Biological Age: Difference between revisions

Revision as of 00:19, 6 February 2024

Biological Age (BA) is a concept used to assess an individual's aging status, offering a more nuanced understanding than Chronological Age (CA). Chronological age refers simply to the amount of time that has elapsed since a person's birth, while biological age provides a measure of aging based on various physiological, biochemical, and molecular factors. This distinction is crucial because individuals of the same chronological age can exhibit significantly different aging processes and health statuses.

Key Aspects of Biological Age

Biological age is estimated as the equivalent age within the same-sex population which corresponds to an individual's mortality risk.

Biomarkers: Biological age is typically determined by analyzing a range of biomarkers. These can include genetic markers, epigenetic alterations, cellular senescence, telomere length, metabolic markers, and more. The specific biomarkers chosen depend on the method of estimation and the focus of the study.
Health and Functionality: Biological age reflects the functional state of an individual's organs and systems. A lower biological age compared to chronological age might indicate better health and lower risk for age-related diseases, whereas a higher biological age suggests accelerated aging and potentially increased health risks.
Variability: Unlike chronological age, which is uniform and progresses at a constant rate (one year per year), biological age can vary significantly between individuals. Factors such as lifestyle, genetics, environment, and disease can influence the rate at which a person's biological systems age. For example, biological age estimated from blood markers ranged between 20-years younger and 20-years older than individuals' chronological age.^[1]

Importance in Research and Medicine

Research Tool: In scientific research, biological age is valuable for understanding the aging process, identifying aging biomarkers, and evaluating the effectiveness of anti-aging interventions.
Clinical Applications: In a clinical setting, biological age can be used to assess an individual's overall health status, predict the risk of age-related diseases, and personalize healthcare and treatment plans.

Biological Age Estimation Methods

Biological age estimation has emerged as a significant tool in gerontology, aiming to provide a more accurate measure of aging than chronological age.

The accurate estimation of biological age has significant implications for clinical practice, including predicting disease onset and prognosis, improving the quality of life for the elderly, and promoting successful aging. Each method offers unique insights, and a comprehensive understanding of these methods can lead to better clinical decision-making and more effective interventions for aging-related conditions.

Various methods have been developed to estimate biological age, each with its unique approach and criteria:^[2]

Multiple Linear Regression (MLR) is a statistical technique that estimates biological age by relating several independent variables (biomarkers) to a dependent variable (chronological age). In this method, chronological age is used as a criterion for selecting biomarkers and is treated as an independent index.
Principal Component Analysis (PCA) is another statistical technique used in biological age estimation. PCA reduces the dimensionality of the data by transforming multiple biomarkers into a set of linearly uncorrelated variables, known as principal components.
Hochschild’s Method differs from MLR and PCA by making chronological age an independent variable. It aims to estimate biological age by adjusting chronological age based on specific biomarkers.
Klemera and Doubal’s Method (KDM) shares a similar concept with Hochschild’s method but uses a more complex statistical approach. It treats chronological age as an independent variable and incorporates multiple biomarkers to estimate biological age.
Deep Learning-based approaches represent an emerging and powerful method for biological age estimation. Leveraging complex neural network architectures, these methods can analyze large datasets of biomarkers, integrating diverse types of biological data. Deep learning models can uncover intricate patterns in the data that may be missed by traditional statistical techniques, offering potentially more accurate and personalized biological age estimations. However, they require large, well-curated datasets to be effective.

Role of Chronological Age

The role of chronological age in the estimation of biological age varies significantly between different methods. In MLR and PCA, chronological age is used as a criterion for selecting biomarkers. This means that the biomarkers are chosen based on how well they correlate with chronological age. In this context, chronological age is not an independent variable in the statistical model but rather a reference point against which the predictive power of biomarkers is measured.

In contrast, Hochschild’s method and KDM treat chronological age as an independent variable. This means chronological age is directly incorporated into the model as one of the variables that predict biological age.

Comparison

Comparisons of various biological age estimation methods

Method	Proposed	Core concept	Advantage	Disadvantage
MLR	1965, Hollingsworth^[3]	Aging biomarkers are determined by the correlation with chronological age using MLR model	MLR is the preliminary method and is easy to operate	The standards of aging biomarkers lead to the paradox of chronological age MLR also distorts the biological age at the regression edge and ignores discontinuity in the aging rate^[4]^[5]^[6]
PCA	1988, Nakamura^[7]	PCA uses fewer uncorrelated variables to explain the main variance	Biomarkers are uncorrelated variables^[8] PCA avoids the influence of regression edge in MLR^[5]	PCA cannot avoid the paradox of chronological age and some statistical deficiencies of MLR^[8]
Hochschild’s method	1989, Hochschild^[9]	Hochschild’s method aims to select aging biomarkers according to their effects on life expectancy^[9]	Hochschild’s method solves the paradox of chronological age Hochschild’s method avoids statistical problems of MLR	Hochschild’s method is nonstandard and relatively complicated Hochschild’s method is not based on the definition of biological age A large number of subjects are required when this approach is adopted for another system^[10]
KDM	2006, Klemera and Doubal^[8]	KDM is based on minimizing the distance between m regression lines and m biomarker points in an m-dimensional space of all biomarkers^[8]	KDM performed better than chronological age^[11] KDM is precise when compared with other methods^[11]^[10]^[12] KDM solves the paradox of chronological age^[11]^[10]	The calculation of KDM is complicated^[10]
Deep learning	2015^[13]	Deep learning is a subfield of machine learning, where good features can be learned automatically using a general-purpose learning procedure^[13]. Deep neural networks (DNNs)^[14]^[15]^[16]^[17], convolutional neural networks (CNNs)^[18]^[19], and recurrent neural networks (RNNs)^[20] have been employed to build BA models in recent years.	Good at handling high-dimensional dataset^[13] The machine extracts features autonomously by learning^[13]	Difficulty in building large data^[14] The existence of a “black box” and uncontrollable results Excellent programming skills and computer hardware and software support required

Biomarkers for Biological Age Estimation

Organ system	PCA	MLR	Hochschild’s	KDM
Cardiovascular system	Pulse pressure^[21]^[22]^[23]
	Systolic blood pressure^[7]^[24]^[25]^[26]^[27]^[28]^[29]^[12]	Systolic blood pressure^[11]^[30]^[3]^[6]^[12]		Systolic blood pressure^[12]
	Heart rate^[7]^[24]
	Intima-media thickness^[21]^[31]^[22]
	Maximum internal diameter of carotid artery^[31]^[22]
	End diastolic velocity^[21]
	Mitral valve annulus ventricular septum of the peak velocity of early filling^[31]
	Mitral valve annulus lateral wall of peak velocity of early filling^[21]
	Mitral annulus peak E anterior wall^[22]
	Ratio of peak velocity of early filling to atrial filling^[21]
Respiratory system	VO2 max^[32]^[26]
	Forced expiratory volume in 1 second^[32]^[25]^[26]^[27]^[28]	Forced expiratory volume in 1 second^[11]^[30]^[12]	Forced expiratory volume in 1 second^[33]	Forced expiratory volume in 1 second^[12]
	Forced vital capacity^[7]^[24]^[29]	Forced vital capacity^[10]	Forced vital capacity^[33]
	Maximal mid expiratory flow rate 75/25^[31]	Vital capacity^[3]
Nervous system	Trail making test^[31]^[23]
	Digital symbol test^[10]^[22]	Digital symbol test^[10]
	Memory test linking names with faces^[10]	Memory test linking names with faces^[10]
	Memory test: which picture is at what place^[10]	Memory test: which picture is at what place^[10]
	Speed test: pointing icons from 1 to 15 sequentially, mixed in random positions^[10]
	Visual reaction time^[10]^[32]		Visual reaction time^[33]
			Sequence of lamps^[33]
			Alternate button tapping time with/without decision^[33]
			Movement time with/without decision^[33]
Renal system	Blood urea nitrogen^[7]^[24]^[25]^[26]^[34]^[28]^[29]	Blood urea nitrogen^[11]^[30]^[12]		Blood urea nitrogen^[12]
		Serum creatinine^[12]		Serum creatinine^[12]
	Cystatin C^[21]^[31]^[22]^[23]
Liver	Serum albumin^[25]^[26]^[34]^[28]	Serum albumin^[6]
	Glutamic oxaloacetic transaminase^[7]^[24]	Glutamic oxaloacetic transaminase^[12]		Glutamic oxaloacetic transaminase^[12]
	Glutamic pyruvic transaminase^[29]
	Ratio of albumin to globulin^[27]
	Lactate dehydrogenase^[7]^[24]^[29]
		Serum globulin^[6]
		Alkaline phosphatase^[6]
Hematologic system	Erythrocyte sedimentation rate^[26]	Erythrocyte sedimentation rate^[6]
	Mean corpuscular hemoglobin^[27]
	Red blood cell count^[28]
	Hematocrit^[25]
	Hemoglobin concentration^[7]^[24]
	Fibrinogen^[21]
		Ferratin^[12]		Ferratin^[12]
Metabolism	Glycosylated hemoglobin^[26]
	Glucose^[27]^[29]	Glucose^[6]
	Low-density cholesterol^[26]
	Atherogenic index^[7]^[24]
	Triglyceride^[24]	Triglycerides^[30]
	Total cholesterol^[29]	Total cholesterol^[11]^[30]^[3]^[12]		Total cholesterol^[12]
Muscle and fat	Grip strength^[10]^[32]	Grip strength^[10]^[3]
	Soft lean mass^[32]
	Waist circumference^[32]^[26]	Waist circumference^[12]		Waist circumference^[12]
	Percent body fat^[26]
Sensory system	Hearing threshold^[26]
	Highest audible pitch^[10]	Highest audible pitch^[10]	Highest audible pitch^[33]
		Light extinction test^[3]
		Visual acuity^[3]
		Auditory function^[3]
		Vibrotactile sensitivity^[3]	Vibrotactile sensitivity^[33]
		Auditory reaction time^[10]	Auditory reaction time^[33]
		Focal range test using a Landolt ring^[10]	Visual accommodation^[33]
Genetic index	Telomere restriction fragment^[22]

Biomarkers 2

The common aging biomarkers of four methods in different systems.

