Aging

Epigenetic Clocks of Biological Aging and Risk of Incident Mild Cognitive Impairment and Dementia: The Women's Health Initiative Memory Study
Steve Nguyen, Ake T Lu, Steve Horvath, Mark A Espeland, Stephen R Rapp, Adam X Maihofer, Caroline M Nievergelt, Andrea Z LaCroix, Linda K McEvoy, Susan M Resnick, Kenneth Beckman, Aladdin H Shadyab

Abstract:

Aging is the strongest risk factor for dementia; however, few studies have examined the association of biological aging with incident dementia. We analyzed 6069 cognitively unimpaired women (mean age = 70.0 ± 3.8 years) in the Women's Health Initiative Memory Study to examine the association of accelerated biological aging, measured with second and third‐generation epigenetic clocks (AgeAccelPheno and AgeAccelGrim2, and DunedinPACE, respectively) with incident mild cognitive impairment (MCI) and probable dementia. Multivariable Cox proportional hazards models adjusted for age, education, race, ethnicity, smoking, hormone therapy regimen, physical activity, body mass index, and estimated white blood cell counts. For comparison, we also examined first‐generation epigenetic clocks (AgeAccelHorvath; AgeAccelHannum). We evaluated effect modification by age, race/ethnicity, hormone therapy regimen, menopause type (natural vs. surgical), and APOE ε4 carriage. There were 1307 incident MCI or probable dementia events over a median follow‐up of 9.3 (25th percentile = 6.1, 75th percentile = 16.1) years. The adjusted HRs (95% CI; p‐value) for incident MCI/probable dementia per one‐standard deviation increment were 1.07 (1.01–1.15; p = 0.03) for DunedinPACE, 1.11 (1.02–1.20; p = 0.01) for AgeAccelGrim2, and 1.01 (0.95–1.07; p = 0.74) for AgeAccelPheno. Only AgeAccelGrim2 remained significant under the Bonferroni‐corrected threshold for significance (p < 0.02). Other epigenetic clocks were not associated with incident MCI/probable dementia. There was no effect modification in most subgroup analyses (p‐interaction ≥ 0.05). In this cohort study of older women, accelerated biological aging measured by AgeAccelGrim2 was associated with higher risk of incident MCI/probable dementia. These findings provide evidence linking epigenetic biomarkers of biological aging with MCI and dementia development, independent of chronological age.

Annotation

The finding: In this study of women in the Women’s Health Initiative Memory Study (an RCT examining the effects of estrogen hormone therapy on cognitive outcomes), the authors examine whether epigenetic age in women without cognitive impairment predicts the development of either mild cognitive impairment (MCI) or dementia. This study expands upon a prior, smaller study (old n = 578, new n = 6069), and included analysis of additional epigenetic clocks. At baseline, the participants were 70 years old, and were followed for a median of 9.3 years, identifying 1307 cases of MCI and/or dementia. The authors find that accelerated GrimAgeV2 at baseline predicted higher risk of a combined MCI/dementia outcome. DunedinPACE nominally predicted higher risk of MCI/dementia, although it did not survive correction for multiple comparisons.

Strength and weaknesses: This study has several strengths, including a large sample size with a long period of follow-up, the examination of multiple 2nd and 3rd generation epigenetic clocks, the use of multiple comparisons correction, and adjustment of a wide range of confounding factors. The authors also examine the potential for interactions between the potential confounding factors (age, race, hormone therapy, menopause, and APOE e4 genotype). However, the relatively controlled nature of the study and the focus on women may limit generalizability.

Relevance: This study found evidence that, in older women, some measures of epigenetic age (specifically, GrimAgeV2 and DunedinPACE) are associated with a higher risk of MCI and/or dementia. The identification of a small but significant additional risk of GrimAgeV2 (the epigenetic clock which often has the strongest associations with morbidity and mortality) is particularly notable, as multiple psychiatric, psychosocial, and socioeconomic variables have also been associated with GrimAge acceleration, lending support and a potential biological mechanism through which these factors might influence risk of dementia.