Social determinants of health, accelerated biological aging, and long-term health outcomes

Jiang Li; Jie Li; Xiaoqin Xu; Wenqi Shen; Ying Sun; Yanqi Fu; Xiao Tan; Ningjian Wang; Lu Qi; Bin Wang; Yingli Lu

doi:10.1038/s41467-025-67622-7

. 2025 Dec 24;17:900. doi: 10.1038/s41467-025-67622-7

Social determinants of health, accelerated biological aging, and long-term health outcomes

Jiang Li ^1,^#, Jie Li ^1,^#, Xiaoqin Xu ¹, Wenqi Shen ¹, Ying Sun ¹, Yanqi Fu ¹, Xiao Tan ^2,³, Ningjian Wang ^1,^4,^✉, Lu Qi ^5,^6,^✉, Bin Wang ^1,^7,^✉, Yingli Lu ^1,^✉

PMCID: PMC12830646 PMID: 41444248

Abstract

Social determinants of health (SDHs) are the primary drivers of health inequalities, but whether biological aging plays a role in linking SDHs to health outcomes remains unclear. Here we utilize detailed information on social determinants across five domains, clinical parameters and electronic health records from the UK Biobank and US NHANES to examine the associations between combined SDHs, accelerated biological aging, and health outcomes. Compared with participants in the favourable SDH group, participants in the unfavourable SDH group had increased KDM-BA and phenotypic age acceleration. Moreover, unfavourable SDHs were associated with elevated risks of mortality and incident diseases. Accelerated biological aging significantly mediated the association between SDHs and all-cause and cause-specific mortality (UK Biobank: mediation proportion 13.46%-25.21%; US NHANES: 7.62%-22.16%). Also, accelerated biological aging served as a mediator between SDHs and incident diseases in the UK Biobank, with the mediation proportions ranging from 6.20% to 30.48%. The estimates were likely specific to the UK Biobank cohort considering its healthy volunteer bias and limited socioeconomic diversity. Overall, our study reveals that the biological aging discrepancy partially explains the associations of combined SDHs with mortality and chronic diseases. Assessing and delaying aging acceleration may be an effective way to narrow the health disparities caused by SDHs.

Subject terms: Health policy, Diseases

Here the authors report an analysis of two large nationwide cohort studies, the UK biobank and NHANES, suggesting that accelerated biological aging mediate a proportion of the association between social determinants of health and all-cause and cause-specific mortality.

Introduction

Aging is a complex process involving a progressive loss of integrity and resilience capacity of cells, tissues, and organs over time, ultimately resulting in impaired function and increased morbidity and death¹. It is estimated that aging-related diseases accounted for 51.3% of all global disease burden in 2017². Several biological age measures have been proposed to capture the aging process^3,4. Among these measures, the algorithms derived from composite clinical parameters like Klemera–Doubal biological age (KDM-BA) and phenotypic age have shown excellent capability of predicting morbidity and mortality risks across diverse populations and proven to be cost-effective^5–7. Given the expanding aging population worldwide, broadening knowledge of the aging process and its contributing factors is increasingly important for informing public health strategies.

Social determinants of health (SDHs) are the conditions where people are born, grow, live, work, and age, which have been associated with differences in morbidity and mortality⁸. Although the globe has witnessed socioeconomic progress over recent decades, the high rate of wealth inequalities exists and even increases; particularly, socioeconomic inequity in survival is rising in the UK and the US^9,10. Apart from their profound influence on health outcomes, including death, cardiometabolic disease, chronic respiratory disease, and chronic kidney disease^11–14, certain individual disadvantaged SDHs (i.e., low income and chronic psychosocial burden) were associated with accelerated biological aging^15–17. Nevertheless, different disadvantaged SDHs tend to cluster in individuals, and the pathways linking them to health are interconnected¹⁸. In addition, disregarding the synergistic effects of combined SDHs will reduce the power to efficiently identify the groups who are socially vulnerable for targeted interventions. Therefore, it is of great importance to perform a comprehensive assessment of SDHs and establish the association between overall SDHs and accelerated biological aging. Previous studies have suggested that biological aging is an indicator of morbidity and mortality in later life^19,20, but whether SDHs could increase the susceptibility to adverse health outcomes through accelerated biological aging remains unclear.

In this work, we use data from the UK Biobank and US National Health and Nutrition Examination Survey (NHANES) to investigate the associations between combined SDHs, accelerated biological aging, and a range of health outcomes (Fig. 1). The results reveal that accelerated biological aging significantly mediate the associations of unfavorable SDHs with increased risks of mortality and chronic diseases, which are consistent across two cohorts from varied social contexts. These findings support biological aging as a promising target when developing interventions aimed at narrowing the SDHs-related health disparities.

Fig. 1 — The icon representing social determinants of health was adapted from the Healthy People 2030³⁴. KDM-BA Klemera–Doubal biological age, NHANES National Health and Nutrition Examination Survey, SDH social determinant of health. Figure created in BioRender. Li, J. (2025) https://BioRender.com/f13r657.

Results

Population characteristics

Among 266,029 participants from UK Biobank (mean [standard deviation, SD] age 56.1 [8.1] years, 52.4% females), 55,610 (20.9%) were categorized into the favorable SDH group, 135,445 (50.9%) into the medium SDH group and 74,974 (28.2%) into the unfavorable SDH group (Table 1). Among 32,018 participants from the US NHANES (mean [SD] age 46.7 [16.8] years, 51.8% females), 8,297 (25.9%) were categorized into the favorable SDH group, 16,040 (50.1%) into the medium SDH group, and 7,681 (24.0%) into the unfavorable SDH group. Overall, participants in the unfavorable SDH group were more likely to be female, to have unhealthy lifestyles, higher body mass index (BMI), higher prevalence of cardiovascular disease and chronic respiratory disease, and increased biological aging acceleration (Tables 1 and S1). Baseline characteristics by weighted SDH scores are presented in Tables S2 and S3. Excluded participants due to missing data were generally comparable to included participants except for slightly higher proportions of comorbidities and medication use (Table S4).

Table 1.

Baseline characteristics of the study participants from UK Biobank and US NHANES stratified by combined SDHs

Characteristics	UK Biobank^a				US NHANES^b
Characteristics	Overall (n = 266,029)	Favorable (n = 55,610)	Medium (n = 135,445)	Unfavorable (n = 74,974)	Overall (n = 32,018)	Favorable (n = 8297)	Medium (n = 16,040)	Unfavorable (n = 7681)
Age, years	56.1 (8.1)	57.3 (7.7)	56.2 (8.1)	55.0 (8.2)	46.7 (16.8)	49.8 (14.3)	46.6 (18.3)	40.2 (15.7)
Female	139,529 (52.4)	27,282 (49.1)	71,305 (52.6)	40,942 (54.6)	16,446 (51.8)	3912 (48.5)	8258 (53.6)	4276 (54.6)
Race
White	255,533 (96.1)	54,819 (98.6)	131,871 (97.4)	68,843 (91.8)	15,275 (72.2)	5869 (87.8)	7428 (70.4)	1978 (42.1)
Black	2923 (1.1)	28 (0.1)	430 (0.3)	2465 (3.3)	6013 (9.9)	425 (2.1)	2790 (9.4)	2798 (29.0)
BMI, kg/m²	27.3 (4.7)	26.5 (4.0)	27.1 (4.5)	28.2 (5.3)	28.5 (6.5)	28.0 (5.7)	28.7 (6.6)	29.3 (7.7)
Waist circumference, cm	90.1 (13.3)	88.6 (12.5)	89.7 (13.0)	92.1 (14.2)	97.9 (16.1)	97.4 (15.0)	98.0 (16.3)	98.4 (17.7)
Never smoking	145,707 (54.8)	34,189 (61.5)	75,917 (56.1)	35,601 (47.5)	17,087 (52.2)	4874 (57.2)	8577 (51.3)	3636 (43.2)
Moderate or no drinking	169,623 (63.8)	32,420 (58.3)	84,172 (62.1)	53,031 (70.7)	17,349 (51.9)	4381 (49.5)	8896 (53.7)	4072 (52.0)
Healthy diet	106,732 (40.1)	24,001 (43.2)	55,111 (40.7)	27,620 (36.8)	12,807 (38.5)	4019 (45.6)	6328 (36.3)	2460 (28.3)
Physically active	144,531 (54.3)	32,155 (57.8)	74,196 (54.8)	38,180 (50.9)	14,071 (48.9)	4527 (56.6)	6776 (45.6)	2768 (40.5)
Prevalent comorbidities
Cardiovascular disease	18,190 (6.8)	3052 (5.5)	8455 (6.2)	6683 (8.9)	3456 (8.5)	752 (7.1)	1819 (8.9)	885 (10.4)
Diabetes	14,022 (5.3)	1901 (3.4)	6253 (4.6)	5868 (7.8)	4348 (9.4)	801 (7.2)	2445 (10.6)	1102 (10.6)
High cholesterol	37,271 (14.0)	6843 (12.3)	18,242 (13.5)	12,186 (16.3)	9729 (29.3)	2971 (34.4)	4935 (28.2)	1823 (20.9)
Hypertension	67,794 (25.5)	12,498 (22.5)	33,285 (24.6)	22,011 (29.4)	10,699 (29.3)	2639 (28.9)	5633 (30.1)	2427 (28.0)
Chronic respiratory disease	7652 (2.9)	1134 (2.0)	3483 (2.6)	3035 (4.0)	2292 (7.2)	413 (5.2)	1133 (7.7)	746 (10.6)
Cancer	23,345 (8.8)	5066 (9.1)	11,931 (8.8)	6348 (8.5)	2915 (9.1)	979 (10.9)	1499 (9.0)	437 (5.4)
Medication use
Diabetes medication	8720 (3.3)	1106 (2.0)	3851 (2.8)	3763 (5.0)	3156 (6.7)	595 (5.2)	1804 (7.8)	757 (6.7)
Antihypertensive medication	51,672 (19.4)	9543 (17.2)	25,321 (18.7)	16,808 (22.4)	8035 (20.9)	2035 (21.3)	4374 (21.7)	1626 (17.6)
Cholesterol-lowering medication	43,046 (16.2)	7921 (14.2)	20,883 (15.4)	14,242 (19.0)	5114 (13.8)	1482 (16.1)	2732 (13.7)	900 (9.1)

