The Liver Cirrhosis Network Cohort Study: Cirrhosis Definition, Study Population, and Endpoints

Authors: Elliot B. Tapper, David Goldberg, Neehar D. Parikh, Norah A. Terrault, Nicole Welch, Suzanne Sharpton, Bilal Hameed, Mandana Khalili, Andrew Stolz, Elizabeth C. Verna, Robert S. Brown, Arun J. Sanyal, Lisa VanWagner, Daniela P. Ladner, Cynthia A. Moylan, Anna Mae Diehl, Patricia D. Jones, Rohit C. Loomba, Srinivasan Dasarathy, Douglas A. Simonetto, Vijay H. Shah, Jasmohan S Bajaj

Abstract

Background:

One of the primary goals of the Liver Cirrhosis Network (LCN) is to develop a cohort study to better understand and predict the risk of hepatic decompensation and other clinical and patient-reported outcomes among patients with Child A cirrhosis.

Methods:

The LCN consists of a Scientific Data Coordinating Center (SDCC) and 10 clinical centers whose investigators populate multiple committees. The LCN Definitions and Measurements Committee developed preliminary definitions of cirrhosis and its complications by literature review, expert opinion, and reviewing definition documents developed by other organizations. The Cohort Committee developed the study protocol with the input of the steering committee.

Results:

The LCN developed a prospective cohort study to describe and predict the rates of incident clinical events pertaining to first decompensation and patient reported outcomes. The LCN developed a pragmatic definition of compensated cirrhosis incorporating clinical, laboratory, imaging, and histological criteria. Definitions of incident and recompensated ascites, overt hepatic encephalopathy, variceal hemorrhage, bleeding due to portal gastropathy, and hepatocellular carcinoma were also codified.

Conclusion:

The LCN Cohort Study design will inform the natural history of cirrhosis in contemporary patients with compensated cirrhosis. The LCN Definitions and Measures Committee developed criteria for the definition of cirrhosis to standardize entry into this multi-center cohort study and standardized criteria for liver-related outcome measures. This effort has produced definitions intended to be both sensitive and specific as well as easily operationalized by study staff such that outcomes critical to the LCN cohort are identified and reported in an accurate and generalizable fashion.

Introduction

The majority of patients with cirrhosis have compensated disease, lacking complications such as ascites, hepatic encephalopathy (HE), a history of variceal hemorrhage, or hepatocellular carcinoma (HCC).(1) As the development of these complications accounts for most of the public health burden of cirrhosis, there is an urgent need to predict and prevent the progression to decompensated cirrhosis. The epidemiology of chronic liver disease has shifted away from chronic viral hepatitis toward predominantly alcohol-associated liver disease and metabolic dysfunction associated steatohepatitis (MASH), and thus patients who are older and have a higher burden of cardiometabolic comorbidities.(2) These factors influence the risk of competing extra-hepatic adverse events and may alter the patterns and risk of decompensation events.(3, 4) To better stratify patients and understand their progression along a continuum of disease severity, longitudinal cohort studies and clinical trials focused on preventing decompensation in patients with cirrhosis require standardized and translatable definitions of compensated and decompensated cirrhosis and related complications.

In August of 2021, The National Institutes of Health (NIH) funded and established the Liver Cirrhosis Network (LCN) with the aims (1) developing a prospective, longitudinal cohort study of patients with compensated cirrhosis in the US, and (2) performing a clinical trial to evaluate statins in preventing disease progression. Herein, we describe the population, key definitions and measures, and outcomes explored by the LCN Cohort Study.

The Cohort Study

The goals of the LCN Cohort Study are to better understand and predict the risk of hepatic decompensation and other clinical and patient-reported outcomes among patients with compensated cirrhosis. The structure of the LCN consists of a Scientific Data Coordinating Center (SDCC) and 10 clinical centers whose investigators populate several committees, including the Definitions and Measurements Committee, the Cohort Committee, and the Trial Committee, among others. Study sites and other details can be found at https://www.lcnstudy.org/about-the-network/. A Steering Committee comprised of all clinical center principal investigators, NIH representatives, and SDCC representatives provided feedback with additional input from the Imaging and Radiology Working Group (a sub-committee within the LCN). The LCN Cohort Study is a prospective observational study with a target enrollment of 1200 participants designed with the primary objective to evaluate improvements in predictive accuracy for the risk of decompensation and develop a composite risk score that includes non-modifiable and modifiable clinical, pathophysiological and behavioral risk factors. The focus is to improve prediction of time-to-(first)-decompensation using standard of care measures and determine the added value of additional assessments such as cognitive testing, patient-reported outcomes, frailty, and transient elastography in this prediction. The protocol can be accessed at ClinicalTrials.gov (NCT05740358) and the synopsis is provided in the Supplement.