System		MLR	PCA	KDM	Deep learning
Cardiovascular system	Systolic Blood Pressure (SBP)	SBP 23213031 5841151 3226152 950448 6667707 17889950 23642770 28110151 30899733	SBP 23213031 3226152 2737197 2282902 8803500 12634284 12672981 17921421 18840798 18597867 22433233 28110151 30899733	SBP 23213031 26150497 28110151 28958059 28977464 30999227 30899733 31566204 32946548 34038024
	Diastolic Blood Pressure (DBP)	DBP 17889950 23642770 31179487	DBP 31179487	DBP 27216811 31179487
	Pulse Pressure	Pulse pressure 23642770	Pulse pressure 29188884 24522464 19940465
	Mean Arterial Pressure		Mean arterial pressure 28203066
	Pulse	Pulse 3226152	Pulse 3226152 2737197
	Pulse Wave Velocity	Pulse wave velocity 6667707
	Heart Rate		Heart rate 2282902
	Intima-Media Thickness		Intima-media thickness 24659482 19940465
	Minimum Intima-Media Thickness		Minimum intima-media thickness 29188884 24522464
	End Diastolic Velocity		End diastolic velocity 19940465
	Mitral Valve E/A Peak		mitral valve E/A peak 29188884 19940465
	MVEL, MVES, MVEA		MVEL 19940465 , MVES 24522464 , MVEA 24659482
	Atherosclerosis Index		Atherosclerosis index 2737197 2282902
	NT-proBNP				NT-proBNP 34453631
	Cardiac Troponin I				Cardiac troponin I 34453631
	Creatine Phosphokinase	Creatine phosphokinase 23642770
	Homocysteine	Homocysteine 23642770
Respiratory system	Forced Vital Capacity (FVC)	FVC 3226152 17889950 23642770 20005245	FVC 3226152 2737197 2282902 8803500 12634284	FVC 20005245
	Forced Expiratory Volume in 1 Second (FEV1)	FEV1 23213031 950448 7162237 17889950 23642770 28110151 30899733 31179487	FEV1 23213031 12634284 12672981 17921421 18840798 18597867 22433233 28110151 30899733 31179487	FEV1 23213031 26150497 28110151 28958059 28977464 30899733 31179487
	Vital Capacity	Vital capacity 5841151 6667707
	Maximal Midexpiratory Flow Rate 75/25		Maximal midexpiratory flow rate 75/25 24522464
	VO2 Max		VO2 max 18597867 22433233
	Chest Radiography				Chest radiography 33744131
Nervous system	Mini-Mental State Examination (MMSE)	MMSE 31179487	MMSE 31179487	MMSE 31179487
	Digital Symbol Test	Digital symbol test 6667707	Digital symbol test 24659482
	Numeric Memory	Numeric memory 20005245		Numeric memory 20005245
	Associated Memory	Associated memory 20005245		Associated memory 20005245
	Topological Memory	Topological memory 20005245		Topological memory 20005245
	Short-Time Memory	Short-time memory 6610563
	Concentration	Concentration 20005245		Concentration 20005245
	Intellectuality - Mental Defect	Intellectuality -mental defect 6667707
	Trail Making Test		Trail making test 29188884 24522464
Endocrine metabolic system	Glucose	Glucose 17889950 23642770	Glucose 8803500 12672981 28203066	Glucose 28977464	Glucose 27191382 29340580 30644411 34453631
	HBA1C	HBA1C 23213031 23642770	HBA1C 23213031 18597867	HBA1C 23213031 26150497 28958059 30999227 31566204 32946548 31693736 34038024	HBA1C 30644411
	C-peptide				C-peptide 34453631
	Insulin				Insulin 34453631
	Triglyceride	Triglyceride 950448 17889950 23642770	Triglyceride 2282902 28203066	Triglyceride 32946548	Triglyceride 29340580 34453631
	Total Cholesterol (TC)	TC 23213031 5841151 3226152 950448 7162237 17889950 23642770 28110151	TC 23213031 3226152 8803500 28110151	TC 23213031 26150497 28110151 28958059 28977464 31566204 32946548 34038024	TC 27191382 29340580
	High-Density Lipoprotein (HDL), Low-Density Lipoprotein (LDL)	HDL 17889950 , LDL 23642770	HDL 28203066 , LDL 18597867		HDL 29340580 34453631 ,LDL 29340580 34453631
	Apolipoprotein A1 and B				Apolipoprotein A1 and B 34453631
	Thyroid-Stimulating Hormone (TSH)	TSH 17889950		TSH 27216811
	Testosterone	Testosterone 17889950			Testosterone 34453631
	Vitamin D	Vitamin D 30899733	Vitamin D 30899733	Vitamin D 30899733	Vitamin D 34453631
	Calcium			Calcium 27216811	Calcium 29340580
	Potassium				Potassium 29340580
	Sodium				Sodium 29340580
	Inorganic Phosphorus			Inorganic phosphorus 27216811
Urinary system	Urea	Urea 23213031 3226152 950448 23642770 28110151	Urea 23213031 3226152 2737197 2282902 8803500 12634284 17921421 18840798 18597867 28110151	Urea 23213031 26150497 27216811 28110151 30999227 31566204 32946548 31693736	Urea 27191382 29340580 30644411
	Creatinine	Creatinine 23213031 17889950 23642770 28110151 30899733	Creatinine 28110151 30899733	Creatinine 23213031 26150497 27216811 28958059 30999227 31566204 32946548 31693736 34038024	Creatinine 29340580 34453631
	Estimated Glomerular Filtration Rate (eGFR)	eGFR 31179487	eGFR 31179487	eGFR 31179487
	Uric Acid			Uric acid 30999227 31693736	Uric acid 34453631
	Cystatin C		Cystatin C 29188884 24659482 24522464 19940465		Cystatin C 34453631
	Creatinine Clearance	Creatinine clearance 23642770
	Urine Specific Gravity	Urine specific gravity 23642770
	Urine pH	Urine pH 23642770
Digestive system	Alanine Aminotransferase (ALT)	ALT 23642770	ALT 8803500		ALT 34453631
	Aspartate Aminotransferase (AST)	AST 3226152 23642770 28110151	AST 3226152 2737197 2282902 28110151	AST 28110151	AST 34453631
	Alkaline Phosphatase (ALP)	ALP 23213031 950448 23642770	ALP 23213031	ALP 23213031 26150497 27216811 28958059 30999227 31693736 34038024	ALP 27191382
	Total Protein	Total protein 23642770		Total protein 27216811	Total protein 29340580
	Albumin	Albumin 23213031 3226152 23642770	Albumin 23213031 3226152 12634284 17921421 18840798 18597867	Albumin 23213031 26150497 27216811 28958059 30999227 31566204 32946548 31693736 34038024	Albumin 27191382 29340580 34453631
	Albumin/Globulin Ratio (A/G)	A/G 3226152 23642770	A/G 3226152 12634284 12672981
	Total Bilirubin	Total bilirubin 23642770			Total bilirubin 29340580
	Direct Bilirubin	Direct bilirubin 23642770
	Amylase	Amylase 23642770
	Lactate Dehydrogenase	Lactate dehydrogenase 7162237 23642770	Lactate dehydrogenase 2282902 8803500
	Alpha 2 Globulin				Alpha 2 globulin 27191382
	Gamma Glutamyl Transpeptidase	Gamma glutamyl transpeptidase 23642770
Hematologic System	Red Blood Cell	Red blood cell 30899733	Red blood cell 12634284 18840798 30899733	Red blood cell 30899733 32946548	Red blood cell 27191382 29340580 34453631
	Red Blood Cell Volume Distribution Width			Red blood cell volume distribution width 30999227	Red blood cell volume distribution width 27191382 34453631
	Hematocrit		Hematocrit 12634284 17921421		Hematocrit 27191382 29340580
	Mean Corpuscular Volume			Mean corpuscular volume 30999227 31566204 31693736	Mean corpuscular volume 29340580 34453631
	Mean Corpuscular Hemoglobin		Mean corpuscular hemoglobin 12672981
	Mean Corpuscular Hemoglobin Concentration				Mean corpuscular hemoglobin concentration 29340580 34453631
	Hemoglobin	Hemoglobin 3226152 31179487	Hemoglobin 3226152 2737197 2282902 8803500 12634284 31179487	Hemoglobin 27216811 31179487	Hemoglobin 29340580 34453631
	White Blood Cell			White blood cell 30999227 31566204 31693736	White blood cell 34453631
	Granulocytes				Granulocytes 34453631
	Neutrophils				Neutrophils 34453631
	Basophils, Eosinophils				Basophils 34453631 , Eosinophils 34453631
	Lymphocytes			Lymphocytes 30999227 31566204 31693736	Lymphocytes 27191382 34453631
	Monocytes	Monocytes 31179487	Monocytes 31179487	Monocytes 31179487	Monocytes 34453631
	Platelet			Platelet 32946548	Platelet 29340580 34453631
	Mean Platelet Volume				Mean platelet volume 34453631
	Platelet Distribution Width				Platelet distribution width 34453631
	Erythrocyte Sedimentation Rate	950448 17889950	18597867
	D-dimer, Fibrinogen		D-dimer 24659482 Fibrinogen 19940465		D-dimer 34453631
	Ferritin	Ferritin 28110151	Ferritin 28110151	Ferritin 28110151 32946548	Ferritin 30644411
	Transferrin			Fransferrin 32946548
Sensory system	Visual Accommodation	Visual accommodation 6667707 20005245		Visual accommodation 20005245
	Visual Reaction Time	Visual reaction time 20005245		Visual reaction time 20005245
	Visual Acuity	Visual acuity 5841151 3226152	Visual acuity 3226152
	Hearing	Hearing 5841151 7162237 6667707 20005245	Hearing 18597867	Hearing 20005245
	Vibrotactile	Vibrotactile 5841151 6610563 20005245		Vibrotactile 20005245
	Retinal Photos				Retinal photos
Inflammatory index	C-Reactive Protein (CRP)	CRP 23213031	CRP 23213031	CRP 23213031 26150497 28958059 28977464 30999227 31566204 32946548 31693736 34038024	CRP 34453631
	Cytomegalovirus Optical Density	Cytomegalovirus optical density 23213031	Cytomegalovirus optical density 23213031	Cytomegalovirus optical density 23213031 26150497 31566204
	Interleukin-6			Interleukin-6 28977464
	P-selectin			P-selectin 28977464
Motion index	Grip Strength	Grip strength 5841151 6610563 31179487 20005245	Grip strength 22433233 31179487	Grip strength 31179487 20005245
	Vertical Jump		Vertical jump 22433233
	Timed Up and Go Test	Timed up and go test 31179487	Timed up and go test 31179487	Timed up and go test 31179487
	Chair Rise Time	Chair rise time 31179487	Chair rise time 31179487	Chair rise time 31179487
	1-Week Physical Activity				1-week physical activity 29581467 31388024
Body morphology index	Waist Circumference (WC)	WC 23642770 28110151	WC 18597867 22433233 28110151 28203066	WC 28110151
	Waist-to-Hip Ratio	Waist-to-hip ratio 17889950 23642770
	Waist-to-Height Ratio	Waist-to-height ratio 30899733	Waist-to-height ratio 30899733	Waist-to-height ratio 30899733
	Body Mass Index (BMI)	Body mass index 17889950 23642770
	Weight	Weight 6667707
	Height	Height 31179487	Height 31179487	Height 31179487
	Body Fat	Body fat 17889950 23642770	Body fat 18597867
	Lean Body Mass	Lean body mass 17889950 23642770
	Soft Lean Mass		Soft lean mass 22433233
Genetic index	Terminal Telomere Restriction Fragment		Terminal telomere restriction fragment 24659482

SBP, systolic blood pressure; DBP, diastolic blood pressure; NT-proBNP, N-terminal pro brain natriuretic peptide; MVEA, mitral annulus peak E anterior wall; MVEL, mitral valve annulus lateral wall of peak velocity of early filling; MVES, mitral valve annulus ventricular septum of the peak velocity of early filling; FEV1, forced expiratory volume in 1.0 s; FVC, forced vital capacity; MMSE, mini-mental state examination; eGFR, estimated glomerular filtration rat; HBA1C, glycosylated hemoglobin; HDL, high density lipoprotein; LDL, low density lipoprotein; TSH, thyroid stimulating hormone; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; ALP, alkaline phosphatase; A/G, ratio of albumin to globulin; CRP, c-reactive protein; WC, waist circumference.

Todo

2017, Biological Age Predictors ^[36]
2022, Biological Age Predictors: The Status Quo and Future Trends ^[37]
2021, Predictors of Biological Age: The Implications for Wellness and Aging Research ^[38]
2023, Biological age estimation using circulating blood biomarkers ^[1]
2021, Deep learning for biological age estimation ^[39]

