Does metformin improve the efficacy of standard epidermal growth factor receptor-tyrosine kinase inhibitor treatment for patients with advanced non-small-cell lung cancer?

Abstract

A best evidence topic in thoracic surgery was written according to a structured protocol. The question addressed was whether metformin improved the efficacy of standard epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) treatment for patients with epidermal growth factor receptor (EGFR)-mutated advanced non-small-cell lung cancer. A total of 99 papers were found using the reported search, of which 4 represented the best evidence to answer this clinical question. The authors, journal, publication date, country, study type, treatment regimen, relevant outcomes and results of these papers are tabulated. We concluded that the addition of metformin to EGFR-TKI might improve the survival of patients with EGFR-mutated non-small-cell lung cancer and diabetes mellitus type 2. However, for non-diabetic non-small-cell lung cancer patients with EGFR mutation, the efficiency of additional metformin in EGFR-TKI treatment remains unclear because of the conflicting results of only 2 available studies.

INTRODUCTION

A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].

THREE-PART QUESTION

In [patients with epidermal growth factor receptor (EGFR)-mutated advanced non-small-cell lung cancer (NSCLC) with or without diabetes mellitus], does [additional metformin plus EGFR-tyrosine kinase inhibitor (TKI)] provide [superior outcomes including progression-free survival (PFS) and overall survival (OS)]?

CLINICAL SCENARIO

A 59-year-old male patient with advanced NSCLC came to your hospital to enquire about his treatments. After a thorough evaluation of the patient, you found him harbouring activating EGFR mutations and did not suffer from diabetes. You recommended him to receive EGFR-TKI treatment. At this time, 1 of your colleagues suggested adding metformin to EGFR-TKI treatment to improve the survival of patients with or without diabetes mellitus. You were unsure about this recommendation and decided to check further.

SEARCH STRATEGY

We searched the PubMed, Medline (via Ovid), Embase (via Ovid), Web of science and Cochrane library from January 2000 to October 2019 with the following search terms: (epidermal growth factor receptor) AND (tyrosine kinase inhibitor) AND (lung cancer) AND (non-small cell) AND (metformin or dimethylbiguanidine or dimethylguanylguanidine or glucophage).

SEARCH OUTCOME

A total of 99 papers were found using the reported search. From these, 4 papers were identified as the best evidence to answer the question. These are presented in Table 1.

Table 1:

Best evidence papers

Author, date, journal and country Study type (level of evidence)	Patient group	Outcomes	Key results (metformin vs control)	Comments
Arrieta et al. (2019), JAMA Oncol, Mexico [2]   RCT (level 2)	Trial information   Study period: 2016–2017   Sample size: n = 139   Metformin group: n = 69   Control group: n = 70   Median follow-up: 16.9 months (95% CI, 7.0–24.9)   Patients   Stage IIIB or IV LUAD   EGFRm: positive   Previous anti-cancer treatment: NA   History of diabetes: none   Intervention   Metformin group: erlotinib, afatinib or gefitinib at standard dosage, plus metformin 500 mg, p.o., b.i.d.   Control group: erlotinib, afatinib or gefitinib at standard dosage	Median PFS	13.1 months (95% CI 9.8–16.3) vs 9.9 months, (95% CI 7.5–12.2)   HR 0.60, 95% CI 0.40–0.94, P = 0.03	This RCT was open labelled and only enrolled non-diabetic NSCLC patients. The allocation in this study was not masked   All patients in this study were EGFR-TKI naive. However, prior anti-cancer treatments were unavailable among the entire population   In the metformin group, the discontinuation rate of metformin usage was 23.2%   The intention-to-treat analysis was not performed
		Median OS	31.7 months (95% CI 20.5–42.8) vs 17.5 months (95% CI 11.4–23.7)   HR 0.52, 95% CI 0.30–0.90, P = 0.04
		ORR	71.0% vs 54.3%   OR 0.48, 95% CI 0.24–0.97, P = 0.04
		DCR	98.6% vs 92.9%   OR 0.19, 95% CI 0.02–1.68, P = 0.14
Li et al. (2019), Clin Cancer Res, China [3]   RCT (level 2)	Trial information   Study period: 2013–2015   Sample size: n = 202   Metformin group: n = 97   Control group: n = 105   Median follow-up: 19.15 months (IQR 12.99–28.44)   Patients   Stage IIIB or IV NSCLC   EGFRm: positive   Previous anti-cancer treatment: none   History of diabetes: none   Intervention   Metformin group: gefitinib, 250mg, p.o., q.d., plus metformin, from 500 to 1000 mg, p.o., b.i.d.   Control group: gefitinib, 250 mg, p.o., q.d., plus placebo, from 500 to 1000 mg, p.o., b.i.d.	Median PFS	10.3 months (95% CI 8.4–13.0) vs 11.4 months (95% CI 10.0–12.2)   HR 1.04, 95% CI 0.75–1.45, P = 0.8087	This RCT only studied non-diabetic NSCLC patients   All the enrolled patients were EGFR-TKI naive and never received any previous anti-cancer treatment   The status of LKB1 expression was not detected   The discontinuation rate of metformin usage was 3.1%   The statistical analysis was based on the intention-to-treat population
		1-year PFS	41.2% (95% CI 30.0–52.2) vs 42.9% (95% CI 32.6–52.7), P = 0.6268
		Median OS	22.0 months (95% CI 19.0–31.5) vs 27.5 months (95% CI 22.8–31.5)   HR 1.15, 95% CI 0.79–1.68, P = 0.4571
		ORR	66.0% (95% CI 55.7–75.3) vs 66.7% (95% CI 56.8–75.6)   OR 0.97, 95% CI 0.52–1.81, P = 1.000
		DCR	97.9% (95% CI 92.7–99.7) vs 97.1% (95% CI 91.9–99.4)   OR 1.40, 95% CI 0.16–17.04, P = 1.000
		AE incidence	91.89% vs 82.88%a
		Grade 3–4 AEb	23.42% vs 18.92%a
Chen et al. (2015), Cancer Lett, China [4]   Retrospective observational study (level 3)	Study information   Study period: 2006–2014   Sample size: n = 90   Metformin group: n = 44   Control group: n = 46   Patients   Stage IIIA, IIIB or IV NSCLC   EGFRm: positive   Previous anti-cancer treatment: none   History of diabetes   Type of diabetes: T2DM   Severity: each patient’s diabetes was well treated   Exposure  b   Metformin group: erlotinib, gefitinib or icotinib, plus metformin   Control group: erlotinib, gefitinib or icotinib, plus non-metformin hypoglycemic agents	Median PFS	19.0 months (95% CI 16.4–21.6) vs 8.0 months (95% CI 5.2–10.8)   HR 0.46, 95% CI 0.28–0.75, P = 0.005	This study focused on NSCLC patients with T2DM   The sample size (90 patients) was relatively small   Bias from the comparison should be considered because some patients used insulin in the control group. Previous studies have indicated that insulin might increase the risk of tumour growth [5]
		Median OS	32.0 months (95% CI 26.5–37.5) vs 23.0 months (95% CI 9.6–36.4)   HR 0.44, 95% CI 0.26–0.76, P = 0.002
		ORR	70.5% (95% CI 55–83) vs 45.7% (95% CI 31–61), P = 0.017
		DCR	97.7% (95% CI 88–100) vs 80.4% (95% CI 66–91), P = 0.009
Hung et al. (2019), Integr Cancer Ther, China [6]   Retrospective observational study, (level 3)	Study information   Study period: 2004–2012   Sample size: n = 1633   Metformin group: n = 373   Control group: n = 1260   Patients   Stage IIIB or IV NSCLC   EGFRm: positive   Anti-cancer treatment (non-EGFR-TKI): chemotherapy, radiotherapy, concurrent chemoradiotherapy or none   History of diabetes   Type of diabetes: T2DM   History of hypoglycemic agents’ usage: patients who received insulin were excluded.   Severity: NA   Exposure   Metformin group: erlotinib, gefitinib or both, plus metformin   Control group: erlotinib, gefitinib or both	Median PFS	9.2 months (95% CI 8.6–11.7) vs 6.4 months (95% CI 5.9–7.2), P < 0.001	This study only focused on NSCLC patients with T2DM   The distribution of comorbidities was a significantly different in the 2 groups   Anti-cancer treatment was heterogeneous between and within the groups
		Median OS	33.4 months (95% CI 29.4–40.2) vs 25.4 months (95% CI 23.7–27.2), P < 0.001
		Comorbidity
		Hypertension	78.3% vs 71.4%, P = 0.0088
		COPD	29.8% vs 39.8%, P = 0.0005
		RI	2.1% vs 5.2%, P = 0.0116
		SRD	17.2% vs 26.3%, P = 0.0003

AE: adverse events; b.i.d.: twice a day; CI: confidence interval; COPD: chronic obstructive pulmonary disease; DCR: disease control rate; EGFR: epidermal growth factor receptor; EGFRm: EGFR mutation; HR: hazard ratio; IQR: interquartile range; LUAD: lung adenocarcinoma; NA: not available; NSCLC: non-small-cell lung cancer; ORR: objective response rate; OR: odd ratio; OS: overall survival; PFS: progression-free survival; p.o.: take orally; q.d.: once a day; RCT: randomized controlled trial; RI: renal insufficiency; SRD: smoking-related disorder; T2DM: diabetes mellitus type 2; TKI: tyrosine kinase inhibitor; TNM: tumour–node–metastasis.

^a n = 111 in each group for safety analysis.

^bAdverse effects were graded according to Common Terminology Criteria for Adverse Events (V.4.0) [7].

RESULTS

Overall, these 4 included papers studied 2 different patient populations. Two prospective randomized controlled trials (RCTs) enrolled non-diabetic patients with EGFR-mutated NSCLC [2, 3], while the other 2 retrospective studies focused on EGFR-mutated NSCLC patients with diabetes mellitus type 2 (T2DM) [4, 6].

