Intended for healthcare professionals


Contraindications to the use of metformin

BMJ 2003; 326 doi: (Published 04 January 2003) Cite this as: BMJ 2003;326:4

Evidence suggests that it is time to amend the list

  1. G C Jones, consultant physician (greg.jones{at},
  2. J P Macklin, senior house officer,
  3. W D Alexander, consultant physician
  1. Department of Diabetes, Gartnavel General Hospital, Glasgow G12 0YN
  2. Diabetes Centre, Western General Hospital, Edinburgh EH4 2XU

    According to the United Kingdom prospective diabetes study, patients with type 2 diabetes randomised to intensive treatment with metformin, sulphonylurea, or insulin had similar degrees of glycaemic control and significantly reduced microvascular end points.1 The study showed that the use of metformin in obese patients reduced cardiovascular events. The group treated with metformin had no hypoglycaemia and less weight gain. Treatment with metformin rather than diet alone produced a significant reduction in relative risk in all cause mortality (36%, P=0.011), diabetes related deaths (42%, P=0.017), any diabetes related end point (32%, P=0.0023), and myocardial infarction (39%, P=0.01). Metformin is the only oral hypoglycaemic agent proved to reduce cardiovascular risk and is now recognised as the treatment of choice in overweight patients with type 2 diabetes.

    Lactic acidosis associated with metformin is a rare condition with an estimated prevalence of one to five cases per 100 000.2 Although classically lactic acidosis associated with metformin has been thought of as lactic acidosis secondary to accumulation of metformin, the evidence for this is poor.

    Metformin does not affect lactate concentrations in patients with type 2 diabetes,3 is excreted solely through the kidney, and has a short half life—accumulation of metformin therefore rarely occurs in the absence of advanced renal failure.4 Accumulation of metformin alone is rarely reported as a cause of lactic acidosis, and tissue hypoxia acting as a “trigger” is found in most instances. Accumulation of metformin does not correlate with lactate concentrations or mortality. Mortality is predicted by the severity of underlying hypoxia.5 Metformin should therefore be discontinued when tissue hypoxia is suspected.

    A recent review of cases of lactic acidosis associated with metformin, which was published between May 1995 and January 2000, concluded that no mortality was associated with metformin alone.6 Another study noted that the rates of lactic acidosis in the United States before the approval of metformin were no different from the rates observed in users of metformin.7 A Cochrane systematic review concluded that treatment with metformin was not associated with an increased risk of lactic acidosis.8

    Suggested revised contraindications and guidelines for withdrawing metformin

    • Stop if serum concentration of creatinine is higher than 150 micromols/l.*

    • Withdraw during periods of suspected tissue hypoxia (for example, due to myocardial infarction, sepsis).

    • Withdraw for three days after contrast medium containing iodine has been given, and start treatment with metformin only after renal function has been checked.

    • Withdraw two days before general anaesthesia and reinstate when renal function is stable.

    If adherence to the published contraindications, all of which relate to the feared risk of lactic acidosis, were to be strict, metformin would, or rather should, be seldom prescribed at all. The British National Formulary says that any predisposition to lactic acidosis is a contraindication ( As diabetes itself is a predisposition to accumulation of lactate,3 should we stop using the drug altogether in the treatment of diabetes? The BNF and other publications also use the terms “renal or hepatic impairment.” These terms are vague and therefore unhelpful.

    The DIGAMI (diabetes mellitus, insulin glucose infusion in acute myocardial infarction) study suggests that treatment with insulin would be the treatment of choice immediately after acute myocardial infarction, but after this no apparent reason exists why metformin should not be reinstated.9 The withdrawal of metformin in stable chronic heart failure has been questioned as it may have an adverse effect on glycaemic control.10

    In the United Kingdom it has been shown that doctors tend not to comply with these contraindications. In Southampton 54% of 89 patients treated with metformin had a published contraindication.11 In Dundee recent analysis of 1847 patients treated with metformin showed that 24.5% (452) had a contraindication to metformin.12

    Although circumstantial evidence shows that treatment with metformin may be linked with lactic acidosis, no causal relation has been proved. Metformin is proved to reduce plasma glucose and complications of diabetes. Uniquely, it is the only hypoglycaemic agent to date that has been shown to reduce the macrovascular complications of diabetes.1 Current published guidelines vary and may limit the use of metformin and cause confusion among doctors. It is essential that the benefits of treatment with metformin be made available to as wide a group of appropriate patients as possible without laying prescribers open to criticism or litigation in the event of concomitant lactic acidosis. A simplified and pragmatic set of guidelines should be adopted, stressing the importance of renal clearance of metformin and withdrawal of metformin in patients with tissue hypoxia.

    As metformin is the only oral hypoglycaemic agent proved to reduce cardiovascular mortality, its use should be as widespread as possible in type 2 diabetes. We hope that these suggested guidelines are less ambiguous than current ones and prevent the current situation of many clinicians, who are having to ignore written contraindications in order to maximise the use of metformin appropriately.


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    View Abstract