BMJ Open. For peer review only -

Transcription

1 Direct targeting of risk factors leads to a two-fold increase in the detection of liver cirrhosis in primary care: Prospective cross-sectional study Journal: BMJ Open Manuscript ID: bmjopen-0-00 Article Type: Research Date Submitted by the Author: -Dec-0 Complete List of Authors: Harman, David; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Ryder, Stephen; Nottingham James, Martin; University of Nottingham, Nottingham Digestive Diseases NIHR Biomedical Research Unit Jelpke, Matthew; NHS Rushcliffe Clinical Commissioning Group, Ottey, Sean; NHS Rushcliffe Clinical Commissioning Group, Wilkes, Emilie; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Card, Tim; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Aithal, Guruprasad; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Guha, Neil; University of Nottingham, <b>primary Subject Heading</b>: Secondary Subject Heading: Diagnostics Keywords: Gastroenterology and hepatology Adult gastroenterology < GASTROENTEROLOGY, Hepatobiliary disease < GASTROENTEROLOGY, GENERAL MEDICINE (see Internal Medicine) BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

2 Page of BMJ Open Direct targeting of risk factors leads to a two-fold increase in the detection of liver cirrhosis in primary care David J Harman, Stephen D Ryder, Martin W James, Matthew Jelpke, Dominic S Ottey, Emilie A Wilkes, Timothy R Card,, Guruprasad P Aithal *, Indra Neil Guha * *=joint senior authors National Institute for Health Research Nottingham Digestive Diseases Biomedical Research Unit (NDDBRU), Nottingham University Hospitals NHS Trust and University Of Nottingham, Nottingham, NG UH, United Kingdom (DJ Harman Clinical Research Fellow in Hepatology; SD Ryder, MW James, EA Wilkes Consultant Hepatobiliary Physician; TR Card - Consultant Gastroenterologist and Clinical Associate Professor of GI epidemiology; GP Aithal - Consultant Hepatobiliary Physician and Professor of Hepatology; IN Guha - Consultant Hepatobiliary Physician and Clinical Associate Professor of Hepatology) NHS Rushcliffe Clinical Commissioning Group, Nottingham, NG LQ, United Kingdom (M Jelpke, DS Ottey General Practitioner) Division of Epidemiology and Public Health, Clinical Sciences Building Phase, City Hospital Campus, University of Nottingham, Nottingham, NG PB, United Kingdom (TR Card - Consultant Gastroenterologist and Clinical Associate Professor of GI epidemiology) Corresponding Author: Dr Neil Guha Nottingham Digestive Diseases Biomedical Research Unit E Floor, West Block Queens Medical Centre, Derby Road, Nottingham NG UH neil.guha@nottingham.ac.uk Telephone: 0 ext 00 Manuscript Word Count (excluding figure legends and references):, words Abstract Word Count: words Tables: + supplementary tables Figures: + supplementary figure Abbreviations: ALD, alcoholic liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CPRD, Clinical Practice Research Datalink; kpa, kilopascals; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; NHS, National Health Service; PHG, portal hypertensive gastropathy; TE, transient elastography; UK, United Kingdom Keywords: Chronic liver disease; cirrhosis; hepatic fibrosis; transient elastography; screening BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

3 Page of Abstract Objectives: To assess the feasibility of a novel diagnostic algorithm targeting patients with risk factors for chronic liver disease in a community setting. Design: Prospective cross-sectional study. Setting: Two primary care practices (adult patient population, ) in Nottingham, United Kingdom. Participants: Adult patients (aged years or over) fulfilling one or more selected risk factors for developing chronic liver disease; i) Hazardous alcohol use, ii) Type Diabetes or iii) Persistently elevated alanine aminotransferase (ALT) liver function enzyme with negative serology. Interventions: A serial biomarker algorithm, using a simple blood-based marker (AST:ALT ratio for hazardous alcohol users, BARD score for other risk groups) and subsequently liver stiffness measurement using Transient Elastography (TE). Main outcome measures: Diagnosis of clinically significant liver disease (defined as liver stiffness kilopascals); definitive diagnosis of liver cirrhosis. Results: We identified 0 patients with the defined risk factors of whom 0 patients agreed to undergo investigation. A normal blood biomarker was found in patients (.%) who required no further investigation. Subsequently, patients agreed to undergo transient elastography, of whom (.% of valid scans) had elevated liver stiffness. Importantly / (.%) patients with elevated liver stiffness had normal liver enzymes and would be missed by traditional investigation algorithms. We identified new patients with definite cirrhosis, representing a 0% increase in the number of diagnosed cases in this population. Conclusions: A non-invasive liver investigation algorithm based in a community setting is feasible to implement. Targeting risk factors using a non-invasive biomarker approach identified a substantial number of patients with previously undetected cirrhosis. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

4 Page of BMJ Open Strengths and Limitations of this Study: - The current study provides an alternative diagnostic algorithm to the usual flawed concept of investigating patients for liver disease on the basis of elevated liver function enzymes. We introduced a novel algorithm in a primary care, which targets established risk factors for chronic liver disease. Patients were investigated patients utilising simple and cheap blood-based biomarkers followed by Transient Elastography (a validated biomarker of liver fibrosis). - Patients with common lifestyle-related risk factors for chronic liver disease (including hazardous alcohol use and type diabetes) were identified using a widely utilised prospective primary care database in the United Kingdom, from general practices with an adult population of, patients. Our patient selection method, and subsequent investigation algorithm, would therefore be easy to adopt at other primary care sites in the United Kingdom. - Further study is required in a number of areas to address limitations of the current study, including validation of the diagnostic algorithm in other regions of the UK, investigation using alternative simple fibrosis markers or Transient Elastography, and health economic analysis of earlier identification of chronic liver disease. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

