Large-Scale Plasma Proteomics Profiles for Predicting Ischemic Stroke Risk in the General Population

Area of Science:

• Biomarkers and Disease Prediction
• Cardiovascular Research
• Proteomics and Genomics

Background:

• Ischemic stroke (IS) remains a leading cause of disability and mortality.
• Accurate prediction of IS risk is crucial for implementing timely preventive measures.
• Existing risk prediction models have limitations in capturing individual susceptibility.

Purpose of the Study:

• To develop and validate a novel protein-based risk score for predicting ischemic stroke (IS).
• To compare the predictive performance of the protein risk score against established clinical risk factors and a polygenic risk score.
• To identify key proteins contributing to IS risk prediction.

Main Methods:

• A prospective cohort study involving over 53,000 participants from the UK Biobank Pharmaceutical Proteomics Project (UKB-PPP).
• Development of an IS protein risk score using least absolute shrinkage and selection operator (LASSO) regression on proteomic data.
• Validation of the score using internal (training/validation sets) and external (Scotland/Wales) cohorts, assessing predictive ability with the C statistic.

Main Results:

• The IS protein risk score demonstrated superior predictive performance (C statistic: 0.765) compared to clinical risk factors (C statistic: 0.753) and the IS polygenic risk score (C statistic: 0.730).
• A simplified model incorporating age, sex, and the IS protein risk score (or its top 5 proteins) showed strong predictive capability.
• Key proteins like GDF15, PLAUR, NT-proBNP, IGFBP4, and BCAN were major contributors to the score's predictive power.

Conclusions:

• A protein risk score offers a highly effective tool for predicting ischemic stroke (IS) risk.
• The developed score surpasses traditional clinical factors and polygenic risk scores in predictive accuracy.
• A simplified model using the protein risk score provides a practical approach for IS risk assessment.