References

↑ ^{Jump up to: 1.0} ^1.1 Bortz J et al.: Biological age estimation using circulating blood biomarkers. Commun Biol 2023. (PMID 37884697) [PubMed] [DOI] [Full text] Abstract
Biological age captures physiological deterioration better than chronological age and is amenable to interventions. Blood-based biomarkers have been identified as suitable candidates for biological age estimation. This study aims to improve biological age estimation using machine learning models and a feature-set of 60 circulating biomarkers available from the UK Biobank (n = 306,116). We implement an Elastic-Net derived Cox model with 25 selected biomarkers to predict mortality risk (C-Index = 0.778; 95% CI [0.767-0.788]), which outperforms the well-known blood-biomarker based PhenoAge model (C-Index = 0.750; 95% CI [0.739-0.761]), providing a C-Index lift of 0.028 representing an 11% relative increase in predictive value. Importantly, we then show that using common clinical assay panels, with few biomarkers, alongside imputation and the model derived on the full set of biomarkers, does not substantially degrade predictive accuracy from the theoretical maximum achievable for the available biomarkers. Biological age is estimated as the equivalent age within the same-sex population which corresponds to an individual's mortality risk. Values ranged between 20-years younger and 20-years older than individuals' chronological age, exposing the magnitude of ageing signals contained in blood markers. Thus, we demonstrate a practical and cost-efficient method of estimating an improved measure of Biological Age, available to the general population.
↑ ^{Jump up to: 2.0} ^2.1 Jia L et al.: Common methods of biological age estimation. Clin Interv Aging 2017. (PMID 28546743) [PubMed] [DOI] [Full text] Abstract
At present, no single indicator could be used as a golden index to estimate aging process. The biological age (BA), which combines several important biomarkers with mathematical modeling, has been proposed for >50 years as an aging estimation method to replace chronological age (CA). The common methods used for BA estimation include the multiple linear regression (MLR), the principal component analysis (PCA), the Hochschild's method, and the Klemera and Doubal's method (KDM). The fundamental differences in these four methods are the roles of CA and the selection criteria of aging biomarkers. In MLR and PCA, CA is treated as the selection criterion and an independent index. The Hochschild's method and KDM share a similar concept, making CA an independent variable. Previous studies have either simply constructed the BA model by one or compared the four methods together. However, reviews have yet to illustrate and compare the four methods systematically. Since the BA model is a potential estimation of aging for clinical use, such as predicting onset and prognosis of diseases, improving the elderly's living qualities, and realizing successful aging, here we summarize previous BA studies, illustrate the basic statistical steps, and thoroughly discuss the comparisons among the four common BA estimation methods.
↑ ^{Jump up to: 3.0} ^3.1 ^3.2 ^3.3 ^3.4 ^3.5 ^3.6 ^3.7 ^3.8 Hollingsworth JW et al.: Correlations between tests of aging in Hiroshima subjects--an attempt to define "physiologic age". Yale J Biol Med 1965. (PMID 5841151) [PubMed] [Full text]
↑ Dubina TL et al.: Biological age and its estimation. II. Assessment of biological age of albino rats by multiple regression analysis. Exp Gerontol 1983. (PMID 6873212) [PubMed] [DOI] Abstract
Multiple regression model of biological age (BA) theoretically gives agreement with the main concept of BA. When assessment of BA is based on the model, the age being in regression center, the method provides satisfactory results, whereas BA estimates of individuals in extreme age groups are erroneous. Investigation of male and female Wistar rats of age 5-29 months showed the BA estimates calculated from 4-10 physiological indices in young (5-7 mo) animals are overestimated, and in old (24-28 mo) animals are underestimated. Coincidence of average BA in one-age group of animals with its chronological age served as a criterion for the correspondence of the estimate to "real" BA. The paper also examines the following questions: the necessary and sufficient number of physiological indices; the sample size from the intact animal population to establish normal aging standard; the relationship between BA and animal weight.
↑ ^{Jump up to: 5.0} ^5.1 Nakamura E et al.: Assessment of biological age by principal component analysis. Mech Ageing Dev 1988. (PMID 3226152) [PubMed] [DOI] Abstract
A method of assessing biological age by the application of principal component analysis is reported. Healthy individuals (462) randomly selected from about 6000 men who had taken a 2-day health examination were studied. Out of the 30 physiological variables examined in routine check-ups, 11 variables were selected as suitable for the assessment of biological age based on the results of factor analysis and the physiological meaning of each test. This variable set was then submitted to principal component analysis, and the 1st principal component obtained from this analysis was used as an equation for assessing one's biological age. However, the biological age calculated from this equation is expressed as a score, so the estimated score was transformed to years (biological age) using the T-score idea. The biological age estimated by this method is practically useful and theoretically valid in contrast with the multiple regression model, because this approach eliminates and overcomes the following 2 big problems of the multiple regression model: (1) the distortion of the individual biological age at the regression edges; and (2) a theoretical contradiction in that a perfect model will merely be predicting the subject's chronological age, not his biological age.
↑ ^{Jump up to: 6.0} ^6.1 ^6.2 ^6.3 ^6.4 ^6.5 ^6.6 Webster IW & Logie AR: A relationship between functional age and health status in female subjects. J Gerontol 1976. (PMID 950448) [PubMed] [DOI] Abstract
A multiple regression equation was used to predict age from seven clinical variables in 1080 apparently well female subjects. A multiple correlation coefficient of R = 0.77 was achieved by five of the variables: timed forced expiratory volume, systolic blood pressure, plasma urea nitrogen, cholesterol, and alkaline phosphatase. On the basis of selection by medical questionnaire responses and other objective criteria, 9% of the subjects were nonsmokers and healthier than the rest. These selected subjects showed a significant reduction in preducted age. Within this group, subjective perception of health was associated with differences in predicted age: poor health with an increase and good health with a decrease in functional age. This study of functional age was based on the healthiest segment of the population in order to minimize the effect of overt pathological processes on the aging rate. An association has been demonstrated between health impairment and predicted age as a measure of the aging rate.
↑ ^{Jump up to: 7.0} ^7.1 ^7.2 ^7.3 ^7.4 ^7.5 ^7.6 ^7.7 ^7.8 Nakamura E et al.: Biological age versus physical fitness age. Eur J Appl Physiol Occup Physiol 1989. (PMID 2737197) [PubMed] [DOI] Abstract
A population of healthy middle-aged (n = 69) and elderly men (n = 12), who participated in a health promotion program, was studied to determine whether really physically fit individuals are in good biological condition, and also whether improvement of physical fitness in the middle-aged and the elderly reduces their "rate of aging". Biological and physical fitness ages of the individuals studied were estimated from the data for 18 physiological function tests and 5 physical fitness tests, respectively, by a principal component model. The correlation coefficient between the estimated biological and physical fitness ages was 0.72 (p less than 0.01). Detailed analyses of the relationship between the estimated biological and physical fitness ages revealed that those who manifested a higher ("older") physical fitness age did not necessarily have a higher biological age, but those who manifested a lower ("younger") physical fitness age were also found to have a lower biological age. These results suggested that there were considerable individual variations in the relationship between biological condition and physical fitness among individuals with an old physical fitness age, but those who were in a state of high physical fitness maintained a relatively good biological condition. The data regarding the elderly men who had maintained a regular exercise program indicated that their estimated biological ages were considerably younger than the expected values. This might suggest that in older individuals regular physical activity may provide physiological improvements which in turn might reduce "the rate of aging".
↑ ^{Jump up to: 8.0} ^8.1 ^8.2 ^8.3 Klemera P & Doubal S: A new approach to the concept and computation of biological age. Mech Ageing Dev 2006. (PMID 16318865) [PubMed] [DOI] Abstract
The lack of exact definition of the concept of biological age (BA) is a typical feature of works concerning BA. That is why comparison of results of various published methods makes little sense and eventual proof of their optimality is impossible. Based on natural and simple presumptions, an attempt to express mathematically the supposed relation between chronological age (CA) and BA has proven to be unexpectedly fruitful. In the present paper, an optimum method of estimation of BA, which is easily applicable even in nonlinear cases, is derived. Moreover, the method allows evaluating the precision of the estimates and also offers tools for validation of presumptions of the method. A special feature of the method is that CA should be used as a standard biomarker, leading to essential improving the precision of BA-estimate and illuminating relativity of the known "paradox of biomarkers". All theoretical results of the method were fully approved by means of a special simulation program. Further, the theory and the results of the simulation have proven that many published results of BA-estimates using multiple linear regression (MLR) are very probably disserviceable because CA is typically more precise estimate of BA than estimates computed by MLR. This unpleasant conclusion also concerns methods, which use MLR as the final step after transformation of the battery of biomarkers by factor analysis or by principal component analysis.
↑ ^{Jump up to: 9.0} ^9.1 Hochschild R: Improving the precision of biological age determinations. Part 1: A new approach to calculating biological age. Exp Gerontol 1989. (PMID 2684676) [PubMed] [DOI] Abstract
In calculating biological age, almost all prior studies used multiple regression of chronological age on scores of biomarkers of aging. Multiple regression is invalid for this purpose for three, and in some circumstances four, reasons. These are: a) weighting of the contribution of each biomarker's scores according to strength of association with chronological age; b) regression of calculated ages to sample mean age and the inadequacy of proposed corrections; c) frequent occurrence of regression coefficients whose sign equates poorer adult performance on a test to younger biological ages; and d) multicollinearity when lung function scores and height are on the same side of the regression equation. An alternative method for calculating biological age is outlined. Regression to sample mean age and its solution are illustrated on data for highest audible pitch, one of 12 biomarkers measured in a study of 2462 office workers. Prior published studies employing multiple regression to calculate biological age appear to have been in error.
↑ ^{Jump up to: 10.00} ^10.01 ^10.02 ^10.03 ^10.04 ^10.05 ^10.06 ^10.07 ^10.08 ^10.09 ^10.10 ^10.11 ^10.12 ^10.13 ^10.14 ^10.15 ^10.16 ^10.17 ^10.18 Cho IH et al.: An empirical comparative study on biological age estimation algorithms with an application of Work Ability Index (WAI). Mech Ageing Dev 2010. (PMID 20005245) [PubMed] [DOI] Abstract
In this study, we described the characteristics of five different biological age (BA) estimation algorithms, including (i) multiple linear regression, (ii) principal component analysis, and somewhat unique methods developed by (iii) Hochschild, (iv) Klemera and Doubal, and (v) a variant of Klemera and Doubal's method. The objective of this study is to find the most appropriate method of BA estimation by examining the association between Work Ability Index (WAI) and the differences of each algorithm's estimates from chronological age (CA). The WAI was found to be a measure that reflects an individual's current health status rather than the deterioration caused by a serious dependency with the age. Experiments were conducted on 200 Korean male participants using a BA estimation system developed principally under the concept of non-invasive, simple to operate and human function-based. Using the empirical data, BA estimation as well as various analyses including correlation analysis and discriminant function analysis was performed. As a result, it had been confirmed by the empirical data that Klemera and Doubal's method with uncorrelated variables from principal component analysis produces relatively reliable and acceptable BA estimates.
↑ ^{Jump up to: 11.0} ^11.1 ^11.2 ^11.3 ^11.4 ^11.5 ^11.6 Levine ME: Modeling the rate of senescence: can estimated biological age predict mortality more accurately than chronological age?. J Gerontol A Biol Sci Med Sci 2013. (PMID 23213031) [PubMed] [DOI] [Full text] Abstract
Biological age (BA) is useful for examining differences in aging rates. Nevertheless, little consensus exists regarding optimal methods for calculating BA. The aim of this study is to compare the predictive ability of five BA algorithms. The sample included 9,389 persons, aged 30-75 years, from National Health and Nutrition Examination Survey III. During the 18-year follow-up, 1,843 deaths were counted. Each BA algorithm was compared with chronological age on the basis of predictive sensitivity and strength of association with mortality. Results found that the Klemera and Doubal method was the most reliable predictor of mortality and performed significantly better than chronological age. Furthermore, when included with chronological age in a model, Klemera and Doubal method had more robust predictive ability and caused chronological age to no longer be significantly associated with mortality. Given the potential of BA to highlight heterogeneity, the Klemera and Doubal method algorithm may be useful for studying a number of questions regarding the biology of aging.
↑ ^{Jump up to: 12.00} ^12.01 ^12.02 ^12.03 ^12.04 ^12.05 ^12.06 ^12.07 ^12.08 ^12.09 ^12.10 ^12.11 ^12.12 ^12.13 ^12.14 ^12.15 ^12.16 ^12.17 Jee H & Park J: Selection of an optimal set of biomarkers and comparative analyses of biological age estimation models in Korean females. Arch Gerontol Geriatr 2017. (PMID 28110151) [PubMed] [DOI] Abstract
To date, an optimal working model which predicts biological age (BA) with a set of working biomarkers has not been devised to represent the Korean female population. Accuracy of prediction and applicability are required of an optimal set of commonly assessed biomarkers to provide information on the health status. The goal of this study was to identify a set of biomarkers that represent the aging process and to develop and compare different BA prediction models to elucidate the most fitting and applicable model for providing information on health status in the Korean female population. Using a series of selection processes, eight clinically assessable variables were selected by analyzing relations between 31 clinical variables and chronologic age in 912 normal, healthy individuals among 3642 female participants with ages ranging from 30 to 80 years. The multiple linear regression (MLR), principal component analysis (PCA), and the Klemera-Doubal (KDM) statistical methods were applied to obtain three different sets of BA prediction models. These three models were assessed by calculating and performing the coefficient determinations (r2), regression slopes, effect sizes, pairwise t-tests, and Bland-Altman plots. The BA models were further compared for the applicability by calculating the BAs of clinical risk groups. MLR showed the narrowing effects at the either ends of the age spectrum with greatest effect sizes. PCA showed the greatest degree of dispersion and deviation from the regression center. These MLR and PCA trends were also exhibited by clinically risk groups. In conclusion, the KDM BA prediction model based on the selected biomarkers was found to provide the most reliable and stable results for the practical assessment of BA.
↑ ^{Jump up to: 13.0} ^13.1 ^13.2 ^13.3 LeCun Y et al.: Deep learning. Nature 2015. (PMID 26017442) [PubMed] [DOI] Abstract
Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.
↑ ^{Jump up to: 14.0} ^14.1 Putin E et al.: Deep biomarkers of human aging: Application of deep neural networks to biomarker development. Aging (Albany NY) 2016. (PMID 27191382) [PubMed] [DOI] [Full text] Abstract
One of the major impediments in human aging research is the absence of a comprehensive and actionable set of biomarkers that may be targeted and measured to track the effectiveness of therapeutic interventions. In this study, we designed a modular ensemble of 21 deep neural networks (DNNs) of varying depth, structure and optimization to predict human chronological age using a basic blood test. To train the DNNs, we used over 60,000 samples from common blood biochemistry and cell count tests from routine health exams performed by a single laboratory and linked to chronological age and sex. The best performing DNN in the ensemble demonstrated 81.5 % epsilon-accuracy r = 0.90 with R(2) = 0.80 and MAE = 6.07 years in predicting chronological age within a 10 year frame, while the entire ensemble achieved 83.5% epsilon-accuracy r = 0.91 with R(2) = 0.82 and MAE = 5.55 years. The ensemble also identified the 5 most important markers for predicting human chronological age: albumin, glucose, alkaline phosphatase, urea and erythrocytes. To allow for public testing and evaluate real-life performance of the predictor, we developed an online system available at http://www.aging.ai. The ensemble approach may facilitate integration of multi-modal data linked to chronological age and sex that may lead to simple, minimally invasive, and affordable methods of tracking integrated biomarkers of aging in humans and performing cross-species feature importance analysis.
↑ Mamoshina P et al.: Population Specific Biomarkers of Human Aging: A Big Data Study Using South Korean, Canadian, and Eastern European Patient Populations. J Gerontol A Biol Sci Med Sci 2018. (PMID 29340580) [PubMed] [DOI] [Full text] Abstract
Accurate and physiologically meaningful biomarkers for human aging are key to assessing antiaging therapies. Given ethnic differences in health, diet, lifestyle, behavior, environmental exposures, and even average rate of biological aging, it stands to reason that aging clocks trained on datasets obtained from specific ethnic populations are more likely to account for these potential confounding factors, resulting in an enhanced capacity to predict chronological age and quantify biological age. Here, we present a deep learning-based hematological aging clock modeled using the large combined dataset of Canadian, South Korean, and Eastern European population blood samples that show increased predictive accuracy in individual populations compared to population specific hematologic aging clocks. The performance of models was also evaluated on publicly available samples of the American population from the National Health and Nutrition Examination Survey (NHANES). In addition, we explored the association between age predicted by both population specific and combined hematological clocks and all-cause mortality. Overall, this study suggests (a) the population specificity of aging patterns and (b) hematologic clocks predicts all-cause mortality. The proposed models were added to the freely-available Aging.AI system expanding the range of tools for analysis of human aging.
↑ Mamoshina P et al.: Blood Biochemistry Analysis to Detect Smoking Status and Quantify Accelerated Aging in Smokers. Sci Rep 2019. (PMID 30644411) [PubMed] [DOI] [Full text] Abstract
There is an association between smoking and cancer, cardiovascular disease and all-cause mortality. However, currently, there are no affordable and informative tests for assessing the effects of smoking on the rate of biological aging. In this study we demonstrate for the first time that smoking status can be predicted using blood biochemistry and cell count results andthe recent advances in artificial intelligence (AI). By employing age-prediction models developed using supervised deep learning techniques, we found that smokers exhibited higher aging rates than nonsmokers, regardless of their cholesterol ratios and fasting glucose levels. We further used those models to quantify the acceleration of biological aging due to tobacco use. Female smokers were predicted to be twice as old as their chronological age compared to nonsmokers, whereas male smokers were predicted to be one and a half times as old as their chronological age compared to nonsmokers. Our findings suggest that deep learning analysis of routine blood tests could complement or even replace the current error-prone method of self-reporting of smoking status and could be expanded to assess the effect of other lifestyle and environmental factors on aging.
↑ Gialluisi A et al.: Exploring domains, clinical implications and environmental associations of a deep learning marker of biological ageing. Eur J Epidemiol 2022. (PMID 34453631) [PubMed] [DOI] Abstract
Deep Neural Networks (DNN) have been recently developed for the estimation of Biological Age (BA), the hypothetical underlying age of an organism, which can differ from its chronological age (CA). Although promising, these population-specific algorithms warrant further characterization and validation, since their biological, clinical and environmental correlates remain largely unexplored. Here, an accurate DNN was trained to compute BA based on 36 circulating biomarkers in an Italian population (N = 23,858; age ≥ 35 years; 51.7% women). This estimate was heavily influenced by markers of metabolic, heart, kidney and liver function. The resulting Δage (BA-CA) significantly predicted mortality and hospitalization risk for all and specific causes. Slowed biological aging (Δage < 0) was associated with higher physical and mental wellbeing, healthy lifestyles (e.g. adherence to Mediterranean diet) and higher socioeconomic status (educational attainment, household income and occupational status), while accelerated aging (Δage > 0) was associated with smoking and obesity. Together, lifestyles and socioeconomic variables explained ~48% of the total variance in Δage, potentially suggesting the existence of a genetic basis. These findings validate blood-based biological aging as a marker of public health in adult Italians and provide a robust body of knowledge on its biological architecture, clinical implications and potential environmental influences.
↑ Pyrkov TV et al.: Extracting biological age from biomedical data via deep learning: too much of a good thing?. Sci Rep 2018. (PMID 29581467) [PubMed] [DOI] [Full text] Abstract
Age-related physiological changes in humans are linearly associated with age. Naturally, linear combinations of physiological measures trained to estimate chronological age have recently emerged as a practical way to quantify aging in the form of biological age. In this work, we used one-week long physical activity records from a 2003-2006 National Health and Nutrition Examination Survey (NHANES) to compare three increasingly accurate biological age models: the unsupervised Principal Components Analysis (PCA) score, a multivariate linear regression, and a state-of-the-art deep convolutional neural network (CNN). We found that the supervised approaches produce better chronological age estimations at the expense of a loss of the association between the aging acceleration and all-cause mortality. Consequently, we turned to the NHANES death register directly and introduced a novel way to train parametric proportional hazards models suitable for out-of-the-box implementation with any modern machine learning software. As a demonstration, we produced a separate deep CNN for mortality risks prediction that outperformed any of the biological age or a simple linear proportional hazards model. Altogether, our findings demonstrate the emerging potential of combined wearable sensors and deep learning technologies for applications involving continuous health risk monitoring and real-time feedback to patients and care providers.
↑ Raghu VK et al.: Deep Learning to Estimate Biological Age From Chest Radiographs. JACC Cardiovasc Imaging 2021. (PMID 33744131) [PubMed] [DOI] Abstract
OBJECTIVES: The goal of this study was to assess whether a deep learning estimate of age from a chest radiograph image (CXR-Age) can predict longevity beyond chronological age. BACKGROUND: Chronological age is an imperfect measure of longevity. Biological age, a measure of overall health, may improve personalized care. This paper proposes a new way to estimate biological age using a convolutional neural network that takes as input a CXR image and outputs a chest x-ray age (in years) as a measure of long-term mortality risk. METHODS: CXR-Age was developed using CXR from 116,035 individuals and validated in 2 held-out testing sets: 1) 75% of the CXR arm of PLCO (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial) (N = 40,967); and 2) the CXR arm of NLST (National Lung Screening Trial) (N = 5,414). CXR-Age was compared to chronological age and a multivariable regression model of chronological age, risk factors, and radiograph findings to predict all-cause and cardiovascular mortality with a maximum 23 years and 13 years of follow-up, respectively. The primary outcome was observed mortality; results are provided for the testing datasets only. RESULTS: In the PLCO testing dataset, a 5-year increase in CXR-Age carried a higher risk of all-cause mortality than a 5-year increase in chronological age (CXR-Age hazard ratio [HR]: 2.26 [95% confidence interval (CI): 2.24 to 2.29] vs. chronological age HR: 1.77 [95% CI: 1.75 to 1.78]; p < 0.001). A similar pattern was found for cardiovascular mortality (CXR-Age cause-specific HR: 2.45 per 5 years [95% CI: 2.34 to 2.56] vs. chronological age HR: 1.82 per 5 years [95% CI: 1.74 to 1.90]). Similar results were seen for both outcomes in the NLST external testing dataset. Adding CXR-Age to the multivariable model resulted in significant improvements for predicting both outcomes in both testing datasets (p < 0.001 for all comparisons). CONCLUSIONS: Based on a CXR image, CXR-Age predicted long-term all-cause and cardiovascular mortality.
↑ Rahman SA & Adjeroh DA: Deep Learning using Convolutional LSTM estimates Biological Age from Physical Activity. Sci Rep 2019. (PMID 31388024) [PubMed] [DOI] [Full text] Abstract
Human age estimation is an important and difficult challenge. Different biomarkers and numerous approaches have been studied for biological age estimation, each with its advantages and limitations. In this work, we investigate whether physical activity can be exploited for biological age estimation for adult humans. We introduce an approach based on deep convolutional long short term memory (ConvLSTM) to predict biological age, using human physical activity as recorded by a wearable device. We also demonstrate five deep biological age estimation models including the proposed approach and compare their performance on the NHANES physical activity dataset. Results on mortality hazard analysis using both the Cox proportional hazard model and Kaplan-Meier curves each show that the proposed method for estimating biological age outperforms other state-of-the-art approaches. This work has significant implications in combining wearable sensors and deep learning techniques for improved health monitoring, for instance, in a mobile health environment. Mobile health (mHealth) applications provide patients, caregivers, and administrators continuous information about a patient, even outside the hospital.
↑ ^{Jump up to: 21.0} ^21.1 ^21.2 ^21.3 ^21.4 ^21.5 ^21.6 Bai X et al.: Evaluation of biological aging process - a population-based study of healthy people in China. Gerontology 2010. (PMID 19940465) [PubMed] [DOI] Abstract