5 Page of Introduction Current strategies to identify liver disease in the general population rely upon the use of liver function enzyme blood tests, which are non-specific markers of liver injury. In the United Kingdom (UK) persistently elevated liver enzymes typically trigger referral to hospital-based gastroenterology and hepatology clinics, where liver biopsy is the reference standard investigation for staging of hepatic fibrosis. This is an inadequate strategy, as the absence of symptoms in early stages of liver disease, and poor sensitivity of liver function blood tests to detect hepatic fibrosis, often result in late diagnosis. A recent study, using data from a large community database, the clinical practice research datalink (CPRD), suggested that 0% of patients with cirrhosis are given their initial diagnosis of liver disease after the first hospitalisation with decompensation which substantially impaired prognosis independent of stage of cirrhosis versus patients who are diagnosed whilst ambulatory (Hazard Ratio (HR)=., % CI.-.0)(). By contrast, many patients referred for investigation of abnormal liver tests have no evidence of significant liver disease when investigated in hospital. A large study from Tayside, Scotland() has shown that at least % of a community population will have their LFTs checked in a decade and around a third will have at least one abnormal value. While abnormal transaminases were predictive of liver disease (HR.) the actual rate of detection was very low with only.% of those with abnormal transaminases being diagnosed with liver disease within years of the abnormal test. Subsequently, the BALLETS study () has shown convincing evidence that most abnormal LFTS in primary care are due to non-alcoholic fatty liver disease (NAFLD) or alcohol, and the yield for detecting other parenchymal liver disease was only %. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

6 Page of BMJ Open These inefficient diagnostic algorithms have contributed to the fact that liverrelated deaths have continued to rise in the United Kingdom (UK) (), in stark contrast to other major causes of death (). Liver cirrhosis is now the third commonest cause of premature death (persons aged less than years)(). Given the majority of chronic liver disease results from lifestyle-related risk factors which are amenable to intervention, deaths from chronic liver disease are largely preventable. In order to tackle the rising cirrhosis prevalence and mortality, we need an investigation strategy which results in early detection of hepatic fibrosis whilst patients are still asymptomatic, and which is both economically efficient and acceptable to patients. Targeting risk factors for chronic liver disease rather than abnormal blood tests alone, performed directly in a community setting, is one such approach which is fundamentally different to current diagnostic pathways. In a community setting, liver biopsy is not a practical reference standard for liver disease stratification as it is invasive and requires specialist provision (). Numerous noninvasive alternatives to liver biopsy for fibrosis staging have been validated in hospital populations, but require further evaluation in primary care before widespread adoption into investigation algorithms for the general population. Despite the widespread recognition that significant liver disease exists in the context of normal liver function tests, current management algorithms are still dependent on abnormal liver enzymes to initiate investigation. Our aim was to challenge this flawed concept by testing the feasibility of a novel diagnostic algorithm which directly targeted risk groups (including alcohol and diabetes) rather than liver functions tests. Importantly, we designed a pathway that integrated non-invasive diagnostic tests and liver specialists within a community setting. The pathway used BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

7 Page of simple tests followed by a validated test of liver fibrosis (transient elastography). We hypothesised this approach would detect a substantial number of hitherto undiagnosed cases of chronic liver disease. Materials and Methods Study Setting This was a prospective study with recruitment from two general medical practices in Nottingham, United Kingdom with a total patient population of,. The study ran for months from February 0 to April 0 and local regulatory approval was obtained (/HO0/ and /EM/0). Both practices utilise the SystmOne general practice records system (TPP, United Kingdom) for live recording of clinical, anthropometric and biochemical patient data. Data is stored as searchable numeric data or prospectively coded with Read Codes (clinical encoding of numerous parameters including patient demographics, diagnoses, clinical signs and laboratory test results). Study Patient Selection From the electronic general practice records, we identified adult patients (aged years or older) with selected risk factors for lifestyle related chronic liver disease at the start of the study period. Patients were eligible for study inclusion, regardless of previous liver function blood test results, if identified with one or more of the following chronic liver disease risk factors: BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

8 Page of BMJ Open a) Hazardous alcohol use - defined as > units per week ethanol consumption for women and > units per week ethanol consumption for men, or alcohol AUDIT questionnaire score, or presence of Read codes related to alcohol misuse*. b) Type Diabetes Presence of Read codes related to type diabetes (a full code list is available from the authors on request). c) Raised ALT As a comparator group, patients with persistently elevated serum alanine aminotransferase levels (defined as consecutive occasions above local laboratory cut-offs) without type diabetes or hazardous alcohol use were assessed. In order to ensure only patients with alcohol-related liver disease or NAFLD were investigated, other causes of chronic liver disease were excluded by specific testing of chronic viral hepatitis B and C, autoimmune liver diseases, haemochromatosis and common genetic liver diseases, and no other relevant reason for transaminase elevation (e.g, congestive cardiac failure, medication use). Local laboratory cut-offs denoting elevated ALT are >U/L for women and >U/L for men. Identified patients were excluded from the study a priori if i) there was definitive evidence of hepatic fibrosis or cirrhosis from previous investigation, ii) there was a contraindication to Transient Elastography (pregnancy, indwelling cardiac device) or iii) patients unable to consent to investigation, or housebound and could not attend the community practice. Diagnostic Algorithm Patients identified with the defined risk factors were invited to the diagnostic algorithm (see Figure ). Patients with type diabetes were invited opportunistically at their diabetes annual review. Patients with hazardous alcohol use were invited BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

9 Page of opportunistically during primary care appointments or via letter where they did not undergo a consultation during the study period. Patients in the raised ALT subgroup were prospectively referred by the investigating general practitioner. Patients initially underwent a simple blood-based biomarker, with a high negative predictive value to accurately rule out hepatic fibrosis and therefore avoid the need for further tests. Hazardous alcohol users had the AST:ALT ratio (aspartate aminotransferase:alanine aminotransferase) measured, a simple and cheap blood biomarker which has been shown to reflect disease severity in a number of chronic liver diseases, including alcoholic liver disease (ALD)(), autoimmune liver disease() and hepatitis C(). As there is limited published data on diagnostic thresholds of AST:ALT ratio identifying alcohol-related liver fibrosis, a conservative AST:ALT cutoff of 0. was used to rule out hepatic fibrosis. Patients with type diabetes or persistently raised ALT had the BARD score() measured, a simple weighted score of body mass index kg/m, AST:ALT ratio 0., and type diabetes. A BARD score of less than in hospital populations of non-alcoholic fatty liver disease (NAFLD) has a high negative predictive value to rule out significant hepatic fibrosis(-). We therefore considered that a score of indicated an increased risk. Patients with a normal simple biomarker test result did not proceed down the algorithm. Patients with a high simple biomarker result underwent transient elastography (TE) at the community practice. TE is a non-invasive imaging method which calculates liver stiffness via measurement of the propagation of an elastic shear wave through the liver substance. TE accurately predicts liver biopsy findings of BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

