Table 2.

Geometric means of absolute density, absolute nondensity and relative density and trend tests over groups of characteristics

nAbsolute dense, cm2Absolute nondense, cm2Percent dense, %
Age at examination, y
    <5413931.372.026.1
    54-5814022.280.118.9
    >5813918.292.314.3
    P for trend418<0.01<0.01<0.01
Weight, kg
    <6314026.253.029.0
    63-7014322.691.317.6
    >7013521.2114.613.8
    P for trend4180.06<0.01<0.01
Height, m
    <1.6013321.384.717.5
    1.60-1.6515123.681.019.4
    >1.6513425.280.220.8
    P for trend4180.150.510.18
BMI, kg/m2
    <201228.321.752.0
    20 to <2518726.662.026.4
    25 to <3017121.6101.115.8
    <304117.3156.18.9
    P for trend4180.07<0.01<0.01
Age at menarche, y
    <1312221.496.616.1
    13-1417823.472.321.1
    >1411625.582.820.5
    P for trend4160.160.070.08
Age at menopause, y
    <4910719.583.116.8
    49-5113919.990.915.8
    >5111527.077.722.5
    P for trend3610.010.420.04
Time since menopause, y
    <412028.374.423.7
    4-912121.181.518.1
    >912017.498.513.5
    P for trend361<0.01<0.01<0.01
Parous
    Yes33222.287.817.7
    No8628.162.526.3
    P4180.04<0.01<0.01
Age at first childbirth, y
    <2511020.892.616.2
    25-2811122.681.719.1
    >2811123.189.617.9
    P for trend3320.420.700.50
Parity (no. children)
    08628.162.526.3
    15326.982.321.3
    29328.577.323.6
    38920.292.415.8
    >39717.298.213.5
    P for trend418<0.01<0.01<0.01
Current oral contraceptive use
    Yes831.366.228.8
    No41023.282.219.1
    P4180.870.360.29
Ever used oral contraceptives
    Yes4328.078.323.5
    No37522.882.318.8
    P4180.180.640.19
Used hormone therapy in the last 12 mo
    Yes4727.875.923.9
    No37122.882.718.7
    P4180.180.410.14
Ever smoked
    Yes11627.467.525.1
    No30221.988.317.3
    P4180.03<0.01<0.01
Family history of breast cancer*
    Yes2623.677.420.7
    No38223.282.719.0
    P4080.920.620.71
  • NOTE: Geometric means were calculated in tertiles of continuous variables or binary groups of dichotomous variables. For BMI, standard classes were used and for parity, nominal classes were used, with a truncation at >3. Tests for linear trends were done through ANOVA with an F test on the geometric data.

  • * Defined as one or more affected mother and/or sister.