Maximum human lifespan: revision 1 to 3

Revision 1

Revision 3

categories: ["longevity"]

tags: ["aging", "demography", "mortality", "supercentenarians", "life extension"]

summary: "The longest duration a human being has lived or could live, distinct from average life expectancy and disputed as to whether any fixed ceiling exists."

summary: "The longest duration a human being has lived or could live, distinct from average life expectancy and disputed as to whether any fixed ceiling exists."

updated: "2026-07-27"

updated: "2026-11-01"

issues: ["Timeline entries for 1992 and for 2021-2022 name results that carry no citation."]

issues: ["The Japan 2010 registry audit and the worldwide centenarian growth claim still need citations."]

---

```infobox

"caption": "Concept in demography and biogerontology",

"rows": [

{ "label": "Verified record", "value": "122 years, 164 days" },

{ "label": "Record holder", "value": "Jeanne Calment (1875–1997)" },

{ "label": "Record holder", "value": "Jeanne Calment (1875-1997)" },

{ "label": "Oldest living (Nov 2026)", "value": "Ethel Caterham, 117" },

{ "label": "Supercentenarian", "value": "Aged 110 or older" },

{ "label": "Key dispute", "value": "Fixed limit vs. open-ended tail" },

{ "label": "Field", "value": "Demography, biogerontology" },

}

```

**Maximum human lifespan** is the greatest age a human being has attained, or in its theoretical sense the greatest age a human could attain under the best achievable conditions. It is a property of the extreme tail of the mortality distribution and behaves very differently from life expectancy, which is a mean. Whether the species has a fixed ceiling, and if so where, has been argued in the demographic literature since the 1990s without resolution.

**Maximum human lifespan** is the greatest age a human being has attained or, in its theoretical

sense, the greatest age a human could attain under the best achievable conditions. It is a property

of the extreme tail of the mortality distribution and behaves very differently from life expectancy,

which is a mean. Whether the species has a fixed ceiling, and if so where, has been argued in the

demographic literature since the 1990s without resolution.

```keyfacts

[

{ "value": "122y 164d", "label": "Longest verified human life", "note": "Jeanne Calment, 1875–1997" },

{ "value": "122y 164d", "label": "Longest verified human life", "note": "Jeanne Calment, 1875-1997" },

{ "value": "119y", "label": "Next two verified cases", "note": "Sarah Knauss and Kane Tanaka" },

{ "value": "119y", "label": "Next two verified cases", "note": "Kane Tanaka and Sarah Knauss" },

{ "value": "117", "label": "Oldest living validated person", "note": "Ethel Caterham, UK, as of November 2026" },

{ "value": "110", "label": "Age defining a supercentenarian", "note": "the threshold for record validation" }

]

```

## Lifespan and life expectancy

The two quantities are routinely confused. Life expectancy at birth is the mean age at death of a hypothetical cohort experiencing current age-specific mortality rates. It has risen dramatically over two centuries, mostly because fewer people die young. Maximum lifespan is a record — an order statistic drawn from the extreme right tail — and is far less responsive to public-health improvement. Saving infants does not lengthen the tail; it only sends more people into it.

The two quantities are routinely confused. Life expectancy at birth is the mean age at death of a

hypothetical cohort experiencing current age-specific mortality rates. It has risen dramatically

over two centuries, mostly because fewer people die young. Maximum lifespan is a record, an order

statistic drawn from the extreme right tail, and it responds far less to public-health improvement.

Saving infants does not lengthen the tail; it only sends more people into it.

This distinction disciplines several common arguments. Classical claims that Bronze Age humans lived to 30 describe an average dragged down by infant mortality, not a species whose members died at 30. It also explains why record ages have crept upward far more slowly than mean ages: the tail grows because the population at risk grows, not because the biology has changed. A third quantity, [[healthspan]], tracks the years lived in good function and moves independently of both; whether added years are healthy ones is the subject of [[compression-of-morbidity]].

The distinction disciplines several common arguments. Classical claims that Bronze Age humans lived

to 30 describe an average dragged down by infant mortality, not a species whose members died at 30.

It also explains why record ages have crept upward far more slowly than mean ages: the tail grows

because the population at risk grows, not because the biology has changed. A third quantity,

[[healthspan]], tracks the years lived in good function and moves independently of both. Whether

added years are healthy ones is the subject of [[compression-of-morbidity]].

> [!note] Terminology

> "Maximum lifespan" is used in at least three senses: the oldest verified individual, the age by which some specified fraction of a cohort has died, and a hypothesized biological ceiling. Papers frequently switch between them without warning, which accounts for a good deal of the apparent disagreement.

