Scientists at the Karolinska Institute in Sweden have developed a new statistical screening method (the CORE model) that uses existing data (such as age, sex, and liver function tests) to predict with high accuracy the risk of developing severe liver diseases, including cirrhosis and cancer, over the next 10 years.
The study, published in “The BMJ,” is potentially revolutionary for primary healthcare, as it will facilitate the early detection of severe liver diseases. This is particularly important given that, statistically, liver problems, including cirrhosis, are more frequent in men globally. For instance, based on 2019 data, the overall incidence rate of cirrhosis was significantly higher in men than in women. This difference further emphasizes the need for early screening for liver diseases.
The new model, named CORE, was created using advanced statistical methods and is based on five variables: age, sex, and the levels of three key liver enzymes (AST, ALT, and GGT), which are routinely measured during periodic health check-ups.
It is important to note that liver enzyme levels alone do not always provide an accurate prediction of the development of severe liver disease. The CORE model’s accuracy is achieved precisely by integrating age and sex as additional statistical factors, allowing for a much more accurate risk assessment than one based solely on biochemical markers.
The researchers note that implementing the CORE model is a significant step forward for providing early screening for liver diseases within primary healthcare. This innovation is especially relevant because effective drug treatment regimens already exist for patients who are at high risk of developing severe liver pathologies (e.g., cirrhosis or cancer).
The research was based on data from over 480,000 individuals in Stockholm who underwent health check-ups between 1985 and 1996. Over 30 years of follow-up on the participants, it was found that approximately 1.5% developed severe liver disease (cirrhosis, cancer) and required a liver transplant.
The CORE model demonstrates high predictive accuracy; it can distinguish individuals at risk of developing the disease from those not at risk in 88% of cases. This rate significantly exceeds the effectiveness of the currently recommended FIB-4 method in the general population. As the researchers emphasize, FIB-4 is not optimally suited for general population screening and is less effective in predicting future risk, which deprived primary healthcare of the opportunity for timely detection.
The model has been successfully validated in population groups in Finland and the United Kingdom. However, further testing is needed in specific groups, such as individuals with type 2 diabetes or obesity. Also, to simplify clinical use, the model must be integrated into electronic medical record systems.