Population

The cohort study includes adults (≥18 years) with compensated cirrhosis who were willing to provide samples at baseline and who are willing to be followed for three years. Key exclusion criteria were selected to focus on stable compensated cirrhosis (where the data gap for risk prediction is greatest), to ensure that underlying etiologies of liver disease are under reasonable control, and free of extrahepatic life-limiting illnesses.(Table 1) Alcohol as the etiology is determined based on the treating hepatologist’s diagnosis. However, alcohol will also be measured using validated questionnaires (AUDIT-C and the timeline follow-back) and phosphatidylethanol testing. The LCN has an intentional approach towards diverse patient recruitment and uses multiple language translations for its materials. The LCN also has a website https://www.lcnstudy.org/for-patients/ that is frequently accessed by patients outside the 10 centers with the opportunity for patients to request the opportunity for enrollment. Finally, many centers tailor recruitment to facilitate enrollment at satellite and community sites.

Definitions

The LCN Definitions and Measurements Committee’s goal was to provide definitions that were sensitive and specific, would be easily operationalized by study staff, and would provide a homogenous/comparable population between different sites. Recognizing that, in the course of clinical care, complications are diagnosed with variable degrees of certainty, we applied a schema of ‘definite,’ ‘highly likely,’ and ‘probable’ to allow for post-hoc sensitivity analyses consistent with definitions from other consortia.(5)

The LCN determined that the definition of cirrhosis must account for the multiplicity of methods used to support this clinical diagnosis while optimizing both specificity and sensitivity. Liver biopsy is highly specific yet can be interpreted incorrectly and is often deferred in the context of other supporting evidence of cirrhosis.(6) Although considered a “gold standard”, it is infrequently used given the availability of noninvasive measures and patient preferences. Therefore, defining and diagnosing cirrhosis in clinical studies requires alternative strategies. Whereas cirrhosis is more certain amongst those with current or past decompensations, the absence of decompensating events makes clinical criteria for defining compensated cirrhosis much more difficult. In this context, the methods used to diagnose cirrhosis are varied. Ultrasound and cross-sectional liver imaging has sensitivity ≤85%(7, 8)) and specificity of 75–95%(8) based on the surface features with low inter-rater reliability.(9) Although a nodular liver is common, additional features are required to increase the accuracy of imaging-based cirrhosis diagnoses as liver nodularity can be seen in conditions such as congestion and non-cirrhotic chronic liver disease (e.g., nodular regenerative hyperplasia). Elastography can be more accurate,(9) but cutoff values require added care in selection since they vary with underlying etiologies owing to the variable burden of inflammation and cholestasis related to different liver disorders. Elastography alone has poor reported positive predictive values at levels consistent with statistically optimal cutoffs (i.e., 12.5 Kilopascals, kPa, for transient elastography)(6) while higher values felt to be clearly suggestive of cirrhosis (i.e., 20 kPa) unacceptably reduce negative predictive value.(10) Thrombocytopenia (platelet count <150,000 per μL) and/or high Fibrosis-4 (FIB-4 >2.67) index are strongly suggestive of cirrhosis, but are inadequately predictive when considered in isolation.(6) Sequential or paired testing of two or more modalities improves the accuracy of non-invasive testing.(6, 11–13) For these reasons, the LCN collectively decided that patients lacking biopsy-proven cirrhosis are required to have at least two non-invasive tests to make a diagnosis for inclusion for entry into the LCN cohort. The LCN developed criteria with an aim of erring on the side of greater fibrosis (i.e., higher FIB-4) and combinations of non-histological criteria to minimize false positives such that all patients enrolled indeed have cirrhosis. Efforts to increase certainty in the diagnosis of cirrhosis reduce the heterogeneity of the population by lowering the chance that individuals without cirrhosis are recruited. In the absence of a biopsy suggestive of cirrhosis within 5 years, at least two of the following five measures were (i) conventional imaging, (ii) elastography (≥12.5kPa by transient elastography or ≥5.0 kPa by magnetic resonance elastography), (iii) biopsy >5 years ago suggestive of cirrhosis, or (iv) varices seen on endoscopy or imaging, and (v) FIB-4>2.67 or platelet count <150,000 per μL. Of special note, conventional imaging features must include nodular liver (which is inadequate alone) plus signs of portal hypertension including either splenomegaly or recanalized umbilical vein. The decision to require additional features for biopsies >5 years old is based on data demonstrating regression of fibrosis with etiologic control. The precise cutoffs and timing of measures supporting a diagnosis of compensated cirrhosis were refined iteratively during the launch of the study and the final version is summarized in Table 2.

We acknowledge that there has been a recent multi-society endorsement of a nomenclature change from Nonalcoholic Fatty Liver Disease (NAFLD) and Nonalcoholic Steatohepatitis (NASH) to metabolic dysfunction associated steatotic liver disease (MASLD) and MASH.(14) While this change that may impact of future of the study of this entity, there is strong evidence that old and new definitions overlap without risk consequence for study decisions which were based on prior literature which utilized the previous NAFLD definition.(15)

Primary outcomes

The primary outcome of the cohort study is time from consent to decompensation. Participants complete standardized study visits every 6 months according to the study protocol. Definitions of decompensating events below are made from the perspective that these would be the (1) first decompensating event, and (2) would be robust enough to be captured remotely, through telephone contact, or using chart review in between the in-person visits. (Table 3) These are also standardized to diagnose subsequent and multiple combinations of decompensating events as the study progresses. We did not include pulmonary portal hypertensive complications due to their relative rarity vis-à-vis the complications discussed below. We also provide definitions for recompensation where appropriate.(Table 4) The network developed a standardized adjudication process that will be reported in primary dissemination materials. All events of interest, including decompensating events, will be independently adjudicated by at minimum 2 network investigators that do not belong to the site at which the event occurred according to the definitions below.