> "Maximum lifespan" is used in at least three senses: the oldest verified individual, the age by

> which some specified fraction of a cohort has died, and a hypothesized biological ceiling. Papers

> often switch between them without warning, which accounts for a good deal of the apparent

> disagreement.

## The record

Jeanne Calment of Arles, France, died in 1997 at 122 years and 164 days, an age validated against parish and civil records by a team led by the demographer Jean-Marie Robine. No verified case has come within three years of it in the intervening decades. Sarah Knauss, who died in the United States in 1999, and Kane Tanaka, who died in Japan in 2022, both reached 119. Later record holders — including Lucile Randon and Maria Branyas Morera — died at 117 or 118.

Jeanne Calment of Arles, France, was born on 21 February 1875 and died on 4 August 1997 at 122

years and 164 days.[^lq2026] Jean-Marie Robine and Michel Allard validated the age when she entered

her 121st year, working mainly from civil and church records and censuses, supplemented by notarial documents and newspaper articles.[^robine1999] In the decades since, no verified case has come within three years

of it. Kane Tanaka, who died in Japan in 2022 at 119 years and 107 days, and Sarah Knauss, who died

in the United States in 1999 at 119 years and 97 days, rank second and third on LongeviQuest's list

of validated ages. Two later holders of the title of oldest living person, Lucile Randon of France

(118 years, 340 days) and Maria Branyas Morera of Spain (117 years, 168 days), died in 2023 and

2024.[^lq2026]

The stability of Calment's record is itself a data point. Over a period in which the number of people surviving past 100 rose by more than an order of magnitude worldwide, the maximum reported age at death did not follow it upward. Interpretations differ sharply: one camp reads this as a ceiling, the other as the ordinary behavior of a sparse extreme tail.

As of November 2026 the oldest living person with a validated age is Ethel Caterham, born in

Hampshire, England, on 21 August 1909, who turned 117 in a care home in Surrey on 21 August

2026.[^grg2026] Guinness World Records lists her as the oldest person living,[^gwr2026] and at the

end of October 2026 LongeviQuest placed her eleventh among the oldest validated people

ever.[^lq2026]

Calment's case has been challenged. In 2019 the Russian researchers Nikolay Zak and Valery Novoselov argued that her daughter Yvonne had assumed her identity, which would place the true age at about 99. A rebuttal by Robine and colleagues reviewed the documentary and testimonial evidence and concluded the substitution hypothesis was inconsistent with the record.[^robine2019] Mainstream demography continues to treat the case as validated, while acknowledging that any single-record claim rests on document quality.

The stability of Calment's record is itself a data point. Over a period in which the number of

people surviving past 100 rose worldwide by more than an order of magnitude, the maximum reported

age at death did not rise with it. Interpretations differ sharply. One camp reads this as a ceiling;

the other reads it as the ordinary behavior of a sparse extreme tail.

Calment's case has been challenged. In a 2019 paper in *Rejuvenation Research*, Nikolay Zak of the

Moscow Society of Naturalists, crediting conversations with the geriatrician Valery Novoselov,

argued that Jeanne Calment died in 1934 and that her daughter Yvonne assumed her

identity.[^zak2019] Because Yvonne was born in 1898, that would put the true age at death at about

99. A 2020 Bayesian analysis by Zak and Peter Gibbs said a definitive conclusion was not possible on

the available evidence, called the validation "unsafe", and proposed DNA tests that could settle the

question.[^zak2020] A rebuttal by Robine and colleagues set out further documentation, argued that

the conspiracy claims rested on inaccurate facts, and presented models in which a 122-year-old is

extremely rare but plausible.[^robine2019] LongeviQuest's validated list still ranks Calment

first.[^lq2026]

