Table 3.
Weighted Odds Ratios and 95% Confidence Intervals Representing the Association between Physical Inactivity and Epithelial Ovarian Cancer among Case-Control Studies included in the Meta-analysesa
| Epithelial Ovarian Cancer (EOC) Histological Type |
Modelb | Summary ORc |
95% CI | Heterogeneity | |||
|---|---|---|---|---|---|---|---|
| Lower | Upper | Q-statistic | P-value | I2 | |||
|
Overall EOC
(all histological types combined) |
Multivariable | 1.34 (RE) | 1.14 | 1.57 | 33.77 | <0.001 | 76.31% |
| Add BMI | 1.32 (RE) | 1.12 | 1.56 | 34.43 | <0.001 | 76.76% | |
|
Invasive EOC
(all invasive cases combined) |
Multivariable | 1.35 (RE) | 1.14 | 1.60 | 34.23 | <0.001 | 76.63% |
| Add BMI | 1.33 (RE) | 1.12 | 1.58 | 33.13 | <0.001 | 75.85% | |
| Invasive High-grade Serous | Multivariable | 1.30 (RE) | 1.10 | 1.54 | 23.40 | 0.003 | 65.82% |
| Add BMI | 1.30 (RE) | 1.08 | 1.53 | 22.60 | 0.004 | 64.60% | |
| Invasive Low-grade Serous | Multivariable | 1.39 (FE) | 1.06 | 1.83 | 15.02 | 0.059 | 46.72% |
| Add BMI | 1.33 (FE) | 1.01 | 1.76 | 13.90 | 0.084 | 42.44% | |
| Invasive Endometrioid | Multivariable | 1.29 (RE) | 1.01 | 1.65 | 17.79 | 0.023 | 55.03% |
| Add BMI | 1.26 (RE) | 0.98 | 1.63 | 18.29 | 0.019 | 56.23 | |
| Invasive Mucinous | Multivariable | 1.54 (FE) | 1.22 | 1.95 | 5.40 | 0.715 | 0.00% |
| Add BMI | 1.50 (FE) | 1.17 | 1.10 | 5.78 | 0.672 | 0.00% | |
| Invasive Clear Cell | Multivariable | 1.45 (FE) | 1.16 | 1.80 | 15.46 | 0.051 | 48.24% |
| Add BMI | 1.40 (FE) | 1.11 | 1.74 | 13.59 | 0.093 | 41.13% | |
|
Borderline EOC
(serous & mucinous combined) |
Multivariable | 1.27 (FE) | 1.10 | 1.46 | 7.84 | 0.250 | 23.50% |
| Add BMI | 1.20 (FE) | 1.04 | 1.39 | 7.64 | 0.265 | 21.51% | |
| Borderline Serous | Multivariable | 1.27 (FE) | 1.07 | 1.52 | 6.59 | 0.360 | 8.95% |
| Add BMI | 1.20 (FE) | 1.01 | 1.44 | 5.59 | 0.470 | 0.00% | |
| Borderline Mucinous | Multivariable | 1.31 (FE) | 1.06 | 1.61 | 9.69 | 0.138 | 38.08% |
| Add BMI | 1.25 (FE) | 1.01 | 1.55 | 9.88 | 0.130 | 39.30% | |
a
For each EOC endpoint, studies with >10 cases are included in meta-analyses
b
Study specific multivariable models are adjusted by age, race, parity, oral contraceptive use, a personal history of endometriosis, and a family history of breast or ovarian cancer
c
Random-effects (RE) ORs reported when Q-statistic is significant (p<0.05) or when I2 is greater than 50%; Fixed-effects (FE) models reported when no significant heterogeneity was observed.