Article Text

Liver posters 410–441
Free

## 410 URINARY TAURINE AND HIPPURATE ARE USEFUL MARKERS OF ALCOHOLIC CIRRHOSIS

K. Dabos, P. Ramachandran, I. Sadler1, J. Plevris, P. Hayes. Liver Cell Biology Laboratory, Department of Medicine, Royal Infirmary of Edinburgh, Edinburgh, UK; 1The Department of Chemistry, University of Edinburgh, King's Buildings, Edinburgh, UK

Both taurine and hippurate are mainly produced by the liver and their excretion rate in the urine could provide us with markers in liver dysfunction. We aimed to evaluate the usefulness of urinary taurine and hippurate levels as markers of cirrhosis.

Materials and Methods: Urine was collected from 40 patients with alcoholic cirrhosis (18 males and 12 females) aged 37 -74 (mean age 52.9) and 20 controls (25 males and 15 females) aged 21–68 years old (mean age 49.9) with normal liver function. All patients had moderate to severe liver disease (mean Child's Pugh score 9.3) due to ethanol abuse. We used 1H NMR spectroscopy to quantify levels of taurine and hippurate in the patients urine. Both taurine and hippurate were expressed as excretion indexes relative to the amount of creatinine in each sample. ANOVA was applied to compare values between the groups.

Results: Taurine excretion index was significantly higher in cirrhotics than controls (0.27 ± 0.04 vs 0.046 ± 0.005) (p<0.009). Hippurate excretion index was significantly lower in patients with cirrhosis than controls ( 0.097 ± 0.016 vs 0.25 ± 0.06) (p<0.014). If the two values were combined then the results were again highly statistically significant (p<0.000126).

Conclusions: A combination of low hippurate and high taurine excretions is highly significant in alcoholic cirrhosis and can be a cheap non invasive marker of the disease.

## 411 HAZARDOUS DRINKING IN HOSPITAL INPATIENTS: STILL COMMON, STILL UNDER DIAGNOSED

E.J. Williams, E. McFarlane, E. Rigney, B. Saward, M.P. Bradley, D. Ray-Chaudhuri, C. Davidson, D. Gleeson. Liver Unit, Sheffield Teaching Hospitals, Sheffield, UK

Introduction: Brief counselling for hazardous drinkers can lead to reduced alcohol consumption, hence detection is important.

Aims: a): To establish the prevalence and detection rate of hazardous drinking in unselected hospital inpatients and b) to ascertain whether, in patients who present with a first episode of decompensated alcoholic liver disease (ALD), opportunities to manage excessive drinking during previous admissions were exploited.

Conclusion: hazardous drinking remains common, under-diagnosed and under-managed in hospital inpatients.

## 412 FUNCTIONAL POLYMORPHISMS IN THE RENIN-ANGIOTENSIN SYSTEM (RAS) ARE NOT RELATED TO FIBROSIS IN CHRONIC HEPATITIS C (HCV) INFECTION

D. Thorburn1, E.H. Forrest1, E. Spence2, K. Oien3, G. Inglis4, C. Smith5, E.A.B. McCruden5, R. Fox6, P.R. Mills2.1Victoria Infirmary;2Gastroenterology Unit and6Brownlee Unit, Gartnavel General Hospital;3Department of Pathology and4Department of Medicine, Western Infirmary;5Institute of Virology, University of Glasgow, Glasgow, UK

Background: Angiotensin II (AngII), the main effector molecule of the RAS, may influence hepatic fibrosis. Functional polymorphisms in components of the RAS [angiotensinogen (Ang), angiotensin converting enzyme (ACE), AngII receptor (AT1R)] which alter gene expression and RAS phenotype are recognised. We aimed to assess the role of functional RAS polymorphisms in the progression of liver fibrosis in Scottish patients with chronic HCV.

Methods: 195 patients with HCV (RT PCR positive) and chronic hepatitis on biopsy were grouped by stage of liver fibrosis. Rates of RAS polymorphism were recorded in each fibrosis group.

Results: Patients homozygous for 2 or 3 high expression RAS mutations had similar stages of fibrosis compared with those homozygous for one or none (p = 0.95). Rates (%) of RAS polymorphisms in each fibrosis group are shown in the table. On multiple linear regression advanced fibrosis was associated with a history of excess alcohol consumption (p<0.01) and the grade of inflammation on liver biopsy (p<0.001).

Abstract 412

Conclusions: RAS polymorphisms are not associated with accelerated progression of fibrosis in chronic HCV infection.

## 413 THYROID HORMONE (T3) INCREASES THE FUNCTIONAL CAPACITY OF THE INTACT LIVER

R. Malik, A. Ala, R. Tootle, H. Hodgson. Royal Free Hospital, Rowland Hill St, London NW3 2PF, UK

Background: We characterised the time course and magnitude of effect of tri-iodothyronine (T3) as a primary mitogen for the liver in rats. A single (4mg/kg) s/c dose of T3 induces a proliferative response within the liver, which peaks at 24h (7% cell proliferation index) followed by a gradual decline. This increase in cell proliferation results in an increase in liver mass that peaks at 10 days. The 15% increase in liver mass at 10 days (compared to controls) is associated with corresponding increases in total DNA and liver protein levels confirming an increased cell number.

Aim: To assess if T3 induced increases in liver mass confer a useful increase in hepatic function, using galactose elimination capacity (GEC).

Method: Two groups of rats (n=5) were assigned to either T3 or control (vehicle only) ten days prior to assessing the galactose elimination capacity by: (1) Administering 0.5 mls of 50% galactose via the internal jugular vein approach. (2) A 0.5ml venesection was performed every ten minutes between 20 and 50 minutes. (3) A bladder puncture performed at the end to collect urine. Galactose elimination capacity was calculated as the ratio of the injected amount of galactose (with correction for urinary excretion) and the extrapolated time to zero concentration. All animals were approximately 250g to eliminate variations in GEC due to bodyweight. Results are given as the mean ± standard deviation of the sample. Statistical differences were determined using the two tailed t-test and reported if p<0.05.

Results: In rats receiving T3 the GEC was 9.3 ± 0.6 μmol/min as compared to control rats in which the figure was 7.9 ± 0.4 μmol/min (p<0.01).

Conclusion: A single injection of thyroid hormone results in an increase in liver mass that peaks at 10 days. This increase in liver mass enhances the metabolic capacity (as assessed by GEC) of the intact rat liver by 20%. The ability to increase functional hepatic mass could be therapeutically valuable if applicable to man.

## 414 TRANSIENT GENERATION OF CORE CD8+ CYTOTOXIC T-CELL ESCAPE MUTANTS DURING PRIMARY HBV INFECTION

S.A. Whalley, G.J.M. Webster, D. Brown, C.G. Teo1, A. Bertoletti, V.C. Emery, G.M. Dusheiko. Royal Free University College;1Central Public Health Laboratory, London, UK

Previous studies have found little evidence that mutations affect B- or T- cell epitopes during the course of acute HBV infection. It has been speculated that the vigorous and broadly reactive nature of the CTL response during acute infection prevents the emergence of CTL escape mutants. We investigated six patients with acute resolving hepatitis and four patients who progressed to chronic infection for the emergence of CD8+ cells. Our methodology included PCR, cloning, denaturing gradient gel electrophoresis and DNA sequence analyses. All non-synonymous mutations detected in the core region occurred in regions previously mapped as B, CD4+ or CD8+ epitopes. Four patients were HLA A*0201. Tetramers containing the A*0201 restricted core epitope 18–27 showed that three patients with acute resolving hepatitis developed CD8+ T-cells directed at this epitope whereas the patient who developed chronic infection did not. All three patients with detectable CD8+ CTL response developed mutations encompassing the core 18–27 epitope while the remaining seven patients showed genetic stability within the same core region (p=0.01). Previous studies have shown that mutation within the core 18–27 epitope is less efficiently recognised by the prototypic anti-core 18–27 CD8+ CTL response than the wild type sequence, confirming that these variants represent CD8+ CTL escape mutants. However, our study demonstrated that patients with acute HBV infection and CD8+ CTL escape mutants ultimately cleared HBV from serum indicating that the broadly reactive nature of the immune response was capable of clearing such evolving mutants and preventing viral persistence.