Two phase II RCTs [2, 3] compared EGFR-TKI plus metformin with EGFR-TKI alone or plus placebo for non-diabetic patients with EGFR-mutated advanced NSCLC. One of the studies was conducted by Arrieta et al. [2] at the Instituto Nacional de Cancerología, Mexico, from 2013 to 2016. This study enrolled 139 patients with EGFR-mutated stage IIIB–IV lung adenocarcinoma without diabetes mellitus. The results indicated that the addition of metformin to EGFR-TKI (erlotinib, afatinib or gefitinib) treatment significantly improved the median PFS (13.1 vs 9.9 months, P = 0.03), median OS (31.7 vs 17.5 months, P = 0.04) and objective response rate (ORR) (71.0% vs 54.3%, P = 0.04). Multivariable analysis showed that the metformin addition was the only significant factor for improved OS and did not lead to increased adverse events in the additional metformin group. In the metformin group, negative liver kinase 1 (LKB1) expression was associated with poorer OS (not reached vs 6.7 months; P = 0.002). However, conflicting results were found in the double-blind placebo-controlled RCT conducted by Li et al. [3], which enrolled 202 non-diabetic patients with stage IIIB–IV NSCLC harbouring EGFR mutations from 9 centres in China from 2013 to 2015. The results showed that the addition of metformin might not improve the treatment efficiency of gefitinib for non-diabetic patients with EGFR-mutated NSCLC, as evident by the comparable median PFS (10.3 vs 11.4 months, P = 0.8087) and OS (22.0 vs 27.5 months, P = 0.4571), as well as similar ORR (66% vs 66.7%, P = 1.00) and disease control rate (97.9% vs 97.1%, P = 1.00) in both groups. Meanwhile, patients with additional metformin were at a slightly higher risk of suffering from grade 3–4 treatment-related adverse events (23.42% vs 18.92%), according to Common Terminology Criteria for Adverse Events, V.4.0 [7]. Diarrhoea was the most commonly reported adverse effect when taking metformin (78.38% vs 43.24%).

Two retrospective studies [4, 6] investigated the role of metformin plus EGFR-TKI in patients with T2DM and EGFR-mutated advanced NSCLC. All the patients in these 2 studies received metformin or other anti-diabetic therapy before EGFR-TKI treatment. Chen et al. [4] retrospectively studied 90 patients with EGFR-mutated stage IIIA–IV NSCLC and T2DM from 6 hospitals in China from 2006 to 2013. Of these patients, 45 underwent second-line EGFR-TKI treatment. The results indicated that the therapeutic efficiency of EGFR-TKI plus metformin was superior to that of EGFR-TKI plus other hypoglycaemic agents (insulin, sulfonylureas, acarbose and thiazolidinedione) in terms of PFS (19.0 vs 8.0 months, P = 0.005), OS (32.0 vs 23.0 months, P = 0.002), ORR (70.5% vs 45.7%, P = 0.017) and disease control rate (97.7% vs 80.4%, P = 0.009). Of note, subgroup analysis revealed that the addition of metformin had improved the PFS (16.0 vs 8 months, P = 0.031) but failed to prolong the OS (25.0 vs 17.0 months, P = 0.169) for patients treated with second-line EGFR-TKI. Hung et al. [6] retrospectively reviewed 1633 patients with T2DM and stage IIIB–IV NSCLC by using the data from the Taiwan National Health Insurance Research Database from 2004 to 2012. Among these patients, 373 were included in the metformin cohort and 1260 were included in the controlled cohort. When considering the baseline, more patients had hypertension in the metformin group, while fewer patients had chronic obstructive pulmonary disease, renal insufficiency and smoking-related disorders in this group. In addition, fewer patients received chemotherapy and/or radiotherapy in the metformin cohort. This study indicated that metformin plus EGFR-TKI was also associated with improved PFS (9.2 vs 6.4 months, P < 0.001) and OS (33.4 vs 25.4 months, P < 0.001) than single EGFR-TKI treatment.

Another single-arm phase 1/2 trial [8] attempted to study the safety of metformin plus erlotinib for non-diabetic patients with stage IV NSCLC. The results showed that the maximum tolerated dose of metformin was 1500 mg/day. However, this trial focused on the second-line therapy of erlotinib in patients with wild-type EGFR, which was not directly relevant to our question.

CLINICAL BOTTOM LINE

The addition of metformin to EGFR-TKI treatment might improve the outcomes of patients with T2DM and EGFR-mutated advanced NSCLC. Hence, metformin might serve as a suitable hypoglycaemic agent for these patients.

However, for non-diabetic patients with EGFR-mutated advanced NSCLC, the efficiency of additional metformin in EGFR-TKI treatment remains controversial because of the conflicting results from the 2 available studies [2, 3]. Some other important factors, including the heterogeneous patient populations (races, EGFR mutation subtypes, status of LKB1 expression) and previous medical prescriptions for cancer, should also be considered in future studies.