Open in a new tab

In the UK Biobank cohort, P values were calculated using analysis of variance and χ2 test for continuous and categorical variables, respectively. All P values were <0.001. In the US NHANES cohort, all estimates accounted for complex survey designs, and P values were calculated using analysis of variance adjusting for sampling weights and Rao Scott χ2 test for continuous and categorical variables, respectively. All P values were <0.001, except for waist circumference (P = 0.040) and hypertension (P = 0.133).

BMI body mass index, NHANES National Health and Nutrition Examination Survey, SDHs social determinants of health.

^aData are presented mean (SD) or n (%).

^bData are presented as mean (SD) or n (%), and the means, SDs, and percentages are population weighted. Statistical tests were two-sided and P value of <0.05 was considered statistically significant.

Association of combined SDHs with accelerated biological aging

As shown in Fig. 2, KDM-BA and phenotypic age acceleration were positively correlated with combined SDH scores in the UK Biobank (Spearman’s R = 0.141 for KDM-BA and R = 0.158 for phenotypic age acceleration, P < 2.2e-16) and the US NHANES (Spearman’s R = 0.105 for KDM-BA and R = 0.123 for phenotypic age acceleration, P < 2.2e-16). Table 2 shows the associations between combined SDHs and accelerated biological aging. In UK Biobank, compared with participants in the favorable SDH group, participants in the medium SDH group had increased KDM-BA and phenotypic age acceleration of 0.56 (95% confidence interval [CI]: 0.49, 0.62) and 0.33 (0.28, 0.38) years, and those in the unfavorable SDH group had increased KDM-BA and phenotypic age acceleration of 1.50 (1.42, 1.57) and 1.07 (1.02, 1.13) years, respectively. In US NHANES, participants in the medium and unfavorable SDH group also had an increase in KDM-BA acceleration (medium SDH, 1.24 [0.77, 1.71] years; unfavorable SDH, 1.78 [1.14, 2.41] years) and phenotypic age acceleration (medium SDH, 0.45 [0.29, 0.60] years; unfavorable SDH, 0.82 [0.60, 1.05] years). The results remained consistent in sensitivity analyses using weighted combined SDH scores, additionally adjusting for waist circumference and sleep score, without adjustment for lifestyle factors, using imputation for all missing variables, and adopting different classifications of SDHs (Tables S5–S9). Across the five SDH domains, financial circumstances showed the strongest association with KDM-BA (0.69 [0.67-0.72] years) and phenotypic age acceleration (0.44 [0.42–0.46] years) in UK Biobank, while education access and quality and social and community context showed the strongest association with KDM-BA (0.77 [0.59-0.94] years) and phenotypic age acceleration (0.33 [0.26−0.40] years) in US NHANES, respectively (Table S10). Furthermore, we examined the associations between SDHs and change in biological aging and found that SDHs were associated with increased KDM-BA and phenotypic age acceleration over time (Table S11). In stratified analyses, consistent associations of combined SDHs with increased KDM-BA and phenotypic age acceleration were found among all subgroups, but seemed to be more pronounced among males, current or former smokers, and individuals with an unhealthy diet or comorbidities in the UK Biobank and among males in the US NHANES (P_interaction < 0.05) (Figs. S1 and S2).

Table 2.

Associations of combined SDHs with accelerated biological aging in the UK Biobank and US NHANES cohorts

SDH score	No. of participants	KDM-BA acceleration, β (95% CI)				Phenotypic age acceleration, β (95% CI)
SDH score	No. of participants	Model 1	P value	Model 2	P value	Model 1	P value	Model 2	P value
UK Biobank
Per 1-point increment	266,029	0.41 (0.40, 0.42)	<0.001	0.22 (0.21, 0.23)	<0.001	0.35 (0.34, 0.35)	<0.001	0.17 (0.16, 0.18)	<0.001
Favorable	55,610	Reference		Reference		Reference		Reference
Medium	135,445	1.05 (0.98, 1.12)	<0.001	0.56 (0.49, 0.62)	<0.001	0.76 (0.71, 0.81)	<0.001	0.33 (0.28, 0.38)	<0.001
Unfavorable	74,974	2.86 (2.78, 2.93)	<0.001	1.50 (1.42, 1.57)	<0.001	2.32 (2.26, 2.38)	<0.001	1.07 (1.02, 1.13)	<0.001
P trend	-	<0.001		<0.001		<0.001		<0.001
US NHANES
Per 1-point increment	32,018	0.71 (0.58, 0.84)	<0.001	0.39 (0.28, 0.51)	<0.001	0.37 (0.33, 0.42)	<0.001	0.17 (0.13, 0.21)	<0.001
Favorable	8297	Reference		Reference		Reference		Reference
Medium	16,040	2.08 (1.57, 2.59)	<0.001	1.24 (0.77, 1.71)	<0.001	1.00 (0.81, 1.19)	<0.001	0.45 (0.29, 0.60)	<0.001
Unfavorable	7681	3.46 (2.72, 4.20)	<0.001	1.78 (1.14, 2.41)	<0.001	1.88 (1.62, 2.14)	<0.001	0.82 (0.60, 1.05)	<0.001
P trend	-	<0.001		<0.001		<0.001		<0.001

Open in a new tab

Model 1 was adjusted for age and sex. Model 2 was additionally adjusted for BMI, smoking status, alcohol drinking, diet, physical activity, cardiovascular disease, cancer, diabetes, high cholesterol, hypertension, respiratory disease, and medication for cholesterol-lowering, anti-hypertensive, or anti-diabetes. Statistical tests were two-sided and P value of <0.05 was considered statistically significant. No adjustments were made for multiple comparisons.

BMI body mass index, KDM-BA Klemera–Doubal biological age, NHANES National Health and Nutrition Examination Survey, SDHs social determinants of health.

Associations of combined SDHs with mortality and incident diseases

During a median follow-up of 15.1 (IQR, 14.4–15.8) years, the UK Biobank documented 20,773 deaths, including 4327 deaths from cardiovascular disease, 1269 from respiratory disease, and 10,359 from cancer. During a median follow-up of 11.1 (IQR, 4.0–15.2) years, the US NHANES documented 5116 deaths, including 155 deaths from cardiovascular disease, 283 from respiratory disease, and 90 from cancer. More unfavorable SDHs were associated with higher risks of all-cause and cause-specific mortality (Figs. 3 and S3). For example, in the UK Biobank, each 1-point increase in the SDH scores was associated with a hazard ratio (HR) of 1.07 (95% CI: 1.07–1.08) for all-cause mortality. In the US NHANES, each 1-point increase in the SDH scores was associated with an HR of 1.22 (95% CI: 1.19–1.25) for all-cause mortality.