## 415 INCREASED INTESTINAL PERMEABILITY IS ASSOCIATED WITH ALTERATIONS IN LIVER FUNCTION TESTS (LFTs) IN HEALTHY SUBJECTS BUT EFFECT IS NOT MEDIATED THROUGH SYSTEMIC ENDOTOXAEMIA

A. Poullis1,2, R. Foster2, A. Shetty1, M.K. Fagerhol3, R. Barclay4, T.C. Northfield2, M.A. Mendall1.1Mayday University Hospital;2St George's Hospital Medical School;3University of Oslo, Norway;4Western General Hospital, Edinburgh, UK

Non-alcoholic steatohepatitis (NASH) is associated with increased small bowel permeability. Nothing is known about the interaction of intestinal permeability, systemic endotoxaemia and LFTs in healthy subjects. Faecal calprotectin is a surrogate marker of intestinal permeability and correlates well with urinary excretion of enterally administered 51Cr-EDTA.

Aims: To assess the interaction between faecal calprotectin, systemic endotoxaemia and LFTs in healthy middle aged subjects.

Methods: 230 subjects (155 male, 75 female) aged between 50 and 70 were recruited at random from GP lists in South London. Patients with known liver disease were excluded. A previously validated lifestyle questionnaire was completed. LFTs were measured by auto-analyser. Endotoxin was analysed using the Limulus amoebocyte lysate (LAL) assay. A stool sample was analysed for calprotectin by ELISA.

Results: Using Spearmans Rank test there was a positive association between calprotectin tertiles and alkaline phosphatase (p=0.004), aspartate transaminase (p=0.02) and γ GT (p=0.007). There was no correlation with bilirubin or alanine transaminase. There was no association between systemic endotoxin and LFTs.

Conclusion: The finding that a surrogate marker of intestinal permeability is associated with alterations in LFTs confirms that this may be of pathophysiological importance in NASH. Systemic endotoxin does not affect LFTs in healthy middle aged subjects. Bacterial translocation across a leaky intestine' into the portal circulation may be the mechanism by which intestinal permeability affects LFTs.

## 416 1% OF ADULTS HOMOZYGOUS FOR THE C282Y MUTATION OF THE HFE GENE HAVE BEEN CLINICALLY DIAGNOSED WITH IRON OVERLOAD: EVIDENCE FROM THE SOUTH WALES HAEMOCHROMATOSIS STUDY

C.A. McCune, M. Worwood, L.N. Al-Jader (introduced by A.B. Hawthorne). University of Wales College of Medicine, Cardiff, UK

Background: The clinical significance of HFE mutations remains uncertain with a large discrepancy between the frequency of the predisposing genotype and clinical disease. To our knowledge this is the first study to examine in detail the hospital burden of disease in the UK within a defined population area.

Aim: To establish accurately the number of patients treated for hereditary haemochromatosis (HH) in Bro Taf and Gwent Health Authorities within a 2-year period (Jan 1998- Dec 1999) and to compare this with the number of subjects homozygous for C282Y calculated from the genotype frequencies of 10,556 healthy blood donors from S Wales. In addition to determine the proportion with moderate to severe iron overload (≥4g iron).

Methods: Hospital patients were identified from: information obtained from PEDW/APC data using ICD10 codes; laboratory data and correspondence with all gastroenterology and haematology consultants.

Results: 81 patients were considered to have HH with varying iron phenotypes. 59 were confirmed C282Y homozygotes (see table). 34 (42%) had moderate to severe iron overload. In S Wales 1 in 147 blood donors are C282Y homozygous. We have calculated that only 1.1% of adult homozygotes have been diagnosed and treated for iron overload.

Abstract 416

Conclusions: Genetic screening would detect many thousands of healthy subjects in S Wales. Of the 1% likely to be diagnosed with HH less than 45% will have moderate to severe iron overload based on established criteria.

## 417 THE PREVALENCE OF PORTAL HYPERTENSION (PHT) IN PRIMARY BILIARY CIRRHOSIS (PBC)

D.E.J. Jones, R. Walter, M.I. Prince, M. Hudson. Centre for Liver Research, University of Newcastle, UK

Data from our patient cohort suggest that the development of PHT in PBC may be associated with a more adverse outcome than has previously been thought to be the case. Estimation of the overall extent of the morbidity associated with PHT in PBC has been hampered by a paucity of accurate data relating to its prevalence in the patient population. Previous studies (which have shown prevalences ranging from 23–75%) have been limited to case series subject to patient inclusion bias. In order to further quantify the scale of the problem posed by PHT in PBC we studied the prevalence of PBC in a comprehensive cohort of patients defined by geographical residence rather than specific clinic attendance.

A comprehensive and exhaustive case finding exercise was performed to identify all prevalent cases of PBC within the study area. At the census point 166 PBC patients were prevalent within the study area. The point prevalences of portal hypertension, endoscopically proven varices and a history of variceal haemorrhage were 25% (42/166), 8% (13/166) and 1% (2/166) respectively. 58/166 (35%) of the prevalent patients had histologically confirmed advanced disease (Scheuer stage III/IV) at the study point. The point prevalence of PHT in this diagnosed advanced disease subgroup was 52% (19/58).

In addition to the prevalent cases a further 146 deceased PBC patients were identified whose last residence was within the study area. The total number of patient years at risk was calculated for the whole patient cohort from the point of diagnosis until the development of the relevant complication (PHT, oesophageal varices or variceal haemorrhage), death , liver transplantation or the study end-point. The total number of at risk years for PHT, oesophageal varices and variceal haemorrhage development were 1710, 2106 and 2266 respectively. The incidence rates for PHT (129/312), oesophageal varices (69/312) and variceal haemorrhage (36/312) development were therefore 75/1000, 33/1000 and 16/1000 patient years at risk respectively.

Conclusions: PHT and its complications are common in patients with histologically advanced PBC. Given the adverse outcome seen with PHT in this disease screening is warranted.

## 418 PROGNOSTIC ACCURACY OF APACHE III SCORING SYSTEM IS GREATER THAN THAT OF THE CONVENTIONAL CHILD-PUGH'S SCORE IN PREDICTING SHORT-TERM HOSPITAL MORTALITY OF NON INTENSIVE CARE UNIT PATIENTS WITH LIVER CIRRHOSIS

C. Chatzikostas1, M. Roussomoustakaki1, G. Notas2, G.I. Vlachonikolis3, E. Matrella1, D. Samonakis1, E. Vardas1, E. Kouroumalis1,2.1Department of Gastroenterology, University Hospital, Heraklion, Crete, Greece;2Liver Research Laboratory and3Biostatistics Laboratory, University of Crete Medical School, Crete, Greece

Objective: The aim of this study was to assess the prognostic accuracy of Child-Pugh's score (CPS) and Acute Physiology, Age and Chronic Health Evaluation (APACHE) II and III scoring systems in predicting short-term, in-hospital mortality of patients with liver cirrhosis admitted to a gastroenterological medical ward.

Methods: 200 consecutive admissions of 147 cirrhotic patients (44% viral-associated liver cirrhosis, 33% alcoholic, 18.5% cryptogenic, 4.5% both viral and alcoholic) were studied prospectively. Clinical and laboratory data conforming to the Child-Pugh, APACHE II and APACHE III scores were recorded on day one for all patients. Statistical analysis for the prognostic variables was performed by using t-test, receiver operating characteristic (ROC) curves and area under a ROC curve (AUC), non-parametric Wilcoxon test and discriminant analysis.

