To assess the value of passive leg raising (PLR) test in predicting fluid responsiveness in patients with sepsis-induced cardiac dysfunction.A prospective observational cohort study was conducted. Thirty-eight patients under mechanical ventilation suffering from sepsis-induced cardiac dysfunction admitted to Department of Surgical Intensive Care Unit of First Affiliated Hospital of Sun Yat-Sen University from September 2013 to July 2014 were enrolled. The patients were studied in four phases: before PLR (semi-recumbent position with the trunk in 45°), PLR (the lower limbs were raised to a 45° angle while the trunk was in a supine position), before volume expansion (VE, return to the semi-recumbent position), and VE with infusing of 250 mL 5% albumin within 30 minutes. Hemodynamic parameters were recorded in every phase. The patients were classified into two groups according to their response to VE: responders (at least a 15% increase in stroke volume, A SV(VE) ≥ 15%), and non-responders. The correlations among all changes in hemodynamic parameters were analyzed by linear correlation analysis, and the receiver operating characteristic curve (ROC) was plotted to assess the value of hemodynamic parameters before and after PLR in predicting fluid responsiveness.Of 38 patients, 25 patients were responders, and 13 non-responders. There was no significant difference in the baseline and hemodynamic parameters at semi-recumbent position between the two groups. The changes in SV and cardiac output (CO) after PLR (Δ SV(PLR) and Δ CO(PLR)) were significantly higher in responders than those of non-responders [ Δ SV(PLR): (14.7 ± 5.7)% vs. (6.4 ± 5.3)%, t = 4.304, P = 0.000; ΔCO(PLR): (11.2 ± 7.5)% vs. (3.4 ± 2.3)%, t = 3.454, P = 0.001], but there was no significant difference in the changes in systolic blood pressure, mean arterial pressure, pulse pressure, and heart rate after PLR (Δ SBP(PLR), Δ MAP(PLR), Δ PP(PLR) and Δ HR(PLR)) between two groups. Δ SV(VE) in responders was significantly higher than that of the non-responders [(20.8 ± 5.5) % vs. (5.0 ± 3.7) %, t = 8.347, P = 0.000]. It was shown by correlation analysis that ΔSV(PLR) was positively correlated with Δ SV(VE) (r = 0.593, P = 0.000), Δ CO(PLR) was positively correlated with ΔSV(VE) (r = 0.494, P = 0.002). The area under ROC curve (AUC) of Δ SV(PLR) ≥ 8.1% for predicting fluid responsiveness was 0.860 ± 0.062 (P = 0.000), with sensitivity of 92.0% and specificity of 70.0%; the AUC of Δ CO(PLR) ≥ 5.6% for predicting fluid responsiveness was 0.840 ± 0.070 (P = 0.000), with sensitivity of 84.0% and specificity of 76.9%; the AUC of Δ MAP(PLR) ≥ 6.9% for predicting fluid responsiveness was 0.662 ± 0.089, with sensitivity of 68.0% and specificity of 76.9%; the AUC of Δ SBP(PLR) ≥ 6.4% for predicting fluid responsiveness was 0.628 ± 0.098, with sensitivity of 76.0% and specificity of 61.5%; the AUC of Δ PP(PLR) ≥ 6.2% for predicting fluid responsiveness was 0.502 ± 0.094, with sensitivity of 56.0% and specificity of 53.8%; the AUC of Δ HR(PLR) ≥ -1.7% for predicting fluid responsiveness was 0.457 ± 0.100, with sensitivity of 56.0% and specificity of 46.2%.In patients with sepsis-induced cardiac dysfunction, changes in SV and CO induced by PLR are accurate indices for predicting fluid responsiveness, but the changes in HR, MAP, SBP and PP cannot predict the fluid responsiveness.