(1447) Hypo- and Hyperinflammatory Subphenotypes in Sepsis Within the Individual

Introduction: Sepsis is a severe and life-threatening condition that affects approximately 10% of patients admitted to the ICU in the United States yearly and results in annual healthcare costs of around $20 billion1. Additionally, mortality in cases of septic shock can be as high as 38% in one recent review and can reach levels of >50% in patients with vasoplegia requiring high doses of vasopressors.1,2 Recently, latent class analysis methods have identified distinct subphenotypes of patients suffering from ARDS, and these subphenotypes have been predictive of responses to therapies. We suspect that a similar LCA model applied to patients suffering from sepsis may identify subsets of patients that may respond differently to the multiple therapies we use to treat sepsis. Our study aims to determine whether such a model would consistently predict subphenotypes over multiple admissions to the ICU.

Methods: This study was a retrospective cohort study from a convenience sample of 42 patients admitted to BJC hospitals from 2017 to 2022. ICD 9 and 10 codes were used to select for patients who underwent multiple admissions to the ICU for sepsis, and a chart review was subsequently done to verify that patients underwent two distinct and separate sepsis admissions. A previously reported set of clinical and biological values used to perform latent class analysis in both ARDS and sepsis populations was collected. These values have previously been shown to determine two classes in these populations of hypo- and hyper-inflammatory4. A one-way ANOVA was used to determine variability in single patients over multiple sepsis admissions.

Results: 34 patients met criteria of admission for sepsis over 30 days apart. There was no significant difference in values in single patients over separate sepsis admissions.

Conclusions: This is a small study which hopes to add to the recent literature on inflammatory subtypes seen in the ICU. If patients always present with a certain subphenotype or if pathogens are more frequently associated with a subphenotype, this could be useful information to have for treatment while in the hospital. This also may lead to further investigation of these subphenotypes at an individual level.