The reliability and stability of biomarkers of aging.

Different types of stability of a biomarker are important properties, influencing the degree of predictability across age (ordinal stability) and the interpretation of quantitative and qualitative change with age (structural stability). These properties may be expected to differ from biomarker to biomarker and may change with age. Any age-related process with individual differences in time of onset of change or in rate of change will necessarily display reduced ordinal stability. Another source of reduced correlation across occasions is the short-term fluctuance of individuals due to cyclic processes and to responsiveness to environmental displacements of biomarker values and recovery therefrom. Structural stability of composite variates may be quite high across relatively short intervals but sufficiently low across longer intervals as to suggest the inappropriateness of simple description of mean changes or differences across these longer time spans. The outcome with multivariate composites raises the issue that single biomarkers may have quite different meanings at different parts of the life span.