Metformin use and mortality in patients with advanced chronic kidney disease: national, retrospective, observational, cohort study

Summary

Background

Metformin is recommended as a first-line treatment for patients with type 2 diabetes. However, use of this drug has been contraindicated in individuals with impaired kidney function because of the perceived risk of lactic acidosis. Evidence now supports cautious use of metformin in people with mild-to-moderate chronic kidney disease. However, studies examining the use of metformin in patients with advanced chronic kidney disease are lacking. We aimed to assess the safety of metformin in patients with type 2 diabetes and advanced (approximately stage 5) chronic kidney disease.

Methods

We did a retrospective, observational, cohort study of patients with type 2 diabetes who were enrolled prospectively in Taiwan's national health insurance research database between Jan 1, 2000, and June 30, 2009, and had follow-up data until Dec 31, 2009. We included individuals with a serum creatinine concentration greater than 530 μmol/L, which is approximately equivalent to stage 5 chronic kidney disease. From a consecutive sample of 12 350 patients with type 2 diabetes and chronic kidney disease, 1005 used metformin and 11 345 were non-users. We matched users and non-users of metformin by propensity score in a 1:3 ratio. Our primary outcome was all-cause mortality.

Findings

813 metformin users were matched by propensity score to 2439 non-users. The two groups of patients did not differ significantly by 30 baseline clinical and socioeconomic variables. Median follow-up in the matched cohort was 2·1 years (range 0·3–9·8). All-cause mortality was reported in 434 (53%) of 813 metformin users and in 1012 (41%) of 2439 non-users. After multivariate adjustment, metformin use was an independent risk factor for mortality (adjusted hazard ratio 1·35, 95% CI 1·20–1·51; p<0·0001). The increased mortality risk was dose-dependent and was consistent across all subgroup analyses. However, metformin use compared with no use was associated with a higher but non-significant risk of metabolic acidosis (1·6 vs 1·3 events per 100 patient-years; adjusted hazard ratio 1·30, 95% CI 0·88–1·93; p=0·19).

Interpretation

Use of metformin in people with type 2 diabetes and a serum creatinine concentration greater than 530 μmol/L is associated with a significantly increased risk of all-cause mortality compared with non-users. Metformin use should not be encouraged in this patient group.

Funding

Taipei Tzu Chi Hospital, Taipei Veterans General Hospital, Foundation for Poison Control, National Health Research Institutes, Ministry of Science and Technology of Taiwan.

Introduction

The prevalence of chronic kidney disease in patients with type 2 diabetes has risen correspondingly with the prevalence of type 2 diabetes over the past two decades.¹ The coexistence of chronic kidney disease confers further increased risk for cardiovascular events and mortality in patients with diabetes.² However, the best treatment strategy for type 2 diabetes in patients who also have chronic kidney disease has not been well defined.

Metformin is a biguanide antidiabetic drug that decreases hepatic glucose production, delays intestinal glucose absorption, and improves insulin sensitivity. In the UK Prospective Diabetes Study (UKPDS), a substantial beneficial effect of metformin was noted in patients with type 2 diabetes, with a relative risk reduction of about a third in all-cause mortality and myocardial infarction, compared with individuals treated with lifestyle changes alone.3, 4 Metformin is generally recommended as the initial pharmacological agent for individuals with type 2 diabetes because of its low cost and potential cardiovascular protection.⁵ However, the drug has been contraindicated in patients with chronic kidney disease because of concerns about the risk of lactic acidosis.⁶ In 1994, the US Food and Drug Administration (FDA) approved metformin but contraindicated its use in men with a serum creatinine concentration of 133 μmol/L or higher, and in women with a serum creatinine concentration of 124 μmol/L or higher, because of concerns about the risk of lactic acidosis.

Metformin is mainly excreted unchanged by the kidney; thus, in people with chronic kidney disease, the drug could accumulate and might cause lactic acidosis. In a retrospective analysis of 77 601 patients with type 2 diabetes in the UK treated with metformin,⁷ most individuals (92%) had some level of renal impairment. However, the incidence of lactic acidosis was very low (10·4 per 100 000 patient-years) and no significant difference was noted among patients with normal, mildly reduced, moderately reduced, and severely reduced kidney function.⁷ Analysis of the Reduction of Atherothrombosis for Continued Health (REACH) registry⁸ suggested that the proposed cardiovascular benefits of metformin could extend to individuals with established atherosclerosis and moderate chronic kidney disease (an estimated glomerular filtration rate [eGFR] of 30–60 mL/min per 1·73 m²). After decades of experience with metformin, these data and safety profiles have led to relaxation of clinical practice to permit use of this drug in patients with mild-to-moderate (stages 1–3) chronic kidney disease (eGFR ≥30 mL/min per 1·73 m²).9, 10 Now, debate is ongoing about whether metformin can be prescribed more widely to people with advanced (stage 4 or 5) chronic kidney disease (eGFR <30 mL/min per 1·73 m²).11, 12, 13 To our knowledge, no studies have assessed the risks and benefits of metformin in this population.¹⁴

Research in context

Evidence before this study

We searched PubMed, with no restriction on date or language, with the keywords “metformin”, “chronic kidney disease”, “mortality”, and “lactic acidosis”. Although metformin is contraindicated in patients with advanced (stages 4–5) chronic kidney disease because of concerns about lactic acidosis, data are scant to support this advice. Moreover, the potential for a higher risk of death with metformin use in patients with advanced chronic kidney disease has not been investigated. We identified two large-scale studies from 2014 that included metformin users with stage 4–5 chronic kidney disease. In a study by Richy et al, the incidence of lactic acidosis did not differ among patients with type 2 diabetes and various levels of kidney function who were treated with metformin, whereas in another study by Eppenga et al, a greater risk of lactic acidosis was reported with estimated glomerular filtration rates less than 60 mL/min per 1·73 m². Neither study reported data for mortality.

Added value of this study

To the best of our knowledge, our study is the first in patients with type 2 diabetes and advanced (approximately stage 5) chronic kidney disease to investigate the association between metformin and clinical outcomes. Metformin use in this population was associated with a significantly higher risk of death from all causes compared with non-users, but no significant difference was noted in the occurrence of metabolic acidosis.

Implications of all the available evidence

An increased risk of mortality associated with metformin use in patients with advanced chronic kidney disease is possible and should be addressed further in future studies. Our study findings have important therapeutic implications and support current recommendations restricting metformin use to patients with mild-to-moderate (stages 1–3) chronic kidney disease.

Perhaps because of the favourable effects of metformin on cardiovascular outcomes, and the notion that the evidence supporting existing precautions was inadequate, metformin was prescribed to patients in Taiwan without renal contraindication until 2009, when its use was restricted by the Taiwan Food and Drug Administration to men and women with serum creatinine concentrations of less than 133 μmol/L and less than 124 μmol/L, respectively. This delay in changing guidance provides a unique opportunity to assess the safety of metformin in patients with type 2 diabetes and advanced chronic kidney disease in a national population-based cohort.