One organ, three models of age and risk
A study with genetic data from 409,000 participants found limited overlap between organ age acceleration and organ-specific mortality risk, with different genetic associations and links between mortality scores and tissue-damage markers such as GDF15.
Three models of age and risk for one organ
Groups identified as having accelerated organ aging and those identified as having a high risk of death from that organ’s diseases overlapped by only 12% to 27%, depending on the organ. This finding comes from a University of Hong Kong preprint published on Research Square, a platform for preliminary research papers, on September 30, 2026.
The researchers built three models for each of 14 organs. The models used the same molecular measurements but predicted different outcomes: chronological age, the risk of death from diseases of a specific organ, or the risk of death from any cause. The first model estimated how far an organ’s molecular age exceeded chronological age. The other two estimated mortality risk.
The differences also appeared in the genetics: of 172 independent genomic regions associated with these measures, 114 had a statistically significant association with only one of the three. Organ-specific mortality scores were also associated with proteins that indicate tissue damage, including GDF15. These scores may therefore partly reflect tissue damage. This association alone does not show that lowering GDF15 will extend life.
Molecular data were available for 26,000 participants in UK Biobank, a British biobank, while genetic data covered 409,000. The main genetic findings were also tested in an independent group of 383,000 people. The work remains a preprint.
For trials of drugs intended to slow aging, the choice of measure affects which effects researchers can detect. Measuring only molecular age acceleration could miss a change in the risk of death from a specific disease. Measuring only mortality risk could miss a change in molecular aging. A reported change in biological age therefore needs to be read alongside the model’s description: what exactly was it trained to predict?
This item was selected because it gives readers a concrete test for biological-age claims: inspect the model target and the biological signals behind it. The linked Eternal Search glossary page provides the relevant entry for GDF15, a tissue-damage marker associated in the study with organ-specific mortality scores.