BACKGROUND: Although there have been cross-sectional and longitudinal studies examining biological age (BA) with chronological age (CA)-related changes in physical, physiological, biochemical, and hormonal variables, few studies have performed echocardiographic evaluation of the cardiovascular system and inflammatory biomarkers. Furthermore, little is known about biomarkers of aging and BA score (BAS) for healthy people in China. OBJECTIVES: The purpose of this study was to identify the biomarkers of healthy aging and to establish BAS for healthy people in China. METHODS: We examined 2,876 men and women aged 30-98 years old in three Chinese cities, and 852 healthy subjects were assessed with 108 physical, morphological, physiological and biochemical variables. After excluding binary variables, variables that had a correlation coefficient with CA of < or =0.25 and redundant variables, eight variables including CA, arterial pulse pressure (PP), intima-media thickness (IMT), end diastolic velocity (EDV), ratio of peak velocity of early filling to atrial filling (E/A), mitral valve annulus lateral wall of peak velocity of early filling (MVEL), cystatin C (CYSC), and fibrinogen (FIB) were selected as candidate biomarkers of aging based on a factor-weighted BAS composite for predicting BA. RESULTS: The BAS equation was 0.248 (CA) + 0.195 (IMT) - 0.196 (EDV) - 0.167 (E/A) - 0.166 (MVEL) + 0.188 (PP) + 0.182 (FIB) + 0.193 (CYSC). Individual BAS were significantly correlated with CA (r = 0.893, p < 0.001). Biological aging rate predicted by BAS was accelerated with increases in CA, and peaked when healthy men and women reached > or =75 years of age. CONCLUSIONS: Our data suggest that BAS is superior to CA in assessing the rate of aging in healthy Chinese people. The cardiovascular variables play a crucial role in the evaluation of biological aging. Biological aging rate appears to be age specific.
↑ ^{Jump up to: 22.0} ^22.1 ^22.2 ^22.3 ^22.4 ^22.5 ^22.6 Zhang WG et al.: Select aging biomarkers based on telomere length and chronological age to build a biological age equation. Age (Dordr) 2014. (PMID 24659482) [PubMed] [DOI] [Full text] Abstract
The purpose of this study is to build a biological age (BA) equation combining telomere length with chronological age (CA) and associated aging biomarkers. In total, 139 healthy volunteers were recruited from a Chinese Han cohort in Beijing. A genetic index, renal function indices, cardiovascular function indices, brain function indices, and oxidative stress and inflammation indices (C-reactive protein [CRP]) were measured and analyzed. A BA equation was proposed based on selected parameters, with terminal telomere restriction fragment (TRF) and CA as the two principal components. The selected aging markers included mitral annulus peak E anterior wall (MVEA), intima-media thickness (IMT), cystatin C (CYSC), D-dimer (DD), and digital symbol test (DST). The BA equation was: BA = −2.281TRF + 26.321CYSC + 0.025DD − 104.419MVEA + 34.863IMT − 0.265DST + 0.305CA + 26.346. To conclude, telomere length and CA as double benchmarks may be a new method to build a BA.
↑ ^{Jump up to: 23.0} ^23.1 ^23.2 Zhang W et al.: Model Construction for Biological Age Based on a Cross-Sectional Study of a Healthy Chinese Han population. J Nutr Health Aging 2017. (PMID 29188884) [PubMed] [DOI] Abstract
OBJECTIVES: Biological age (BA) has been proposed to evaluate the aging status in an objective way instead of chronological age (CA). The purpose of our study is to construct a more precise formula of BA in the cross-sectional study based on a largest-ever sample of our studies. This formula aims at better evaluation of body function and exploring the disciplines of aging in different genders and age stages. METHODS: A total of 1,373 healthy Chinese Han (age range, 19-93 years) were recruited from five cities in China, including 581 males and 792 females. Physical examination, blood routine, blood chemistry, and other lab tests were performed to obtain a total of 74 clinical variables. Then, the principal component analysis (PCA) was used to select variables and estimate BA. The BA formula was further validated in a population with some diseases (n=266), including cardiovascular diseases, type 2 diabetes, kidney diseases, pulmonary diseases, cancer and disorders in nervous system. RESULTS: The BA formula was constructed as follows: BA = 0.358 (pulse pressure) + 0.258 (trail making test) - 11.552 (mitral valve E/A peak) + 26.383 (minimum intima-media thickness) + 31.965 (Cystatin C) + 0.163 (CA) - 3.902. In validation of the formula, BAs of patients were older than those of healthy persons. The BA accelerates faster in the middle-aged population than in the elderly population (>75 years old). CONCLUSION: This BA formula can reflect health condition changes of aging better than CA in a Chinese Han population.
↑ ^{Jump up to: 24.0} ^24.1 ^24.2 ^24.3 ^24.4 ^24.5 ^24.6 ^24.7 ^24.8 Nakamura E et al.: Biological age versus physical fitness age in women. Eur J Appl Physiol Occup Physiol 1990. (PMID 2282902) [PubMed] [DOI] Abstract
The purpose of this study was to determine whether adult women who are in a state of high physical fitness are in a good state biologically, in terms of biological and physical fitness ages as estimated by statistical means. The subjects were 65 healthy Japanese women (aged 20-64 years). Biological and physical fitness ages were estimated from the data for 18 physiological function tests and 5 physical fitness tests, respectively, by a principal component model. The correlation coefficient between biological and physical fitness ages was 0.70 (P less than 0.01), which was generally regarded as a high correlation. Therefore, those who were in a state of high physical fitness were considered to be in good biological condition. This result is in good agreement with the results (r = 0.72) from adult men, on whom we reported previously. A statistical analysis to ascertain the relative importance of each contributory variable associated with the variance in biological age suggested that routine clinical evaluation of blood pressure and lipid metabolism might play an important role in determining not only the presence and severity of vascular disease but also the rate of biological aging in women.
↑ ^{Jump up to: 25.0} ^25.1 ^25.2 ^25.3 ^25.4 Nakamura E & Miyao K: A method for identifying biomarkers of aging and constructing an index of biological age in humans. J Gerontol A Biol Sci Med Sci 2007. (PMID 17921421) [PubMed] [DOI] Abstract
This study was conducted to identify biomarkers of aging and to construct an index of biological age in humans. Healthy adult men (n = 86) who had received an annual health examination from 1992 through 1998 were studied. From 29 physiological variables, five variables (forced expiratory volume in 1 second, systolic blood pressure, hematocrit, albumin, blood urea nitrogen) were selected as candidate biomarkers of aging. Five candidate biomarkers expressed substantial covariance along one principal component. The first principal component obtained from a principal component analysis was used to calculate biological age scores (BAS). Individual BAS showed high longitudinal stability of age-related changes. Age-related changes of BAS are characterized by three components: age, peak functional capacity, and aging rate. A logistic regression analysis suggested that aging rate was influenced by environmental factors, but peak functional capacity was almost independent of environmental factors.
↑ ^{Jump up to: 26.00} ^26.01 ^26.02 ^26.03 ^26.04 ^26.05 ^26.06 ^26.07 ^26.08 ^26.09 ^26.10 Park J et al.: Developing a biological age assessment equation using principal component analysis and clinical biomarkers of aging in Korean men. Arch Gerontol Geriatr 2009. (PMID 18597867) [PubMed] [DOI] Abstract
The purpose of the present study is to find clinically useful candidate biomarkers of aging, and using these to develop an equation measuring biological age (BA) in Korean men, then to validate the clinical usefulness of it. Among 4288 men who received medical health examinations, we selected 1588 men who met the normality criteria of each variable. We assumed that chronological ages (CA) of healthy persons represent the BA of them. Variables showing significant correlations with CA were selected. Redundant variables were excluded. We selected 11 variables: VO(2)max, percent body fat (%BF), waist circumference (WC), forced expiratory volume in 1 s (FEV1), systolic blood pressure (SBP), low density cholesterol (LDLCH), blood urea nitrogen (BUN), serum albumin (SA), erythrocyte sedimentation rate(ESR) hearing threshold (HT), and glycosylated hemoglobin (HBA1C). These 11 variables were then submitted into principal component analysis (PCA) and standardized BA scores were obtained. Using them and T-scale idea, the following equation to assess BA was developed: BA=-28.7+0.83(%BF)+0.48(WC)+0.13(SBP)-0.27(VO(2)max)+0.19(HT)-3.1(FEV1)+0.32(BUN)+0.06(LDLCH)-3.0(SA)+0.34(ESR)+4.6(HBA1C). We compared the BA of 3122 men by their fasting glucose and age level. The BA of the higher glucose level group was significantly higher than that of others at all CA levels. The selected 11 biomarkers encompassed known clinically important factors of adult diseases and functional disabilities. This BA assessment equation can be used in the general Korean male population and it proved to be clinically useful.
↑ ^{Jump up to: 27.0} ^27.1 ^27.2 ^27.3 ^27.4 Ueno LM et al.: Biomarkers of aging in women and the rate of longitudinal changes. J Physiol Anthropol Appl Human Sci 2003. (PMID 12672981) [PubMed] [DOI] Abstract
The purposes of this study were (1) to estimate biological age score (BAS) in Japanese healthy women based on the 4-7 years longitudinal data for physiological, hematological and biochemical examinations and (2) to examine the rate of aging changes in adult women based on the estimated BAS. The samples consisted of cross-sectional (n=981) and longitudinal (n=110) groups. Out of 31 variables examined, five variables (forced expiratory volume in 1.0 s, systolic blood pressure, mean corpuscular hemoglobin, glucose, albumin/globulin ratio) that met the following criteria: 1) significant cross-sectional correlation with age; 2) significant longitudinal change in the same direction as the cross-sectional correlation; and (3) assessment of redundancy, were selected as candidate biomarkers of aging. This variable set was then submitted into a principal component analysis, and the first principal component obtained from this analysis was used as an equation for assessing one's BAS. Individual BAS showed a high longitudinal stability of age-related changes, suggesting high predictive validity of our newly developed aging measurement equation. However, changes in the aging rate based on the estimated BAS were not constant. The mean slopes of the regression lines of BAS for the three age groups (age<45, 45</=age<65 yrs, 65</=age) were 0.095, 0.065, 0.138, respectively. One-way analysis of variance detected a significant difference (F=5.14, p<0.01) among the three age groups. These results suggest that the rate of aging in adult women is relatively slower until 65 years of age, but after 65, the rate of aging shows a rapid increase. We concluded that the longitudinal method used for selection of variables to compute the BAS was useful and theoretically valid compared to those obtained from cross-sectional data analysis.
↑ ^{Jump up to: 28.0} ^28.1 ^28.2 ^28.3 ^28.4 Nakamura E & Miyao K: Sex differences in human biological aging. J Gerontol A Biol Sci Med Sci 2008. (PMID 18840798) [PubMed] [DOI] Abstract
This study aims to clarify sex differences in human biological aging and to explore the gender gaps in health and longevity. Eighty-six men and 93 women who received a 2-day routine health checkup for 6-7 years beginning in 1992 at the Kyoto Red Cross Hospital were selected. Five candidate biomarkers of aging (forced expiratory volume in 1.0 second per square of height [FEV(1)/Ht(2)], systolic blood pressure [SBP], red blood cells [RBC], albumin [ALBU], and blood urea nitrogen [BUN]) were selected from 29 physiological variables. Individual biological ages (BAS) were estimated from these five biomarkers by a principal component model. From the investigation of the longitudinal changes of individual BAS, it was suggested that (i) beyond 65 years, the rate of aging showed a rapid increase, and (ii) women had relatively lower functional capabilities compared with men, but the rate of aging was slower than that of men, suggesting that these differences might present both disadvantages and advantages for women with regard to health and longevity.
↑ ^{Jump up to: 29.0} ^29.1 ^29.2 ^29.3 ^29.4 ^29.5 ^29.6 Nakamura E et al.: Effects of habitual physical exercise on physiological age in men aged 20-85 years as estimated using principal component analysis. Eur J Appl Physiol Occup Physiol 1996. (PMID 8803500) [PubMed] [DOI] Abstract
A population of 221 healthy adult men (aged 20-85 years) was studied to determine whether those who exercised regularly were in good biological condition, and also whether those who were in a state of high physical fitness were in a good state biologically, in terms of physiological age (PA) and physical fitness age (FA) as estimated by principal component analysis. A group of 17 physiological function tests and 5 physical fitness tests were employed to estimate PA and FA, respectively. The results of this study indicated that those who maintained high physical fitness at all age decade groups from 20 to 79 years had a trend towards maintaining a relatively lower PA (physiologically younger). Mean PA and FA of the trained group were younger by 4.7 and 7.3 years, respectively than those of the untrained group. In addition, the slope of regression line of PA on chronological age was more gentle in the trained group than that in the untrained group. These results would suggest that those who are in a state of high physical fitness maintain a relatively good physiological condition, and that regular physical exercise may delay physiological changes normally seen with aging, and consequently may increase the life span.
↑ ^{Jump up to: 30.0} ^30.1 ^30.2 ^30.3 ^30.4 Borkan GA & Norris AH: Assessment of biological age using a profile of physical parameters. J Gerontol 1980. (PMID 6967883) [PubMed] [DOI] Abstract
The present study describes a new approach to the assessment of biological age in adults using a profile of physical parameters. The investigation was based on data from 1086 adult male participants in the aging study of the Gerontology Research Ctr., Baltimore, MD. For each of 24 age-related variables, data were transformed into biological age scores reflecting a man's status relative to his chronological age peers. Analysis of the mean biological age profiles of men who were estimated by physicians as looking older than their age showed them to be biologically older on the profile parameters as well. Comparison of age-corrected scores of 166 men who have died with those of survivors reveals the deceased group to have been biologically older than the survivors at the time they were measured. These results suggest the value of this technique in investigating interindividual variation in the aging process.
↑ ^{Jump up to: 31.0} ^31.1 ^31.2 ^31.3 ^31.4 ^31.5 Zhang WG et al.: Construction of an integral formula of biological age for a healthy Chinese population using principle component analysis. J Nutr Health Aging 2014. (PMID 24522464) [PubMed] [DOI] Abstract
BACKGROUND: Whereas chronological age (CA) cannot distinguish functional differences among individuals of the same age, the biological age (BA) may be used to reflect the functional state of the body. The purpose of this study was to construct an integral formula of the BA, by using principle component analysis (PCA). METHODS: The vital organ function of 505 healthy individuals of Han origin (age 35-91 years) was examined. A total of 114 indicators of cardiovascular, pulmonary, and brain functions, and clinical, inflammatory, genetic, psychological, and life habit factors were assessed as candidate indicators of aging. Candidate indicators were submitted with CA to correlation and redundancy analyses. The PCA method was used to build an integral formula of the BA for the population. RESULTS: Seven biomarkers were selected in accordance with a certain load standard. These biomarkers included the trail making test (TMT), pulse pressure (PP), mitral valve annulus ventricular septum of the peak velocity of early filling (MVES), minimum carotid artery intimal-medial thickness (IMTmin), maximum internal diameter of the carotid artery (Dmax), maximal midexpiratory flow rate 75/25 (MMEF75/25), and Cystatin C (CysC). The formula for the BA was: BA = 0.0685 (TMT) + 0.267 (PP) - 1.375 (MVES) + 22.443 (IMTmin) + 2.962 (Dmax) - 2.332 (MMEF75/25) + 16.104 (CysC) + 0.137 (CA) + 0.492. CONCLUSION: Several genetic and lifestyle indicators were considered as candidate markers of aging. However, ultimately, only markers reflecting the function of the vital organs were included in the BA formula. This study represents a useful attempt to employ multiple indicators to build a comprehensive BA evaluation formula of aging populations.
↑ ^{Jump up to: 32.0} ^32.1 ^32.2 ^32.3 ^32.4 ^32.5 Jee H et al.: Development and application of biological age prediction models with physical fitness and physiological components in Korean adults. Gerontology 2012. (PMID 22433233) [PubMed] [DOI] Abstract
BACKGROUND: Several biological age (BA) prediction models have been suggested with a variety of biomarkers. Valid models should be able to measure BA in a relatively short time period and predict subsequent physiological capability. Physiological and physical fitness variables have been shown to be distinctive markers for predicting BA and morbidity. The practical and noninvasive nature of such variables makes them useful as clinical assessment tools in estimating BA for in-depth diagnosis and corresponding intervention. OBJECTIVE: To identify, develop and evaluate biomarkers and BA prediction models and validate their clinical usefulness for the practical diagnosis of functional aging. METHODS: Fourteen variables were measured in 3,112 male and 1,233 female participants aged 30 and older between the years 2004 and 2007. Through a series of parsimonious stepwise elimination processes, two sets of 8 gender-specific variables were selected as candidate biomarkers for 1,604 men and 760 women. Principal component analysis, linear regression analysis and adjustment methods were further applied to obtain two sets of true BA (TBA) prediction models. The TBA models were examined for validity by comparing TBA to the corresponding chronological age (CA) with clinical risk factors. RESULTS: TBA prediction models with r(2) values of 0.638 and 0.672 were developed, each unique to men and women, respectively. The overall mean TBA and CA of the participants were 53.9 and 51.8 years, respectively, with a marginal difference of -2.1 and -1.3 years. The regression slopes or rates of TBA as a function of CA were 1.00 and 1.28 for men and women with r values of 0.799 and 0.820 (p < 0.001), respectively. In comparing TBA to CA rates between healthy and clinical risk groups, both sarcopenic and obese groups showed significant increases in TBA. CONCLUSIONS: The selected biomarkers encompass various complex physiopathological factors related to intrinsic and extrinsic physiological and functional aging. The BA prediction models based on the selected biomarkers could be practical in assessing BA for Korean adults.
↑ ^{Jump up to: 33.0} ^33.1 ^33.2 ^33.3 ^33.4 ^33.5 ^33.6 ^33.7 ^33.8 ^33.9 Hochschild R: Improving the precision of biological age determinations. Part 2: Automatic human tests, age norms and variability. Exp Gerontol 1989. (PMID 2583248) [PubMed] [DOI] Abstract
In order to eliminate variability due to test operators, procedures for measuring 12 physiological functions that are candidate biomarkers of aging have been automated. Data was collected from a norm group of 2462 male and female office workers using an instrument which requires no test operators, administers all 12 tests in about 45 min. per subject, computes biological age, prints out results, and stores data on floppy disks for transfer to other computers for analysis. This report a) describes the instrumentation and test procedures, b) presents normal age/sex standards for each of the 12 biomarkers, c) reports the variance of the data for each biomarker by sex, d) lists sources of biomarker variance, e) discusses criteria for biomarker selection and f) examines implications for information loss when biomarker data is combined to calculate biological age. After eliminating chronological age as a variable, the standard deviations of the frequency distributions of predicted age for individual biomarkers were found to vary from .226 to 1.075, a range of more than 4 to 1. Procedures are discussed for improving the ratio of useful-to-useless variance in calculating biological age.
↑ ^{Jump up to: 34.0} ^34.1 Zhang WG et al.: Peripheral Blood Leukocyte Telomere Length Is Associated with Age but Not Renal Function: A Cross-Sectional Follow-Up Study. J Nutr Health Aging 2018. (PMID 29380856) [PubMed] [DOI] Abstract
OBJECTIVES: We aimed to evaluate the relationship between baseline renal function and changes in telomere length in Han Chinese. METHODS: The telomere restriction fragment (TRF) length of leukocytes in the peripheral blood was measured in healthy volunteers recruited in 2014. The estimated glomerular filtration rate (eGFR) was calculated based on serum creatinine (Scr) and serum cystatin C (CysC)-eGFRcys and eGFRScr-cys through the Cockcroft-Gault formula (eGFRC-G) or the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI / eGFRCKD-EPI) equation. The correlation between telomere length changes over time and renal function was analyzed. RESULTS: Leukocyte TRF lengths were negatively correlated to age (r = -0.393, p < 0.001) and serum CysC (r = -0.180, p < 0.01), while positively associated with eGFRCKD-EPI, eGFRC-G, eGFRcys, and eGFRScr-cys (r = 0.182, 0.122, 0.290, and 0.254 respectively, p < 0.01). The 3-year change of telomere length was 46 bp/years. When adjusted for age, the associations between telomere length changes and baseline, subsequent TRF lengths, and serum CysC were no longer present. No association was observed between TRF length changes and renal function. CONCLUSION: The rate of telomere length changes was affected by age and baseline telomere length. The telomere length changes might be important markers for aging.
↑ Li Z et al.: Progress in biological age research. Front Public Health 2023. (PMID 37124811) [PubMed] [DOI] [Full text] Abstract
Biological age (BA) is a common model to evaluate the function of aging individuals as it may provide a more accurate measure of the extent of human aging than chronological age (CA). Biological age is influenced by the used biomarkers and standards in selected aging biomarkers and the statistical method to construct BA. Traditional used BA estimation approaches include multiple linear regression (MLR), principal component analysis (PCA), Klemera and Doubal's method (KDM), and, in recent years, deep learning methods. This review summarizes the markers for each organ/system used to construct biological age and published literature using methods in BA research. Future research needs to explore the new aging markers and the standard in select markers and new methods in building BA models.
↑ Jylhävä J et al.: Biological Age Predictors. EBioMedicine 2017. (PMID 28396265) [PubMed] [DOI] [Full text] Abstract
The search for reliable indicators of biological age, rather than chronological age, has been ongoing for over three decades, and until recently, largely without success. Advances in the fields of molecular biology have increased the variety of potential candidate biomarkers that may be considered as biological age predictors. In this review, we summarize current state-of-the-art findings considering six potential types of biological age predictors: epigenetic clocks, telomere length, transcriptomic predictors, proteomic predictors, metabolomics-based predictors, and composite biomarker predictors. Promising developments consider multiple combinations of these various types of predictors, which may shed light on the aging process and provide further understanding of what contributes to healthy aging. Thus far, the most promising, new biological age predictor is the epigenetic clock; however its true value as a biomarker of aging requires longitudinal confirmation.
↑ Erema VV et al.: Biological Age Predictors: The Status Quo and Future Trends. Int J Mol Sci 2022. (PMID 36499430) [PubMed] [DOI] [Full text] Abstract
There is no single universal biomarker yet to estimate overall health status and longevity prospects. Moreover, a consensual approach to the very concept of aging and the means of its assessment are yet to be developed. Markers of aging could facilitate effective health control, more accurate life expectancy estimates, and improved health and quality of life. Clinicians routinely use several indicators that could be biomarkers of aging. Duly validated in a large cohort, models based on a combination of these markers could provide a highly accurate assessment of biological age and the pace of aging. Biological aging is a complex characteristic of chronological age (usually), health-to-age concordance, and medically estimated life expectancy. This study is a review of the most promising techniques that could soon be used in routine clinical practice. Two main selection criteria were applied: a sufficient sample size and reliability based on validation. The selected biological age calculators were grouped according to the type of biomarker used: (1) standard clinical and laboratory markers; (2) molecular markers; and (3) epigenetic markers. The most accurate were the calculators, which factored in a variety of biomarkers. Despite their demonstrated effectiveness, most of them require further improvement and cannot yet be considered for use in standard clinical practice. To illustrate their clinical application, we reviewed their use during the COVID-19 pandemic.
↑ Lohman T et al.: Predictors of Biological Age: The Implications for Wellness and Aging Research. Gerontol Geriatr Med 2021. (PMID 34595331) [PubMed] [DOI] [Full text] Abstract
As healthspan and lifespan research breakthroughs have become more commonplace, the need for valid, practical markers of biological age is becoming increasingly paramount. The accessibility and affordability of biological age predictors that can reveal information about mortality and morbidity risk, as well as remaining years of life, has profound clinical and research implications. In this review, we examine 5 groups of aging biomarkers capable of providing accurate biological age estimations. The unique capabilities of these biomarkers have far reaching implications for the testing of both pharmaceutical and non-pharmaceutical interventions designed to slow or reverse biological aging. Additionally, the enhanced validity and availability of these tools may have increasingly relevant clinical value. The authors of this review explore those implications, with an emphasis on lifestyle modification research, and provide an overview of the current evidence regarding 5 biological age predictor categories: Telomere length, composite biomarkers, DNA methylation "epigenetic clocks," transcriptional predictors of biological age, and functional age predictors.
↑ Ashiqur Rahman S et al.: Deep learning for biological age estimation. Brief Bioinform 2021. (PMID 32363395) [PubMed] [DOI] [Full text] Abstract
Modern machine learning techniques (such as deep learning) offer immense opportunities in the field of human biological aging research. Aging is a complex process, experienced by all living organisms. While traditional machine learning and data mining approaches are still popular in aging research, they typically need feature engineering or feature extraction for robust performance. Explicit feature engineering represents a major challenge, as it requires significant domain knowledge. The latest advances in deep learning provide a paradigm shift in eliciting meaningful knowledge from complex data without performing explicit feature engineering. In this article, we review the recent literature on applying deep learning in biological age estimation. We consider the current data modalities that have been used to study aging and the deep learning architectures that have been applied. We identify four broad classes of measures to quantify the performance of algorithms for biological age estimation and based on these evaluate the current approaches. The paper concludes with a brief discussion on possible future directions in biological aging research using deep learning. This study has significant potentials for improving our understanding of the health status of individuals, for instance, based on their physical activities, blood samples and body shapes. Thus, the results of the study could have implications in different health care settings, from palliative care to public health.

