Limitations of global end-diastolic volume index as a parameter of cardiac preload in the early phase of severe sepsis: a subgroup analysis of a multicenter, prospective observational study.

Background: In patients with severe sepsis, depression of cardiac performance is common and is often associated with left ventricular (LV) dilatation to maintain stroke volume. Although it is essential to optimize cardiac preload to maintain tissue perfusion in patients with severe sepsis, the optimal preload remains unknown. This study aimed to evaluate the reliability of global end-diastolic volume index (GEDI) as a parameter of cardiac preload in the early phase of severe sepsis.

Methods: Ninety-three mechanically ventilated patients with acute lung injury/acute respiratory distress syndrome secondary to sepsis were enrolled for subgroup analysis in a multicenter, prospective, observational study. Patients were divided into two groups-with sepsis-induced myocardial dysfunction (SIMD) and without SIMD (non-SIMD)-according to a threshold LV ejection fraction (LVEF) of 50% on the day of enrollment. Both groups were further subdivided according to a threshold stroke volume variation (SVV) of 13% as a parameter of fluid responsiveness.

Results: On the day of enrollment, there was a positive correlation (r = 0.421, p = 0.045) between GEDI and SVV in the SIMD group, whereas this paradoxical correlation was not found in the non-SIMD group and both groups on day 2. To evaluate the relationship between attainment of cardiac preload optimization and GEDI value, GEDI with SVV ≤13% and SVV >13% was compared in both the SIMD and non-SIMD groups. SVV ≤13% implies the attainment of cardiac preload optimization. Among patients with SIMD, GEDI was higher in patients with SVV >13% than in patients with SVV ≤13% on the day of enrollment (872 [785-996] mL/m(2) vs. 640 [597-696] mL/m(2); p < 0.001); this finding differed from the generally recognized relationship between GEDI and SVV. However, GEDI was not significantly different between patients with SVV ≤13% and SVV >13% in the non-SIMD group on the day of enrollment and both groups on day 2.

Conclusions: In the early phase of severe sepsis in mechanically ventilated patients, there was no constant relationship between GEDI and fluid reserve responsiveness, irrespective of the presence of SIMD. GEDI should be used as a cardiac preload parameter with awareness of its limitations.