Review: Metformin: Potential benefits and use in chronic kidney disease


Assoc Professor HL Pilmore, Department of Renal Medicine, Auckland City Hospital, Park Road, Grafton, Auckland ••, New Zealand. Email:


Diabetes mellitus is the commonest cause of end-stage renal failure in both Australia and New Zealand. In addition, the burden of diabetes is prominent in those with chronic kidney disease who have not yet reached the requirement for renal replacement therapy. While diabetes is associated with a higher incidence of mortality and morbidity in all populations studied with kidney disease, little is known about optimal treatment strategies for hyperglycaemia and the effects of glycaemic treatment in this large group of patients. Metformin is recommended as the drug of first choice in patients diagnosed with type 2 diabetes in the USA, Europe and Australia. There are potential survival benefits associated with the use of metformin in additional to recent studies suggesting benefits in respect to cardiovascular outcomes and metabolic parameters. The use of metformin has been limited in patients with renal disease because of the perceived risk of lactic acidosis; however, it is likely that use of this drug would be beneficial in many with chronic kidney disease. Thus the potential benefits and harms of metformin are outlined in this review with suggestions for its clinical use in those with kidney disease.


Diabetes mellitus is the commonest cause of end-stage renal failure in both Australia and New Zealand accounting for 31% and 41%, respectively, of patients starting dialysis in 2008.1 While it is clear that diabetes is associated with a higher incidence of mortality and morbidity in those with chronic kidney disease, the renal transplant population2 and those on dialysis, there are few trials examining treatment options for hyperglycaemia in these patient groups.

Metformin is recommended as the drug of first choice in patients diagnosed with type 2 diabetes in a consensus document issued by the American Diabetic Association and the European Association for the Study of Diabetes.3,4 The Diabetes Australia Guideline Consortium also recommended metformin as first-line treatment in type 2 diabetes.5 As a result of the potential risk of lactic acidosis with metformin in those with renal impairment however, it's use in patients with chronic kidney disease and after renal transplantation is limited.


The major effect of metformin is to reduce hepatic glucose production.6 Until recently, its major mechanism of action has been unclear; however, recent data have shown that phosphorylation of the transcriptional coactivator cAMP response element-binding (CREB) protein occurs with metformin, thus reducing the expression of genes inducing gluconeogenesis.7 In addition, metformin increases the insulin-mediated utilization of glucose in peripheral tissue thereby improving glycaemic control8 while also reducing free fatty acid concentrations resulting in less substrate available for gluconeogenesis.

In comparison to other hypoglycaemic agents, metformin is much less likely to result in hypoglycaemic episodes, rendering this agent safer from this perspective.9 Elimination is reduced in those with renal impairment thereby lengthening the plasma half life of the drug, which is increased in proportion to the degree of impairment in creatinine clearance.10


Metformin is generally well tolerated but gastroenterological side-effects are common, occurring in at least 10% of patients. These include anorexia, nausea, abdominal pain and diarrhoea. These symptoms can be mild and transient but are severe in some necessitating discontinuation of the drug in only 5%. A reduction in Vitamin B12 absorption can also occur after a long period of metformin use11 and although this is uncommon, some have recommended vitamin B12 screening.12


The greatest perceived risk associated with metformin is that of lactic acidosis. A number of reports in the literature link biguanides with the development of lactic acidosis. Initial reports with phenformin showed a high incidence of lactic acidosis with an event rate of 40–64 per 100 000 patient years.13 Phenformin was removed from the US market because of the risk of lactic acidosis in 1977. The incidence of lactic acidosis with metformin is markedly lower than with phenformin, with two recent meta-analyses showing no evidence of an increased risk of lactic acidosis associated with the use of metformin compared with non-metformin therapies.14,15 Similarly, a study from Western Australia found only five cases of lactic acidosis in over 12 000 patient years.16 Of these, only three patients were taking metformin. All patients had evidence of significant systemic disease associated with the development of lactic acidosis and there was no increased risk for the condition demonstrated with metformin.

The risk of lactic acidosis has been reported to be increased in patients with renal impairment, heart failure, liver disease, high alcohol intake or a previous history of lactic acidosis.17 Renal dysfunction appears to be the most common risk factor implicated with lactic acidosis and many current guidelines suggest discontinuation of metformin at a glomerular filtration rate (GFR) of <60 mL/min. Despite this, there are a large number of patients with renal impairment using metformin with no reported increase in the incidence of lactic acidosis.18 For these reasons, the recently published National Evidence Based Guidelines for Blood Glucose Control in type 2 diabetes5 have stated that lactic acidosis is rare and have suggested that an estimated glomerular filtration rate (eGFR) cut-off of <60 mL/min/1.73 m2 is overly conservative, recommending that although metformin is contraindicated in those with an eGFR of less than 30 mL/min per 1.73 m2, it can be used with caution in those with a GFR of 30–45 mL/min per 1.73 m2. While there is no clear data to define specifically at which level of renal impairment metformin should be contraindicated, the risk of lactic acidosis in those with mild to moderate renal impairment is believed to be less than in those with more severe renal impairment.


