Hepatology Clinical.

Abstract

Journal of Gastroenterology and HepatologyVolume 31, Issue S2 p. 89-120 Supplement ArticleFree Access Hepatology Clinical First published: 05 October 2016 https://doi.org/10.1111/jgh.13520AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Prevalence and predictors of indeterminate IFN-γ release assays for latent TB at liver transplant assessment S Anwar1, RJ Woodman2 and AJ Wigg1,2 1Hepatology Unit, Flinders Medical Centre; 2School of Medicine, Flinders University, Adelaide Background: Interferon gamma release assay (IGRA) is the most widely used test for screening of latent tuberculosis (TB). However, indeterminate IGRA tests may occur, reducing the utility of the test, and have been associated with impaired immune status. The performance of the IGRA in patients with end stage liver disease, a disease state with known, multifaceted immune dysfunction, has not been well studied. The aim of this study was therefore to evaluate the prevalence and predictors of indeterminate IGRAs in patients with end stage liver disease being assessed for liver transplantation. Methods: Prospective study of 49 consecutive patients undergoing liver transplantation assessment who underwent IGRA (Quantiferon-TB Gold) to exclude latent TB. Groups (indeterminate versus determinate) were compared and tested for association with clinically relevant variables (age, aetiology of liver disease, MELD score, Child–Pugh score and absolute lymphocyte count). Groups were compared using the Mann–Whitney U-test. Testing for independent associations with an indeterminate test was performed using multivariate logistic regression. An ROC curve was also created to define the optimum probability cut off. Results: Of the 49 IGRAs performed, 12 (24%) were indeterminate and 37 (76%) were determinate. Of the determinate IGRAs, three (6 %) were positive and 34 (70%) were negative. Patients who had an indeterminate test were significantly older than those with determinate tests (57.3 vs. 50.5, p = 0.03). There were no other statistically significant differences between the indeterminate versus determinate group: mean MELD score (16.1 vs. 15.8, p = 0.43), mean Child–Pugh score (9.8 vs. 8.8, p = 0.21), mean absolute lymphocyte count (1.03 vs. 0.95, p = 0.30). Multivariate logistic regression testing for independent associations with an indeterminate test, together with changes in the area under the ROC curve, are shown in Table 1. The combination of these variables could provide a high degree of accuracy for predicting an indeterminate test with a c statistic of 0.80 for the ROC curve (Fig. 1). Table 1 Multivariate analysis for factors associated with an indeterminate IGRA. Odds ratio (95% CI) P value Cumulative c-statistic Lymphocyte count 0.78 (0.18 to 3.37) 0.74 0.685 Age 1.08 (0.98 to 1.19) 0.10 0.689 Aetiology (HCV vs. other) 3.24 (0.47 to 22.43) 0.23 0.706 MELD score 0.87 (0.61 to 1.25) 0.46 0.766 Child–Pugh score 1.91 (0.90 to 4.01) 0.94 0.800 Figure 1 ROC curve Conclusions: Patients with advanced liver disease had a high rate of indeterminate IGRA in this study relative to other groups described in the literature, which may limit the utility of this test in this population. Indeterminate tests could be predicted with moderate to high accuracy using five routinely collected clinical variables. The reason for the high prevalence of indeterminate tests may relate to immune dysfunction associated with advanced liver disease but results require validation in larger studies. Detection of hereditary haemochromatosis using haematology parameters P Bentley1, M Ferman2, D Trinder3,4, A Chua3,4, S Hazeldine2 and JK Olynyk2,5,6 1Australian Red Cross Blood Service, Perth, Western Australia, Australia; 2Department of Gastroenterology, Fiona Stanley Hospital, Murdoch, Western Australia, Australia; 3Harry Perkins Institute of Medical Research, Nedlands, Western Australia, Australia; 4School of Medicine and Pharmacology, University of Western Australia, Western Australia, Australia; 5Faculty of Health Sciences, Curtin University, Bentley, Western Australia, Australia; 6School of Veterinary and Life Science, Murdoch