Metformin associated lactic acidosisBMJ 2009; 339 doi: http://dx.doi.org/10.1136/bmj.b3660 (Published 16 September 2009) Cite this as: BMJ 2009;339:b3660
- Emma Fitzgerald, specialist trainee year 2 in anaesthetics 1,
- Stephen Mathieu, specialist registrar in anaesthetics and intensive care medicine1,
- Andrew Ball, consultant in anaesthesia and intensive care medicine1
- Correspondence to: E Fitzgerald
- Accepted 30 March 2009
Metformin, a dimethylbiguanide, is a widely used oral antihyperglycaemic drug used in the long term treatment of type 2 diabetes mellitus. More recently it has also been used to improve fertility and weight reduction in patients with polycystic ovary syndrome.
Many large studies have shown that intensive glucose control with metformin in overweight patients with type 2 diabetes is associated with risk reductions of 32% (P=0.002) for any diabetes related end point, 42% (P=0.017) for diabetes related death, and 36% (P=0.011) for all cause mortality compared with diet alone.1 Furthermore, metformin reduces microvascular end points, and its degree of glycaemic control is similar to that sulphonylureas and insulin. Metformin is considered to be first line treatment in overweight patients with type 2 diabetes whose blood glucose is inadequately controlled by lifestyle interventions alone and should be considered as a first line glucose lowering treatment in non-overweight patients with type 2 diabetes because of its other beneficial effects.2 It may also be useful in overweight patients with type 1 diabetes.
A potential complication of metformin is the development of type B (non-hypoxic) lactic acidosis. Although metformin associated lactic acidosis is a rare condition, with an estimated prevalence of one to five cases per 100 000 population,3 it has a reported mortality of 30-50%.4 Prognosis seems to be unrelated to plasma metformin concentration or lactate level.5
We present a report on a patient with type 2 diabetes who was receiving long term treatment with metformin and developed severe metformin associated lactic acidosis after dehydration, which resulted in renal impairment and consequent accumulation of metformin. This case illustrates the importance of stopping metformin treatment during intercurrent illness, especially dehydration. It also raises the point that continuation of other potentially nephrotoxic drugs should be carefully considered in these circumstances.
A 49 year old woman with type 2 diabetes mellitus presented with a five day history of diarrhoea, vomiting, and lower abdominal pain after a “day case” general anaesthetic for manipulation of a frozen shoulder. She had not stopped taking her regular medications (metformin, losartan, bendroflumethiazide, and atenolol) and had started a postoperative course of non-steroidal anti-inflammatory drugs.
She had presented to her general practitioner complaining of nausea and vomiting and was prescribed an antiemetic (prochlorperazine). That evening she presented to her general practitioner again, this time with acute dyspnoea secondary to pulmonary oedema, and she was referred to hospital.
On admission, she was severely hypoxic and hypotensive but was alert and oriented. Her blood results showed severe lactic acidosis (pH<6.8) and acute renal failure (serum creatinine 769 μmol/l) (table⇓). Chest radiography confirmed pulmonary oedema. Non-invasive ventilatory support and early goal directed treatment was instigated. She remained hypoxic despite continuous positive airway pressure and was intubated that evening. Haemofiltration was started within six hours of arrival. She had a diagnostic laparoscopy to exclude ischaemic bowel in view of her abdominal symptoms and severe lactic acidosis.
She responded to treatment, and within 48 hours of admission she was extubated and inotropic support was discontinued. She needed continuous haemofiltration for four days before being discharged for ongoing intermittent dialysis. One month after presentation to hospital she had returned to her normal function and was back at work.
We know of two other cases of women with type 2 diabetes who presented to us in a similar way with vague abdominal symptoms, vomiting, and dehydration and in whom metformin associated lactic acidosis was also diagnosed.
The patient had symptoms of vomiting and dehydration that were severe enough for her to contact a general practitioner and be prescribed antiemetics. In addition, the she was taking an angiotensin II receptor antagonist and a non-steroidal anti-inflammatory. However, she had not been advised to stop either the metformin or the other nephrotoxic medications while she was dehydrated. Combination treatment with metformin and angiotensin converting enzyme inhibitors or angiotensin II receptor antagonists is common owing to the propensity of hypertension in non-insulin dependent diabetes, and it is an especially dangerous combination in the event of dehydration. Acute renal failure secondary to dehydration and nephrotoxic medications will cause retention of metformin as a result of reduced renal elimination. Although confirmatory laboratory metformin levels were not obtainable, the patient had a raised anion gap (38.5 mEq/l) with no ketonuria or evidence of ingestion of a toxic substance (such as ethylene glycol, methanol). The predominant cause of the hyperlactataemia is therefore the result of acute renal failure and metformin accumulation (type B lactic acidosis). Tissue hypoxia (type A lactic acidosis) as in this case, where pulmonary oedema and resulting hypoxia featured, may have also contributed to lactate production but was not severe enough alone to explain the observed degree of acidosis.