[pmid37884697-1] {Jump up to: 1.0} ^1.1 Bortz J et al.: Biological age estimation using circulating blood biomarkers. Commun Biol 2023. (PMID 37884697) [PubMed] [DOI] [Full text] Abstract
Biological age captures physiological deterioration better than chronological age and is amenable to interventions. Blood-based biomarkers have been identified as suitable candidates for biological age estimation. This study aims to improve biological age estimation using machine learning models and a feature-set of 60 circulating biomarkers available from the UK Biobank (n = 306,116). We implement an Elastic-Net derived Cox model with 25 selected biomarkers to predict mortality risk (C-Index = 0.778; 95% CI [0.767-0.788]), which outperforms the well-known blood-biomarker based PhenoAge model (C-Index = 0.750; 95% CI [0.739-0.761]), providing a C-Index lift of 0.028 representing an 11% relative increase in predictive value. Importantly, we then show that using common clinical assay panels, with few biomarkers, alongside imputation and the model derived on the full set of biomarkers, does not substantially degrade predictive accuracy from the theoretical maximum achievable for the available biomarkers. Biological age is estimated as the equivalent age within the same-sex population which corresponds to an individual's mortality risk. Values ranged between 20-years younger and 20-years older than individuals' chronological age, exposing the magnitude of ageing signals contained in blood markers. Thus, we demonstrate a practical and cost-efficient method of estimating an improved measure of Biological Age, available to the general population.

[pmid28546743-2] {Jump up to: 2.0} ^2.1 Jia L et al.: Common methods of biological age estimation. Clin Interv Aging 2017. (PMID 28546743) [PubMed] [DOI] [Full text] Abstract
At present, no single indicator could be used as a golden index to estimate aging process. The biological age (BA), which combines several important biomarkers with mathematical modeling, has been proposed for >50 years as an aging estimation method to replace chronological age (CA). The common methods used for BA estimation include the multiple linear regression (MLR), the principal component analysis (PCA), the Hochschild's method, and the Klemera and Doubal's method (KDM). The fundamental differences in these four methods are the roles of CA and the selection criteria of aging biomarkers. In MLR and PCA, CA is treated as the selection criterion and an independent index. The Hochschild's method and KDM share a similar concept, making CA an independent variable. Previous studies have either simply constructed the BA model by one or compared the four methods together. However, reviews have yet to illustrate and compare the four methods systematically. Since the BA model is a potential estimation of aging for clinical use, such as predicting onset and prognosis of diseases, improving the elderly's living qualities, and realizing successful aging, here we summarize previous BA studies, illustrate the basic statistical steps, and thoroughly discuss the comparisons among the four common BA estimation methods.

[pmid5841151-3] {Jump up to: 3.0} ^3.1 ^3.2 ^3.3 ^3.4 ^3.5 ^3.6 ^3.7 ^3.8 Hollingsworth JW et al.: Correlations between tests of aging in Hiroshima subjects--an attempt to define "physiologic age". Yale J Biol Med 1965. (PMID 5841151) [PubMed] [Full text]

[pmid6873212-4] Dubina TL et al.: Biological age and its estimation. II. Assessment of biological age of albino rats by multiple regression analysis. Exp Gerontol 1983. (PMID 6873212) [PubMed] [DOI] Abstract
Multiple regression model of biological age (BA) theoretically gives agreement with the main concept of BA. When assessment of BA is based on the model, the age being in regression center, the method provides satisfactory results, whereas BA estimates of individuals in extreme age groups are erroneous. Investigation of male and female Wistar rats of age 5-29 months showed the BA estimates calculated from 4-10 physiological indices in young (5-7 mo) animals are overestimated, and in old (24-28 mo) animals are underestimated. Coincidence of average BA in one-age group of animals with its chronological age served as a criterion for the correspondence of the estimate to "real" BA. The paper also examines the following questions: the necessary and sufficient number of physiological indices; the sample size from the intact animal population to establish normal aging standard; the relationship between BA and animal weight.

