# Development of Age‐ and Sex‐Specific Metabolomics‐Based Biological Ageing Clocks for 10‐Year Mortality Prediction

## TL;DR

Researchers developed new metabolomics-based biological aging clocks that improve 10-year mortality prediction in European adults.

## Contribution

The study introduces sex- and age-specific metabolomic risk scores and biological aging clocks for mortality prediction.

## Key findings

- 68 metabolites were significantly associated with 10-year all-cause mortality in two cohorts.
- Metabolite-augmented models improved c-statistics by 0.036–0.084 across subgroups.
- Each year of age acceleration is linked to an 8–9% higher 10-year mortality risk in external validation.

## Abstract

Metabolite concentrations vary by age and sex, yet age‐ and sex‐specific metabolomic risk scores and biological ageing clocks for mortality prediction remain undeveloped. Nuclear magnetic resonance (NMR)‐based metabolomic profiling is conducted in 209144 UK Biobank participants (12347 deaths) and 6820 from the German ESTHER study (804 deaths). Mortality risk scores are derived using least absolute shrinkage and selection operator (LASSO)‐regularized Cox regression, and metabolomics‐based mortality risk clocks (MetaboMR clocks) are constructed using elastic net regression in sex‐ and age‐stratified subgroups (50–59 and 60–69 years). Models are trained in 70% of UK Biobank and validated internally (30%) and externally in ESTHER. 68 metabolites are significantly associated with 10‐year all‐cause mortality in both cohorts. 20, 18, 12, and 13 metabolites improved 10‐year mortality prediction in younger and older men, younger and older women. Metabolite‐augmented models improved c‐statistics by 0.036–0.084 across subgroups. In the external validation set, each year of age acceleration is associated with an 8% and 9% higher 10‐year mortality risk for MetaboMR clock1 and clock2. Sex‐ and age‐specific metabolomic risk scores significantly enhance 10‐year mortality prediction beyond traditional models. The MetaboMR clocks may serve as measures of biological ageing and support personalized risk stratification in clinical settings.

Metabolomics‐based ageing clocks developed using blood biomarkers from over 215 000 European adults, improved the prediction of 10‐year all‐cause mortality compared to traditional models. Biological age acceleration is associated with higher mortality risk, as each additional year of metabolic age corresponded to at least 8% increased risk. These tools may support personalized risk assessment and provide insights into the biological ageing process.