Twelve-Month Estrogen Levels in Premenopausal Women With Hormone Receptor-Positive Breast Cancer Receiving Adjuvant Triptorelin Plus Exemestane or Tamoxifen in the Suppression of Ovarian Function Trial (SOFT): The SOFT-EST Substudy

Abstract

Purpose: To describe estradiol (E2), estrone (E1), and estrone sulfate (E1S) levels during the first year of monthly triptorelin plus exemestane or tamoxifen and to assess possible suboptimal suppression while receiving exemestane plus triptorelin.

Patients and methods: Premenopausal patients with early breast cancer on the Suppression of Ovarian Function Trial who selected triptorelin as the ovarian suppression method and were randomly assigned to exemestane plus triptorelin or tamoxifen plus triptorelin were enrolled until the target population of 120 patients was reached. Blood sampling time points were 0, 3, 6, 12, 18, 24, 36, and 48 months. Serum estrogens were measured with a highly sensitive and specific assay. This preplanned 12-month analysis evaluated E2, E1, E1S, follicle-stimulating hormone, and luteinizing hormone levels in all patients and the proportion of patients with E2 levels greater than 2.72 pg/mL at any time point during treatment with exemestane plus triptorelin.

Results: One hundred sixteen patients (exemestane, n = 86; tamoxifen, n = 30; median age, 44 years; median E2, 51 pg/mL; 55% prior chemotherapy) started triptorelin and had one or more samples drawn. With exemestane plus triptorelin, median reductions from baseline E2, E1, and E1S levels were consistently ≥ 95%, resulting in significantly lower levels than with tamoxifen plus triptorelin at all time points. Among patients on exemestane plus triptorelin, 25%, 24%, and 17% had an E2 level greater than 2.72 pg/mL at 3, 6, and 12 months, respectively. Baseline factors related to on-treatment E2 level greater than 2.72 pg/mL were no prior chemotherapy (P = .06), higher body mass index (P = .05), and lower follicle-stimulating hormone and luteinizing hormone (each P < .01).

Conclusion: During the first year, most patients on exemestane plus triptorelin had E2 levels below the defined threshold of 2.72 pg/mL, consistent with levels reported in postmenopausal patients on aromatase inhibitors, but at each time point, at least 17% of patients had levels greater than the threshold.

Fig 1.

Consort diagram of patient flow from random assignment on the parent Suppression of Ovarian Function Trial (SOFT) to inclusion in the analytic cohort for the SOFT Estrogen substudy (SOFT-EST). E, exemestane; OFS, ovarian function suppression; T, tamoxifen.

Fig 2.

Distributions of (A) estradiol (E2), (B) estrone (E1), and (C) estrone sulfate (E1S) over time according to treatment assignment. Boxes indicate the 25th, 50th, and 75th percentiles. (A) The horizontal dashed line indicates the threshold of 2.72 pg/mL. Levels were significantly different at each postbaseline time point (each P < .001).

Fig 3.

Percentages of patients in the exemestane plus triptorelin group with estradiol (E2) values greater than the predefined threshold (> 2.72 pg/mL, which defines a strict threshold to indicate E2 inconsistent with postmenopausal levels on an aromatase inhibitor) and greater than two additional exploratory thresholds (> 10 and > 20 pg/mL, representing a less strict threshold above which E2 was clearly inconsistent with postmenopausal levels on an aromatase inhibitor and a threshold above which E2 was inconsistent with gonadotropin-releasing hormone agonist–related postmenopausal status, respectively) at each time point. The number of patients tested at each time point is shown at the bottom of the bars.

Fig 4.

Estradiol (E2) levels over time for 27 patients with at least one E2 level greater than 2.72 pg/mL according to the number (one, two, or three) of postbaseline samples with E2 level greater than 2.72 pg/mL. The horizontal dashed line indicates the threshold of 2.72 pg/mL.

Fig A1.

Estimated mean fold-difference between treatment groups (tamoxifen plus triptorelin relative to exemestane plus triptorelin) for each hormone level over time. Levels (log10-transformed) were modeled using generalized estimating equation (GEE) as a function of time point, treatment assignment, the treatment-by-time interaction, and patient characteristics, accounting for correlation of longitudinal values. Mean fold-differences are plotted with 95% CIs. SE used robust (sandwich) variance calculation. The horizontal dashed line at 1 indicates no difference. NS=not statistically significant.

Fig A2.

Distribution of follicle-stimulating hormone (FSH) and luteinizing hormone (LH) levels over time according to treatment assignment. Boxes indicate the 25th, 50th, and 75th percentiles.

Fig A3.

Distribution of estrogen and gonadotropin levels at each scheduled sampling time grouped by sampling days from last triptorelin injection.

Fig A3.

Distribution of estrogen and gonadotropin levels at each scheduled sampling time grouped by sampling days from last triptorelin injection.