Depressive symptoms are associated with DNA methylation age acceleration in a cross-sectional analysis of adults over age 50 in the United States ================================================================================================================================================= * Herong Wang * Kelly M. Bakulski * Freida Blostein * Brittany R. Porath * John Dou * Cesar Higgins Tejera * Erin B. Ware ## Abstract **Background** Major depressive disorder affects mental well-being and accelerates DNA methylation age, a marker of biological aging. Subclinical depressive symptoms and DNA methylation aging have not been explored. **Objective** To assess the cross-sectional association between depressive symptoms and accelerated DNA methylation aging among United States adults over age 50. **Methods** We included 3,793 participants from the 2016 wave of the Health and Retirement Study. Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression scale and operationalized as high versus low/no. Blood DNA methylation GrimAge was regressed on chronologic age to obtain acceleration. Multiple linear regression assessed the relationship between high depressive symptoms and GrimAge acceleration, controlling for demographic factors, health behaviors, and cell type proportions. We investigated sex and race/ethnicity stratified associations. **Results** Participants were 42% male, 14% had high depressive symptoms, 44% had accelerated GrimAge, and were mean age 70 years. In our fully adjusted model, those with high depressive symptoms had 0.40 (95%CI: 0.06, 0.73) years accelerated GrimAge, compared to those with low/no depressive symptoms. The association between depressive symptoms and GrimAge acceleration was larger in male participants (*P* = 0.04). **Conclusion** Higher depressive symptoms were associated with accelerated DNA methylation age among older adults. Keywords * Epigenetic clock * DNA methylation age * epigenetic age acceleration * depressive symptoms * epidemiology * older adults ## Introduction Major depressive disorder is characterized by feelings of sadness, hopelessness, and discouragement, or the loss of pleasure or interest in activities for more than 2 weeks1. In the United States, approximately 9% of adults aged 65 years and older have a diagnosis of major depressive disorder2. Also in the United States, nearly 18% of adults aged 65 and over experience depressive symptoms each year3, and older adults experience more depressive symptoms than middle-aged adults4. High levels of depressive symptoms are associated with greater mortality risk and increased burden of medical conditions5,6. Additionally, depressive symptomatology is associated with poorer social functioning among older people compared to chronic medical conditions such as lung disease, hypertension or diabetes7. Thus, understanding the multiple effects of depressive symptoms in vulnerable subgroups of the population is of public health importance. There are existing research documenting disparities in depressive symptoms across sex and among racialized social groups8. Specifically in the United States, higher levels of depressive symptoms are reported by women relative to men, as well as Black, Hispanic, and Asian individuals relative to White individuals8–10. Therefore, understanding underlying molecular mechanisms between depressive symptoms and the aging process and how sex and race/ethnicity may interact with depressive symptoms is of public health importance. DNA methylation-based biomarkers are one way to measure biological aging. DNA methylation most commonly occurs at cytosine nucleotides upstream of guanine nucleotides, termed CpG sites11,12. Levels of DNA methylation at sets of CpG sites and traits are used to train DNA methylation age estimators11. When a person’s estimated DNA methylation age is higher than their chronological age, accelerated DNA methylation age is said to occur. GrimAge is a DNA methylation age estimator trained to predict lifespan13. GrimAge predicts all-cause mortality and showed strong association with numerous age-related conditions14 and is therefore likely to be sensitive for assessing biological aging related to psychiatric conditions15. However, there are few publications investigating the association of DNA methylation age acceleration with depressive symptoms using GrimAge. DNA methylation age acceleration has been observed in middle-age adults with major depressive disorder. Among 1,130 Dutch participants aged 41 on average, major depressive disorder was associated with 0.64 years accelerated blood DNA methylation age, using a custom estimator16. Similarly, among 109 participants aged 40 on average, major depressive disorder was associated 2 years accelerated GrimAge measured in blood15. These studies in younger age participants were limited however by small sample size and strict diagnosis of the condition. The role of more prevalent sub-clinical depressive symptoms on biological aging among more vulnerable and diverse samples of older adults is less clear. In light of the insufficient evidence on DNA methylation aging related to depressive symptoms among older adults, we examined the association between depressive symptoms and GrimAge age acceleration in a large and diverse cross-sectional sample (n=3,793) of the United States Health and Retirement Study. We also assessed the effect modification of race/ethnicity and sex, and provided race/ethnicity and sex-specific estimations, given that the lack of evidence on the disparities across sex and among racialized social groups in the association between depressive symptoms and biological aging. ## Methods ### Study Sample The Health and Retirement Study is an ongoing, nationally-representative, panel study of Americans over 50 years old17. The cohort began in 1992 