[pmid3226152-5] {Jump up to: 5.0} ^5.1 Nakamura E et al.: Assessment of biological age by principal component analysis. Mech Ageing Dev 1988. (PMID 3226152) [PubMed] [DOI] Abstract
A method of assessing biological age by the application of principal component analysis is reported. Healthy individuals (462) randomly selected from about 6000 men who had taken a 2-day health examination were studied. Out of the 30 physiological variables examined in routine check-ups, 11 variables were selected as suitable for the assessment of biological age based on the results of factor analysis and the physiological meaning of each test. This variable set was then submitted to principal component analysis, and the 1st principal component obtained from this analysis was used as an equation for assessing one's biological age. However, the biological age calculated from this equation is expressed as a score, so the estimated score was transformed to years (biological age) using the T-score idea. The biological age estimated by this method is practically useful and theoretically valid in contrast with the multiple regression model, because this approach eliminates and overcomes the following 2 big problems of the multiple regression model: (1) the distortion of the individual biological age at the regression edges; and (2) a theoretical contradiction in that a perfect model will merely be predicting the subject's chronological age, not his biological age.

[pmid950448-6] {Jump up to: 6.0} ^6.1 ^6.2 ^6.3 ^6.4 ^6.5 ^6.6 Webster IW & Logie AR: A relationship between functional age and health status in female subjects. J Gerontol 1976. (PMID 950448) [PubMed] [DOI] Abstract
A multiple regression equation was used to predict age from seven clinical variables in 1080 apparently well female subjects. A multiple correlation coefficient of R = 0.77 was achieved by five of the variables: timed forced expiratory volume, systolic blood pressure, plasma urea nitrogen, cholesterol, and alkaline phosphatase. On the basis of selection by medical questionnaire responses and other objective criteria, 9% of the subjects were nonsmokers and healthier than the rest. These selected subjects showed a significant reduction in preducted age. Within this group, subjective perception of health was associated with differences in predicted age: poor health with an increase and good health with a decrease in functional age. This study of functional age was based on the healthiest segment of the population in order to minimize the effect of overt pathological processes on the aging rate. An association has been demonstrated between health impairment and predicted age as a measure of the aging rate.

[pmid2737197-7] {Jump up to: 7.0} ^7.1 ^7.2 ^7.3 ^7.4 ^7.5 ^7.6 ^7.7 ^7.8 Nakamura E et al.: Biological age versus physical fitness age. Eur J Appl Physiol Occup Physiol 1989. (PMID 2737197) [PubMed] [DOI] Abstract
A population of healthy middle-aged (n = 69) and elderly men (n = 12), who participated in a health promotion program, was studied to determine whether really physically fit individuals are in good biological condition, and also whether improvement of physical fitness in the middle-aged and the elderly reduces their "rate of aging". Biological and physical fitness ages of the individuals studied were estimated from the data for 18 physiological function tests and 5 physical fitness tests, respectively, by a principal component model. The correlation coefficient between the estimated biological and physical fitness ages was 0.72 (p less than 0.01). Detailed analyses of the relationship between the estimated biological and physical fitness ages revealed that those who manifested a higher ("older") physical fitness age did not necessarily have a higher biological age, but those who manifested a lower ("younger") physical fitness age were also found to have a lower biological age. These results suggested that there were considerable individual variations in the relationship between biological condition and physical fitness among individuals with an old physical fitness age, but those who were in a state of high physical fitness maintained a relatively good biological condition. The data regarding the elderly men who had maintained a regular exercise program indicated that their estimated biological ages were considerably younger than the expected values. This might suggest that in older individuals regular physical activity may provide physiological improvements which in turn might reduce "the rate of aging".

[pmid16318865-8] {Jump up to: 8.0} ^8.1 ^8.2 ^8.3 Klemera P & Doubal S: A new approach to the concept and computation of biological age. Mech Ageing Dev 2006. (PMID 16318865) [PubMed] [DOI] Abstract
The lack of exact definition of the concept of biological age (BA) is a typical feature of works concerning BA. That is why comparison of results of various published methods makes little sense and eventual proof of their optimality is impossible. Based on natural and simple presumptions, an attempt to express mathematically the supposed relation between chronological age (CA) and BA has proven to be unexpectedly fruitful. In the present paper, an optimum method of estimation of BA, which is easily applicable even in nonlinear cases, is derived. Moreover, the method allows evaluating the precision of the estimates and also offers tools for validation of presumptions of the method. A special feature of the method is that CA should be used as a standard biomarker, leading to essential improving the precision of BA-estimate and illuminating relativity of the known "paradox of biomarkers". All theoretical results of the method were fully approved by means of a special simulation program. Further, the theory and the results of the simulation have proven that many published results of BA-estimates using multiple linear regression (MLR) are very probably disserviceable because CA is typically more precise estimate of BA than estimates computed by MLR. This unpleasant conclusion also concerns methods, which use MLR as the final step after transformation of the battery of biomarkers by factor analysis or by principal component analysis.

[pmid2684676-9] {Jump up to: 9.0} ^9.1 Hochschild R: Improving the precision of biological age determinations. Part 1: A new approach to calculating biological age. Exp Gerontol 1989. (PMID 2684676) [PubMed] [DOI] Abstract
In calculating biological age, almost all prior studies used multiple regression of chronological age on scores of biomarkers of aging. Multiple regression is invalid for this purpose for three, and in some circumstances four, reasons. These are: a) weighting of the contribution of each biomarker's scores according to strength of association with chronological age; b) regression of calculated ages to sample mean age and the inadequacy of proposed corrections; c) frequent occurrence of regression coefficients whose sign equates poorer adult performance on a test to younger biological ages; and d) multicollinearity when lung function scores and height are on the same side of the regression equation. An alternative method for calculating biological age is outlined. Regression to sample mean age and its solution are illustrated on data for highest audible pitch, one of 12 biomarkers measured in a study of 2462 office workers. Prior published studies employing multiple regression to calculate biological age appear to have been in error.

[pmid20005245-10] {Jump up to: 10.00} ^10.01 ^10.02 ^10.03 ^10.04 ^10.05 ^10.06 ^10.07 ^10.08 ^10.09 ^10.10 ^10.11 ^10.12 ^10.13 ^10.14 ^10.15 ^10.16 ^10.17 ^10.18 Cho IH et al.: An empirical comparative study on biological age estimation algorithms with an application of Work Ability Index (WAI). Mech Ageing Dev 2010. (PMID 20005245) [PubMed] [DOI] Abstract
In this study, we described the characteristics of five different biological age (BA) estimation algorithms, including (i) multiple linear regression, (ii) principal component analysis, and somewhat unique methods developed by (iii) Hochschild, (iv) Klemera and Doubal, and (v) a variant of Klemera and Doubal's method. The objective of this study is to find the most appropriate method of BA estimation by examining the association between Work Ability Index (WAI) and the differences of each algorithm's estimates from chronological age (CA). The WAI was found to be a measure that reflects an individual's current health status rather than the deterioration caused by a serious dependency with the age. Experiments were conducted on 200 Korean male participants using a BA estimation system developed principally under the concept of non-invasive, simple to operate and human function-based. Using the empirical data, BA estimation as well as various analyses including correlation analysis and discriminant function analysis was performed. As a result, it had been confirmed by the empirical data that Klemera and Doubal's method with uncorrelated variables from principal component analysis produces relatively reliable and acceptable BA estimates.

[pmid23213031-11] {Jump up to: 11.0} ^11.1 ^11.2 ^11.3 ^11.4 ^11.5 ^11.6 Levine ME: Modeling the rate of senescence: can estimated biological age predict mortality more accurately than chronological age?. J Gerontol A Biol Sci Med Sci 2013. (PMID 23213031) [PubMed] [DOI] [Full text] Abstract
Biological age (BA) is useful for examining differences in aging rates. Nevertheless, little consensus exists regarding optimal methods for calculating BA. The aim of this study is to compare the predictive ability of five BA algorithms. The sample included 9,389 persons, aged 30-75 years, from National Health and Nutrition Examination Survey III. During the 18-year follow-up, 1,843 deaths were counted. Each BA algorithm was compared with chronological age on the basis of predictive sensitivity and strength of association with mortality. Results found that the Klemera and Doubal method was the most reliable predictor of mortality and performed significantly better than chronological age. Furthermore, when included with chronological age in a model, Klemera and Doubal method had more robust predictive ability and caused chronological age to no longer be significantly associated with mortality. Given the potential of BA to highlight heterogeneity, the Klemera and Doubal method algorithm may be useful for studying a number of questions regarding the biology of aging.

[pmid28110151-12] {Jump up to: 12.00} ^12.01 ^12.02 ^12.03 ^12.04 ^12.05 ^12.06 ^12.07 ^12.08 ^12.09 ^12.10 ^12.11 ^12.12 ^12.13 ^12.14 ^12.15 ^12.16 ^12.17 Jee H & Park J: Selection of an optimal set of biomarkers and comparative analyses of biological age estimation models in Korean females. Arch Gerontol Geriatr 2017. (PMID 28110151) [PubMed] [DOI] Abstract
To date, an optimal working model which predicts biological age (BA) with a set of working biomarkers has not been devised to represent the Korean female population. Accuracy of prediction and applicability are required of an optimal set of commonly assessed biomarkers to provide information on the health status. The goal of this study was to identify a set of biomarkers that represent the aging process and to develop and compare different BA prediction models to elucidate the most fitting and applicable model for providing information on health status in the Korean female population. Using a series of selection processes, eight clinically assessable variables were selected by analyzing relations between 31 clinical variables and chronologic age in 912 normal, healthy individuals among 3642 female participants with ages ranging from 30 to 80 years. The multiple linear regression (MLR), principal component analysis (PCA), and the Klemera-Doubal (KDM) statistical methods were applied to obtain three different sets of BA prediction models. These three models were assessed by calculating and performing the coefficient determinations (r2), regression slopes, effect sizes, pairwise t-tests, and Bland-Altman plots. The BA models were further compared for the applicability by calculating the BAs of clinical risk groups. MLR showed the narrowing effects at the either ends of the age spectrum with greatest effect sizes. PCA showed the greatest degree of dispersion and deviation from the regression center. These MLR and PCA trends were also exhibited by clinically risk groups. In conclusion, the KDM BA prediction model based on the selected biomarkers was found to provide the most reliable and stable results for the practical assessment of BA.

[pmid26017442-13] {Jump up to: 13.0} ^13.1 ^13.2 ^13.3 LeCun Y et al.: Deep learning. Nature 2015. (PMID 26017442) [PubMed] [DOI] Abstract
Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.

[pmid27191382-14] {Jump up to: 14.0} ^14.1 Putin E et al.: Deep biomarkers of human aging: Application of deep neural networks to biomarker development. Aging (Albany NY) 2016. (PMID 27191382) [PubMed] [DOI] [Full text] Abstract
One of the major impediments in human aging research is the absence of a comprehensive and actionable set of biomarkers that may be targeted and measured to track the effectiveness of therapeutic interventions. In this study, we designed a modular ensemble of 21 deep neural networks (DNNs) of varying depth, structure and optimization to predict human chronological age using a basic blood test. To train the DNNs, we used over 60,000 samples from common blood biochemistry and cell count tests from routine health exams performed by a single laboratory and linked to chronological age and sex. The best performing DNN in the ensemble demonstrated 81.5 % epsilon-accuracy r = 0.90 with R(2) = 0.80 and MAE = 6.07 years in predicting chronological age within a 10 year frame, while the entire ensemble achieved 83.5% epsilon-accuracy r = 0.91 with R(2) = 0.82 and MAE = 5.55 years. The ensemble also identified the 5 most important markers for predicting human chronological age: albumin, glucose, alkaline phosphatase, urea and erythrocytes. To allow for public testing and evaluate real-life performance of the predictor, we developed an online system available at http://www.aging.ai. The ensemble approach may facilitate integration of multi-modal data linked to chronological age and sex that may lead to simple, minimally invasive, and affordable methods of tracking integrated biomarkers of aging in humans and performing cross-species feature importance analysis.

[pmid29340580-15] Mamoshina P et al.: Population Specific Biomarkers of Human Aging: A Big Data Study Using South Korean, Canadian, and Eastern European Patient Populations. J Gerontol A Biol Sci Med Sci 2018. (PMID 29340580) [PubMed] [DOI] [Full text] Abstract
Accurate and physiologically meaningful biomarkers for human aging are key to assessing antiaging therapies. Given ethnic differences in health, diet, lifestyle, behavior, environmental exposures, and even average rate of biological aging, it stands to reason that aging clocks trained on datasets obtained from specific ethnic populations are more likely to account for these potential confounding factors, resulting in an enhanced capacity to predict chronological age and quantify biological age. Here, we present a deep learning-based hematological aging clock modeled using the large combined dataset of Canadian, South Korean, and Eastern European population blood samples that show increased predictive accuracy in individual populations compared to population specific hematologic aging clocks. The performance of models was also evaluated on publicly available samples of the American population from the National Health and Nutrition Examination Survey (NHANES). In addition, we explored the association between age predicted by both population specific and combined hematological clocks and all-cause mortality. Overall, this study suggests (a) the population specificity of aging patterns and (b) hematologic clocks predicts all-cause mortality. The proposed models were added to the freely-available Aging.AI system expanding the range of tools for analysis of human aging.

[pmid30644411-16] Mamoshina P et al.: Blood Biochemistry Analysis to Detect Smoking Status and Quantify Accelerated Aging in Smokers. Sci Rep 2019. (PMID 30644411) [PubMed] [DOI] [Full text] Abstract
There is an association between smoking and cancer, cardiovascular disease and all-cause mortality. However, currently, there are no affordable and informative tests for assessing the effects of smoking on the rate of biological aging. In this study we demonstrate for the first time that smoking status can be predicted using blood biochemistry and cell count results andthe recent advances in artificial intelligence (AI). By employing age-prediction models developed using supervised deep learning techniques, we found that smokers exhibited higher aging rates than nonsmokers, regardless of their cholesterol ratios and fasting glucose levels. We further used those models to quantify the acceleration of biological aging due to tobacco use. Female smokers were predicted to be twice as old as their chronological age compared to nonsmokers, whereas male smokers were predicted to be one and a half times as old as their chronological age compared to nonsmokers. Our findings suggest that deep learning analysis of routine blood tests could complement or even replace the current error-prone method of self-reporting of smoking status and could be expanded to assess the effect of other lifestyle and environmental factors on aging.

[pmid34453631-17] Gialluisi A et al.: Exploring domains, clinical implications and environmental associations of a deep learning marker of biological ageing. Eur J Epidemiol 2022. (PMID 34453631) [PubMed] [DOI] Abstract
Deep Neural Networks (DNN) have been recently developed for the estimation of Biological Age (BA), the hypothetical underlying age of an organism, which can differ from its chronological age (CA). Although promising, these population-specific algorithms warrant further characterization and validation, since their biological, clinical and environmental correlates remain largely unexplored. Here, an accurate DNN was trained to compute BA based on 36 circulating biomarkers in an Italian population (N = 23,858; age ≥ 35 years; 51.7% women). This estimate was heavily influenced by markers of metabolic, heart, kidney and liver function. The resulting Δage (BA-CA) significantly predicted mortality and hospitalization risk for all and specific causes. Slowed biological aging (Δage < 0) was associated with higher physical and mental wellbeing, healthy lifestyles (e.g. adherence to Mediterranean diet) and higher socioeconomic status (educational attainment, household income and occupational status), while accelerated aging (Δage > 0) was associated with smoking and obesity. Together, lifestyles and socioeconomic variables explained ~48% of the total variance in Δage, potentially suggesting the existence of a genetic basis. These findings validate blood-based biological aging as a marker of public health in adult Italians and provide a robust body of knowledge on its biological architecture, clinical implications and potential environmental influences.

