Association between lactate-to-albumin ratio and mortality in patients with sepsis-induced coagulopathy: a retrospective cohort study

This study aims to investigate the association between the lactate-to-albumin ratio (LAR) and adverse outcomes in critically ill patients with sepsis-induced coagulopathy (SIC). Clinical data from 1960 patients admitted to the Intensive Care Unit (ICU) with SIC were extracted from the MIMIC-IV (3.1) database. The study population was divided into four groups according to quartiles of LAR. The primary outcome was 28-day all-cause mortality, and the secondary outcome was ICU mortality. Kaplan-Meier curves, Cox proportional hazards regression, restricted cubic spline analyses, and subgroup analyses were used to investigate the association between LAR and adverse outcomes. Additionally, receiver operating characteristic (ROC) curve analysis was performed to evaluate the predictive ability of LAR for adverse outcomes in patients with SIC. Finally, a nomogram model was constructed using the Boruta algorithm to integrate key predictors. This study included a total of 1960 patients with SIC. The LAR was independently associated with both 28-day all-cause mortality and ICU mortality (both p < 0.01). Kaplan-Meier survival curve analysis demonstrated that patients with higher LAR levels had significantly lower 28-day and ICU survival rates (Log-rank test, p < 0.001).Restricted cubic spline (RCS) analysis indicated a linear relationship between LAR and mortality risk (overall p < 0.001; nonlinearity p = 0.068). ROC analysis showed that LAR had moderate predictive accuracy. When combined with the SOFA score, its predictive performance improved further. Building upon this, the final nomogram achieved the highest predictive accuracy for 28-day mortality. Compared with traditional scoring systems, the newly developed model demonstrated significant improvements in both reclassification and discrimination, as reflected by the Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI).Subgroup analyses revealed that this association remained consistent across all subgroups. Using Boruta to select predictive factors, a nomogram incorporating LAR was constructed, demonstrating good predictive performance for 28-day all-cause mortality (AUC 0.779, 95% CI 0.756-0.802).An online web server was built (https://shaojian.shinyapps.io/sic-mortality-predictor/) to facilitate the use of the nomogram. The LAR exhibits a linear relationship with 28-day all-cause mortality and ICU mortality in SIC patients and can serve as an independent predictive factor.