with surveys occurring every two years. The Health and Retirement Study uses a multi-stage area probability design with geographical stratification and oversampling of African-American and Hispanic households. During the 2016 wave, the Health and Retirement Study collected venous blood samples for DNA methylation measures18, thus we selected this wave for our cross-sectional analyses. The Health and Retirement Study data is sponsored by the National Institute on Aging (grant number U01AG009740) and is conducted by the University of Michigan. This secondary analysis was approved by the University of Michigan Institutional Review Board (HUM00128220). Participants provided written informed consent. Data are publicly available at [https://hrsdata.isr.umich.edu/data-products](https://hrsdata.isr.umich.edu/data-products). ### Depressive Symptom Measures Depressive symptoms were assessed by self-report on the modified 8-item Center for Epidemiologic Studies Depression (CES-D) scale19. The 8-item version of the CES-D scale is consistent, validated, and reliable in measuring depressive symptoms in older adults20,21. Individual item responses were summed for a final score ranging from 0 to 8. For our primary analyses, we created a binary depressive symptoms variable, indicating whether participants had high (4-8 symptoms) or low/no depressive symptoms (0-3 symptoms), consistent with previous studies22. In sensitivity analyses, we considered continuous depressive symptoms and an alternatively dichotomized CES-D score (no symptoms versus 1 or more). ### DNA Methylation Age Acceleration Measures Venous blood was collected by a trained phlebotomist, using a whole blood EDTA vacutainer blood tube18. Samples were shipped overnight to the University of Minnesota Advanced Research and Diagnostic Laboratory and processed within 48 hours of collection12. DNA methylation levels were quantified using the Infinium MethylationEPIC BeadChip (Illumina Inc.) for 4,104 samples randomized across plates by key demographic variables including age, cohort, sex, education, and race/ethnicity. Measures from 86 samples were excluded during quality control assessments12. DNA methylation age estimates were calculated and released by the Health and Retirement Study12. Our primary DNA methylation age estimate was GrimAge. GrimAge is trained on health-related and disease-related indicators like sex, education, race, and smoking and is therefore likely to be a stronger and more sensitive predictor for all-causes mortality, age-related health status and biological aging in psychiatric conditions13–15. To create the age acceleration values, we regressed the DNA methylation age estimate against chronological age and captured the residuals from this model13,23. We modelled these residuals as continuous in linear regressions and for bivariate analyses and we dichotomized the residuals into accelerated aging (residuals > 0) or no age acceleration (residuals ≤ 0). In sensitivity analyses, we used four additional DNA methylation age estimates calculated and released by the Health and Retirement Study. These additional age estimates included Horvath, Hannum, Levine, and DunedinPoAm38 (MPOA). Horvath and Hannum were selected as DNA methylation estimators of chronologic age. The Horvath age estimator was trained on DNA methylation at 353 CpG sites across 51 different tissues and cell types24. The Hannum age estimator was trained on 71 CpGs selected in blood to capture changes in chronological age25. Levine and MPOA were selected as measures of phenotypic aging. The Levine age estimator was trained on 513 CpGs to capture the combination of age-related and disease phenotypes26. The MPOA algorithm was trained to predict the rate of aging27 and we transformed the MPOA rates into years for alignment with the other age estimators. We regressed each DNA methylation age estimate on chronological age and took the residuals from the linear model to obtain age acceleration values. Residuals were modelled as continuous in linear regressions for use in sensitivity analysis. ### Covariate Measures Sociodemographic information was provided through self-reported baseline interviews including sex (male/female), race/ethnicity (non-Hispanic White, non-Hispanic Black or Other, and Hispanic), childhood socioeconomic status (financial situation before age 16 of well off financially, about average, poor, varied), and education (no degree, GED/high school diploma, and college or more than college). Interviews at the 2016 wave provided age (years), marital status (currently married/not currently married), alcohol use (number of days per week consuming alcohol), cigarette smoking (never, former, current), number of chronic health conditions, and physical activity (levels of energy expended during exercise weighted by Metabolic Equivalent of Task (METs)) (Supplemental Methods). The covariate dataset used for this study was from the RAND Center for the Study of Aging with funding from the National Institute on Aging and the Social Security Administration. Santa Monica, CA produced (July 2022)28. Cell type composition is an important precision variable for DNA methylation analyses29. Percentages of granulocytes, monocytes, and lymphocytes were assessed from a complete blood count using a Sysmex XE-21000 instrument18 and were released by the Health and Retirement Study with the DNA methylation age estimates12. ### Statistical Methods The included analytic sample was restricted to participants who had complete data for depressive symptoms, DNA methylation age, and all covariates. We visualized sample inclusion and exclusion using a flow chart. We assessed the sample distributions of continuous variables using mean and standard deviation and for categorical variables using count and frequency. To assess