndings consistent of PSC was made. The median duration of PSC calculated from the time of ®rst presentation of abnormal serum liver function tests was 5 years (range 0±24 years) in the BDC group and 15 years (range 0±18 years) in the HCC group (NS). Both the median S-ALP (1380 vs. 620 units/l) and S-bil (180 vs. 63 mmol/l) were somewhat higher in patients with BDC compared with HCC patients, but the differences were not statistically signi®cant. Varices and variceal bleeding were encountered in all of the HCC patients but in none of the BDC patients (P , 0:00005). The median serum CEA was 10.4 mg/l (range 1±120 mg/l) in the BDC patients and 4.4 mg/l (1.5± 7.2 mg/l) in the HCC patients (NS). The median serum CA 19-9 was higher in the BDC patients (120 units/l, range 35± 8100 units/l) than in the HCC patients (6 units/l, range 2±10 units/l) (P , 0:05) (Table 3). 4. Discussion The major ®nding in this case-control study is that HBC, especially BDC, may develop also in patients with an early phase PSC and thus is not necessarily a late complication of end-stage PSC as has been previously assumed [2,13,14]. Patients with HBC had a ®ve year shorter median duration of PSC than the control patients without HBC. The HBC patients had a tendency towards onset of PSC at a later age. They also seemed to have a longer duration of IBD before the onset of PSC. In a recent study by Bergqvist et al., there were no differences in the duration of PSC between the PSC patients with and without HBC [17]. They de®ned the onset of PSC as the time of ®rst presentation of abnormal serum liver function tests consistent with the diagnosis of PSC. It should be mentioned that some of the patients in the study by Bergqvist et al. (12 HBC patients and four control patients) are actually also included in the present study. In the present study the onset of PSC was de®ned as the time point when cholangiography with ®ndings consistent of PSC was made. In addition, the duration of PSC was
795
calculated in two different ways: from the time of cholangiography and from the time of ®rst presentation of abnormal serum liver function tests. Abnormal liver function tests do not necessarily stand for PSC. Majority of PSC patients have IBD (over 80% in the present study), usually extensive UC. Transient elevations in liver function tests are encountered in patients with UC in up to 50% [27]. Usually these transient abnormalities in liver function tests are associated with active colitis or surgery [28,29]. At least mild, persistent liver involvement has been found in 21% of UC patients at the time of colectomy although the criteria of PSC was ful®lled only in one third of them [30]. These milder forms of hepatobiliary dysfunction did not show any progression in the follow-up [30]. The golden standard in the diagnosis of PSC is endoscopic retrograde cholangiography (ERC). ERC is an invasive procedure with signi®cant complication risks. Therefore it has not always been carried out at the time of ®rst presentation of abnormal liver function tests. This may cause delay in the diagnosis of PSC in some patients. In addition, PSC has even been encountered without laboratory abnormalities [31]. Therefore it is impossible to determine the exact onset of PSC by serum liver function tests alone. Presence of esophageal or peristomal varices or variceal bleedings were rare in the HBC patients compared with the control patients. This ®nding supports our presumption that HBC may develop before disease progression to end-stage PSC. Correspondingly, in an earlier study by Miros et al. PSC patients with HBC seemed to have a less advanced histological stage of PSC than those without HBC [16]. Wee et al. have also detected HBC in patients with small duct PSC [2]. Unfortunately, suf®cient data regarding the histological stage of PSC patients with HBC was not available in the present study. With regard to patients with PSC and HCC the ®ndings of the present study are in close agreement with a study by Harnois et al., suggesting HCC to be a complication of end-stage PSC [9]. All HCC patients had bled from esophageal varices as a sign of chronic, end-stage liver disease while none of BDC patients had peristomal or esophageal varices. Patients with HCC had a longer median duration