Background. Although most patients with septic shock have a poor outcome, some may survive after blood purification treatment such as polymyxin B cartridge hemoperfusion (PMX). Objective. To explore the most significant characteristic associated with 28-day mortality in patients with septic shock receiving PMX. Methods. Between April 2006 and March 2008, 116 patients with septic shock who had received PMX in a prospectively collected multicenter collaborative study were enrolled. Uni- and multivariate analyses Wing the Cox proportional hazard model were performed to assess the most significant clinical characteristic that was associated with 28-day mortality. Results. Among 33 clinicolaboratory characteristics, receiver operating characteristic (ROC) curve analyses Selected 12 characteristics with recommended cutoff values such as HCO3- (<= 19.81/>19.8; mEq/L), base excess (<=-5.35/>-5.35; mEq/L), diastolic blood pressure (<= 48/>48 mmHg), mean arterial pressure (<= 73/>73 mmHg), pH (<= 7.29/>7.29), interleukin-6 (<= 19,150/>19,150 pg/dL), prothrombin time International Normalized Ratio (PT-INR; <= 2.05/>2.05), predictive value of Acute Physiology and Chronic Health Evaluation II (APACHE II; <= 0.4/>0.4), pyruvate (<= 1.82/>1.82 mg/dL),. APACHE II score (<= 21/>21), acetate/pyruvate ratio (<= 19/>19), and acetate (<= 44.8/>44.8 mg/dL) on the basis of large area under the ROC curves for 28-day mortality. The results of uni- and multivariate analyses using these selected characteristics revealed that only PT-INR (<= 2.05/>2.05; hazard ratio, 2.823; 95% CI, 1.243-6.412; P = .013) was associated with 28-day mortality. Survival curve analysis demonstrated a significant difference in 28-day mortality between patients with lower (<= 2.05) and higher (>2.05) PT-INR (P < .001). Conclusion. Prolonged PT-INR is an independent risk factor for 28-day mortality in patients receiving PMX for septic shock.