the possibility of selection bias, included participants were compared to participants excluded due to missing variables using Wilcoxon rank sum test for continuous variables and Pearson’s chi-squared test for categorical variables. We calculated two-sided *P*-values from these tests at an alpha of 0.05 in all models. In the included sample, we examined the bivariate characteristics similarly based on whether the participant had high or low/no depressive symptoms. We also calculated bivariate characteristics based on dichotomous GrimAge age acceleration. To check the validity of age-related variables, we examined the pairwise relationships between chronological age, GrimAge, GrimAge age acceleration, and the difference between chronological age and GrimAge. We created scatterplots to visualize these relationships and calculated Pearson correlation coefficients for each pairing. We constructed three linear regression models to test the association between high depressive symptoms and GrimAge age acceleration. The initial unadjusted model included the binary variable for high depressive symptoms as the exposure and GrimAge age acceleration as the outcome. In the primary adjusted model, we included chronological age, sex, race/ethnicity, highest degree obtained, marital status, percentage of granulocytes, and percentage of monocytes as predictors. We removed percent of lymphocytes from the models to avoid collinearity. In a health behavior adjusted model, we added the number of days per week of alcohol use, smoking status, and number of chronic health conditions to the primary adjusted model. To evaluate effect modification by race/ethnicity and sex, we applied separate primary and health behavior adjusted models in strata of race/ethnicity and sex and pooled the estimates using generic inverse variance meta-analyses. To examine the significance of the effect modification, we included product terms between depressive symptoms and race/ethnicity and sex in the models. We assessed heterogeneity of associations across strata of race/ethnicity and sex using Cochran’s Q test. If we observed heterogeneity, we used stratification models and reported the estimates in each stratum. Otherwise, we reported the pooled estimates. We performed all statistical analyses using R version 4.0.3. Code to perform all analyses is provided ([https://github.com/bakulskilab](https://github.com/bakulskilab)). ### Sensitivity Analyses We conducted several sensitivity analyses to assess the robustness of our results. For each sensitivity analysis, we produced an unadjusted model, a primary adjusted model, and a health behavior adjusted model using the same covariates for each model as described above. First, to assess the relationship between the presence of any depressive symptoms and GrimAge age acceleration, we tested the association between an indicator for any depressive symptoms and GrimAge age acceleration. Second, to assess whether depressive symptoms had a linear relationship with accelerated DNA methylation aging, we modelled depressive symptoms as a continuous variable. To determine whether the relationship between depressive symptoms and accelerated DNA methylation aging was specific to the GrimAge estimator, we examined four alternate DNA methylation age estimators including Horvath, Hannum, Levine, and MPOA. Fourth, we examined whether the results differed after further adjustment for physical activity and financial situation before age 16. Finally, we excluded participants older than 85 to assess the robustness of the linear regression model we used to derive GrimAge age acceleration. ## Results ### Descriptive statistics There were 4,018 eligible participants for our study and the final analytic sample consisted of 3,793 participants with complete data for all analyses (**Figure 1**). Included participants were similar to excluded participants in chronological age, GrimAge, Horvath, Levine, MPOA, and Hannum age (**Supplemental Table 1**); however, included participants were more likely to be non-Hispanic White, married, have a GED/high school diploma as their highest educational attainment, smoke currently, consume alcohol more days per week and have less chronic health condition compared to excluded participants. ![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2023/04/25/2023.04.24.23289052/F1.medium.gif) [Figure 1.](http://medrxiv.org/content/early/2023/04/25/2023.04.24.23289052/F1) Figure 1. Flow diagram of inclusion and exclusion of participants from the United States Health and Retirement Study 2016 wave who had available DNA methylation data from the 2016 Venous Blood Study. Note: N is the number of observations included. n is the number of observations excluded. CES-D is score on the modified Center for Epidemiologic Studies Depression (CES-D) scale; HRS Health and Retirement Study Among the analytic sample, 544 (14.4%) of participants had high levels of depressive symptoms based on CES-D score (**Table 1**). Those with high depressive symptoms were significantly more likely to be younger, have higher GrimAge age acceleration, be female, report a race/ethnicity other than non-Hispanic White participants, be single, have more chronic health conditions, be current or former smokers, and have a lower level of physical activity than those with low/no reported depressive symptoms. View this table: [Table 1.](http://medrxiv.org/content/early/2023/04/25/2023.04.24.23289052/T1) Table 1. Characteristics of the included analytic sample of the 2016 Health and Retirement Study Venous Blood Study participants with DNA methylation data by level of depressive symptoms and GrimAge age acceleration (n=3,793) We present the distribution of GrimAge age acceleration as were all other age acceleration variables in **Supplemental Figure 1**. GrimAge