[pmid29581467-18] Pyrkov TV et al.: Extracting biological age from biomedical data via deep learning: too much of a good thing?. Sci Rep 2018. (PMID 29581467) [PubMed] [DOI] [Full text] Abstract
Age-related physiological changes in humans are linearly associated with age. Naturally, linear combinations of physiological measures trained to estimate chronological age have recently emerged as a practical way to quantify aging in the form of biological age. In this work, we used one-week long physical activity records from a 2003-2006 National Health and Nutrition Examination Survey (NHANES) to compare three increasingly accurate biological age models: the unsupervised Principal Components Analysis (PCA) score, a multivariate linear regression, and a state-of-the-art deep convolutional neural network (CNN). We found that the supervised approaches produce better chronological age estimations at the expense of a loss of the association between the aging acceleration and all-cause mortality. Consequently, we turned to the NHANES death register directly and introduced a novel way to train parametric proportional hazards models suitable for out-of-the-box implementation with any modern machine learning software. As a demonstration, we produced a separate deep CNN for mortality risks prediction that outperformed any of the biological age or a simple linear proportional hazards model. Altogether, our findings demonstrate the emerging potential of combined wearable sensors and deep learning technologies for applications involving continuous health risk monitoring and real-time feedback to patients and care providers.

[pmid33744131-19] Raghu VK et al.: Deep Learning to Estimate Biological Age From Chest Radiographs. JACC Cardiovasc Imaging 2021. (PMID 33744131) [PubMed] [DOI] Abstract
OBJECTIVES: The goal of this study was to assess whether a deep learning estimate of age from a chest radiograph image (CXR-Age) can predict longevity beyond chronological age. BACKGROUND: Chronological age is an imperfect measure of longevity. Biological age, a measure of overall health, may improve personalized care. This paper proposes a new way to estimate biological age using a convolutional neural network that takes as input a CXR image and outputs a chest x-ray age (in years) as a measure of long-term mortality risk. METHODS: CXR-Age was developed using CXR from 116,035 individuals and validated in 2 held-out testing sets: 1) 75% of the CXR arm of PLCO (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial) (N = 40,967); and 2) the CXR arm of NLST (National Lung Screening Trial) (N = 5,414). CXR-Age was compared to chronological age and a multivariable regression model of chronological age, risk factors, and radiograph findings to predict all-cause and cardiovascular mortality with a maximum 23 years and 13 years of follow-up, respectively. The primary outcome was observed mortality; results are provided for the testing datasets only. RESULTS: In the PLCO testing dataset, a 5-year increase in CXR-Age carried a higher risk of all-cause mortality than a 5-year increase in chronological age (CXR-Age hazard ratio [HR]: 2.26 [95% confidence interval (CI): 2.24 to 2.29] vs. chronological age HR: 1.77 [95% CI: 1.75 to 1.78]; p < 0.001). A similar pattern was found for cardiovascular mortality (CXR-Age cause-specific HR: 2.45 per 5 years [95% CI: 2.34 to 2.56] vs. chronological age HR: 1.82 per 5 years [95% CI: 1.74 to 1.90]). Similar results were seen for both outcomes in the NLST external testing dataset. Adding CXR-Age to the multivariable model resulted in significant improvements for predicting both outcomes in both testing datasets (p < 0.001 for all comparisons). CONCLUSIONS: Based on a CXR image, CXR-Age predicted long-term all-cause and cardiovascular mortality.

[pmid31388024-20] Rahman SA & Adjeroh DA: Deep Learning using Convolutional LSTM estimates Biological Age from Physical Activity. Sci Rep 2019. (PMID 31388024) [PubMed] [DOI] [Full text] Abstract
Human age estimation is an important and difficult challenge. Different biomarkers and numerous approaches have been studied for biological age estimation, each with its advantages and limitations. In this work, we investigate whether physical activity can be exploited for biological age estimation for adult humans. We introduce an approach based on deep convolutional long short term memory (ConvLSTM) to predict biological age, using human physical activity as recorded by a wearable device. We also demonstrate five deep biological age estimation models including the proposed approach and compare their performance on the NHANES physical activity dataset. Results on mortality hazard analysis using both the Cox proportional hazard model and Kaplan-Meier curves each show that the proposed method for estimating biological age outperforms other state-of-the-art approaches. This work has significant implications in combining wearable sensors and deep learning techniques for improved health monitoring, for instance, in a mobile health environment. Mobile health (mHealth) applications provide patients, caregivers, and administrators continuous information about a patient, even outside the hospital.

[pmid19940465-21] {Jump up to: 21.0} ^21.1 ^21.2 ^21.3 ^21.4 ^21.5 ^21.6 Bai X et al.: Evaluation of biological aging process - a population-based study of healthy people in China. Gerontology 2010. (PMID 19940465) [PubMed] [DOI] Abstract
BACKGROUND: Although there have been cross-sectional and longitudinal studies examining biological age (BA) with chronological age (CA)-related changes in physical, physiological, biochemical, and hormonal variables, few studies have performed echocardiographic evaluation of the cardiovascular system and inflammatory biomarkers. Furthermore, little is known about biomarkers of aging and BA score (BAS) for healthy people in China. OBJECTIVES: The purpose of this study was to identify the biomarkers of healthy aging and to establish BAS for healthy people in China. METHODS: We examined 2,876 men and women aged 30-98 years old in three Chinese cities, and 852 healthy subjects were assessed with 108 physical, morphological, physiological and biochemical variables. After excluding binary variables, variables that had a correlation coefficient with CA of < or =0.25 and redundant variables, eight variables including CA, arterial pulse pressure (PP), intima-media thickness (IMT), end diastolic velocity (EDV), ratio of peak velocity of early filling to atrial filling (E/A), mitral valve annulus lateral wall of peak velocity of early filling (MVEL), cystatin C (CYSC), and fibrinogen (FIB) were selected as candidate biomarkers of aging based on a factor-weighted BAS composite for predicting BA. RESULTS: The BAS equation was 0.248 (CA) + 0.195 (IMT) - 0.196 (EDV) - 0.167 (E/A) - 0.166 (MVEL) + 0.188 (PP) + 0.182 (FIB) + 0.193 (CYSC). Individual BAS were significantly correlated with CA (r = 0.893, p < 0.001). Biological aging rate predicted by BAS was accelerated with increases in CA, and peaked when healthy men and women reached > or =75 years of age. CONCLUSIONS: Our data suggest that BAS is superior to CA in assessing the rate of aging in healthy Chinese people. The cardiovascular variables play a crucial role in the evaluation of biological aging. Biological aging rate appears to be age specific.

[pmid24659482-22] {Jump up to: 22.0} ^22.1 ^22.2 ^22.3 ^22.4 ^22.5 ^22.6 Zhang WG et al.: Select aging biomarkers based on telomere length and chronological age to build a biological age equation. Age (Dordr) 2014. (PMID 24659482) [PubMed] [DOI] [Full text] Abstract
The purpose of this study is to build a biological age (BA) equation combining telomere length with chronological age (CA) and associated aging biomarkers. In total, 139 healthy volunteers were recruited from a Chinese Han cohort in Beijing. A genetic index, renal function indices, cardiovascular function indices, brain function indices, and oxidative stress and inflammation indices (C-reactive protein [CRP]) were measured and analyzed. A BA equation was proposed based on selected parameters, with terminal telomere restriction fragment (TRF) and CA as the two principal components. The selected aging markers included mitral annulus peak E anterior wall (MVEA), intima-media thickness (IMT), cystatin C (CYSC), D-dimer (DD), and digital symbol test (DST). The BA equation was: BA = −2.281TRF + 26.321CYSC + 0.025DD − 104.419MVEA + 34.863IMT − 0.265DST + 0.305CA + 26.346. To conclude, telomere length and CA as double benchmarks may be a new method to build a BA.

[pmid29188884-23] {Jump up to: 23.0} ^23.1 ^23.2 Zhang W et al.: Model Construction for Biological Age Based on a Cross-Sectional Study of a Healthy Chinese Han population. J Nutr Health Aging 2017. (PMID 29188884) [PubMed] [DOI] Abstract
OBJECTIVES: Biological age (BA) has been proposed to evaluate the aging status in an objective way instead of chronological age (CA). The purpose of our study is to construct a more precise formula of BA in the cross-sectional study based on a largest-ever sample of our studies. This formula aims at better evaluation of body function and exploring the disciplines of aging in different genders and age stages. METHODS: A total of 1,373 healthy Chinese Han (age range, 19-93 years) were recruited from five cities in China, including 581 males and 792 females. Physical examination, blood routine, blood chemistry, and other lab tests were performed to obtain a total of 74 clinical variables. Then, the principal component analysis (PCA) was used to select variables and estimate BA. The BA formula was further validated in a population with some diseases (n=266), including cardiovascular diseases, type 2 diabetes, kidney diseases, pulmonary diseases, cancer and disorders in nervous system. RESULTS: The BA formula was constructed as follows: BA = 0.358 (pulse pressure) + 0.258 (trail making test) - 11.552 (mitral valve E/A peak) + 26.383 (minimum intima-media thickness) + 31.965 (Cystatin C) + 0.163 (CA) - 3.902. In validation of the formula, BAs of patients were older than those of healthy persons. The BA accelerates faster in the middle-aged population than in the elderly population (>75 years old). CONCLUSION: This BA formula can reflect health condition changes of aging better than CA in a Chinese Han population.

[pmid2282902-24] {Jump up to: 24.0} ^24.1 ^24.2 ^24.3 ^24.4 ^24.5 ^24.6 ^24.7 ^24.8 Nakamura E et al.: Biological age versus physical fitness age in women. Eur J Appl Physiol Occup Physiol 1990. (PMID 2282902) [PubMed] [DOI] Abstract
The purpose of this study was to determine whether adult women who are in a state of high physical fitness are in a good state biologically, in terms of biological and physical fitness ages as estimated by statistical means. The subjects were 65 healthy Japanese women (aged 20-64 years). Biological and physical fitness ages were estimated from the data for 18 physiological function tests and 5 physical fitness tests, respectively, by a principal component model. The correlation coefficient between biological and physical fitness ages was 0.70 (P less than 0.01), which was generally regarded as a high correlation. Therefore, those who were in a state of high physical fitness were considered to be in good biological condition. This result is in good agreement with the results (r = 0.72) from adult men, on whom we reported previously. A statistical analysis to ascertain the relative importance of each contributory variable associated with the variance in biological age suggested that routine clinical evaluation of blood pressure and lipid metabolism might play an important role in determining not only the presence and severity of vascular disease but also the rate of biological aging in women.

[pmid17921421-25] {Jump up to: 25.0} ^25.1 ^25.2 ^25.3 ^25.4 Nakamura E & Miyao K: A method for identifying biomarkers of aging and constructing an index of biological age in humans. J Gerontol A Biol Sci Med Sci 2007. (PMID 17921421) [PubMed] [DOI] Abstract
This study was conducted to identify biomarkers of aging and to construct an index of biological age in humans. Healthy adult men (n = 86) who had received an annual health examination from 1992 through 1998 were studied. From 29 physiological variables, five variables (forced expiratory volume in 1 second, systolic blood pressure, hematocrit, albumin, blood urea nitrogen) were selected as candidate biomarkers of aging. Five candidate biomarkers expressed substantial covariance along one principal component. The first principal component obtained from a principal component analysis was used to calculate biological age scores (BAS). Individual BAS showed high longitudinal stability of age-related changes. Age-related changes of BAS are characterized by three components: age, peak functional capacity, and aging rate. A logistic regression analysis suggested that aging rate was influenced by environmental factors, but peak functional capacity was almost independent of environmental factors.

[pmid18597867-26] {Jump up to: 26.00} ^26.01 ^26.02 ^26.03 ^26.04 ^26.05 ^26.06 ^26.07 ^26.08 ^26.09 ^26.10 Park J et al.: Developing a biological age assessment equation using principal component analysis and clinical biomarkers of aging in Korean men. Arch Gerontol Geriatr 2009. (PMID 18597867) [PubMed] [DOI] Abstract
The purpose of the present study is to find clinically useful candidate biomarkers of aging, and using these to develop an equation measuring biological age (BA) in Korean men, then to validate the clinical usefulness of it. Among 4288 men who received medical health examinations, we selected 1588 men who met the normality criteria of each variable. We assumed that chronological ages (CA) of healthy persons represent the BA of them. Variables showing significant correlations with CA were selected. Redundant variables were excluded. We selected 11 variables: VO(2)max, percent body fat (%BF), waist circumference (WC), forced expiratory volume in 1 s (FEV1), systolic blood pressure (SBP), low density cholesterol (LDLCH), blood urea nitrogen (BUN), serum albumin (SA), erythrocyte sedimentation rate(ESR) hearing threshold (HT), and glycosylated hemoglobin (HBA1C). These 11 variables were then submitted into principal component analysis (PCA) and standardized BA scores were obtained. Using them and T-scale idea, the following equation to assess BA was developed: BA=-28.7+0.83(%BF)+0.48(WC)+0.13(SBP)-0.27(VO(2)max)+0.19(HT)-3.1(FEV1)+0.32(BUN)+0.06(LDLCH)-3.0(SA)+0.34(ESR)+4.6(HBA1C). We compared the BA of 3122 men by their fasting glucose and age level. The BA of the higher glucose level group was significantly higher than that of others at all CA levels. The selected 11 biomarkers encompassed known clinically important factors of adult diseases and functional disabilities. This BA assessment equation can be used in the general Korean male population and it proved to be clinically useful.

[pmid12672981-27] {Jump up to: 27.0} ^27.1 ^27.2 ^27.3 ^27.4 Ueno LM et al.: Biomarkers of aging in women and the rate of longitudinal changes. J Physiol Anthropol Appl Human Sci 2003. (PMID 12672981) [PubMed] [DOI] Abstract
The purposes of this study were (1) to estimate biological age score (BAS) in Japanese healthy women based on the 4-7 years longitudinal data for physiological, hematological and biochemical examinations and (2) to examine the rate of aging changes in adult women based on the estimated BAS. The samples consisted of cross-sectional (n=981) and longitudinal (n=110) groups. Out of 31 variables examined, five variables (forced expiratory volume in 1.0 s, systolic blood pressure, mean corpuscular hemoglobin, glucose, albumin/globulin ratio) that met the following criteria: 1) significant cross-sectional correlation with age; 2) significant longitudinal change in the same direction as the cross-sectional correlation; and (3) assessment of redundancy, were selected as candidate biomarkers of aging. This variable set was then submitted into a principal component analysis, and the first principal component obtained from this analysis was used as an equation for assessing one's BAS. Individual BAS showed a high longitudinal stability of age-related changes, suggesting high predictive validity of our newly developed aging measurement equation. However, changes in the aging rate based on the estimated BAS were not constant. The mean slopes of the regression lines of BAS for the three age groups (age<45, 45</=age<65 yrs, 65</=age) were 0.095, 0.065, 0.138, respectively. One-way analysis of variance detected a significant difference (F=5.14, p<0.01) among the three age groups. These results suggest that the rate of aging in adult women is relatively slower until 65 years of age, but after 65, the rate of aging shows a rapid increase. We concluded that the longitudinal method used for selection of variables to compute the BAS was useful and theoretically valid compared to those obtained from cross-sectional data analysis.

[pmid18840798-28] {Jump up to: 28.0} ^28.1 ^28.2 ^28.3 ^28.4 Nakamura E & Miyao K: Sex differences in human biological aging. J Gerontol A Biol Sci Med Sci 2008. (PMID 18840798) [PubMed] [DOI] Abstract
This study aims to clarify sex differences in human biological aging and to explore the gender gaps in health and longevity. Eighty-six men and 93 women who received a 2-day routine health checkup for 6-7 years beginning in 1992 at the Kyoto Red Cross Hospital were selected. Five candidate biomarkers of aging (forced expiratory volume in 1.0 second per square of height [FEV(1)/Ht(2)], systolic blood pressure [SBP], red blood cells [RBC], albumin [ALBU], and blood urea nitrogen [BUN]) were selected from 29 physiological variables. Individual biological ages (BAS) were estimated from these five biomarkers by a principal component model. From the investigation of the longitudinal changes of individual BAS, it was suggested that (i) beyond 65 years, the rate of aging showed a rapid increase, and (ii) women had relatively lower functional capabilities compared with men, but the rate of aging was slower than that of men, suggesting that these differences might present both disadvantages and advantages for women with regard to health and longevity.

[pmid8803500-29] {Jump up to: 29.0} ^29.1 ^29.2 ^29.3 ^29.4 ^29.5 ^29.6 Nakamura E et al.: Effects of habitual physical exercise on physiological age in men aged 20-85 years as estimated using principal component analysis. Eur J Appl Physiol Occup Physiol 1996. (PMID 8803500) [PubMed] [DOI] Abstract
A population of 221 healthy adult men (aged 20-85 years) was studied to determine whether those who exercised regularly were in good biological condition, and also whether those who were in a state of high physical fitness were in a good state biologically, in terms of physiological age (PA) and physical fitness age (FA) as estimated by principal component analysis. A group of 17 physiological function tests and 5 physical fitness tests were employed to estimate PA and FA, respectively. The results of this study indicated that those who maintained high physical fitness at all age decade groups from 20 to 79 years had a trend towards maintaining a relatively lower PA (physiologically younger). Mean PA and FA of the trained group were younger by 4.7 and 7.3 years, respectively than those of the untrained group. In addition, the slope of regression line of PA on chronological age was more gentle in the trained group than that in the untrained group. These results would suggest that those who are in a state of high physical fitness maintain a relatively good physiological condition, and that regular physical exercise may delay physiological changes normally seen with aging, and consequently may increase the life span.

[pmid6967883-30] {Jump up to: 30.0} ^30.1 ^30.2 ^30.3 ^30.4 Borkan GA & Norris AH: Assessment of biological age using a profile of physical parameters. J Gerontol 1980. (PMID 6967883) [PubMed] [DOI] Abstract
The present study describes a new approach to the assessment of biological age in adults using a profile of physical parameters. The investigation was based on data from 1086 adult male participants in the aging study of the Gerontology Research Ctr., Baltimore, MD. For each of 24 age-related variables, data were transformed into biological age scores reflecting a man's status relative to his chronological age peers. Analysis of the mean biological age profiles of men who were estimated by physicians as looking older than their age showed them to be biologically older on the profile parameters as well. Comparison of age-corrected scores of 166 men who have died with those of survivors reveals the deceased group to have been biologically older than the survivors at the time they were measured. These results suggest the value of this technique in investigating interindividual variation in the aging process.

[pmid24522464-31] {Jump up to: 31.0} ^31.1 ^31.2 ^31.3 ^31.4 ^31.5 Zhang WG et al.: Construction of an integral formula of biological age for a healthy Chinese population using principle component analysis. J Nutr Health Aging 2014. (PMID 24522464) [PubMed] [DOI] Abstract
BACKGROUND: Whereas chronological age (CA) cannot distinguish functional differences among individuals of the same age, the biological age (BA) may be used to reflect the functional state of the body. The purpose of this study was to construct an integral formula of the BA, by using principle component analysis (PCA). METHODS: The vital organ function of 505 healthy individuals of Han origin (age 35-91 years) was examined. A total of 114 indicators of cardiovascular, pulmonary, and brain functions, and clinical, inflammatory, genetic, psychological, and life habit factors were assessed as candidate indicators of aging. Candidate indicators were submitted with CA to correlation and redundancy analyses. The PCA method was used to build an integral formula of the BA for the population. RESULTS: Seven biomarkers were selected in accordance with a certain load standard. These biomarkers included the trail making test (TMT), pulse pressure (PP), mitral valve annulus ventricular septum of the peak velocity of early filling (MVES), minimum carotid artery intimal-medial thickness (IMTmin), maximum internal diameter of the carotid artery (Dmax), maximal midexpiratory flow rate 75/25 (MMEF75/25), and Cystatin C (CysC). The formula for the BA was: BA = 0.0685 (TMT) + 0.267 (PP) - 1.375 (MVES) + 22.443 (IMTmin) + 2.962 (Dmax) - 2.332 (MMEF75/25) + 16.104 (CysC) + 0.137 (CA) + 0.492. CONCLUSION: Several genetic and lifestyle indicators were considered as candidate markers of aging. However, ultimately, only markers reflecting the function of the vital organs were included in the BA formula. This study represents a useful attempt to employ multiple indicators to build a comprehensive BA evaluation formula of aging populations.

[pmid22433233-32] {Jump up to: 32.0} ^32.1 ^32.2 ^32.3 ^32.4 ^32.5 Jee H et al.: Development and application of biological age prediction models with physical fitness and physiological components in Korean adults. Gerontology 2012. (PMID 22433233) [PubMed] [DOI] Abstract
BACKGROUND: Several biological age (BA) prediction models have been suggested with a variety of biomarkers. Valid models should be able to measure BA in a relatively short time period and predict subsequent physiological capability. Physiological and physical fitness variables have been shown to be distinctive markers for predicting BA and morbidity. The practical and noninvasive nature of such variables makes them useful as clinical assessment tools in estimating BA for in-depth diagnosis and corresponding intervention. OBJECTIVE: To identify, develop and evaluate biomarkers and BA prediction models and validate their clinical usefulness for the practical diagnosis of functional aging. METHODS: Fourteen variables were measured in 3,112 male and 1,233 female participants aged 30 and older between the years 2004 and 2007. Through a series of parsimonious stepwise elimination processes, two sets of 8 gender-specific variables were selected as candidate biomarkers for 1,604 men and 760 women. Principal component analysis, linear regression analysis and adjustment methods were further applied to obtain two sets of true BA (TBA) prediction models. The TBA models were examined for validity by comparing TBA to the corresponding chronological age (CA) with clinical risk factors. RESULTS: TBA prediction models with r(2) values of 0.638 and 0.672 were developed, each unique to men and women, respectively. The overall mean TBA and CA of the participants were 53.9 and 51.8 years, respectively, with a marginal difference of -2.1 and -1.3 years. The regression slopes or rates of TBA as a function of CA were 1.00 and 1.28 for men and women with r values of 0.799 and 0.820 (p < 0.001), respectively. In comparing TBA to CA rates between healthy and clinical risk groups, both sarcopenic and obese groups showed significant increases in TBA. CONCLUSIONS: The selected biomarkers encompass various complex physiopathological factors related to intrinsic and extrinsic physiological and functional aging. The BA prediction models based on the selected biomarkers could be practical in assessing BA for Korean adults.

[pmid2583248-33] {Jump up to: 33.0} ^33.1 ^33.2 ^33.3 ^33.4 ^33.5 ^33.6 ^33.7 ^33.8 ^33.9 Hochschild R: Improving the precision of biological age determinations. Part 2: Automatic human tests, age norms and variability. Exp Gerontol 1989. (PMID 2583248) [PubMed] [DOI] Abstract
In order to eliminate variability due to test operators, procedures for measuring 12 physiological functions that are candidate biomarkers of aging have been automated. Data was collected from a norm group of 2462 male and female office workers using an instrument which requires no test operators, administers all 12 tests in about 45 min. per subject, computes biological age, prints out results, and stores data on floppy disks for transfer to other computers for analysis. This report a) describes the instrumentation and test procedures, b) presents normal age/sex standards for each of the 12 biomarkers, c) reports the variance of the data for each biomarker by sex, d) lists sources of biomarker variance, e) discusses criteria for biomarker selection and f) examines implications for information loss when biomarker data is combined to calculate biological age. After eliminating chronological age as a variable, the standard deviations of the frequency distributions of predicted age for individual biomarkers were found to vary from .226 to 1.075, a range of more than 4 to 1. Procedures are discussed for improving the ratio of useful-to-useless variance in calculating biological age.

[pmid29380856-34] {Jump up to: 34.0} ^34.1 Zhang WG et al.: Peripheral Blood Leukocyte Telomere Length Is Associated with Age but Not Renal Function: A Cross-Sectional Follow-Up Study. J Nutr Health Aging 2018. (PMID 29380856) [PubMed] [DOI] Abstract
OBJECTIVES: We aimed to evaluate the relationship between baseline renal function and changes in telomere length in Han Chinese. METHODS: The telomere restriction fragment (TRF) length of leukocytes in the peripheral blood was measured in healthy volunteers recruited in 2014. The estimated glomerular filtration rate (eGFR) was calculated based on serum creatinine (Scr) and serum cystatin C (CysC)-eGFRcys and eGFRScr-cys through the Cockcroft-Gault formula (eGFRC-G) or the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI / eGFRCKD-EPI) equation. The correlation between telomere length changes over time and renal function was analyzed. RESULTS: Leukocyte TRF lengths were negatively correlated to age (r = -0.393, p < 0.001) and serum CysC (r = -0.180, p < 0.01), while positively associated with eGFRCKD-EPI, eGFRC-G, eGFRcys, and eGFRScr-cys (r = 0.182, 0.122, 0.290, and 0.254 respectively, p < 0.01). The 3-year change of telomere length was 46 bp/years. When adjusted for age, the associations between telomere length changes and baseline, subsequent TRF lengths, and serum CysC were no longer present. No association was observed between TRF length changes and renal function. CONCLUSION: The rate of telomere length changes was affected by age and baseline telomere length. The telomere length changes might be important markers for aging.

[pmid37124811-35] Li Z et al.: Progress in biological age research. Front Public Health 2023. (PMID 37124811) [PubMed] [DOI] [Full text] Abstract
Biological age (BA) is a common model to evaluate the function of aging individuals as it may provide a more accurate measure of the extent of human aging than chronological age (CA). Biological age is influenced by the used biomarkers and standards in selected aging biomarkers and the statistical method to construct BA. Traditional used BA estimation approaches include multiple linear regression (MLR), principal component analysis (PCA), Klemera and Doubal's method (KDM), and, in recent years, deep learning methods. This review summarizes the markers for each organ/system used to construct biological age and published literature using methods in BA research. Future research needs to explore the new aging markers and the standard in select markers and new methods in building BA models.

[pmid28396265-36] Jylhävä J et al.: Biological Age Predictors. EBioMedicine 2017. (PMID 28396265) [PubMed] [DOI] [Full text] Abstract
The search for reliable indicators of biological age, rather than chronological age, has been ongoing for over three decades, and until recently, largely without success. Advances in the fields of molecular biology have increased the variety of potential candidate biomarkers that may be considered as biological age predictors. In this review, we summarize current state-of-the-art findings considering six potential types of biological age predictors: epigenetic clocks, telomere length, transcriptomic predictors, proteomic predictors, metabolomics-based predictors, and composite biomarker predictors. Promising developments consider multiple combinations of these various types of predictors, which may shed light on the aging process and provide further understanding of what contributes to healthy aging. Thus far, the most promising, new biological age predictor is the epigenetic clock; however its true value as a biomarker of aging requires longitudinal confirmation.

[pmid36499430-37] Erema VV et al.: Biological Age Predictors: The Status Quo and Future Trends. Int J Mol Sci 2022. (PMID 36499430) [PubMed] [DOI] [Full text] Abstract
There is no single universal biomarker yet to estimate overall health status and longevity prospects. Moreover, a consensual approach to the very concept of aging and the means of its assessment are yet to be developed. Markers of aging could facilitate effective health control, more accurate life expectancy estimates, and improved health and quality of life. Clinicians routinely use several indicators that could be biomarkers of aging. Duly validated in a large cohort, models based on a combination of these markers could provide a highly accurate assessment of biological age and the pace of aging. Biological aging is a complex characteristic of chronological age (usually), health-to-age concordance, and medically estimated life expectancy. This study is a review of the most promising techniques that could soon be used in routine clinical practice. Two main selection criteria were applied: a sufficient sample size and reliability based on validation. The selected biological age calculators were grouped according to the type of biomarker used: (1) standard clinical and laboratory markers; (2) molecular markers; and (3) epigenetic markers. The most accurate were the calculators, which factored in a variety of biomarkers. Despite their demonstrated effectiveness, most of them require further improvement and cannot yet be considered for use in standard clinical practice. To illustrate their clinical application, we reviewed their use during the COVID-19 pandemic.

[pmid34595331-38] Lohman T et al.: Predictors of Biological Age: The Implications for Wellness and Aging Research. Gerontol Geriatr Med 2021. (PMID 34595331) [PubMed] [DOI] [Full text] Abstract
As healthspan and lifespan research breakthroughs have become more commonplace, the need for valid, practical markers of biological age is becoming increasingly paramount. The accessibility and affordability of biological age predictors that can reveal information about mortality and morbidity risk, as well as remaining years of life, has profound clinical and research implications. In this review, we examine 5 groups of aging biomarkers capable of providing accurate biological age estimations. The unique capabilities of these biomarkers have far reaching implications for the testing of both pharmaceutical and non-pharmaceutical interventions designed to slow or reverse biological aging. Additionally, the enhanced validity and availability of these tools may have increasingly relevant clinical value. The authors of this review explore those implications, with an emphasis on lifestyle modification research, and provide an overview of the current evidence regarding 5 biological age predictor categories: Telomere length, composite biomarkers, DNA methylation "epigenetic clocks," transcriptional predictors of biological age, and functional age predictors.

[pmid32363395-39] Ashiqur Rahman S et al.: Deep learning for biological age estimation. Brief Bioinform 2021. (PMID 32363395) [PubMed] [DOI] [Full text] Abstract
Modern machine learning techniques (such as deep learning) offer immense opportunities in the field of human biological aging research. Aging is a complex process, experienced by all living organisms. While traditional machine learning and data mining approaches are still popular in aging research, they typically need feature engineering or feature extraction for robust performance. Explicit feature engineering represents a major challenge, as it requires significant domain knowledge. The latest advances in deep learning provide a paradigm shift in eliciting meaningful knowledge from complex data without performing explicit feature engineering. In this article, we review the recent literature on applying deep learning in biological age estimation. We consider the current data modalities that have been used to study aging and the deep learning architectures that have been applied. We identify four broad classes of measures to quantify the performance of algorithms for biological age estimation and based on these evaluate the current approaches. The paper concludes with a brief discussion on possible future directions in biological aging research using deep learning. This study has significant potentials for improving our understanding of the health status of individuals, for instance, based on their physical activities, blood samples and body shapes. Thus, the results of the study could have implications in different health care settings, from palliative care to public health.

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]