of PSC compared with those with BDC. Also the ®ndings at liver biochemistry were slightly different in BDC patients, showing a greater degree of cholestasis in them. Correspondingly, itching was encountered in 61% of BDC patients but in none of the HCC patients. However, the differences between HCC and BDC patients in duration of PSC or in parameters representing cholestasis were not statistically signi®cant, probably due to the small number of HCC patients. To our knowledge, these ®ndings on HCC in PSC patients have not been reported earlier. Sixteen patients had had PSC just for 1 year or less before the detection of HBC, when the onset of PSC was de®ned as the time point when cholangiography with ®ndings consistent of PSC was made. This may give rise to doubt about the diagnosis of PSC in these patients. One may argue that in
796
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
some patients the cholangiographic ®ndings which have led to the diagnosis of PSC, may actually have been due to the HBC itself. The diagnosis of PSC seems to be justi®ed also in these HBC patients with a very short duration of preexisting PSC. Twelve of these 16 patients also had IBD. HBC is assumed to be very rare among patients with IBD without any co-existing large or small duct PSC [2]. Three of the four patients without IBD had undergone liver transplantation with thorough histological examination of the removed liver, with ®ndings typical of PSC and complicating HBC [2], and not to a sclerosing cholangiocarcinoma without pre-existing PSC [32] The number of the HBCpatients with very short duration of PSC was 11, when the duration of PSC was calculated from the time of ®rst presentation of abnormal serum liver function tests. We found that epigastric pain was more common in HBC patients. There was also a tendency towards a greater weight loss in this group. Both symptoms have also been related to HBC in previous studies [3,16]. However, epigastric pain as well as weight loss are signs of advanced malignancy and do not help in the early diagnosis of HBC. In addition, the presence of ascites and the use of diuretics render it dif®cult to evaluate the actual wasting of a patient merely by measuring the weight. Jaundice, itching or abnormalities in liver function tests do not distinguish between the HBC and the control group, because these could be either due to HBC or to the PSC, itself. The ®ndings of earlier studies have been controversial in regard to sudden clinical deterioration or rise in serum bilirubin as signs of HBC in PSC patients [3,16,17]. The rapid clinical or biochemical deterioration can be due to HBC as well as liver decompensation or a dominant benign stricture [17,33]. MacCarty et al. have found that markedly dilated ducts or ductal segments, large (1 cm or greater) polypoid masses and progressive stricture formation or dilatation suggest cholangiocarcinoma in PSC [14]. Van Laethem et al. considered dominant intrahepatic duct strictures as malignant and extrahepatic strictures (with benign brush cytology) to be due to the PSC itself [33]. On the contrary, Bergqvist et al. found it impossible to differentiate malignant strictures from benign by cholangiographic appearances [17]. PSC patients with dominant strictures may bene®t from endoscopic stenting [33]. However, liver transplantation must be considered in all patients with dominant strictures even in those with less advanced PSC, because it is impossible to rule out HBC. Elevated serum levels of tumor markers CEA and Ca19-9 have been detected in PSC patients with BDC compared with patients with stable or progressive PSC without BDC [17±19]. Our own results are in close agreement with these previous studies. Tumor markers CEA and CA 19-9 are widely used in follow-up of PSC patients in order to detect BDC. However, at least moderately elevated serum tumor marker levels have also been found in PSC patients without malignancies [18,19], probably due to obstructive jaundice [34]. Normal serum tumor marker levels do not exclude