and chronological age were positively correlated in the analytic sample (Pearson rho = 0.83). We verified the effect of chronologic age was successfully removed from our measure of GrimAge acceleration by plotting GrimAge acceleration against chronological age (Pearson rho=0.002, **Supplemental Figure 2**). In the analytic sample, 44% of participants had GrimAge acceleration greater than 0 (**Table 1**). Prevalence of high depressive symptoms was higher among those with accelerated GrimAge (17%) relative to those without accelerated GrimAge (12%) (*P* < 0.01). Participants with accelerated GrimAge were more likely to be male, non-Hispanic Black or Other race/ethnicity, and have lower educational attainment. These participants also had more chronic health conditions and higher levels of drinking compared to those without accelerated GrimAge. ### Multivariable associations between depressive symptoms and DNA methylation age acceleration We performed linear regression models to test whether high depressive symptoms are associated with GrimAge acceleration (**Table 2**). In the unadjusted model, those with high depressive symptoms had 1.29 (95% CI: 0.86, 1.72) years higher GrimAge acceleration on average compared to those with low/no depressive symptoms. The strength of the association was attenuated after controlling for demographic factors and cell-type composition (*β* = 1.15 years, 95% CI: 0.77, 1.54). We observed further attenuation after additional adjustment for health behaviors (*β*= 0.40 years, 95% CI: 0.06, 0.73). Model fit based on adjusted R2 value and likelihood ratio test both improved significantly after adjustment for these factors. View this table: [Table 2.](http://medrxiv.org/content/early/2023/04/25/2023.04.24.23289052/T2) Table 2. Linear regression models predicting GrimAge acceleration based on presence of high depressive symptoms in the analytic sample of the 2016 Health and Retirement Study Venous Blood Study participants with DNA methylation data (n=3,792) ### Stratified sample regression results After adjusting for sociodemographic covariates and cell type composition in primary model, the association between depressive symptoms and GrimAge acceleration was larger in male participants (*P* = 0.04) (**Supplemental Table 5**). While we found significant effect modification of the association between depressive symptoms and GrimAge age acceleration by sex in primary models, we did not observe effect modification by sex when adding the health behaviour covariates. Nor did we observe effect modification of the relationship between depressive symptoms and GrimAge age acceleration by self-reported race/ethnicity. We show these results in **Figure 2**. Therefore, we do not report stratified models or explore them further in sensitivity models. ![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2023/04/25/2023.04.24.23289052/F2.medium.gif) [Figure 2.](http://medrxiv.org/content/early/2023/04/25/2023.04.24.23289052/F2) Figure 2. Forest plots for the meta-analysis of race/ethnicity and sex specific regression models of the association between high depressive symptoms (>4) and GrimAge age acceleration among the analytic sample of the 2106 Health and Retirement Study Venous Blood Study participants with DNA methylation data (n = 3,793) Abbreviation: CI, confidence interval. The heterogeneity test and *P*-value were performed using Cochran’s Q test A The forest plot for race/ethnicity and sex stratified effect estimates and 95% confidence interval in primary models adjusted for age, gender, highest degree obtained, marital status, percentage of granulocytes, and percentage of monocytes B The forest plot for race/ethnicity and sex stratified estimates and 95% confidence interval in health behavior models adjusted for all covariates in the primary model and number of chronic health conditions, number of days per week with alcohol consumption, and smoking status ### Sensitivity analysis results In sensitivity analyses examining the parameterization of depressive symptoms, the association between the depressive symptom characterization and GrimAge acceleration remained significant for both 1) any versus no depressive symptoms and for 2) the continuous depressive symptoms measure (**Supplemental Table 2**). When examining other DNA methylation age estimators, the presence of high depressive symptoms was not associated with DNA methylation age acceleration for the Horvath or Hannum clock in any models. The presence of high depressive symptoms was significantly associated with Levine (*β* = 0.73 years, 95% CI: 0.12, 1.34) and MPOA (*β* = 0.93 years, 95% CI: 0.39, 1.47) acceleration in primary models but not in health behavior adjusted models (**Supplemental Table 3**). Smaller and nonsignificant association between high level of depressive symptoms and GrimAge acceleration was observed (0.26, 95% CI: -0.08, 0.59) after further adjustment for physical activity and financial situation before age 16 (**Table 2**) The association between depressive symptoms and GrimAge acceleration in the health behavior model was robust to the exclusion of participants older than 85 (**Supplemental Table 4**). ## Discussion This cross-sectional study examined the relationship between depressive symptoms and accelerated DNA methylation age among a large, diverse sample of Americans over 50 years old. Among 3,793 participants, the presence of high depressive symptoms was associated with accelerated DNA methylation aging, measured using GrimAge, after adjusting for demographic factors, health status, and health behaviors. We found the association between depressive symptoms and accelerated DNA methylation aging to be robust to different parameterization of depressive symptoms and to other second-generation DNA methylation age estimates, like