New Algorithm to Enhance Prediction of Cardiovascular Disease Risk

Current Algorithms Do Not Include All Risks

Cardiovascular diseases are the most common cause of death worldwide. Consequently, the World Health Organization, the European Union, and individual countries use algorithms to predict the risk of cardiovascular diseases in various demographic groups of the population. The goal is to identify risk groups for whom specific measures, such as screening programs, can then be set up.

However, recent research has shown that the most commonly used algorithms – the American ASCVD, the European SCORE2, and the British QRISK3 – fail to consider seven significant risk factors for cardiovascular diseases: cancers of the brain, lungs, or oral cavity, leukemia, Down syndrome, COPD, and learning disabilities. In addition, two other predictors found exclusively in women, namely preeclampsia and postpartum depression, have been identified.

Development and Testing

For the development of the new QR4 algorithm, scientists utilized the QResearch and Clinical Practice Research Datalink databases, which contain anonymized data from more than 16 million individuals in the United Kingdom, under the British National Health Service (NHS). The study itself only included health data of adults aged 18 to 84 from the period from January 1, 2010, to December 31, 2021.

In addition to the newly discovered and existing factors, scientists also investigated which age groups are more and less affected by specific risk factors. Based on this, they could assign the correct coefficients to each factor and age group for calculations.

Compared to previous models, the new QR4 algorithm also considers the risks of death from causes other than cardiovascular in nature. Especially in older age groups, this reduces the current overestimation of cardiovascular risk factors.

Subsequent testing and calibration showed that the performance of the new QR4 model is significantly more accurate than current algorithms. Compared to the QRISK3 model, the new model showed much better agreement between predicted and subsequently observed risk. Comparisons with the SCORE2 and ASCVD models showed that the European model underestimated some factors, while the American model overestimated them.

Limits Open the Door for Further Research

However, as the authors of the study themselves point out, in the case of risk factors associated with the occurrence of other diseases, especially cancers, it is necessary to assess the risk of cardiovascular diseases in the context of the original diagnosis. According to them, the use of the new algorithm in patients with cancerous diseases will need to be further considered and its contribution to specific cases evaluated.

For instance, the risk should be assessed differently in patients with lung tumors, where the 5-year survival rate from diagnosis is approximately only 15%, and differently in patients with leukemia or oral cavity cancers, from which up to 90% survive 5 years.

According to the study's authors, this opens up opportunities for further research that would better explore and describe the connections between cancer treatment and subsequent risk of cardiovascular diseases.

Editorial Team, Medscope.pro

Source: Hippisley-Cox J., Coupland C. A. C., Bafadhel M. et al. Development and validation of a new algorithm for improved cardiovascular risk prediction. Nat Med 2024; 30 (5): 1440–1447, doi: 10.1038/s41591-024-02905-y.