Development and Validation of 2 Composite Aging Measures Using Routine Clinical Biomarkers in the Chinese Population: Analyses From 2 Prospective Cohort Studies

Alcohol consumption and accelerated biological ageing in middle-aged and older people: A longitudinal study from two cohorts

BACKGROUND AND AIMS

The relationship between alcohol consumption and age-related diseases is inconsistent. Biological age (BA) serves as both a precursor and a predictor of age-related diseases; however, longitudinal associations between alcohol consumption and BA in middle-aged and older people remain unclear. We measured whether there was a longitudinal association between drinking frequency and pure alcohol intake with BA among middle-aged and older people.

DESIGN AND SETTING AND PARTICIPANTS

This study involved two prospective cohort studies, set in Southwestern China and the United Kingdom. A total of 8046 participants from the China Multi-Ethnic Cohort study (CMEC) and 5412 participants from the UK Biobank (UKB), aged 30-79 years, took part, with complete data from two waves of clinical biomarkers.

MEASUREMENTS

BA was calculated by the Klemera Doubal's method. Accelerated BA equalled BA minus chronological age. Drinking frequency and pure alcohol intake were obtained through self-reported questionnaires. Drinking frequency in the past year was classified as current non-drinking, occasional (monthly drinking) and regular (weekly drinking).

FINDINGS

Compared with consistent current non-drinkers, more frequent drinkers [CMEC: β = 0.46, 95% confidence interval (CI) = 0.13-0.80; UKB: β = 0.65, 95% CI = 0.01-1.29)], less frequent drinkers (CMEC: β = 0.62, 95% CI = 0.37-0.87; UKB: β = 0.54, 95% CI = -0.01-1.09), consistent occasional drinkers (CMEC: β = 0.51, 95% CI = 0.23-0.79; UKB: β = 0.63, 95% CI = 0.13-1.13) and consistent regular drinkers (CMEC: β = 0.56, 95% CI = 0.17-0.95; UKB: β = 0.46, 95% CI = 0.00-0.91) exhibited increased accelerated BA. A non-linear relationship between pure alcohol intake and accelerated BA was observed among consistent regular drinkers.

CONCLUSIONS

In middle-aged and older people, any change in drinking frequency and any amount of pure alcohol intake seem to be positively associated with acceleration of biological ageing, compared with maintaining abstinence.

Social Determinants of Health and Biological Age among Diverse U.S. Adults, NHANES 2011-2018

We examined the sex-specific association between education and income with biological age (BA) and by race/ethnicity. The Klemera-Doubal method was used to calculate BA among 6,213 females and 5,938 males aged 30-75 years who were Hispanic, non-Hispanic (NH) White, NH Black (NHB), or NH Asian (NHA). Compared with a college education, less than a high school education was associated with greater BA by 3.06 years (95% CI: 1.58, 4.54) among females only; associations were strongest among NHB, Hispanic, and NHA females. Compared with an annual income of ≥$75,000, an income <$25,000 was associated with greater BA by 4.95 years (95% CI: 3.42, 6.48) among males and 2.76 years among females (95% CI: 1.51, 4.01); associations were strongest among NHW and NHA adults, and Hispanic males. Targeting upstream sources of structural disadvantage among racial/ethnic minority groups, in conjunction with improvements in income and education, may promote healthy aging in these populations.

Accelerated biological aging mediated associations of ammonium, sulfate in fine particulate matter with liver cirrhosis

BACKGROUND

Although both air pollution and aging are related to the development of liver cirrhosis, the role of biological aging in association of the mixture of fine particulate matter (PM2.5) and its constituents with liver cirrhosis was unknown.

METHODS

This case-control retrospective study included 100 liver cirrhosis patients and 100 control subjects matched by age and sex. The concentrations of PM2.5 and its constituents were estimated for patients using machine-learning methods. The clinical biomarkers were used to calculate biological age using the Klemera-Doubalmethod (KDM) algorithms. Individual associations of PM2.5 and its constituents or biological age with liver cirrhosis were analyzed by generalized linear models. WQS and BKMR were applied to analyze association of mixture of PM2.5 and its constituents with liver cirrhosis. The mediation effect of biological age on associations of PM2.5 and its constituents with liver cirrhosis was further explored.