Specific complications that define

Ascites

We developed our definition of ascites to reflect the development of clinically meaningful free fluid in the peritoneum (and not simply trace free fluid seen on imaging) that reflected a change in clinical status. We initially based this on prior and current American Association for the Study of Liver Disease (AASLD) guidance,(16, 17) and we further specified terms required for data capture and inclusion of outcomes that were medically managed. As abdominal imaging is standard of care for hepatocellular carcinoma (HCC) screening, we considered the potential that patients would have incidentally detected low volume perihepatic or pelvic fluid that is not of clinical significance and does not reflect a significant change in the patient’s status. We considered specifying a grade of ascites as an endpoint, referring to the recent AASLD guidance on ascites grading.(17) However, each grade is defined in part by response to therapy over a period of months, which renders grading a retrospective classification. We therefore defined ascites as free fluid which requires initiation of diuretics and/or a paracentesis procedure. Definitions of recompensation are provided with certainty levels commensurate with the time since withdrawal of therapy. (Table 4)

Overt hepatic encephalopathy (HE)

Overt HE is a clinical diagnosis. While many processes such as sepsis, drug adverse events, alcohol and other substance misuse-related syndromes, and nutritional impairment can confound the etiologic interpretation of an episode of encephalopathy,(18) the response to HE-directed therapy is an essential determinant of the etiology of hepatic encephalopathy. The International Society for Hepatic Encephalopathy and Nitrogen Metabolism (ISHEN) consensus defined overt HE as grade 2 or higher as defined by asterixis, disorientation, or inappropriate behavior.(19) This was also subsequently used by AASLD/EASL (European Association for the Study of the Liver) guidelines as the definition for overt HE.(20, 21) We operationalized this further One or more episodes of acute disorientation (unaware of person, place, or time; somnolence; coma) with resolution after HE-directed therapy (lactulose or rifaximin) documented by a gastroenterologist or hepatologist as overt HE. Many patients, however, are managed without expert consultation(22) and often HE therapy is initiated on the basis of a report detailing an episode consistent with HE. Applying a grading of definite, highly likely, or probable, we therefore require confirmation of HE by a medical professional and would exclude those situations where HE is suspected based on report from family members or caregivers. Definitions of recompensation are provided with certainty levels commensurate with the time since withdrawal of therapy. (Table 4)

Portal Hypertensive bleeding (variceal hemorrhage and portal gastropathy-associated bleeding):

Variceal hemorrhage is often intermittent and can be temporarily controlled with vasoactive medications. Patients may or may not have active bleeding at endoscopy or even high-risk stigmata such as red wale or platelet plugs (“nipple sign”). Accordingly, the definition of variceal hemorrhage must account for a range of presentations. We therefore applied a grading of definite, highly likely, or probable to address three scenarios with the certainty of attribution variable dependent on the findings at endoscopy. The current definition accounts for uncertainties by specifying key measures for data capture and focusing on two types of bleeding. As bleeding from portal gastropathy may be harder to ascertain, we define definite portal gastropathy-associated bleeding as “hematemesis and/or melena and endoscopy performed with 24 hours of admission demonstrating moderate or severe mosaic like-pattern (e.g., discrete cherry red spots and/or diffuse hemorrhagic gastropathy(23)) with active bleeding/oozing.”(24, 25) Scenarios accounting for lesser degrees of certainty are described in Table 3. Recompensation after variceal hemorrhage was not considered as this typically requires lifelong prophylactic therapy.

Secondary Outcomes

Secondary outcomes include the number of decompensating events, all-cause mortality, adjudicated liver-related mortality, all-cause hospitalizations, liver transplantation, hepatocellular carcinoma (defined using standard criteria(26–29)), mesenteric thrombosis, health-related quality of life (defined using the PROMIS-29+2)(30), liver stiffness (using transient elastography), cognitive function (using the EncephalApp Stroop)(31, 32), and frailty (using the Liver Frailty Index). (32)

Conclusion

The LCN cohort study aims to define the contemporary risk of decompensation for patients with compensated cirrhosis as well as characterize multiple facets of the patient experience and phenotype to increase reliability of the diagnoses and outcomes for this cohort study rather than general clinical practice. The definitions provided here support the aims of the LCN cohort study by 1) allowing for the enrollment of patients with clinically diagnosed cirrhosis while preserving rigor through standardized criteria and 2) empowering the adjudication of incident liver-related outcomes among patients receiving standard clinical care.

Supplementary Material