Worldwide study shows limits to how long humans can live

There are few questions in science that grip the public imagination quite like the limits of human longevity, and a new study published in the journal Biogerontology now adds a substantial piece of evidence to one of the most contentious debates in demography and aging research. Marta Gonçalves and Byung Mook Weon of Sungkyunkwan University in the Republic of Korea have carried out a global analysis of upper-tail lifespan dynamics across 46 countries, using reliable period life-table data spanning from the 1990s onward. Their conclusion is striking: while typical lifespans continue to rise steadily around the world, the extreme upper end of human longevity appears to be behaving in a fundamentally different way—one that points toward a bounded, rather than unlimited, expansion of the human lifespan.

The debate over whether there is a hard ceiling on how long humans can live has raged for decades. On one side, researchers such as those who published the 2016 Nature paper “Evidence for a limit to human lifespan” have argued that the record books show signs of saturation, with maximum reported ages plateauing since the 1990s. On the other, statisticians applying extreme value theory, including analysts whose work appeared in the Journal of the American Statistical Association and the journal Extremes, have argued that the available data cannot rule out unbounded lifespans—famously summarized in one paper’s title as “human life is unlimited—but short.” Into this contested space, the new study by Gonçalves and Weon brings an unusually broad and systematic empirical approach, analyzing not isolated record-holders but the statistical behavior of the entire upper tail of the mortality distribution across dozens of national populations.

The technical core of the study rests on three quantities extracted annually from period life tables: the characteristic life, denoted alpha; a model-based upper-tail parameter, denoted omega; and the interval between them, delta, defined as omega minus alpha. The characteristic life captures the age around which the bulk of deaths in a population concentrate—the “typical” age of death in a society that has completed its demographic transition. The upper-tail parameter, by contrast, describes where the survival distribution effectively terminates, the mathematical expression of how far the extreme end of the lifespan distribution extends. The interval delta, therefore, measures how much room remains between the age at which most people die and the theoretical far edge of survival. By tracking these three parameters year by year, country by country, the researchers could observe how the entire geometry of human survival is shifting in real time.

What they found is a pattern of profound asymmetry. Across the 46 countries analyzed, the characteristic life alpha increases steadily—a demographic signature of continued improvements in public health, medicine, and living standards pushing the age at which most people die ever higher. Yet the upper-tail parameter omega does not follow. Instead, it declines or stabilizes across countries, and the interval delta narrows correspondingly. In plain terms, the gap between the typical age of death and the extreme edge of survival is shrinking. The researchers interpret this pattern as evidence of postponed and increasingly concentrated late-life mortality: deaths are being pushed later into life, but they are also clustering more tightly around a common age, rather than stretching out toward ever more distant extremes. This is the demographic fingerprint of what demographers call mortality compression, a phenomenon first theorized by James Fries in his influential 1980 paper on the compression of morbidity in the New England Journal of Medicine.

To project where these trends lead, the authors employed a baseline projection model denoted L=120, which uses the observed dynamics of the three parameters to extrapolate their trajectories to the end of the century. The result is remarkable in its convergence: under this model, both females and males approach an upper-tail parameter of approximately 120 years and a characteristic life of approximately 103.5 years by 2100. If the projection holds, the majority of deaths in the most long-lived societies would occur beyond the age of one hundred by the end of the century, while the far edge of survival would settle near 120 years. Such a scenario represents the continuation and near-completion of a century-long process of survival-curve rectangularization—the straightening of the survival curve into a rectangle, as described in classic work by Wilmoth and Horiuchi in Demography in 1999.

Crucially, however, the authors are careful to delimit exactly what their findings do and do not establish. Sensitivity analyses performed in the study show that the inferred asymptotic boundary is model-dependent, meaning that the projected value near 120 years emerges from the specific mathematical structure of the L=120 projection rather than from an incontrovertible identification of a biological wall. The researchers explicitly state that the projected value near 120 years should be interpreted as a mathematical upper-tail estimate rather than a definitive biological maximum. Their findings, they write, are consistent with—but do not establish—an increasingly bounded upper-tail survival regime within the present demographic and modeling framework. This methodological caution places the study squarely within the tradition of statistical reviews, such as the 2022 Annual Review of Statistics and Its Application paper asking “Is there a cap on longevity?”, which have emphasized that every claimed lifespan limit rests on contestable modeling assumptions.

The data underpinning the analysis come from the Human Mortality Database, a widely trusted resource maintained by the Max Planck Institute for Demographic Research, the University of California, Berkeley, and the French Institute for Demographic Studies. Period life tables, the study’s raw material, summarize the mortality conditions of a population in a given calendar year as if a hypothetical cohort were to live through those conditions at every age. By relying on period rather than cohort data, the analysis captures the contemporary mortality regime directly, though it also means the parameters reflect conditions experienced by living populations rather than the completed lifespans of those born today. The use of 46 countries gives the analysis a genuinely global character, moving beyond the single-country studies—such as the well-known 2000 Science paper tracking maximum lifespan increases in Sweden from 1861 to 1999—that have historically dominated this literature.

The new findings resonate with, and in some respects sharpen, other recent warnings in the longevity literature. A 2024 paper in Nature Aging by S. Jay Olshansky and colleagues argued that radical life extension in humans during the twenty-first century is implausible, noting that gains in life expectancy at older ages have slowed and that a scenario in which most people live past one hundred would require breakthroughs far beyond current medicine. Similarly, Gavrilova and Gavrilov have argued in Biogerontology that the compensation effect of mortality—a robust regularity in old-age death rates—poses a challenge to any scenario of substantial human lifespan extension. The convergence of these independent lines of evidence with the new global analysis of upper-tail dynamics suggests that the weight of demographic data is increasingly tilting toward a bounded view of human longevity, even as the exact location and nature of any boundary remain uncertain.

The study also connects to a broader biological and public-health framework. The characteristic life and its steady rise reflect what demographers describe as the horizontalization of the survival curve—survival stretching rightward as more people survive to older ages—while the narrowing delta speaks to verticalization, the compression of deaths into an ever-narrower age window. Recent work in Nature Communications on interventions that steepen the survival curve has shown that compression of morbidity, the shrinking of the period of life spent in poor health, can be actively promoted through medical and behavioral interventions. If the lifespan upper tail is indeed bounded near 120 years, then the practical prize for public health is not more years of life at the extreme but a higher proportion of people reaching old age in good health—healthspan, rather than extended lifespan, becomes the realistic target of policy and research.

For a world grappling with the consequences of societal aging—from pension sustainability to healthcare capacity to, as one 2022 study in Nature Climate Change noted, the challenges that aging developed societies pose for carbon mitigation—the question of whether lifespans will continue to expand without limit is far from academic. The new analysis suggests that the answer is likely no: populations are converging on a regime in which nearly everyone survives to old age, deaths concentrate around an age just above one hundred, and the extreme edge of survival holds near 120 years. Yet the authors’ own insistence on the model-dependent character of that boundary serves as a reminder that the biology of aging could still surprise us. What the study establishes with high confidence is the shape of the current trend—bounded upper-tail growth, concentrated late-life mortality, and steadily rising characteristic lifespans—and that shape, replicated across 46 countries and three decades of data, is the strongest global evidence yet that human lifespan growth, however remarkable its past trajectory, may finally be approaching its asymptote.