RESULTS

we found that each 1-unit increment in NH4+, NO3-, SO42- and biological age were related to 3.618-fold (95%CI: 1.896, 6.904), 1.880-fold (95%CI: 1.319, 2.680), 2.955-fold (95%CI: 1.656, 5.272) and 1.244-fold (95%CI: 1.093, 1.414) increased liver cirrhosis. Both WQS and BKMR models showed that the mixture of PM2.5 and its constituents was related to increased liver cirrhosis. Furthermore, the mediated proportion of biological age on associations of NH4+ and SO42- with liver cirrhosis were 14.7 % and 14.6 %, respectively.

CONCLUSIONS

Biological aging may partly explain the exposure to PM2.5 and its constituents in association with increased risk for liver cirrhosis, implying that delaying the aging process may be a key step for preventing PM2.5-related liver cirrhosis risk.

A biological age model based on physical examination data to predict mortality in a Chinese population

Biological age could be reflective of an individual's health status and aging degree. Limited estimations of biological aging based on physical examination data in the Chinese population have been developed to quantify the rate of aging. We developed and validated a novel aging measure (Balanced-AGE) based on readily available physical health examination data. In this study, a repeated sub-sampling approach was applied to address the data imbalance issue, and this approach significantly improved the performance of biological age (Balanced-AGE) in predicting all-cause mortality with a 10-year time-dependent AUC of 0.908 for all-cause mortality. This mortality prediction tool was found to be effective across different subgroups by age, sex, smoking, and alcohol consumption status. Additionally, this study revealed that individuals who were underweight, smokers, or drinkers had a higher extent of age acceleration. The Balanced-AGE may serve as an effective and generally applicable tool for health assessment and management among the elderly population.

The mediating role of accelerated biological aging in the association between blood metals and cognitive function

Aging is a key risk factor in cognitive diseases. Recently, metal exposures were found associated with both biological aging and cognitive function. Here, we aim to evaluate the associations of blood metals with cognitive function and the mediated effect of biological aging. Fourteen metals were detected and biological age was calculated through Klemera and Doubal method among 514 adults in Beijing, China. The generalized linear models indicated that the copper (Cu), molybdenum (Mo), and strontium (Sr) were positively associated with biological aging [βCu (95% CI): 12.76 (9.26, 16.27); βMo (95% CI): 1.50 (0.15, 2.85)], and βSr (95% CI): 1.86 (0.68, 3.03)], while vanadium (V) was inversely related to biological aging [βV (95% CI): -0.76 (-1.48, -0.05)]. Subsequently, Cu, lead (Pb), selenium (Se), and biological aging were associated with cognitive function and further mediation analyses confirmed that biological aging partially mediated (33.98%, P = 0.019) the association of Cu and cognitive function. Additionally, we constructed a lifestyle index that implied the modifiable healthy lifestyle could slow aging to attenuate the detrimental effect of metals on cognition. Our findings provide insights into the potential pathways linking multiple metals exposure to aging and cognition and underscore the importance of adopting healthy lifestyles.

Physiological Dysregulation Proceeds and Predicts Health Outcomes Similarly in Chinese and Western Populations

BACKGROUND

A decade ago, we proposed an index of physiological dysregulation based on Mahalanobis distance (DM) that measures how far from the norm an individual biomarker profile is. While extensive validation has been performed, focus was mostly on Western populations with little comparison to developing countries, particularly at a physiological system level. The degree to which the approach would work in other sociocultural contexts and the similarity of dysregulation signatures across diverse populations are still open questions.

METHODS

Using 2 data sets from China and 3 from Western countries (United States, United Kingdom, and Italy), we calculated DM globally and per physiological system. We assessed pairwise correlations among systems, difference with age, prediction of mortality and age-related diseases, and sensitivity to interchanging data sets with one another as the reference in DM calculation.

RESULTS

Overall, results were comparable across all data sets. Different physiological systems showed distinct dysregulation processes. Association with age was moderate and often nonlinear, similarly for all populations. Mahalanobis distance predicted most health outcomes, although differently by physiological system. Using a Chinese population as the reference when calculating DM for Western populations, or vice versa, led to similar associations with health outcomes, with a few exceptions.