Results: The in-hospital mortality was 11.5%. The mean CPS, APACHE II and III scores for survivors were found to be significantly lower than those of nonsurvivors. When ROC curves were plotted, no significant differences between Child-Pugh's (AUC, 0.85), APACHE II (AUC, 0.75), and APACHE III (AUC, 0.81) overall performances were noticed, however the overall correctness of prediction of APACHE III was 9% greater than that of the CPS (Wilcoxon test: z = 2.846, p = 0.004) and 11% greater than that of the APACHE II, namely, 87%, 78% and 76% respectively (cutoff values, 62,10 and 15 respectively).

Conclusions: All three scores were proven to be of value in risk stratifying patients with liver cirrhosis. Although the overall performance of APACHE III system as assessed by ROC curve analysis is no superior than that of the conventional Child-Pugh's score in predicting short term outcome of hospitalized patients with liver cirrhosis, APACHE III correctly statifies a significantly greater number of patients.

## 419 HYPERURICAEMIA, GOUT AND CARDIOVASCULAR RISK AFTER LIVER TRANSPLANTATION

D.A.J. Neal1, B.D.M. Tom2, A.E.S. Gimson1, P. Gibbs3, G.J.M. Alexander1.1Department of Medicine and3University Department of Surgery, Addenbrooke's NHS Trust, Cambridge, CB2 2QQ, UK;2Centre for Applied Medical Statistics, Department of Public Health and Primary Care, Cambridge CB2 2SR, UK

Background: Hyperuricaemia and gout are recognised complications of renal and cardiac transplantation. In contrast the development of hyperuricaemia following liver transplantation has received less attention. Elevated serum uric acid has been cited as an independent risk factor for cardiovascular disease in the general population. To evaluate the prevalence of hyperuricaemia and its association with cardiovascular risk factors we reviewed the case records of 134 consecutive liver transplant recipients with a mean follow up of 52 months (range 6 – 92 months).

Results: 47% had hyperuricaemia after liver transplant. Peak uric acid correlated significantly with corresponding serum creatinine (rs = 0.694). 6% of patients developed an acute episode of gout. Hypertension, hypercholesterolaemia and a body mass index > 25kg/m2 were present in 53, 46 and 48 % of hyperuricaemic patients respectively and in 47, 54 and 52% of patients with normal serum urate. None of these differences were significant. Cardiovascular events comprised 1 myocardial infarct and 1 incident angina, each patient having hyperuricaemia, and 2 strokes, one of which had hyperuricaemia.

Conclusions: There is an important association between liver transplantation and hyperuricaemia. Gout is a significant cause of morbidity but occurs less frequently than after renal or cardiac transplants. There was no association between hyperuricaemia and other cardiovascular risk factors. Too few cardiac events occurred to draw any conclusions about an association with uric acid.

## 420 OBSTETRIC CHOLESTASIS IN SOUTH WALES

C.L. Ch'ng1, M. Morgan2, J.G.C. Kingham1. Department of Gastroenterology1and Obstetrics2, Singleton Hospital, Swansea, Wales, UK

Aims: A prospective study of the incidence, clinical and biochemical features, management and outcome of obstetric cholestasis (OC) in a defined population in South Wales, UK.

Methods: All pregnancies in an obstetric unit serving 270,000 were screened for OC between March 1999 and June 2001. Diagnosis of OC was based on pruritus, abnormal liver tests and exclusion of other hepatobiliary diseases.

Results: 45 OC patients were identified among 8142 pregnancies – incidence 0.5%. Age ranged from 16 - 40 years (Median 30). There were seven twin and one triplet pregnancies. Sixteen were primiparae while 29 had had 1 to 6 previous pregnancies, ten of which were complicated by OC with two associated stillbirth. Three had family history of OC. All were symptomatic: pruritus in 45, vomiting in four, diarrhoea in four and severe malaise in two. Two patients suffered hyperemesis earlier in pregnancy, two pre-eclampsia and one HELLP syndrome. Symptoms started between 8 and 39 weeks gestation (median 34). Eighteen had proven urinary tract infection either just before or after diagnosis of OC. 43 had elevated AST (31–519; median 141 U/l). Serum bile acids raised in 38 of 41 tested (15–179; median 34 μmol/l). γGT was modestly elevated (42–292; median 71 U/l) in only 14 episodes while bilirubin was raised in 17 (16 - 34; median 20 μmol/l). Leucocytosis was seen in 28 patients. The interval from diagnosis of OC to delivery ranged from 1 to 142 days (median 8). Ursodeoxycholic acid was given to eleven patients with improvement in symptoms in six and biochemistry in ten. Induction of labour or Caesarean section was undertaken in 33 because of OC. There was no maternal or fetal death but 15 of 52 babies required admission to SCBU. Symptoms and abnormal liver function resolved rapidly in 75% within 2 weeks of delivery.

Conclusion: OC complicates 1 in 180 pregnancies in South Wales. It is characterised by elevated transaminases rather than cholestasis; jaundice is neither a necessary nor a common component. Combined medical and obstetric care with early delivery prevents fetal loss.

## 421 SURVEILLANCE FOR HEPATOCELLULAR CARCINOMA IN CIRRHOSIS: A NATIONAL AUDIT OF THE PRACTICE IN THE UK

C.W.Y. Lai, M. Wardle, V. Birkett, E. Powell, G.A.O. Thomas. Department of Gastroenterology, University Hospital of Wales, Cardiff, UK

Background: Hepatocellular carcinoma (HCC) is a significant cause of mortality in cirrhotic patients. Its detection at an early stage increases the chance of curative therapy. Clear evidence of survival benefit with surveillance is limited. However, anecdotally, many clinicians undertake some form of surveillance program.

Aim: To evaluate the current practice of HCC surveillance in cirrhosis in the UK.

Methods: 1080 postal questionnaires were sent to the members of the British Society of Gastroenterologists (BSG), excluding radiologists, pathologists and paediatricians.

Results: 525 replied (49%); of these, 120 did not look after adult cirrhotics and were excluded from analysis. Of the 405 remaining respondents, 296 (73%) surveyed for HCC, 123/296 had protocols. Hepatologists and those with a large cirrhotic practice were more likely to survey. 96/296 quoted an age limit (median 70) for surveillance. 107/296 (36%) surveyed all cirrhotics regardless of their suitability for curative therapy; in contrast, 127/296 surveyed only those suitable for liver transplantation or partial hepatectomy. 166/296 (56%) chose to survey all causes of cirrhosis, whereas a smaller group (83/296, [28%]) were more selective and surveyed those with Hepatitis B, C, haemachromatosis and alcohol liver disease. The commonest mode of surveillance was a combination of abdominal ultrasound and alpha fetoprotein, with a wide range of test intervals (3 to 24 months). In those choosing to survey, 130/296 believed it increased survival, while 95/296 felt that non-surveillance might leave them legally liable. 109/405 did not survey, 58 of who quoted the lack of evidence for survival benefit, and 46 the lack of guidelines, as the reasons for their practice.

Conclusions: Despite the lack of guidelines or clear evidence for benefit, a majority of clinicians who responded performed some form of surveillance. Practice was variable with a significant number being unselective in the types of patients to survey. This has obvious resource implications. Guidelines are needed to rationalise and clarify practice.

## 422 HYDROXYCHLOROQUINE REDUCES LIVER RELATED MORTALITY IN HEPATITIS C ASSOCIATED (HCV) COMPENSATED CIRRHOSIS

E.A. Kouroumalis, J. Moschandreas, G. Alexandrakis, E.Matrella. Department of Gastroenterology, University Hospital and Medical School Crete, Greece

Hydroxychloroquine, a lysosomotropic and macrophage modulating agent has been reported to improve liver biochemistry in chronic viral hepatitis.

