The Effect of Metformin on Chemotherapy-Induced Toxicities in Non-diabetic Breast Cancer Patients: A Randomised Controlled Study

This is a multicentre randomised controlled trial conducted at the Medical Research Institute and Alexandria Main University Hospital, Alexandria University, Egypt. Institutional ethical approvals by Alexandria University (IRB no. 00012098), and Damanhour University (IRB no. 919PP18) were obtained before commencement of the study. The study was conducted in compliance with the Declaration of Helsinki. All study participants provided a written informed consent. This study is a part of the METNEO study registered in ClinicalTrials.gov under registration number: NCT04170465 (https://​clinicaltrials.​gov/​ct2/​show/​NCT04170465).

Patients were eligible if they were aged 18–65 years, candidates for neoadjuvant chemotherapy, received no prior chemotherapy, and had baseline left ventricular ejection fraction (LVEF) greater than 50%, normal liver functions, and Eastern Cooperative Oncology Group Performance Status (ECOG-PS) from zero to two. Patients with diabetes, history of anti-diabetic medication use, pregnancy, breastfeeding, metastatic or recurrent breast cancer, or increased risk of metformin-induced lactic acidosis, such as those with heart failure or estimated glomerular filtration rate (eGFR) ≤ 45 mL/min/1.73 m² were excluded from the study. Fasting blood glucose level test was done for patients with no history of diabetes to make sure they were not underdiagnosed and were eligible for participation in the study. Also, baseline LVEF%, renal and liver function, and complete blood count with differentials were determined.

Eligible patients were randomised in a 1:1 ratio to either metformin or control arms using blocked randomisation with random block sizes of two, four, and six. The randomisation scheme was generated by using the website “http://​www.​randomization.​com”. Allocation to either arm was concealed using sealed numbered envelopes by an independent third-party individual. The control arm received AC-T chemotherapy 16 cycles protocol (i.e., doxorubicin 60 mg/m² slow intravenous [IV] push over 10–15 min and cyclophosphamide 600 mg/m² IV infusion over 1 h every 21 days for a total of four cycles followed by paclitaxel 80 mg/m² IV infusion over 1 h every week for a total of 12 cycles). The metformin arm received the same chemotherapy protocol in addition to oral metformin 850 mg once daily for 1 week followed by 850 mg twice a day during the entire neoadjuvant treatment period and stopped metformin only 3–7 days before surgery. Metformin was slowly titrated and was prescribed with meals to reduce the incidence of associated gastrointestinal (GI) side effects. Moreover, vitamin B12 was supplemented to study participants when peripheral neuropathy symptoms develop in both metformin and control arms.

Patients were interviewed in person and data about birthdate, weight, height, ECOG-PS, menopausal status, history of gestation, pregnancy and abortion, breastfeeding, method of contraception, other comorbidities, and family history were collected.

Complete blood picture with differentials was obtained, and patients were monitored for the incidence of haematological side effects (i.e., anaemia, thrombocytopenia, neutropenia, and febrile neutropenia) before each of the 16 cycles. Also, non-haematological side effects (i.e., fatigue, nausea, vomiting, diarrhoea, constipation, oral mucositis, and peripheral sensory neuropathy) were monitored after each chemotherapy session during the whole treatment period. At baseline, echocardiography and ultrasonography for abdomen and pelvis were done. Moreover, renal and liver function tests were assessed before each cycle.

Before the commencement of neoadjuvant therapy, patients were interviewed in person and patient education about the proper measures for preventing and self-management of oral mucositis was provided equally in both arms. After each chemotherapy session, each participant was assessed within the first week and again after 18–21 days from the session by an independent physician and the study investigator by telephone or face to face. Patients were assessed for the occurrence of any AE based on the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 5.0 [30], using an assessment form extracted from the NCI-CTCAE v.5 that contains the CTCAE terms for the studied AEs and their severity grades from 1 to 5 with a checkbox underneath each grade, and a key providing the definitions for the AEs and the corresponding severity grades (Electronic supplementary file 1). Moreover, echocardiography and ultrasonography were repeated after completing treatment, to assess the state of the patient before surgery. The haematological toxicities, cardiotoxicity, and hepatotoxicity were assessed by a blind investigator.

All statistical analyses were performed using Stata version 14.2. For interval variables, Shapiro–Wilk test was used to inspect data normality, Levene’s test was used to check equality of variances between arms, followed by the two-sided two-sample independent t test with equal or unequal variances, which was used as appropriate to test for differences between treatment arms. Pearson’s Chi-square, or Fisher’s exact tests were used as appropriate for categorical variables. To detect the difference in the severity grades of different AEs (i.e., peripheral neuropathy, oral mucositis, fatigue, gastrointestinal, and haematological adverse effects) between the two arms at each time point, the Mann–Whitney U test was used, and data were expressed as frequencies and percentages of occurrence of different severity grades. Then ordered logistic regression was performed to detect the effect size of treatment per cycle, and the odds ratio (OR) with its 95% confidence intervals (CI) was reported. On the other hand, the paired t test was used to compare pre-treatment to post-treatment values of LVEF% in each arm, and data were expressed as means ± SD, then the multiple linear regression was performed to detect the effect size of metformin and baseline LVEF% on post-treatment LVEF% values, and the β-coefficient with its 95% CIs was reported. Binary logistic regression was used to estimate the effect size of metformin treatment and other parameters on fatty liver development, and the OR with its 95% CIs was reported. Two-sided p values of < 0.05 were considered significant. The sample size was estimated based on a previous study [31], in which 60 patients (i.e., 30 per arm) were needed to detect a change in the incidence of peripheral neuropathy using Chi-square test with an alpha error of 5%, a beta error of 10%, and a 20% drop-out rate. As indicated above, peripheral neuropathy is one of the most common toxicities encountered in BC patients treated with AC-T chemotherapy [32, 33].

Between November 2019 and February 2021, 663 patients with biopsy-proven BC were screened for eligibility to participate in the study. Of these, 586 patients did not meet the inclusion criteria and 7 patients refused to participate. Seventy patients met the eligibility criteria and were randomised to either metformin or control arms in a 1:1 ratio. All patients were assessed for study outcomes. The CONSORT diagram summarises the study flow (Fig. 1).

Patient and tumour characteristics are summarised in Table 1. Patients in both metformin and control arms showed similar baseline characteristics. Most females were aged < 40 years and were premenopausal (68.57% and 71.43%, respectively). Most patients were obese, representing 67.14% of participants. Most patients (67.14%) had no family history of BC. Moreover, the luminal molecular subtype (65.71%) and clinical prognostic stage of IIIB (44.29%) were more prevalent among patients.

The frequency and proportion of AEs during the whole treatment period were summarised in Table 2. No AE-related deaths (i.e., grade 5) were observed during the entire study. The most frequently occurring AEs observed were anaemia, fatigue, neutropenia, peripheral sensory neuropathy, oral mucositis, nausea, and diarrhoea (Table 2). Most AEs were of grades 1 and 2. The metformin arm showed a lower incidence and severity of fatigue, oral mucositis, and peripheral neuropathy (Electronic supplementary file 2, Fig. 1). On the other hand, no significant differences were observed between the two arms in diarrhoea, constipation, nausea, vomiting, anaemia, thrombocytopenia, neutropenia, and febrile neutropenia.

This study was an open-label randomised controlled study. However, the implementation of allocation concealment and the blind outcome assessment for haematological toxicity, cardiotoxicity and hepatotoxicity may have reduced bias in the estimated effect of treatment. Also, results may have been reinforced by measuring specific confirmatory biomarkers for AC-T-induced toxicities.