When Does Human Longevity Start?: Demarcation of the Boundaries for Human Longevity

Interestingness:5

By Natalia S. Gavrilova and Leonid A. Gavrilov. Journal of Anti-Aging Medicine. June 2001, 4(2): 115-124. doi:10.1089/10945450152466161.

More Gavrilovs.  This one is another paper where they continue to extract results from the historical records of the european aristocracy from the 1600s onwards. In this one, they focus on the period 1800-1880s so as to have complete data for women (recent records are more complete) but also to make sure everyone in it is dead.

They analyse longevity of 5800 daughters that made it to adulthood (30 years) compared to longevity of their mothers and fathers (separately). They avoid sons due to frequent military deaths. They find almost no correlation if the mother died before 85 years of age, but strong correlation otherwise, with a line of best fit with a slope of 0.412±0.204.  Similarly for the father, with no correlation if he died before 75 years of age, but a strong correlation with a slope of 0.236±0.078 otherwise. They don't show results with a combined mother plus father variable. The graphs are quite nice.

There's something somewhat dodgy about the method since they were looking for a piecewise linear regression, and they determined where to start the significant line by inspecting the graph visually. I don't know how much that detracts from the result.

Like they mention, this does "resolve" the contradictions found where some research says that the genetic component of longevity is small, while others show that people with very old age are bunched in families.

Season of Birth and Human Longevity

Summary: Adult women over thirty live three years longer if they were born in May or December, rather than in August.

Interestingness: 2

Paper by Leonid A Gavrilov and Natalia S Gavrilova in the Journal of Anti-Aging Medicine, Volume 2, Issue 4, Winter 1999.

(((
Short two-page paper analysing a subset of the same data they used for their longevity vs fertility paper, ie European aristocratic families. In this case, they analysed a lot of variables until they found one that correlated with longevity.

On the relevant subsample of 4911 women, adult women over 30 born between 1800 and 1880 lived shortest if they were born in August and longest if they were born in May with the difference at about three and a half years. This is after correcting for a whole heap of variables that have nothing to do with month of birth but are related to longevity: year of birth, maternal and paternal life spans, age of parents at birth, birth order, nationality, whether the death was violent, loss of either or both parents before age twenty.

They have a graph and it doesn't look good to me. The only reasons I can think of, and that they propose, for the effect would be availability of vitamins (or calories but they are unlikely to be a problem for this group since they were all from aristocratic families) at specific points in the pregnancy or early life, but the graph is very noisy and it mostly jumps up and down. For example, the difference between July and August births is two and a bit years, and between August and September the difference is about a year and a half, with August at the minimum. The other bad months are February and March, on the other side of the year, with women born then living about a year longer than those in August. If the effect is real, then the critical periods during pregnancy must be very short.

I'm not buying any of it until it gets replicated.
)))

Is There a Reproductive Cost for Human Longevity? and Human Longevity and Reproductive Success: Response to Gavrilov and Gavrilova

Summary: Critique and response of a study of an implicit tradeoff between number of offspring and mortality.

Interestingness: 1

Paper by Leonid A Gavrilov and Natalia S Gavrilova in the Journal of Anti-Aging Medicine, Volume 2, Issue 2, Summer 1999.
Response by Rudi GJ Westendorp and Thomas BL Kirkwood in the same issue.

(((
This is a critic of a paper that appeared in Nature about a trade off between number of offspring and longevity among the Brittish aristocracy going back to the year 740. Reasonable sounding critique, and reasonable sounding response. The abstract below is a good summary of the critique. The reply shows that adjusting for each of the items in the critique doesn't change the resultant mortality increase associated with having 2 or more children (1.15), and that they get different results from other studies because they restricted themselves to a homogenous group.
)))


Abstract of first paper follows (second one lacked an abstract):

This is a critical review of the recent claims by Westendorp and Kirkwood that human longevity is achieved at the cost of reproductive success. The criticism could be summarized in four statements. (1) Declaring that long-lived women have less progeny and older age at first childbirth, the authors failed to adjust the data for the age at marriage—the most important explanatory variable both for the number of children and for the age at first childbirth; (2) they also overlooked another important confounding variable—the husband's fertility; (3) the authors used the data that are inappropriate for fertility studies—extremely ancient and incomplete genealogies with many underreported records for daughters that led to incorrect estimates for the number of progeny and for the age at childbirth; and (4) the authors presented their study as completely new for humans and did not quote the opposite results from the earlier study by Le Bourg et al. where no trade-off between human longevity and fertility was observed. They also ignored findings of Bideau and of Knodel that the most fertile women live longer than the remainder or at least not shorter contrary to the author's claims. Thus, the conclusions of Westendorp and Kirkwood are inconsistent with the existing knowledge and should be reanalyzed using more appropriate methods and data.