Associations of Epigenetic Age Estimators with Cognitive Function Trajectories in the Women's Health Initiative Memory Study

Association Between Epigenetic Age Acceleration Estimates and Cognitive Trajectories in the Women’s Health Initiative Memory Study


Maintaining cognitive function is crucial for independent living among older adults. With the growth of the aging population, cognitive decline has become a significant public health challenge. Many studies have found that various modifiable and non-modifiable risk factors, such as educational attainment, socioeconomic status, physical activity, cardiovascular diseases (CVD), cardiovascular disease risk factors, hearing ability, and genetic factors (e.g., APOE e4), are associated with cognitive decline. However, the molecular mechanisms underlying the aging process, particularly the relationship between epigenetic age acceleration and cognitive decline, are not fully understood.

Epigenetic Age Estimation

Epigenetic age estimators are biomarkers based on DNA methylation (DNAm) that capture aging-related biological processes. Studies have shown that epigenetic age acceleration is prospectively associated with higher mortality, coronary heart disease (CHD) risk, and various age-related phenotypes, such as poorer physical function, frailty, and cancer. However, research on the association between epigenetic age acceleration and cognitive function is limited, with most studies being cross-sectional or longitudinal with few repeated measurements and shorter follow-up periods. Therefore, exploring the association between epigenetic age acceleration and cognitive trajectories in older adults is crucial.

Study Source

This research was conducted by Steve Nguyen, Linda K. McEvoy, Mark A. Espeland, Eric A. Whitsel, Ake Lu, Steve Horvath, Joann E. Manson, Stephen R. Rapp, and Aladdin H. Shadyab, among others, from various renowned institutions in the United States, including the University of California San Diego, Wake Forest University, University of North Carolina, Altos Labs, Harvard Medical School, and Kaiser Permanente Washington Health Research Institute. The paper was published in the journal Neurology in 2024.

Study Process

Study Population and Data Source

The study included older women from the Women’s Health Initiative Memory Study (WHIMS), with DNA methylation data collected between 1995 and 1998, followed by a 13-year follow-up period. These women underwent annual assessment of global cognitive function using the Modified Mini-Mental State Examination (3MS, 1995-2007) and the Modified Telephone Interview for Cognitive Status (TICs-M, 2008-2021).

The study population consisted of 795 participants, with a mean baseline age of 70.8 ± 4 years, of whom 10.7% were Black, 3.9% were Hispanic or Latina, and 85.4% were White.

Epigenetic Age Estimation

The study calculated five epigenetic age estimators:

  1. Extrinsic Epigenetic Age Acceleration (EEAA)
  2. Intrinsic Epigenetic Age Acceleration (IEAA)
  3. AgeAccelPheno
  4. AgeAccelGrim2
  5. Dunedin Pace of Aging (DunedinPACE) derived from the epigenome

EEAA and IEAA are first-generation epigenetic age estimators, while AgeAccelPheno and AgeAccelGrim2 are second-generation estimators. DunedinPACE captures the pace of aging across multiple organ systems using a panel of biomarkers.

Data Processing and Analysis Methods

The study used Linear Mixed-Effects Models (LMMs) to estimate the longitudinal associations between epigenetic age estimators and cognitive function. The models included minimally adjusted and fully adjusted variables, with the latter including age, education, race/ethnicity, and further adjustments for alcohol consumption, smoking, body mass index (BMI), and comorbidities. Additionally, the study examined the potential modifying effect of APOE e4 genotype on the results.

Study Results

Key Findings

  1. Association between DunedinPACE and Cognitive Decline: The study found that a one-standard deviation (0.12 unit) increase in DunedinPACE was associated with a faster annual decline in TICs-M scores. This association was significant in both minimally adjusted and fully adjusted models. Furthermore, an increase in DunedinPACE was also associated with a decline in 3MS scores among APOE e4 carriers, but not among non-carriers.

  2. Other Epigenetic Age Estimators: AgeAccelPheno was associated with a decline in TICs-M scores in the minimally adjusted model, but the association was not significant in the fully adjusted model. EEAA, IEAA, and AgeAccelGrim2 did not show significant associations with changes in 3MS or TICs-M scores.

  3. AgeAccelGrim2 Components: Various AgeAccelGrim2-related DNAm plasma protein surrogates, such as GDF15, B2M, Cystatin C, TIMP1, and Adm, were associated with declines in 3MS and TICs-M scores. Among these, dnAm GDF15 and dnAm B2M showed relatively stronger associations.


DunedinPACE may help identify older women at higher risk of future cognitive decline. This finding suggests the potential application of epigenetic techniques in cognitive function prediction. However, the study has limitations, such as the sample being derived from a substudy and the data not being specifically designed to investigate the relationship between epigenetics and cognitive function. Furthermore, 3MS and TICs-M are measures of global cognitive function, and future studies should explore the association between epigenetics and specific cognitive domains.


This study extends the understanding of the relationship between epigenetic age acceleration and cognitive decline. The results indicate that specific epigenetic age indicators, particularly DunedinPACE, could serve as potential tools for predicting future cognitive health. For older adults, especially APOE e4 carriers, these findings have important implications.

By integrating these research findings and methods, this paper provides a valuable foundation and direction for further research on epigenetic age estimators. Future large-scale longitudinal studies are needed to validate these findings and explore their broader application potential.