## The limit debate

```timeline

[

  { "year": "1992", "title": "Mortality deceleration in insects", "text": "Carey and colleagues report that death rates in very large medfly cohorts stop rising at the oldest ages, contradicting a simple exponential model of aging." },

  { "year": "1992", "title": "Mortality deceleration in insects", "text": "Carey and colleagues report that death rates in medfly cohorts totalling about 1.2 million flies level off and decline at the oldest ages, contradicting a simple exponential model of aging." },

  { "year": "2016", "title": "A claimed natural limit", "text": "Dong, Milholland and Vijg argue in Nature that the maximum reported age at death plateaued in the mid-1990s and that human lifespan is capped near 115 years." },

  { "year": "2016", "title": "A claimed natural limit", "text": "Dong, Milholland and Vijg argue in Nature that the maximum reported age at death plateaued around 1995 at about 115 years." },

  { "year": "2017", "title": "Multiple technical rebuttals", "text": "Nature publishes several comments attacking the 1995 breakpoint as arbitrary and the statistical treatment of a very small sample as unsound." },

  { "year": "2017", "title": "Five rebuttals", "text": "Nature publishes five comments disputing the 1995 breakpoint, the statistical methods and the data used." },

  { "year": "2018", "title": "The Italian plateau", "text": "Barbi and colleagues report in Science that mortality hazard in Italian cohorts flattens after about age 105, implying no hard wall at 115." },

  { "year": "2018", "title": "The Italian plateau", "text": "Barbi and colleagues report in Science that the mortality hazard of Italians aged 105 and older is essentially constant, implying no hard wall at 115." },

  { "year": "2018", "title": "Errors as an explanation", "text": "Saul Newman argues that late-life mortality plateaus can be generated entirely by age-reporting error, without any biological deceleration." },

  { "year": "2018", "title": "Errors as an explanation", "text": "Saul Newman argues that late-life mortality plateaus can be generated entirely by age-reporting and cohort-blending errors." },

  { "year": "2021–2022", "title": "Statistical reviews", "text": "Extreme-value analyses find the data consistent with an exponential tail and no identifiable finite cap within the observable range, while resilience-based modelling suggests an outer bound somewhere past 120." }

  { "year": "2019", "title": "A finite Dutch endpoint", "text": "An extreme-value analysis of about 285,000 Dutch deaths at 92 or older reports an upper limit for each sex, with no upward trend over 1986-2015." },

  { "year": "2021", "title": "No cap below 130", "text": "Belzile and colleagues find a constant force of mortality beyond 108 in Italian and French data and call an upper bound below 130 implausible." },

  { "year": "2021", "title": "A resilience extrapolation", "text": "Pyrkov and colleagues extrapolate a loss of physiological resilience in blood-count and step-count data to a critical point at 120-150 years." },

  { "year": "2022", "title": "A statistical review", "text": "A review of the extreme-value literature concludes that any upper limit lies well beyond the highest lifetime reliably recorded." },

  { "year": "2024", "title": "Slowing life-expectancy gains", "text": "Olshansky and colleagues report that life-expectancy gains in the longest-lived populations have decelerated since 1990." }

]

```

The 2016 paper by Dong, Milholland and Vijg is the reference point for the "hard limit" position.[^dong2016] Using the International Database on Longevity and national datasets, they reported that the maximum reported age at death rose until roughly 1995 and then stopped, and inferred an average ceiling near 115 with 125 as a practical absolute. The rebuttals published the following year objected that the breakpoint was chosen after inspecting the data, that the analysis discarded most of the available information, and that the same dataset supports a continued slow rise.

The idea that mortality might stop accelerating at the oldest ages predates the human data. In 1992

Carey and colleagues reported that death rates leveled off and then fell at older ages in about 1.2

million caged medflies, and argued that this cast doubt on the idea that species have absolute

life-span limits.[^carey1992] The 2016 paper by Dong, Milholland and Vijg is the reference point for

the "hard limit" position in humans.[^dong2016] Using 534 supercentenarian deaths from France, Japan, the United Kingdom and

the United States in the International Database on Longevity, they reported that the yearly maximum

reported age at death rose until about 1995 and then stopped, plateauing at 114.9 years (95%

confidence interval 113.1-116.7). Modelling that maximum as a Poisson variable, they put the

probability of anyone exceeding 125 in a given year at less than 1 in 10,000.[^dong2016] Nature

published five rebuttals in June 2017.[^rw2017] Siegfried Hekimi, co-author of one, said there was

no justification for splitting the data at 1995; James Vaupel told Retraction Watch that the authors had used weak statistical methods on inappropriate data; and Saul Newman, in a separate critique published in F1000Research, argued that the later decline

reflected populations dropping out of the data, not a trend.[^rw2017][^hughes2017][^lenart2017] Another

group argued that maximum lifespan may rise to 125 years.[^debeer2017] The original authors

defended their partition of the data as "data-driven".[^rw2017]

Barbi and colleagues took a different route, fitting hazard curves to Italian semi-supercentenarians and finding that the annual probability of death flattens after 105 rather than continuing to climb.[^barbi2018] A flat hazard implies a long, thin tail rather than a wall: if the annual risk of death holds near one in two, ages beyond 120 remain possible but exceedingly rare. Whether the plateau is real remains open. Saul Newman has argued that plateaus of exactly this shape are produced by clerical error in age records, an argument closely related to his later work on [[blue-zones]].[^newman2018]

Barbi and colleagues took a different route. They estimated hazard rates for all 3,836 Italians aged

105 and older between 2009 and 2015 and found that the annual risk of death flattens after 105

instead of continuing to climb.[^barbi2018] For the 1904 birth cohort the estimated plateau

corresponds to an annual probability of dying of 0.475 and a further life expectancy of 1.55 years.