Fig. 3 — A HRs (95% CIs) of combined SDHs for health outcomes in UK Biobank. B HRs (95% CIs) of combined SDHs for health outcomes in US NHANES. Hazard ratios (HRs) of medium and unfavourable SDHs were calculated with favourable SDHs as reference (not shown in the figure). The blue squares represent the estimated HRs and error bars represent the 95% confidence intervals (CIs). The vertical line represents HR = 1.0. Model was adjusted for age, sex, BMI, smoking status, alcohol drinking, diet, physical activity, cardiovascular disease, cancer, diabetes, high cholesterol, hypertension, respiratory disease, and medication for cholesterol-lowering, anti-hypertensive, or anti-diabetes. For incident diseases, only participants free from the corresponding disease at baseline were included and prevalence of corresponding disease was not included in the model. Undiagnosed prevalent diabetes was identified using HbA1c levels ≥6.5% or the use of diabetes medication. Incident diseases were only available in UK Biobank. MI body mass index, NHANES National Health and Nutrition Examination Survey, SDHs social determinants of health. Source data are provided as a Source Data file.

Over a median follow-up of up to 15.1 (IQR, 14.3–15.8) years in the UK Biobank, 34,350 participants developed cardiovascular disease, 10,468 developed type 2 diabetes, 8507 developed chronic liver disease, 11,159 developed chronic respiratory disease, and 12,545 developed chronic kidney disease. Similar graded associations between unfavorable SDHs and risks of incident diseases were detected (Figs. 3 and S3). For instance, in UK Biobank, each 1-point increase in the SDH scores was associated with an HR of 1.04 (95% CI: 1.03–1.04) for cardiovascular disease, 1.10 (1.09-1.11) for type 2 diabetes, 1.08 (1.07–1.09) for liver disease, 1.13 (1.13–1.14) for chronic respiratory disease, and 1.06 (1.05–1.07) for chronic kidney disease. These associations remained robust when using weighted SDH scores, adjusting for waist circumference and sleep score, without adjustment for lifestyle factors, or taking into account the competing risk (Tables S12–S16). Furthermore, the combined effects of genetic risk and overall SDHs on incident diseases expressed in a dose-response manner, and significant interactions were observed between combined SDHs and genetic risk for type 2 diabetes and chronic respiratory disease (Table S17).

Mediation analysis

Both KDM-BA and phenotypic age acceleration showed positive associations with risks of mortality and incident diseases (Table S18). We further performed mediation analyses and found that accelerated biological aging significantly mediated the associations between combined SDHs and all-cause and cause-specific mortality (Fig. 4). For all-cause mortality, the estimates for the indirect effect of KDM-BA and phenotypic age acceleration were 0.50 (95% CI: 0.46, 0.54) and 0.69 (0.64, 0.74) in UK Biobank, respectively; the proportion mediated by KDM-BA acceleration was 13.46% (12.13%, 14.97%) and the proportion mediated by phenotypic age acceleration was 18.55% (16.89%, 20.45%). The mediation proportions of KDM-BA and phenotypic age acceleration in US NHANES were 7.62% (5.98%, 9.42%) and 9.41% (7.72%, 11.26%), respectively. In addition, accelerated biological aging significantly mediated the relationships between combined SDHs and incident diseases in the UK Biobank. The mediation proportions attributed to KDM-BA (22.19% [19.46%, 25.61%]) and phenotypic age acceleration (30.48% [26.93%, 34.91%]) were largest for chronic kidney disease, followed by cardiovascular disease (KDM-BA: 15.38% [13.36%, 17.83%]; phenotypic age: 15.51% [13.49%, 17.96%]). Results were largely unchanged in all sensitivity analyses (Tables S19–S22). Additionally, structural equation modeling revealed that KDM-BA and phenotypic age acceleration exhibited distinct mediation patterns across different SDH domains. For instance, in the UK Biobank cohort, 10.5% (95% CI: 9.4–11.7%) of the association between financial circumstance and all-cause mortality was mediated by KDM-BA acceleration, and 13.1% (11.7–14.5%) was mediated by phenotypic age acceleration. In the US NHANES cohort, for the association between social and community context and all-cause mortality, 9.5% (4.6–14.4%) was mediated by KDM-BA acceleration, and 20.2% (12.2–28.2%) was mediated by phenotypic age acceleration (Tables S23–S26).

Fig. 4 — A Mediating role of KDM-BA acceleration and phenotypic age acceleration in the associations between combined SDHs and mortality and incident diseases in UK Biobank. B Mediating role of KDM-BA acceleration and phenotypic age acceleration in the associations between combined SDHs and mortality and incident diseases in US NHANES. The effects are presented as the number of additional cases per 10,000 person-years for each 1-point increment in SDH score using the Aalen additive hazard model. The 95% confidence intervals (CIs) for direct and indirect effects and proportions mediated were estimated using simulation with 100,000 repeats. The size of the squares or bars and the internal center represent point estimation. The error bar and horizontal lines indicate the corresponding 95% CI. Red denotes mediation by KDM-BA acceleration, while blue denotes mediation by Phenotypic age acceleration. The number of participants across different health outcomes in the UK Biobank ranged from 247,839 to 266,029 (247,839 for cardiovascular disease, 251,860 for type 2 diabetes, 263,682 for liver disease, 258,377 for chronic respiratory disease, 262,500 for chronic kidney disease, and 266,029 for all-cause and cause-specific mortality), while the number of participants in the US NHANES was 32,018. Model was adjusted for age, sex, BMI, smoking status, alcohol drinking, diet, physical activity, cardiovascular disease, cancer, diabetes, high cholesterol, hypertension, respiratory disease, and medication for cholesterol-lowering, anti-hypertensive, or anti-diabetes. For incident diseases, only participants free from the corresponding disease at baseline were included and prevalence of corresponding disease was not included in the model. Undiagnosed prevalent diabetes was identified using HbA1c levels ≥6.5% or the use of diabetes medication. Incident diseases were only available in the UK Biobank. BMI body mass index, KDM-BA Klemera-Doubal biological age, NHANES National Health and Nutrition Examination Survey, SDHs social determinants of health. Source data are provided as a Source Data file.

Discussion

In the two nationwide cohort studies, combined SDHs were significantly associated with KDM-BA and phenotypic age acceleration. Compared with individuals in the favorable SDH group, individuals in the unfavorable SDH group had an increase in accelerated biological aging. Unfavorable SDHs increased the risks of mortality and incident diseases. In addition, accelerated biological aging was detected as a significant mediator in the associations between combined SDHs and all-cause and cause-specific mortality, as well as in the associations between combined SDHs and incident diseases, particularly for chronic kidney disease and cardiovascular disease.

With the fast growth of aging population, efforts to delay biological aging and disclose the population heterogeneity in aging are required for preventing diseases and improving health. Our results are aligned with studies that reported a positive association between disadvantaged SDHs and accelerated biological aging. However, previous studies mostly focused on a single SDH or socioeconomic status^15,16,21,22. Since different disadvantaged SDHs often coexist and the causal pathways are numerous, interconnected, and complex¹⁸, it is difficult to accurately quantify the impact of individual SDHs on accelerated biological aging. Moreover, focusing on certain factors in isolation will preclude the identification of socially vulnerable population, consequently weakening the effectiveness of interventions¹⁸. Furthermore, most previous studies have limited sample sizes (n < 3000) and assessed aging metrics at epigenetic, mRNA, or other molecular levels, which may restrict their wide use in large population-based cohort studies due to concerns about cost and practicality^7,23. By contrast, composite algorithms combining information from standard clinical parameters serve as a more cost-effective alternative while ensuring measurement accuracy in predicting health outcomes, making them more applicable and clinically relevant^7,24.