CONCLUSIONS

While small differences were noticeable, they did not systematically emerge between Chinese and Western populations, but rather diffusively across all data sets. These findings suggest that DM presents similar properties, notwithstanding sociocultural backgrounds, and that it is equally effective in capturing the loss of homeostasis that occurs during aging in diverse industrial human populations.

Daytime napping, biological aging and cognitive function among middle-aged and older Chinese: insights from the China health and retirement longitudinal study

Objective

The complicated association of daytime napping, biological aging and cognitive function remains inconclusive. We aimed to evaluate the cross-sectional and longitudinal associations of daytime napping and two aging measures with cognition and to examine whether napping affects cognition through a more advanced state of aging.

Methods

Data was collected from the China Health and Retirement Longitudinal Study. Napping was self-reported. We calculated two published biological aging measures: Klemera and Doubal biological age (KDM-BA) and physiological dysregulation (PD), which derived information from clinical biomarkers. Cognitive z-scores were calculated at each wave. Linear mixed models were used to explore the longitudinal association between napping, two aging measures, and cognitive decline. Mediation analyses were performed to assess the mediating effects of biological age acceleration on the association between napping and cognition.

Results

Participants aged over 45 years were included in the analyses. Non-nappers had greater KDM-BA and PD [LS means (LSM) = 0.255, p = 0.007; LSM = 0.085, p = 0.011] and faster cognitive decline (LSM = -0.061, p = 0.005)compared to moderate nappers (30-90 min/nap). KDM-BA (β = -0.007, p = 0.018) and PD (β = -0.034, p < 0.001) showed a negative association with overall cognitive z scores. KDM-BA and PD partially mediated the effect of napping on cognition.

Conclusion

In middle-aged and older Chinese, compared to moderate nappers, non-nappers seem to experience a more advanced state of aging and increased rates of cognitive decline. The aging status possibly mediates the association between napping and cognition. Moderate napping shows promise in promoting healthy aging and reducing the burden of cognitive decline in Chinese middle-aged and older adults.

Effects of accelerated biological age on depressive symptoms in a causal reasoning framework

BACKGROUND

Depression in middle-aged and elderly individuals is multifaceted and heterogeneous, linked to biological age (BA) based on aging-related biomarkers. However, due to confounding with chronological age and the absence of subgroup analysis and causal reasoning, the association between BA and depressive symptoms (DS) might be unstable and requires further investigation.

METHODS

We utilized data from the China Health and Retirement Longitudinal Study (N = 9478) to perform association analysis, causal inference, and subgroup analysis. BA acceleration (BAA) was derived using machine learning and adjusted for chronological age. A generalized linear mixed-effects model (GLMM) tree algorithm was employed to identify subgroups. The causal reasoning frame included propensity score matching and fast large-scale almost matching exactly.

RESULTS

In the longitudinal analysis, BAA exhibited a consistent and significant positive association with DS, even after controlling for demographic characteristics, lifestyle factors, health status, and physical functions. This association remained unchanged within the causal framework. GLMM tree analysis identified three partitioning variables (sex, satisfaction, and BMI) and five subgroups. Further subgroup analysis revealed that BAA exerted the strongest effect on DS among women with less satisfying lives.

LIMITATIONS

Depressive symptoms were evaluated through scale measurements rather than clinical diagnosis. The sample was derived from the general population, not the clinically depressed population.

CONCLUSIONS

This study provided the first longitudinal evidence that biological age acceleration increases depressive symptoms under causal reasoning and subgroup analysis, particularly among less satisfied women. And the association between BAA and DS was independent of known risk factors.

Biomarkers selection and mathematical modeling in biological age estimation

Biological age (BA) is important for clinical monitoring and preventing aging-related disorders and disabilities. Clinical and/or cellular biomarkers are measured and integrated in years using mathematical models to display an individual's BA. To date, there is not yet a single or set of biomarker(s) and technique(s) that is validated as providing the BA that reflects the best real aging status of individuals. Herein, a comprehensive overview of aging biomarkers is provided and the potential of genetic variations as proxy indicators of the aging state is highlighted. A comprehensive overview of BA estimation methods is also provided as well as a discussion of their performances, advantages, limitations, and potential approaches to overcome these limitations.