In critical care, severe lactic acidosis usually results from overwhelming sepsis. This patient had little evidence of a septic focus but had vague abdominal symptoms. (Concerns over the presence of ischaemic bowel led one of the two patients not reported here to require diagnostic laparoscopy in the critical care unit as she was too sick to be transferred to the operating theatre or computed tomography scanner. Early recognition of the possibility of metformin associated lactic acidosis may thus prevent unnecessary surgery.) In addition, the measurement of mixed venous oxygen saturations from the tip of the central line (central venous oxygen saturation) may help to exclude sepsis as a cause of a severe lactic acidosis. In sepsis, oxygen consumption is in excess of the oxygen delivery and therefore more oxygen is extracted from the blood as it passes through the capillary beds. Thus in sepsis the central venous oxygen saturation would have an abnormally low value (<75%).
Metformin is thought to act by increasing glucose transport into cells and by decreasing hepatic gluconeogenesis. It decreases the activity of pyruvate dehydrogenase and the transport of mitochondrial reducing agents, and thus enhances anaerobic metabolism even in the presence of adequate oxygen. The presence of reduced insulin or insulin resistance increases the production of precursors for the tricarboxylic acid (Krebs) cycle. Inhibition of pyruvate dehydrogenase then channels the conversion of pyruvate into lactate rather than the aerobic pathway.6 As the figure⇓ shows, coexisting sepsis will accelerate the formation of lactate. Metformin has a short half-life (6.2 hours), and so in normal circumstances it will not accumulate. However, if ingested in toxic doses7 8 or in situations of impaired renal elimination,9 lactic acidosis does occur even in the absence of tissue hypoxia or any other risk factors. Clearly, additional precipitants affecting the glomerular filtration rate,6 10 renal perfusion pressure, or general tissue perfusion will increase the propensity for developing lactic acidosis. This is reflected by the current recommendations on contraindications and guidelines for withdrawing metformin (box). 2 11
Review dose of metformin
If serum creatinine is >130 μmol/l or estimated glomerular filtration rate is <45 ml/min/1.73 m2
If serum creatinine is >150 μmol/l or estimated glomerular filtration rate is <30 ml/min/1.73 m2
During periods of suspected tissue hypoxia (such as myocardial infarction, sepsis)
For three days after use of contrast medium that contains iodine
Two days before general anaesthesia
*Reinstate when renal function stabilises
Although liver failure will reduce the excretion of lactate, the results of liver function tests were normal in our patient (and in the two cases not reported here), and therefore impaired liver function could not be a cause for the accumulation of lactate. Serum metformin levels were not available locally for us to obtain definitive confirmation of the metformin toxicity; however, it may be possible to arrange these with the pharmaceutical manufacturer.
Metformin associated lactic acidosis is a severe and potentially fatal condition that can be easily avoided. Its prevention is imperative, and we emphasise the importance of discontinuing metformin and any potentially renal toxic medications, especially angiotensin converting enzyme inhibitors and angiotensin II receptor antagonists, in any patient who is at risk of dehydration or of developing renal failure.
Once patients become unwell, recognition of metformin associated lactic acidosis requires a high index of suspicion as severe lactic acidosis may suggest that sepsis is overwhelming and likely to be untreatable. Rapid supportive treatment and removal of toxic levels of metformin by haemofiltration is paramount. Measurement of mixed central venous oxygen concentration may be helpful in excluding sepsis as a cause of lactic acidosis and may prevent unnecessary investigational surgery and confirm that continued aggressive supportive treatment is appropriate. Patients can survive prolonged severe acidosis for more than 12 hours without any long term consequences.
The guidelines for reducing or stopping the dose of metformin are well documented in patients with laboratory evidence of worsening renal function, in the presence of intercurrent illnesses causing tissue hypoxia, and when iodinated contrast agents are used (box). However, when a previously well patient presents with an apparently less important insult such as dehydration, the guidance is less clear. Although we do not advocate the withdrawal of metformin in all patients in this situation, a careful clinical evaluation of renal function (particularly with coexisting use of nephrotoxic agents) and early consideration for treatment adjustment may help to avoid the life threatening adverse events of metformin.
We suggest that all patients taking metformin (or indeed any nephrotoxic agent) should seek medical advice if they develop dehydration as a result of either an inadequate intake or prolonged (>24 hours) gastrointestinal disturbance. They should be assessed clinically for evidence of intravascular volume depletion together with urgent blood biochemistry tests. Although our patient (and the two cases not reported here) had a serum creatinine concentration <130 μmol/l and an estimated glomerular filtration rate of >45 ml/min/1.73 m2 within the preceding year, these values had not been checked in the period after the acute illness and before admission to hospital.
A thorough review of all medications is essential (especially non-steroidal anti-inflammatory drugs, angiotensin converting enzyme inhibitors, and angiotensin II receptor antagonists). Careful consideration of temporarily omitting metformin and any nephrotoxic agents may significantly decrease the risk of worsening renal function and consequently decrease the risk of metformin associated lactic acidosis. Follow-up, perhaps by a community district nurse, to ensure that glycaemic control remains stable may be necessary. Although the absolute risk of continuing the metformin in such a situation is not clear, and the risk of developing metformin associated lactic acidosis is reportedly rare, it is associated with very high mortality.
We reported our patient’s case (and the other two cases not reported here) to the Committee on Safety of Medicine, and we suggest that dehydration be added to the list of relative contraindications to metformin therapy.
Cite this as: BMJ 2009;339:b3660
Contributors: All authors analysed and interpreted the data and drafted and revised the article.
Competing interests: None declared.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent obtained.