European Journal of Pediatrics

, Volume 164, Issue 6, pp 362–365

Severe metformin intoxication with lactic acidosis in an adolescent


  • Martin Lacher
    • Department of PaediatricsUniversity Hospital Freiburg
  • Maren Hermanns-Clausen
    • Poisons Information CentreUniversity Hospital Freiburg
  • Karsten Haeffner
    • Department of PaediatricsUniversity Hospital Freiburg
  • Matthias Brandis
    • Department of PaediatricsUniversity Hospital Freiburg
    • Department of PaediatricsUniversity Hospital Freiburg
Original Paper

DOI: 10.1007/s00431-005-1634-y

Cite this article as:
Lacher, M., Hermanns-Clausen, M., Haeffner, K. et al. Eur J Pediatr (2005) 164: 362. doi:10.1007/s00431-005-1634-y


A 15-year-old healthy girl ingested 38.25 g (0.55 g/kg body weight) of metformin in a suicide attempt. Subsequently she developed lactic acidosis and moderate renal failure. An initial session of haemodialysis was able to treat the acidosis and reduce the toxic level of metformin. Nevertheless, a further increase in serum lactate was observed during and after the first dialysis treatment. A second session of haemodialysis was started 5 h after the end of the first session and resulted in a lowering of the lactate level and an almost total elimination of metformin. During the further clinical course, reversible acute renal failure with a maximum creatinine of 2.4 mg/dl was observed. Conclusion: Despite sufficient haemodialysis, the production of lactate can be greater than the elimination in the case of severe metformin intoxication. Therefore haemodialysis should be continued even in the situation of rising lactate levels during the treatment.


HaemodialysisIntoxicationLactic acidosisMetformin


Biguanides have been used for many years as oral anti-hyperglycaemic agents in the treatment of type 2 diabetes mellitus. Lactic acidosis is a rare but serious adverse effect of metformin, especially in patients with renal failure [4]. The exact mechanism whereby metformin causes acidosis is not entirely known. Metformin is absorbed relatively quickly by the intestine and is not metabolised. About 90% of the drug is eliminated by glomerular filtration and tubular secretion with a serum half-life between 1.5 and 5 h [17]. The mechanism thought to be responsible for the development of lactic acidosis is suppression of gluconeogenesis from several substrates, including lactate, pyruvate, glycerol and amino acids leading to lactate accumulation [9]. The risk of metformin associated lactic acidosis is increased by either chronic or acute renal insufficiency [16]. There are several well documented descriptions of metformin overdose in the literature [5, 7, 13, 15,20]. Nearly all of them occurred in adult diabetic patients. In three cases the outcome was fatal: a 25-year-old man after ingestion of an unknown number of metformin tablets, a 33-year-old woman after ingestion of 35 g metformin and a 42-year-old man after ingestion of 63 g metformin [15,20]. Intoxications with metformin in childhood are rare. In one multicentre case series ( n =55) of paediatric metformin ingestion, none of the patients showed imbalances of blood sugar levels or evidence of lactic acidosis [18]. In this report we describe a 15-year-old healthy girl, who developed lactic acidosis after ingestion of a metformin overdose in an attempt to commit suicide.

Case report

The 15-year-old girl took 38.25 g (0.55 g/kg body weight) of metformin in a suicide attempt. Between 1–2 h after ingestion she was found somnolent (9/15 on the Glasgow Coma Scale). After oral adminstration of glucose and 60 g activated charcoal by the rescue team, the girl vomited several times and was brought to an emergency care facility.

Initial laboratory investigations showed a metabolic acidosis with a pH of 7.29, serum bicarbonate of 17 mmol/l and a base excess of −10 mmol/l. The blood glucose was 166 mg/dl and serum lactate 6.4 mmol/l. Three hours later the girl was transferred to our centre with a Glasgow Coma Scale level of 7/15. The laboratory investigations at this time showed a worsening metabolic acidosis with a pH of 7.2, serum bicarbonate of 15 mmol/l and a base excess of −12 mmol/l. The blood glucose was 49 mg/dl and serum lactate 8.7 mmol/l. Her renal function was impaired (serum creatinine 1.48 mg/dl) and the blood metformin level was 165 mg/l (therapeutic levels <2 mg/l) (Fig. 1a). Electrolytes, inflammatory markers, liver and pancreas enzymes were normal. Ethanol was not detectable in the blood.
Fig. 1

a Serum metformin level and b serum pH and lactate levels during hospitalisation

Approximately 5–6 h after ingestion and 4 h after she was found, the girl underwent haemodialysis for a duration of 4 h. In addition, 300 ml of 8.4% sodium bicarbonate was given intravenously. Despite efficient dialysis, the serum lactate level rose from 8.7 to 16.3 mmol/l (Fig 1b). Eight hours after admission and 5 h after the end of the first session of haemodialysis, the metformin level was lower (37 mg/l) but the serum lactate level had increased to 20.6 mmol/l with a pH of 7.33, serum bicarbonate of 20 mmol/l and a base excess of −5 mmol/l (Fig. 1). Therefore a second haemodialysis was performed over 5 h which succeeded in lowering the serum lactate to 4.4 mmol/l and in further reducing the metformin level to 14 mg/l (Fig. 1).

During the further clinical course, a reversible acute renal failure with a maximum creatinine of 2.4 mg/dl was observed. Blood glucose levels were maintained within the reference range. The girl was discharged from the intensive care unit 19 h after admission and transferred to the psychiatric ward 2 days later for further psychological work-up.


