Table 2.

Adjusted geometric mean hormone levels by frequency of analgesic use in 1990 among postmenopausal women in the Nurses' Health Study

NFrequency of use (d/mo)*Percent differencePtrend
    Maximum n63830013969130
    Estradiol, pg/mL6176.−3.60.37
    Free estradiol, pg/mL5880.−2.70.45
    Estrone, pg/mL48524.124.724.226.510.10.11
    Estrone sulfate, pg/mL616203211234194−4.50.51
    Testosterone, ng/dL62121.422.020.522.02.80.77
    Free testosterone, ng/dL6060.
    Androstenedione, ng/dL48454.456.052.361.312.80.11
    Ratio of estrone/androstenedione4620.470.460.480.470.300.85
    Ratio of estradiol/testosterone5880.360.350.370.33−2.70.24
Nonaspirin NSAIDs
    Maximum n596387953579
    Estradiol, pg/mL5817.−8.90.14
    Free estradiol, pg/mL5530.−11.40.12
    Estrone, pg/mL46424.823.821.324.0−3.10.66
    Estrone sulfate, pg/mL573207230176183−11.30.12
    Testosterone, ng/dL57921.522.519.919.8−7.80.18
    Free testosterone, ng/dL5640.−12.10.09
    Androstenedione, ng/dL46255.−6.80.34
    Ratio of estrone/androstenedione4430.490.450.440.47−1.10.84
    Ratio of estradiol/testosterone5510.360.350.310.34−1.70.44
    Maximum n6063391624758
    Estradiol, pg/mL5887.
    Free estradiol, pg/mL5580.−7.90.47
    Estrone, pg/mL47224.925.424.222.7−8.90.19
    Estrone sulfate, pg/mL583213213176190−10.80.14
    Testosterone, ng/dL58921.021.322.321.73.70.55
    Free testosterone, ng/dL5720.
    Androstenedione, ng/dL47154.457.658.653.0−2.70.79
    Ratio of estrone/androstenedione4510.490.460.470.45−4.30.26
    Ratio of estradiol/testosterone5590.370.350.340.34−3.20.18

NOTE: Adjusted for age at blood draw, laboratory batch, fasting status at blood draw, date and time of blood draw, parity, AFB, BMI at blood draw, physical activity, smoking history, duration of PMH use among past users, age at menopause, alcohol intake, and frequency of use of other analgesics.

  • *Top two frequency categories (15-21 and ≥22 d/mo) are combined due to the small number of women who reported 15 to 21 d of use per month.

  • Percent difference for highest vs lowest category of use (15+ vs 0 d/mo), calculated using (eβ − 1) × 100; no statistically significant differences.

  • Weighted by the midpoint of each frequency category (0, 2.5, 9.5, or 23 d/mo) and calculated using the Wald test.