In Practice with Dr. Denver: Rethinking Liver Fibrosis Screening in the General Population: CLDI vs FIB-4

Introduction:

• Liver disease mortality in the UK has risen by more than 400% since 1970¹, in contrast to other common chronic diseases such as heart disease, lung disease and diabetes.
• Despite this burden, there is no UK national screening programme for liver disease or liver cancer – unlike breast, bowel or AAA screening.
• This matters because liver disease is typically silent and progressive with many patients first coming to medical attention during an emergency hospital admission, by which time cirrhosis is often already advanced and prognosis is poor².
• Earlier identification allows effective treatment to be initiated in time to improve patient outcomes.
• This urgency is now matched with opportunity as there are new emerging therapies for MASH on the horizon. Health care systems need to be equipped with the right tools and pathways to select the right patient cohorts to benefit.

Population Screening:

• In population screening, it is essential to optimise the balance between test sensitivity and specificity to minimise false-negative and false-positive results³.
• The overall discriminatory performance of a screening test is commonly assessed using the area under the receiver operating characteristic (ROC) curve (AUROC or AUC), where a value of 1.0 indicates perfect discrimination and a value of 0.5 indicates performance no better than chance⁴.
• Conventional liver tests, such as serum aminotransferases, have poor sensitivity and specificity for identifying cirrhosis, and a liver biopsy is too invasive for a screening test at population level.
• Other non-invasive fibrosis markers such a Fibrosis-4 Index (FIB-4), present a challenge when applied at population level.
• Current UK, guidance keeps FIB-4 as the first line filter but proposes applying it only to higher-risk groups and acting on a low threshold as a way of ruling out lower risk patients.
• FIB-4 was developed and validated in populations with established or suspected chronic liver disease (Hepatitis C/ HIV co-infection)⁵ rather than in asymptomatic individuals from the general population.
• As a result, its diagnostic accuracy declines when applied to unselected populations, where the prevalence of advanced fibrosis is substantially lower.
• Performance of FIB-4 is strongly influenced by age. Age is a component of the FIB-4 equation, causing scores to increase independently of fibrosis. Consequently, older adults are more likely to be classified as intermediate- or high-risk despite not having advanced fibrosis, whereas younger patients may have falsely reassuring low scores.
• Population-based studies have also shown that many individuals with elevated liver stiffness have FIB-4 values below the low-risk threshold. In one multinational study by Graupera et al., approximately 43% of participants with liver stiffness ≥8 kPa had a normal FIB-4 score⁶, indicating an unacceptable false-negative rate for a screening programme.
• The ROC-AUC for the general population cohort for identifying elevated liver stiffness (≥8 kPa, used as a threshold for possible significant fibrosis) was 0.572⁶.
• The performance of FIB-4 has shown low sensitivity with a significant proportion of false negatives when applied to at-risk subgroups.
• In a meta-analysis of 12 studies (5,624 patients), pooled sensitivity in T2DM cohorts was only 0.74⁷ — meaning roughly 1 in 4 patients with true advanced fibrosis were mis-classified as low risk.

Cumulative Liver Damage Index:

• In Somerset, a prospective study (LiveWell) tested whether longitudinal blood test data combined with polygenic risk score could identify clinically significant fibrosis from the general population.
• A case-finding tool called hepatoSIGHT® was implemented, which uses the Cumulative Liver Damage Index (CLDI) ⁸ – a longitudinal integral of routine ALT – to risk stratify the population, based on existing blood tests.
• CLDI significantly predicted F2+ fibrosis (OR 3.87, p < 0.001) and F4 cirrhosis (OR 5.02, p<0.01).
• The combined- cohort hepatoSIGHT model (age, sex, BMI, CLDI, platelet mean) achieved a ROC-AUC 0.884 (95% CI 0.79 – 0.967)⁹ outperforming other standard assessment approaches at population level.
• Further validation of this approach is in progress across other sites in the South West region.

Conclusion

• The UK’s rising liver disease mortality — up over 400% since 1970, with half of deaths in working-age adults — combined with the absence of a national screening programme, means the case for earlier detection has never been stronger, particularly as new MASH therapies create a real opportunity to intervene before cirrhosis develops.
• FIB-4 was developed and validated in patients with established chronic liver disease, not asymptomatic individuals in the general population and does not work as a whole-population screening test.
• Even in preselected high-risk cohorts, FIB-4 approximately 25- 33% of patients.
• The Somerset LiveWell study offers a promising alternative: the hepatoSIGHT tool, which utilises the Cumulative Liver Damage Index (CLDI) and combined with demographic and longitudinal routine biomarker variables (age, sex, BMI, platelet count), achieved a ROC-AUC of 0.884 — substantially outperforming FIB-4 at population level, and significantly predicting both F2+ fibrosis and F4 cirrhosis.
• With further validation of hepatoSIGHT now underway across other South West sites, this approach represents a meaningful step toward a screening strategy for population-level liver screening.