BDC [18,19] and neither is monitoring CEA and CA 19-9 helpful in the diagnosis of HCC. Anyway, by combined use of both markers it may be possible to detect most occult bile duct malignancies and exclude inappropriate cases for liver transplantation more effectively [19]. Unfortunately, malignancies detected by the measurement of CEA or CA 19-9 serum levels or by combined use of these markers, are usually at an incurable stage and little can be done to help these patients [19]. The optimal timing of liver transplantation has been addressed by Wiesner et al. who developed the Mayo PSC natural history model [35]. An important limitation to applying this risk score model concerns the risk of HBC among PSC patients. The survival of patients with PSC and HBC has not been signi®cantly improved by liver transplantation, because it has been impossible to include such components as `undiagnosed HBC' in the Mayo model [15]. In Hannover, Nashan et al. compared Mayo risk scores with the incidence of HBC in PSC patients [36]. They found that all HBC patients had a risk score of 4.3 or higher. Correspondingly, they assumed that the risk of HBC is very low in PSC patients with a risk score of 4.0 or lower. Unfortunately they had only very few patients with a risk score under 4.3 in their study to support their assumption. They also noticed that most HBC patients were included in the medium and not in the high-risk score group. The most probable explanation to this con¯icting ®nding is that the components included in the Mayo model (age, serum bilirubin concentration, splenomegaly, histological stage) can hardly to be expected to distinguish between PSC patients with and without HBC. Broome et al. have noted that the risk of bile duct carcinoma in PSC was concentrated on patients with concomitant UC and colorectal neoplasia [11]. A similar tendency was observed by Leidenius et al., but the difference in incidence of hepatobiliary malignancy between PSC patients with and without colorectal dysplasia or carcinoma was not so striking [12]. The same tendency was seen in the present study (39% in patients with PSC, IBD and HBC vs. 28% in control PSC patients with IBD, NS). The difference between the incidences of colorectal dysplasia and cancer was increased (46% in patients with HBC vs. 26% in controls, NS) when patients and controls with Mb Crohn or indeterminate colitis were excluded (data not shown). However, we also found that the duration of IBD was somewhat longer (18 vs. 14 years in controls) in the HBC group and may have in¯uenced these ®ndings. The results are also biased by the fact that PSC patients with UC advanced colorectal carcinoma were not included in the present study, because such patients were not referred for liver transplantation. The detection of HBC among patients with PSC using established clinical, biochemical, radiological or cytological methods is inaccurate or even impossible at least at an early stage. The elevated serum levels in tumor markers as well as weight loss and epigastric pain are indicators of manifest carcinoma and do not provide any clues on how to identify
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
PSC patients prone to develop hepatobiliary carcinoma. New, more precise methods for detecting HBC are urgently needed to optimize the timing of liver transplantation and to improve the survival of patients with PSC after transplantation. Positron emission tomography may turn out to be a promising method for the detection even smaller foci of HBC in PSC patients [37]. The need for an early diagnosis of BDC is emphasized, because liver transplantation combined with preoperative chemotherapy has been found to be associated with prolonged survival in highly selected patients with early-stage PSC [38]. In conclusion, the ®ndings of the present study suggest that disease progression in PSC patients with BDC is different, leading to a malignancy instead of an end-stage liver disease. The relatively short duration of PSC and the absence of varices in patients with BDC suggest that BDC is not necessarily a late complication of end-stage PSC as has been assumed previously. On the contrary, HCC seems to be a complication of chronic end-stage liver disease caused by PSC. Neither clinical symptoms, signs nor biochemical ®ndings are helpful in detecting occult HBC among PSC patients.
[11] [12]
[13] [14] [15]
[16] [17]
[18]
Acknowledgements This study was in part supported by grants from Finnish Cancer Foundation and Suomen Astra.