Glycaemic control

The primary indication for metformin use is treatment of hyperglycaemia although it is also potentially useful for promotion of ovulation in polycystic ovary syndrome19 and is used for the treatment of obesity.20 The effects of metformin have been compared with those of other diabetes treatment in a recent Cochrane review examining 29 trials with 37 treatment arms.21 This systematic review demonstrated that metformin is highly efficacious at improving glycaemic control with a significant improvement in HbA1c compared with placebo or diet. Comparisons with sulphonylureas are varied, with the Cochrane review demonstrating a benefit in HbA1c and fasting plasma glucose in patients treated with metformin compared with sulphonylureas.21 A summary of metformin's effects on glycaemia is appended in Table 1.

Table 1.  Effects of metformin on glycaemic control assessed by HbA1c and fasting plasma glucose (summarized from Cochrane Review21)
 Number of studiesNumber of patientsStudy mean difference95% confidence interval
 Studies comparing metformin and placebo121587−0.97[1.25, −0.69]
 Studies comparing metformin and sulphonlyureas122376−0.14[−0.28, −0.01]
 Studies comparing metformin and diet3914−1.06[−1.89, −0.22]
 Studies comparing metformin and thiazolidinediones3260−0.28[−0.52, −0.03]
 Studies comparing metformin and insulin28110.26[−0.22, 0.74]
Fasting plasma glucose    
 Studies comparing metformin and placebo121587−0.87[−1.13, −0.61]
 Studies comparing metformin and sulphonlyureas132409−0.16[−0.27, −0.05]
 Studies comparing metformin and diet2814−2.08[−4.87, 0.71]
 Studies comparing metformin and thiazolidinediones32600.15[−0.09, 0.39]
 Studies comparing metformin and insulin28110.46[−0.25, 1.17]

The risks and benefits of intensive glycaemic control have been extensively studied in both type 1 and type 2 diabetes. Intensive glycaemic control has been shown to reduce both microvascular and macrovascular disease in those with type 1 diabetes.22,23

In type 2 diabetes, however, the benefits of tight glycaemic control are less clear. While good glycaemic control has been shown to reduce the development and progression of microvascular disease, in particular retinopathy and nephropathy;24,25 recent studies have failed to show a reduction in macrovascular events with intensive glucose lowering.25,26 Indeed, the ACCORD trial,26 which was specifically designed to examine the effect of targeting normoglycaemia in patients with type 2 diabetes and previous cardiovascular disease, was terminated prematurely because of the finding of an increase in mortality in the intensive arm although there was no difference in the primary end point of cardiovascular events at the time of study termination. Similar results were found in the ADVANCE study.26 This issue, however, remains somewhat unclear however, with a recent meta-analysis27 demonstrating a significant reduction in coronary events with intensive glucose monitoring although there was no reduction in all-cause mortality or stroke.

Although it is clear that metformin has excellent hypoglycaemic efficacy, its durability of effect, while greater than that of sulphanylureas, may not be as sustained as that of thiazolidinediones.28

Metformin and survival

Demonstration of a survival benefit with different hypoglycaemic medications is difficult because of the ability to adequately power studies and is confounded by factors such as glycaemic control. Nevertheless, there are suggestions of a survival benefit associated with metformin. In the UKPDS study,24 newly diagnosed patients with type 2 diabetes and obesity were randomized to intensive treatment with a sulphonylurea or insulin, or metformin compared with conventional treatment with diet. Patients allocated to intensive glycaemic control with metformin showed a greater benefit than intensive treatment with sulphonylureas or insulin for any diabetic-related outcome and for all-cause mortality (RR 0.73; 95% CI 0.55–0.97) with a number needed to treat of 19 to prevent one case of all-cause mortality.

In comparison to the placebo arm in this trial, the use of metformin was associated with a significant reduction in diabetes-related death and all-cause mortality although this was somewhat confounded by differences in glycaemic control.

Effect of metformin on cardiovascular (CVS) disease

Macrovascular disease is prevalent in patients with diabetes mellitus and the commonest cause of mortality.29 There is increasing evidence that metformin use results in a reduction in cardiovascular events although this effect may not be clinically apparent for many years. A recently published follow-up study of UKPDS30 studied patients for a further 5 years with no attempt made to maintain their previously assigned therapy. While the differences in glycaemic control between the two groups were lost in the follow-up phase, as more events emerged over time, there was a significant reduction in the risk of myocardial infarction with metformin of 33%, and a 30% reduction in diabetes-related death compared with those in the original conventionally treated arm.