A flat hazard implies a long, thin tail, not a wall: if the annual risk of death holds near one in

two, ages beyond 120 remain possible but exceedingly rare. Whether the plateau is real remains open.

Saul Newman has argued that clerical error in age records produces plateaus of exactly this shape,

and that age-estimation and cohort-blending errors at rates below 1 in 10,000 are enough to create

them, an argument closely related to his later work on [[blue-zones]].[^newman2018]

Extreme-value statisticians have reached opposite verdicts from different data. An analysis of about

285,000 Dutch residents who died at 92 or older between 1986 and 2015 reported "compelling

statistical evidence" of an upper limit for each sex, with no sign that the limits rose over those

30 years.[^einmahl2019] Belzile and colleagues, analysing Italian and French

semi-supercentenarians, found that a constant force of mortality beyond 108 describes the data well

and that an upper bound below 130 years is implausible.[^belzile2021] Their 2022 review put the

lower 95% confidence limits for the human lifespan at about 130 years, with point estimates that

typically show no upper limit.[^belzile2022] Using a related exponential model with population

projections, Pearce and Raftery estimated a greater than 99% probability that Calment's record will

be broken by 2100, and probabilities of 89%, 44% and 13% that someone this century reaches 126, 128

or 130.[^pearce2021]

A physiological estimate arrived from outside demography. Pyrkov and colleagues tracked fluctuations

in a mortality score built from complete blood counts in a consumer laboratory database, and in

step counts from wearable devices, and found that recovery from perturbations slowed steadily with

age. Extrapolating that trend, they placed the complete loss of resilience at a critical point of

120-150 years and described it as an absolute limit of human lifespan.[^pyrkov2021] The figure is an

extrapolation well beyond the ages in the data, and five of the six authors were employees of Gero, the company that funded the study.[^pyrkov2021]

A separate line of work concerns averages rather than records. Olshansky and colleagues, using

national data for 1990-2019 from the eight countries with the longest-lived populations plus Hong

Kong and the United States, found that improvements in life expectancy have decelerated since 1990.

They estimated that survival to 100 is unlikely to exceed 15% for females and 5% for males this

century and concluded that, unless biological aging can be markedly slowed, radical life extension

is implausible in the twenty-first century.[^olshansky2024]

> [!debate] What each side is actually claiming

> The "limit" position holds that the tail is truncated by biology — that something goes wrong at an age no intervention currently touches. The "no limit" position holds only that the data cannot distinguish a truncated tail from a very thin one, not that people will live to 150. Almost nobody argues that current medicine can produce a 130-year-old.

> The "limit" position holds that biology truncates the tail: something goes wrong at an age no

> current intervention touches. The "no limit" position holds only that the data cannot distinguish

> a truncated tail from a very thin one, not that people will live to 150. Almost nobody argues that

> current medicine can produce a 130-year-old.

## Verification as the binding constraint

Every claim in this area is limited by record quality. Validating an age of 110 requires a birth record created before the person could have had any reason to falsify it, an unbroken chain of identity documents, and an official death record. Only a small minority of claims to a very high age survive that process, and the validated lists maintained by longevity researchers are correspondingly short.

Record quality limits every claim in this area. Validating an age of 110 requires a birth record

created before the person could have had any reason to falsify it, an unbroken chain of identity

documents, and an official death record. Only a small minority of claims to a very high age survive

that process, and the validated lists maintained by longevity researchers are correspondingly short.

The failure modes are systematic rather than random. Age exaggeration rises where pensions are age-linked, where birth registration was late or incomplete, and where the reward for a false record is largest. Japan's 2010 audit of its centenarian registry found very large numbers of people recorded as living who were dead or untraceable. Because errors are one-directional — records overstate age far more often than they understate it — they contaminate precisely the tail that limit studies analyse.

The failure modes are systematic, not random. Age exaggeration rises where pensions are linked to

age, where birth registration was late or incomplete, and where a false record brings the largest

reward. Japan's 2010 audit of its centenarian registry found very large numbers of people recorded

as living who were dead or untraceable. Because the errors run in one direction, with records

overstating age far more often than understating it, they contaminate exactly the tail that limit

studies analyse.