Levine and MPOA. We ultimately found no effect modification by sex or self-reported race/ethnicity in the association between depressive symptoms and GrimAge acceleration in the fully adjusted model. We investigated depressive symptoms as an exposure to capture the effects of potential preclinical indicators of depressive disorder on DNA methylation aging. Depressive symptoms affect a larger share of the population than clinically diagnostic forms of major depressive disorder. Previous studies have generally focused on major depressive disorder. Existing studies have found varying effect sizes for the relationship between major depressive disorder and accelerated DNA methylation age. These effects range from an average DNA methylation age estimate of 0.64 years higher for those with major depressive disorder than controls16, to around a 2 year increased median GrimAge acceleration for those with major depressive disorder compared to controls15. These samples were from middle-aged participants with a mean age around 40 years15,16. Additionally, postmortem brain tissue for those with major depressive disorder has shown an average increase of 1.11 years for the Horvath DNA methylation age acceleration compared to controls16. Our observation of an increase of 0.40 years in GrimAge DNA methylation age acceleration for those with high depressive symptoms is in line with these results. Major depressive disorder involves an increased burden of depressive symptoms for an extended period that interfere with daily life and represent a more severe phenotype than our dichotomy of depressive symptom counts. In sensitivity analyses, we also observed a significant, linear association between depressive symptoms treated as a continuous variable and accelerated GrimAge DNA methylation where each additional depressive symptom increased the GrimAge age acceleration by 0.08 years. We again would expect a decreased effect size for an increase of individual depressive symptoms compared to a clinical diagnosis of major depressive disorder, especially in older adults where there may be a greater accumulation of environmental effects on DNA methylation age acceleration, which may reduce the individual effect of depressive symptoms. In our sensitivity analyses, a continuous measure of depressive symptoms (CES-D score) was associated with accelerated GrimAge, after full adjustment. Since the majority of studies in this area focus on major depressive disorder and/or use clocks other than GrimAge, this is a novel finding. These results point to a small dose-response relationship between depressive symptoms and GrimAge acceleration. We performed a second sensitivity analysis to assess the cut-point we selected for determining the presence of depressive symptoms. Any depressive symptoms remained associated with accelerated GrimAge, with an effect size only slightly smaller when compared to the effect size of high depressive symptoms (≥4) on accelerated GrimAge. This suggests that having any level of depressive symptoms may be associated with GrimAge acceleration and that depressive symptoms may be association linearly with increased DNA methylation age acceleration. Consistent with previous literature, we found that depressive symptoms were associated with “second generation” DNA methylation aging measures, whereas we found no association for the “first generation” DNA methylation aging measures. These second-generation measures of phenotypic aging such as GrimAge, Levine and MPOA, are trained on health indicators such as sex, education, race, and smoking and more highly associated with age-related disease, and thus is likely to be sensitive to psychiatric disorders while measures of chronological aging such as Horvath and Hannum are not. In prior studies, GrimAge was a strong predictor of time-to multiple outcomes, including death, coronary heart disease, and cancer13. When examining depressive symptomology among a cohort of 724 Danish monozygotic twins, depressive symptomology score was not associated with Horvath age acceleration, after adjustment for chronological age30. These results are consistent with the sensitivity analysis we conducted using the Horvath clock, where we did not observe an association between depressive symptoms and accelerated DNA methylation aging. The biological mechanism through which depressive symptoms may lead to accelerated DNA methylation age is unknown. Chronic diseases such as major depressive disorder could be prolonged stress exposures, and researchers found chronic stress have multiple physical and epigenetic effects throughout the body31,32. Increases in depressive symptoms could also result in behavioral changes, such as smoking, less physical activity, and dietary changes33–36 which are risk factors for epigenetic effects37–39 and likely to mediate depressive symptoms in DNA methylation aging. When controlling for some such potential health behaviors, we did observe an attenuation of the association, but a significant statistical association remained. Prolonged exposure and increased severity of depressive symptoms may produce chronic stress throughout the body that affects DNA methylation and other cellular characteristics. This study has several limitations that should be considered when interpreting these results. The cross-sectional nature of this analysis prevents us from establishing that depressive symptoms precede DNA methylation age acceleration, and future work in longitudinal settings is necessary. In addition, more research is needed to understand how the cumulative experience of depressive symptoms across an individual’s lifetime may accelerate DNA methylation aging. Tissue specificity of DNA methylation signatures may be a concern in considering these results. We used DNA methylation age measured from whole