In a smaller study, patients with type 2 diabetes on insulin randomized to the addition of either metformin or placebo31 had a 39% reduction in macrovascular events with a number needed to treat of 16 (CI 9.2–66.6). A retrospective analysis of a randomized controlled trial in diabetic patients undergoing coronary interventions found a 79% reduction in myocardial infarction with metformin use compared with a sulphonylurea or insulin.32

There is also evidence of beneficial effects of metformin on vascular function, with improvements in endothelium-dependent vasodilatation of the brachial artery in patients with the metabolic syndrome on metformin compared with placebo.33 In addition, there are improvements in markers of endothelial activation and coagulation in patients with impaired glucose tolerance treated with metformin compared with placebo.34

While the literature suggests a macrovascular benefit from metformin, some controversy remains. In patients in the UKPDS sub-study,29 the early addition of metformin in patients already on a sulphonylurea was associated with a significant increase in diabetes-related death suggesting the necessity for further investigation into the optimal glycaemic treatment in type 2 diabetes.

Metabolic effects of metformin

The improvement in cardiovascular outcomes potentially associated with metformin, may, at least in part, be due to improvements in metabolic factors implicated in the development of cardiovascular disease. A number of metabolic benefits have been demonstrated with metformin (Table 2). In particular, there are benefits over sulphonylureas, in terms of weight and BMI, while more modest benefits are shown for lipid levels and measures of coagulation.

Table 2.  Metabolic effects of metformin
ParameterMetabolic effect of metforminReference
  1. BMI, body mass index; HDL, high density lipoprotein; LDL, low density lipoprotein.

Weight gaininline image21
BMIinline image21
Waist circumferenceinline image35
Leptininline image36
Plasminogen activator inhibitor-1inline image36,37
LDL cholesterolinline image21,36
Triglyceridesinline image21,36
Lipoprotein lipaseinline image37
HDL cholesterolinline image38
Fibrinogeninline image37
Platelet aggregationinline image37
Metabolic syndromeinline image39

Recent data have shown a reduction in the development of the metabolic syndrome with metformin in patients at high risk.39 In the Diabetes Prevention Program, the use of metformin was associated with a 17% reduction in the incidence of the metabolic syndrome in comparison to placebo although this was superseded by the benefits of lifestyle modification that resulted in a 41% reduction. While lifestyle modification resulted in benefits in all parameters of the metabolic syndrome, metformin use was associated with benefits in waist circumference, and High Density Lipoprotein (HDL) cholesterol levels in addition to glucose levels.

Additionally, there have been recent reports of a reduction in the incidence of cancer in diabetics on metformin compared with those who have never used this class of medication. In a matched cohort study, there was a 37% reduction in the likelihood of diagnosis of cancer in patients treated with metformin40 in addition to a reduction in the incidence of cancer-related deaths. Certainly, there is a plausible tumour suppressor mechanism associated with metformin, with its activation of AMP-activated protein kinase (AMPK) resulting in cell growth suppression.41

Metformin in heart failure

Heart failure is seen as a relative contraindication to the use of metformin. This is largely due to the perceived increased risk of lactic acidosis in this patient group. Nevertheless, there have been a number of trials examining the use of metformin in patients with heart failure. A recent systematic review reported eight studies examining the effects of glycaemic treatment in patients with diabetes and heart failure.42 In this review, three studies examined the use of metformin in 3327 patients and while none of these studies were randomized controlled trials, metformin was associated with a 14% reduction in mortality compared with other anti-diabetic drugs and insulin. In addition, there was no increase in hospital admissions for any cause in patients treated with metformin suggesting that this agent appears safe in patients with heart failure.

Metformin and prevention of diabetes

The Diabetes Prevention Program43 is the largest randomized controlled trial aiming to prevent the development of diabetes in high-risk patients. Patients with impaired glucose tolerance were randomized to placebo, metformin or a lifestyle modification programme and followed for a mean of 2.8 years. Lifestyle modification resulted in a 58% reduction in the development of diabetes and was significantly superior to both metformin and placebo. The use of metformin, however, did result in a significant reduction in diabetes compared with placebo (31%) with a number needed to treat with metformin of 13.9 to prevent one case of diabetes in this high-risk group.

In a recent comparison of women in this study who had a history of gestational diabetes, the effects of metformin were the same as lifestyle modification,44 suggesting that some groups may benefit more from the use of metformin than others.

