A new study from Karolinska Institutet, published in the scientific journal The BMJ, shows how a simple blood analysis can predict the risk of developing severe liver disease. The method may already start to be applied in primary care to enable the earlier detection of cirrhosis and cancer of the liver.
"These are diseases that are growing increasingly common and that have a poor prognosis if detected late," says Rickard Strandberg , affiliated researcher at Karolinska Institutet's Department of Medicine, Huddinge , who has developed the test with his departmental colleague Hannes Hagström. "Our method can predict the risk of severe liver disease within 10 years and is based on three simple routine blood tests."
For the study, the researchers at Karolinska Institutet and their colleagues in Finland evaluated how well the method can estimate the risk of severe liver disease. The model, which is called CORE, was produced with advanced statistical methods and is based on five factors: age, sex and levels of three common liver enzymes (AST, ALT and GGT), which are commonly measured during regular health checks.
A web-based calculator
Their aim has been to produce a tool that is easy to use in primary care, where most patients first seek medical attention. A web-based calculator is already available for doctors and nurses at www.core-model.com .
"This is an important step towards being able to offer early screening for liver disease in primary care," says principal investigator Hannes Hagström , adjunct professor at the Department of Medicine, Huddinge, and senior consultant at Karolinska University Hospital. "Drug treatment is now available, soon hopefully also in Sweden, for treating people at a high risk of developing liver diseases such as cirrhosis or liver cancer."
The study is based on data from over 480,000 people in Stockholm who underwent health checks between 1985 and 1996. On following the participants for up to 30 years, the researchers could see that some 1.5 per cent developed severe liver disease, such as liver cirrhosis and liver cancer, or required a liver transplant.
Highly accurate risk prediction
The CORE model proved highly accurate and was able to differentiate between people who either did or did not develop the disease in 88 per cent of cases, which is an improvement on the currently recommended FIB-4 method.
"Primary care hasn't had the tools to detect the risk of severe liver disease in time," says Professor Hagström. "FIB-4 is not suited for the general population and is less effective at predicting the future risk of severe liver disease."
The model was also tested on two other population groups in Finland and the UK, where it again demonstrated a high accuracy in predicting this risk. The researchers make the point, however, that it needs to be further tested on groups at especially high risk, such as people with type 2 diabetes or obesity. They also recognise a need to integrate the model into medical records systems to facilitate its clinical use.
The study was a collaboration between Karolinska Institutet, Helsinki University Hospital, Helsinki University and the Finnish Institute for Health and Welfare. It was financed by the Swedish Research Council, Region Stockholm (CIMED) and the Swedish Cancer Society. Hannes Hagström is engaged in several collaborations with the pharmaceutical industry regarding liver disease prognosis, but none that is relevant to the current study.
Publication
Rickard Strandberg, Fredrik Åberg, Juho V Asteljoki, Panu K Luukkonen, Veikko Salomaa, Antti Jula, Annamari Lundqvist, Satu Männistö, Markus Perola, Mats Talbäck, Niklas Hammar, Hannes Hagström
The BMJ, online 29 September 2025, doi: 10.1136/bmj-2024-083182