Contribution of CKD to mortality in middle-aged and elderly people with diabetes: the China Health and Retirement Longitudinal Study : CKD was a chronic stressor for diabetics

BACKGROUND

The contribution of chronic kidney disease (CKD) to mortality in diabetic patients is unclear. This study aimed to explore the association between diabetics with CKD and mortality in middle-aged and elderly people of different ages.

METHODS

Data were obtained from the China Health and Retirement Longitudinal Study, including 1,715 diabetic individuals, 13.1% of whom also had CKD. Diabetes and CKD were assessed by combining the physical measurements and self-reports. We fitted Cox proportional hazards regression models to examine the effect of diabetics with CKD on mortality in middle-aged and elderly people. The risk factors for death were further predicted based on age stratification.

RESULTS

The mortality rate of diabetic patients with CKD (29.3%) was increased as compared to that of diabetic patients without CKD (12.4%). Diabetics with CKD were at a higher risk of all-cause mortality than those without CKD, with a hazard ratio of 1.921 (95% CI: 1.438, 2.566). Additionally, for participants 45 to 67 years of age, the hazard ratio was 2.530 (95% CI: 1.624, 3.943).

CONCLUSIONS

Our findings suggested that, for diabetics, CKD was a chronic stressor that led to death in middle-aged and elderly people, especially among participants aged 45 to 67 years.

Association of PM2.5 Components with Acceleration of Aging: Moderating Role of Sex Hormones

Fine particulate matter (PM_2.5) has been linked to aging risk, and a lack of knowledge about the relationships between PM_2.5 components and aging risk impeded the development of healthy aging. Participants were recruited through a multicenter cross-sectional study in the Beijing-Tianjin-Hebei region in China. Middle-age and older males and menopausal women completed the collection of basic information, blood samples, and clinical examinations. The biological age was estimated by Klemera-Doubal method (KDM) algorithms based on clinical biomarkers. Multiple linear regression models were applied to quantify the associations and interactions while controlling for confounders, and a restricted cubic spline function estimated the corresponding dose-response curves of the relationships. Overall, KDM-biological age acceleration was associated with PM_2.5 component exposure over the preceding year in both males and females, with calcium [females: 0.795 (95% CI: 0.451, 1.138); males: 0.712 (95% CI: 0.389, 1.034)], arsenic [females: 0.770 (95% CI: 0.641, 0.899); males: 0.661 (95% CI: 0.532, 0.791)], and copper [females: 0.401 (95% CI: 0.158, 0.644); males: 0.379 (95% CI: 0.122, 0.636)] having greater estimates of the effect than total PM_2.5 mass. Additionally, we observed that the associations of specific PM_2.5 components with aging were lower in the higher sex hormone scenario. Maintaining high levels of sex hormones may be a crucial barrier against PM_2.5 component-related aging in the middle and older age groups.

Associations between multiple metals exposure and biological aging: Evidence from the Dongfeng-Tongji cohort

Aging is related to a progressive decline in physiological functions and is affected by environmental factors. Metal exposures are linked with many health effects, but have poorly understood associations with aging. In this study, a total of 33,916 participants from the Dongfeng-Tongji cohort were included to establish biological age (BA) predictors by using recent advanced algorithms, Klemera and Doubal method (KDM) and Mahalanobis distance. Two biological aging indexes (BAIs), recorded as KDM-accel [the residual from regressing KDM-BA on chronological age] and physiological dysregulation (PD), were separately defined and tested on their associations with mortality by using Cox proportional hazard models. Among 3320 subjects with laboratory determinations of 23 metals in plasma, the individual and overall associations between these metals and BAIs were evaluated by using multiple-linear regression and weighted quantile sum (WQS) models. Both BAIs were prospectively associated with all-cause mortality among the whole participants [KDM-accel: HR(95%CI) = 1.23(1.18, 1.29); PD: HR(95%CI) = 1.37(1.31, 1.42)]. Each 1-unit increment in ln-transformed strontium and molybdenum were cross-sectionally associated with a separate 0.71- and 0.34-year increase in KDM-accel, and each 1 % increment in copper, rubidium, strontium, cobalt was cross-sectionally associated with a separate 0.10 %, 0.10 %, 0.09 %, 0.02 % increase in PD (all FDR < 0.05). The WQS models observed mixture effects of multi-metals on aging, with a 0.20-year increase in KDM-accel and a 0.04 % increase in PD for each quartile increase in ln-transformed concentrations of all metals [KDM-accel: β(95%CI) = 0.20(0.08, 0.32); PD: β(95%CI) = 0.04(0.02, 0.06)]. Our findings revealed that plasma strontium, molybdenum, copper, rubidium and cobalt were associated with accelerated aging. Multi-metals exposure showed mixture effects on the aging process, which highlights potential preventative interventions.

