Table 3

Risk ratios (95% CIs) for breast cancer by tertile of serum hormone concentrations

HormoneTertile CategoriesaP for trendb
1 (referent)23
11βOHANo. cases/No. controls15/3418/3620/36
 RRc1.01.2 (0.5–2.8)1.3 (0.5–3.1)0.22
 Adjusted RRd1.01.2 (0.5–3.1)1.4 (0.5–3.6)0.20
AndrostenedioneNo. cases/No. controls7/2412/2324/28
 RRc1.02.0 (0.6–6.3)2.4 (1.0–6.2)0.05e
 Adjusted RRd1.01.6 (0.5–5.8)2.4 (0.9–6.5)0.02e
Δ4A:11βOHAfNo. cases/No. controls11/2113/2215/23
 RRg1.01.1 (0.4–3.0)1.3 (0.5–3.6)0.95
 Adjusted RRh1.01.2 (0.4–3.6)1.2 (0.4–3.5)0.99
  • a Tertile cutpoints were: 11βOHA (nm), 3.3, 5.0; androstenedione (nm), 2.4, 3.4; Δ4A:11βOHA, 0.6, 0.9.

  • b P (2-sided) for trend from model with log-transformed hormone concentration included as a continuous variable.

  • c RRs from conditional logistic regression matched on age, date, and time of day of blood collection.

  • d RRs from conditional logistic regression matched on age, date, and time of day of blood collection and adjusted for years since menopause, height, weight, parity, and family history of breast cancer.

  • e One influential outlier deleted.

  • f Δ4A: 11βOHA, excludes 5 cases and 12 controls who had bilateral oophorectomies.

  • g RRs from logistic regression adjusted for age, date, and time of day of blood collection.

  • h RRs from logistic regression adjusted for age, date, time of day of blood collection, years since menopause, height, weight, parity, and family history of breast cancer.