Table 1.

Measures of biological aging included in analysis. The table reports the 6 measures of biological aging included in the analysis. For each measure, the table reports the criterion used to develop the measure and the interpretation of the measure's values. Criterion refers to the quantity the biological aging algorithm was developed to predict. Interpretation refers to the inference about biological aging that can be made on the basis of the values of the measure.

Measure		Criterion	Interpretation
Blood-chemistry measures. All algorithms were parameterized using data from NHANES III and included the following blood chemistries: albumin, alkaline phosphatase, creatinine, C-reactive protein (log), white blood cell count, lymphocyte %, mean cell volume, and red cell distribution width. PhenoAge additionally included glucose. KDM Biological Age and Homeostatic Dysregulation additionally included glycated hemoglobin (HbA1C). PhenoAge and KDM Biological Age algorithms included information about chronological age. For analysis, PhenoAge and KDM Biological Age were differenced from chronological age to calculate biological-age advancement values.
	PhenoAge	Mortality	Age at which the participant's biomarker-predicted mortality risk matches the norm in the reference population (NHANES III).
	KDM Biological Age	Chronological Age	Age at which the participant's biomarker-predicted physiological integrity matches the norm in the reference population (NHANES III).
	Homeostatic Dysregulation	Deviation from healthy youth	Log biomarker-Mahalanobis distance of participant from young, healthy reference population (nonobese NHANES III participants aged 20–30 years).
DNAm measures. DNAm measures were developed from analysis of genome-wide DNAm measured on Illumina 27 k and 450 k arrays in a range of different datasets. The Horvath Clock was developed from the analysis of 82 different datasets. The Hannum Clock was developed from analysis of research volunteers at UC San Diego, University of Southern California, and West China Hospital. The PhenoAge Clock was developed from analysis of NHANES III and the InCHIANTI Study. The GrimAge clock was developed from analysis of the Framingham Heart Study Offspring Cohort. The DunedinPoAm Pace of Aging was developed from analysis of the Dunedin Study. DNAm measures were calculated by the HRS investigators. For analysis, DNAm clocks were residualized on chronological age to calculate biological-age advancement values.
Second generation DNAm clocks
	PhenoAge Clock	Blood-chemistry PhenoAge	DNAm prediction of the age at which the participant's biomarker-predicted mortality risk matches the norm in the NHANES III reference population (based on analysis of the InCHIANTI Study).
	GrimAge Clock	Mortality	Age at which the participants’ DNAm-predicted mortality risk matches the norm in the reference population (Framingham Heart Study Offspring cohort). The GrimAge clock was derived by first developing DNAm surrogates for blood proteins and smoking history, and then developing a mortality prediction model based on these DNAm surrogates, sex, and chronological age.
Pace of aging
	DunedinPoAm Pace of Aging	Change over 12 years of follow-up in 18 system-integrity biomarkers	Years of physiological decline experienced per 1 year of calendar time over the recent past. DunedinPoAm was developed by modeling a composite of change scores for 18 biomarkers of organ system integrity from DNAm data. The expected value of DunedinPoAm in midlife adults is 1. Values > 1 indicate accelerated aging. Values < 1 indicate slowed aging.