In this study, each 1-point increment in overall SDH score was associated with increased KDM-BA and phenotypic age acceleration by 0.22 and 0.17 years in the UK Biobank and 0.39 and 0.17 years in US NHANES. This finding underscores the significant benefits of interventions targeting combined SDHs in delaying biological aging and preventing aging-related diseases. Additionally, the effects of combined SDHs on accelerated biological aging were stronger in individuals with unhealthy lifestyles, including smoking and an unhealthy diet. Considering the relationships between SDHs, unhealthy lifestyles, and biological aging metrics^25,26, public health initiatives aiming to improve SDHs may be prioritized in those with unhealthy lifestyles. We also found more pronounced associations of combined SDHs with accelerated biological aging in men than women, possibly due to sex differences in hormones and psychology. Further studies are required to validate our findings.

In line with previous studies, we found that unfavorable SDHs elevated the risk of mortality and incident cardiometabolic disease, liver disease, chronic respiratory disease, and chronic kidney disease^11–14. Our study extended the findings by demonstrating the mediation effect of accelerated biological aging on linking combined SDHs to health outcomes. This finding not only supports previous reports regarding the associations between accelerated biological aging and multiple health outcomes^19,27, but also highlights the importance of measuring biological aging for prediction and prevention of major chronic diseases and early mortality risk in individuals with disadvantaged SDHs. Our results based on KDM-BA and phenotypic age were largely comparable, despite the slight variations in estimated mediation proportions between the two measures. The aging-related mechanisms underpinning the connection between combined SDHs and health outcomes warrant further exploration. Of note, the mediation proportion of biological aging in the association between SDHs and mortality risk ranged from 13.46 to 25.21% in the UK Biobank and 7.62 to 22.16% in US NHANES, which were lower than those shown in prior studiesGraf et al. reported that biological aging measures, including the blood-chemistry-based phenotypic age, DNA-methylation (DNAm) GrimAge clock, and DNAm pace-of-aging measure (DunedinPoAm), accounted for 52–80% of the Black-White differences in mortality in the Health and Retirement Study²⁸. Additionally, Graf et al. reported that DunedinPoAm mediated 50% of the association between educational mobility and mortality risk in the Framingham Heart Study²⁹. The observed discrepancies may be partly explained by differences in SDH measurement. Compared with previous studies focusing on isolated SDHs (e.g., race or educational mobility), the composite SDH score used in our study reflects a more complicated interconnection of different SDHs with biological aging. This potentially leads to an attenuation of mediation effects through diffuse biological pathways. Additionally, the weaker associations observed between SDHs, biological aging, and health outcomes in the UK Biobank may be attributed to healthy volunteer bias, limited socioeconomic diversity, and predominantly European ancestry within the cohort³⁰. In this case, the mediation estimates generated are specific to this population and may not be applicable to more diverse groups. The relatively low mediation proportions also suggest that KDM-BA and phenotypic age may not fully capture the aging processes mediating SDHs and health outcomes in our cohorts^6,7.

We observed stronger associations of SDHs with KDM-BA acceleration in the US NHANES, whereas the associations with phenotypic age acceleration were more pronounced in the UK Biobank. These differences may be relevant to variations in the construction of the combined SDH scores, which were based on the availability of SDH variables within each cohort. Moreover, KDM-BA and phenotypic age capture distinct biological aging dimensions, which may differentially reflect the impact of SDHs across populations. However, the consistent association between unfavorable SDHs and accelerated biological aging across two cohorts with diverse cultural backgrounds and social contexts has important public health implications. The combined SDH score could be used as a positive frame of reference for screening socially vulnerable populations. Besides, the use of clinical parameters-derived aging algorithms would help determine high-risk individuals among socially disadvantaged groups who are most in need of practical interventions to promote healthy longevity. Such approaches would facilitate health management to address health inequalities caused by SDHs.

Our study comprehensively investigates the associations between combined SDHs, accelerated biological aging, and multiple health outcomes within the same cohort populations. Major strengths of this study include the large sample size, long follow-up, and consistent results from two well-established nationwide cohorts. Furthermore, the use of genetic data enables us to account for genetic risk and explore gene-environment interaction. Nevertheless, several limitations should be noted. First, information on SDHs was self-reported and collected once at baseline, which might have been subject to measurement bias. Also, we only estimated biological age at baseline and could not capture the long-term trajectory or explore the effect of SDHs on the rates of biological aging changes over time. Second, due to the differences in study design and data collection, combined SDH scores were constructed based on available data from the UK Biobank and US NHANES, making it difficult to generate an identical score for both cohorts. The selection of SDHs in the two cohorts might have reflected different aspects related to social health, thus, misclassification bias in different populations might exist. Likewise, biological aging metrics were established using available clinical biomarkers in each cohort, which may limit direct comparability across the two cohorts. However, the use of cohort-specific validated measures ensures internal validity within each cohort, and the consistent association between SDHs and accelerated biological aging supports the broader relevance and generalizability of our findings. Third, the differences in certain characteristics between included and excluded participants may raise the concern about selection bias. Nevertheless, we conducted a sensitivity analysis using multiple imputation for all missing variables, including SDHs, biological aging markers, and covariates. The results from the imputed dataset were largely consistent with our primary findings, suggesting that the impact of exclusion of participants with missing data on SDHs and biological aging would be minor. Fourth, since combined SDHs and biomarker data were collected at baseline, the temporal relationship between SDHs and biological aging was uncertain. To address this concern, we used the first repeat measurement of biomarkers (2012–2013) in the UK Biobank to assess the change in biological aging and found that SDHs were associated with accelerated biological aging over time. Fifth, the UK Biobank is acknowledged to have a low response rate (5%), with participants being more likely to come from less deprived areas and have healthier lifestyle behaviors compared with the general UK population. Consequently, the healthy volunteer bias might be introduced³¹. The fact that people who were not recruited tended to have both less favorable SDHs and unhealthy conditions could lead to an underestimation of our results. Moreover, the cohort’s limited socioeconomic diversity and predominantly European ancestry further restrict the representativeness of the sample. Finally, due to the observational nature of the present study, the possibility of residual confounding could not be eliminated, although we have adjusted for multiple potential confounders.

In two large nationwide cohorts from the UK and US, more unfavorable SDHs were dose-dependently associated with increased KDM-BA and phenotypic age acceleration. Furthermore, accelerated biological aging significantly mediated the associations between combined SDHs and mortality and incident diseases. Our findings suggest that targeting accelerated biological aging may represent a promising strategy to narrow the health inequalities caused by SDHs.

Methods

Study design and population

The UK Biobank is a large-scale cohort that recruited over 500,000 participants aged 37–73 years from 22 assessment centers in England, Wales, and Scotland between 2006 and 2010. The study design and data collection procedures have been reported in detail previously³². Among the 502,415 participants, we excluded those without information on SDHs and biological aging. A total of 266,029 participants were included to examine the associations of SDHs with accelerated biological aging and mortality (Fig. S4). For the analysis of the joint effects of SDHs and genetic susceptibility on disease outcomes, we restricted the sample to individuals of European descent who had genetic information (n = 253,783). Additionally, participants with outcomes of interest at baseline were excluded for the analysis of incident diseases.

The US NHANES is an ongoing cross-sectional survey with a nationally representative sample of the civilian resident population of the USA using a complex, multistage probability design in 2-year cycles since 1999–2000. Details have been described elsewhere³³. We included 55,081 participants aged 20 years and older from cycles 1999 to 2018. After excluding those with missing data on SDHs and biological aging, 32,018 participants were retained in the analysis (Fig. S5).

Assessment of SDHs

We systematically reviewed standardized questionnaires from the UK Biobank and US NHANES, mapping variables to the five domains guided by the Healthy People 2030³⁴ and World Health Organization³⁵: economic stability, education access and quality, health care access and quality, neighborhood and built environment, and social and community context. The selection of specific variables within these domains was informed by prior studies, which demonstrated strong associations between such variables and health outcomes, including mortality and major chronic diseases^11,12,36,37. Specifically, we chose 18 SDH variables (household income, employment status, income quality, education attainment, education quality, healthcare, accommodation stability, housing quality, local crime rate, natural environment, psychosocial problems, race/ethnicity, living alone/with partners, social support, social activity, social isolation, loneliness, and emotional distress) in UK Biobank. Loneliness, the subjective experience of social disconnection, has also been recognized as a significant health risk factor and was therefore included in this study³⁸. For the US NHANES, 10 SDH variables (household income, employment status, food security, education attainment, healthcare, health insurance, accommodation stability, psychosocial problems, race/ethnicity, and marital status) were considered. All variables were collected from standardized questionnaires at baseline or linkage to the index of multiple deprivation for the corresponding Lower-layer Super Output Area in the UK. Details of SDHs assessment are provided in Table S27.