10 Page of BMJ Open fibrosis and cirrhosis in all major aetiologies of chronic liver disease(). TE was performed by one of three trained nurses, all of whom had performed more than 0 examinations in the hospital prior to study commencement, using the portable Fibroscan FS0 device (Echosens, Paris). The technique for liver stiffness measurement using elastography has been previously described in detail(). The median value of successful measurements, measured in kilopascals (kpa) was representative of liver stiffness. A scan failure was defined as the inability to obtain valid elastography measurements on a single patient. A successful TE result was deemed unreliable if liver stiffness was. kilopascals and the interquartile range/median was greater than 0. as per recent guidance(, ). Patients with a body mass index of greater than kg/m underwent transient elastography at the hospital using the XL probe (FS0 device), due to high rate of scan failures using the standard M probe above this BMI threshold(). All patients, regardless of TE threshold result, received lifestyle advice from the nurses and were given the British Liver Trust Looking After Your Liver leaflet. A TE threshold of.0 kilopascals or greater, which has been previously shown to accurately determine presence of hepatic fibrosis during community screening(), was used to define elevated liver stiffness, and hence clinically significant liver disease. Patients with high liver stiffness results, including high but unreliable acquisitions, were reviewed by a visiting consultant hepatologist in the community. Where appropriate, further investigations including ultrasonography, liver biopsy and enrolment into cirrhosis surveillance programmes, were organised after this consultation. Cirrhosis was definitively diagnosed in all cases based upon established clinical, radiological (including TE result) and/or histological assessment. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

11 Page of Both local cutoffs for elevated ALT (>U/L for women and >U/L for men), and suggested alternative cutoffs from the literature(0) (>U/L for women and >0U/L for men) were compared for identifying elevated liver stiffness and cirrhosis in the study population. Statistics Statistical analysis was performed using SPSS version.0 (IBM). Categorical data are presented as number (percentage). Continuous data are presented as mean (standard deviation (SD)) for parametric data, and medians (range) for non-parametric data. Anthropometric and biochemical data were compared between patients with normal and elevated liver stiffness - continuous variables were compared using the two sample t test for parametric variables, and Mann-Whitney test for non-parametric variables. Categorical variables were compared using chi-squared test, or Fisher s exact test where appropriate. Results Study Population The total patient population of the two participating general medical practices was,, of whom, patients were adults. In total,, adult patients (.%) had alcohol consumption documented, with.% of the total GP population ( patients) meeting our definition for hazardous alcohol use. The adult prevalence of type diabetes was.% (0 patients). Both of these risk factors were found in patients and thus,0 patients were identified for the study. We excluded patients and therefore 0 were invited to the study (see Figure ). Of the excluded BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

12 Page of BMJ Open patients had prior definitive staging of liver disease due to alcohol ( patients), hepatitis B (), NASH (), haemochromatosis () and primary biliary cirrhosis (). Overall 0 patients (.%) underwent the simple blood-based biomarker; the baseline characteristics for these patients compared to all adult patients in the primary care practices is shown in table. Patients who underwent the simple bloodbased biomarker were older than the comparative total adult primary care population, more likely to be male, obese and have ischaemic heart disease. Blood test uptake differed significantly between patients with type diabetes and hazardous alcohol users (.% vs.%, p<0.00). Normal blood-based biomarker results were seen in patients (.%). In total, patients underwent transient elastography, of whom portable M-probe readings were performed in patients (.%), whilst XL probe readings were performed in the remaining patients. It was not possible to obtain a valid liver stiffness measurement in patients (.%), and a further patients (.%) had an unreliable measurement. Chronic Liver Disease Risk Factors In Those With Raised Blood-Based Biomarkers Hazardous alcohol use was present in patients (.0%), type diabetes in patients (.%), raised ALT in patients (.%; includes patients without hazardous alcohol use or type diabetes). Overall, patients (.%) had more than one chronic liver disease risk factor (see Figure ). Baseline characteristics for all patients undergoing TE and the individual risk factor groups are shown in Table. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

13 Page of Significant Liver Disease Detection Valid liver stiffness acquisition was possible in patients (.%). A new diagnosis of clinically significant liver disease was made in patients (.%) with valid TE measurement. This represents a substantial increase in diagnoses for these practices. A significantly greater percentage of patients with type diabetes had elevated liver stiffness compared to hazardous alcohol users (.0% vs.%, p=<0.00) (Figure ). Patients with more than one defined risk factor (n= with successful TE measurement) had a significantly greater percentage of elevated liver stiffness compared with patients with type diabetes alone (.% vs.%, p=0.0) or with hazardous alcohol use alone (.% vs.%, p<0.00). In the subgroup of patients with raised ALT but neither hazardous alcohol use nor type diabetes (Figure ), elevated liver stiffness was observed in / (0.%) patients with raised ALT but neither hazardous alcohol use nor type diabetes. A diagnosis of NASH with early fibrosis was assigned in all cases. Patients with elevated liver stiffness (n=) were older, had a higher BMI and higher prevalence of the metabolic syndrome than those with normal liver stiffness (n=). On laboratory testing, patients with elevated liver stiffness had lower platelet count, higher serum ALT and lower AST:ALT ratio than patients with normal liver stiffness (see Supplementary Table ). Of patients referred for community assessment by the hepatologists, liver biopsy was performed in patients for whom there was diagnostic uncertainty on review of clinical data and TE results. Hepatic fibrosis was confirmed in 0 patients BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