blood, though the brain is the preferred target tissue for these measures when looking at depressive symptoms40. However, blood DNA methylation may correlate with brain DNA methylation and whole blood is useful in cases where brain tissue would not be available, such as when using large, population-based studies of living participants41. Future research into this area could focus on longitudinal designs while controlling for additional confounders and potentially investigating DNA methylation in disease-specific tissue. Finally, for the purposes of this analysis, we did not use sampling weights, so our results are only generalizable to people over 50 years old living in the United States with similar demographic characteristics to the sample. Future research should investigate whether this association is robust across different scales of measuring depressive symptoms and in other populations. We were able to interrogate a large number of participants in this study due to the availability of our measures of interest in a population-representative study. Additionally, we were able to adjust for cell type proportions in our analysis, an important precision variable in measuring DNA methylation aging. Our focus on depressive symptoms rather than major depressive disorder diagnosis makes these results more generalizable to those experiencing varying levels and severity of depressive symptoms. We also provided race/ethnicity and sex specific estimates, providing insight on race/ethnicity and sex differences in the association between depressive symptoms and DNA methylation age acceleration. Our use of the GrimAge clock is another strength of this study, since GrimAge predicts lifespan13. People with high levels of depressive symptoms typically have higher mortality and more comorbidities, so the GrimAge DNA methylation age estimate is particularly relevant in estimating the potential effect of depressive symptoms5,6. Due to these strengths, our study can provide some continuing evidence for an association between depressive symptoms and accelerated DNA methylation age. In summary, in this cross-sectional analysis we observed high depressive symptoms were associated with accelerated GrimAge among participants in the Health and Retirement Study. Future studies could try to replicate these findings in other populations and establish a temporal relationship between experiencing depressive symptoms and GrimAge acceleration for the purpose of understanding the effects of depressive symptoms over the life course. Depressive symptoms have likely increased since the beginning of the COVID-19 pandemic, with one study indicating a prevalence of 8.5% among adults in the United States pre-pandemic and a prevalence of 27.8% in April 202042. As depressive symptoms increase in the population, it may be important to study their long-term impact on epigenetic effects. Identifying risk groups, like those with high depressive symptoms, for biological correlates that may lead to earlier mortality could aid in creating intervention targets and highlight the impact of improving mental health management to help reduce early mortality. ## Data Availability All data produced in the present study are available upon reasonable request to the authors ## Funding The HRS (Health and Retirement Study) is sponsored by the National Institute on Aging (NIA U01AG009740) and is conducted by the University of Michigan. These analyses were supported by the National Institutes on Aging (R01 AG067592, R01 AG055406, P30 AG072931). * Received April 24, 2023. * Revision received April 24, 2023. * Accepted April 25, 2023. * © 2023, Posted by Cold Spring Harbor Laboratory The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission. ## References 1. 1.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Association; 2013. doi:10.1176/appi.books.9780890425596 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.books.9780890425596&link_type=DOI) 2. 2.Hasin DS, Goodwin RD, Stinson FS, Grant BF. Epidemiology of Major Depressive Disorder: Results From the National Epidemiologic Survey on Alcoholism and Related Conditions. Arch Gen Psychiatry. 2005;62(10):1097–1106. doi:10.1001/archpsyc.62.10.1097 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/archpsyc.62.10.1097&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16203955&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000232430600005&link_type=ISI) 3. 3.Villarroel MA, Terlizzi EP. Symptoms of Depression Among Adults: United States, 2019. NCHS Data Brief Natl Cent Health Stat. 2020;No. 379. [https://www.cdc.gov/nchs/products/databriefs/db379.htm](https://www.cdc.gov/nchs/products/databriefs/db379.htm) 4. 4.Sutin AR, Terracciano A, Milaneschi Y, An Y, Ferrucci L, Zonderman AB. The Trajectory of Depressive Symptoms Across the Adult Lifespan. JAMA Psychiatry Chic Ill. 2013;70(8):803–811. doi:10.1001/jamapsychiatry.2013.193 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jamapsychiatry.2013.193&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23760442&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 5. 5.Young JQ, Kline-Simon AH, Mordecai DJ, Weisner C. Prevalence of behavioral health disorders and associated chronic disease burden in a commercially insured health system: findings of a case–control study. Gen Hosp Psychiatry. 2015;37(2):101–108. doi:10.1016/j.genhosppsych.2014.12.005 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.genhosppsych.2014.12.005&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25578791&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 6. 6.Cuijpers P, Smit F. Excess mortality in depression: a meta-analysis of community studies. J Affect Disord. 