Each SDH was dichotomized into advantaged and disadvantaged levels using conventional cut points (Table S28) established in prior literature to enhance interpretability and facilitate the application of SDH score^11,12,36,37. We constructed an unweighted score by assigning 0 points for the advantaged level of each SDH and 1 point for the disadvantaged level. As a sensitivity analysis, we also constructed weighted SDH scores to account for varying magnitude of the associations (absolute β coefficients) between individual SDHs and biological aging acceleration (Table S29). In UK Biobank, weighted SDHs score = (β₁ × SDH₁ + … + β₁₈ × SDH₁₈) × (18/sum of β coefficients). In the US NHANES, weighted SDHs score = (β₁ × SDH₁ + … + β₁₀ × SDH₁₀) × (10/sum of β coefficients). The unweighted and weighted SDH score ranged from 0 to 18 in the UK Biobank and from 0 to 10 in US NHANES (Fig. S6), with higher scores indicating more unfavorable SDHs. Participants were classified into three groups by tertiles: favorable SDH group, medium SDH group, and unfavorable SDH group. Furthermore, we examined the associations of different SDH domains with biological aging. SDH domain scores were calculated by summing up individual scores for each SDH within specific domains and were standardized to facilitate comparisons across different SDH domains.

Assessment of biological age and age acceleration

Two best-trained biological age algorithms, KDM-BA and phenotypic age, were used. These algorithms have been originally developed and validated in the US NHANES^23,39,40 and subsequently adapted for application in the UK Biobank using available biomarker data^19,41,42. The clinical parameters used for biological age calculation, including anthropometric measurements and biochemical markers, were collected at baseline. Due to the differences in study designs and populations, different parameters were collected in the UK Biobank (See Table S30 for the corresponding data field IDs) and US NHANES, which referred to previous studies conducted within these two cohorts^19,42,43.

An individual’s KDM-BA prediction reflects the chronological age at which their physiological state would be considered approximately normal. KDM-BA is derived from a series of regressions of individual biomarkers on chronological age within a reference population⁴⁴. In the UK Biobank, we utilized nine biomarkers to calculate KDM-BA: systolic blood pressure, forced expiratory volume in one second (FEV1), total cholesterol, glycated hemoglobin, blood urea nitrogen, albumin, creatinine, C-reactive protein, and alkaline phosphatase^19,42. The US NHANES included four additional biomarkers: uric acid, lymphocyte percentage, mean cell volume, and white blood cell count, while omitting FEV1, resulting in a total of twelve biomarkers⁴³. Phenotypic age, initially developed using elastic-net Gompertz regression of mortality on 42 biomarkers among NHANES participants, can be interpreted as the age at which an individual’s mortality risk aligns with the predicted average mortality risk²⁷. In this study, we calculated an individual’s phenotypic age using chronological age and nine biomarkers: lymphocyte percentage, mean cell volume, glucose, red cell distribution width, white blood cell count, albumin, creatinine, C-reactive protein, and alkaline phosphatase in the UK Biobank^19,42. In the US NHANES, we utilized chronological age and the same twelve biomarkers for KDM-BA calculation⁴³.

Computation of KDM-BA and phenotypic age values was conducted using the R package BioAge (https://github.com/dayoonkwon/BioAge). Age acceleration was calculated as the residual by regressing KDM-BA and phenotypic age on chronological age⁴². An age acceleration value greater than 0 indicated an advanced state of biological aging, while an age acceleration value less than 0 indicated a delayed biological aging state⁴⁵.

Polygenic risk scores

Genetic data were obtained from the UK Biobank, and details of the single-nucleotide polymorphisms (SNPs) regarding genotyping, imputation, and quality control have been reported previously⁴⁶. For cardiovascular disease and type 2 diabetes, we used polygenic risk scores (PRS) that were calculated by using external genome-wide association studies (GWAS) data and were made available in the UK Biobank (Category 301, https://biobank.ndph.ox.ac.uk/showcase/label.cgi?id=301, Table S31)⁴⁷. For other diseases with no PRS provided in the UK Biobank, we retrieved Polygenic Score (PGS) catalog⁴⁸ or GWAS summary statistics derived in predominantly European populations (Table S31). The number of associated effect alleles was weighted by the strengths of associations with diseases and summed to derive a PRS in UK Biobank individuals of European ancestry (Supplementary Methods). Participants were then categorized into low genetic risk (tertile 1), medium genetic risk (tertile 2), and high genetic risk (tertile 3) according to PRS tertiles.

Assessment of health outcomes

Health outcomes were ascertained using International Classification of Diseases 10th Revision (ICD-10) codes. The primary outcomes included all-cause mortality and cause-specific mortality (from cardiovascular disease, respiratory disease, and cancer). In the UK Biobank, information on the date and cause of death was obtained from death certificates held within the National Health Service (NHS) Information Centre (England and Wales) and the NHS Central Register (Scotland) to April 30, 2024⁴⁹ (Table S32). In the US NHANES, deaths were ascertained through linkage to death certificates from the National Death Index records until December 31, 2019⁵⁰. For cardiovascular disease mortality, only deaths from heart disease were available in the US NHANES.

The secondary outcomes were incident diseases, including cardiovascular disease, type 2 diabetes, liver disease, chronic respiratory disease, and chronic kidney disease, which were only available from the UK Biobank. Disease diagnoses were obtained from hospital inpatient data, primary care records, and death register linkage according to the ICD-10 codes up to April 30, 2024⁵¹ (Table S32).

Covariates

Information on demographic characteristics, lifestyle factors, and medication use was collected through self-report questionnaires. The demographic characteristics included age and sex. Lifestyle factors included BMI, waist circumference, smoking status (never, previous, or current), alcohol drinking (moderate or no drinking, or heavy drinking), diet (healthy or unhealthy), and physical activity (active or inactive). Prevalent diseases, including cardiovascular disease, cancer, diabetes, high cholesterol, hypertension, and respiratory disease, were obtained from self-reports (both cohorts) and disease registers (UK Biobank only). Medication use for high cholesterol, hypertension, and diabetes was extracted. Weight, height, and waist circumference were measured by trained technicians, and BMI was calculated as weight in kg divided by height in m². Details of covariates are shown in Supplementary Methods.

Statistical analysis

All statistical analyses in the US NHANES accounted for the complex survey design and yielded population-weighted estimates representative of the US adults. Baseline characteristics of participants were described as mean (SD) for continuous variables and absolute numbers with percentages for categorical variables across SDH groups. In the US NHANES, means, SDs, and percentages were population-weighted. Differences across SDH groups were tested using analysis of variance for continuous variables and the Rao-Scott χ² test for categorical variables, whereby sampling weights were adjusted. In the UK Biobank, analysis of variance and the χ² test were employed for continuous and categorical variables, respectively. Correlations between SDH score and biological aging acceleration were assessed using Spearman’s correlation coefficient. Multiple imputation by chained equations was utilized to impute the missing values of covariates.

We used multivariable linear regression models to assess the associations of combined SDH score with KDM-BA acceleration and phenotypic age acceleration. Model 1 was adjusted for age and sex. Model 2 was further adjusted for BMI, smoking status, alcohol drinking, diet, physical activity, cardiovascular disease, cancer, diabetes, high cholesterol, hypertension, respiratory disease, and medication use for diabetes, high cholesterol, and hypertension. The P value for trend across SDH groups was calculated using integer values. Moreover, we explored the associations of individual SDHs with biological aging.

We performed stratified analyses to test the potential modification effects on the association between combined SDH score and accelerated biological aging by age (<50, 50–59, vs ≥60 years), sex (female vs male), cardiovascular disease (yes vs no), diabetes (yes vs no), high cholesterol (yes vs no), hypertension (yes vs no), BMI (<30 vs ≥30 kg/m²), smoking (never, previous, or current), moderate or no drinking (yes vs no), physically active (yes vs no), and healthy diet (yes vs no).