University, Murdoch, Western Australia, Australia Introduction: Elevated serum transferrin saturation and ferritin levels are the commonest screening abnormalities for detecting C282Y homozygous hereditary haemochromatosis (HH). Previous studies suggest that subjects with HH exhibit different peripheral blood erythrocyte parameters compared to non-HH subjects. The aim of this study was to evaluate haemoglobin concentration (Hb), mean cell volume (MCV), mean cell haemoglobin (MCH), and mean cell Hb concentration (MCHC) as screening tests for asymptomatic HH. Methods: Peripheral blood Hb, MCV, MCH and MCHC values were obtained from consecutive asymptomatic subjects with HH (42 male and 30 female) attending a large community-based clinic. Values were obtained prior to and immediately following phlebotomy treatment. Healthy plasma donors from the Australian Red Cross Blood Service (220 male and 216 female) served as a control group. Comparisons between groups were performed using parametric t tests. Receiver-operator curve (ROC) analysis was performed to determine the optimal sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for each parameter. Results: The mean ages for male and female HH subjects at diagnosis (45 y and 45 y, respectively) were similar to those of control male (42 y) and female (40 y) subjects. Serum ferritin levels at diagnosis for male HH (1182 ± 250 µg/L) and female HH subjects (479 ± 64 µg/L) were significantly higher than control values (male 164 ± 27 µg/L, female 108 ± 19 µg/L, p < 0.01)). Prior to treatment, and compared with controls, male and female HH subjects had significantly higher Hb values (male HH 155 ± 2, female HH 138 ± 2, male control 147 ± 1, female control 128 ± 1 g/L, p < 0.0001), MCV values (male HH 92.7 ± 0.6, female HH 95.4 ± 0.7, male control 88.7 ± 0.3, female control 89.6 ± 0.3 fL, p < 0.0001), MCH (male HH 32.1 ± 0.3, female HH 32.1 ± 0.3, male control 30.0 ± 0.2, female control 29.7 ± 0.2 pg, p < 0.0001) and MCHC (male HH 346 ± 2, female HH 337 ± 2, male control 338 ± 1, female control 332 ± 1 g/L, p < 0.01). Following treatment of HH subjects (end of treatment male ferritin 59 ± 8 µg/L, female ferritin 70 ± 8 µg/L), MCV, MCH and MCHC values all declined significantly compared with pre-treatment values but were still all significantly elevated compared with control values. Areas under the curve from ROC analysis were as follows – Hb (male 0.71, female 0.77), MCV (male 0.77, female 0.77), MCH (male 0.81, female 0.82), MCHC (male 0.67, female 0.67). For both genders MCH > 31 pg had sensitivity (male 78%, female 83%), specificity (male and female 76%), PPV (male 76%, female 77%), NPV (male 77%, female 81%). Conclusions: Asymptomatic male and female HH subjects have significantly higher Hb, MCV, MCH and MCHC values compared with control subjects. These differences persist following treatment. A MCH value greater than 31 pg should prompt further assessment for HH. Non-alcoholic fatty liver disease related cirrhosis is frequently an incidental diagnosis L Calzadilla Bertot1, GP Jeffrey1,2, G MacQuillan1,2, G Garas1,2, M Wallace1,2 and LA Adams1,2 1School of Medicine and Pharmacology, University of Western Australia, Nedlands, WA, Australia; 2Department of Hepatology, Sir Charles Gairdner Hospital, Nedlands, WA, Australia Background and Aim: Non-alcoholic fatty liver disease (NAFLD) is the commonest cause of liver test abnormalities in the community and is frequently encountered by general practitioners and specialists. Determination of the presence of cirrhosis is important as it significantly alters prognosis and mandates screening for hepatocellular carcinoma (HCC). We sought to determine in a cohort of patients with NAFLD related cirrhosis, whether cirrhosis had been discovered incidentally or had been diagnosed intentionally using non-invasive fibrosis measures or liver biopsy and whether patients had been screened for HCC. Methods: Patients with NAFLD cirrhosis seen at a tertiary academic centre were identified from prospectively collected databases. Cirrhosis was confirmed by liver biopsy or by imaging evidence (nodular liver with evidence of portal hypertension). Medical