Table 2. Risk of ovarian cancer by use of oral contraceptives using conditional logistic regression and generalised estimating equations, Breast Cancer Family Registry.
Conditional logistic |
GEE |
GEE |
|||
---|---|---|---|---|---|
Cases with unaffected sister controls |
Cases with unaffected sister controls |
All cases and all controls |
|||
|
Age-adjusted model |
Multivariable modela |
Age-adjusted model |
Multivariable modelb |
Multivariable modelc |
OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | |
Oral contraceptive use | |||||
Never | Ref | Ref | Ref | Ref | Ref |
Ever |
0.51 (0.23, 1.15) |
0.52 (0.23, 1.17) |
0.71 (0.47, 1.06) |
0.58 (0.37, 0.91) |
0.35 (0.27, 0.45) |
Duration of oral contraceptive use (years) | |||||
0 | Ref | Ref | Ref | Ref | Ref |
< 3 | 0.77 (0.32, 1.85) | 0.85 (0.34, 2.10) | 1.16 (0.68, 1.97) | 0.99 (0.54, 1.80) | 0.41 (0.30, 0.55) |
3 to <6 | 0.18 (0.04, 0.88) | 0.19 (0.04, 0.90) | 0.30 (0.10, 0.88) | 0.27 (0.09, 0.83) | 0.23 (0.15, 0.38) |
⩾ 6 | 0.26 (0.08, 0.84) | 0.26 (0.08, 0.83) | 0.44 (0.23, 0.84) | 0.33 (0.16, 0.67) | 0.34 (0.24, 0.49) |
P–value for trend |
<0.01 |
<0.01 |
<0.01 |
<0.01 |
<0.01 |
Age at first oral contraceptive use (years) | |||||
Never | Ref | Ref | Ref | Ref | Ref |
<20 | 0.66 (0.18, 2.47) | 0.68 (0.18, 2.56) | 0.87 (0.48, 1.57) | 0.71 (0.37, 1.39) | 0.26 (0.18, 0.39) |
20–25 | 0.20 (0.06, 0.68) | 0.21 (0.06, 0.71) | 0.45 (0.24, 0.82) | 0.35 (0.18, 0.66) | 0.26 (0.19, 0.37) |
>25 | 0.61 (0.19, 1.96) | 0.61 (0.19, 1.96) | 0.66 (0.28, 1.57) | 0.55 (0.23, 1.30) | 0.46 (0.31, 0.69) |
P– value for trend | 0.07 | 0.07 | 0.03 | <0.01 | <0.01 |
Adjusted for age, parity.
Adjusted for age, race, parity, age at last birth.
Adjusted for age, race, parity.