Evidence-based criteria for identifying at-risk individuals requiring liver disease screening.

Abstract

Background: Liver fibrosis screening is recommended in at-risk groups, but a clear definition of "at risk" for entry criteria is lacking. We analyzed different combinations of established risk factors to define specific screening entry criteria with a prespecified sensitivity requirement.

Methods: Data regarding individuals aged 40-70 years from Finnish health-examination surveys (FINRISK 2002-2012 and Health 2000, n=15,057) and the UK Biobank (n=454,990) were linked with healthcare registries for liver cirrhosis-related events (LREs; liver-related hospitalizations, cancer, or death). The predictive performance of 1919 combinations of risk factors, including alcohol consumption, metabolic disturbances, abnormal liver function tests, and Chronic Liver Disease risk score, was assessed for 10-year LRE risk requiring a minimum 90% sensitivity. Validations were performed using liver stiffness measurement (LSM) >12 kPa in the NHANES 2017-2020 sample (n=3367).

Results: Optimal entry criteria for predicting 10-year LRE risk with >90% sensitivity included any one of: hazardous alcohol use, severe obesity, metabolic syndrome, an AST-to-ALT ratio >0.8 with elevated ALT, and an intermediate-to-high Chronic Liver Disease risk score. The sensitivity and specificity for this strategy were 91% and 51% for LREs, respectively, in the Finnish cohort, and 91% and 41% for LSM >12 kPa in the US sample. In the US sample, applying these entry criteria followed by fibrosis-4 ≥1.3 for predicting LSM >12 kPa reduced the sensitivity to 45% (specificity: 85%), which was attributed to the suboptimal sensitivity of fibrosis-4.

Conclusions: This study identifies an inexpensive risk factor-based strategy with >90% sensitivity for predicting LRE and LSM >12 kPa, which is practical and scalable for targeted liver fibrosis screening to improve population outcomes. However, a more sensitive first-line noninvasive fibrosis test is needed.