907: Associations Between Ppi Trajectories and Clinical Outcomes in Shock Patients in Icu

Introduction: One of the core pathological features of shock is microcirculatory dysfunction, where impaired tissue perfusion drives progressive organ failure. The Perfusion Pressure Index (PPI) enables non-invasive, real-time hemodynamic assessment, offering critical insights into microvascular perfusion. Although requiring further validation, PPI demonstrates strong prognostic utility for guiding resuscitation strategies and predicting patient outcomes. This study leverages Group-Based Trajectory Modeling (GBTM) to decode PPI dynamics and their clinical implications. Methods: We conducted a prospective multicenter observational study across 20 hospitals in Southwest China, enrolling ICU patients with shock. All participants underwent comprehensive monitoring including shock-related hemodynamic parameters, laboratory tests, and ultrasonographic indicators. Regular PPI monitoring was performed for all patients. Group-Based Trajectory Modeling (GBTM) was employed to analyze the dynamic patterns of PPI and their association with clinical outcomes in shock patients. Results: A total of 421 patients were included. Using the GBTM model, PI dynamics were classified into 5 groups with optimal fit (average posterior probability >0.7, relative entropy=0.819):Class1 (21.14%): Low baseline, slow increase, Class2 (50.36%): Low baseline, moderate increase, Class3 (21.85%): Moderate baseline, slow increase,Class4 (1.19%): Low baseline, rapid increase, Class5 (5.46%): High baseline, slow increase. Trajectory groups were significantly associated with discharge outcomes (χ2=27.10, P< 0.001) and ICU discharge outcomes (χ2=26.84, P< 0.001). Class1 (21.14%) had the highest mortality rate (48.31%), while no deaths occurred in Class4 (1.19%). For discharge outcomes, KM curves showed significant intergroup differences at 7, 10, and 15 days (P< 0.05). Cox regression indicated lower discharge risk in Classes2–5 versus Class1; at 28 days, unadjusted HR=0.632 (P=0.017) and adjusted HR=0.658 (P=0.032) for Class2, and unadjusted HR=0.415 (P=0.002) and adjusted HR=0.452 (P=0.005) for Class3. Conclusions: Our study identified five distinct PPI trajectory patterns during shock resuscitation, with their early-phase dynamic changes demonstrating predictive value for clinical outcomes.