Cox proportional hazards regression models were used to estimate the associations of combined SDH score and biological aging acceleration with health outcomes. Person years were calculated from baseline until the diagnosis of the incident diseases, death, or the end of follow-up, whichever occurred first. Proportional hazard assumptions were satisfied, as indicated by Schoenfeld residual test. All the covariates in Model 2 were included for adjustment. Additionally, we explored the joint effects of combined SDHs and genetic risk on incident diseases, with additional adjustment for genetic kinship and the first 10 genetic principal components. The interaction between combined SDHs and genetic risk was assessed by adding a cross-product term in the models.

We further investigated whether accelerated biological aging mediated the associations of combined SDH score with mortality and incident diseases. Causal mediation analyses were conducted using the Aalen additive hazards model to estimate the total effect of SDH on the outcome, adjusting for covariates included in Model 2, which was decomposed into the natural direct effect and the natural indirect effect^52,53. The effect estimates are presented as the number of additional events per person-time, offering a more interpretable metric from a public health perspective. The 95% CIs for direct and indirect effects and proportions mediated were estimated using simulation with 100,000 repeats, as previously described⁵². In addition, generalized structural equation modeling was performed to assess the mediating role of biological aging in the relationships between specific SDH domains and health outcomes. The 95% CIs for direct and indirect effects and proportions mediated were estimated by 10,000-iteration nonparametric bootstrap approach⁵⁴.

A series of sensitivity analyses was conducted. First, a weighted SDH score was constructed to test the robustness of the results. Second, we additionally adjusted for waist circumference and healthy sleep score to reduce potential residual confounding. Third, as lifestyle behaviors may act as potential mediators in the associations between specific SDHs, biological aging, and health outcomes, we repeated the analyses without adjusting for lifestyle behaviors to detect the differences in effect estimates. Fourth, Fine and Gray’s proportional subhazards models were applied to account for competing risk of death for cause-specific mortality (with deaths from other causes as competing events) and incident diseases (with death as a competing event). Fifth, to reduce temporal uncertainty, we assessed the association between SDHs and changes in biological aging among 4273 UK Biobank participants with biomarker data at both baseline and the first repeat visit (2012–2013). The changes in biological aging were calculated as the difference between KDM-BA and phenotypic age acceleration at two time points. Sixth, we performed multiple imputation for all variables with missing data, including SDHs, biological aging markers, and covariates. Last, to mitigate potential information loss from dichotomizing SDHs, we reclassified SDHs according to the following principles: SDHs obtained from questionnaires were categorized into three or more groups based on the original questions, while those derived from area-level scores were classified into three categories according to the tertiles of their distributions (Tables S33 and S34).

All analyses were performed using R (version 4.4.1). A two-tailed P value of less than 0.05 was considered statistically significant.

Ethics statement

The UK Biobank received approval from the North West Multicenter Research Ethics Committee and all participants provided electronic informed consent. This study has been conducted under application number 77740. The study protocols for the US NHANES were approved by the ethics review board of the National Center for Health Statistics and written informed consent was provided by all participants.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(1.8MB, pdf)}

Reporting Summary^{(2.9MB, pdf)}

Transparent Peer Review File^{(3MB, pdf)}

Source data

Source data^{(11.1MB, xlsx)}

Acknowledgements

We thank all participants and staff of the UK Biobank and US NHANES cohorts for their dedication and contribution to this study. Prof. Y.L. was supported by National Natural Science Foundation of China (82404337, 82170870, and 82120108008), Science and Technology Commission of Shanghai Municipality (24ZR1443400 and 22015810500), and Major Science and Technology Innovation Program of Shanghai Municipal Education Commission (2019–01-07–00-01-E00059).

Author contributions

Y.L. and B.W. conceived and designed the study. Jiang.L. and Jie.L. performed the statistical analysis and drafted the manuscript. X.X., W.S., Y.S., and Y.F. participated in data collection. X.T., N.W., L.Q., B.W., and Y.L. critically revised the manuscript. All authors reviewed the manuscript drafts, critically revised the manuscript, and approved the final manuscript. B.W. and Y.L. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

For the UK Biobank study, the datasets generated and analyzed are available at https://www.ukbiobank.ac.uk/ and our research has been conducted under application number 77740. The raw UK Biobank data are protected and are not available due to data privacy laws. Researchers can apply to use the UK Biobank resource for health-related research and public interest via the UK Biobank Access Management System (https://ams.ukbiobank.ac.uk/ams/). For the US NHANES study, the data are public available at https://www.cdc.gov/nchs/nhis/index.htm. Source data are provided with this paper.

Code availability

Biological age estimation was conducted using the R package BioAge (v0.1.0) (https://github.com/dayoonkwon/BioAge). Additional analyses were conducted using packages including stats (v4.4.1), survival (v3.6-4), cmprsk (2.2-12), survey (v4.4-2), timereg (v2.0.6). No customized code was developed in this study. The analytic code used will be made available from the corresponding author upon request.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Jiang Li, Jie Li.

Contributor Information

Ningjian Wang, Email: wnj486@126.com.

Lu Qi, Email: lqi1@tulane.edu.

Bin Wang, Email: binwang1126@163.com.

Yingli Lu, Email: luyingli2008@126.com.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-67622-7.

