Skip to main content

Table 3 The associations between CV risk factors and the ECG-based cardiac ages and the role of DNAm AgeAccel as a mediator of these relationships

From: Accelerated DNA methylation age plays a role in the impact of cardiovascular risk factors on the human heart

ECG-based cardiac age

Model

Effect type

Body Mass Index

Diabetes

High cholesterol

Hypertension

Cardiovascular disease

Any CV risk factor

β (95% CI)

p value

β (95% CI)

p value

β (95% CI)

p value

β (95% CI)

p value

β (95% CI)

p value

β (95% CI)

p value

Bayesian A-ECG age

M1

Total

0.32

(0.21, 0.44)

 < 2e−16

1.30

(0.09, 2.62)

0.034

2.31

(0.89, 3.57

 < 2e−16

2.67

(1.70, 3.74)

 < 2e−16

1.81

(− 0.58, 4.27)

0.150

2.70

(1.61, 3.74)

 < 2e−16

M2

(AgeAccelHannum)

ACME

0.00

(0.00, 0.01)

0.940

− 0.01

(− 0.11, 0.09)

0.950

0.01

(− 0.09, 0.11)

0.910

− 0.01

(− 0.08, 0.09)

0.940

0.01

(− 0.11, 0.23)

0.820

− 0.02

(− 0.17, 0.11)

0.740

ADE

0.32

(0.21, 0.44)

 < 2e−16

1.31

(0.08, 2.64)

0.036

2.30

(0.88, 3.55)

 < 2e−16

2.68

(1.70, 3.75)

 < 2e−16

1.80

(− 0.60, 4.23)

0.150

2.72

(1.61, 3.74)

 < 2e−16

M3

(AgeAccelHorvath)

ACME

0.02

(− 0.01, 0.04)

0.19

0.02

(− 0.17, 0.22)

0.818

0.03

(− 0.23, 0.35)

0.850

0.12

(− 0.02, 0.38)

0.160

0.07

(− 0.10, 0.54)

0.540

0.15

(− 0.07, 0.45)

0.170

ADE

0.30

(0.19, 0.43)

 < 2e−16

1.28

(0.07, 2.62)

0.036

2.28

(0.91, 3.57)

 < 2e−16

2.55

(1.60, 3.61)

 < 2e−16

1.74

(− 0.70, 4.15)

0.160

2.55

(1.41, 3.59)

 < 2e−16

M4

(AgeAccelPheno)

ACME

0.02

(− 0.01, 0.06)

0.160

0.23

(0.01, 0.52)

0.028

0.34

(0.03, 0.74)

0.026

0.08

(− 0.07, 0.28)

0.290

0.26

(− 0.05, 0.88)

0.150

0.17

(0.01, 0.41)

0.048

ADE

0.30

(0.21, 0.44)

 < 2e−16

1.08

(− 0.16, 2.45)

0.098

1.97

(0.53, 3.21)

0.002

2.59

(1.62, 3.64)

 < 2e−16

1.55

(− 0.81, 4.01)

0.150

2.53

(1.40, 3.61)

 < 2e−16

M5

(AgeAccelGrim)

ACME

0.02

(− 0.01, 0.05)

0.064

0.24

(0.01, 0.60)

0.030

0.34

(− 0.02, 0.74)

0.068

0.19

(0.00, 0.42)

0.048

0.33

(− 0.06, 0.91)

0.100

0.26

(− 0.04, 0.57)

0.092

ADE

0.30

(0.18, 0.43)

 < 2e−16

1.06

(− 0.19, 2.43)

0.118

1.97

(− 0.02, 0.74)

0.068

2.48

(1.50, 3.74)

0.04

1.48

(− 0.91, 3.92)

0.250

2.44

(1.28, 3.50)

 < 2e−16

DNN

ECG-based age

M1

Total

0.59

(0.37, 0.82)

 < 2e−16

1.72

(− 0.93, 4.35)

0.200

2.71

(− 0.05, 5.49)

0.054

2.72

(0.43, 5.01)

0.020

4.16

(0.18, 8.69)

0.040

3.20

(0.84, 5.74)

0.008

M2

(AgeAccelHannum)

ACME

0.00

(− 0.02, 0.02)

0.920

0.02

(− 0.16, 0.31)

0.850

0.01

(− 0.23, 0.28)

0.910

0.05

(− 0.11, 0.33)

0.560

− 0.10

(− 0.58, 0.43)

0.752

0.09

(− 0.11, 0.44)

0.500

ADE

0.59

(0.37, 0.82)

 < 2e−16

1.70

(− 0.92, 4.36)

0.200

2.70

(− 0.05, 5.49)

0.058

2.67

(0.33, 4.99)

0.024

4.26

(0.16, 8.68)

0.042

3.11

(0.76, 5.71)

0.010

M3

(AgeAccelHorvath)

ACME

0.02

(− 0.02, 0.08)

0.280

0.04

(− 0.28, 0.36)

0.840

0.05

(− 0.48, 0.67)

0.860

0.23

(− 0.04, 0.72)

0.152

0.21

(0.30, 1.14)

0.524

0.31

(− 0.16, 0.80)

0.206

ADE

0.57

(0.35, 0.80)

 < 2e−16

1.68

(− 0.96, 4.34)

0.210

2.66

(− 0.03, 5.29)

0.054

2.49

(0.09 4.83)

0.034

3.95

(0.06, 8.47)

0.046

2.89

(0.49, 5.54)

0.020

M4

(AgeAccelPheno)

ACME

0.03

(− 0.05, 0.11)

0.530

0.24

(− 0.15, 0.74)

0.270

0.36

(− 0.24, 1.14)

0.276

0.13

(− 0.12, 0.53)

0.370

0.51

(− 0.12, 1.61)

0.176

0.27

(− 0.08, 0.82)

0.140

ADE

0.56

(0.34, 0.80)

 < 2e−16

1.48

(− 0.93, 4.35)

0.250

2.35

(− 0.23, 5.05)

0.084

2.59

(0.25, 4.95)

0.030

2.65

(− 0.24, 8.01)

0.064

2.93

(0.50, 5.52)

0.016

M5

(AgeAccelGrim)

ACME

0.04

(− 0.01, 0.11)

0.100

0.48

(0.02, 1.22)

0.028

0.76

(0.04, 1.71)

0.036

0.46

(0.01, 1.07)

0.046

0.69

(− 0.10, 1.97)

0.100

0.68

(− 0.04, 1.47)

0.064

ADE

0.55

(0.33, 0.79)

 < 2e−16

1.24

(− 1.36, 4.03)

0.342

1.95

(0.04, 1.71)

0.160

2.26

(− 0.05, 4.70)

0.060

3.47

(− 0.40, 7.85)

0.070

2.52

(− 0.08, 5.18)

0.056

  1. All reported analyses consisted of generalized linear models with Gaussian distributions and identity links. Significant p-values are in bold
  2. M1 aimed to evaluate the total effects, and was adjusted for chronological age, sex, SEP (as a binary variable: manual and non-manual), smoking status (as current smoker, ex-smoker and never-smoked), alcohol comsunption (as the average number of alcohol units drank per day), physical activity status (as active or inactive) and white cell counts (i.e., naïve and exhausted CD8 + T-lymphocytes, CD4 + T-lymphocytes, B-cells, natural killer cells, granulocytes, and monocytes). Compared to M1, M2 was in addition adjusted for AgeAccelHannum, M3 for AgeAccelHorvath, M4 for AgeAccelPheno and M5 for AgeAccelGrim
  3. ACME = average causal mediation effect; ADE = average direct effect; M = model. Other abbreviations as in Table 2