Fig. 1
HR and 95%CI for B2 (A), M (B), G2 (C) and VB (D) stratified by urinary equol levels. Data are HR with 95% confidence intervals. Cox proportional hazard regression models were used, adjusted for parental education level, mother’s age at menarche, energy intake at baseline, dietary fibre intakes (residuals) at baseline and percent body fat at baseline, with the group of those in both lowest tertile of total soy intakes and lower equol level (< median values) serving as the reference group. p for interactions refers to the 2-way interactions of urinary equol level on the relations between dietary soy intake and puberty markers. A, B Ngirl=589. Range of total soy intakes: 1st tertile (1.5–11.3), 2nd tertile (12.1–38.9) and 3rd tertile (39.3–68.6). Participants in groups: low soy, low equol: n=109; low soy, high equol: n=103; medium soy, low equol: n=95; medium soy, high equol: n=86; high soy, low equol: n=108; high soy, high equol: n=88. C, D Nboy=722. Range of soybean intakes: 1st tertile (0–4.5), 2nd tertile (5.2–46.4) and 3rd tertile (50.2–80.9). Participants in groups: low soy, low equol: n=146; low soy, high equol: n=121; medium soy, low equol: n=123; medium soy, high equol: n=98; high soy, low equol: n=126; high soy, high equol: n=108