Living to 90 and reaching 90 in relatively good health are related achievements, but they are not genetically identical. A whole-genome study comparing very old adults with younger controls found that some variants tracked survival itself, while others were more closely associated with healthy longevity. The distinction matters because extending life without preserving function is not the goal most people have in mind.
Evidence at a glance
- Study type: Whole-genome sequencing and genome-wide association analysis.
- Population: 3,703 adults aged 90 or older and 22,354 controls aged 18 to 75.
- Evidence level: Human genetic association evidence.
- Main caution: Associated variants are not deterministic and do not identify a treatment by themselves.
Two outcomes, two genetic patterns
The investigators analyzed 26,057 genomes. For longevity, variants near APOE, APOC1 and CFAP46 showed negative associations, meaning some versions were less common among the long-lived group. APOE is already well known for its links to Alzheimer’s disease and cardiovascular risk, which makes its appearance biologically plausible without making it simple.
When the researchers narrowed the outcome to healthy longevity, different positive signals appeared near MYO18B, TBC1D28 and LOC105376454. This suggests that surviving to advanced age and avoiding major disease or disability along the way can reflect overlapping but partly distinct biology.
Why longevity genetics is difficult
Exceptional survival is a highly selected outcome. People who reach their nineties have passed through decades of changing environments, healthcare, diet, infections and social conditions. The oldest participants are therefore not a random sample of their birth cohorts. This survivor effect can make variants appear important because they interact with a particular historical environment.
The definition of “healthy” also matters. A study may use diagnosed disease, medication, mobility, cognition or self-reported function, and each choice changes the phenotype. Genetic associations can also vary across ancestry groups. Results discovered in one population need replication elsewhere before they are treated as broadly representative.
Genes are probabilities, not instructions
Longevity is polygenic. Many variants each contribute small effects, and those effects interact with one another and with the environment. Carrying a favorable variant does not guarantee a long life, just as carrying an unfavorable one does not dictate an early death. Smoking, blood pressure, physical activity, vaccination, medical care and chance still matter enormously.
An earlier multivariate genomic analysis also pointed to distinct pathways, including haem metabolism, depending on how aging-related traits were combined. Together, the studies argue against a single “longevity gene.” They instead describe a network of risk management, maintenance and disease resistance.
What would increase confidence
Replication in larger, ancestrally diverse cohorts is essential. Functional experiments must then show what the implicated variants do in relevant cells and tissues. The most useful findings will be those that connect a reproducible genetic signal to a modifiable biological pathway without sacrificing other aspects of health.
The Lifespan Brief assessment
The study strengthens a practical idea: lifespan and healthspan should be measured separately. Genetics can help reveal the biology shared by both outcomes and the biology that separates them. It cannot yet tell an individual how long they will live or provide a clinically validated route to healthy longevity.