Aim: To investigate the effect of hydroxychloroquine on survival of HCV active cirrhotic patients

Methods: 162 patients ( 31% male, 69% female) with compensated HCV cirrhosis were prospectively evaluated. All were Child-Pugh A or B at entry into the study. 52 with active cirrhosis (increased aminotranferases) were treated with hydroxychloroquine 200 mg tid for 6 months. A 3 month treatment was reinstituted if during follow up, aminotranferases were again high. 110 patients with inactive cirrhosis were the non treated controls. Time to decompensation and death were recorded. Patients were followed up between 3 and 136 months.

Results: Median time to decompensation was 81 months (95 % C, 45–117 months) and was not different between the two groups (p=0.7, Kaplan-Meier analysis). However, variceal bleeding, hepatic encephalopathy and spontaneous bacterial peritonitis were signs of decompensation only in the controls, ascites being the only decompensation sign in the treated group. Overall mortality, although lower in the treatment group just failed to reach significance (p< 0.07, 356 patients should be enrolled to reach significance). However, when only liver related deaths were assessed, the survival of the treated patients was significantly better (p< 0.05).

Conclusions: Hydroxychloroquine treatment reduces liver related mortality in active HCV cirrhosis, despite the fact that the control group were inactive cirrhotics where a more favourable prognosis is to be expected.

## 423 MEASUREMENT OF AFP IN PATIENTS WITH HEPATITIS C CIRRHOSIS DURING HEPATOMA SURVEILLANCE

T. Woodall, S. Sen, A.E. Gimson, G.J.M. Alexander, D.J. Lomas, S. Vowler. Department of Medicine, University of Cambridge, Hills Road, Cambridge CB2 2QQ, UK

Background: Patients with chronic hepatitis C virus ( HCV) infection and cirrhosis have an increased risk of hepatocellular carcinoma (HCC).Surveillance to identify early tumours based on a periodic estimation of serum alpha-fetoprotein (AFP) is unproven.

Methods: Surveillance was undertaken in 100 consecutive patients with chronic HCV and cirrhosis, based on a minimum of 6 monthly serum AFP and ultrasound (US), for a median of 2 years (range 0.5 – 6 ). 186 US and 662 AFP measurements were undertaken. 11 patients were lost to follow up. 40 (45%) had an elevated AFP and 18 had a suspicious US. 11 patients (12%) developed HCC during surveillance.

Results: Of 40 with an elevated AFP, 10 (25%) also had a suspicious US and HCC was identified in 7 of those (17.5%). The level of AFP however, was non-discriminatory. 30 of 40 (75%) had an elevated AFP with a clear US and HCC was identified in 1. The level of AFP was again non-discriminatory. Of 49 with a normal AFP, 8 (16 %) had a suspicious US and HCC was identified in 3 (6%). 22 CT Scans, 7 MR scans and 11 Angiograms were undertaken in patients who did not have HCC. As a predictive test for HCC, elevated AFP in isolation had a positive predictive value (PPV) of only 0.18, a sensitivity of 0.7 and a specificity of 0.58. In comparison, US had a PPV of 0.8, sensitivity of 0.91 and specificity of 0.89.

Conclusion: The incidence of HCC in this group of Hepatitis C cirrhotic patients is high. Surveillance is a considerable commitment but the combination of serum AFP and US has traditionally been used for the detection of HCC . However, an elevated AFP in isolation was common, had a poor predictivity for the detection of HCC and led to numerous (expensive) investigations. Restricting surveillance to US alone would be as effective.

## 424 COMBINATION PROPHYLAXIS WITH LAMIVUDINE (LAM) AND HEPATITIS B IMMUNOGLOBULIN (HBIG) PREVENTS GRAFT RE-INFECTION BY HEPATITIS B (HBV)

V. Lai, E. Elias, K. O'Donnell, D. Mutimer. Liver Unit, Queen Elizabeth Hospital, Birmingham, Edgbaston B15 2TH, UK

Prevention of graft re-infection remains the most important issue in transplantation of HBV infected patients. Re-infection by HBV is almost inevitable if no prophylaxis is given. The course of recurrent infection is aggressive causing sub-acute liver failure or rapid progression to cirrhosis. In the early 90's, prophylaxis comprised of regular and repeated administration of HBIg. Studies of LAM prophylaxis in this setting began in 1994. Both LAM and HBIg proved ineffective for patients with high serum HBV DNA titre.

Subsequently the combination of LAM and HBIg has been used and preliminary reports suggest the combination is effective and safe. The aim of this study is to review the QEH liver unit's experience with combination LAM/HBIg for prevention of HBV re-infection.

This is a retrospective review of the clinical and virological outcome of these patients. Since introduction of LAM/HBIg prophylaxis in 1997, 16 HBV patients have undergone liver transplantation (15Male, median age 51). Pretreatment serum HBV DNA titres range from <400 copies/ml to >4 x 107. 3 patients died (12 days, 6 weeks and 8 month post transplant). Median follow up of survivors is 2 years (6months- 4 years). All survivors are serum HBsAg and HBV DNA (Roche PCR-based assay) negative.

Conclusion: LAM/HBIg prevents HBV re-infection even in high-risk patients.

## 425 DISTRIBUTION OF IL-6 IN LIVER CIRRHOSIS

N.A. Mohammed, S. Abd El-Aleem, H. Abdel Hafiz, R.F.T. McMahon. Department of Histopathology, Clinical Sciences Building, Manchester Royal Infirmary, Oxford Road, Manchester M13 9WL, UK

Interleukin-6 (IL-6) has been proposed as one of the main inflammatory mediators. It is a major mediator of the acute phase response in infectious diseases and inflammatory processes inducing fever, leukocytosis, and increase in the hepatic synthesis of acute phase proteins. In this study, sera and liver biopsies from fifteen patients with clinically and pathologically diagnosed liver cirrhosis were taken. In addition sera from 7 and liver biopsies from 3 healthy controls were used. Serum levels of IL-6 were measured using ELISA kits and the cellular distribution was investigated using immunohistochemistry. We have shown that the serum IL-6 levels in cirrhotic patients (25.46 ± 11.6 pg/ml) were significantly (P <0.05) increased by comparison with the control group (12.14 ± 6.6 pg/ml). Immunohistochemically, in the control group, IL-6 was seen only in occasional sinusoidal cells. However it was widely distributed in the cirrhotic liver. In the latter, it was mostly seen in the inflammatory cells infiltrating the liver but also expressed in the sinusoidal cells, Kupffer cells, vascular endothelial lining cells and hepatocytes. Upregulation of IL-6 in cirrhotic patients could be due to active synthesis or defective clearance by non-functioning hepatocytes. However our findings suggest that both mechanisms are operating since we have shown high expression in inflammatory cells which could be due to oversynthesis and high expression in hepatocytes which could be due to accumulation in non-functioning cells. It is therefore clear that IL-6 showed systemic and local augmentation in cirrhotic patients and is mainly produced by inflammatory cells. Taken together these findings suggest that IL-6 production in liver cirrhosis is dependent on the inflammatory stage and the local production of IL-6 could contribute to the inflammation, fibrosis and immunological responses in the cirrhotic liver. Moreover the characteristic distribution of IL-6 in cirrhotic lobules could implicate it in the development of cirrhosis. In the near future, the appropriate manipulation of IL-6 may provide a novel strategy for the treatment of patients with liver cirrhosis or at least improve the fate of cirrhosis.

## 426 QUANTITATIVE STUDIES OF LIVER ATROPHY FOLLOWING PORTACAVAL SHUNT IN RATS

A.M. Zaitoun1, G. Apelqvist2, H. Al-Mardini3, T. Gray1, F. Bengtsson2, C.O. Record3.1Department of Histopathology, Queen's Medical Centre, University Hospital, Nottingham NG7 2UH, UK;2Department of Clinical Pharmacology, Lund University, S-221 85 Lund, Sweden;3Department of Medicine, The Royal Victoria Infirmary, Newcastle-upon-Tyne NE1 4LP, UK

Background: To evaluate the morphological changes in the liver that occur in portal-systemic encephalopathy after portacaval shunting (PCS).

