Since the early 20th century, sustained declines in mortality have driven steady increases in life expectancy, particularly in high-income countries. Several studies have reported that this improvement occurred in a near-linear fashion: a narrative often echoed in media headlines as if it must continue indefinitely.

However, a recent study in the Proceedings of the National Academy of Sciences using human cohort mortality data challenges that assumption, using human cohort mortality data.

Mortality Database

The research team, led by Héctor Pifarré i Arolas, PhD, of the University of Wisconsin-Madison, analyzed cohort mortality in 23 high-income countries using the Human Mortality Database.

Researchers collected mortality data for cohorts born between 1850 and 1938, which were considered fully observed for practical purposes. They estimated the historical rates of increase in cohort life expectancy from these mortality rates. For cohorts born between 1939 and 2000, predictive models were applied to the available data to generate life expectancy estimates from the resulting life tables.

The study used six different methods to forecast future mortality rates, including Lee-Carter, compositional data analysis and cohort segmented transformation age-death distributions, adding robustness to the conclusions. Sensitivity analyses included cross-validation, alternative prediction scenarios, and decomposition by age group and cause of death. Together, these findings suggest that the observed slowdown reflects real patterns rather than methodological artifacts.

Cohort Life Expectancy

Unlike the more commonly reported period of life expectancy, which reflects age at the year of death, cohort life expectancy measures the longevity of individuals born in a specific year. Speaking to Univadis Spain, a Medscape Network platform, Pifarré explained, “Previous studies have examined and predicted life expectancy over a period, not a cohort. Period life expectancy measures the years a hypothetical cohort would live if they faced current mortality conditions throughout life.”

He added, “This means that when there is a mortality crisis, such as the recent pandemic, period life expectancy suffers a greater decline than actual cohorts will experience. This is not a hypothesis: many recent, now extinct, cohorts experienced pandemics and wars yet still achieved large improvements in longevity.”

Cohort life expectancy provides a more accurate picture of longevity across successive birth years. The drawback is that it can only be measured once a cohort has completed its life cycle; therefore, predictive models are required for living cohorts. 

Pifarré emphasized, “Ultimately, the longevity of people born in a certain year can only be directly assessed through cohort life expectancy. Our research uses statistical prediction techniques to complete mortality data for cohorts born up to 2000 to estimate their life expectancy.”

Longevity Slowdown

Data from cohorts born between 1850 and 1938 show an average life expectancy gain of 0.46 years per cohort. In contrast, projections for cohorts born between 1939 and 2000 indicate a reduction in improvement of 37%-52%, depending on the method. Average gains per cohort shrink to 2.5-3.5 months, compared with 5.5 months in earlier cohorts.

Discussing the slowdown, Pifarré noted, “Perhaps the most surprising aspect is that most of the slowdown in longevity is explained not by future mortality but by past mortality. To a large extent, the large improvements in life expectancy in the first cohorts of the 20th century are explained by the reduction in mortality at young ages (under 20 years).” 

Many of these gains, such as better nutrition, medical treatment, and pediatric care, have already been achieved, and infant and child mortality rates are now low.

“The cohorts alive today had little room for improvement at younger ages,” he said. “Predicted gains at older ages cannot offset the loss of this driver of life expectancy growth.” Future improvements may come from reducing midlife mortality, including biomedical advances in age-related conditions and behavioral interventions, such as smoking cessation, diet, and exercise.

No cohort born between 1939 and 2000 is projected to reach a life expectancy of 100 years. 

Pifarré clarified, “Our results do not indicate that life expectancy will stop increasing. Growth will likely continue, but at a slower pace.” He cautioned that unexpected events, such as epidemics, social crises, or major medical breakthroughs, could significantly alter mortality trends.

Another unresolved question is the potential biological limits of human life expectancy. Considerations include public health improvements, medical advances, and socioeconomic development, raising the debate about whether human longevity has biological limits and whether these can be overcome.

This story was translated from Univadis Spain .