This study determined if the use of metformin affected the prognostic value of hyperlactatemia in predicting 28-day mortality among patients with sepsis and bacteremia. We enrolled adult diabetic patients with sepsis and bacteremia. Of 590 patients, 162 and 162 metformin users and nonusers, respectively, were selected in propensity matching. The mean serum lactate levels in metformin users were higher than those in nonusers (4.7 vs. 3.9 mmol/L, p = 0.044). We divided the patients into four groups based on quick Sepsis-related Organ Failure Assessment (qSOFA) scores. No significant difference was found among nonusers with qSOFA score <2, nonusers with qSOFA score ≥2, and metformin users with qSOFA score <2. The lactate levels in metformin users with qSOFA score ≥2 were higher than those in other groups, and significant differences were found in both nonsurvivors (8.9 vs. 4.6 mmol/L, p = 0.027) and survivors (6.4 vs. 3.8 mmol/L, p = 0.049) compared with metformin users with qSOFA score <2. The best cut-off point to predict 28-day mortality in metformin users (5.9 mmol/L; area under the receiver operating characteristic curve (AUROC), 0.66; 95% confidence interval (CI), 0.55⁻0.77) was higher than that in nonusers (3.6 mmol/L; AUROC 0.63; 95% CI, 0.56⁻0.70). Metformin users had higher lactate levels than nonusers in increasing sepsis severity. Serum lactate levels could be useful in predicting mortality in patients using metformin, but higher levels are required to obtain more precise results.
Keywords: lactate; metformin; mortality; sepsis.