14 Page of BMJ Open undergoing liver biopsy; the remainder all had steatohepatitis (see Supplementary Figure ). Overall, patients were newly diagnosed with liver cirrhosis during the study period based on clinical, radiological and/or histological assessment; of these patients had additional evidence of portal hypertension (see Supplementary Table ). Prior to commencing the study, only patients were known to have cirrhosis in this primary care population. The new observed cirrhosis prevalence of patients after the study period therefore represents a 0% increase compared to before the study; usual care methods have missed the majority of patients with very advanced liver disease. For local ALT cut-offs (>U/L for women and >U/L for men),.% of patients with elevated liver stiffness, 0% with liver fibrosis on biopsy and 0.% with liver cirrhosis had normal ALT levels and would have been missed by standard diagnostic algorithms. Using conservative cutoffs (>U/L for women and >0U/L for men), would still have missed.% of patients with elevated liver stiffness and.% with cirrhosis (see Table ). Discussion Statement of Principal Findings This study uses a novel approach to detect clinically significant but asymptomatic chronic liver disease, based upon a) directly targeting established risk factors and not simply liver function tests and b) the integration of validated noninvasive tests of liver fibrosis and liver specialists within a community setting. The utilisation of a simple, low cost algorithm detected a large number of previously BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

15 Page of undiscovered subclinical but significant chronic liver disease cases, including patients with elevated liver stiffness (.% of all those with successful liver stiffness measurements), and the number of observed cases of liver cirrhosis in the primary care population of, adult patients from which our high risk patients were drawn was increased by 0% compared to before the study. The study highlights the inadequacy of liver function enzymes as a stratification tool in primary care, % of patients with elevated liver stiffness, 0% with liver fibrosis on biopsy and % diagnosed with cirrhosis had normal ALT. In the absence of liver function abnormalities, or specific liver-related symptoms, their disease would have been missed by the standard diagnostic algorithms. We have demonstrated this diagnostic approach is feasible to implement and acceptable to patients, with more than % of investigations occurring in the community setting and non-attendance rates for TE appointments of less than %. This nurse-led TE service, which we originally established in the hospital setting (), has now been implemented in the community with equivalent quality assurance as evidenced by successful TE acquisition in.% and reliable measurements in 0.% of cases. Strengths and Weaknesses of the Study This study is the first to utilise blood biomarkers and Transient Elastography to stratify patients at risk of liver disease in a community population of the United Kingdom. We have utilised a novel diagnostic algorithm, which is rapidly applicable in clinical practice, in a large community population of more than,000 adults. Targeting patients with well-defined risk factors for liver disease resulted in a high pre-test probability of detecting hepatic fibrosis, with.% of patients with TE BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

16 Page of BMJ Open results exceeding the threshold for clinically significant liver disease. This is a far greater diagnostic yield than traditional algorithms using elevated liver enzymes alone to stratify fibrosis risk (, ), and is likely therefore to be a cost efficient approach; formal health economic modelling studies are required to confirm this. The use of simple and inexpensive tests as the initial diagnostic, followed by more specialist tests rationalised resource use and enhanced the feasibility of stratifying the large community population. Moreover, we extrapolated data from hospital-based studies and used conservative thresholds for the diagnostic tests. These prior studies demonstrated a high negative predictive value ( %)(, ) for ruling out significant fibrosis using our chosen fibrosis markers. In a community population, where the disease prevalence is lower than prior hospital-based studies, we would expect their negative predictive value to be higher (albeit in exchange for a reduction in positive predictive value). Thus, we believe it is unlikely that we have missed significant disease within the risk groups that we have stratified. However long term follow up of these patients will be necessary to confirm this. As it is clear that the type of risk factor and presence of multiple risk factors alter the diagnostic yield, a further strength of the pathway is that it can be applied to diverse risk factors, with either the tests or thresholds adjusted depending on the initial risk factor(s), and this improves the generalizability of our pathway to other healthcare regions. However, there are limitations to discuss from our approach. We investigated patients from specific medical practices within a distinct socio-demographic area of the United Kingdom. It is possible that both patient attendance and the detection of clinically significant liver disease may differ elsewhere, and further study of our algorithm in other regions is necessary. The pragmatic study design, both in terms of BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

17 Page of biomarker selection and investigation of selected risk factors for liver disease, means we also cannot formally assess the sensitivity of the algorithm, or the total fibrosis and cirrhosis prevalence in this community population. Alcohol use was only screened in % of this community population, which is lower than the approximately 0% documentation of alcohol intake in primary care practices in the East Midlands region(). We also cannot account for those patients who have underreported their alcohol intake to be within normal limits. Furthermore, additional risk factors such as obesity and the metabolic syndrome, whilst incorporated partially in this pathway (e.g. BARD score) were not specifically included to ensure feasible stratification of the defined at-risk patient groups during the study period. Taken together, the detected prevalence of significant disease, including cirrhosis may represent an underestimate and the presence of cirrhosis in the community is likely to be higher than we report. Also of note, patient uptake of screening was not optimal (% of targeted patients were investigated with the algorithm). However, this is comparable to uptake rates of colorectal cancer screening (, ), and greater than other community liver disease stratification studies where screening uptake is explicitly reported (-). In the UK, patients with diabetes undergo an annual health check which made them more accessible for opportunistic invitation for screening, and this is one explanation for this significant difference in uptake compared to patients with hazardous alcohol use (% vs % undergoing investigation respectively). Lastly, there is scope to improve our diagnostic algorithm further. In total only % of patients were excluded from Transient Elastography by the simple blood-based biomarkers, and future studies using alternative simple serum fibrosis tests or TE alone are warranted to develop an optimal significant liver disease detection algorithm in the community. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

18 Page of BMJ Open Strengths and Limitations In Relation to Other Studies Previous studies of chronic liver disease stratification in the general population support our findings of an appreciable burden of previously undetected chronic liver disease. A study of presumed well army personnel and families in the USA() screened participants with ultrasound and liver biopsy for NAFLD. This approach identified a high prevalence of both NASH (.%) and significant fibrosis (.%). An alternative, non-invasive approach, using transient elastography was employed to screen a defined general population for significant liver disease in France(). TE was performed in, asymptomatic patients over the age of from a well person clinic. An elevated reading, above a validated threshold of kilopascals was found in % of the cohort and cases of cirrhosis were detected. A further study of, patients aged over 0 years and investigated sequentially with Fibrotest (a serum fibrosis marker) and TE detected a fibrosis prevalence of.% and cirrhosis prevalence of 0.%(). Importantly these prior studies used an unselected approach to detect liver disease in the community. We deliberately focussed our diagnostic tests on individuals with risk factors for liver disease. The pre-test probability of detecting disease is increased using this approach and this is evidenced by the fact that % of our cohort undergoing TE had elevated liver stiffness readings compared to % in the Roulot et al study(). A risk factor in combination with the elevated biomarker tests also rationalises the use of liver biopsy with a high yield of clinically significant liver disease. In our study the positive predictive value of selected patients undergoing liver biopsy was 0% for detecting any fibrosis. This compares favourably to the % positive predictive value of finding BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