records were reviewed to determine the following factors at the time of cirrhosis diagnosis: mode of cirrhosis diagnosis (incidental or by intent), specialty of diagnostician, and evidence of previous clinical, imaging or laboratory features of unrecognized cirrhosis. Results: Of 100 NAFLD related cirrhosis patients (mean age 66 years, 62% male), the majority (n = 66) were discovered incidentally, and the majority of these (74%) had their NAFLD diagnosed at the same time of cirrhosis diagnosis. Incidental diagnosis was most commonly made on imaging performed for reasons unrelated to liver disease (n = 21, 32%), following liver tests performed for reasons unrelated to NAFLD or liver disease (n = 17, 26%), following incidental finding of varices on endoscopy (n = 14, 21%), unexpected finding at surgery (n = 9, 14%) or other (n = 5, 8%). By-intent diagnoses were predominantly made by gastroenterologists/hepatologists (27/34 patients), whereas general practitioners, surgeons and physicians were more likely to diagnose cirrhosis incidentally (p < 0.001). A high proportion of patients who were diagnosed incidentally (n = 48/66, 73 %) had previous evidence of cirrhosis (unexplained thrombocytopenia or splenomegaly) during previous outpatient clinic review or hospitalization, although no patient had undergone non-invasive fibrosis testing. When calculated retrospectively, the majority had high non-invasive scores prior to the diagnosis (81% had Hepascore >0.84 and NAFLD Fibrosis Score >0.676, and 80% had FIB-4 >1.5). Thirty-nine of the total cohorts were diagnosed with HCC, half of whom had not received previous screening. Patients not in a HCC screening program were more likely to have a higher BCLC, (Stage C, 52%) than those in a screening program (Stage C, 20%) (p = 0.01). Conclusions: The majority of NAFLD cirrhosis patients are diagnosed incidentally despite prior clinical and biochemical evidence. Screening with non-invasive fibrosis markers is not done outside specialty clinics but would detect 80% of these patients. Screening of HCC in NAFLD cirrhotics is associated with a lower BCLC stage at time of diagnosis. Increased awareness of NAFLD cirrhosis and appropriate cirrhosis and HCC screening methods is needed. Validation of non-invasive scoring systems for the prediction of overall mortality and liver related events in non-alcoholic fatty liver disease L Calzadilla Bertot1, GP Jeffrey1,2, G MacQuillan1,2, G Garas1,2, I Huang1 and LA Adams1,2 1School of Medicine and Pharmacology, University of Western Australia, Nedlands, WA, Australia; 2Department of Hepatology, Sir Charles Gairdner Hospital, Nedlands, WA, Australia Background and aims: Long term predictors of overall mortality and liver decompensation are important to determine prognosis and thus influence management of patient with non-alcoholic fatty liver disease (NAFLD). We aimed to examine the accuracy of four non-invasive clinical models in the prediction of overall mortality and liver related events in NAFLD patients. Methods: We conducted a retrospective cohort study of 249 patients diagnosed with NAFLD between 2006 and 2014 in a tertiary centre in Western Australia. Diagnosis of NAFLD was made in 212 (85%) by liver biopsy and 37 (15%) using non-invasive techniques. Patients with available data for Hepascore, NAFLD fibrosis score, APRI and FIB-4 scores were included. Outcomes included overall mortality/liver transplantation and occurrence of a first event of hepatic decompensation (ascites, variceal bleed, encephalopathy and hepatocellular carcinoma). Multivariate forward stepwise Cox proportional hazards model was used to determine the association between the models and outcomes. C statistics were used to assess and compare the accuracy of the four biomarker models to predict overall mortality and liver related events. Kaplan–Meier survival analysis was performed, assigning patients to 3 groups (low, intermediate and high risk) based on original predefined cut off values of the non-invasive biomarkers panels and compared by log-rank testing. Model calibration was assessed by goodness of fit (Hosmer-Lemeshow test (H-L)) comparing observed