Progress in biological age research

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.

Familial aggregation of the aging process: biological age measured in young adult offspring as a predictor of parental mortality

Measures of biological age (BA) integrate information across organ systems to quantify "biological aging," i.e., inter-individual differences in aging-related health decline. While longevity and lifespan aggregate in families, reflecting transmission of genes and environments across generations, little is known about intergenerational continuity of biological aging or the extent to which this continuity may be modified by environmental factors. Using data from the Jerusalem Perinatal Study (JPS), we tested if differences in offspring BA were related to mortality in their parents. We measured BA using biomarker data collected from 1473 offspring during clinical exams in 2007-2009, at age 32 ± 1.1. Parental mortality was obtained from population registry data for the years 2004-2016. We fitted parametric survival models to investigate the associations between offspring BA and parental all-cause and cause-specific mortality. We explored potential differences in these relationships by socioeconomic position (SEP) and offspring sex. Participants' BAs widely varied (SD = 6.95). Among those measured to be biologically older, parents had increased all-cause mortality (HR = 1.10, 95% CI: 1.08, 1.13), diabetes mortality (HR = 1.19, 95% CI: 1.08, 1.30), and cancer mortality (HR = 1.07, 95% CI: 1.02, 1.13). The association with all-cause mortality was stronger for families with low compared with high SEP (P_interaction = 0.04) and for daughters as compared to sons (P_interaction < 0.001). Using a clinical-biomarker-based BA estimate, observable by young adulthood prior to the onset of aging-related diseases, we demonstrate intergenerational continuity of the aging process. Furthermore, variation in this familial aggregation according to household socioeconomic position (SEP) at offspring birth and between families of sons and daughters proposes that the environment alters individuals' aging trajectory set by their parents.

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

Background

Biological age (BA) has been recognized as a more accurate indicator of aging than chronological age (CA). However, the current limitations include: insufficient attention to the incompleteness of medical data for constructing BA; Lack of machine learning-based BA (ML-BA) on the Chinese population; Neglect of the influence of model overfitting degree on the stability of the association results.

Methods and results

Based on the medical examination data of the Chinese population (45–90 years), we first evaluated the most suitable missing interpolation method, then constructed 14 ML-BAs based on biomarkers, and finally explored the associations between ML-BAs and health statuses (healthy risk indicators and disease). We found that round-robin linear regression interpolation performed best, while AutoEncoder showed the highest interpolation stability. We further illustrated the potential overfitting problem in ML-BAs, which affected the stability of ML-Bas’ associations with health statuses. We then proposed a composite ML-BA based on the Stacking method with a simple meta-model (STK-BA), which overcame the overfitting problem, and associated more strongly with CA (r = 0.66, P < 0.001), healthy risk indicators, disease counts, and six types of disease.

Conclusion

We provided an improved aging measurement method for middle-aged and elderly groups in China, which can more stably capture aging characteristics other than CA, supporting the emerging application potential of machine learning in aging research.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12859-022-04966-7.

Contribution of life course circumstances to the acceleration of phenotypic and functional aging: A retrospective study

Background

Accelerated aging leads to increasing burdens of chronic diseases in late life, posing a huge challenge to the society. With two well-developed aging measures (i.e., physiological dysregulation [PD] and frailty index [FI]), this study aimed to evaluate the relative contributions of life course circumstances (e.g., childhood and adulthood socioeconomic status) to variance in aging.

Methods

We assembled data for 6224 middle-aged and older adults in China from the 2014 life course survey (June to December 2014), the 2015 biomarker collection (July 2015 to January 2016), and the 2015 main survey (July 2015 to January 2016) of the China Health and Retirement Longitudinal Study. Two aging measures (PD and FI) were calculated, with a higher value indicating more accelerated aging. Life course circumstances included childhood (i.e., socioeconomic status, war, health, trauma, relationship, and parents' health) and adulthood circumstances (i.e., socioeconomic status, adversity, and social support), demographics, and behaviours. The Shapley value decomposition, hierarchical clustering, and general linear regression models were performed.