## Biology of the extreme tail

Comparative work offers little support for a species-specific hard stop. Mortality rates in some organisms do not rise measurably with age at all, the subject of [[negligible-senescence]], and lifespan across mammals varies by more than two orders of magnitude with no obvious mechanism setting an individual ceiling. Invertebrate genetics makes the point from the other side: a *daf-2* mutation roughly doubles lifespan in *Caenorhabditis elegans*, the result reported by [[cynthia-kenyon|Cynthia Kenyon's]] laboratory, so a species-typical maximum is not necessarily a fixed quantity — though the mammalian versions of the same manipulation move lifespan far less. Within humans, the identified genetic contributions to extreme longevity are modest and inconsistently replicated, and heritability of lifespan estimated from pedigrees is low once assortative mating is accounted for.

Comparative work offers little support for a species-specific hard stop. In some organisms mortality

rates do not rise measurably with age at all, the subject of [[negligible-senescence]], and lifespan

across mammals varies by more than two orders of magnitude with no obvious mechanism setting an

individual ceiling. Invertebrate genetics makes the point from the other side. A *daf-2* mutation

roughly doubles lifespan in *Caenorhabditis elegans*, the result reported by

[[cynthia-kenyon|Cynthia Kenyon's]] laboratory,[^kenyon1993] so a species-typical maximum is not

necessarily fixed, though the mammalian versions of the same manipulation move lifespan far less.

Within humans, the identified genetic contributions to extreme longevity are modest and

inconsistently replicated. An analysis of pedigrees covering hundreds of millions of people found

that the heritability of lifespan falls well below 10% once assortative mating is accounted

for.[^ruby2018]

Mechanistic candidates for a ceiling have been proposed and none has held up cleanly. [[telomeres-and-telomerase|Telomere attrition]] sets a replicative limit in cultured cells but correlates weakly with human survival; accumulated [[cellular-senescence|senescent cells]] and [[stem-cell-exhaustion|stem cell exhaustion]] progress steadily rather than hitting a threshold; [[inflammaging|chronic sterile inflammation]] rises with age but is at least partly downstream of other damage. Each explains decline, none explains a wall.

Mechanistic candidates for a ceiling have been proposed, and none has held up cleanly.

[[telomeres-and-telomerase|Telomere attrition]] sets a replicative limit in cultured cells but

correlates weakly with human survival. Accumulated [[cellular-senescence|senescent cells]] and

[[stem-cell-exhaustion|stem cell exhaustion]] progress steadily instead of hitting a threshold.

[[inflammaging|Chronic sterile inflammation]] rises with age but is at least partly downstream of

other damage. Each explains decline; none explains a wall.

What supercentenarian biology does show is a compression of terminal decline: the oldest old tend to die of general frailty, systemic amyloidosis, and organ failure rather than of the cancers and cardiovascular events that kill people in their seventies. This is sometimes read as evidence that [[hallmarks-of-aging|aging processes]] themselves, rather than any specific disease, become the proximate cause of death at the tail — the core premise of the [[geroscience-hypothesis]].

Supercentenarian biology does show a compression of terminal decline. The oldest old tend to die of

general frailty, systemic amyloidosis and organ failure, not of the cancers and cardiovascular

events that kill people in their seventies. This is sometimes read as evidence that at the tail the

[[hallmarks-of-aging|aging processes]] themselves, and not any specific disease, become the

proximate cause of death, the core premise of the [[geroscience-hypothesis]].

No intervention has extended maximum human lifespan. [[caloric-restriction]] extends maximum lifespan in rodents; the human data address [[biological-age|biomarkers]] and cardiometabolic risk, not the tail. [[senolytics]], [[rapamycin]], and other geroprotector candidates have not been tested on a timescale that could show such an effect, and no [[aging-biomarkers|validated surrogate]] exists that would let them be tested faster. Arguments about [[longevity-escape-velocity]] advanced by [[aubrey-de-grey]] and others concern whether that will change; they are not evidence that it has. For those who expect no such change within their own lifetime, [[cryonics]] is the fallback bet.

## Forecasts of radical life extension

Forecasts of lifespans in the hundreds of years rest on assumed future medicine, not on

extrapolation of the records above. The best-known come from [[aubrey-de-grey]]. In a 2004 essay he

wrote that if mortality rates at all ages fell by 10% a year and that decline were sustained

indefinitely, the first 1,000-year-old would probably be only 5-10 years younger than the first

150-year-old, the reasoning behind [[longevity-escape-velocity]].[^degrey2004] The arithmetic of

such estimates is discussed under negligible senescence.