Metformin in renal disease

There have been no randomized controlled trials examining hypoglycaemic agents or insulin in patients with chronic kidney disease. Kidney Disease Outcomes Quality Initiative (K/DOQI), which has developed guidelines for the management of hyperglycaemia in patients with chronic kidney disease,45 is explicit in stating that the guidelines are extrapolated from trials of patients with normal renal function or Chronic Kidney Disease (CKD) 1 and 2 because of the paucity of trials in patients with advanced CKD. Treatment options often need to be altered in patients with worsening kidney disease for a number of reasons. Patients with renal impairment have an increased risk of hypoglycaemia as a result of reduced renal clearance of insulin and impaired gluconeogenesis in the kidney. Additionally, a number of agents are not recommended or are contraindicated in renal impairment. Metformin has been included in this group because of the perceived risk of lactic acidosis although hypoglycaemia is not a significant issue with this drug. In dialysis patients, K/DOQI recommends that patients follow the ADA guidelines, however, make the caveat that dialysis patients are not targeted in the trials and further research is required in this group.

Effect of diabetes after renal transplantation

Development of new onset diabetes after transplantation (NODAT) is common in patients after renal transplantation. Early studies had varying definitions of diabetes and many reported the development of diabetes only when the use of insulin was required with a recent systematic review reporting an incidence from 2% to 50%. A large observational study, however, suggests a rate of NODAT of 14–16% in first post-transplantation year, declining to an annual incidence of 4–6% thereafter.46

There are a large number of risk factors for the development of NODAT. These include standard risk factors such as increasing age, male gender, non-white ethnicity and BMI. Substantial weight gain occurs in the first 1–2 years post transplantation47 and this has been shown to be associated with an increased risk of NODAT. There are, however, a number of additional risk factors more specific to transplantation. These include Hepatitis C with a recent48 meta-analysis showing OR 3.97 for the development of NODAT with Hepatitis C infection and the use of a number of immunosuppressive agents.

The use of corticosteroids49 and calcineurin inhibitors, in particular tacrolimus,50 has been shown to increase the incidence of NODAT. The DIRECT trial51 randomized patients to tacrolimus or cyclosporine after renal transplantation and found a significantly higher incidence of NODAT and a nearly twofold risk of insulin requirement with tacrolimus compared with cyclosporine. Additionally, the use of sirolimus appears to be implicated in the development of NODAT resulting in reductions in insulin sensitivity, beta cell function and overall glucose tolerance.52

The development of diabetes after renal transplantation has a significant impact on outcomes after transplantation. There is a marked increased risk of cardiovascular events in patients both with impaired glucose tolerance and with NODAT53 while both pre-existing diabetes and NODAT are associated with reductions in long-term patient survival.2 There has also been an increased risk of acute rejection reported in those with poor glycaemic control after transplantation.54

Despite this, there are very few trials examining prevention and treatment of patients with diabetes after kidney transplantation. One study55 examined the effects of lifestyle modification (dietician referral, exercise, weight loss advice) in patients with impaired glucose tolerance (IGT) or NODAT demonstrating a 15% improvement in 2 h postprandial glucose in this group. Thiazolidinediones have been used after transplantation but not in clinical trials. While they appear to be safe in case reports, troglitazone induces P450 and lowers cyclosporine levels.56 After renal transplantation, there has been one retrospective review of patients with either NODAT or pre-existing diabetes being treated with metformin.57 A total of 32 patients had been treated with metformin with a mean GFR of 74 mL/min at the start of treatment. In those patients with pre-existing diabetes, there was a reduction in the GFR at a mean of 16 months follow up; however, the mean GFR remained relatively high at 60 mL/min. Five patients, however, discontinued metformin because of an increase in the serum creatinine with a cut-off of 1.6 mg/dL (142 µmol/L). While the cause of the elevation in creatinine was unclear, there were no cases of lactic acidosis and the HBA1c was well controlled throughout the study.


Despite metformin being internationally recommended as the first-line drug in patients with newly diagnosed diabetes, its use in those with kidney disease is limited by the perceived risk of lactic acidosis. This risk appears to be largely due to other co-morbid events resulting in tissue hypoxia, and is extremely rare. Metformin is, however, extremely efficacious in the management of hyperglycaemia and has metabolic effects that are likely to be beneficial in those with kidney disease. Similarly, metformin appears to have beneficial effects on survival and potentially on macrovascular events, especially in overweight and obese patients.

While the use of metformin should remain contraindicated in dialysis patients, it is possible that its use in patients with CKD and after renal transplantation would result in cardiovascular and survival benefits. Thus the recommendations of the Australian Diabetes Guidelines to liberalize the GFR guidelines for the use of metformin appear sensible. A clear GFR cut-off has not been established in the literature; however, the risk of lactic acidosis is extremely low while the potential benefits are substantial.

Finally, the institution of clinical trials examining treatment options for hyperglycaemia in patients with renal disease will increase our understanding of management of this important patient group and should be encouraged and facilitated.