Findings

The Shapley value decomposition revealed that all included life course circumstances accounted for about 6·3% and 29·7% of variance in PD and FI, respectively. We identified six subpopulations who shared similar patterns in terms of childhood and adulthood circumstances. The most disadvantaged subpopulation (i.e., subpopulation 6 [more childhood trauma and adulthood adversity]) consistently exhibited accelerated aging indicated by the two aging measures. Relative to the most advantaged subpopulation (i.e., subpopulation 1 [less childhood trauma and adulthood adversity]), PD and FI in the most disadvantaged subpopulation were increased by an average of 0·14 (i.e., coefficient, by one-standard deviation, 95% confidence interval [CI] 0·06-0·21; p < 0·0001) and 0·10 (by one-point, 95% CI 0·09-0·11; p < 0·0001), respectively.

Interpretation

Our findings highlight the different contributions of life course circumstances to phenotypic and functional aging. Special attention should be given to promoting health for the disadvantaged subpopulation and narrowing their health gap with advantaged counterparts.

Funding

National Natural Science Foundation of China, Milstein Medical Asian American Partnership Foundation, Natural Science Foundation of Zhejiang Province, Fundamental Research Funds for the Central Universities, National Institute on Aging, National Centre for Advancing Translational Sciences, and Yale Alzheimer's Disease Research Centre.

All-Subset Analysis Improves the Predictive Accuracy of Biological Age for All-Cause Mortality in Chinese and U.S. Populations

BACKGROUND

Klemera-Doubal's method (KDM) is an advanced and widely applied algorithm for estimating biological age (BA), but it has no uniform paradigm for biomarker processing. This article proposed all subsets of biomarkers for estimating BAs and assessed their association with mortality to determine the most predictive subset and BA.

METHODS

Clinical biomarkers, including those from physical examinations and blood assays, were assessed in the China Health and Nutrition Survey (CHNS) 2009 wave. Those correlated with chronological age (CA) were combined to produce complete subsets, and BA was estimated by KDM from each subset of biomarkers. A Cox proportional hazards regression model was used to examine and compare each BA's effect size and predictive capacity for all-cause mortality. Validation analysis was performed in the Chinese Longitudinal Healthy Longevity Survey (CLHLS) and National Health and Nutrition Examination Survey (NHANES). KD-BA and Levine's BA were compared in all cohorts.

RESULTS

A total of 130 918 panels of BAs were estimated from complete subsets comprising 3-17 biomarkers, whose Pearson coefficients with CA varied from 0.39 to 1. The most predictive subset consisted of 5 biomarkers, whose estimated KD-BA had the most predictive accuracy for all-cause mortality. Compared with Levine's BA, the accuracy of the best-fitting KD-BA in predicting death varied among specific populations.

CONCLUSION

All-subset analysis could effectively reduce the number of redundant biomarkers and significantly improve the accuracy of KD-BA in predicting all-cause mortality.

An objective metric of individual health and aging for population surveys

BACKGROUND

We have previously developed and validated a biomarker-based metric of overall health status using Mahalanobis distance (DM) to measure how far from the norm of a reference population (RP) an individual's biomarker profile is. DM is not particularly sensitive to the choice of biomarkers; however, this makes comparison across studies difficult. Here we aimed to identify and validate a standard, optimized version of DM that would be highly stable across populations, while using fewer and more commonly measured biomarkers.

METHODS

Using three datasets (the Baltimore Longitudinal Study of Aging, Invecchiare in Chianti and the National Health and Nutrition Examination Survey), we selected the most stable sets of biomarkers in all three populations, notably when interchanging RPs across populations. We performed regression models, using a fourth dataset (the Women's Health and Aging Study), to compare the new DM sets to other well-known metrics [allostatic load (AL) and self-assessed health (SAH)] in their association with diverse health outcomes: mortality, frailty, cardiovascular disease (CVD), diabetes, and comorbidity number.