Materials and Methods: Male rats underwent either PCS (n=35) or control sham operations (n=31). Blood samples were taken prior to sacrifice (weeks5–7) for measurement of hormone concentrations. Morphological assessments were carried out on 5μm thick haematoxylin and eosin stained sections and digital electron micrographs using the Prodit 5.2 image analysis system.

Results: There was a significant reduction in the liver mass of PCS rats (4.98v13.65g). Testosterone was significantly reduced (2.49v10.1ng/ml;p=0.002) and oestrogen significantly increased (77.9v51.6ng/ml;p=0.0004). Morphometric analysis showed significant reductions in the average distance between peri and postsinusoidal vessels (298v499□B%m) in PCS rats while in zone 3 the mean area of cytoplasm was also reduced (74.3v112.9□B%m2). Within the PCS group there was a significant reduction in the mean area of hepatocyte nuclei from 55.52□B%m2 in zone 1 to 45.31□B%m2 in zone 3 and a marked difference in the mean area of cytoplasm (113.25□B%m2 in zone 1 to 74.30□B%m2 in zone 3). Electron microscopy revealed reduction in the cytoplasmic organelles in PCS rats in comparison with sham rats. Apoptosis was increased in zone 3 to 2.4v0.3 counts per mm2 in PCS rats (P=0.00000) while mitosis was significantly increased in zone 1.

Conclusion: This study shows that liver atrophy after portacaval shunting has a complex aetiology. The microcirculatory disturbances and hormonal changes after portacaval shunting induce apoptosis that further contributes to liver atrophy in this animal model.

## 427 ROLE OF TECHNICAL VARIABLES IN EARLY SHUNT INSUFFICIENCY FOLLOWING TIPSS FOR THE TREATMENT OF PORTAL HYPERTENSION

S. Balata1, H.F. Lui1, A. Helmy1, D. Tripathi1, D.N. Redhead2, P.C. Hayes1.1Liver Unit, Department of Medicine, Royal Infirmary of Edinburgh, Lauriston Place, Edinburgh EH3 9YW, Scotland, UK;2Department of Radiology, Royal Infirmary of Edinburgh, Lauriston Place, Edinburgh EH3 9YW, Scotland, UK

Summary: Transjugular intrahepatic portosystemic stent shunt (TIPSS) has increasingly been used for the treatment of complications of portal hypertension. This study evaluated the technical predictors of early shunt insufficiency after TIPSS placement in patients with advanced liver disease.

A retrospective analysis of 399 patients undergoing TIPSS over a period of 9 years between July 1991-July 2000 was carried out. The indications for TIPSS were bleeding oesophageal varices in 288 patients, gastric varices in 52, refractory ascites in 42, portal hypertensive gastropathy in 8 and ectopic varices in 9 patients. TIPSS placement was successful in 383 (96.0%) patients, and failed in 16 (4.0%). The shunt remained patent without assistance in 40 (10.0 %) (primary patency) and 44 (11.0%) patients developed early shunt insufficiency within 30 days from TIPSS insertion. The univariate association of 21 prognostic clinical and technical variables between patients with early shunt insufficiency and primary patency was tested by the Chi-squared or Wilcoxon rank-sum tests. Multiple logistic regression analysis was utilized to determine the relationship of multiple technical and clinical variables in predicting early shunt outcome.

The two groups were comparable and representative of the whole TIPSS cohort patients. Of the 21 technical and clinical variables, stent diameter, distance of shunt from IVC, duration of the procedure, and portal pressure gradients post TIPSS were independent predictors of early shunt insufficiency.

Based on our analysis, four technical variables at index TIPSS can reliably predict early shunt insufficiency. Total protection from early complications of TIPSS requires awareness of the risk factors and a low threshold maintained for early shunt revision.

## 428 CHARACTERISATION OF LIVER INFILTRATING T-CELL ANTIGEN SPECIFICITY IN A MOUSE MODEL OF PRIMARY BILIARY CIRRHOSIS (PBC)

A.J. Robe, J.M. Palmer, J.A. Kirby, M.F. Bassendine, D.E.J. Jones. Centre for Liver Research, University of Newcastle, UK

PBC is characterised histologically by damage to the intra-hepatic bile ducts accompanied by a T-cell rich portal tract mononuclear cell (MNC) infiltrate. PBC is characterised immunologically by autoreactive antibody and T-cell responses to the self-antigen pyruvate dehydrogenase complex (PDC). Human studies suggest that portal tract T-cells seen in PBC liver are specific for PDC, implicating autoreactive T-cell responses directed at this antigen in disease pathogenesis. We have recently demonstrated that SJL/J mice can be induced to break down both B-cell and splenic T-cell tolerance to self-PDC, and that this tolerance breakdown is associated with the development of portal tract inflammation and bile duct damage. In this study we set out to characterise the antigen specificity of the infiltrating portal tract T-cells in this model.

Female SJL/J mice of 10–12 weeks were sensitised (n=7) with a mixture of murine (m) and bovine (b) PDC (study group). Control mice (n=4) received mPDC only. Mice were sacrificed at 8 weeks post-sensitisation. The spleen was removed and splenic T-cell responses characterised by FACS and primary proliferation. Livers were perfused in situ with collagenase via the portal vein prior to removal. The liver infiltrating mononuclear cell population was isolated by further collagenase digestion and density centrifugation. T-cell responses were characterised as for splenic T-cells.

As before, splenic T-cells from the study group, but not the controls showed T-cell reactivity with mPDC. Liver T-cells from the study group, but not controls showed a proliferative response to mPDC (study group: mean incorporated counts 6943±2831 v background 5616±2408,p<0.01; control 6603±2290 v back 6628±2035,p=ns). The liver mononuclear cell infiltrate in the study group was significantly expanded (6.5x106 v controls 2.7x106;p<0.01). and skewed towards to the CD8+ subset.

Conclusions: The anti-PDC liver T-cell auto-reactivity seen in this model closely resembles that seen in human PBC, further confirming the potential value of this model for the study of PBC pathogenesis.

## 429 FOCAL LIVER LESIONS: TO BIOPSY OR TO IMAGE?

B.S.F. Stacey, D.R. Fine. Southampton General Hospital, Tremona Rd, Southampton SO16 6YD, UK

Background: There is much debate about whether a focal liver lesion should be biopsied or only imaged, given the risk of tumour seeding in the biopsy track. There is however only anecdotal evidence coming from case reports and small series to support this view. As imaging methods improve many authors feel that a confident diagnosis need no longer rely on tissue. Practice varies between doctors and strong opinions are held.

Methods: A questionnaire was sent to 73 consultants and SpRs around the region. Four hypothetical cases were given and participants asked to offer a diagnosis, preferred next investigation and whether or not they would biopsy the lesion. The cases included an asymptomatic 60 year old man with a solitary liver lesion, an elderly lady with an apparent colonic liver metastasis, a young woman with a likely pill related adenoma and a middle aged man with a clear hepatocellular carcinoma.

Results: 38 (52%) of those canvassed responded. Several diagnoses were offered for the first case (benign and malignant) but there was more agreement on the initial investigation with CT +/- contrast (83%). There was no agreement on whether to biopsy with 46% saying no'. All respondents gave the correct diagnosis for the case of metastasis and 77% would use CT as the next test. However, only 54% would not biopsy and this figure was significantly higher in teaching hospitals (90% v 40%). Most respondents thought the third case was an adenoma or focal nodular hyperplasia and CT or MRI were the investigations of choice (89%). About half would consider biopsy but this figure was much higher in teaching hospitals (80% v 40%). All participants identified the HCC but there was a large range of next investigations including CT, MRI, biopsy, cytology and angiography. 57% would not biopsy but this figure was higher in teaching hospitals (90% v 44%).

Conclusions: There is a wide discrepancy in biopsy policy between teaching hospital and DGH's but also between specialities. Significantly, 46% would consider biopsying a metastasis and 43% would consider biopsying an HCC. Without a Grade A evidence base this highlights the need for a large scale randomized controlled trail into management of focal liver lesions with and without biopsy.

