Abstract Rationale Metabolic syndrome (MetS) is associated with chronic low-grade inflammation and metabolic dysregulation. We previously reported that MetS and its component diagnoses were associated with hypoinflammatory sepsis (PMID: 41072401). As these phenotypes may have differential therapeutic responses to metabolically active drugs such as statins, we evaluated the association between MetS, pre-hospital metabolic drugs, and phenotypes in a large cohort of critical ill adults with sepsis. Methods We performed a retrospective observational cohort study of adults admitted with sepsis to Vanderbilt’s medical intensive care unit from 2004-2024. Sepsis was identified from the electronic health record (EHR) using the Rhee et al. algorithm (PMCID: PMC5710396). Premorbid MetS component diagnoses (obesity, hypertension, diabetes, and/or dyslipidemia) were identified from diagnostic billing codes in the EHR. The primary outcome was sepsis phenotype (hyperinflammatory vs. hypoinflammatory) determined by a validated Clinical Classifier Model using vital signs and laboratory data (PMCID: PMC5710396). Prescriptions for MetS medications (Statins, ACE-inhibitors, ARBs, Beta-blockers, Calcium Channel Blockers CCB, Thiazides, Insulin, and Metformin) prior to hospitalization were identified from the EHR. To account for confounding by indication, we used propensity matching for receipt of each drug class using demographics, MetS diagnoses, and chronic kidney disease. Multivariable logistic regression was used to evaluate associations between number of premorbid MetS component diagnoses and MetS drugs with development of sepsis phenotypes. We performed stratified analyses of MetS and Non-MetS subgroups to test if medication use modified the association between MetS drugs and manifested phenotype. Results We included 13,875 patients: 2,480 (17.9%) had metabolic syndrome and 4,316 (31.11%) had the hyperinflammatory phenotype. In unmatched analyses, each additional MetS component diagnosis was associated with lower odds of hyperinflammatory phenotype (Adjusted OR:0.80, 95%CI 0.74-0.86, p 0.001). Pre-hospital metformin was associated with the hyperinflammatory phenotype (OR:1.50, 95%CI 1.24-1.81, p 0.001), while statins (OR:0.81, 95%CI 0.71-0.92, p = 0.0016) and CCBs (OR:0.83, 95%CI 0.72-0.96, p = 0.015) were associated with the hypoinflammatory phenotype. Propensity matched analyses showed pre-hospital statins, insulin, beta-blockers, ACE inhibitors, and aspirin were associated with hypoinflammatory sepsis only among patients with MetS, whereas these drugs had no effect on phenotype in non-MetS patients (Figure 1). Conclusion Premorbid MetS and exposure to MetS-targeted drugs were associated with hypoinflammatory sepsis. These findings suggest that chronic metabolic dysfunction and pharmacologic modulation may attenuate excessive inflammatory responses, either through improved MetS disease control or via drug effects unrelated to their role in treating MetS, potentially conferring a survival advantage during critical illness. This abstract is funded by: K01 HL157755 (VEK); F30HL170483 (AMB); R01 HL171809, R01 AG069900, R01 AG084550 (WQW); R01HL173531, R35GM142992 (PS); U01 HL168412, R01 HL164937, HL087738 (LBW); UL1 TR002243 (VUMC)
Sharie et al. (Fri,) studied this question.