Effect of metformin versus placebo on metabolic factors in the MA.32 randomized breast cancer trial

Abstract

Metformin may exert anticancer effects through indirect (mediated by metabolic changes) or direct mechanisms. The goal was to examine metformin impact on metabolic factors in non-diabetic subjects and determine whether this impact varies by baseline BMI, insulin, and rs11212617 SNP in CCTG MA.32, a double-blind placebo-controlled randomized adjuvant breast cancer (BC) trial. 3649 subjects with T1-3, N0-3, M0 BC were randomized; pretreatment and 6-month on-treatment fasting plasma was centrally assayed for insulin, leptin, highly sensitive C-reactive protein (hsCRP). Glucose was measured locally and homeostasis model assessment (HOMA) calculated. Genomic DNA was analyzed for the rs11212617 SNP. Absolute and relative change of metabolic factors (metformin versus placebo) were compared using Wilcoxon rank and t-tests. Regression models were adjusted for baseline differences and assessed interactions with baseline BMI, insulin, and the SNP. Mean age was 52 years. The majority had T2/3, node positive, hormone receptor positive, HER2 negative BC treated with (neo)adjuvant chemotherapy and hormone therapy. Median baseline body mass index (BMI) was 27.4 kg/m² (metformin) and 27.3 kg/m² (placebo). Median weight change was -1.4 kg (metformin) vs +0.5 kg (placebo). Significant improvements were seen in all metabolic factors, with 6 month standardized ratios (metformin/placebo) of 0.85 (insulin), 0.83 (HOMA), 0.80 (leptin), and 0.84 (hsCRP), with no qualitative interactions with baseline BMI or insulin. Changes did not differ by rs11212617 allele. Metformin (vs placebo) led to significant improvements in weight and metabolic factors; these changes did not differ by rs11212617 allele status.

Fig. 1. CONSORT diagram—identification of subjects included in the metabolic and metabolic SNP populations from the full MA.32 study population.

SNP single-nucleotide polymorphism.

Fig. 2. Assessment of differential study drug effects on the metabolic outcomes weight/BMI, glucose, insulin, HOMA, leptin, and hsCRP, by baseline BMI, insulin, and the rs11212617 SNP.

Depicted is the standardized metformin to placebo ratio of the metabolic factor at 6 months, with 95% confidence intervals, obtained from adjusted interaction regression models. BMI body mass index, SNP single-nucleotide polymorphism, Intx interaction. a Interaction of study drug by baseline BMI (continuous). b Interaction of study drug by baseline insulin (continuous). c Interaction of study drug by genotype any C versus AA of the rs11212617 SNP.