A Study of Associations Between Plasma Metformin Concentration, Lactic Acidosis, and Mortality in an Emergency Hospitalization Context.

OBJECTIVES: To determine the plasma metformin concentration threshold associated with lactic acidosis and analyze the outcome in metformin-treated patients with lactic acidosis hospitalized in an emergency context. DESIGN: A retrospective, observational, single-center study. SETTING: Emergency department and ICUs at Amiens University Hospital (Amiens, France). PATIENTS: All consecutive patients with data on arterial lactate and pH up to 12 hours before or after a plasma metformin assay within 24 hours of admission, over a 9.7-year period. INTERVENTION: None. MEASUREMENTS AND MAIN RESULTS: The study population consisted of 194 metformin-treated diabetic patients (median age: 68.6; males: 113 [58.2%]); 163 (84%) had acute kidney injury, which was associated variously with dehydration (45.4%), sepsis (41.1%), cardiogenic shock (20.9%), and diabetic ketoacidosis (16%). Eighty-seven patients (44.8%) had lactic acidosis defined as an arterial blood pH less than 7.35 and a lactate concentration greater than or equal to 4 mM, and 38 of them (43.7%) died in the ICU. A receiver operating characteristic curve analysis showed that a metformin concentration threshold of 9.9 mg/L was significantly associated with the occurrence of lactic acidosis (specificity: 92.9%; sensitivity: 67.1%; area under the receiver operating characteristic curve: 0.83; p < 0.0001). Among lactic acidosis-positive patients, however, in-ICU death was less frequent when the metformin concentration was greater than or equal to 9.9 mg/L (33.9% vs 61.3% for < 9.9 mg/L; p = 0.0252). After adjustment for the Simplified Acute Physiology Score II, in-ICU death was positively associated with prothrombin activity less than 70% and negatively associated with the initiation of renal replacement therapy at admission. CONCLUSIONS: In metformin-treated patients admitted in an emergency context, a plasma metformin concentration greater than or equal to 9.9 mg/L was strongly associated with the presence of lactic acidosis. This threshold may assist with the delicate decision of whether or not to initiate renal replacement therapy. Indeed, the outcome of lactic acidosis might depend on the prompt initiation of renal replacement therapy-especially when liver failure reduces lactate elimination.