Mainstream biogerontologists have rejected the timescale. In 2005, 28 of them wrote that each

proposal in de Grey's programme was, "at our present stage of ignorance, exceptionally optimistic",

and that reversing aging within our lifetime "is so far from plausible that it commands no respect

at all within the informed scientific community."[^warner2005] A long-running wager marks a

narrower disagreement. In 2001 Steven Austad bet S. Jay Olshansky that someone would live to 150 by 2150,

cognizant of their surroundings throughout, while Olshansky bet that 130 years is the top end of the

human lifespan; three scientists are to settle it on 1 January 2150.[^uidaho2001]

## Evidence from interventions

No intervention has extended maximum human lifespan, and none has been tested in a way that could

show it. The animal record is real but modest, and it differs sharply by species.

- **Mice.** In the [[interventions-testing-program]], [[rapamycin]] fed from 600 days of age

increased the age at 90% mortality by 14% in female and 9% in male genetically heterogeneous

mice.[^harrison2009] [[caloric-restriction]] also extends maximum lifespan in rodents.

- **Dogs.** In a paired study of 48 Labrador retrievers, dogs fed 25% less than their pair-mates

from eight weeks of age had a significantly longer median lifespan.[^kealy2002] As of November 2026 no drug

trial in dogs has reported a lifespan result. The [[dog-aging-project]]'s rapamycin trial had enrolled 170

dogs by December 2024 with a target of 580,[^tamu2024] and [[loyal]]'s STAY study of its drug

LOY-002 is fully enrolled at 1,300 dogs and remains blinded.[^loyal2026]

- **Primates.** Two long-running studies of calorie restriction in rhesus monkeys disagreed: the

Wisconsin study reported fewer age-related deaths,[^colman2009] while the National Institute on

Aging study found no improvement in survival.[^mattison2012] A joint 2017 analysis attributed the

difference partly to study design and concluded that the health benefits are conserved in

monkeys.[^mattison2017] For rapamycin in common marmosets, the researcher Adam Salmon told

public radio in 2024 that his team had concluded the drug extended lifespan; that report gave no

effect size and was not a peer-reviewed paper.[^tpr2024]

- **Humans.** Calorie-restriction trials address [[biological-age|biomarkers]] and cardiometabolic

risk, not the tail. [[senolytics]], rapamycin and other geroprotector candidates have not been

tested on a timescale that could show such an effect, and no [[aging-biomarkers|validated surrogate]] exists

that would let them be tested faster.

The mammalian effects reported so far are increments: percentages of lifespan in mice, a longer

median in one dog study. Arguments about whether that will change concern future medicine; they are

not evidence that it has. [[cryonics]] is the fallback bet for those who expect no such change within

their own lifetime.

## Outlook

The question that would settle the debate is empirical and slow: as the number of people reaching 105 continues to grow, does the maximum reported age at death grow with it? Extreme-value theory makes a testable prediction — a thin exponential tail produces a slowly rising record, a truncated one produces a flat record — and the datasets needed to distinguish them are accumulating at a rate set by human generations. Better birth registration in the cohorts now entering old age will do more to resolve it than any new statistical method, since the current dispute is at least as much about data quality as about biology. A second, faster route would be a molecular measure that tracked the tail directly: if an [[epigenetic-clock|epigenetic]] or proteomic measure could distinguish a 112-year-old who will reach 120 from one who will not, the ceiling question would become answerable without waiting for the deaths.

The question that would settle the debate is empirical and slow: as the number of people reaching

105 keeps growing, does the maximum reported age at death grow with it? Extreme-value theory makes a

testable prediction. A thin exponential tail produces a slowly rising record, and a truncated one

produces a flat record. The datasets needed to tell them apart accumulate at a rate set by human

generations. Better birth registration in the cohorts now entering old age will do more to resolve

the question than any new statistical method, since the current dispute concerns data quality at

least as much as biology. A second, faster route would be a molecular measure that tracked the tail

directly. If an [[epigenetic-clock|epigenetic]] or proteomic measure could distinguish a

112-year-old who will reach 120 from one who will not, the ceiling question would become answerable

without waiting for the deaths.