References:

1. British Liver Trust (2024). Liver Disease in Numbers – Key Facts and Statistics. [online] British Liver Trust. Available at: https://britishlivertrust.org.uk/information-and-support/statistics/.
2. Williams, R., Aithal, G., Alexander, G.J., Allison, M., Armstrong, I., Aspinall, R., Baker, A., Batterham, R., Brown, K., Burton, R., Cramp, M.E., Day, N., Dhawan, A., Drummond, C., Ferguson, J., Foster, G., Gilmore, I., Greenberg, J., Henn, C., Jarvis, H., Kelly, D., Mathews, M., McCloud, A., MacGilchrist, A., McKee, M., Moriarty, K., Morling, J., Newsome, P., Rice, P., Roberts, S., Rutter, H., Samyn, M., Severi, K., Sheron, N., Thorburn, D., Verne, J., Vohra, J., Williams, J. and Yeoman, A. (2020). Unacceptable failures: The final report of the lancet commission into liver disease in the UK. The lancet, 395(10219), pp.226–239. doi:10.1016/s0140-6736(19)32908-3.
3. Principles, methods, applications and organisation of screening for early detection, prevention, treatment and control of disease | health knowledge. (2017). [online] org.uk. Available at: https://www.healthknowledge.org.uk/public-health-textbook/disease-causation-diagnostic/2c-diagnosis-screening/principles-methods-applications?utm_source=chatgpt.com.
4. Ja, H. and Bj, M. (1982). The meaning and use of the area under a receiver operating characteristic (ROC) curve. [online] Radiology. Available at: https://pubmed.ncbi.nlm.nih.gov/7063747/.
5. Castera, L., Friedrich-Rust, M. and Loomba, R. (2019). Noninvasive assessment of Liver disease in Patients with nonalcoholic fatty liver disease. Gastroenterology, [online] 156(5), pp.1264–1281.e4. doi:10.1053/j.gastro.2018.12.036.
6. Graupera, I., Thiele, M., Serra-Burriel, M., Caballeria, L., Roulot, D., Wong, G.L.-H., Fabrellas, N., Guha, I.N., Arslanow, A., Expósito, C., Hernández, R., Aithal, G.P., Galle, P.R., Pera, G., Wong, V.W.-S., Lammert, F., Ginès, P., Castera, L. and Krag, A. (2021). Low accuracy of FIB-4 and nafld fibrosis scores for Screening for Liver fibrosis in the population. Clinical gastroenterology and hepatology, [online] 0(0). doi:10.1016/j.cgh.2021.12.034.
7. Han, J.W., Kim, H.Y., Yu, J.H., Kim, M.N., Chon, Y.E., An, J.H., Jin, Y.-J., Choi, M., Kim, S.U., Lee, H.A. and Jun, D.W. (2024). Diagnostic accuracy of the Fibrosis-4 index for advanced liver fibrosis in nonalcoholic fatty liver disease with type 2 diabetes: A systematic review and meta-analysis. Clinical and molecular hepatology, [online] 30(Suppl), pp.S147-S158. doi:10.3350/cmh.2024.0330.
8. Wesley, E., Matull, W.-R., Kitchin, A., Hutchison, K., Madge, S. and Jobson, T. (2022). O06 a cumulative liver damage index (CLDI) identifies patients at risk of significant liver disease. Abstracts, [online] pp.A4-A4. doi:10.1136/gutjnl-2022-basl.6.com.
9. Jobson, T., Denver, C., Owen, C., Stevens, N., Caudrillier, A., Hamadache, Y., O’Brien, H., Short, P., Webb, J. and Barnes, K. (2026). LBP-017 Longitudinal analysis of routine laboratory data prospectively identifies clinically significant hepatic fibrosis from the general population, outperforming standard fibrosis scores: Validation from the LiveWell multi-modal screening study. Journal of Hepatology, [online] 84(Suppl 1), p.S75. doi:10.1016/S0168-8278(26)00432-0.