[19] [20]
References [21] [1] Bansal P, Sonnenberg A. Sclerosing cholangitis as a risk factor for gastrointestinal cancer. Gastroenterology 1995;4:A405. [2] Wee A, Ludwig J, Coffey R, LaRusso NF, Wiesner RH. Hepatobiliary carcinoma associated with primary sclerosing cholangitis and chronic ulcerative colitis. Hum Pathol 1985;16:719±726. [3] Rosen CB, Nagorney DM, Wiesner RH, Coffey RJ, LaRusso NF. Cholangiocarcinoma complicating primary sclerosing cholangitis. Ann Surg 1991;213:21±25. [4] Schrumpf E, Abdelnoor M, Fausa O, Elgjo K, Jenssen E, Kolmannskog F. Risk factors in primary sclerosing cholangitis. J Hepatol 1994;21:1061±1066. [5] Broome U, Olsson R, LoÈoÈf L, Bodemar G, Hultcrantz R, Danielsson Ê , et al. Natural history and prognostic factors in 305 Swedish A patients with primary sclerosing cholangitis. Gut 1996;38:610±615. [6] Farrant MJ, Hayllar KM, Wilkinson ML, Karani J, Portmann BC, Westaby D, et al. Natural history and prognostic variables in primary sclerosing cholangitis. Gastroenterology 1991;100:1710±1717. [7] Farges O, Malassagne B, Sebagh M, Bismuth H. Primary sclerosing cholangitis: liver transplantation or biliary surgery. Surgery 1995;117:146±155. [8] Knechtle S, D'Alessandro A, Harms B, Pirsch J, Belzer F, Kalayoglu M. Relationship between sclerosing cholangitis, in¯ammatory bowel disease, and cancer in patients undergoing liver transplantation. Surgery 1995;118:615±620. [9] Harnois D, Gores G, Ludwig J, Steers J, LaRusso N, Wiesner R. Are patients with cirrhotic stage primary sclerosing cholangitis at risk for the development of hepatocellular carcinoma? J Hepatol 1997;27:512±516. [10] Brentnall T, Haggit R, Rabinovitch P, Kimmey M, Bronner M, Levine D, et al. Risk and natural history of colonic neoplasia in patients with
[22] [23]
[24] [25]
[26]
[27] [28] [29] [30]
797
primary sclerosing cholangitis and ulcerative colitis. Gastroenterology 1996;110:331±338. Broome U, LoÈfberg, Veress B, Eriksson LS. Primary sclerosing cholangitis and ulcerative colitis: evidence for increased neoplastic potential. Hepatology 1995;5:1404±1408. Leidenius MHK, FaÈrkkilaÈ MA, KaÈrkkaÈinen P, Taskinen EI, Kellokumpu IH, HoÈckerstedt KAV. Colorectal dysplasia and carcinoma in patients with ulcerative colitis and primary sclerosing cholangitis. Scand J Gastroenterol 1997;32:1521±1525. Coffey R, Wiesner R, Beaver S, Grif®n M, Ludwig J, MacCarty R, et al. Bile duct carcinoma: a late complication of end stage primary sclerosing cholangitis. Hepatology 1984;4:1056. MacCarty R, LaRusso N, May G, Bender C, Wiesner R, King J, et al. Cholangiocarcinoma complicating primary sclerosing cholangitis: cholangiographic appearances. Radiology 1985;156:43±46. Abu-Elmagd K, Malinchoc M, Dickson R, Fung J, Murtaugh P, Langworthy A, et al. Ef®cacy of hepatic transplantation in patients with primary sclerosing cholangitis. Surg Gynecol Obstet 1993;177:335± 344. Miros M, Kerlin P, Walker N, Harper J, Lynch S, Strong R. Predicting cholangiocarcinoma in patients with primary sclerosing cholangitis before transplantation. Gut 1991;33:1369±1373. Bergqvist A, Glaumann H, Persson B, Broome U. Risk factors and clinical presentation of hepatobiliary carcinoma in patients with primary sclerosing cholangitis: a case-control study. Hepatology 1998;27:311±316. Nichols J, Gores G, LaRusso N, Wiesner R, Nagorney D, Ritts R. Diagnostic role of serum CA 19-9 for cholangiocarcinoma in patients with primary sclerosing cholangitis. Mayo Clin Proc 1993;68:874± 879. Ramage J, Donaghy A, Farrant J, Iorns R, Willliams R. Serum tumor markers for the diagnosis of