and predicted risk. Results: A total of 249 patients were included (mean age 54 ± 13 years, 56% women, body mass index, 37 (±9.0)). At baseline, 129 (52%) had type 2 diabetes. Over a mean follow-up of 3.2 years, (range 1–9) 9.6% of patients died or underwent liver transplantation, and 14.5% developed a liver related event. Hepascore (hazard ratio (HR) 4.6, 95% confidence intervals (CI) (2.27–9.34)), NAFLD fibrosis score (HR 3.1, CI: 1.82–5.27), APRI (HR 2.3, CI: 1.33–4.0) and FIB-4 (HR 2.8, CI: 1.63–5.13) were associated with overall mortality. The HRs for liver related events were 5.0 (95% CI: 2.78–9.0), 4.0 (95% CI: 2.53–6.43), 2.87 (95% CI: 1.8–4.5) and 4.1 (95 % CI: 2.4–7.2) for Hepascore, NAFLD fibrosis score, APRI and FIB-4, respectively. Unadjusted cumulative probabilities of death/liver transplantation at 9 years of follow-up were significantly different among high risk groups for Hepascore >0.55 (31%), NAFLD fibrosis score >1.455 (40%), FIB-4 >1.30 (36%) and APRI >0.5 (30%) as compared with the comparative low risk group, Hepascore <0.55 (1.3%), NAFLD fibrosis score <1.455 (2.6%) FIB-4 <1.30 (3.3%) and APRI <0.5 (10%) (all p<0.001 by log rank test). Unadjusted cumulative probability of having an event of hepatic decompensation at 9 years of follow up were 77, 87, 91 and 92% for patients in high category of Hepascore, NAFLD fibrosis score, APRI and FIB-4, respectively. NAFLD fibrosis score, Hepascore and FIB-4 showed a similar predictive accuracy, for both overall mortality/liver transplant and liver related events outcomes with similar C statistics (Table 1). Each model calibrated well (Hepascore, H-L = 6.77, p = 0.56), NAFLD fibrosis score, H-L = 5.86, p = 0.66), APRI, H-L = 6.77, p = 0.56 and FIB-4, H-L = 10.70, p = 0.21) with no differences between observed and predicted risk for overall morality/liver transplantation. Conclusions: Non-invasive scoring systems are useful tools to predict overall mortality and liver related events in NAFLD patients according to predefined risk groups. The use of these biomarkers panels should be included in daily clinical practice. Table 1 C index of four non-invasive scoring systems to predict outcomes C statistics ± SE (95% CI) Score events Overall mortality/liver transplant Liver related NAFLD fibrosis score 0.79 ± 0.04 (0.71–0.87) 0.84 ± 0.03 (0.78–0.91) Hepascore 0.79 ± 0.05 (0.70–0.89) 0.83 ± 0.03 (0.76–0.91) FIB-4 0.79 ± 0.04 (0.70–0.88) 0.85 ± 0.04 (0.79–0.92) APRI 0.70 ± 0.05 (0.60–0.80) 0.77 ± 0.04 (0.69–0.86) A study of nutritional support alone for alcoholic hepatitis: a case for more aggressive therapy T Bhatia1, K Changela2 and S Anand2 1Carepoint Health; 2The Brooklyn Hospital Center Background: Alcoholic liver disease (ALD) comprises a wide spectrum of injury ranging from fatty liver to frank cirrhosis. Alcoholic hepatitis (AH) encompasses mild to severe life threatening injury and often presents acutely in the setting of chronic liver disease. Protein calorie malnutrition is a common finding in patients with ALD, which puts them at increased risk for complications. The current guidelines recommend nutritional support for mild–moderate AH and the addition of steroids in cases of severe AH. In this study, we examined the outcome of nutritional support alone in patients with AH who had a mean determinant factor (MDF) of less than 32 or had a contraindication to steroids. Methods: Our study is a retrospective observational study. All hospitalized adult patients with diagnosis of AH were identified using ICD-9 code. The study period included patients who were admitted between January 2007 and December 2012. Demographics, clinical, nutritional, and laboratory data were collected and analyzed. The subjects included received only high-calorie–high-protein nutritional support and did not receive steroid or pentoxyphiline therapy. Ineligibility to steroid therapy was based on MDF <32 or having a contraindication to steroids. Results: Seventy-three patients fulfilled the inclusion criteria (52 male and 21 female). Mean age was 50.5 years. Mean length of stay was 9.3 days. Eighteen percent developed hepatic encephalopathy, 31% had ascites and 4% developed