Subgroup analysis.

Objective
To investigate whether early administration of serum albumin infusion in the acute pancreatitis (AP) patients admitted to the intensive care unit (ICU) reduces the risk of sepsis.
Methods
Data were collected from the Medical Information Mart for Intensive Care III and IV databases for this retrospective cohort study. The primary outcome was the occurrence of sepsis in AP patients. We used univariate and multivariate logistic regression models to evaluate the association between albumin infusion and sepsis risk of AP patients. Additional subgroup analysis by stratification to serum albumin were also performed.
Results
The study included 779 patients with AP. They were divided into a sepsis group (83 patients) and a non-sepsis group (696 patients) according to whether they developed sepsis, and the prevalence of sepsis was approximately 10.65%. Multivariate logistic regression model indicated that albumin infusion was associated with decreased risk of sepsis in the AP patients [adjusted odds ratio (OR)=0.37, 95% confidence interval (CI)=0.13–0.88]. Subgroup analysis showed a negative association between the albumin infusion and sepsis risk in the AP patients with serum albumin ≤ 3.5g/L (adjusted OR=0.29, 95%CI = 0.08–0.77).
Conclusion
In this study, we found an association between albumin administration and a lower risk of sepsis in AP patients, which persisted after multivariate adjustment. This suggests that albumin infusion may have unique potential benefits for this population.