The contribution of genetics and the environment to human lifespan has long been debated. Twin and pedigree studies over the past century, largely based on historical cohorts, have generally placed the heritability of lifespan at around 20-25%. More recent large pedigree analyses have suggested it may be even lower. These findings have fueled skepticism about whether genetic studies have anything meaningful to say about variation in human aging.
A new study published in Science challenges this view1. Shenhar et al. (2026) argue that widely cited figures underestimate the genetic contribution to lifespan because they fail to account for deaths driven by external causes. By separating extrinsic mortality from intrinsic biological aging, the authors suggest that the heritability of intrinsic human lifespan is closer to 50%.
The difference between extrinsic and intrinsic mortality is important. Extrinsic mortality includes causes that originate outside the body, such as accidents, violence, infections, and environmental hazards. Intrinsic mortality reflects biological aging and age-related diseases.
Historical cohorts used in classic twin studies were born in periods with high extrinsic mortality. In the late 19th and early 20th centuries, deaths from infections, workplace accidents, and poor living conditions were common. These deaths introduce random variation, weakening correlations between genetically related individuals.
The authors show mathematically that extrinsic mortality both compresses differences between genetic groups and increases intra-group variability. The result, the authors claim, is a systematic reduction in estimated heritability.
To test this idea, Shenhar et al. analyzed multiple datasets, including Danish and Swedish twin cohorts born between the late 1800s and early 1900s, the Swedish Adoption Twin Study of Aging (SATSA), and siblings of US centenarians. Together, these datasets span different populations, study designs, and levels of extrinsic mortality. The SATSA cohort includes twins raised apart that had not previously been analyzed for lifespan heritability, allowing genetic effects to be assessed under reduced shared environmental influence.
Because cause-of-death data are incomplete or unavailable for many historical cohorts, the authors used mortality modeling to separate intrinsic and extrinsic components. They fitted two established mortality models to historical population data and incorporated genetic variation by allowing key aging-related parameters to vary between individuals.
Monozygotic twins were simulated with identical parameters, while dizygotic twins and siblings shared parameters with partial correlation. The models were calibrated to reproduce observed lifespan correlations in each cohort, then rerun with progressively lower extrinsic mortality.
Across all cohorts and models, the same pattern emerged. As extrinsic mortality was reduced, correlations between genetically related individuals increased. When extrinsic mortality approached zero, estimated heritability converged at around 50-55%.
In the SATSA cohort, uncorrected heritability estimates rose across birth cohorts as extrinsic mortality declined during the early 20th century. The modeling results mirrored this trend, supporting the conclusion that external causes of death masked genetic effects in earlier studies.
Using a standard twin-based estimate of heritability and correcting for extrinsic mortality, the authors report a consistent estimate of intrinsic lifespan heritability of approximately 55%. This value was stable across Scandinavian twins, adopted twins raised apart (SATSA), and US centenarian sibling data.
In this context, a heritability of around 50% means that, within a population and under low extrinsic mortality, genetic differences account for around half of the variation in lifespan on top of well-established non-genetic influences on aging, not that genes determine how long any one person will live.
The SATSA dataset includes cause-of-death information, allowing the authors to explore whether genetic contributions vary by disease category. They found heritability varied by cause and age.
Cancer-related mortality showed moderate and relatively stable heritability. Cardiovascular disease showed higher heritability at earlier old ages, but this genetic influence largely evaporated among those reaching their 100th year. Finally, Dementia-related mortality showed the highest heritability by age 80, before stabilising later in life.
These findings suggest that genetic influences on lifespan are not uniform but depend on which biological processes dominate mortality risk at different ages.
This study may reframe how lifespan heritability should be interpreted in a field where many non-genetic drivers of aging are already well established. It does not claim that genes determine how long an individual will live, but suggests that genetic differences may explain some remaining variation in lifespan. Roughly half of lifespan variation remains unexplained by additive genetic effects, reflecting environment, lifestyle, healthcare, random biological processes, and non-additive genetic effects.
However, it does show that genetic influences on intrinsic aging are comparable in magnitude to those seen for many other human traits. The low estimates reported in earlier studies are thought to largely reflect the environments in which historical cohorts lived, rather than a weak genetic contribution to aging itself.
The authors also emphasize that heritability is not a fixed property. It depends on population and environment, as well as the time period. As extrinsic mortality declines in modern societies, genetic contributions to lifespan become easier to detect.
Modern populations experience lower extrinsic mortality than those born a century ago. In these settings, deaths increasingly reflect chronic disease and biological aging rather than external hazards. This shift makes genetic studies of aging more informative, but also raises new challenges.
A higher heritability does not translate into simple prediction. Longevity remains highly polygenic, with many variants of small effect interacting with the environment over decades. Identifying genetic mechanisms will require larger cohorts, careful modeling, and integrating molecular and clinical data.
The study also serves as a cautionary note for interpreting historical data. Estimates derived from past populations cannot be assumed to apply unchanged to present conditions. As environments continue to change, so too will the balance between genetic and non-genetic influences on lifespan. By accounting explicitly for extrinsic mortality, this work helps clarify a long-standing debate and repositions genetics as an important, though incomplete, contributor to human aging.