spontaneous bacterial peritonitis. Infections were observed in 31% of the cohort. Out of the patients with complete laboratory panels on admission and discharge, MELD worsened in 66% (mean MELD on admission – 14.3), albumin worsened in 38% (mean albumin on admission – 3.3). Total deaths were 6 within 1 month and 7 within 3 months of admission. Conclusion: This study indicates that nutritional support alone was highly inadequate in managing AH even when the MDF was less than 32. A more aggressive approach with the introduction of steroids or pentoxyphiline in borderline cases appears to be a good strategy. Are indirect biomarkers enough for community treatment? S Bloom1,2,10, J Lubel1,10, A Dev3,8, P Gow4,9, A Nicoll1,5,9,10, S Roberts2,8, S Bell6,8, I Kronborg7, V Knight3, S Sood5, D Lewis1,10, W Kemp2,8 and on behalf of Melbourne Liver Group (MLG) 1Departments of Gastroenterology and Hepatology at Eastern Health, Melbourne, Australia; 2Alfred Health, Melbourne, Australia; 3Monash Health, Melbourne, Australia; 4Austin Health, Melbourne, Australia; 5Royal Melbourne Hospital, Melbourne, Australia; 6St. Vincent's Hospital, Melbourne, Australia; 7Western Health, Melbourne, Australia; 8Monash University, Melbourne, Australia; 9University of Melbourne, Victoria, Australia; 10Eastern Health Clinical School Monash University, Melbourne, Australia Introduction: The recently released Australian Hepatitis C (CHC) Management Guidelines recognize the importance of fibrosis assessment prior to direct acting anti-viral therapy (DAA). In the absence of a liver biopsy gold standard, it is recommended to use a non-invasive tool such as transient elastography (TE) or an indirect biochemical marker (APRI) to determine treatment duration and need for specialist care. This study aims to explore the accuracy of commonly used indirect biochemical markers (APRI and FIB4) as an alternative to TE prior to DAA therapy in CHC patients. Methods: Individuals undergoing initial assessment at tertiary centres and through a community-screening program were prospectively recruited over 18 months. The inclusion criteria included age greater than 18 years and evidence of HCV infection duration greater than 6 months. Baseline demographic and biometric information were collected in conjunction with liver stiffness measurement (LSM) and biochemical assessment. All patients were fasted for at minimum two hours prior to assessment. APRI and LSM cut-offs of 1.0 and 12.5 kPa, respectively, were taken as cut-offs to exclude cirrhosis as per current Australian consensus guidelines. Results: A total of 677 patients were included in the analysis. The prevalence of LSM ≥12.5 kPa was 16.8% (114 participants). An APRI ≥1.0 was observed in 23.8% (161 participants). APRI and LSM agreed in 83.8% of participants and were moderately correlated (Pearson's 0.587, p < 0.005). FIB4 ≥1.45 was observed in 36.8% (249 participants) and agreed with LSM in 75.8% (Pearson's 0.612, p < 0.005). Using TE as the reference standard based on superior diagnostic performance,1 APRI had an 11.5% false positive rate (FP) (13.9% of LSM <12.5 kPa) and a 4.6% false negative rate (FN) (27.2% of LSM ≥12.5 kPa). APRI had an area under the receiver operator curve (AUROC) of 0.861. Of those at risk of cirrhosis on TE (LSM > 12.5 kPa) with an APRI <1 (n = 31), the median LSM was 17.3 kPa (range 13.20–53.30 kPa). Ten (32.3% of FN group) had a LSM ≥21 with 60% having subsequent radiological evidence of portal hypertension. FIB4 was associated with a higher FP of 22.5% (26.9% of LSM <12.5 kPa) compared with APRI (p=0.045) but a lower FN of 2.6% (n = 17, p = 0.029) (15.2% of LSM ≥12.5 kPa). Three patients with LSM ≥21 were misclassified (v. APRI p = 0.051). FIB4 had a AUROC of 0.881 and a higher negative predictive value than APRI (Table 1). In our cohort an APRI of 0.853 was the optimal cut-off as a screening tool. This cut-off had a FP of 15.2%, however, a lower FN of 3.5%. The optimal threshold for FIB4 was 1.531 with 18.3% FP and 2.7% FN. Using either a positive APRI or FIB4 as an indicator for further assessment, reduced the FN to 2.0% with a higher FP rate (24.2%). Table 1 Sensitivity (95% CI) Specificity (95% CI) PPV NPV +LR −ve LR APRI ≥1 72.81% (63.67–80.72) 86.15% (83.01–88.89) 51.55% 93.99% 5.26 0.32 FIB4 ≥1.45 85.09% (77.20–91.07) 73.89% (70.05–77.47) 39.75% 96.07% 3.26 0.20 APRI ≥0.86 78.95% (70.31–86.02) 81.71% (78.26–84.81) 46.63% 95.05% 4.32 0.26 FIB4 ≥1.531 83.93% (75.79–90.19) 77.98% (74.32–81.33) 43.12% 96.06% 3.81 0.21 APRI + FIB4 87.72% (80.25–93.12) 70.82% (66.87–74.55) 37.88% 96.60% 3.01 0.17 Conclusion: The real world use of indirect biomarkers requires further optimization to minimize the number with cirrhosis under diagnosed and undertreated with current DAAs, particularly genotype 3. Furthermore, 16.4% at risk of portal hypertension would be missed with the current APRI cut-off in our cohort. No single test in isolation is sufficient. In the absence of a gold standard, all clinical and radiological information should be used to assess for the presence of cirrhosis. References 1Castera L, Vergniol J, Foucher J et al. Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C. Gastroenterology 2005; 128 (2): 343– 350. Community approach targeting cirrhosis and hepatocellular carcinoma (CATCH) - community cirrhosis prevalence in viral hepatitis S Bloom1,2,10, W Kemp2,8, A Dev3,8, P Gow4,9, A Nicoll1,5,9,10, S Roberts2,8, S Bell6,8, I Kronborg7, V Knight3, S Sood5, D Lewis1,10, J Lubel1,10 and on behalf of Melbourne Liver Group (MLG) 1Eastern Health, Melbourne, Australia; 2Alfred Health, Melbourne, Australia; 3Monash Health, Melbourne, Australia; 4Austin Health, Melbourne, Australia; 5Royal Melbourne Hospital, Melbourne, Australia; 6St. Vincent's Hospital, Melbourne, Australia; 7Western Health, Melbourne, Australia; 8Monash University, Melbourne, Australia; 9University of Melbourne, Victoria, Australia; 10Eastern Health Clinical School Monash University, Melbourne, Australia Introduction: The population prevalence of cirrhosis is unknown with an urgent need to detect those in the community at risk of liver related morbidity and mortality. Fibroscan is a suitable screening tool with predetermined cut-offs validated in chronic hepatitis B (CHB) and C (CHC). The aim of this study was to estimate the prevalence of significant fibrosis and cirrhosis as assessed by liver stiffness measurement (LSM) in at-risk populations and determine the practicality of community screening for cirrhosis and ultimately the role of this intervention for detection of hepatocellular carcinoma (HCC). Methods: Participants were recruited from 21 primary care practices across Melbourne, Australia. Inclusion criteria included age 18–80 years, with CHB or CHC (duration >6months), absence of prior or recent (<18months) specialist input and no current or prior HCC. Clinical assessment, transient elastography (Fibroscan® 402) and blood analysis were performed in the primary care setting. Scans were considered reliable if the success rate was >60% and IQR/median stiffness <0.3. LSMs of 8.0 and 12.5 kPa were taken as cut-offs to represent those at risk of significant fibrosis (>F2) and cirrhosis (F4), respectively. Individuals meeting these cut-offs were referred to specialist centres for ongoing management. For comparison the community cohort (c) were compared with a consecutively recruited control group (h) comprising patients referred to one tertiary center for identical initial assessment. Results: From October 2014 to April 2016, 1098 patients were invited to participate. Failure to attend rate was 31.5%. In total, 751 participants were assessed (498 CHC/253 CHB). LSM failure occurred in one patient (0.13%) and phlebotomy failure in 6 (0.80%). Assessment was performed using the M probe in 95.5% (n = 717) and the XL probe in 4.5% (n = 34). Our control cohort (179 CHC/68 CHB) only differed on demographic data on one account; older median age in hCHC (p < 0.001). In the community CHC cohort (cCHC), the mean LSM was 9.92 kPa, and in the community CHB cohort (cCHB), it was 5.6 kPa. There was no significant difference between mean LSM in the community compared with the control cohort (CHC p = 0.773, CHB p = 0.067). An LSM ≥8 kPa in the community was observed in 29.8% (208 cCHC [41.8%], 16 cCHB [6.3%] p = <0.001), and an LSM ≥12.5 kpa was seen in 9.98% (all cCHC [15.1%]). Three cCHB patients were subsequently diagnosed as cirrhotic (1.2% of cCHB) on clinical and radiological grounds (all had LSM >11.0). This was at a lower prevalence than our hCHB (5.9% – p = 0.039); however, indications for treatment were present in a significant proportion (24.1%) of the cCHB cohort. The distribution of LSM ≥12.5 kPa was not significantly difference between the cCHC group and our hCHC (p = 0.739). On univariate analysis, advanced age, duration of infection, genotype 6, skin to capsule depth, body mass index (BMI), at risk alcohol consumption (>140 gram per week) and smoking (>10 p.y. history) were associated with an LSM ≥12.5 kPa in cCHC. Age (OR 1.062 p < 0.001), elevated BMI (OR 1.128 p = 0.002), alcohol consumption (OR 2.549 p = 0.001) and genotype 6 (OR 8.347 p = 0.009) remained significant on multivariate analysis. In cCHB; elevated viral load (OR 2.063 p = 0.023), BMI (OR 1.518 p = 0.019) and viral phase 2 (OR 21.818 p = 0.033) were associated with increased risk of cirrhosis, however, did not remain significant on multivariate analysis. Six malignancies have been detected through the program, one non-small lung cancer, two colorectal cancers (DUKE B & D) and three hepatocellular carcinomas (all BCLC A). Conclusion: This study demonstrates that a community-based screening program using LSM is practical with an acceptable uptake in the CHC and CHB population. Based on the LSM data, the prevalence of fibrosis and cirrhosis in the community CHC patients is significant and comparable with that seen in a tertiary hospital referral cohort, highlighting the need of HCC surveillance programs in the community. Hepatocellular carcinoma surveillance in primary care S Bloom1,2,10, J Lubel1,10, A Dev3,8, P Gow4,9, A Nicoll1,5,9,10, S Roberts2,8, S Bell6,8, I Kronborg7, V Knight3, S Sood5, D Lewis1,10, W Kemp2,8 and on behalf of Melbourne Liver Group (MLG) 1Eastern Health, Melbourne, Australia; 2Alfred Health, Melbourne, Australia; 3Monash Health, Melbourne, Australia; 4Austin Health, Melbourne, Australia; 5Royal Melbourne Hospital, Melbourne, Australia; 6St. Vincent's Hospital, Melbourne, Australia; 7Western Health, Melbourne, Australia; 8Monash University, Melbourne, Australia; 9University of Melbourne, Victoria, Australia; 10Eastern Health Clinical School Monash University, Melbourne, Australia Introduction: There are approximately 218,000 people with chronic hepatitis B (CHB) in Australia with around 1000 cases of hepatocellular carcinoma (HCC) diagnosed each year; less than half of these HCCs are diagnosed through surveillance programs. The uptake of HCC screening in the CHB population managed in primary care is unknown. This study examines the HCC screening practices in primary care. Methods: Patients with CHB over the age of 18 and managed in primary care were invited to participate. Each patient underwent a thorough history, examination and comprehensive assessment of hepatic fibrosis using transient elastography and serological tests. Patients were also evaluated for HCC risk using accepted criteria (demographic criteria as per AASLD recommendations and transient elastography). Results: Over 18 months, 322 patients were recruited (male 47.4% and 94.1% of Asian ethnicity). The median age was 44 with median estimated duration of infection 32.8 ± 14.5 years. Median liver stiffness was 5.1 ± 1.9 kPa with 1.9% LSM ≥ 11.5 kPa. The majority of patients (64.8%) were in phase 3 (immune control) with 12.5% unable to be clearly classified into a phase. The average viral load was 6,344,886 IU/L and mean ALT was 31 IU/L. Yearly biochemical and virological assessment was performed in 68.3%. A prior specialist review had occurred in 23.1% with 36.7% having had an ultrasound (US) in the last 6 months. No radiological assessment of liver disease had occurred in 30.4%. HCC surveillance was indicated in 172 (53.6%); only 66 (38.4%) had a US in the last 6 months and 46 (32.4%) had no prior radiological assessment. Surveillance was indicated due to family history in 37 (21.5%), a high probability of cirrhosis (LSM ≥11.5 kPa) in 5 (2.9%) or bas