The older we get, the greater our likelihood of dying. Or at least that’s what we thought. New research suggests mortality rates level off after we turn 105, and that no upper limit exists for the human lifespan. It’s an extraordinary conclusion — one that’s not going over very well with other ageing researchers.
Back in 1825, British mathematician and actuary Benjamin Gompertz noticed a strange quirk about human aging and mortality. Our risk of dying, he observed, increases exponentially as we get older. Scientists now refer to this as the Gompertz Law of Mortality, and it has largely remained intact since it was first documented nearly 200 years ago.
Once we enter into adulthood, our chance of dying doubles every eight years or so. But as Gompertz himself admitted, there simply isn’t enough data to prove that this trend holds true for the oldest of the old. Research published over the last 10 years, however, continues to affirm Gompertz’s model, including a 2017 study which showed that supercentenarians — people who live 110 years or more — still experience accelerated mortality rates over time.
In recent years, however, some scientists have started to question the Gompertzian paradigm. A controversy has emerged that suggests mortality plateaus—in which death rates slow down and even grind to halt — exist for both humans and animals beyond a certain age. There’s now an active debate among ageing experts as to whether mortality rates continue to accelerate and/or rise exponentially into extreme age, or if they eventually plateau.
New research published today in Science suggests a mortality plateau does indeed exist for humans, and that the Gompertzian model breaks down after the age of 105. The authors, led by Elisabetta Barbi from the Department of Statistics at Sapienza University of Rome, say their data suggests there’s no upper limit to life—at least, not any limit that has yet been observed.
Some of the experts Gizmodo spoke to, however, said this rather jaw-dropping conclusion is completely unwarranted, and that the researchers made numerous mistakes and misinterpretations to achieve their result.
A photo from 2013 showing President Barack Obama meeting with Richard A. Overton. At the time, Overton was 107 years old. Today, he’s 112, and the oldest living man in the United States.
A fundamental problem in ageing research is the extreme dearth of data. Only a handful of people live beyond the age of 105, and even fewer live beyond 110. For the new study, Barbi and her colleagues sought to overcome this limitation by taking a look at data collected by the Italian National Institute of Statistics (ISTAT). This unique dataset contains recently collected and validated information about individual survival rates of all Italian residents aged 105 and older from January 1, 2009 to December 31, 2015. The researchers used this data to create their own unique database of super-old people, and to produce estimates of mortality rates at extreme ages.
For the analysis, the researchers looked at 3,836 individuals, of which only 463 were male (12 percent). Fewer than 4 percent of these individuals were born abroad, and a significant portion came from clear Italian heritage. Some 2,880 deaths were observed in this sample pool over the seven years, meaning the remaining individuals were still alive at the time the study was completed. Cause of death was not a factor, as this study was strictly concerned with mortality rates.
Graph showing chances of death (hazard) by age (years), with the plateau (shown in orange) after age of 105.
The data revealed that death rates increased exponentially up to about the age of 80, but decelerated afterwards, reaching or closely approaching a plateau after the age of 105. Across both women and men, the chance of dying at 68 was around 2 percent, and the chance of dying at 76 was about 4 percent. At 97, the odds jumped to around 30 percent, and by 105 the chance of death reached 60 percent—but it held steady beyond that.
“[The] Gompertz model does not appear to hold, fails at older ages and far overshoots our estimated plateau beyond age 105,” write the researchers in the study. What’s more, the data suggests mortality rates actually decline after 105, though ever so slightly, among those born in the same year.”
They go on: “The increasing number of exceptionally long-lived people and the fact that their mortality beyond 105 is seen to be declining across cohorts [those born in the same year] — lowering the mortality plateau or postponing the age when it appears—strongly suggest that longevity is continuing to increase over time and that a limit, if any, has not been reached. Our results contribute to a recently re-kindled debate about the existence of a fixed maximum lifespan for humans, underwriting doubt that any limit is as yet in view.”
The oldest known person to have ever lived died at age 122, but the reason we haven’t seen anyone live to 140, or 200, or beyond, is that a 60 percent chance of dying each year all but guarantees our eventual demise — at least, according to this paper.
This analysis was purely statistical and demographic in nature, so the authors didn’t provide any tangible or meaningful explanations for the observed mortality plateau, aside from the suggestion that structural (e.g. improved healthcare for the extreme elderly) and evolutionary factors (e.g. the mutation accumulation theory and age-related genetic factors) are likely at play.
