Predicting the probability of death using proteomics.

Area of Science:

• Biochemistry
• Genomics
• Epidemiology

Background:

• Predicting all-cause mortality is complex, often requiring extensive patient data.
• Large-scale proteomics datasets are emerging as valuable tools for health outcome prediction.

Purpose of the Study:

• To develop and validate plasma protein-based predictors for short- and long-term all-cause mortality risk.
• To assess the performance of proteomics predictors against conventional mortality risk factors.

Main Methods:

• Analysis of plasma protein levels (4,684 proteins) in a large cohort (22,913 Icelanders).
• Longitudinal follow-up (mean 13.7 years) with mortality data collection.
• Comparison of proteomics predictor performance with traditional risk factor models.

Main Results:

• A novel predictor based on plasma protein levels was developed for all-cause mortality.
• The proteomics predictor demonstrated superior survival prediction compared to conventional risk factors.
• Identified distinct high-risk (88% 10-year mortality) and low-risk (1% 10-year mortality) groups based on protein levels.
• Predicted mortality risk correlated with frailty measures in an independent dataset.

Conclusions:

• Plasma proteome analysis offers a powerful method for assessing general health status.
• Proteomics-based prediction of mortality risk is feasible and outperforms traditional methods.
• This approach can significantly improve the identification of individuals at high risk of death.