19 Page of any fibrosis on biopsy following investigation of raised liver enzymes in seronegative individuals(0). A recent study in the UK investigated, patients with elevated liver enzymes in primary care; in those with suspected NAFLD,. % were found to have significant disease based upon an elevated NAFLD fibrosis score(). An important finding from our study is that normal ALT levels, within the local laboratory range, were found in % of patients with elevated liver stiffness and % of patients with cirrhosis. Thus the reliance on abnormal liver enzymes to detect chronic liver disease, which represents standard clinical practice, is inherently flawed. In NAFLD, the presence of significant liver disease in the context of normal enzymes is well documented. The entire spectrum of NAFLD severity has been shown to exist in patients with normal liver enzymes; multiple studies have observed no difference in disease severity between this group and those with abnormal liver enzymes(-). Interestingly, if more conservative thresholds are utilised in our population at the /0U/L cut-off as advocated by the work of Prati et al (0) then a normal ALT is still observed in % of individuals with elevated liver stiffness and % with cirrhosis. In those patients with an elevated ALT, patients (%) did not have type diabetes or hazardous alcohol use as a co-morbidity. Only out of these patients had an abnormal TE reading. Thus, our work strongly supports the concept that the risk for chronic liver disease is better encapsulated by the underlying risk factor as opposed to liver enzyme tests alone. Implications For Clinicians and Policymakers BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

20 Page of BMJ Open The utilisation of a simple, low cost and patient acceptable algorithm in the community resulted in the detection of a large number of cases of previously unidentified chronic liver disease. Whilst the detection of these new cases is currently of unproven benefit, there are likely to be a number of advantages. Early identification of asymptomatic liver disease is likely to reduce future liver-related morbidity and mortality. Interventions for lifestyle-related disease risk factors form established treatments (such as Brief Interventions for alcohol, or dietary and exercise advice), can be performed in a community setting(, ) and are cost effective (, ). A decision analysis study which modelled early identification of NAFLD and subsequent lifestyle or medical intervention showed a % relative risk reduction in liver-related mortality at years, compared with a strategy of not investigating these patients(). Furthermore detection of compensated cirrhosis whilst patients are asymptomatic and ambulatory, with enrolment into surveillance programmes, is of proven benefit to reduce mortality and morbidity (-), and also reduces the expensive healthcare utilisation of the decompensated state. Secondly, early identification of chronic liver disease is likely to be cost effective compared to traditional investigation algorithms. The investigation of liver disease in specialist centres can be both expensive for the healthcare system and inconvenient to the patient. Recent economic modelling from the UK has suggested that the substitution of elastography for liver biopsy in a hospital setting would save 0 (approximately $0) per patient investigated(). Implementing these services in the community is likely to improve the health economic benefits of early liver disease detection yet further, although future health-economic modelling studies are required to confirm this. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

21 Page 0 of Lastly, there are no current medical treatments with widely proven efficacy to prevent fibrosis progression in either alcohol-related liver disease or NAFLD. Novel investigation algorithms, such as the one we have presented, identify a representative population of patients with clinically significant liver disease in whom appropriate therapies may be tested and developed. These, in conjunction with wider adoption of health-related lifestyle changes, may enable reduction of disease progression, symptomatic liver disease, and liver disease related deaths at a general population level. Conclusions We have utilised a novel diagnostic approach in the general population by identifying patients with chronic liver disease risk factors and stratifying using noninvasive fibrosis biomarkers. This algorithm was feasible to implement and resulted in increased detection of asymptomatic but clinically significant liver disease compared to traditional liver-enzyme based algorithms, including a 0% increase in observed cirrhosis prevalence. Importantly, the majority of patients diagnosed with significant liver disease had normal liver function enzymes. Future studies are warranted to ensure the generalizability of our findings to other community populations, to identify the optimal biomarker algorithm for detecting clinically significant liver disease, and subsequently the health-economic benefits of this approach compared to standard care. 0 BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

22 Page of BMJ Open Acknowledgements The authors would like to thank all of the doctors and staff at the participating General Practices for their help in implementing the community investigation algorithm. The authors would also like to acknowledge and thank the staff of the Queen s Day Case Unit, Queens Medical Centre, Nottingham for their help in coordinating the algorithm and performing community transient elastography. Conflicts of Interest We have read and understood BMJ Open policy on declaration of interests and declare that we have no competing interests. All authors have completed the ICMJE uniform disclosure form at and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work. Funding Source and NIHR Disclaimer Internal funding for study was provided by the NIHR Nottingham Digestive Diseases Biomedical Research Unit, part of the University of Nottingham and Nottingham University Hospitals NHS Trust. The study sponsor is the University of Nottingham, who are data custodians but had no role in the design, analysis or interpretations of the data. All authors declare that they are free from other sources of external funding related to this study. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

23 Page of Authorship Statement: DJ Harman, SD Ryder, MW James, M Jelpke, DS Ottey, EA Wilkes, GP Aithal and IN Guha were involved in the study design and concept, implementation of the study in primary care, interpretation of results and editing of the manuscript. Additionally DJ Harman analysed the data set and wrote the initial manuscript draft. TR Card interpreted results and critically revised and edited the draft manuscript. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. Moreover, all authors approve the final version of the article, including the authorship list. GP Aithal and IN Guha are guarantors. NIHR Disclaimer This article/paper/report presents independent research funded by the National Institute for Health Research (NIHR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. Trial registration The diagnostic algorithm utilised for this study can be found on clincialtrials.gov (NCT00), and is part of a continuing longitudinal cohort study. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

