Differences in prior estrogen exposure may explain the difference in magnitude of breast cancer risk reduction observed in raloxifene breast cancer prevention trials

Abstract

1519 Background: Sep 2007, raloxifene (RLXF) approved for invasive breast cancer (IBC) risk reduction (RR) in postmenopausal women (PMW) with osteoporosis (OP) based on two placebo-controlled trials: RLXF Use for The Heart (RUTH) in PMW at increased risk of coronary events and Multiple Outcomes of RLXF Evaluation (MORE) in PMW with OP, later: Continuing Outcomes Relevant to Evista (CORE). Each trial showed statistically significant reduction in IBC risk, but Number Needed to Treat (NNT: women who need to take RLXF for 1 year to prevent one IBC) was 862 and 323 in RUTH and MORE: a difference >500. Impressive MORE results, attributed to RLXF, later on, led to Study of Tamoxifen and Raloxifene (STAR trial). Methods: IRs of IBC in RLXF and placebo (PLCB) arms of RUTH and MORE were compared and trial publications searched for history of prior estrogen exposure. Results: IRs of IBC in RUTH and MORE = 2.66 and 4.36 in PLCB arms & 1.50 and 1.26 in RLXF arms; corresponding absolute RR = 1.16 and 3.10 per 1,000 person-years, respectively. IRs of IBC are similar in RLXF arms but remarkably different in PLCB arms, and derive the observed difference in RR magnitude: 2.66 to 1.50 (RUTH) vs 4.36 to 1.26 (MORE). Prior to enrollment, a smaller fraction of women was on estrogen replacement in RUTH than MORE: 20% vs 28%; differences in duration of estrogen therapy and estrogen exposure free time prior to randomization are not known. Conclusions: It is the higher IBC IR in MORE PLCB arm that seems to derive the effect size of the observed risk reduction. High IBC IR in MORE PLCB arm is unexpected: publications show lower IBC incidence in women with OP (attributed to lower lifetime estrogen exposure) or no difference. Differences in prior OP estrogen treatment might have led to different IBC IRs in RUTH and MORE PLCB arms, ie, the observed remarkable RR of IBC in MORE resulted from interplay of two factors: 1) Persisting estrogen exposure effect in MORE PLCB arm with the consequent higher than expected IBC IR, & 2) Starting RLXF in RLFX arms. Possibility remains that impressive RR in MORE was a random finding: IBC IR was a safety endpoint in MORE. Notably, since Women's Health Initiative publication, lowering of IBC IR has been reported: attributed to a decline in estrogen use. No significant financial relationships to disclose.