A combined model based on lactate clearance rate and SOFA score in prognostic assessment of patients with severe pneumonia

To investigate the role of lactate clearance rate (LCR) and Sequential Organ Failure Assessment (SOFA) score in prognostic assessment of patients with severe pneumonia (SP). A retrospective study of 185 SP patients was conducted. Baseline characteristics were compared between survivors (n = 128) and non-survivors (n = 57). Following univariate and correlation analyses, significant variables were used to build a combined prediction model via binary logistic regression. Model calibration and predictive performance for 28-day mortality were assessed using Hosmer-Lemeshow and ROC curve analyses, respectively. Dynamic changes in LCR and SOFA scores at admission (T0), 6 h (T6), and 24 h (T24) were compared. The 28-day mortality rate was 30.81% (57/185). Univariate analysis revealed statistically significant differences in age, chronic lung disease, procalcitonin (PCT), LCR, and SOFA score (P < 0.05). Spearman correlation analysis showed a moderate negative correlation between LCR and SOFA score (r = -0.401, P < 0.001). Multivariate analysis identified LCR 0.911 (0.861–0.965) and SOFA score 1.508 (1.278–1.779) as independent predictors (P < 0.05). The combined model demonstrated superior predictive accuracy (AUC: 0.882, 95% CI: 0.833–0.932) compared to LCR (AUC: 0.793) or SOFA score (AUC: 0.832) alone (Delong test, P < 0.05). Survivors exhibited significantly higher LCR and lower SOFA scores at T6 and T24, with improving trends over time, unlike non-survivors. The combined model integrating LCR and SOFA score outperforms either parameter alone in predicting 28-day mortality risk in SP, offering a potential tool for early risk stratification.