24 Page of BMJ Open References. Ratib S, Fleming KM, Crooks CJ, et al. and year survival estimates for people with cirrhosis of the liver in England, -00: a large population study. J Hepatol. 0 Feb;0():-. PubMed PMID:. Epub 0//. eng.. Donnan PT, McLernon D, Dillon JF, et al. Development of a decision support tool for primary care management of patients with abnormal liver function tests without clinically apparent liver disease: a record-linkage population cohort study and decision analysis (ALFIE). Health technology assessment. 00 Apr;():iii-iv, ix-xi, -. PubMed PMID:.. Armstrong MJ, Houlihan DD, Bentham L, et al. Presence and severity of nonalcoholic fatty liver disease in a large prospective primary care cohort. J Hepatol. 0 Jan;():-0. PubMed PMID: 0. Epub 0/0/. eng.. Bhala N, Aithal G, Ferguson J. How to tackle rising rates of liver disease in the UK. Brit Med J. 0 Feb ;. PubMed PMID: ISI: English.. Murray CJ, Richards MA, Newton JN, et al. UK health performance: findings of the Global Burden of Disease Study 0. Lancet. 0 Mar ;():-0. PubMed PMID:. Epub 0/0/. eng.. Myers RP, Fong A, Shaheen AA. Utilization rates, complications and costs of percutaneous liver biopsy: a population-based study including biopsies. Liver Int. 00 May;():0-. PubMed PMID:. Epub 00/0/. eng.. Nyblom H, Berggren U, Balldin J, et al. High AST/ALT ratio may indicate advanced alcoholic liver disease rather than heavy drinking. Alcohol Alcoholism. 00 Jul- Aug;():-. PubMed PMID: ISI: English.. Nyblom H, Nordlinder H, Olsson R. High aspartate to alanine aminotransferase ratio is an indicator of cirrhosis and poor outcome in patients with primary sclerosing cholangitis. Liver Int. 00 Jun;():-. PubMed PMID: ISI: English.. Giannini E, Risso D, Botta F, et al. Validity and clinical utility of the aspartate aminotransferase-alanine aminotransferase ratio in assessing disease severity and prognosis in patients with hepatitis C virus-related chronic liver disease. Arch Intern Med. 00 Jan ;():-. PubMed PMID:. Epub 00/0/. eng.. Harrison SA, Oliver D, Arnold HL, et al. Development and validation of a simple NAFLD clinical scoring system for identifying patients without advanced disease. Gut. 00 Oct;():-. PubMed PMID: ISI: English.. McPherson S, Stewart SF, Henderson E, et al. Simple non-invasive fibrosis scoring systems can reliably exclude advanced fibrosis in patients with non-alcoholic fatty liver disease. Gut. 0 Sep;():-. PubMed PMID: ISI: English.. McPherson S, Anstee QM, Henderson E, et al. Are simple noninvasive scoring systems for fibrosis reliable in patients with NAFLD and normal ALT levels? Eur J Gastroenterol Hepatol. 0 Jan. PubMed PMID:. Epub 0/0/. Eng.. Yoneda M, Imajo K, Eguchi Y, et al. Noninvasive scoring systems in patients with nonalcoholic fatty liver disease with normal alanine aminotransferase levels. J Gastroenterol. 0 Nov. PubMed PMID: 0. Epub 0//. Eng.. Friedrich-Rust M, Ong MF, Martens S, et al. Performance of transient elastography for the staging of liver fibrosis: a meta-analysis. Gastroenterology. 00 Apr;():0-. PubMed PMID: 0. Epub 00/0/0. eng.. Foucher J, Chanteloup E, Vergniol J, et al. Diagnosis of cirrhosis by transient elastography (FibroScan): a prospective study. Gut. 00 Mar;():0-. PubMed PMID: 00. Pubmed Central PMCID: 0. Epub 00/0/. eng.. Boursier J, Zarski JP, de Ledinghen V, et al. Determination of reliability criteria for liver stiffness evaluation by transient elastography. Hepatology. 0 Mar;():-. PubMed PMID:. Epub 0/0/. eng. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

25 Page of Echosens. Fibroscan Recommendations - Updated recommendations for a reliable Fibroscan liver stiffness measurement 0. Available from: Myers RP, Pomier-Layrargues G, Kirsch R, et al. Feasibility and diagnostic performance of the FibroScan XL probe for liver stiffness measurement in overweight and obese patients. Hepatology. 0 Jan;():-0. PubMed PMID:. Epub 0/0/0. eng.. Roulot D, Costes JL, Buyck JF, et al. Transient elastography as a screening tool for liver fibrosis and cirrhosis in a community-based population aged over years. Gut. 0 Jul;0():-. PubMed PMID: ISI: English. 0. Prati D, Taioli E, Zanella A, et al. Updated definitions of healthy ranges for serum alanine aminotransferase levels. Ann Intern Med. 00 Jul ;():-. PubMed PMID: 0. Epub 00/0/0. eng.. McCorry RB, Palaniyappan N, Chivinge A, et al. Development and evaluation of a nurse-led transient elastography service for the staging of hepatic fibrosis in patients with suspected chronic liver disease. QJM. 0 Aug;():-. PubMed PMID: 0. Epub 0/0/. eng.. Khadjesari Z, Marston L, Petersen I, et al. Alcohol consumption screening of newlyregistered patients in primary care: a cross-sectional analysis. Br J Gen Pract. 0 Oct;():e0-. PubMed PMID:. Pubmed Central PMCID: 0.. Group UKCCSP. Results of the first round of a demonstration pilot of screening for colorectal cancer in the United Kingdom. BMJ. 00 Jul ;():. PubMed PMID: 0. Pubmed Central PMCID:.. Meissner HI, Breen N, Klabunde CN, et al. Patterns of colorectal cancer screening uptake among men and women in the United States. Cancer Epidem Biomar. 00 Feb;():-. PubMed PMID: WOS: English.. Sheron N, Moore M, Ansett S, et al. Developing a 'traffic light' test with potential for rational early diagnosis of liver fibrosis and cirrhosis in the community. Brit J Gen Pract. 0 Sep;(0):0-. PubMed PMID: ISI: English.. Wong VW, Chu WC, Wong GL, et al. Prevalence of non-alcoholic fatty liver disease and advanced fibrosis in Hong Kong Chinese: a population study using proton-magnetic resonance spectroscopy and transient elastography. Gut. 0 Mar;():0-. PubMed PMID:. Epub 0/0/. eng.. Zelber-Sagi S, Ratziu V, Zvibel I, et al. The association between adipocytokines and biomarkers for nonalcoholic fatty liver disease-induced liver injury: a study in the general population. Eur J Gastroen Hepat. 0 Mar;():-. PubMed PMID: WOS: English.. Williams CD, Stengel J, Asike MI, et al. Prevalence of Nonalcoholic Fatty Liver Disease and Nonalcoholic Steatohepatitis Among a Largely Middle-Aged Population Utilizing Ultrasound and Liver Biopsy: A Prospective Study. Gastroenterology. 0 Jan;0():-. PubMed PMID: ISI: English.. Poynard T, Lebray P, Ingiliz P, et al. Prevalence of liver fibrosis and risk factors in a general population using non-invasive biomarkers (FibroTest). BMC Gastroenterol. 0;:0. PubMed PMID: 0. Pubmed Central PMCID: 0. Epub 0/0/. eng. 0. Skelly MM, James PD, Ryder SD. Findings on liver biopsy to investigate abnormal liver function tests in the absence of diagnostic serology. Journal of Hepatology. 00 Aug;():-. PubMed PMID: ISI: English.. Mofrad P, Contos MJ, Haque M, et al. Clinical and histologic spectrum of nonalcoholic fatty liver disease associated with normal ALT values. Hepatology. 00 Jun;():-. PubMed PMID: ISI: English.. Verma S, Jensen D, Hart J, et al. Predictive value of ALT levels for non-alcoholic steatohepatitis (NASH) and advanced fibrosis in non-alcoholic fatty liver disease (NAFLD). Liver Int. 0 Oct;():-0. PubMed PMID: 0. Epub 0/0/. eng. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

