Skip to main content

Table 2 Summary of the main findings from the meta-analysis

From: Clinicohistopathological implications of MMP/VEGF expression in retinoblastoma: a combined meta-analysis and bioinformatics analysis

Outcome-molecular

No. of patients/studies

Random-effects model

Quality-effects modela

Fixed-effects modelb

OR(s) or r (95% CI)

Test for overall effects (P)

Overall heterogeneity (I2), %

OR(s) or r (95% CI)

Overall Heterogeneity (I2), %

OR(s) or r (95% CI)

Test for overall effects (P)

Overall heterogeneity (I2), %

Tumor invasion

 

OR

  

OR

 

OR

  

 MMP-1

113/3

4.21 (1.86–9.54)

0.0006

0

4.14 (1.83–9.39)

0

4.21 (1.87.9/47)

0.0005

0

 MMP-2

140/3

11.18 (4.26–29.30)

< 0.00001

0

11.24 (4.29–29.50)

0

11.46 (4.41–29.79)

< 0.00001

0

 MMP-9

431/10

10.41 (4.26–25.47)

< 0.00001

59

6.37 (2.48–16.35)

57

–

–

–

 VEGF

409/10

8.09 (4.03–16.20)

< 0.00001

0

8.08 (4.03–16.19)

0

9.93 (5.11–19.30)

< 0.00001

0

Tumor differentiation

 

OR

  

OR

 

OR

  

 MMP-1

113/3

3.58 (1.48–8.71)

0.005

13

3.42 (1.40–8.33)

13

3.55 (1.60–7.89)

0.002

13

 MMP-2

140/3

2.96 (1.32–6.64)

0.009

13

3.03 (1.35–6.81)

13

3.01 (1.45–6.25)

0.003

13

 MMP-9

474/10

5.49 (3.55–8.48)

< 0.00001

0

6.37 (2.48–16.35)

57

5.77 (3.77–8.83)

< 0.00001

0

 VEGF

510/11

5.30 (2.93–9.60)

< 0.00001

38

4.93 (2.71–8.96)

0

5.04 (3.23–7.87)

< 0.00001

38

Clinical stage

 

OR

  

OR

 

OR

  

 MMP-9

227/4

5.17 (2.85–9.38)

< 0.00001

0

5.22 (2.88–9.48)

0

5.19 (2.87–9.39)

< 0.00001

0

Gender

 

OR

  

OR

 

OR

  

 MMP-9

152/3

0.60 (0.26–1.34)

0.21

32

0.60 (0.27–1.36)

32

0.60 (0.31–1.15)

0.12

32

Coexpressiona

 

r

  

r

    

 MMP-9 and VEGF

287/5

0.61 (0.38–0.77)

–

84

0.59 (0.34–0.77)

84

–

–

–

  1. No number, OR odds ratio, r summary correlation coefficients, CI confidence intervals
  2. aThese results are calculated using statistical add-in software MetaXL, in which the p value of testing for overall effects is not provided
  3. bThe fixed-effects model is applied only when I2 ≤ 50%