Brandon Milholland, an ageing researcher in the Department of Genetics at Albert Einstein College of Medicine who wasn’t involved with the new study, said the new paper is interesting, but he believes it’s only a small piece of a much larger puzzle.
“It’s based on seven years of data in one country, and most of the data applies for ages 105 to 108,” Milholland told Gizmodo. “I would say this paper does not tell us much, if anything, about supercentenarian mortality. It is also questionable if the results are generalisable to other countries.”
The finding that late-life mortality differs from the Gompertz law, he said, is not especially surprising.
“For most of adult life, mortality doubles every eight years or so,” said Milholland. “This paper makes sense in that the doubling cannot continue indefinitely. If mortality is at 60 percent, it cannot eventually double to 120 percent — that is mathematically impossible. However, this paper models late-life mortality as suddenly stopping in its tracks and remaining completely flat. That is highly implausible.”
Milholland said a levelling off is understandable, but that doesn’t mean it has to remain flat.
“In fact, it seems rather far-fetched that after increasing exponentially, the chance of dying should suddenly stop in its tracks,” he said. “It is possible to imagine any number of curves that continue to increase without exceeding 100 percent. I am disappointed that a false dichotomy has been presented between the Gompertz law and a complete plateau. This criticism is not to single out this particular paper; to my knowledge, no one has examined late-life mortality with a more sophisticated model. However, I do not consider the evidence for a plateau presented in this paper to be especially strong.”
S. Jay Olshansky, a professor in the School of Public Health at the University of Illinois at Chicago, also has serious misgivings about the new paper — though he said the authors did an excellent job of creating a new set of data for researchers to evaluate.
“The key thing, however, is that researchers had to work really hard to create such a dataset because so few people live out to extreme old age,” Olshansky told Gizmodo. “Why is that you ask? Because there’s a limit to human longevity that is fundamentally influenced by our basic biology and the body design we inherited from our ancestors, so when people live past age 105, they’ve approached about the maximum that humans can live. So the very problem they’ve tried to overcome — which is to create a larger dataset of rare extreme survivors — tells us that the conclusion that they’ve come to, which is that there is no upper limit to life, is an unreasonable conclusion. I don’t understand why the authors can’t see this.”
Ultimately, the observation that death rates at extreme old age level off around 105, or continue to rise, doesn’t matter, he said, because by the time people reach these ages, at least half disappear every year.
“If 100 people survive to age 110 out of billions, what difference does it make if 50 or 60 die before their next birthday? Human bodies are not intended for long-term use, and when we do manage to get them to operate past a century, plenty of age-related diseases accumulate,” said Olshansky. “Debating about the mortality rates of a small number of people at extreme ages, and concluding anything about the longevity of humans in general, is equivalent to describing the average running speed for the one-mile run for all of humanity by examining world records for running set by Olympic athletes. The tail of the survival distribution tells us practically nothing about human longevity.”
“There just aren’t enough people surviving past age 110 to generate reliable death rates. Why? Because there’s a limit to duration of life. How obvious is that?”
Olshansky said the observation that death rates might level off at older ages is not unexpected, but evidence exists to the contrary.
“Remember, these folks have been highly selected by time, so the only people that can live this long had to have won the genetic lottery at birth,” he said. “The rest of us have almost no chance of living that long. The conclusion that there is no limit to life does not follow from these data. The fact that there is a limit to life should come from the fact that it’s just so difficult to find enough people at extreme old age to generate reliable death rates. For this dataset, there just aren’t enough people surviving past age 110 to generate reliable death rates. Why? Because there’s a limit to duration of life. How obvious is that?”
As far as Olshansky is concerned, these findings should have no bearing on how we study ageing and develop therapeutic interventions. Instead, he said scientists should try to figure out why people under the age of 100 are dying, and find ways to slow down their ageing. “I’m less interested in the tails of the distribution — except to figure out what’s so unique about them — I’m much more interested in the rest of humanity that won’t live that long,” he said.
But not all scientists see it this way.
Siegfried Hekimi, a biology professor at McGill University in Montreal, said the conclusion of the paper, that age-specific mortality does not increase any more after a certain age, is reasonable.
“This conclusion is very firm for the dataset which is examined,” Hekimi told Gizmodo. “And the dataset is very good. If age-specific mortality is not increasing anymore after a certain age, then some individuals can make it to a very great age indeed.”
Regardless of where one stands on this paper, the next steps for scientists are reasonably clear: get better data, and don’t just rely on statistics and demography. What’s more, the only way to find out the true answer to this mystery will be to delve into biology itself. [Science]