## See also

- [[compression-of-morbidity]]

- [[hallmarks-of-aging]]

- [[geroscience-hypothesis]]

- [[interventions-testing-program]]

## References

[^barbi2018]: `paper` Barbi, E., Lagona, F., Marsili, M., Vaupel, J.W. and Wachter, K.W. "The plateau of human mortality: Demography of longevity pioneers." *Science*, 2018. {Italian records only, with each individual age validated; whether the plateau appears in other national datasets is disputed.}

[^barbi2018]: `paper` Barbi, E., Lagona, F., Marsili, M., Vaupel, J.W. and Wachter, K.W. "The plateau of human mortality: Demography of longevity pioneers." *Science*, 2018. {Italian records only, with each individual age validated; whether the plateau appears in other national datasets is disputed.}

[^newman2018]: `paper` Newman, S.J. "Errors as a primary cause of late-life mortality deceleration and plateaus." *PLoS Biology*, 2018. {A modelling argument that plateaus can be generated by age-reporting error; it does not present newly validated ages.}

[^newman2018]: `paper` Newman, S.J. "Errors as a primary cause of late-life mortality deceleration and plateaus." *PLoS Biology*, 2018. {A modelling argument that plateaus can be generated by age-reporting error; it does not present newly validated ages.}

[^robine2019]: `paper` Robine, J.-M., Allard, M., Herrmann, F.R. and Jeune, B. "The real facts supporting Jeanne Calment as the oldest ever human." *The Journals of Gerontology: Series A*, 2019. {Written by members of the team that validated the record originally, replying to the identity-substitution claim.}

[^robine2019]: `paper` Robine, J.-M., Allard, M., Herrmann, F.R. and Jeune, B. "The real facts supporting Jeanne Calment as the oldest ever human." *The Journals of Gerontology: Series A*, 2019. {Written by members of the team that validated the record originally, replying to the identity-substitution claim.}

[^lq2026]: `report` LongeviQuest. "Oldest People Ever." Validated-age database, accessed 1 November 2026. https://longeviquest.com/atlas/oldest-people-ever/ {A continuously updated list; living people's ages and ranks change daily.}

[^robine1999]: `book` Robine, J.-M. and Allard, M. "Jeanne Calment: Validation of the duration of her life." In Jeune, B. and Vaupel, J.W. (eds), *Validation of Exceptional Longevity*. Odense University Press, 1999. https://www.demogr.mpg.de/books/odense/6/09.htm

[^grg2026]: `report` Gerontology Research Group. "Ethel Caterham, world's oldest living person, celebrated her 117th birthday." 21 August 2026. https://www.grg-supercentenarians.org/2026/08/21/ethel-caterham-worlds-oldest-living-person-celebrated-her-117th-birthday/ {States that the GRG validated her case on 28 January 2023.}

[^gwr2026]: `report` Guinness World Records. "Oldest person living." Record entry verified 1 April 2026. https://www.guinnessworldrecords.com/world-records/84549-oldest-person-living

[^zak2019]: `paper` Zak, N. "Evidence that Jeanne Calment died in 1934, not 1997." *Rejuvenation Research*, 2019. doi:10.1089/rej.2018.2167 {The challenge to the record; single author, assembled from interviews, biographies, photographs and archival documents rather than new vital records.}

[^zak2020]: `paper` Zak, N. and Gibbs, P. "A Bayesian assessment of the longevity of Jeanne Calment." *Rejuvenation Research*, 2020. doi:10.1089/rej.2019.2227

[^rw2017]: `news` Han, A.P. "Widely publicized Nature study on human age limit draws fire." *Retraction Watch*, 28 June 2017. https://retractionwatch.com/2017/06/28/widely-publicized-nature-study-human-age-limit-draws-fire

[^hughes2017]: `paper` Hughes, B.G. and Hekimi, S. "Many possible maximum lifespan trajectories." *Nature*, 2017. doi:10.1038/nature22786

[^lenart2017]: `paper` Lenart, A. and Vaupel, J.W. "Questionable evidence for a limit to human lifespan." *Nature*, 2017. doi:10.1038/nature22790

[^debeer2017]: `paper` de Beer, J., Bardoutsos, A. and Janssen, F. "Maximum human lifespan may increase to 125 years." *Nature*, 2017. doi:10.1038/nature22792

[^einmahl2019]: `paper` Einmahl, J.J., Einmahl, J.H.J. and de Haan, L. "Limits to human life span through extreme value theory." *Journal of the American Statistical Association*, 2019. doi:10.1080/01621459.2018.1537912 {Dutch-born residents only; the analysis uses all deaths at 92 or older rather than only the oldest few.}

[^belzile2021]: `paper` Belzile, L.R., Davison, A.C., Rootzén, H. and Zholud, D. "Human mortality at extreme age." *Royal Society Open Science*, 2021. doi:10.1098/rsos.202097

[^belzile2022]: `paper` Belzile, L.R., Davison, A.C., Gampe, J., Rootzén, H. and Zholud, D. "Is there a cap on longevity? A statistical review." *Annual Review of Statistics and Its Application*, 2022. doi:10.1146/annurev-statistics-040120-025426

