The Secrets of Long Life from Longitudinal Data

Anatoliy I. Yashin, Duke University
Igor Akushevich, Duke University
Konstantin G. Arbeev, Duke University
Liudmila Akushevich, Duke University
Svetlana V. Ukraintseva, Duke University

In this paper we perform analysis of extreme longevity using longitudinal data collected in the Framingham Heart Study (FHS). For these purposes we identified sub-cohorts of individuals (males and females), having potential to live long life. We evaluated age trajectories of physiological indices for these sub-cohorts and compared them with those for the rest of the FHS participants. We found that average trajectories of such indices as well as survival curves in these groups of individuals differ significantly. Then we developed mathematical and computer models of human mortality and aging appropriate for the analysis of longitudinal data and evaluated dynamic characteristics of risk functions. We found that the notion of “normal” physiological state (corresponding to minimal risk) depends on age and estimated respective age trajectories. The relations between the dynamic properties of risk functions and changes in stress resistance with age are discussed.