The Safety and Pharmacokinetics of Metformin in Patients With Chronic Liver Disease

Felicity C. Smith; Sophie L. Stocker; Mark Danta; Jane E. Carland; Shaun S. Kumar; Zhixin Liu; Jerry R. Greenfield; Hannah E. Braithwaite; Tim S. Cheng; Garry G. Graham; Kenneth M. Williams; Richard O. Day


Aliment Pharmacol Ther. 2020;51(5):565-575. 

In This Article

Abstract and Introduction


Background: The FDA approved 'label' for metformin lists hepatic insufficiency as a risk for lactic acidosis. Little evidence supports this warning.

Aims: To investigate the safety and pharmacokinetics of metformin in patients with chronic liver disease (CLD).

Methods: Chronic liver disease patients with and without type 2 diabetes mellitus (T2DM) were studied by a cross-sectional survey of patients already prescribed metformin (n = 34), and by a prospective study where metformin (500 mg, immediate release, twice daily) for up to 6 weeks was prescribed (n = 24). Plasma metformin and lactate concentrations were monitored. Individual pharmacokinetics were obtained and compared to previously published values from healthy and T2DM populations without CLD.

Results: All plasma metformin and lactate concentrations remained below the putative safety thresholds (metformin, 5 mg/L; lactate, 5 mmol/L). Lactate concentrations were unrelated to average steady-state metformin concentrations. In patients with CLD, T2DM was associated with higher plasma lactate concentrations (48% higher than those without T2DM, P < 0.0001). CLD patients with cirrhosis had 23% higher lactate concentrations than those without cirrhosis (P = 0.01). The pharmacokinetics of metformin in CLD patients were similar to patients with T2DM and no liver disease. The ratio of apparent metformin clearance (CLMet/F) to creatinine clearance was marginally lower in CLD patients compared to healthy subjects (median, interquartile range; 12.6, 9.5–15.9 vs 14.9, 13.4–16.4; P = 0.03).

Conclusions: The pharmacokinetics of metformin are not altered sufficiently in CLD patients to raise concerns regarding unsafe concentrations of metformin. There were no unsafe plasma lactate concentrations observed in CLD patients receiving metformin (ACTRN12619001292167; ACTRN12619001348145).


There is a well-documented association between chronic liver disease (CLD) and type 2 diabetes mellitus (T2DM).[1,2] T2DM is associated with the development and progress of liver disease and is a risk factor for hepatocellular carcinoma.[3] There is a need to optimise pharmacotherapy in CLD patients with T2DM to ensure appropriate and adequate glucose control. Metformin is the first-line therapy for T2DM as it provides effective glycaemic control without causing hypoglycaemia,[4] may assist in weight loss,[5,6] and is suggested to be cardioprotective[7] although stronger evidence for the latter is needed.[8] Additionally, there is a growing body of research focusing upon metformin's potentially positive pleiotropic effects in CLD patients. These include restoring metabolic derangements in patients with non-alcoholic fatty liver disease,[9–11] reducing insulin resistance in hepatitis C,[12] decreasing mortality in cirrhosis patients[13] and decreasing the incidence of disease progression to hepatocellular carcinoma.[14] For these reasons, metformin may be the initial drug treatment of choice for CLD patients with T2DM. Metformin also has a well-characterised safety profile. It is eliminated unchanged by the kidneys.[15] It may, therefore, be reasonable to use metformin in patients with CLD as its disposition should be unaffected by alterations in liver function.

Despite the many potential benefits of metformin, the FDA official 'label', under 'Warnings and Precautions', advises against its use in hepatic impairment because of the risk of inducing lactic acidosis. This warning stems from historical concerns that the reduced hepatic clearance of lactate in CLD may increase the risk of lactic acidosis associated with metformin therapy.[16,17] Metformin-associated lactic acidosis is a life-threatening condition which has been widely reported but is rare and the nature and importance of its association with metformin is unclear.[18] There are very little data on the safety and pharmacokinetics of metformin in patients with CLD. To date, only two observational studies have investigated the effect of metformin on all-cause mortality in CLD patients, both demonstrating increased survival with metformin therapy compared to those who discontinued metformin.[13,14] Neither study measured lactate or metformin concentrations. In the absence of sufficient safety and pharmacokinetic data, and the current proscriptive regulatory labelling, many clinicians remain hesitant to prescribe metformin to patients with CLD[19] although reports from some studies indicate up to 30% of patients with CLD and T2DM are prescribed the drug.[20,21] Additional evidence on the disposition of metformin and lactate concentrations in this patient population is, therefore, required if clinicians are to feel confident in prescribing metformin to patients with CLD. Reassurance is needed not only for prescribers and patients with CLD but also regulators if they are to be motivated to relax the warnings regarding prescribing the drug in CLD.

The aim of the present study was to investigate the safety of metformin in patients with all-cause CLD, with varying degrees of liver fibrosis and cirrhosis, with or without T2DM. A secondary, related aim was to compare the pharmacokinetics of metformin in patients with CLD with healthy subjects, and patients with T2DM and no known history of CLD.