26 Page of BMJ Open Fracanzani AL, Valenti L, Bugianesi E, et al. Risk of severe liver disease in nonalcoholic fatty liver disease with normal aminotransferase levels: a role for insulin resistance and diabetes. Hepatology. 00 Sep;():-. PubMed PMID:. Epub 00/0/0. eng.. Bertholet N, Daeppen JB, Wietlisbach V, et al. Reduction of alcohol consumption by brief alcohol intervention in primary care - Systematic review and meta-analysis. Archives of Internal Medicine. 00 May ;():-. PubMed PMID: ISI: English.. Thoma C, Day CP, Trenell MI. Lifestyle interventions for the treatment of nonalcoholic fatty liver disease in adults: a systematic review. J Hepatol. 0 Jan;():-. PubMed PMID:. Epub 0/0/0. eng.. Fleming MF, Mundt MP, French MT, et al. Benefit-cost analysis of brief physician advice with problem drinkers in primary care settings. Med Care. 000 Jan;():-. PubMed PMID: 0. Epub 000/0/. eng.. Fleming MF, Mundt MP, French MT, et al. Brief physician advice for problem drinkers: long-term efficacy and benefit-cost analysis. Alcohol Clin Exp Res. 00 Jan;():-. PubMed PMID:. Epub 00/0/0. eng.. Gaidos JK, Hillner BE, Sanyal AJ. A decision analysis study of the value of a liver biopsy in nonalcoholic steatohepatitis. Liver Int. 00 May;():0-. PubMed PMID:. Epub 00/0/. eng.. Merkel C, Marin R, Angeli P, et al. A placebo-controlled clinical trial of nadolol in the prophylaxis of growth of small esophageal varices in cirrhosis. Gastroenterology. 00 Aug;():-. PubMed PMID: 000. Epub 00/0/. eng. 0. Imperiale TF, Chalasani N. A meta-analysis of endoscopic variceal ligation for primary prophylaxis of esophageal variceal bleeding. Hepatology. 00 Apr;():0-. PubMed PMID:. Epub 00/0/0. eng.. Thompson Coon J, Rogers G, Hewson P, et al. Surveillance of cirrhosis for hepatocellular carcinoma: a cost-utility analysis. Br J Cancer. 00 Apr ;():-. PubMed PMID:. Pubmed Central PMCID:. Epub 00/0/0. eng.. Consortium YHE. An Economic Evaluation of Ultrasound Elastography in the Diagnosis of Liver Fibrosis 0. Available from: BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

27 Page of Tables Table. Baseline characteristics of all adult patients at studied General Practice sites compared to patients undergoing study Variable All Adult GP Patients (n=,) P Value Age (n%) --0 years - -0 years - -0 years - -0 years - -0 years - -0 years ->0 years 0 (0.%) (.%) 0 (.%) (.%) (.%) (.%) 0 (.%) Patients Undergoing Study* (n=0) 0 (.0%) (.%) (.%) (.%) (.0%) (0.%) (.%) <0.00 < <0.00 <0.00 < Male Gender n(%) (.0%) (.%) <0.00 Body Mass Index n(%) Missing Hazardous Alcohol Use n(%) Type Diabetes n(%) Ischaemic Heart Disease n(%) (.%) 0 (.%) (.0%) 0 (.%) 0 (.%) (.%) <0.00 <0.00 <0.00 (.%) (.0%) < (.%) (.%) <0.00 (.0%) (.%) <0.00 Hypertension n(%) (.%) (.%) <0.00 Hyperlipidaemia n(%) (0.%) (.%) <0.00 *Patients undergoing study refers to all patients having the simple blood biomarker step of stratification. Categorical variables are displayed as n(%) and compared using chi-squared test. P values reaching statistical significance (p 0.0) are shown in bold. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

28 Page of BMJ Open Table : Baseline population characteristics of patients with raised blood biomarker result undergoing transient elastography Variable All Patients Undergoing TE (n=) Hazardous Alcohol Use * (n=) Type Diabetes * (n=) Raised ALT * (any case) (n=) Age Years. (.0). (.). (.).0 (.) Male - n(%) (.%) (.%) 0 (.%) (.%) Body Mass Index (kg/m ). (.). (.) 0.0 (.). (.) Obesity - n(%) 0 (.%) (.%) 0 (.%) (.%) Hypertension - n(%) (.%) (.0%) (.%) (.%) Metabolic Syndrome - n(%) Ischaemic Heart Disease - n(%) Type Diabetes - n(%) Hazardous Alcohol Use - n(%) (.0%) (.%) (.%) (.%) (.0%) (.%) (0.%) (.0%) (.%) 0 (.%) (0%) (0.0%) (.0%) (0%) 0 (.%) (.%) Platelets ( /L) 0. (.) 0. (.). (.0) 0. (.) Creatinine (µmol/l) 0. (.). (0.). (.). (.0) Bilirubin (µmol/l). (.0). (.). (.). (.) Albumin (g/l). (.). (.). (.0).0 (.) ALT (IU/L).0 (.0). (.). (.) 0. (0.) AST (IU/L). (.) 0. (.).0 (.0).0 (.) AST:ALT ratio.0 (0.). (0.).0 (0.) 0. (0.) Numerical variables are displayed as mean (standard deviation(sd)), categorical variables are displayed as n(%). ALT = alanine aminotransferase, AST = aspartate aminotransferase. **patient numbers refer to patients fulfilling the single risk factor alone (see Figure ). BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

