The usual story about longevity involves people in their eighties and nineties. A study published in Nature Communications found the signal much earlier than that, in the first month of life.
Researchers led by Matthew Thomas Keys at the University of Southern Denmark tracked four generations descending from 659 Danish families selected for exceptional longevity. The founding families each contained two or more siblings who had reached at least 88 and were alive when recruited between 2006 and 2009, and 99.5 percent eventually had two or more siblings who survived to at least 90.
Their grandchildren, born decades later, had roughly half the infant mortality of demographically matched Danish babies. The hazard ratio was 0.53, with a confidence interval running from 0.36 to 0.77.
The Survival Curves Split Within 30 Days
The study compared 5,637 third-generation grandchildren and 14,908 great-grandchildren with 41,090 matched control births, using registries that covered every live birth in Denmark from 1973 to 2018. Matching accounted for sex, birth year, birth season, maternal birth year, parity, and whether the mother was born in Denmark.
The shape of the survival curves is the interesting part. Grandchildren separated from controls within the first 30 days of life, then ran parallel afterward. Whatever advantage they carried was concentrated in the neonatal period rather than accumulating gradually through infancy.
Other birth outcomes moved in the same direction. Grandchildren had lower odds of preterm birth (0.82), of being small for gestational age (0.83), and of neonatal respiratory disorders (0.77), and all three survived correction for multiple comparisons. Every other outcome examined pointed in the same direction without reaching significance, which the authors interpreted as a broad advantage rather than protection against a particular condition.
The absolute differences are small, which is worth holding onto. Infant mortality in modern Denmark is low for everyone, so halving a small risk still leaves a small risk.
The Advantage Thinned Between Generations, and Not Because Everyone Else Caught Up
In the fourth generation, the effect largely dissolved. Great-grandchildren had an infant mortality hazard ratio of 0.90, with a confidence interval from 0.70 to 1.17 that crosses 1.0.
The obvious explanation is that Danish neonatal care improved so much over the study period that the general population caught up. The authors anticipated the objection and tested it directly, comparing the third and fourth generations against each other during the years when their birth cohorts overlapped, roughly 1973 to 2010.
The gap held. Grandchildren had a hazard ratio of 0.32 (95% CI, 0.16-0.64) compared with great-grandchildren, independent of secular trends in the background population. A battery of negative control comparisons supported this: fourth-generation children did not differ in survival from third-generation controls, and the two control groups did not differ from each other. The researchers read this as genuine dilution, a partial loss of protective factors as each generation married outside the original families, rather than a catch-up effect.
That fits earlier work from the same group showing that spouses who marry into these families survive better than the general population but less well than their partners. The analysis was first posted as a preprint and is indexed in PubMed.
Why Money and Cigarettes Do Not Explain It
The most testable alternative was that these are simply advantaged families. They are, somewhat. Parents in both generations were more likely to hold higher qualifications, with odds ratios of 1.28 in the third generation and 1.34 in the fourth, and mothers smoked less during pregnancy, with odds ratios of 0.80 and 0.75.
Socioeconomic position does shape lifespan, though the picture is more specific than it is often described. A Mendelian randomization analysis in Nature Human Behavior found a causal effect of education on longevity but no independent effect from income or occupation, a result summarized in an accompanying Nature Aging commentary.
In this Danish cohort, adjusting for parental education barely moved the survival estimates. And the behavioral advantages stayed stable across generations while the infant survival advantage collapsed, which is hard to reconcile with behavior driving the effect.
There was also no difference depending on whether the mother or father descended from a long-lived family. The interaction estimate was 1.06, with a confidence interval of 0.50 to 2.28, which argues against explanations resting solely on the maternal environment during pregnancy.
The authors are careful about what remains. This is observational, so it cannot establish that any single early-life factor determines lifespan. Their socioeconomic and behavioral measures were limited, smoking data were missing for much of the third generation, and the fourth generation is too young to know how its adult health will unfold. Nothing here identifies a gene, a mechanism, or anything an expectant parent could act on.
What it does suggest is that whatever runs in these families is already at work before the first birthday, and that it is diluting generation by generation.
Key Questions Answered
What is a longevity-enriched family?
Here, a family with two or more siblings who reached at least 88 and were alive at recruitment. Nearly all eventually had two siblings reach 90.
What was the main finding?
Grandchildren of these families had about half the infant mortality of matched Danish babies, with survival curves separating within the first 30 days of life.
How large is the effect in absolute terms?
Small. Infant mortality is low across the Danish population, so a halved relative risk still translates into a modest absolute difference.
Why did the advantage weaken in the next generation?
The researchers compared generations directly during overlapping birth years and still found a gap, suggesting real dilution of protective factors rather than improved care lifting everyone.
Could wealth and education explain it?
The families were better educated and smoked less, but adjusting for education barely changed the results, and those advantages held steady while the survival effect faded.
Does this identify a longevity gene?
No. The study is observational and register-based, implicating heritable genetic or epigenetic factors without identifying any specific gene or mechanism.
