Lactate and lactylation in sepsis-associated acute kidney injury: clinical evidence from the MIMIC-IV database and mechanistic insights

Sujing Zhang¹Mengyuan Luo¹Zhijie Lu^1*Qiqing Shi^1,2*

• ¹Department of Anesthesiology, Fudan University Minhang Hospital, Shanghai, China
• ²Shanghai Fourth People's Hospital Department of Anesthesiology and Perioperative Medicine, Shanghai, China

Background: Sepsis-associated acute kidney injury (SA-AKI) is a common and severe complication of sepsis, yet early predictors remain limited. Lactate, beyond being a marker of tissue hypoperfusion, may act as a signaling molecule through protein lactylation. This study aimed to investigate the association between lactate levels and SA-AKI risk using the MIMIC-IV database and to explore the potential mechanistic role of lactylation in septic mice. Methods: Adult sepsis patients were identified from the MIMIC-IV (v3.1) database. Patients were stratified by lactate tertiles within 24 h of ICU admission, and lactate clearance (LC) was assessed as a dynamic indicator. The primary outcome was SA-AKI, and secondary outcomes included CRRT use and 28-day mortality. Kaplan–Meier analysis, Cox regression, and restricted cubic spline (RCS) models were performed. In parallel, a murine cecal ligation and puncture (CLP) model was used to evaluate tissue-specific protein lactylation by Western blot, along with serum lactate and hematological parameters. Results: A total of 11,431 patients were included. Higher lactate levels were associated with increased disease severity, higher incidence and severity of SA-AKI, greater use of CRRT, and elevated 28-day ICU and in-hospital mortality. In Cox regression, lactate as both a continuous and categorical variable was independently associated with SA-AKI risk. RCS analysis revealed nonlinear dose–response relationships, with sharply increased risk of SA-AKI above 5.7 mmol/L. In sensitivity analyses (n=6,593), higher LC was associated with lower risks of SA-AKI, CRRT use, and mortality. Cox regression confirmed LC as an independent protective factor, while RCS suggested a downward but non-significant trend. In the CLP mouse model, protein lactylation was markedly elevated in the kidney but not in the heart, liver, or spleen, accompanied by higher serum lactate. Conclusions: Elevated lactate levels are independently associated with increased risk of SA-AKI in sepsis patients, whereas higher lactate clearance is linked to improved renal outcomes. Moreover, kidney-specific upregulation of protein lactylation in septic mice suggests a possible molecular link between lactate metabolism and renal vulnerability. These findings highlight lactate and lactylation as both prognostic markers and potential mechanistic contributors in SA-AKI.