Blood Protein Profiles Predict Mortality Risk for Earlier Medical Intervention

Elevated levels of specific proteins in the blood can signal increased risk of mortality, according to new evidence showing that five proteins involved in cancer, inflammation, and cell regulation strongly predict long-term survival. Researchers also identified hundreds of additional proteins linked to death within five or ten years, even after accounting for lifestyle and medical factors. These findings point to a potential early warning system that could prompt timely, preventive medical care.

In the study led by the University of Surrey (Guildford, UK, researchers analyzed blood-protein data from more than 38,000 adults enrolled in the UK Biobank. By examining large-scale proteomic profiles, the team aimed to determine whether circulating proteins could reliably reflect the risk of early non-accidental mortality. The dataset included middle-aged and older adults, some of whom died within five or ten years of providing blood samples.

Image: The new analysis of blood samples links specific protein patterns to five- and ten-year mortality risk (Photo courtesy of Adobe Stock)

To identify predictive markers, investigators evaluated thousands of proteins and catalogued associations with near- and longer-term death risk. The research, published in Nature Communications, revealed 392 proteins predictive of death within five years and 377 proteins linked to ten-year mortality. Five proteins—PLAUR, SERPINA3, CRIM1, DDR1, and LTBP2—emerged as consistent indicators across both timeframes, suggesting a central role in underlying biological pathways that contribute to disease severity and poor outcomes.

Short-term mortality was most strongly associated with proteins involved in inflammation, cellular communication, and rapid cellular changes. One example is SERPINA1, a lung-protective protein whose elevated levels signaled heightened five-year mortality risk. Meanwhile, persistently high levels of SERPINA3 have been previously linked to reduced survival in certain cancers, reinforcing its potential as a powerful marker of biological stress.

These discoveries could help guide the development of clinical protein-based risk models capable of flagging individuals who would benefit from preventive monitoring or targeted intervention. Researchers say such models may support healthcare planning by identifying early physiological disruption before overt disease develops. The next steps involve validating the predictive proteins in additional populations and assessing how they perform across diverse clinical settings.

“Certain blood proteins reflect more gradual, chronic processes that affect long-term survival, whilst others are reflective of more immediate, acute health risks, like troponin for heart disease,” said Natalia Koziar, lead author. “The new information could be useful to clinicians in creating a scorecard of protein levels that help to identify individuals most at risk of developing conditions with poor outcomes.”

Related Links:
University of Surrey