29 Page of Table : Serum alanine aminotransferase (ALT) levels of patients with successful liver stiffness acquisition Liver Stratification Normal Liver Stiffness (n=) Elevated Liver Stiffness (n=) Cirrhosis (n=) Raised ALT Local Lab* Normal ALT Local Lab* Raised ALT /0 Normal ALT /0 (.%) (0.%) (.%) 0 (.%) (.%) (.%) (.%) (.%) (.%) (0.%) (.%) (.%) Serum ALT levels of patients with elevated liver stiffness undergoing subsequent percutaneous liver biopsy No Hepatic Fibrosis (n=) Any Hepatic Fibrosis (n=0) (0%) (0%) (0%) (0%) (0%) (0%) (%) (%) Numerical values are displayed as n(%).*locally utilised cutoffs: ALT>IU/L for women and >IU/L for men. Alternative cutoffs: ALT>IU/L for women and >0IU/L for men. Values are displayed as n(%) in all cases. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

30 Page of BMJ Open Supplementary Table : Comparison of Biochemical and Clinical Features Between Patients with Normal and Elevated Liver Stiffness Variable Normal Liver Stiffness (n=) Elevated Liver Stiffness (n=) P Value Age (Years) 0. (.0). (.) 0.0 Male - n(%) (.%) (.%) 0. Body Mass Index (g/m ). (.). (.0) <0.00 Type Diabetes - n(%) (0.%) (0.%) 0.00 Hazardous Alcohol - n(%) (.%) (.%) 0.00 Raised ALT - n(%) (.%) (.%) <0.00 Multiple Risk Factors n(%) (.%) (.%) <0.00 Metabolic Syndrome n(%) (.%) (.0%) <0.00 Ischaemic Heart Disease n(%) (.%) (.%) 0.0 Platelet Count ( /L). (0.). (.) 0.0 ALT (U/L). (.0). (.) <0.00 Bilirubin (µmol/l). (.). (.) 0. Albumin (g/l). (.). (.) 0. AST:ALT Ratio. (0.).0 (0.) 0.0 Numerical variables are displayed as mean(sd) and compared using two sample t test. Categorical variables are displayed as n(%) and compared using chi-squared test. P values reaching statistical significance (p 0.0) are shown in bold. BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

31 Page 0 of Supplementary Table : Laboratory, Imaging and Histopathology Findings of patients diagnosed with cirrhosis during study Liver Disease Risk Factor* Alcohol and TDM Age Liver Stiffness (kpa) Platelet Count ( /L) Ultrasound Abnormality. Splenomegaly (cm) Histopathology Not Performed TDM. Cirrhosis (NASH) TDM. 0 Cirrhosis (NASH) TDM. Cirrhosis Not Performed Alcohol.0 Cirrhosis Not Performed TDM.0 Cirrhosis (NASH) TDM and raised ALT. Cirrhosis, splenomegaly (cm) TDM. Cirrhosis, splenomegaly (cm) Cirrhosis (NASH) Not Performed Endoscopy Abnormality Grade Varices TDM. Not Performed PHG TDM 0. Cirrhosis Not Performed TDM 0 Cirrhosis Not Performed *Patients with alcohol as risk factor were assigned diagnosis of cirrhosis due to alcoholic liver disease. Patients with type diabetes without alcohol excess were assigned diagnosis of cirrhosis due to NASH. PHG = portal hypertensive gastropathy, TDM = type diabetes 0 BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

32 Page of BMJ Open Figure Legends Fig : Diagnostic algorithm and patient flowchart through non-invasive biomarker pathway Fig : Distribution of chronic liver disease risk factors for all patients (n=) undergoing Transient Elastography Examination Fig : Transient Elastography results of patients with successful liver stiffness acquisition (n=) Supplementary Fig : Histopathology findings of patients with elevated liver stiffness undergoing liver biopsy (n=) (Figures are supplied separately in.tiff format). BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

33 Page of x0mm (00 x 00 DPI) BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

34 Page of BMJ Open x0mm (00 x 00 DPI) BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

35 Page of x0mm (00 x 00 DPI) BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

36 Page of BMJ Open x0mm (00 x 00 DPI) BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.

37 Direct targeting of risk factors significantly increases the detection of liver cirrhosis in primary care: A cross-sectional diagnostic study utilising Transient Elastography Journal: BMJ Open Manuscript ID: bmjopen-0-00.r Article Type: Research Date Submitted by the Author: 0-Mar-0 Complete List of Authors: Harman, David; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Ryder, Stephen; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit James, Martin; University of Nottingham, Nottingham Digestive Diseases NIHR Biomedical Research Unit Jelpke, Matthew; NHS Rushcliffe Clinical Commissioning Group, Ottey, Sean; NHS Rushcliffe Clinical Commissioning Group, Wilkes, Emilie; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Card, Tim; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Aithal, Guruprasad; University of Nottingham, NIHR Nottingham Digestive Diseases Biomedical Research Unit Guha, Neil; University of Nottingham, <b>primary Subject Heading</b>: Secondary Subject Heading: Diagnostics Keywords: Gastroenterology and hepatology Adult gastroenterology < GASTROENTEROLOGY, Hepatobiliary disease < GASTROENTEROLOGY, GENERAL MEDICINE (see Internal Medicine) BMJ Open: first published as./bmjopen-0-00 on May 0. Downloaded from on May 0 by guest. Protected by copyright.