To the Editor:-We appreciate the case report by Mercker et al. (Anesthesiology 1997; 87:1003–5) regarding postoperative lactic acidosis in a patient receiving metformin. Although we praise the authors for wisely stopping the metformin after diagnosing the patient's systemic inflammatory response syndrome, we disagree with their conclusion that metformin contributed to the acidosis and with their decision to routinely stop metformin administration for several days perioperatively.
As the authors noted, the association of biguanides (metformin, phenformin, and buformin) with metabolic acidosis led to their removal from the United States in 1977. An initial review contained 330 cases of lactic acidosis in diabetic patients taking biguanides; however, only 12 cases involved patients taking metformin, and each of the 12 patients had significant renal insufficiency (the lowest creatinine was 3.0 mg/dl). Renal insufficiency is now a contraindication for metformin because it is excreted by the kidneys. Accumulation of metformin, as in a patient with renal insufficiency who continues to receive the drug, can result in elevated lactate levels and death. 
It is unlikely that Mercker et al.'s patient had an elevated metformin level after receiving only 500 mg/day for postoperative days 1–4, with normal renal function on postoperative days 1 and 2. Given that the half-life of metformin in patients with normal renal function is only 1.7–4.5 h and that doses up to 2,550 mg/day result in therapeutic metformin levels, this patient's total dose of 2,000 mg during 4 days would not cause elevated levels. [3,4]The authors considered hemodialysis and used venovenous filtration to eliminate metformin in their patient, but we believe that these treatments were not indicated without elevated metformin levels.
Metformin has never been proven to be detrimental in patients without contraindications. This is not surprising: at therapeutic levels, metformin does not significantly inhibit oxidative phosphorylation, as does phenformin; therefore, metformin causes, at most, a minimal increase in lactate. [5,6]Metformin was used in Canadians without contraindications for 56,000 patient-years without any reported cases of lactic acidosis. Also, in a study of diabetic patients presenting to the emergency room, the incidence of lactic acidosis was not higher in patients taking metformin than those given sulfonylureas or insulin. 
The effects of therapeutic levels of metformin are unknown during type A lactic acidosis (e.g., hypoperfusion from sepsis). The high mortality of patients with type A acidosis probably is caused by the degree of hypoperfusion, as reflected by the lactate ion level, not a result of the hyperlactatemia itself. [9,10]Lalau et al. reported metformin-associated lactic acidosis in 14 patients; the 4 patients without elevated metformin levels had worse prognoses because their lactic acidoses were purely caused by hypoperfusion compared with patients who had lactic acidoses partly because of metformin accumulation. Similarly, a study of 20 patients with metformin- associated lactic acidosis showed that most patients with elevated metformin levels did well, whereas those with therapeutic or low levels had a high mortality. Therefore, even if a therapeutic metformin level contributed to a worsening lactic acidosis, it is doubtful that this would adversely affect morbidity.
In conclusion, we believe that the risks of continuing metformin perioperatively are minimal. There are potential benefits to continuing the drug: patients are more likely to arrive for surgery with better diabetic control, which could result in improved wound healing, decreased infection rates, and decreased need for perioperative insulin. For patients scheduled for major surgery (with the possibility of hypoperfusion), we recommended eliminating only the day of surgery dose because this should result in a minimal metformin level at the time of surgery. Even if a patient were to suffer a serious insult, their metformin level would be low. If concern develops that a perioperative event may have resulted in renal compromise, metformin should not be restarted until renal function is determined to be normal. For ambulatory surgery (with negligible risk of hypoperfusion), we continue administration of metformin on the day of surgery and recommend that patients take it as usual after discharge.
Stewart J. Lustik, M.D.
Alison Vogt, M.D.
Ashwani K. Chhibber, M.D.
Assistant Professors of Anesthesiology; University of Rochester Medical Center; Rochester, New York;SLUSTIK@CCMAIL.ANES.ROCHESTER.EDU