[^pearce2021]: `paper` Pearce, M. and Raftery, A.E. "Probabilistic forecasting of maximum human lifespan by 2100 using Bayesian population projections." *Demographic Research*, 2021. doi:10.4054/DemRes.2021.44.52 {Builds on an exponential survival model for supercentenarians; the probabilities assume that model holds.}

[^pyrkov2021]: `paper` Pyrkov, T.V. et al. "Longitudinal analysis of blood markers reveals progressive loss of resilience and predicts human lifespan limit." *Nature Communications*, 2021. doi:10.1038/s41467-021-23014-1 {Funded by Gero, which employed most authors; the 120-150 figure is a linear extrapolation beyond the observed ages.}

[^olshansky2024]: `paper` Olshansky, S.J., Willcox, B.J., Demetrius, L. and Beltrán-Sánchez, H. "Implausibility of radical life extension in humans in the twenty-first century." *Nature Aging*, 2024. doi:10.1038/s43587-024-00702-3 {About period life expectancy and survival to 100, not the maximum age at death.}

[^carey1992]: `paper` Carey, J.R., Liedo, P., Orozco, D. and Vaupel, J.W. "Slowing of mortality rates at older ages in large medfly cohorts." *Science*, 1992. doi:10.1126/science.1411540

[^kenyon1993]: `paper` Kenyon, C., Chang, J., Gensch, E., Rudner, A. and Tabtiang, R. "A C. elegans mutant that lives twice as long as wild type." *Nature*, 1993.

[^ruby2018]: `paper` Ruby, J.G. et al. "Estimates of the heritability of human longevity are substantially inflated due to assortative mating." *Genetics*, 2018. doi:10.1534/genetics.118.301613 {Built on Ancestry public family trees covering birth cohorts of the 1800s and early 1900s.}

[^degrey2004]: `statement` de Grey, A.D.N.J. "Escape velocity: Why the prospect of extreme human life extension matters now." *PLoS Biology*, 2004. doi:10.1371/journal.pbio.0020187 {An essay setting out the author's own position, conditional on a sustained 10% annual fall in mortality at all ages.}

[^warner2005]: `paper` Warner, H. et al. "Science fact and the SENS agenda: What can we reasonably expect from ageing research?" *EMBO Reports*, 2005. {A signed commentary by 28 biogerontologists rather than a research paper; it reports their judgment of the SENS proposals.}

[^uidaho2001]: `statement` University of Idaho. "Scientists bet $500 million on whose aging theory will prove best in 2150." Press release, 11 January 2001. https://www.newswise.com/articles/scientists-bet-500m-on-whose-aging-theory-will-prove-best-in-2150

[^harrison2009]: `paper` Harrison, D.E. et al. "Rapamycin fed late in life extends lifespan in genetically heterogeneous mice." *Nature*, 2009. doi:10.1038/nature08221

[^kealy2002]: `paper` Kealy, R.D. et al. "Effects of diet restriction on life span and age-related changes in dogs." *Journal of the American Veterinary Medical Association*, 2002. {Control dogs were their pair-mates fed 25% more, so they were heavier and fatter; the study ran in one pet-food company's colony.}

[^tamu2024]: `statement` Texas A&M School of Veterinary Medicine. "Dog Aging Project receives $7 million NIH grant to expand clinical trial of anti-aging drug." Press release, 9 December 2024.

[^loyal2026]: `statement` Loyal. "Clinical trials." Company web page, accessed 1 November 2026. https://loyal.com/clinical-trials {The company describing its own trial; treatment assignments stay blinded until the study ends.}

[^colman2009]: `paper` Colman, R.J. et al. "Caloric restriction delays disease onset and mortality in rhesus monkeys." *Science*, 2009. doi:10.1126/science.1173635 {Reported a lower incidence of age-related deaths; at the time point reported, 80% of restricted and 50% of control animals had survived.}

[^mattison2012]: `paper` Mattison, J.A. et al. "Impact of caloric restriction on health and survival in rhesus monkeys from the NIA study." *Nature*, 2012. doi:10.1038/nature11432

[^mattison2017]: `paper` Mattison, J.A. et al. "Caloric restriction improves health and survival of rhesus monkeys." *Nature Communications*, 2017. doi:10.1038/ncomms14063

[^tpr2024]: `news` "Science & Medicine: Anti-aging drug shows promise in marmosets." *Texas Public Radio*, 28 July 2024. {Segment co-produced with UT Health San Antonio, the researcher's employer; reports no paper or effect size.}