Skip to main content
. 2016 Feb 29;7(3):178–187. doi: 10.1007/s12672-016-0258-1

Table 4.

Odds ratios for endometrial cancer associated with androgen and SHBG concentrations on the continuous scale

All women Type I endometrial cancer Age <55 years at diagnosis Age ≥55 years at diagnosis Age at diagnosis p-interactionc
Total testosterone
 N cases/controls 155/284 103/189 49/86 106/198
 OR (95 % CI)a 1.48 (1.04, 2.12) 1.46 (0.95, 2.23) 0.93 (0.55, 1.58) 2.09 (1.29, 3.39)
p valuea 0.03 0.08 0.79 0.003 0.03
 BMI-adjusted OR (95 % CI)b 1.34 (0.93, 1.94) 1.31 (0.84, 2.03) 0.83 (0.47, 1.45) 2.08 (1.25, 3.44)
 BMI-adjusted p valueb 0.12 0.24 0.51 0.005
Free testosterone
 N cases/controls 130/228 87/151 47/81 83/147
 OR (95 % CI)a 1.44 (1.10, 1.90) 1.44 (1.04, 1.99) 1.03 (0.68, 1.58) 1.81 (1.25, 2.63)
p valuea 0.009 0.03 0.88 0.002 0.049
 BMI-adjusted OR (95 % CI)b 1.24 (0.92, 1.67) 1.23 (0.86, 1.75) 0.92 (0.57, 1.48) 1.55 (1.04, 2.31)
 BMI-adjusted p valueb 0.16 0.25 0.73 0.03
DHEAS
 N cases/controls 160/300 109/205 51/93 109/207
 OR (95 % CI)a 1.13 (0.89, 1.44) 1.12 (0.84, 1.50) 0.86 (0.59, 1.23) 1.38 (1.00, 1.90)
p valuea 0.33 0.44 0.40 0.05 0.05
 BMI-adjusted OR (95 % CI)b 1.03 (0.80, 1.32) 1.01 (0.74, 1.36) 0.81 (0.55, 1.19) 1.27 (0.92, 1.76)
 BMI-adjusted p valueb 0.81 0.97 0.29 0.15
Androstenedione
 N cases/controls 159/298 108/203 50/92 109/206
 OR (95 % CI)a 1.15 (0.78, 1.69) 1.11 (0.70, 1.75) 0.66 (0.33, 1.32) 1.48 (0.93, 2.36)
p valuea 0.48 0.66 0.24 0.10 0.06
 BMI-adjusted OR (95 % CI)b 1.06 (0.71, 1.58) 0.97 (0.60, 1.57) 0.62 (0.31, 1.27) 1.41 (0.86, 2.30)
 BMI-adjusted p valueb 0.78 0.89 0.19 0.17
SHBG
 N cases/controls 136/247 93/168 49/89 87/158
 OR (95 % CI)a 0.91 (0.81, 1.01) 0.92 (0.80, 1.04) 0.98 (0.80, 1.20) 0.88 (0.77, 1.00)
p valuea 0.08 0.18 0.84 0.05
 BMI-adjusted OR (95 % CI)b 0.98 (0.87, 1.11) 1.00 (0.86, 1.16) 1.01 (0.81, 1.25) 0.98 (0.84, 1.14)
 BMI-adjusted p valueb 0.75 0.99 0.95 0.76 0.41

All analyses were performed on log2 transformed measurements, except SHBG for which square-root transformation was used

aConditional logistic regression (matching factors: cohort, age, and date at blood donation)

bConditional logistic regression (matching factors: cohort, age, and date at blood donation) adjusting for BMI (missing values for BMI were imputed using multiple imputation)

cp-interaction calculated with androgens modeled on the continuous (log2) scale (or square root scale for SHBG)