## 430 MR-GUIDED LASER THERMAL ABLATION OF PRIMARY AND SECONDARY LIVER TUMOURS

E.A. Dick, R. Joarder, M. de Jode, S. Taylor-Robinson, H. Thomas, G. Foster, W.M.W. Gedroyc. St Mary's Hospital, London W2 1NY, UK

Purpose: To test the hypothesis that MR guided hepatic tumour ablation is (i)safe & feasible and (ii) improves patient survival and (iii) decreases viable tumour voalume.

Methods and Materials: 125 MR guided Laser Thermal Ablations (LTA) have been performed on 40 patients (9 females, 31 males, average age 59.1 years) between 1997 and 2001. The liver tumours include Hepatocellular Carcinoma (HCC, n=19), metastases (n=11, mainly colorectal), carcinoid (n=5) and two benign liver tumours. 3 patients were excluded from follow-up.

Results: Mean survival for all patients was 15.2 months, with an adjusted mean survival of 16 months for HCCs and 15.2 months for metastases. There were three major and five minor post-procedural complications but no deaths. An average of 57% of tumour was ablated as assessed by per-procedural thermal mapping, with an average of 49.4% of tumour ablated assessed by pre and post ablation gadolinium-enhanced MRIs. Average tumour size was unchanged after ablation. In patients with multiple liver tumours ablated tumours grew significantly less than untreated tumours over the same time period (108% compared to 196% growth over an average follow up period of 5.8 months).

Conclusions: MR guided laser thermal ablation of primary and secondary liver tumours is safe and feasible and produces a better survival in patients with HCC than would be expected in untreated patients, as well as a mean survival in patients with metastases at least equal to the longest median survival in untreated patients. Percentage viable tumour was decreased by a mean of 49.4% per LTA session.

## 431 INTRAHEPATIC CHOLESTASIS OF PREGNANCY: PATTERN OF PRESENTATION AND RESPONSE TO URSODEOXYCHOLIC ACID

E. Jwarah1, J.P. Dwyer1, D. Pring1, A.J. Turnbull, S.M. Kelly.1Dept of Obstetrics & Gynaecology and Dept of Gastroenterology, York District Hospital, Wigginton Rd, York, UK

Intraheaptic cholestasis of pregnancy (IHCP) can adversely affect maternal well-being and foetal outcomes. Early identification, monitoring and prompt delivery is ideal. We report our experience of IHCP with respect to presentation, pattern of LFT abnormalities, maternal and foetal outcomes and the effect of treatment with ursodeoxycholic acid (UDCA).

Results: Over a period of 18 months 29 cases of IHCP were identified, complete data was available from 18 patients. All cases presented with pruritis and were found to have abnormal LFTs. Other causes of liver disease were excluded. Symptoms started at a mean of 33.5 weeks (range 26–38). Serum bile acids were clearly elevated in 14 (mean 62 umol/L, normal <14) and borderline in the other 4 (mean 12umol/L). Only one patient became jaundiced (bilirubin 64 umol/L). At presentation serum alkaline phosphatase ranged from 114 – 456 IU/L (mean 255) and ALT 13–468 IU/L (mean 140). Of note in 10 patients the ALT was particularly high at over 100, mean 220 IU/L (range 121–468). 14 patients were treated with UDCA with a subsequent marked improvement in LFTs, bile acids and symptoms. In 10 patients the pregnancy was actively managed with induction at 38 weeks. 6 underwent a spontaneous labour (33–39 weeks) and two had caesarean sections for obstetric reasons. 2 babies were jaundiced at delivery and one went to SCBU. Both made a full recovery. There was no maternal morbidity.

Conclusions: IHCP presents with pruritis and abnormal LFTs. It is readily recognisable and responds well to UDCA with good maternal and foetal outcomes. Of particular note, the ALT is often markedly elevated, a feature that is not widely recognised and which may represent a potential source for diagnostic confusion.

## 432 AN AUDIT OF SPECIALITY MANAGEMENT OF ALCOHOLIC LIVER DISEASE WITH JAUNDICE

E.H. Forrest (introduced by K. Cochran). Victoria Infirmary, Langside Road, Glasgow G42 9TY, UK

The development of jaundice in the alcohol abuser may indicate acute alcoholic hepatitis (AAH) or the progression to end-stage chronic alcoholic liver disease (ALD). These patients may be cared for by either General Medicine (GM) or Gastroenterology (GI) physicians. This audit aimed to identify if there were differences in the treatment and outcome of patients managed in these different contexts.

Methods: Patients with ALD and serum Bilirubin > 80 μmol/l on admission were identified through discharge coding over a period of 27 months (June 1999 – August 2001). Only patients drinking to excess until the two weeks prior to admission were included. Differences in management were identified in the use of corticosteroids (CS), broad-spectrum antibiotics (AB), and nutritional support (N). GI care assumed the patient's ongoing care for 55% or more of the admission episode.

Results: 79 patients were included in the study: 46 GI, 33 GM. The mean age was 51.6 ± 1.2 years with mean admission serum bilirubin of 230.4 ± 18.4 μmol/l, prothrombin time ratio of 1.57 ± 0.05, and Discriminant Function of 51.4 ± 3.6. The median time to GI review of a patient initially admitted under GM was 3 days (range 0 – 12). Patients managed by GI had a longer median hospital stay (18 [2–87] vs. 10 [2–89] days; p=0.001). Differences in management and 30-day mortality are shown in the table.

Abstract 432

Conclusion: Significant differences may exist between GI and GM physicians in the management of ALD patients with jaundice. More aggressive medical therapy may translate into an improved outcome. Speciality triage of these patients should be encouraged.

## 433 LIVER BIOPSY IN HEPATITIS C: HAVE WE FOLLOWED NICE GUIDELINES?

D.A.J. Lloyd, S. Subramanian, A. Poullis, C.J. Tibbs, J.D. Maxwell. St George's Hospital, London SW17 0QT, UK

Aim: To audit whether patients with hepatitis C undergoing liver biopsy were subsequently treated according to current NICE guidelines.

Methods: Patients with treatment naive RNA PCR +ve hepatitis C undergoing liver biopsy between 1997 and 2000 inclusive were identified using histology and appointment records. Histology was used to categorize the degree of liver disease into mild, moderate and cirrhotic based on a modified HAI scoring (Ishak et al. 1995) and recent BSG management guidelines (Booth et al. 2001). Patient notes were reviewed to assess whether patients received anti-viral treatment by 31st October 2001.

Results: Between 1997 and 2000 113 patients with treatment naive RNA PCR +ve hepatitis C underwent liver biopsy; this constituted 41% of all liver biopsies performed. 38 patients had mild liver disease, 65 had moderate liver disease and 10 had cirrhosis. 8 of the patients with mild liver disease (21%) initiated anti-viral therapy, 3 as part of a UK based clinical trial. 20 of the patients with moderate liver disease (31%) initiated anti-viral treatment and 23 patients (35%) were still awaiting treatment on 31st October 2001; treatment was medically contraindicated in 5 patients, withheld in 7 patients, declined by 4 patients and 6 patients were lost to follow-up. 2 patients with cirrhosis initiated anti-viral treatment and 1 patient was awaiting treatment on 31st October 2001; treatment was contraindicated in 4 patients, withheld in 1 patient and 2 patients were lost to follow-up. The median time from liver biopsy to non-trial treatment was 5 months (range 1 to 40 months). The median length of wait in those patients still awaiting anti-viral therapy was 21 months (range 9 to 51 months).

Conclusions: Of the 113 patients who underwent liver biopsy during the study period 66 (58%) would be considered eligible for treatment with combination therapy according to current NICE guidelines. Although 46 of these patients were offered treatment only 22 had started by the end of the follow-up period. Lack of funding was the principal explanation for delay to treatment. Proposed government support to implement NICE guidelines should reduce this delay.

