Skip to main content

Table 3 Studies investigating the association between the epigenetic clock and mortality

From: The epigenetic clock as a predictor of disease and mortality risk: a systematic review and meta-analysis

Paper

Study (Country)

Characteristics: n, mean age (SD), sex

Outcome, deaths

Sample; platform

Clock*

DNAmAge associations

[correlation with chronological age]

Adj.

[17]

Meta-analysis of 4 cohorts:

LBC1921 and LBC1936 (Scotland),

FHS and NAS

(USA)

LBC1921: 446, age 79.1 (0.6), 60% ♀; LBC1936: 920, age 69.5 (0.8), 49% ♀;

FHS: 2635, age 66.3 (8.9), 54% ♀;

NAS: 657, age 72.9 (6.9), 0% ♀.

2–15 years FU

All-cause mortality, n = 862

Blood; 450 K

IEAAH1

IEAAHa1

16% and 9% ↑ mortality risk for a 5-year higher IEAAHa and IEAA respectively (p < 0.05).

[EAH: r = 0.75, EAHa:0.83]

Chronological age, sex, smoking, education, childhood IQ, social class, hypertension, diabetes, cardiovascular disease, and APOE status.

[18]

Meta-analysis of 12 cohorts: WHI (× 3), NAS, ARIC, FHS, BLSA (USA);

LBC1921 and 1936

(Scotland);

KORA (Germany); InCHIANTI (Italy); Rotterdam (Netherlands)

Overall 12,284.

Age range 52–79, % ♀ not stated.

From 4 to 21 years FU

All-cause mortality, n = 2704

Blood; 450 K

AAH

AAHa

IEAAH

EEAA

All measures of age acceleration are significantly associated with ↑ mortality risk (p ≤ 5.4e-05). EEAA outperformed all measures with the smallest p value.

[EAH: r = 0.15–0.87,

EAHa: r = 0.13–0.89, all depending on cohort, 10 of 26 > 0.80]

Age, BMI, various others

[72]

Twin study

LSADT

(Denmark)

LSADT: 86 same-sex twins,

aged 73–82, 72% ♀

All-cause mortality

n = 55

Blood; 450 K

AAH

AAHa

35% ↑ mortality risk per 5-year ↑ in AA-Horvath (P = 0.02). 2-fold ↑ mortality risk for the twin with the oldest biological age. AAHa NS

[EAH: r = 0.97]

Age, sex, cell composition, twin pairing.

[75]

Cohort study

LBC1921 and LBC1936 (Scotland)

LBC1921: 414, age 79.1 (0.6), 58% ♀, 13 years FU

LBC1936: 920, age 69.5 (0.8), 49%♀,

6 years FU

All-cause mortality, n = 280 and 135

Whole blood; 450 K

AAHa

25% ↑ mortality risk for each 1 SD ↑ in AAHa (p = 0.0001) (Pooled result)

[NS]

Age (at baseline), sex

[76]

Cohort Study

Louisiana Healthy Ageing Study Cohort (USA)

262, age 86 (10), 60.7% ♀, all Caucasian. Average 4.4 years FU

All-cause mortality, n = 206

Blood; 450 K

AAH

NS

[r = 0.63]

Age, cell %

[77]

Cohort Study

NAS

(USA)

241, age 52.6 (10.7), 13% ♀.

White non-Hispanic.

6.5 years FU, average 3.4 years

All-cause mortality, n = 17

Buffy coat; 450 K

AAHa

13% ↑ mortality risk over FU (p = 0.03).

[r = 0.90]

Age (baseline), sex, cell %, PTSD, ancestry

[11]

Seven case-control studies nested in MCCS

Case-control(Australia)

2818 healthy controls matched to cancer cases, age 27–76, 39% ♀. Median 10.7 years FU

All-cause mortality, n = 831;

cancer, n = 240; CVD, n = 203; other-cause mortality, n = 249.

