Effect of metformin vs placebo on and metabolic factors in NCIC CTG MA.32

Abstract

Background: Metformin may improve metabolic factors (insulin, glucose, leptin, highly sensitive C-reactive protein [hs-CRP]) associated with poor breast cancer outcomes. The NCIC Clinical Trials Group (NCIC CTG) MA.32 investigates effects of metformin vs placebo on invasive disease-free survival and other outcomes in early breast cancer. Maintaining blinding of investigators to outcomes, we conducted a planned, Data Safety Monitoring Committee-approved, analysis of the effect of metformin vs placebo on weight and metabolic factors at six months, including examination of interactions with baseline body mass index (BMI) and insulin, in the first 492 patients with paired blood samples.

Methods: Eligible nondiabetic subjects with T1-3, N0-3, M0 breast cancer who had completed surgery and (neo)adjuvant chemotherapy (if given) provided fasting plasma samples at random assignment and at six months. Glucose was measured locally; blood was aliquoted, frozen, and stored at -80°C. Paired plasma aliquots were analyzed for insulin, hs-CRP, and leptin. Spearman correlation coefficients were calculated and comparisons analyzed using Wilcoxon signed rank test. All statistical tests were two-sided.

Results: Mean age was 52.1±9.5 years in the metformin group and 52.6 ± 9.8 years in the placebo group. Arms were balanced for estrogen/progesterone receptor, BMI, prior (neo)adjuvant chemotherapy, and stage. At six months, decreases in weight and blood variables were statistically significantly greater in the metformin arm (vs placebo) in univariate analyses: weight -3.0%, glucose -3.8%, insulin -11.1%, homeostasis model assessment -17.1%, leptin -20.2%, hs-CRP -6.7%; all P values were less than or equal to .03. There was no statistically significant interaction of change in these variables with baseline BMI or insulin.

Conclusions: Metformin statistically significantly improved weight, insulin, glucose, leptin, and CRP at six months. Effects did not vary by baseline BMI or fasting insulin.

© The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.

Figures

Figure 1.

Consolidated Standards of Reporting Trials diagram for metabolic substudy: MA.32. HOMA = homeostasis model assessment.

Figure 2.

Correlation of baseline body mass index (BMI) and insulin with baseline metabolic parameters. Scatterplots of (A) baseline BMI with baseline fasting insulin, glucose, leptin, and hs-CRP and (B) baseline insulin with baseline BMI, fasting glucose, leptin, and hs-CRP (“r” = Spearman correlation coefficient, 95% confidence intervals for r are based on Fisher’s Z transformation). All correlation coefficients statistically significantly different from zero (calculated using Fisher’s z-transformation method) (P < .01). BMI = body mass index; CI = confidence interval; hs-CRP = highly sensitive C-reactive protein.

Figure 3.

Correlation of baseline body mass index (BMI) and insulin with change in metabolic parameters. Scatterplots of baseline BMI and fasting insulin with change (at six months) in BMI and fasting insulin, glucose, leptin, and hs-CRP. A) Associations with baseline BMI. B) Associations with baseline fasting insulin (“r” = Spearman correlation coefficient, 95% confidence intervals for r are based on Fisher’s z-transformation). No correlation coefficients were statistically significantly different from zero (calculated using Fisher’s z-transformation method). BMI = body mass index; CI = confidence interval; hs-CRP = highly sensitive C-reactive protein.