Table 3 Subgroup analyses of pooled relative risks of colorectal cancer per 20 mg/dL increase in fasting blood glucose.

From: A Linear Dose-Response Relationship between Fasting Plasma Glucose and Colorectal Cancer Risk: Systematic Review and Meta-analysis

Subgroup

Number of study

Relative Risk (95%CI)

P value

Test for heterogeneity*

I2 (%)

P value

Cancer type

     

 CRC

6

1.016 (1.012–1.019)

0.000

0

0.904

 CC

2

1.035 (1.008–1.062)

0.011

11

0.295

 RCψ

2

1.031 (0.189–5.628)

0.972

11

0.345

Regionξ

     

 North America

2

1.041 (1.010–1.072)

0.008

2

0.381

 Europe

6

1.010 (0.992–1.029)

0.284

16

0.26

 Asia

2

1.015 (1.011–1.019)

0.000

0

0.694

Follow-up time (years)

     

 <10

5

1.015 (0.998–1.032)

0.076

46

0.030

 ≥10

5

1.015 (1.011–1.019)

0.000

0

0.928

Gender

     

 Both

3

1.013 (0.996–1.031)

0.122

3

0.405

 Male

4

1.016 (1.012–1.020)

0.000

0

0.847

 Female

3

1.011 (0.995–1.027)

0.164

0

0.555

Fasting status

     

 Fasting

6

1.016 (1.012–1.019)

0.000

0

0.585

 Mix

4

1.008 (0.989–1.027)

0.421

23

0.208

Risk type

     

 HR

7

1.015 (1.011–1.019)

0.000

34

0.068

 RR

3

1.036 (1.008–1.063)

0.010

0

0.941

  1. *For the test of heterogeneity in each subgroup, we also calculated the I2 statistic and 50% was regarded as the cutoff point for non-significant and significant levels. No significant heterogeneity was detected in our dose-response meta-analysis and we did not further test the heterogeneity between subgroups.
  2. ψThe total sample size and the number of rectal cancer cases included were small.
  3. Both studies using HR and RR as the risk type showed a significant dose-response relationship and we assigned RR as the risk type to illustrate our results.
Back to article page