## 434 THE COMBINATION OF PEGYLATED INTERFERON (PEG-IFN) AND RIBAVIRIN (RIBA) IN THE TREATMENT OF CHRONIC HEPATITIS C (HCV) INFECTION: PRELIMINARY REPORT OF A SINGLE CENTRE EXPERIENCE

G. Kanagasabai, M. Heydtmann, K. O'Donnell, R. Harrison, D. Mutimer. Queen Elizabeth Hospital Liver and Hepatobiliary Unit, and Birmingham University, Edgbaston, Birmingham B15 2TH, UK

Background: The HCV prevalence in the United Kingdom may be as high as 1%. Patients with chronic HCV infection have been treated with antiviral therapy at the QE Liver Unit. Since September 2000, antiviral treatment comprised pegylated interferon (PEG-IFN, 1–1.5 μg/kg weekly) and ribavirin (RIBA, 1–1.2 g/day). Treatment was intended for 6 months (non genotype 1 and non-cirrhotic patients) or 12 months.

Patients: At present, 79 patients (median age 44; 52 male; 56 non-cirrhotic, 13 cirrhotic) had commenced treatment. Genotype distribution was 34 type 1, 38 type 2 / 3, 2 type 4,1 type 5, 1 type 6. Median pretreatment titre was 6.7x105 copies/ml.

Results: 14 patients were withdrawn from therapy after a median of 16.5 weeks. Indications were: intolerance of symptoms (7), neutropenia (3), psoriasis (1), unstable angina (1), HBV co-infection(1) and thyrotoxicosis (1). At time of analysis 53, 35 and 3 patients completed 3, 6 and 12 months of treatment respectively. Biochemical response (normal ALT) was observed for 32/48 at 3 months and 21/32 patients at 6 months. Of those completing 6 months treatment, HCV RNA negativity was achieved for 22/24 (92%), including 4 cirrhotics with genotype 1 (n=1), 2 (n=1) and 3 (n=2). 25 patients required reduction of PEG-IFN and 16 of RIBA. Mean pre-treatment / 3 month values for haemoglobin and neutrophils were 14.8 / 12.1 g/dl and 3.7 / 1.9 x 109/l.

Conclusions: Combination treatment with PEG-IFN and RIBA achieves biochemical response and early virological clearance for a majority of treated patients. However, side effects may be frequent and severe, and dose reduction of PEG-IFN and/or RIBA is often required.

## 435 EVOLUTION OF HEPATITIS B VIRUS DURING PRIMARY INFECTION

S.A. Whalley, D. Brown, G.J.M. Webster, R. Jacobs, C.G. Teo1, V.C. Emery, G.M. Dusheiko. Royal Free Hospital;1Central Public Health Laboratory, London, UK

To study HBV evolution in vivo, we evaluated over 1MB-pairs of genetic sequences from PCR-clones obtained prospectively in six patients with acute resolving hepatitis and four patients who progressed to chronic infection. We analysed two genetic loci (3 ORF: core, surface, and polymerase). 93 mutations were observed, consistent with a relatively stable genetic population throughout infection. The majority of mutations were non-synonymous (average for all regions 81%). Patients with acute resolving hepatitis had a similar frequency of non-synonymous mutations within the core (2 x 10–4/codon) but a higher frequency of surface/polymerase mutations (1.8 x 10–4/codon) when compared to the chronic group (1.5 x 10–4/codon and 0.65 x 10–4/codon, respectively; p=0.01). Genetic diversity was quantified by the mean Hamming distances (Hd) and the number of strains (Ns) present within a serum sample. Overall, the core region had the greatest genetic diversity when compared to the surface or polymerase regions. A rapid expansion in Hd and Ns occurred in the core region between the earliest time during acute infection and the peak viral load and then a subsequent contraction in genetic diversity occurred as viral load declined. Patients with acute resolving HBV consistently exhibited a higher genetic diversity in all three genetic loci and an increased mutation frequency in surface/polymerase region compared to chronic infection. We speculate that these emerging variants contribute to the evolution of increased virulence, as indicated by a rise in serum LFT's, due to the increased immunological burden they present to the host.

## 436 ALBUMIN DIALYSIS (MARS DEVICE) RELIEVES CHOLESTATIC PRURITUS IN A DRAMATIC BUT NON-SUSTAINED MANNER

N. Zakaria, J. Wendon, D. Creamer, N. Heaton, J. Devlin.King's College Hospital, London, UK

Background: Intractable pruritus, a debilitating symptom related to intrahepatic cholestasis, can seriously impair quality of life. The efficacy and tolerability of presently available remedies is limited and unpredictable. Case reports have shown that plasmapheresis, and charcoal haemoperfusion are of some benefit. Molecular Adsorbent Re-circulating System [MARS] is a novel dialysis technology which efficiently removes albumin bound substances (including bile salts), with promising preliminary results in the management of decompensated liver disease. In this present study, we hypothesized that Albumin dialysis (MARS) would relieve intractable, drug-resistant cholestatic pruritus.

Patients and Methods: Five patients (2M:3F) with intractable pruritus were studied. The underlying conditions were PSC, chronic rejection, PBC, TPN-cholestasis and drug induced cholestasis in a donor liver, respectively. All patients had received standard and second-line anti-pruritic medication without symptomatic relief. Quality of Life Scores (QOL) scores [SF-12] and pruritus scores were calculated in each patient before and after MARS treatment. The four patients underwent a median of 24 hours (range 16–48 hrs) of albumin dialysis over a 2 to 45 day period.

Results: Pruritus resolved completely in 4/5 patients. QOL scores improved in 4/5 patients at the end of MARS treatment. Pruritus recurred in three patients at a median of 10 days (range 2–28 d). No direct relationship between changes in serum bilirubin and the interval of remission was evident. Presently, one patient underwent liver transplantation, one was removed from the transplant list due to marked improvement in clinical condition, one patient died after 6 weeks from variceal haemorrhage and two patients remain completely asymptomatic and are no longer being considered for transplantation.

Conclusion: Albumin dialysis (MARS) is a novel therapeutic tool which offers safe and effective, but non sustained, relief of pruritus in patients with intrahepatic cholestasis resistant to conventional therapy. The place of this technique in the long-term management of pruritus needs to be further defined.

## 437 ABNORMAL VASCULAR FUNCTION IN HEREDITARY HAEMOCHROMATOSIS: INVESTIGATING THE “IRON HYPOTHESIS”

I.S. Cadden, B.M. Mullan, M.E.Callender, D.R.McCance, I.S.Young. Royal Victoria Hospital, Belfast, Northern Ireland

Background: Iron excess has been linked with development of coronary heart disease. Haemochromatosis is a common inherited disorder of iron overload. Studies have shown carriage of the mutation for haemochromatosis is linked to cardiovascular risk. We used the non-invasive technique of Pulse Wave Analysis (PWA) to study the arterial stiffness in patients with haemochromatosis. Using applanation tonometery, the radial artery pressure waveform is recorded. Transfer functions are then applied producing the central aortic waveform and deriving the central aortic pressure. Augmentation of this pressure thus provides a measurement of the compliance of the vascular tree, which can be expressed quantitatively as the augmentation index (AIx).

Methods: The AIx of patients with Haemochromatosis was recorded as a measure of the arterial stiffness following a 10hr fast. Age- and sex-matched healthy individuals served as controls.

Results: 10 subjects were recruited to each group (7M:3F). There was no significant difference between the ages of the haemochromatosis and control subjects, 56.6yr V 54.6yr (p=0.62). The AIx was significantly higher in the haemochromatosis group compared to controls (mean=30.1%, range=24%-38%; mean=20.5%, range=11%-31%: P=0.001).

Conclusions: These results suggest, using a non-invasive, in vivo technique that there is abnormal vascular stiffness in haemochromatosis The tendency to excess iron stores may increase oxidative burden on the vascular endothelium causing injury. The resultant endothelial dysfunction may cause the increased cardiovascular risk which studies have suggested is associated with this condition.

