Evaluation of plasma lactate parameters for predicting mortality of septic patients

OBJECTIVE: To compare the accuracy of serum lactate parameters, including lactate peak concentration (LACpeak), lactate time area (LACarea), and lactate clearance (LC) for predicting mortality of the septic patients, and to compare with the predictive accuracy of National Early Warning Score (NEWS) and Sequential Organ Failure Assessment (SOFA) scores. METHODS: This study retrospectively screened the septic patients admitted to the ICU in the Medical Information Mart for Intensive Care IV (MIMIC-IV) from 2008 to 2019. The baseline data and outcomes of patients were gathered. The subjects were divided into the non-survival group and the survival group. SOFA, NEWS, LACpeak, and LACarea were recorded. The LC was calculated 6 h after LACpeak. The above parameters were compared by the T-test and Mann-Whitney U test, and odds ratios were calculated adjusting for age and sex. The receiver operating characteristic curves (ROCs) of subjects were plotted according to SOFA, NEWS, LACpeak, and LACarea within 24h, and LC at 6h of ICU admission. The Areas under the ROC curve (AUCs), sensitivity, and specificity were compared with R version 4.1.1. RESULTS: 1,169 septic patients were involved, and 366 (31.3%) patients died within 28 days. Compared to the survival group, the LACpeak of the non-survival group was higher [4.85 (3.2, 7.9) vs. 3.4 (2.6, 5.25) mmol/L, adjusted odds ratio 1.18, P < 0.001], and the LACarea of the non-survivals was higher than the survivals too [18.44 (10.36, 27.63) vs. 13.65 (9.01, 21.73), adjusted odds ratio 1.03, P < 0.001)]. The LC of the survivals at 6 h after LACpeak was significantly higher than that of the non-survivals [0.26 (0.14.0.42) vs. 0.19 (0.10, 0.33), adjusted odds ratio 0.06, P < 0.01]. Within 24h of ICU admission, the AUCs of mortality prediction in descending order were NEWS [0.73 (0.70, 0.76)], SOFA [0.69 (0.66, 0.73)], LACpeak [0.64 (0.61, 0.68)], and LACarea [0.60 (0.56, 0.63)]. There were 204 patients with 6-hour LC after LACpeak the AUCs of LACarea, LACpeak and LC were 0.73(0.65, 0.80), 0.71(0.62,0.78) and 0.65 (0.56, 0.73), respectively. CONCLUSIONS: The predictive accuracy of LC was not superior to LACpeak and LACarea for the mortality of the septic patients and the predictive value of all the above lactate parameters for mortality maybe not better than SOFA and NEWS.