If the Census Bureau is right, the number of people aged 85 or older will grow from 3 million today to 18.7 million in 2080. But if a researcher working with the National Institute on Aging is correct, the number of these oldest old in 2080 could actually be 72 million, a figure nearly four times as great as the Census Bureau's projection.
This striking disparity involves a fundamental disagreement about the nature of aging.
"If you assume that little or no progress can be made in reducing the death rates of the oldest old, the cohort will grow a lot," said Dr. James Vaupel, a demographer at Duke University who directed research groups for the Institute on Aging. "But if you assume the death rates among the oldest old will come down, the numbers will explode."
The more conventional view, held by the Census Bureau, Social Security and some gerontologists, is that there is a natural limit to the human life span -- about 85 years on average -- and that number is never going to budge much.
But other researchers believe there is no biological limit to how long people can live. The average life span has been steadily increasing since the 1950s, they note. Because the population is generally healthier, people are entering old age in better shape than they used to, so life spans can be expected to continue to increase into the next century.
A leading advocate of this view, Dr. Vaupel projects that by 2080 the average life span will be 94 for men and 100 for women.
If these researchers are correct, the trend will have major effects on retirement, Social Security funds and pensions.
"That's why this debate is taking place," said Dr. S. Jay Olshansky, a demographer at the University of Chicago.
The Institute on Aging has supported four research projects looking at subjects from fruit flies to Danish twins, to ask which hypothesis of aging is correct. The studies, now reaching completion, are consistently failing to show any evidence for a pre-programmed limit to human life span, said Dr. Vaupel, who directed the research. He added that, in his opinion, a life span of 100 was not out of the question for people, like his own 8-year-old daughter, who are alive today.
Not everyone is convinced. Dr. Olshansky, for example, says it is extremely difficult to keep increasing the human life span. Some gains were made in recent years, he said, but it becomes harder and harder to keep up the pace. In fact, he said, even to get to an average life span of 85 from today's average span of 77.5, "you would have to reduce death rates by 55 percent at every age. That's not easy to achieve."
Dr. James Fries of Stanford University comes to a similar conclusion. "Almost every curve you have -- health or economics -- is a curve of diminishing returns. It just gets harder and harder" to push out the average age of death, he said.
But, said Dr. Richard Suzman, director of the office of demography at the NIA, the institute's four projects argue otherwise.
The research, which involves scientists from the United States, Sweden, Denmark and Finland, was initiated four years ago. "When we started out," Dr. Vaupel said, "I didn't hold to either hypothesis."
One project asked whether it was true, as some scientists have said, that little progress has been or can be made in bringing down the death rates of people over age 85. There is little lTC evidence to confirm or refute that belief because most countries do not publish data on death rates after age 85, Dr. Vaupel said. In the United States, it is particularly difficult to test the hypothesis because birth certificates were not required until 1933.
But Sweden has much better records with long-term data -- going back to 1750 -- on population sizes and death rates at even the very oldest ages, Dr. Vaupel said. And when he and a colleague, Dr. Hans Lundstrom of the Swedish statistics office, meticulously checked the Swedish records, combing data in the archives of the Swedish national statistics office in Stockholm, they came to a surprising conclusion. The mortality rate among those over age 85 remained steady until about 1950 but has dropped steadily since then at a rate of about 2 percent a year.
The Swedish figures were "a gold standard data set," Dr. Vaupel said. Now, Dr. Vaino Kannisto, a Finnish demographer, has gathered mortality data for the oldest old from 27 countries and finds that, like the elderly in Sweden, the oldest old elsewhere have had a steady, sustained increase in their life spans since 1950.
Another project asked whether there is a genetically preordained time of death. Or, in other words, is the old nostrum correct that the way to live to a ripe old age is to pick long-lived ancestors?
To test this hypothesis, Dr. Niels Holm, of the Odense Medical School in Denmark and director of the Danish Twin Registry, and Dr. Vaupel examined the registry, with data on 55,000 twins, identical and fraternal, born since 1870. "We could test the conjecture that identical twins will die at about the same time," Dr. Vaupel said. "It's just not true."
He added that genes do matter. Some people are genetically frail and others are genetically robust. And there is a genetic predisposition to certain causes of death. But, he said, "there is no evidence that genes dictate a maximum life span." Only 2 to 3 percent of the variance in ages of death in the population is explained by genetics, he said.