2002;72(3):227–236. doi:10.1016/S0165-0327(01)00413-X [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0165-0327(01)00413-X&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=12450639&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000180471100002&link_type=ISI) 7. 7.World Health Organization. Mental health of older adults. Published December 12, 2017. Accessed March 29, 2023. [https://www.who.int/news-room/fact-sheets/detail/mental-health-of-older-adults](https://www.who.int/news-room/fact-sheets/detail/mental-health-of-older-adults) 8. 8.Hargrove TW, Halpern CT, Gaydosh L, et al. Race/Ethnicity, Gender, and Trajectories of Depressive Symptoms across Early- and Mid-Life among the Add Health Cohort. J Racial Ethn Health Disparities. 2020;7(4):619–629. doi:10.1007/s40615-019-00692-8 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s40615-019-00692-8&link_type=DOI) 9. 9.Mcleod JD, Owens TJ. Psychological Well-Being in the Early Life Course: Variations by Socioeconomic Status, Gender, and Race/Ethnicity. Soc Psychol Q. 2004;67(3):257–278. doi:10.1177/019027250406700303 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/019027250406700303&link_type=DOI) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000227847700003&link_type=ISI) 10. 10.Ross CE. Social Causes of Psychological Distress. 2nd Edition. Routledge; 2017. [https://doi.org/10.4324/9781315129464](https://doi.org/10.4324/9781315129464) 11. 11.Horvath S, Raj K. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet. 2018;19(6):371–384. doi:10.1038/s41576-018-0004-3 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41576-018-0004-3&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29643443&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 12. 12.Crimmins E, Kim JK, Fisher J, Faul J. HRS Epigenetic Clocks. Surv Res Cent Univ Mich. Published online September 2020. doi:[https://hrsdata.isr.umich.edu/sites/default/files/documentation/data-descriptions/EPICLOCKS\_DD.pdf](https://hrsdata.isr.umich.edu/sites/default/files/documentation/data-descriptions/EPICLOCKS_DD.pdf) 13. 13.Lu AT, Quach A, Wilson JG, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging. 2019;11(2):303–327. doi:10.18632/aging.101684 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.18632/aging.101684&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 14. 14.Hillary RF, Stevenson AJ, McCartney DL, et al. Epigenetic measures of ageing predict the prevalence and incidence of leading causes of death and disease burden. Clin Epigenetics. 2020;12(1):115. doi:10.1186/s13148-020-00905-6 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/s13148-020-00905-6&link_type=DOI) 15. 15.Protsenko E, Yang R, Nier B, et al. “GrimAge,” an epigenetic predictor of mortality, is accelerated in major depressive disorder. Transl Psychiatry. 2021;11(1):1–9. doi:10.1038/s41398-021-01302-0 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41398-021-01302-0&link_type=DOI) 16. 16.Han LKM, Aghajani M, Clark SL, et al. Epigenetic Aging in Major Depressive Disorder. Am J Psychiatry. 2018;175(8):774–782. doi:10.1176/appi.ajp.2018.17060595 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.2018.17060595&link_type=DOI) 17. 17.Juster FT, Suzman R. An Overview of the Health and Retirement Study. J Hum Resour. 1995;30:S7–S56. doi:10.2307/146277 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.2307/146277&link_type=DOI) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1995TP88600002&link_type=ISI) 18. 18.Crimmins E, Faul J, Thyagarajan B, Weir D. Venous blood collection and assay protocol in the 2016 Health and Retirement Study 2016 Venous Blood Study (VBS). Published online December 2017. doi:[https://hrsdata.isr.umich.edu/sites/default/files/documentation/data-descriptions/HRS2016VBSDD.pdf](https://hrsdata.isr.umich.edu/sites/default/files/documentation/data-descriptions/HRS2016VBSDD.pdf) 19. 19.Radloff LS. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Appl Psychol Meas. 1977;1(3):385–401. doi:10.1177/014662167700100306 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/014662167700100306&link_type=DOI) 20. 20.Karim J, Weisz R, Bibi Z, ur Rehman S. Validation of the Eight-Item Center for Epidemiologic Studies Depression Scale (CES-D) Among Older Adults. Curr Psychol. 2015;34(4):681–692. doi:10.1007/s12144-014-9281-y [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s12144-014-9281-y&link_type=DOI) 21. 21.Andresen EM, Malmgren JA, Carter WB, Patrick DL. Screening for Depression in Well Older Adults: Evaluation of a Short Form of the CES-D. Am J Prev Med. 1994;10(2):77–84. doi:10.1016/S0749-3797(18)30622-6 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0749-3797(18)30622-6&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=8037935&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1994NL05500003&link_type=ISI) 22. 22.Powell VD, Abedini NC, Galecki AT, Kabeto M, Kumar N, Silveira MJ. Unwelcome Companions: Loneliness Associates with the Cluster of Pain, Fatigue, and Depression in Older Adults. Gerontol Geriatr Med. 2021;7:2333721421997620. doi:10.1177/2333721421997620 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/2333721421997620&link_type=DOI) 23. 23.Krieger N, Chen JT, Testa C, et al. Use of Correct and Incorrect Methods of Accounting for Age in Studies of Epigenetic Accelerated Aging: Implications and Recommendations for Best Practices. Am J Epidemiol. Published online January 30, 2023:kwad025. doi:10.1093/aje/kwad025 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/aje/kwad025&link_type=DOI) 24. 24.Horvath S. DNA methylation age of human tissues and cell types. Genome Biol. 2013;14(10):3156. doi:10.1186/gb-2013-14-10-r115 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/gb-2013-14-10-r115&link_type=DOI) 25. 