References

1.Moqri, M. et al. Biomarkers of aging for the identification and evaluation of longevity interventions. Cell186, 3758–3775 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Chang, A. Y., Skirbekk, V. F., Tyrovolas, S., Kassebaum, N. J. & Dieleman, J. L. Measuring population ageing: an analysis of the Global Burden of Disease Study 2017. Lancet Public Health4, e159–e167 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Ferrucci, L. et al. Measuring biological aging in humans: a quest. Aging Cell19, e13080 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Kothari, M. & Belsky, D. W. Unite to predict. Elife10, e66223 (2021). [DOI] [PMC free article] [PubMed]
5.Belsky, D. W. et al. Eleven telomere, epigenetic clock, and biomarker-composite quantifications of biological aging: do they measure the same thing? Am. J. Epidemiol.187, 1220–1230 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Li, X. et al. Longitudinal trajectories, correlations and mortality associations of nine biological ages across 20-years follow-up. Elife9, e51507 (2020). [DOI] [PMC free article] [PubMed]
7.Xiang, Y. et al. Tea consumption and attenuation of biological aging: a longitudinal analysis from two cohort studies. Lancet Reg. Health West Pac.42, 100955 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.WHO. Social determinants of health. https://www.who.int/health-topics/social-determinants-of-health#tab=tab_1 (2023).
9.Marmot, M., Allen, J., Boyce, T., Goldblatt, P. & Morrison, M. Health equity in England: the Marmot review ten years on (Institute of Health Equity, 2020).
10.Bor, J., Cohen, G. H. & Galea, S. Population health in an era of rising income inequality: USA, 1980-2015. Lancet389, 1475–1490 (2017). [DOI] [PubMed] [Google Scholar]
11.Bundy, J. D. et al. Social determinants of health and premature death among adults in the USA from 1999 to 2018: a national cohort study. Lancet Public Health8, e422–e431 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Zhong, J. et al. Associations of social determinants of health with life expectancy and future health risks among individuals with type 2 diabetes: two nationwide cohort studies in the UK and USA. Lancet Healthy Longev.5, e542–e551 (2024). [DOI] [PubMed] [Google Scholar]
13.Shah, N. S. et al. Social determinants of cardiovascular health in Asian Americans: a scientific statement from the American Heart Association. Circulation150, e296–e315 (2024). [DOI] [PubMed] [Google Scholar]
14.Hill-Briggs, F. et al. Social determinants of health and diabetes: a scientific review. Diabetes Care44, 258–279 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Potente, C. et al. Socioeconomic inequalities and molecular risk for aging in young adulthood. Am. J. Epidemiol.192, 1981–1990 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Cabeza de Baca, T. et al. Chronic psychosocial and financial burden accelerates 5-year telomere shortening: findings from the coronary artery risk development in young adults study. Mol. Psychiatry25, 1141–1153 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Steptoe, A. & Zaninotto, P. Lower socioeconomic status and the acceleration of aging: an outcome-wide analysis. Proc. Natl. Acad. Sci. USA117, 14911–14917 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Figueroa, J. F., Frakt, A. B. & Jha, A. K. Addressing social determinants of health: time for a polysocial risk score. JAMA323, 1553–1554 (2020). [DOI] [PubMed] [Google Scholar]
19.Jiang, M. et al. Accelerated biological aging elevates the risk of cardiometabolic multimorbidity and mortality. Nat. Cardiovasc. Res.3, 332–342 (2024). [DOI] [PubMed] [Google Scholar]
20.Kuo, C. L. et al. Proteomic aging clock (PAC) predicts age-related outcomes in middle-aged and older adults. Aging Cell23, e14195 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Maunakea, A. K. et al. Socioeconomic status, lifestyle, and DNA methylation age among racially and ethnically diverse adults: NIMHD Social Epigenomics Program. JAMA Netw. Open7, e2421889 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Shen, B. et al. Association of race and poverty status with DNA methylation-based age. JAMA Netw. Open6, e236340 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Liu, Z. et al. A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: a cohort study. PLoS Med.15, e1002718 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Chan, M. S. et al. A biomarker-based biological age in UK Biobank: composition and prediction of mortality and hospital admissions. J. Gerontol. A Biol. Sci. Med. Sci.76, 1295–1302 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Astuti, Y., Wardhana, A., Watkins, J. & Wulaningsih, W. Cigarette smoking and telomere length: a systematic review of 84 studies and meta-analysis. Environ. Res.158, 480–489 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Goeminne, L. J. E. et al. Plasma protein-based organ-specific aging and mortality models unveil diseases as accelerated aging of organismal systems. Cell Metab.37, 205-222.e6 (2024). [DOI] [PubMed]
27.Parker, D. C. et al. Association of blood chemistry quantifications of biological aging with disability and mortality in older adults. J. Gerontol. A Biol. Sci. Med. Sci.75, 1671–1679 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Graf, G. H. et al. Testing black-white disparities in biological aging among older adults in the United States: analysis of DNA-methylation and blood-chemistry methods. Am. J. Epidemiol.191, 613–625 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Graf, G. H. J. et al. Educational mobility, pace of aging, and lifespan among participants in the Framingham Heart Study. JAMA Netw. Open7, e240655 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Keyes, K. M. & Westreich, D. UK Biobank, big data, and the consequences of non-representativeness. Lancet393, 1297 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Fry, A. et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am. J. Epidemiol.186, 1026–1034 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Sudlow, C. et al. UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med.12, e1001779 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.National Center for Health Statistics. NHANES survey methods and analytic guidelines. https://wwwn.cdc.gov/Nchs/Nhanes/AnalyticGuidelines.aspx.
34.US Department of Health and Human Services. Social Determinants of Health. Healthy People 2030. https://odphp.health.gov/healthypeople/priority-areas/social-determinants-health (2023).
35.Marmot, M., Friel, S., Bell, R., Houweling, T. A. & Taylor, S. Closing the gap in a generation: health equity through action on the social determinants of health. Lancet372, 1661–1669 (2008). [DOI] [PubMed] [Google Scholar]
36.Zhu, R. et al. Prevalence of cardiovascular-kidney-metabolic syndrome stages by social determinants of health. JAMA Netw. Open7, e2445309 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Huang, H. et al. Association between social determinants of health and survival among the US cancer survivors population. BMC Med.22, 343 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Wang, F. et al. A systematic review and meta-analysis of 90 cohort studies of social isolation, loneliness and mortality. Nat. Hum. Behav.7, 1307–1319 (2023). [DOI] [PubMed] [Google Scholar]
39.Klemera, P. & Doubal, S. A new approach to the concept and computation of biological age. Mech. Ageing Dev.127, 240–248 (2006). [DOI] [PubMed] [Google Scholar]
40.Levine, M. E. Modeling the rate of senescence: can estimated biological age predict mortality more accurately than chronological age? J. Gerontol. A Biol. Sci. Med. Sci.68, 667–674 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Kuo, C. L., Pilling, L. C., Liu, Z., Atkins, J. L. & Levine, M. E. Genetic associations for two biological age measures point to distinct aging phenotypes. Aging Cell20, e13376 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Gao, X. et al. Accelerated biological aging and risk of depression and anxiety: evidence from 424,299 UK Biobank participants. Nat. Commun.14, 2277 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Kwon, D. & Belsky, D. W. A toolkit for quantification of biological age from blood chemistry and organ function test data: BioAge. Geroscience43, 2795–2808 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Hastings, W. J., Shalev, I. & Belsky, D. W. Comparability of biological aging measures in the National Health and Nutrition Examination Study, 1999–2002. Psychoneuroendocrinology106, 171–178 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Cui, F. et al. Early-life exposure to tobacco, genetic susceptibility, and accelerated biological aging in adulthood. Sci. Adv.10, eadl3747 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Bycroft, C. et al. The UK Biobank resource with deep phenotyping and genomic data. Nature562, 203–209 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Thompson, D. J. et al. A systematic evaluation of the performance and properties of the UK Biobank Polygenic Risk Score (PRS) Release. PLoS ONE19, e0307270 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Lambert, S. A. et al. The polygenic score catalog as an open database for reproducibility and systematic evaluation. Nat. Genet.53, 420–425 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.UK Biobank. Mortality data: linkage to death registries. https://biobank.ndph.ox.ac.uk/showcase/ukb/docs/DeathLinkage.pdf (2023).
50.National Center for Health Statistics. 2019 public-use linked mortality files. https://www.cdc.gov/nchs/data-linkage/mortality-public.htm (2022).
51.UK Biobank. First Occurrence of health outcomes defined by 3-character ICD10 code. https://biobank.ndph.ox.ac.uk/showcase/ukb/docs/first_occurrences_outcomes.pdf (2019).
52.Lange, T. & Hansen, J. V. Direct and indirect effects in a survival context. Epidemiology22, 575–581 (2011). [DOI] [PubMed] [Google Scholar]
53.Li, H. et al. Smoking-induced risk of future cardiovascular disease is partly mediated by cadmium in tobacco: Malmö diet and cancer cohort study. Environ. Health18, 56 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Li, Y. et al. The brain structure and genetic mechanisms underlying the nonlinear association between sleep duration, cognition and mental health. Nat. Aging2, 425–437 (2022). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(1.8MB, pdf)}

Reporting Summary^{(2.9MB, pdf)}

Transparent Peer Review File^{(3MB, pdf)}

Source data^{(11.1MB, xlsx)}

Data Availability Statement

[CR1] 1.Moqri, M. et al. Biomarkers of aging for the identification and evaluation of longevity interventions. Cell186, 3758–3775 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Chang, A. Y., Skirbekk, V. F., Tyrovolas, S., Kassebaum, N. J. & Dieleman, J. L. Measuring population ageing: an analysis of the Global Burden of Disease Study 2017. Lancet Public Health4, e159–e167 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Ferrucci, L. et al. Measuring biological aging in humans: a quest. Aging Cell19, e13080 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Kothari, M. & Belsky, D. W. Unite to predict. Elife10, e66223 (2021). [DOI] [PMC free article] [PubMed]

[CR5] 5.Belsky, D. W. et al. Eleven telomere, epigenetic clock, and biomarker-composite quantifications of biological aging: do they measure the same thing? Am. J. Epidemiol.187, 1220–1230 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Li, X. et al. Longitudinal trajectories, correlations and mortality associations of nine biological ages across 20-years follow-up. Elife9, e51507 (2020). [DOI] [PMC free article] [PubMed]

[CR7] 7.Xiang, Y. et al. Tea consumption and attenuation of biological aging: a longitudinal analysis from two cohort studies. Lancet Reg. Health West Pac.42, 100955 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.WHO. Social determinants of health. https://www.who.int/health-topics/social-determinants-of-health#tab=tab_1 (2023).

[CR9] 9.Marmot, M., Allen, J., Boyce, T., Goldblatt, P. & Morrison, M. Health equity in England: the Marmot review ten years on (Institute of Health Equity, 2020).

