C24-02 Cardiovascular Response Patterns in Hypo- and Hyperinflammatory Sepsis Phenotypes

Abstract Rationale Two molecular phenotypes (Hypoinflammatory and Hyperinflammatory) of sepsis with divergent clinical outcomes and biological characteristics have been consistently identified. Previously, we observed that during sepsis, echocardiographic findings were minimally perturbed compared to baseline non-sepsis studies1. However, given the Hyperinflammatory phenotype is characterized by profound inflammation and endothelial dysfunction, we hypothesized that cardiac dysfunction would be most pronounced in this phenotype. Method This was a retrospective cohort study of critically-ill patients admitted to the ICU with sepsis (2010-2020). We analyzed all patients with echocardiograms within 24-hours of ICU admission (N = 2,701), and a subset with preceding non-critical illness echocardiogram within a year of index hospitalization (N = 1,069). We classified molecular phenotypes using a previously validated machine learning model using only vital signs and laboratory variables2. Adjusted logistic regression and linear mixed-effect models were used to evaluate the association between mortality and echocardiographic findings during sepsis and over time, respectively. Result In the complete cohort, 41% were Hyperinflammatory. Hyperinflammatory patients were younger (6151-70 vs. 6655-75 years), with greater norepinephrine (0.45 0.22-0.84 vs. 0.24 0.12-0.46 µg/kg/min), and mortality (46% vs 19%)(p 0.001 all). In Hyperinflammatory phenotype, LVEF(%) was 5952-64, TAPSE(cm) 1.901.55-2.30 and TV S′(cm/s) 12.2 10.0-15.0 and did not differ significantly compared to the Hypoinflammatory phenotype. In both phenotypes, modelling LVEF as a restricted cubic spline revealed a nonlinear pattern, with mortality increasing at both low (OR 0.98, 95% CI 0.96-0.99; p = 0.002) and high LVEF values (OR 1.03, 95% CI 1.02-1.05; p 0.001)), whereas TAPSE (OR 0.81, 95% CI 0.60-1.09; p = 0.159) showed no significant association. In patients with preceding echocardiogram, longitudinal changes from baseline to sepsis were most pronounced in the Hyperinflammatory phenotype compared to Hypoinflammatory, septic shock, or the whole cohort (Figure). Hyperinflammatory patients showed greater decline in LVEF (3.9% vs. 1.3%), TAPSE (0.2cm vs. 0.06cm), and LVOT VTI (2.6cm vs. 0.37cm)(p 0.001 all). Longitudinal decreases in TAPSE (p = 0.017) and stroke volume (p = 0.012) were significantly associated with mortality in the Hyperinflammatory phenotype, but not the Hypoinflammatory phenotype. Conclusion Echocardiographic findings at baseline were similar across molecular phenotypes, but longitudinal analysis revealed divergent decline in cardiac trajectories. The Hyperinflammatory phenotype showed progressive biventricular dysfunction associated with increased mortality, whereas the Hypoinflammatory phenotype exhibited relatively stable function without prognostic significance, suggesting therapies aimed at restoring cardiac function may be beneficial in the Hyperinflammatory. 1. Dugar SP et al. Am J Respir Crit Care Med. 2025;211:A7699.2. Sinha P et al. Am J Respir Crit Care Med. 2020;202:996-1004. This abstract is funded by: none