25.Hannum G, Guinney J, Zhao L, et al. Genome-wide Methylation Profiles Reveal Quantitative Views of Human Aging Rates. Mol Cell. 2013;49(2):359–367. doi:10.1016/j.molcel.2012.10.016 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.molcel.2012.10.016&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23177740&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000314379400018&link_type=ISI) 26. 26.Levine ME, Lu AT, Quach A, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging. 2018;10(4):573–591. doi:10.18632/aging.101414 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.18632/aging.101414&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29676998&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 27. 27.1. Hagg S, 2. Tyler JK, 3. Hagg S, 4. Justice J, 5. Suderman M Belsky DW, Caspi A, Arseneault L, et al. Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm. Hagg S, Tyler JK, Hagg S, Justice J, Suderman M, eds. eLife. 2020;9:e54870. doi:10.7554/eLife.54870 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.7554/eLife.54870&link_type=DOI) 28. 28.Bugliari D, Carroll J, Hayden O, et al. RAND HRS Longitudinal File 2018 (V2) Documentation: Includes 1992-2018 (Final Release). RAND Corporation; 2022. doi:10.7249/TLA2097-1 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.7249/TLA2097-1&link_type=DOI) 29. 29.Campbell KA, Colacino JA, Park SK, Bakulski KM. Cell Types in Environmental Epigenetic Studies: Biological and Epidemiological Frameworks. Curr Environ Health Rep. 2020;7(3):185–197. doi:10.1007/s40572-020-00287-0 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s40572-020-00287-0&link_type=DOI) 30. 30.Starnawska A, Tan Q, Soerensen M, et al. Epigenome-wide association study of depression symptomatology in elderly monozygotic twins. Transl Psychiatry. 2019;9(1):214. doi:10.1038/s41398-019-0548-9 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41398-019-0548-9&link_type=DOI) 31. 31.Ridout KK, Ridout SJ, Price LH, Sen S, Tyrka AR. Depression and telomere length: A meta-analysis. J Affect Disord. 2016;191:237–247. doi:10.1016/j.jad.2015.11.052 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.jad.2015.11.052&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26688493&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 32. 32.McEwen BS. Neurobiological and Systemic Effects of Chronic Stress. Chronic Stress. 2017;1:2470547017692328. doi:10.1177/2470547017692328 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/2470547017692328&link_type=DOI) 33. 33.Boden JM, Fergusson DM, Horwood LJ. Cigarette smoking and depression: tests of causal linkages using a longitudinal birth cohort. Br J Psychiatry J Ment Sci. 2010;196(6):440–446. doi:10.1192/bjp.bp.109.065912 [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTA6ImJqcHJjcHN5Y2giO3M6NToicmVzaWQiO3M6OToiMTk2LzYvNDQwIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjMvMDQvMjUvMjAyMy4wNC4yNC4yMzI4OTA1Mi5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 34. 34.Jacka FN, Pasco JA, Mykletun A, et al. Association of Western and traditional diets with depression and anxiety in women. Am J Psychiatry. 2010;167(3):305–311. doi:10.1176/appi.ajp.2009.09060881 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.2009.09060881&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=20048020&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000275056300012&link_type=ISI) 35. 35.Jerstad SJ, Boutelle KN, Ness KK, Stice E. Prospective Reciprocal Relations between Physical Activity and Depression in Adolescent Females. J Consult Clin Psychol. 2010;78(2):268–272. doi:10.1037/a0018793 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1037/a0018793&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=20350037&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 36. 36.McManus, Sally, Meltzer, Howard, Campion, Jonathan. Cigarette smoking and mental health in England: Data from the Adult Psychiatric Morbidity Survey 2007. Lond UK Natl Cent Soc Res. Published online 2010. 37. 37.Kim Y, Huan T, Joehanes R, et al. Higher diet quality relates to decelerated epigenetic aging. Am J Clin Nutr. 2022;115(1):163–170. doi:10.1093/ajcn/nqab201 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ajcn/nqab201&link_type=DOI) 38. 38.Klopack ET, Carroll JE, Cole SW, Seeman TE, Crimmins EM. Lifetime exposure to smoking, epigenetic aging, and morbidity and mortality in older adults. Clin Epigenetics. 2022;14(1):72. doi:10.1186/s13148-022-01286-8 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/s13148-022-01286-8&link_type=DOI) 39. 39.Kresovich JK, Garval EL, Martinez Lopez AM, et al. Associations of Body Composition and Physical Activity Level With Multiple Measures of Epigenetic Age Acceleration. Am J Epidemiol. 2021;190(6):984–993. doi:10.1093/aje/kwaa251 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/aje/kwaa251&link_type=DOI) 40. 40.Li Z, He Y, Ma X, Chen X. Epigenetic age analysis of brain in major depressive disorder. Psychiatry Res. 2018;269:621–624. doi:10.1016/j.psychres.2018.09.001 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.psychres.2018.09.001&link_type=DOI) 41. 41.Hannon E, Lunnon K, Schalkwyk L, Mill J. Interindividual methylomic variation across blood, cortex, and cerebellum: implications for epigenetic studies of neurological and neuropsychiatric phenotypes. Epigenetics. 2015;10(11):1024–1032. doi:10.1080/15592294.2015.1100786 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1080/15592294.2015.1100786&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26457534&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F04%2F25%2F2023.04.24.23289052.atom) 42. 42.Ettman CK, Abdalla SM, Cohen GH, Sampson L, Vivier PM, Galea S. Prevalence of Depression Symptoms in US Adults Before and During the COVID-19 Pandemic. JAMA Netw Open. 2020;3(9):e2019686–e2019686. doi:10.1001/jamanetworkopen.2020.19686 [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jamanetworkopen.2020.19686&link_type=DOI)