Blood; 450 K

All

5–8% ↑ risk of all-cause mortality per 5-year ↑ Horvath (both predictor types). 11–14% ↑ risk of cancer-related mortality (all predictors). NS with the risk of CVD or other-cause mortality (all predictors).

[EAH: r = 0.73, EAHa:0.76]

Age, various others

[13]

Seven case-control studies nested in MCCS

Case-control(Australia)

3216 cancer cases (different types) and controls, 27–76, 59% ♀. Median 8.3 years FU

All-cause mortality, n = 1726;

cancer-related,

n = 1271;

other causes of death,

n = 309.

Blood; 450 K

All

4–14% ↑ risk of all-cause, cancer-related and other causes of death per 5-year ↑ Hannum (all 3 predictors). Horvath NS predictive of mortality (all measured types).

[EAH: r = 0.73, EAHa:0.78]

Age and sex, matched. Various others

[70]

Cohort study

NAS

(USA)

442 males, age 71.7 (6.7),

370 seen pre-2003,

306 in cohort 2003 to 2013.

3–5 year FU

Cancer mortality,

n = 34

Buffy coat (3 years FU); 450 K

IEAAHa

17% ↑ cancer mortality per year (p = 0.001) in 2003–2013 cohort only.

[NS]

Age at first blood draw, BMI, education, smoking, alcohol, and top 3 principal components

[78]

Cohort study

ESTHER

(Germany)

1863 total, aged 62.5 (6.6).

1260 survivors, 57.2% ♀,

602 cases (316 selected, 43.7% ♀ and 286 deceased in sub-sample, 39.5% ♀).

Maximum 13 years FU

All-cause mortality, n = 602;

cancer-causing mortality, n = 235;

CVD-related mortality, n = 194.

Blood, baseline; 450 K

AAH AAHa

22–23% ↑ risk of all-cause and cancer-causing mortality per 5-year ↑ in AA-Horvath (p < 0.05). AA-Horvath NS with CVD mortality when the model was fully adjusted. NS AA-Hannum with any of the 3 measures of mortality

[r = 0.77, 0.73)

[EAH: r = 0.73, EAHa: r = 0.77]

Age, sex, BMI, various others

[67]

Two cohorts of ischemic stroke (IS) patients (Spain)

Discovery:

551, 65–81 (NS), 41.9% ♀;

replication:

85, 66–80 (NS), 29.4% ♀

IS outcome at 3 months (from 0 = no symptoms to 6 = death)

Whole blood; 450 K

AAHa

↑ worse 3-month stroke outcome (OR: 1.04, p < 0.001 and OR: 1.06, p = 0.007 in 2 cohorts).

[r = 0.81]

Biological age, sex, recanalization treatment, and basal NIHSS, various others

  1. AAH, Age Acceleration (Horvath); AAHa, age acceleration (Hannum); CVD, cardiovascular disease; DNAmAge, DNA methylation age; EEAA, extrinsic epigenetic age acceleration (Hannum Measurement); EAH, epigenetic age (Horvath); EAHa, epigenetic age (Hannum); EPIC, European Prospective Investigation into Cancer and Nutrition; ESTHER, epidemiological investigations of the changes of preventing, recognising early and optimally treating chronic diseases in an elderly population; FHS, Framingham Heart Study; FU, follow-up; IEAAH, intrinsic epigenetic age acceleration (Horvath); IEAAHa, intrinsic epigenetic age acceleration (Hannum); IS, ischemic stroke; LBC1921, Lothian Birth Cohort 1921; LBC1936, Lothian Birth Cohort 1936; LSADT, Longitudinal Study of Ageing Danish Twins; MCCS, Melbourne Collaborative Cohort Study; NAS, US Department of Veterans Affairs’ Normative Ageing Study; NIHSS, NIH Stroke Scale; NS, not shown; WHI, Women’s Health Initiative (WHI)
  2. *Includes all five measures of interest to this review, including AAH, AAHa, IEAAH, IEAAHa, and EEAA
  3. 1Although not specifically stated, these measures of AA adjusted for cell type proportions in blood; hence both measures are intrinsic epigenetic age acceleration