Table 1 Mendelian Randomisation estimates between accelerometer-measured physical activity and cancer risk.

From: Physical activity and risks of breast and colorectal cancer: a Mendelian randomisation analysis

Methods

Genome-wide significant SNPs (n = 5) from the GWAS by Doherty et al.11

Extended number of SNPs (n = 10) from the GWAS by Klimentidis et al.12

 

No. Cases

Estimates (OR)a

95% CI

P-value

Q-value

P-value for pleiotropyb or heterogeneityc

Estimates (OR)a

95% CI

P-value

Q-value

P-value for pleiotropyb or heterogeneityc

Breast cancer

Inverse-variance weightedd

122,977

0.51

0.27, 0.98

0.04

0.062

4.4 × 10−8

0.59

0.42, 0.84

0.003

0.012

6.8 × 10−7

MR-Egger

0.01

0.00, 2.01

0.09

 

0.16

0.55

0.09, 3.20

0.5

0.9

Weighted median

0.61

0.42, 0.87

0.006

  

0.76

0.59, 0.98

0.03

ER+ve subset

Inverse-variance weightedd

69,501

0.45

0.20, 1.01

0.054

0.077

8.5 × 10−9

0.53

0.35, 0.82

0.004

0.004

3.1 × 10−7

MR-Egger

0.03

0.00, 40

0.34

 

0.46

0.61

0.07, 5.26

0.65

0.9

Weighted median

0.55

0.35, 0.85

0.008

  

0.66

0.48, 0.90

0.008

ER-ve subset

Inverse-variance weightedd

21,468

0.95

0.44, 2.04

0.89

0.89

0.002

0.78

0.51, 1.22

0.27

0.3

0.01

MR-Egger

0.01

0.00, 4.48

0.15

 

0.15

0.24

0.03, 1.81

0.17

0.24

Weighted median

0.84

0.47, 1.47

0.53

  

0.7

0.47, 1.04

0.08

Colorectal cancer

Inverse-variance weighted

52,775

0.66

0.48, 0.90

0.01

0.022

0.39

0.6

0.47, 0.76

2.4 × 10−5

0.0002

0.5

MR-Egger

0.32

0.01, 6.69

0.46

 

0.64

0.24

0.08, 0.72

0.011

0.1

Weighted median

0.6

0.39, 0.92

0.02

  

0.61

0.44, 0.85

0.003

Colorectal cancer in men

Inverse-variance weighted

28,207

0.79

0.50, 1.23

0.29

0.31

0.22

0.76

0.55, 1.07

0.11

0.14

0.62

MR-Egger

16.4

0.32, 812

0.16

 

0.13

0.59

0.12, 2.81

0.51

0.74

Weighted median

0.64

0.34, 1.19

0.16

  

0.8

0.51, 1.27

0.34

Colorectal cancer in women

Inverse-variance weighted

24,568

0.57

0.36, 0.90

0.02

0.036

0.08

0.49

0.35, 0.68

3.0 × 10−5

0.0002

0.19

MR-Egger

0.01

0.00, 0.54

0.02

 

0.045

0.11

0.02, 0.55

0.007

0.06

Weighted median

0.61

0.32, 1.16

0.13

  

0.47

0.29, 0.75

0.002

Colon cancer

Inverse-variance weighted

27,817

0.64

0.44, 0.94

0.02

0.036

0.17

0.56

0.42, 0.73

4.4 × 10−5

0.0002

0.57

MR-Egger

0.42

0.00, 40.5

0.71

 

0.86

0.35

0.09, 1.29

0.11

0.47

Weighted median

0.62

0.36, 1.06

0.08

  

0.49

0.34, 0.72

3.0 × 10−4

 

Proximal colon cancer

Inverse-variance weighted

12,360

0.66

0.41, 1.06

0.09

0.12

0.72

0.6

0.42, 0.86

0.005

0.014

0.9

MR-Egger

0.62

0.01, 33.12

0.82

 

0.98

0.33

0.06, 1.71

0.18

0.46

Weighted median

0.67

0.36, 1.22

0.19

  

0.56

0.35, 0.89

0.01

Distal colon cancer

Inverse-variance weighted

14,016

0.51

0.31, 0.83

0.007

0.018

0.74

0.45

0.31, 0.64

1.7 × 10−5

0.0002

0.72

MR-Egger

0.32

0.00, 121

0.71

 

0.88

0.34

0.06, 1.89

0.22

0.75

Weighted median

0.5

0.25, 1.00

0.051

  

0.45

0.28, 0.75

0.002

Rectal cancer

Inverse-variance weighted

13,713

0.7

0.43, 1.14

0.15

0.18

0.13

0.68

0.47, 0.98

0.04

0.062

0.24

MR-Egger

3.49

0.01, 1635

0.69

 

0.6

0.43

0.06, 3.26

0.41

0.65

Weighted median

0.94

0.49, 1.79

0.85

  

0.76

0.47, 1.27

0.3

  1. CI confidence intervals, MR Mendelian randomisation, OR odds ratio, SNPs Single nucleotide polymorphisms
  2. aThe estimates correspond to a standard deviation increase in physical activity
  3. Q-value: False discovery rate (FDR) correction performed using the Benjamini–Hochberg method
  4. bP-value or pleiotropy based on MR-Egger intercept
  5. cP-value for heterogeneity based on Q statistic
  6. dThe estimates were derived from a random effects model due to the presence of heterogeneity based on Cochran’s Q statistic
Back to article page