## 438 THE IMPACT OF PORTAL HYPERTENSION IN PRIMARY BILIARY CIRRHOSIS (PBC)

D.E.J. Jones, R. Walter, M.I. Prince, M. Hudson. Centre for Liver Research, University of Newcastle, UK

Inter-patient variability in the rate of progression of the chronic liver disease PBC hampers the identification of high risk individuals suitable for invasive therapies such as liver transplantation. In this retrospective study we examined the role played by portal hypertension (PHT) and the presence of oesophageal varices (OV) in disease outcome in a cohort of 438 PBC patients followed up from diagnosis in a single centre. Median follow-up was 72 months with 262 subjects (60%) still alive at the study point. Survival from diagnosis of disease to death or transplant was significantly worse in patients developing PHT (Kaplan Meier analysis (KMA) p<0.001; 5 year survival (5ys) 68% v 87% for patients free of PHT at the study point). PHT development was not acting simply as a surrogate marker for development of histologically advanced disease as PHT retained its adverse prognostic value when analysis was restricted to patients with histologically advanced disease (Scheuer stage III/IV; KMA p <0.001; 5 ys 66% for stage III/IV disease with PHT v 81% for stage III/IV disease without PHT). Survival was significantly worse in PHT patients developing OV than in non-OV PHT patients (KMA p<0.001; 5ys 63% v 75%). In fact, stage III/IV disease patients with PHT but not OV had similar survival to stage III/IV patients not developing PHT. The adverse outcome associated with OV development did not result simply from the consequences of variceal bleeding as survival following PBC diagnosis was the same in patients developing OV with and without bleeding on follow-up (KMA p=ns; 5ys following PBC diagnosis 62% v 63%, 5ys following diagnosis of varices 25% v25%).

In the 159 patients developing OV (93 of whom bled), the one year transplant free survival following OV diagnosis was 63%. Excluding the 34 patients undergoing transplantation (leaving death as the study end-point) 1ys was 68%. In the 67 patients developing OV below the age of 65 who did not undergo transplantation 1ys was 71%.

Conclusion: The development of portal hypertension is associated with an adverse outcome in PBC. The risk maps to patients developing varices regardless of bleeding. The one year survival following development of varices is less than that currently being reported for transplanted PBC.

## 439 NO ASSOCIATION BETWEEN NOD2 POLYMORPHISMS AND SUSCEPTIBILITY TO, OR PROGRESSION OF, PRIMARY SCLEROSING CHOLANGITIS

S.N. Cullen, T. Ahmad, R.W. Chapman, D.P. Jewell. Gastroenterology Unit, University of Oxford, UK

Background: Nod2 is a member of a protein family which regulates apoptosis and NF-κB activation. Nod2 polymorphisms have recently been found to confer susceptibility to Crohn's disease (CD) either by altering the recognition of components of microbial pathogens and/or via the activation of NF-κB. NF-κB plays a crucial role in the activation of the hepatic stellate cell which is the first step in the development of liver damage and fibrosis. Primary sclerosing cholangitis (PSC) is a chronic cholestatic liver disease closely associated with inflammatory bowel disease (IBD), and is characterised by concentric obliterative fibrosis and bile duct strictures. This study examined the effects of 6 Nod2 polymorphisms on susceptibility to, and progression of PSC.

Method: DNA was extracted from 83 patients with well-documented PSC and 349 control patients. 65 of the PSC patients had ulcerative colitis (UC), 6 had CD, and 12 had no associated IBD. Primers were designed to examine 6 polymorphisms in the NOD2 gene using an SSP-PCR method. PSC and control patients were compared using 2x2 contigency tables and a χ2 test with Yates correction.

Results: None of the polymorphisms in the NOD2 gene was significantly associated with susceptibility to PSC. Further analyses to determine whether NOD2 polymorphisms might contribute to the development of cirrhosis or need for transplantation, were also negative. See table.

Abstract 439

Conclusions: Polymorphisms in the Nod2 gene are not significant factors in determining the susceptibility of individuals to developing primary sclerosing cholangitis. In addition, despite the role of the Nod2 gene in NF-κB activation, these polymorphisms do not determine the rate of progression or outcome of the disease.

## 440 PEAK PROPORTION AREA CHANGE (PPAC): A NOVEL METHOD OF THRESHOLD DETERMINATION FOR DIGITAL IMAGE FIBROSIS AREA ESTIMATION ON LIVER BIOPSIES

M. Wright, R. Goldin, H. Thomas, M. Thursz. Hepatology Section, Faculty of Medicine, Imperial College, London W2 1NY, UK

Introduction: Digital image analysis allows quantitative assessment of fibrosis on liver biopsy. Accurate determination of a threshold greyscale level representing fibrous tissue is critical. We present and describe a novel and objective method of determining threshold and compare it with the standard interactive method.

Methods: Digital images of picro sirius stained liver biopsy sections were captured by microscopy and converted to greyscale. Pixel counts at 256 grey levels were measured. Differences between each grey level were calculated and corrected for the number of pixels remaining. Grey level corresponding to maximum change in pixel count was chosen as threshold at which fibrosis area was calculated. Reproducibility was tested in comparison with interactive thresholding by repeating the process 4 times with 30 biopsies for each method.

Results: For the PPAC method correlations of R>0.7, p<0.001 were obtained between each repeated measurement, no significant difference was observed with repeated measures ANOVA (p=0.73), intra-class correlation co-efficient 0.96. For interactive thresholding results were less impressive with correlations (R=0.19-0.64) not all reaching significance, repeated measures ANOVA (p=0.076), intra-class correlation co-efficient 0.77.

Conclusions: The Peak Proportion area change (PPAC) method, described here, for determining threshold in image analysis is objective, reproducible and superior to standard interactive thresholding.

## 441 ABNORMAL LIVER FUNCTION TESTS FOLLOWING BONE MARROW TRANSPLANTATION: WHAT IS THE AETIOLOGY AND THE ROLE OF LIVER BIOPSY?

G.T. Ho, A. Parker1, J.F. MacKenzie, A.J. Morris, A.J. Stanley. Dept of Gastroenterology, Glasgow Royal Infirmary;1Bone Marrow Transplant Unit, Glasgow Royal Infirmary, UK

Introduction: Liver dysfunction is common in bone marrow transplant (BMT) recipients. Common causes are drugs, graft versus host disease (GVHD), infection and iron overload. We studied the prevalence of liver abnormalities, the causes and the use of liver biopsy to aid clinical management in these patients.

Methods: All the allogeneic and autologous BMTs undertaken in our institution between Jan 1997 and December 1998 were studied. Subsequent liver function tests, the use and indication of liver biopsy and the final cause of liver dysfunction were determined in each case.

Results: 121 patients (63 autologous) with BMT were studied. Abnormal LFTs were found in 71% allogeneic and 33% autologous BMTs. Final diagnoses were made without resorting to liver biopsy and these are shown in the table. Liver biopsy was required in 18 allogeneic and only 1 autologous BMT. 63% biopsies were undertaken greater than 100 days post BMT for persistently abnormal LFTs to assess possible GVHD. 16 out of 18 biopsies revealed significant iron overload with only one confirming GVHD. No fibrosis/cirrhosis was found on biopsy. No adverse effects occurred as a result of liver biopsy. In the patients with histological evidence of iron overload, the serum ferritin was also persistently elevated.

Abstract 441 Table showing cause of deranged LFTs

Conclusion: Liver biopsy was required in a minority of patients with abnormal LFTs post BMT, with most causes of liver dysfunction diagnosed clinically. Liver biopsy was most commonly undertaken late post BMT in the group with persistently abnormal LFTs in which chronic GVHD was suspected. However, iron overload was the commonest finding in this subgroup; and serum markers of iron excess could determine this.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.