[Effective arterial elastance in evaluating the fluid challenge in septic shock patients]

To explore the validity of the effective arterial elastance (Ea) before and after fluid challenge in evaluating the fluid challenge in septic shock patients.

A retrospective study was conducted in the medical intensive care unit (MICU) of Peking Union Medical College Hospital from October 2016 to October 2020. 116 septic shock patients were enrolled. All patients received fluid challenge by 500 mL Gelatin or normal saline under invasive hemodynamic monitoring. Heart rate (HR), mean arterial pressure (MAP), cardiac output (CO) and other hemodynamic variables were collected at 10 minutes before and immediately after fluid challenge. An increase in CO greater than 10% after fluid challenge was defined as the positive preload responsiveness, as well as the definition of positive pressure responsiveness was an increase in MAP greater than 10%. Receiver operating characteristic curves (ROC curves) were established to evaluate the predictive abilities of baseline Ea and other arterial load indices in detecting the preload responders and pressure responders. The correlation of the baseline Ea with CO changes after fluid challenge as well as MAP changes were tested by Pearson correlation analysis. Patients with positive preload responsiveness were divided into two groups according to the pressure responsiveness. The changes in Ea and other arterial load indices were analyzed.

A total of 116 patients were finally analyzed. Sixty-three patients were preload responders and 53 patients were preload non-responders. There was no significant difference in demographics and baseline physical variables between the two groups. Ea in preload responders was higher than that in preload non-responders (mmHg/mL: 2.51±1.08 vs. 1.87±0.68, P < 0.01). ROC curve analysis showed that the baseline Ea could predict the preload responsiveness at an area under ROC curve (AUC) = 0.71 [95% confidence interval (95%CI) was 0.62-0.81, P < 0.001]. The cut-off value was 1.97 mmHg/mL with a sensitivity of 71.4% and a specificity of 60.4%. The baseline Ea did not present the predictive ability to detect the pressure responders and pressure non-responders (AUC = 0.52, 95%CI was 0.41-0.63, P = 0.73). Pearson correlation analysis showed that the changes in CO after fluid challenge was moderately correlated to the baseline Ea (r = 0.47, P < 0.001), meanwhile a weak positive correlation between the changes in MAP and baseline Ea was found (r = 0.20, P = 0.03). In preload responders, 27 (42.9%) of 63 patients were pressure responders and 36 (57.1%) patients were pressure non-responders. No statistical difference was found in the baseline Ea or other arterial load indices between the two groups. Fluid challenge decreased Ea both in pressure non-responders and pressure responders (mmHg/mL: 2.13±0.94 vs. 2.51±1.08, P < 0.01; 2.47±1.18 vs. 2.69±1.30, P < 0.05). Moreover, the changes in CO and changes in MAP were strongly correlated with the changes in Ea (r values were -0.50 and 0.58, respectively, both P < 0.001).

The Ea > 1.97 mmHg/mL before fluid challenge could predict fluid responsiveness in septic shock patients. The baseline Ea was not able to predict the subsequent changes in arterial pressure through fluid challenge. A significant decrease in Ea inducing by fluid administration explained why patients increased their CO without improving blood pressure.