We present the first report of lactic acidosis associated with metformin exposure in childhood. There are only a few descriptions of metformin intoxications and their effects in paediatric patients, probably due to the fact that this drug is only rarely used in children with type 2 diabetes mellitus and not licensed for use in children. One multicentre case series of paediatric metformin ingestion [18] is based on case reports to American Poison Control Centers. In this study, 55 cases were collected from which 37 children were evaluated in a healthcare facility. The absolute doses ingested in this study ranged from 250 mg to 16.5 g with a mean of 1.71 g. None of these children experienced hypoglycaemia or acidosis. Therefore Spiller et al. [18] concluded that an ingestion of ≤1700 mg of metformin appears not to pose a significant health risk in healthy paediatric patients. There are several limitations to this study: in 10 of the 11 older children there was no laboratory confirmation of metformin exposure. Of all 55 children, 41 ingested a maximum of two tablets of metformin (<1700 mg) and only 37 children were evaluated in a healthcare facility. The metformin dose was unknown in five cases. In only 37 cases was glucose measured and in only 21 cases were pH measurements done.

Compared with the doses in this study, the ingested dose of 38.25 g metformin (0.55 g/kg body weight) in our patient was high and correlated with the massively elevated initial serum metformin level of 165 mg/l. Measurements of serum metformin concentrations were neither done in the multicentre study of Spiller et al. [18] nor in any of the other reports on metformin ingestions in childhood. In adult patients, a metformin level of 0.5–2 mg/l is regarded as therapeutic [17]. Metformin concentrations between 4.1 to 84.9 mg/l were measured in 10 of 14 adult patients with lactic acidosis [12]. To our knowledge, the initial metformin level of 165 mg/l is the highest reported metformin level so far in a survivor. There are reports on non-survivors with similar high metformin levels: A 42-year-old diabetic male died due to metformin intoxication with a metformin level of 188 mg/l despite veno-venous haemodialysis and a 50-year-old diabetic male died due to severe lactic acidosis with a metformin level of 166 mg/l without haemodialysis treatment [1,15]. Considering these reports, our patient had a high risk for a fatal outcome. Possibly, the lack of additional risk factors and the rapid initiation of haemodialysis were important for her survival. However, in a recently published review of case reports by Stades et al. [19], no relationship could be established between lactate concentration, metformin levels and mortality during metformin therapy. In an adult case series, Lalau and Race [11] found that a high metformin concentration had a prognostically favourable effect on the survival of patients with lactic acidosis. In both of the above case series, all but one patient had one or more risk factors that could have induced lactic acidosis even without metformin therapy. Our patient did not have any risk factors and the acute renal failure clearly developed subsequently to the ingestion of metformin. Therefore, we feel confident to attribute the lactic acidosis to the metformin overdose.

As sodium bicarbonate infusions alone are not able to correct the acid-base metabolism sufficiently, haemodialysis is recommended for clearance of metformin and treatment of acidosis [3]. The use of sodium bicarbonate is well known but controversial [8]. There are theoretical disadvantages of using intravenous bicarbonate including a left shift of the oxyhaemoglobin dissociation curve, excess sodium load, rebound metabolic alkalosis, disturbances in serum potassium and calcium levels and reflex vasodialation after bolus injection [8]. The advantage of haemodialysis with bicarbonate dialysate is the possibility of metformin clearance without the associated risk of intravenous administration. We have found no clinical trials comparing therapeutic modalities in the situation of metformin induced lactic acidosis.

Remarkably, the serum lactate level rose constantly during and after the first session of haemodialysis. It increased from 8.7 to 16.3 mmol/l during dialysis and continued to rise after the end of dialysis up to 20.6 mmol/l. This increase in serum lactate during haemodialysis was not observed before in case reports of metformin intoxications in adult patients [12, 14,20]. At the same time a substantial clearance of metformin was observed. A second session of dialysis was necessary to decrease the lactate level to 4.4 mmol/l. This observation can possibly be explained by the pharmacokinetics of metformin. Gastrointestinal absorption of metformin is incomplete. The bioavailability of a therapeutic dose is 50%–60%, and about 20%–30% of an oral dose can be recovered from the faeces [2,6]. The discrepancy between the amount of drug absorbed and the amount present in the serum may result from presystemic clearance or binding to the intestinal wall [6]. Metformin accumulates in the oesophagus, stomach and small intestine and significantly increases lactate production from the gut [21]. This mechanism could account for the prolonged production and continuing increase of serum lactate during haemodialysis in our case. An additional component of the delayed clearance of serum lactate could be the association of the drug with red blood cells. The concentration of metformin in erythrocytes is known to be more persistent than in serum, in which the concentration of the drug was measured in our patient [8, 10,12]. Despite sufficient clearance from the serum, the metformin concentration could have remained higher in the erythrocytes and could therefore have contributed to the prolonged production of lactate.

As a consequence of one or both of these mechanisms, lactate production might have initially exceeded its elimination by haemodialysis. With a decreasing lactate production rate, haemodialysis would have been able to remove lactate more efficiently.


The authors whish to thank all involved medical personnel, expecially Dr. J.U. Folkens from the referring hospital in Offenburg, Germany, and Dr. H. Neurath from the University of Göttingen, Germany, for measuring metformin levels.

Copyright information

© Springer-Verlag 2005