Table 3.
Model variablesMean mammographic percent density
Spearman correlation coefficientP*R2
Q1Q2Q3Q4
A. Mean percent of breast density by quartiles of circulating estrogens among postmenopausal women not on exogenous hormones, Nurses' Health Study (1989-1990)
Estrone
    Age and batch (n = 505)23.820.218.515.3−0.150.00050.03
    +BMI20.818.817.920.60.030.550.25
    +MV20.218.718.420.60.050.300.31
Estradiol
    Age and batch (n = 515)25.619.817.814.4−0.22<0.00010.05
    +BMI21.618.317.420.40.010.810.25
    +MV21.218.417.520.60.020.610.30
Free estradiol§
    Age and batch (n = 485)26.620.916.113.9−0.26<0.00010.06
    +BMI22.018.716.719.70.0070.880.25
    +MV21.618.517.319.80.020.680.30
Non-SHBG–bound estradiol
    Age and batch (n = 370)26.523.016.814.6−0.28<0.00010.06
    +BMI21.821.117.720.30.0070.900.26
    +MV21.121.118.220.30.020.710.35
Estrone sulfate
    Age and batch (n = 501)22.618.920.115.4−0.100.020.01
    +BMI19.718.619.619.50.050.220.25
    +MV19.518.819.919.10.060.220.30
B. Mean percent of breast density by quartiles of circulating androgens among postmenopausal women not on exogenous hormones, Nurses' Health Study (1989-1998)
DHEA
    Age and batch (n = 488)19.718.819.218.60.040.390.009
    +BMI20.018.719.118.60.030.500.25
    +MV19.918.818.818.80.040.370.29
DHEA sulfate
    Age and batch (n = 512)19.820.218.219.60.0090.830.008
    +BMI20.519.818.319.30.020.690.26
    +MV20.020.018.719.30.030.560.31
Testosterone
    Age and batch (n = 505)20.418.320.417.7−0.0090.850.007
    +BMI19.418.920.218.30.0020.960.25
    +MV18.719.320.218.60.010.830.30
Free testosterone§
    Age and batch (n = 485)21.822.316.716.0−0.120.0080.01
    +BMI19.521.217.318.80.0060.900.25
    +MV19.320.618.218.70.010.750.30
Androstenedione
    Age and batch (n = 501)19.219.321.416.7−0.0130.770.008
    +BMI19.118.421.617.50.030.520.25
    +MV18.918.821.017.90.040.360.30
C. Mean percent of breast density by quartiles of circulating SHBG, progesterone, and prolactin among postmenopausal women not on exogenous hormones, Nurses' Health Study (1989-1998)
SHBG
    Age and batch (n = 500)14.118.020.925.60.21<0.00010.04
    +BMI18.718.219.022.40.0080.860.26
    +MV18.318.418.822.70.010.810.30
Progesterone
    Age and batch (n = 441)19.420.719.219.90.020.690.01
    +BMI19.721.619.419.1−0.020.640.29
    +MV19.520.619.419.6−0.0010.980.33
Prolactin
    Age and batch (n = 493)16.619.920.321.20.070.110.007
    +BMI16.819.919.921.40.090.040.26
    +MV17.120.020.020.80.070.120.31
  • * P value for partial Spearman correlation between hormone and percent mammographic density.

  • R2 based on regression models of continuous hormone measurements regressed on square root–transformed average percent breast density.

  • Adjusted for the following variables at the time of blood collection: age (continuous), laboratory batch, alcohol consumption (none, <5, 5-14.9, and 15+ g/d), age at first birth (nulliparous, age at first birth <25, 25-29, and 30+ years), BMI (continuous), family history of breast cancer (yes/no), history of benign breast disease (yes/no), and prior use of postmenopausal hormones (yes/no).

  • § Calculated using Sodergard's method.