Meta analysis of the predictive efficacy of various derived indicators of sequential organ failure assessment in outcomes of patients with sepsis / 中华危重病急救医学

ABSTRACT

Objective:To systematically review and evaluate the predictive efficacy of various derived indicators of sequential organ failure assessment (SOFA) in mortality rate of sepsis patients.Methods:Literature on sepsis and SOFA scores were searched in PubMed, Embase and Cochrane Library. The retrieval time will be set to the time of database-building to February, 2023. The main outcome measures included 28-day mortality, 30-day mortality, in-hospital mortality, intensive care unit (ICU) mortality and long-term mortality. Literature screening, data extraction and quality evaluation were carried out independently by 2 researchers. Data were analyzed by Revman 5.3.5, Meta-disc and Stata software. Deek funnel plots were used to assess publication bias in the included studies.Results:A total of 40 articles including 51 trials were included. Of these, 32 were in English and 8 in Chinese, 17 were in prospective trials and 34 were in retrospective trials, 38 were in initial SOFA-related trials and 9 were in the change of SOFA score (ΔSOFA)-related studies, a total of 59?962 patients were enrolled. ① The area under the receiver operator characteristic curve (AUC) of initial SOFA and ΔSOFA for predicting outcome in sepsis was 0.773 and 0.787 ( Z = 0.115, P > 0.05), respectively. There was no significant difference between the two indexes in predicting the outcome of patients with sepsis. ② In subgroup analysis, due to limitations in the number of literature articles, the 28-day mortality rate and 30-day mortality rate were merged for discussion. The predictive power of ΔSOFA for 28-day or 30-day mortality was significantly higher than that of initial SOFA (AUC was 0.854, 0.787, Z = 2.603, P ≤ 0.01). ③ There were few studies onΔSOFA in predicting in-hospital mortality, ICU mortality and long-term mortality of sepsis patients. The AUC of the initial SOFA for predicting the study endpoints described above was: ICU mortality (0.814) > 28-day or 30-day mortality (0.787) > in-hospital mortality (0.697) > long-term mortality (0.646). ④ Initial SOFA and ΔSOFA in patients with sepsis of non-Han original had good predictive performance and there was no significant difference between them (AUC was 0.766, 0.811, respectively). However, the pooled sensitivity of ΔSOFA was higher (92%). ⑤ In prospective studies, initial SOFA was better at predicting outcomes in patients with sepsis (AUC was 0.804, pooled sensitivity 64%). The sensitivity of ΔSOFA indicators in predicting the outcome of sepsis patients was significantly higher than the initial SOFA (78% vs. 64%). The funnel plot showed that there was no significant publication bias in the included literature. Conclusion:ΔSOFA has a relatively high diagnostic efficacy in predicting short-term (28-day or 30-day) mortality in patients with sepsis.