cholangiocarcinoma in primary sclerosing cholangitis. Gastroenterology 1995;108:865±869. Nakeeb A, Lipsett P, Lillemoe K, Fox-Talbot M, Coleman J, Cameron J, et al. Biliary carcinoembryonic antigen levels are a marker for cholangiocarcinoma. Am J Surg 1996;171:147±153. Cobb C, Floyd W. Usefulness of bile cytology in the diagnostic management of patients with biliary tract obstruction. Acta Cytol 1983;29:93±100. Martins E, Fleming K, Garrido M, Hine K, Chapman R. Super®cial thrombophlebitis, dysplasia and cholangiocarcinoma in primary sclerosing cholangitis. Gastroenterology 1994;107:537±542. Ê M, Rhodes JM, Summer®eld JA, Dick Chapman RWG, Arborgh BA R, Scheuer PJ, et al. Primary sclerosing cholangitis: a review of its clinical features, cholangiography, and hepatic histology. Gut 1980;21:870±877. Wiesner RH, LaRusso NF. Clinicopathologic features of the syndrome of primary sclerosing cholangitis. Gastroenterology 1980;79:2000±2006. Truelove SC. Natural history of ulcerative colitis. In: Allan RN, Keighley MRB, Alexander-Williams J, Hawkins C, editors. In¯ammatory bowel disease, 1st ed. Birmingham, UK: Churchill Livingstone, 1984. pp. 94±100. Nixon JB, Riddel RH. Histopathology of ulcerative colitis. In: Allan RN, Keighley MRB, Alexander-Williams J, Hawkins C, editors. In¯ammatory bowel disease, 1st ed. Birmingham, UK: Churchill Livingstone, 1984. pp. 194±209. Lupinetti M, Mehigan D, Cameron J. Hepatobiliary complications of ulcerative colitis. Am J Surg 1980;139:113±118. Dew M, Thompson H, Allan R. The spectrum of hepatic dysfunction in in¯ammatory bowel disease. Q J Med 1979;48:113±135. Broome U, Glaummann H, Hellers G, Nilsson B, SoÈrstad J, Hultcrantz R. Liver disease in ulcerative colitis: an epidemiological and follow up study in the county of Stockholm. Gut 1994;35:84±89. Mikkola K, Kiviluoto T, RiihelaÈ M, Taavitsainen M, JaÈrvinen H. Liver involvement and its course in patients operated on for ulcerative colitis. Hepato-Gastroenterology 1995;42:68±72.
798
M. Leidenius et al. / Journal of Hepatology 34 (2001) 792±798
[31] Aadland E, Schrumpf E, Fausa O, Elgjo K, Heilo A, Askhus T, Gjone E. Primary sclerosing cholangitis: a long-term follow-up study. Scand J Gastroenterol 1987;22:655±664. [32] Weinbren K, Mutum SS. Pathological aspects of cholangiocarcinoma. J Pathol 1983;139:217±238. [33] Van Laethem J, Deviere J, Love J, Gelin M, Cremer M, Adler M. Cholangiographic ®ndings in deteriorating primary sclerosing cholangitis. Endoscopy 1995;27:223±228. [34] Jalanko H, Kuusela P, Roberts P, Sipponen P, Haglund C, MaÈkelaÈ O. comparison of a new tumour marker, CA 19-9, with a-fetoprotein and carcinoembryonic antigen in patients with upper gastrointestinal diseases. J Clin Pathol 1984;37:218±222. [35] Wiesner R, Porayko M, Dickson R, Gores G, LaRusso N, Hay E, et al.
Selection and timing of liver transplantation in primary biliary cirrhosis and primary sclerosing cholangitis. Hepatology 1992;16:1290± 1299. [36] Nashan B, Schlitt H, Tusch G, Oldhafer K, Ringe B, Wagner S, et al. Biliary malignancies in primary sclerosing cholangitis: timing for liver transplantation. Hepatology 1996;23:1105±1111. [37] Keiding S, Hansen S, Rasmussen H, Gee A, Kruse A, Roelsgaard A, et al. Detection of cholangiocarcinoma in primary sclerosing cholangitis by positron emission tomography. Hepatology 1998;28:700±706. [38] De Vreede I, Steers J, Burch P, Rosen C, Gunderson L, Haddock M, Burgart L, Gores G. Prolonged disease-free survival after orthotopic liver transplantation plus adjuvant chemoirradiation for cholangiocarcinoma. Liver Transpl 2000;6:309±316.