Cancer antigen 15-3/mucin 1 Levels in CCTG MA.32: a breast cancer randomized trial of metformin vs placebo

Pamela J Goodwin, Ryan J O Dowling, Marguerite Ennis, Bingshu E Chen, Wendy R Parulekar, Lois E Shepherd, Karen A Gelmon, Timothy J Whelan, Jennifer A Ligibel, Dawn L Hershman, Ingrid A Mayer, Timothy J Hobday, Priya Rastogi, Manuela Rabaglio-Poretti, Julie Lemieux, Alastair M Thompson, Daniel W Rea, Vuk Stambolic

Research output: Contribution to journalArticlepeer-review

127 Downloads (Pure)


Background: Circulating levels of cancer antigen (CA) 15-3, a tumor marker and regulator of cellular metabolism, were reduced by metformin in a nonrandomized neoadjuvant study. We examined the effects of metformin (vs placebo) on CA 15-3 in participants of MA.32, a phase III randomized trial in early-stage breast cancer.

Methods: A total of 3649 patients with T1-3, N0-3, M0 breast cancer were randomly assigned; pretreatment and 6-month on-treatment fasting plasma were centrally assayed for CA 15-3. Genomic DNA was analyzed for the rs11212617 single nucleotide polymorphism. Absolute and relative change of CA 15-3 (metformin vs placebo) were compared using Wilcoxon rank and t tests. Regression models adjusted for baseline differences and assessed key interactions. All statistical tests were 2-sided.

Results: Mean (SD) age was 52.4 (10.0) years. The majority of patients had T2/3, node-positive, hormone receptor-positive, HER2-negative breast cancer treated with (neo)adjuvant chemotherapy and hormone therapy. Mean (SD) baseline CA 15-3 was 17.7 (7.6) and 18.0 (8.1 U/mL). At 6 months, CA 15-3 was statistically significantly reduced in metformin vs placebo arms (absolute geometric mean reduction in CA 15-3 = 7.7% vs 2.0%, P < .001; relative metformin: placebo level of CA 15-3 [adjusted for age, baseline body mass index, and baseline CA 15-3] = 0.94, 95% confidence interval = 0.92 to 0.96). This reduction was independent of tumor characteristics, perioperative systemic therapy, baseline body mass index, insulin, and the single nucleotide polymorphism status (all Ps > .11).

Conclusions: Our observation that metformin reduces CA 15-3 by approximately 6% was corroborated in a large placebo-controlled randomized trial. The clinical implications of this reduction in CA 15-3 will be explored in upcoming efficacy analyses of breast cancer outcomes in MA.32.

Original languageEnglish
Article numberpkab066
JournalJNCI cancer spectrum
Issue number5
Early online date28 Jul 2021
Publication statusPublished - Oct 2021


Dive into the research topics of 'Cancer antigen 15-3/mucin 1 Levels in CCTG MA.32: a breast cancer randomized trial of metformin vs placebo'. Together they form a unique fingerprint.

Cite this