Needham Belinda L, Gladish Nicole, Shen Hanyang, Liu Yongmei, Smith Jennifer A, Mukherjee Bhramar, Zhou Xiang, Rehkopf David H
Department of Epidemiology, University of Michigan, Ann Arbor, Michigan, United States of America.
Department of Epidemiology and Population Health, Stanford University, Stanford, California, United States of America.
PLoS One. 2025 Jul 16;20(7):e0327010. doi: 10.1371/journal.pone.0327010. eCollection 2025.
Accelerated biological aging due to differences in socially patterned exposures has been proposed as a mechanism underlying racial and ethnic disparities in morbidity and mortality. Research exploring this hypothesis has been limited by a lack of consensus regarding the measurement of biological aging.
The goal of this study is to examine self-reported race and ethnicity as a predictor of 13 measures of epigenetic aging.
Data are from the National Health and Nutrition Examination Survey (1999-2002), a nationally representative study of US residents aged two months and older. The analytic sample includes 2,402 adults aged 50-84 with epigenetic data. The exposure is self-reported race and ethnicity, and the outcomes are 13 measures of epigenetic aging trained on different aging phenotypes.
In linear regression models controlling for age, age-squared, gender, and nativity, White respondents had higher epigenetic aging than Black respondents (the reference group) for six out of seven measures trained on chronological age (Hannum: b = 1.98, 95% CI = 1.43, 2.54; Horvath: b = 0.75, 95% CI = 0.09, 1.40; Weidner: b = 1.15, 95% CI = 0.30, 2.01; Vidal-Bralo: b = 2.30, 95% CI = 1.76, 2.84; SkinBlood: b = 0.85, 95% CI = 0.28, 1.43; Zhang: b = 0.58, 95% CI = 0.40, 0.76) and for one measure trained on telomere length (b = -0.17, 95% CI = -0.20, -0.14). In contrast, White respondents had lower epigenetic aging than Black respondents for three out of four measures trained on physiological age (GrimAge: b = -1.33, 95% CI = -2.01, -0.64; DunedinPoAm: b = -0.03, 95% CI = -0.04, -0.01; GrimAge2: b = -1.97, 95% CI = -2.74, -1.20) and for one measure trained on stem cell divisions (b = -0.01, 95% CI = -0.01, -0.01). Fewer differences in epigenetic aging were observed when comparing Mexican American, other Hispanic, and another race or ethnicity respondents to Black respondents.
White respondents had higher epigenetic aging than Black respondents for measures trained on chronological age, whereas the opposite was true for measures trained on physiological age. More work is needed to validate measures of epigenetic aging in non-White populations and to determine whether these measures are associated with health-related outcomes similarly across racial and ethnic groups.
社会模式化暴露差异导致的生物衰老加速被认为是发病率和死亡率方面种族和族裔差异的潜在机制。探索这一假设的研究因在生物衰老测量方面缺乏共识而受到限制。
本研究的目的是检验自我报告的种族和族裔作为13种表观遗传衰老测量指标的预测因素。
数据来自国家健康与营养检查调查（1999 - 2002年），这是一项针对美国两个月及以上居民的全国代表性研究。分析样本包括2402名年龄在50 - 84岁且有表观遗传数据的成年人。暴露因素是自我报告的种族和族裔，结果是基于不同衰老表型训练的13种表观遗传衰老测量指标。
在控制年龄、年龄平方、性别和出生地的线性回归模型中，对于基于实足年龄训练的七种测量指标中的六种，白人受访者的表观遗传衰老高于黑人受访者（参考组）（汉纳姆：b = 1.98，95%置信区间 = 1.43，2.54；霍瓦特：b = 0.75，95%置信区间 = 0.09, 1.40；魏德纳：b = 1.15，95%置信区间 = 0.30，2.01；维达尔 - 布拉洛：b = 2.30，95%置信区间 = 1.76，2.84；皮肤血液：b = 0.85，95%置信区间 = 0.28，1.43；张：b = 0.58，95%置信区间 = 0.40，0.76），以及对于基于端粒长度训练的一种测量指标（b = -0.17，95%置信区间 = -0.20，-0.14）。相比之下，对于基于生理年龄训练的四种测量指标中的三种，白人受访者的表观遗传衰老低于黑人受访者（格里姆年龄：b = -1.33，95%置信区间 = -2.01，-0.64；达尼丁PoAm：b = -0.03，95%置信区间 = -0.04，-0.01；格里姆年龄2：b = -1.97，95%置信区间 = -2.74，-1.20），以及对于基于干细胞分裂训练的一种测量指标（b = -0.01，95%置信区间 = -0.01，-0.01）。在将墨西哥裔美国人、其他西班牙裔以及其他种族或族裔受访者与黑人受访者进行比较时，观察到的表观遗传衰老差异较少。
对于基于实足年龄训练的测量指标，白人受访者的表观遗传衰老高于黑人受访者，而对于基于生理年龄训练的测量指标则相反。需要开展更多工作来验证非白人人群中表观遗传衰老测量指标，并确定这些指标在不同种族和族裔群体中与健康相关结局的关联是否相似。