[CR10] 10.Bor, J., Cohen, G. H. & Galea, S. Population health in an era of rising income inequality: USA, 1980-2015. Lancet389, 1475–1490 (2017). [DOI] [PubMed] [Google Scholar]

[CR11] 11.Bundy, J. D. et al. Social determinants of health and premature death among adults in the USA from 1999 to 2018: a national cohort study. Lancet Public Health8, e422–e431 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Zhong, J. et al. Associations of social determinants of health with life expectancy and future health risks among individuals with type 2 diabetes: two nationwide cohort studies in the UK and USA. Lancet Healthy Longev.5, e542–e551 (2024). [DOI] [PubMed] [Google Scholar]

[CR13] 13.Shah, N. S. et al. Social determinants of cardiovascular health in Asian Americans: a scientific statement from the American Heart Association. Circulation150, e296–e315 (2024). [DOI] [PubMed] [Google Scholar]

[CR14] 14.Hill-Briggs, F. et al. Social determinants of health and diabetes: a scientific review. Diabetes Care44, 258–279 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Potente, C. et al. Socioeconomic inequalities and molecular risk for aging in young adulthood. Am. J. Epidemiol.192, 1981–1990 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Cabeza de Baca, T. et al. Chronic psychosocial and financial burden accelerates 5-year telomere shortening: findings from the coronary artery risk development in young adults study. Mol. Psychiatry25, 1141–1153 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Steptoe, A. & Zaninotto, P. Lower socioeconomic status and the acceleration of aging: an outcome-wide analysis. Proc. Natl. Acad. Sci. USA117, 14911–14917 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Figueroa, J. F., Frakt, A. B. & Jha, A. K. Addressing social determinants of health: time for a polysocial risk score. JAMA323, 1553–1554 (2020). [DOI] [PubMed] [Google Scholar]

[CR19] 19.Jiang, M. et al. Accelerated biological aging elevates the risk of cardiometabolic multimorbidity and mortality. Nat. Cardiovasc. Res.3, 332–342 (2024). [DOI] [PubMed] [Google Scholar]

[CR20] 20.Kuo, C. L. et al. Proteomic aging clock (PAC) predicts age-related outcomes in middle-aged and older adults. Aging Cell23, e14195 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Maunakea, A. K. et al. Socioeconomic status, lifestyle, and DNA methylation age among racially and ethnically diverse adults: NIMHD Social Epigenomics Program. JAMA Netw. Open7, e2421889 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Shen, B. et al. Association of race and poverty status with DNA methylation-based age. JAMA Netw. Open6, e236340 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Liu, Z. et al. A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: a cohort study. PLoS Med.15, e1002718 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Chan, M. S. et al. A biomarker-based biological age in UK Biobank: composition and prediction of mortality and hospital admissions. J. Gerontol. A Biol. Sci. Med. Sci.76, 1295–1302 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Astuti, Y., Wardhana, A., Watkins, J. & Wulaningsih, W. Cigarette smoking and telomere length: a systematic review of 84 studies and meta-analysis. Environ. Res.158, 480–489 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Goeminne, L. J. E. et al. Plasma protein-based organ-specific aging and mortality models unveil diseases as accelerated aging of organismal systems. Cell Metab.37, 205-222.e6 (2024). [DOI] [PubMed]

[CR27] 27.Parker, D. C. et al. Association of blood chemistry quantifications of biological aging with disability and mortality in older adults. J. Gerontol. A Biol. Sci. Med. Sci.75, 1671–1679 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Graf, G. H. et al. Testing black-white disparities in biological aging among older adults in the United States: analysis of DNA-methylation and blood-chemistry methods. Am. J. Epidemiol.191, 613–625 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Graf, G. H. J. et al. Educational mobility, pace of aging, and lifespan among participants in the Framingham Heart Study. JAMA Netw. Open7, e240655 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Keyes, K. M. & Westreich, D. UK Biobank, big data, and the consequences of non-representativeness. Lancet393, 1297 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Fry, A. et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am. J. Epidemiol.186, 1026–1034 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Sudlow, C. et al. UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med.12, e1001779 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.National Center for Health Statistics. NHANES survey methods and analytic guidelines. https://wwwn.cdc.gov/Nchs/Nhanes/AnalyticGuidelines.aspx.

[CR34] 34.US Department of Health and Human Services. Social Determinants of Health. Healthy People 2030. https://odphp.health.gov/healthypeople/priority-areas/social-determinants-health (2023).

[CR35] 35.Marmot, M., Friel, S., Bell, R., Houweling, T. A. & Taylor, S. Closing the gap in a generation: health equity through action on the social determinants of health. Lancet372, 1661–1669 (2008). [DOI] [PubMed] [Google Scholar]

[CR36] 36.Zhu, R. et al. Prevalence of cardiovascular-kidney-metabolic syndrome stages by social determinants of health. JAMA Netw. Open7, e2445309 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Huang, H. et al. Association between social determinants of health and survival among the US cancer survivors population. BMC Med.22, 343 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Wang, F. et al. A systematic review and meta-analysis of 90 cohort studies of social isolation, loneliness and mortality. Nat. Hum. Behav.7, 1307–1319 (2023). [DOI] [PubMed] [Google Scholar]

[CR39] 39.Klemera, P. & Doubal, S. A new approach to the concept and computation of biological age. Mech. Ageing Dev.127, 240–248 (2006). [DOI] [PubMed] [Google Scholar]

[CR40] 40.Levine, M. E. Modeling the rate of senescence: can estimated biological age predict mortality more accurately than chronological age? J. Gerontol. A Biol. Sci. Med. Sci.68, 667–674 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Kuo, C. L., Pilling, L. C., Liu, Z., Atkins, J. L. & Levine, M. E. Genetic associations for two biological age measures point to distinct aging phenotypes. Aging Cell20, e13376 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Gao, X. et al. Accelerated biological aging and risk of depression and anxiety: evidence from 424,299 UK Biobank participants. Nat. Commun.14, 2277 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Kwon, D. & Belsky, D. W. A toolkit for quantification of biological age from blood chemistry and organ function test data: BioAge. Geroscience43, 2795–2808 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Hastings, W. J., Shalev, I. & Belsky, D. W. Comparability of biological aging measures in the National Health and Nutrition Examination Study, 1999–2002. Psychoneuroendocrinology106, 171–178 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Cui, F. et al. Early-life exposure to tobacco, genetic susceptibility, and accelerated biological aging in adulthood. Sci. Adv.10, eadl3747 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Bycroft, C. et al. The UK Biobank resource with deep phenotyping and genomic data. Nature562, 203–209 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Thompson, D. J. et al. A systematic evaluation of the performance and properties of the UK Biobank Polygenic Risk Score (PRS) Release. PLoS ONE19, e0307270 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Lambert, S. A. et al. The polygenic score catalog as an open database for reproducibility and systematic evaluation. Nat. Genet.53, 420–425 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.UK Biobank. Mortality data: linkage to death registries. https://biobank.ndph.ox.ac.uk/showcase/ukb/docs/DeathLinkage.pdf (2023).

[CR50] 50.National Center for Health Statistics. 2019 public-use linked mortality files. https://www.cdc.gov/nchs/data-linkage/mortality-public.htm (2022).

[CR51] 51.UK Biobank. First Occurrence of health outcomes defined by 3-character ICD10 code. https://biobank.ndph.ox.ac.uk/showcase/ukb/docs/first_occurrences_outcomes.pdf (2019).

[CR52] 52.Lange, T. & Hansen, J. V. Direct and indirect effects in a survival context. Epidemiology22, 575–581 (2011). [DOI] [PubMed] [Google Scholar]

[CR53] 53.Li, H. et al. Smoking-induced risk of future cardiovascular disease is partly mediated by cadmium in tobacco: Malmö diet and cancer cohort study. Environ. Health18, 56 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Li, Y. et al. The brain structure and genetic mechanisms underlying the nonlinear association between sleep duration, cognition and mental health. Nat. Aging2, 425–437 (2022). [DOI] [PubMed] [Google Scholar]

PERMALINK

Social determinants of health, accelerated biological aging, and long-term health outcomes

Jiang Li

Jie Li

Xiaoqin Xu

Wenqi Shen

Ying Sun

Yanqi Fu

Xiao Tan

Ningjian Wang

Lu Qi

Bin Wang

Yingli Lu

Abstract

Introduction

Fig. 1. Overview of the study design.

Results

Population characteristics

Table 1.

Association of combined SDHs with accelerated biological aging

Fig. 2. Distribution of accelerated biological aging across SDH scores.

Table 2.

Associations of combined SDHs with mortality and incident diseases

Fig. 3. Associations of combined SDHs with health outcomes.

Mediation analysis

Fig. 4. Mediating analyses of accelerated biological aging in the associations between combined SDHs and health outcomes.

Discussion

Methods

Study design and population

Assessment of SDHs

Assessment of biological age and age acceleration

Polygenic risk scores

Assessment of health outcomes

Covariates

Statistical analysis

Ethics statement

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Data availability

Code availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases