Early Vasoactive‐Inotropic Response Trajectories After Polymyxin B Hemoperfusion in Septic Shock: A Single‐Center Retrospective Cohort Study

PURPOSE: To characterize early weight-indexed vasoactive-inotropic score (VIS) trajectories after polymyxin B hemoperfusion (PMX-HP) in septic shock and examine association with 28-day mortality. METHODS: We retrospectively studied 64 adult intensive care unit patients treated with PMX-HP from July 2013 to December 2019. VIS was calculated after body-weight indexing. T1 was before the first session, T2 after the final session, and T3 24 h later. Immediate sustained response required T2 reduction without T3 rebound; delayed response required later reduction; adverse trajectory included no response, rebound, or death before T3. RESULTS: Twenty-nine patients (45.3%) had immediate sustained response, two (3.1%) delayed response, and 33 (51.6%) adverse trajectory. Twelve died before planned T3. Among 60 verified timestamps, median shock-to-treatment interval was 17.3 h (interquartile range, 6.6-27.7); median session duration was 2.0 h (interquartile range, 2.0-2.3). Twenty-eight-day mortality was 31.0%, 0.0%, and 63.6%, respectively. Adverse trajectory had higher unadjusted odds of 28-day mortality than immediate sustained response (odds ratio, 3.89; 95% confidence interval, 1.35-11.22). In an age- and Acute Physiology and Chronic Health Evaluation II-adjusted model excluding delayed response, the adjusted odds ratio was 4.17 (95% confidence interval, 1.35-12.86; p = 0.013). Landmark sensitivity analysis was directionally consistent but imprecise. CONCLUSIONS: Failure to achieve sustained early reduction in weight-indexed vasoactive support, rebound, or death before reassessment identified an adverse early clinical course. Findings support cautious risk stratification, not causal evidence of treatment efficacy.