A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: A cohort study

• DOI: 10.1371/journal.pmed.1002718
• Publication Date: 2018-12-31T13:29:47Z
• Authors: Zuyun Liu; Pei-Lun Kuo; Steve Horvath; Eileen Crimmins; Luigi Ferrucci; Morgan Levine
• Keywords: Adult; Male; Aging; Biomedical and clinical sciences; 610; Medical and Health Sciences; Cohort Studies; Young Adult; 03 medical and health sciences; Clinical Research; Risk Factors; General & Internal Medicine; Health Sciences; Behavioral and Social Science; 80 and over; Humans; Aetiology; Mortality; Aged; Aged, 80 and over; 0303 health sciences; Prevention; R; Health sciences; Middle Aged; Nutrition Surveys; 3. Good health; Good Health and Well Being; Population Surveillance; Medicine; Female; Public Health; Morbidity; 2.4 Surveillance and distribution; Research Article; Follow-Up Studies

Background A person's rate of aging has important implications for his/her risk death and disease; thus, quantifying using observable characteristics applications clinical, basic, observational research. Based on routine clinical chemistry biomarkers, we previously developed a novel measure, Phenotypic Age, representing the expected age within population that corresponds to estimated mortality risk. The aim this study was assess its applicability differentiating variety health outcomes diverse subpopulations include healthy unhealthy groups, distinct persons with various race/ethnic, socioeconomic, behavior characteristics. Methods findings Age calculated based linear combination chronological 9 multi-system biomarkers in accordance our established method. We also Acceleration (PhenoAgeAccel), which represents after accounting (i.e., whether person appears older [positive value] or younger [negative than expected, physiologically). All analyses were conducted NHANES IV (1999–2010, an independent sample from originally used develop measure). Our analytic consisted 11,432 adults aged 20–84 years 185 oldest-old top-coded at 85 years. observed total 1,012 deaths, ascertained over 12.6 follow-up (based National Death Index data through December 31, 2011). Proportional hazard models receiver operating characteristic curves evaluate all-cause cause-specific predictions. Overall, participants more diseases had Age. For instance, among young adults, those 1 disease 0.2 phenotypically disease-free persons, 2 3 about 0.6 phenotypically. After adjusting sex, significantly associated (with exception cerebrovascular mortality). Results robust stratifications by age, race/ethnicity, education, count, behaviors. Further, seemingly participants—defined as who reported being normal BMI—as well even adjustment prevalence. main limitation lack longitudinal incidence. Conclusions In nationally representative US adult population, age. association across different stratifications, particularly behaviors, cause death. strong between count individual had. These suggest new measure may serve useful tool facilitate identification at-risk individuals evaluation efficacy interventions, investigation into potential biological mechanisms aging. Nevertheless, further other cohorts is needed.