RESULTS

A nine- (DM9) and a seventeen-biomarker set (DM17) were identified as highly stable regardless of the chosen RP (e.g.: mean correlation among versions generated by interchanging RPs across dataset of r = 0.94 for both DM9 and DM17). In general, DM17 and DM9 were both competitive compared with AL and SAH in predicting aging correlates, with some exceptions for DM9. For example, DM9, DM17, AL, and SAH all predicted mortality to a similar extent (ranges of hazard ratios of 1.15-1.30, 1.21-1.36, 1.17-1.38, and 1.17-1.49, respectively). On the other hand, DM9 predicted CVD less well than DM17 (ranges of odds ratios of 0.97-1.08, 1.07-1.85, respectively).

CONCLUSIONS

The metrics we propose here are easy to measure with data that are already available in a wide array of panel, cohort, and clinical studies. The standardized versions here lose a small amount of predictive power compared to more complete versions, but are nonetheless competitive with existing metrics of overall health. DM17 performs slightly better than DM9 and should be preferred in most cases, but DM9 may still be used when a more limited number of biomarkers is available.

A Machine Learning-Based Aging Measure Among Middle-Aged and Older Chinese Adults: The China Health and Retirement Longitudinal Study

Objective: Biological age (BA) has been accepted as a more accurate proxy of aging than chronological age (CA). This study aimed to use machine learning (ML) algorithms to estimate BA in the Chinese population. Materials and methods: We used data from 9,771 middle-aged and older Chinese adults (≥45 years) in the 2011/2012 wave of the China Health and Retirement Longitudinal Study and followed until 2018. We used several ML algorithms (e.g., Gradient Boosting Regressor, Random Forest, CatBoost Regressor, and Support Vector Machine) to develop new measures of biological aging (ML-BAs) based on physiological biomarkers. R-squared value and mean absolute error (MAE) were used to determine the optimal performance of these ML-BAs. We used logistic regression models to examine the associations of the best ML-BA and a conventional aging measure-Klemera and Doubal method-BA (KDM-BA) we previously developed-with physical disability and mortality, respectively. Results: The Gradient Boosting Regression model performed the best, resulting in an ML-BA with an R-squared value of 0.270 and an MAE of 6.519. This ML-BA was significantly associated with disability in basic activities of daily living, instrumental activities of daily living, lower extremity mobility, and upper extremity mobility, and mortality, with odds ratios ranging from 1 to 7% (per 1-year increment in ML-BA, all P < 0.001), independent of CA. These associations were generally comparable to that of KDM-BA. Conclusion: This study provides a valid ML-based measure of biological aging for middle-aged and older Chinese adults. These findings support the application of ML in geroscience research and may help facilitate preventive and geroprotector intervention studies.

Association of lifestyle with mortality and the mediating role of aging among older adults in China

OBJECTIVES

1) examine the association between lifestyle and mortality; 2) examine the association between two aging measures and mortality; 3) evaluate the mediating effect of the two aging measures on the association between lifestyle and mortality among older Chinese adults.

METHODS

We used data from 2039 older adults (≥ 65 years) from the 2011/2012 biomarker substudy of the Chinese Longitudinal Healthy Longevity Survey (CLHLS). We created a healthy lifestyle index based on 5 factors (exercise, smoking, drinking, diet, and BMI, range: 0-5). We calculated two aging measures, the Klemera and Doubal method-biological age (KDM-BA) and physiological dysregulation (PD), based on 10 blood-based biomarkers using algorithms developed previously. A Cox proportional hazards model, general linear regression model, and formal mediation analysis were performed.

RESULTS

After adjustment for age and sex, compared to participants without any healthy lifestyle factors, those with 5 healthy lifestyle factors had an 85% lower risk of mortality (hazard ratio [HR] = 0.15, 95% confidence interval [CI]: 0.04, 0.60). PD, but not KDM-BA, was significantly associated with mortality (HR = 1.69, 95% CI: 1.25, 2.29). The healthy lifestyle index was negatively associated with PD (β = -0.021, P = 0.012). PD mediated 9% (95% CI: 1%, 52%, P = 0.043) of the total effect of the healthy lifestyle index on mortality.

CONCLUSIONS

In the older Chinese population, healthy lifestyle reduces mortality risk and aging partially mediates this association. The findings highlight the importance of adherence to a healthy lifestyle for promoting phenotypic aging even in late life.

A machine learning-based biological aging prediction and its associations with healthy lifestyles: the Dongfeng-Tongji cohort

This study aims to establish a biological age (BA) predictor and to investigate the roles of lifestyles on biological aging. The 14,848 participants with the available information of multisystem measurements from the Dongfeng-Tongji cohort were used to estimate BA. We developed a composite BA predictor showing a high correlation with chronological age (CA) (r = 0.82) by using an extreme gradient boosting (XGBoost) algorithm. The average frequency hearing threshold, forced expiratory volume in 1 second (FEV₁ ), gender, systolic blood pressure, and homocysteine ranked as the top five important features for the BA predictor. Two aging indexes, recorded as the AgingAccel (the residual from regressing predicted age on CA) and aging rate (the ratio of predicted age to CA), showed positive associations with the risks of all-cause (HR (95% CI) = 1.12 (1.10-1.14) and 1.08 (1.07-1.10), respectively) and cause-specific (HRs ranged from 1.06 to ∼1.15) mortality. Each 1-point increase in healthy lifestyle score (including normal body mass index, never smoking, moderate alcohol drinking, physically active, and sleep 7-9 h/night) was associated with a 0.21-year decrease in the AgingAccel (95% CI: -0.27 to -0.15) and a 0.4% decrease in the aging rate (95% CI: -0.5% to -0.3%). This study developed a machine learning-based BA predictor in a prospective Chinese cohort. Adherence to healthy lifestyles showed associations with delayed biological aging, which highlights potential preventive interventions.

The association of proBNPage with manifestations of age-related cardiovascular, physical, and psychological impairment in community-dwelling older adults

NT-proB-type natriuretic peptide (NT-proBNP) serum concentration can be transformed by simple formulas into proBNPage, a surrogate of biological age strongly associated with chronological age, all-cause mortality, and disease count. This cross-sectional study aimed to assess whether proBNPage is also associated with other manifestations of the aging process in comparison with other variables. The study included 1117 noninstitutionalized older adults (73.1 ± 5.6 years, 537 men). Baseline measurements of serum NT-proBNP, erythrocyte sedimentation rate, hemoglobin, lymphocytes, and creatinine, which have previously been shown to be highly associated with both age and all-cause mortality, were performed. These variables were compared between subjects with and without manifestations of cardiovascular impairment (myocardial infarction (MI), stroke, peripheral artery disease (PAD), arterial revascularizations (AR)), physical impairment (long step test duration (LSTD), walking problems, falls, deficit in one or more activities of daily living), and psychological impairment (poor self-rating of health (PSRH), anxiety/depression, Mini Mental State Examination (MMSE) score < 24). ProBNPage (years) was independently associated (OR, 95% CI) with MI (1.08, 1.07-1.10), stroke (1.02, 1.00-1.05), PAD (1.04, 1.01-1.06), AR (1.06, 1.04-1.08), LSTD (1.03, 1.02-1.04), walking problems (1.02, 1.01-1.03), and PSRH (1.02, 1.01-1.02). For 5 of these 7 associations, the relationship was stronger than that of chronological age. In addition, proBNPage was univariately associated with MMSE score < 24, anxiety/depression, and falls. None of the other variables provided comparable performances. Thus, in addition to the known associations with mortality and disease count, proBNPage is also associated with cardiovascular manifestations as well as noncardiovascular manifestations of the aging process.

Robust Physiological Metrics From Sparsely Sampled Networks

Physiological and biochemical networks are highly complex, involving thousands of nodes as well as a hierarchical structure. True network structure is also rarely known. This presents major challenges for applying classical network theory to these networks. However, complex systems generally share the property of having a diffuse or distributed signal. Accordingly, we should predict that system state can be robustly estimated with sparse sampling, and with limited knowledge of true network structure. In this review, we summarize recent findings from several methodologies to estimate system state via a limited sample of biomarkers, notably Mahalanobis distance, principal components analysis, and cluster analysis. While statistically simple, these methods allow novel characterizations of system state when applied judiciously. Broadly, system state can often be estimated even from random samples of biomarkers. Furthermore, appropriate methods can detect emergent underlying physiological structure from this sparse data. We propose that approaches such as these are a powerful tool to understand physiology, and could lead to a new understanding and mapping of the functional implications of biological variation.