Researchers Use Statistical Technique To Find Mix Of Biomarkers Predicting Mortality

In this study, the researchers analyzed 13 biomarkers representing neuroendocrine and vascular functions and immune and metabolic activity over 12 years in 1,189 high-functioning men and women aged 70-79 enrolled in the MacArthur Study of Successful Aging.

The goals were to identify biomarker combinations (or high-risk pathways) associated with high levels of mortality in men and women, determine whether the biomarkers that most closely predicted mortality differed in men and women and develop prediction rules based on combinations of biomarker conditions.

The researchers also sought to present "recursive partitioning," a statistical technique for identifying multiple and interacting predictors of an outcome, as a useful analytical tool for addressing research questions in the biomedical sciences. Using recursive partitioning, they found that combinations of neuroendocrine and immune markers frequently appeared in high-risk male pathways, while systolic blood pressure was present, in combination with other biomarkers, in high-risk female pathways.

This work suggests that clinicians and researchers may be able to use recursive partitioning to identify the biological regulatory system's importance in predicting mortality in later life.

Authors of the study include Tara L. Gruenewald, Teresa E. Seeman and Arun S. Karlamangla of UCLA; Carol D. Ryff of the University of Wisconsin, Madison; and Burton H. Singer of the Institute on Aging at the University of Wisconsin, Madison, and the Office of Population Research, Princeton University.

The study appears in the Sept. 19 issue of the Proceedings of the National Academy of Sciences, available online at http://www.pnas.org/cgi/doi/10.1073/pnas.0606215103

Funders of the research include the National Institute on Aging and the John D. and Catherine T. MacArthur Foundation.