Dynamic arterial elastance (Ea _dyn ) has been extensively considered as a functional parameter of arterial load. However, conflicting evidence has been obtained on the ability of Ea _dyn to predict mean arterial pressure (MAP) changes after fluid expansion. This meta-analysis sought to assess the predictive performance of Ea _dyn for the MAP response to fluid expansion in mechanically ventilated hypotensive patients.

We systematically searched electronic databases through November 28, 2020, to retrieve studies that evaluated the association between Ea _dyn and fluid expansion-induced MAP increases in mechanically ventilated hypotensive adults. Given the diverse threshold value of Ea _dyn among the studies, we only reported the area under the hierarchical summary receiver operating characteristic curve (AUHSROC) as the primary measure of diagnostic accuracy.

Eight observational studies that included 323 patients with 361 fluid expansions met the eligibility criteria. The results showed that Ea _dyn was a good predictor of MAP increases in response to fluid expansion, with an AUHSROC of 0.92 [95% confidence interval (CI) 0.89 to 0.94]. Six studies reported the cut-off value of Ea _dyn , which ranged from 0.65 to 0.89. The cut-off value of Ea _dyn was nearly conically symmetrical, most data were centred between 0.7 and 0.8, and the mean and median values were 0.77 and 0.75, respectively. The subgroup analyses indicated that the AUHSROC was slightly higher in the intensive care unit (ICU) patients (0.96; 95% CI 0.94 to 0.98) but lower in the surgical patients in the operating room (0.72; 95% CI 0.67 to 0.75). The results indicated that the fluid type and measurement technique might not affect the diagnostic accuracy of Ea _dyn . Moreover, the AUHSROC for the sensitivity analysis of prospective studies was comparable to that in the primary analysis.

Ea _dyn exhibits good performance for predicting MAP increases in response to fluid expansion in mechanically ventilated hypotensive adults, especially in the ICU setting.