Bioreactance reliably detects preload responsiveness by the end-expiratory occlusion test when averaging and refresh times are shortened

Annals of Intensive Care

Table 3 Ability of the end-expiratory occlusion test to detect preload responsiveness using three different methods for measuring cardiac index

Variable	AUROC ± SE	Sensitivity (95% CI)	Specificity (95% CI)	LR + (95% CI)	LR− (95% CI)	Cutoff	p
EEXPO—Relative ΔCI_pulse	0.983 ± 0.018	1.00 (0.87–1.00)	0.94 (0.70–1.00)	16.0 (2.4–106.7)	–	3.3%	< 0.0001
EEXPO—Relative ΔCI_{Starling-24.4}	0.680 ± 0.086	0.62 (0.41–0.80)	0.94 (0.70–1.00)	9.9 (1.4–67.3)	0.4 (0.2–0.7)	0.1%	0.036
EEXPO—Relative ΔCI_Starling-8.1	0.899 ± 0.049	0.79 (0.59–0.93)	0.86 (0.57–0.98)	5.54 (1.5–20.3)	0.24 (0.1–0.5)	5.1%	< 0.0001

AUROC area under the receiver operating characteristic curve, EEXPO end-expiratory occlusion, LR + positive likelihood ratio, LR− negative likelihood ratio, PLR passive leg raising; SE standard error, 95% CI 95% confidence interval, ΔCI_pulse changes in cardiac index measured through the pulse contour analysis method, ΔCI_{Starling-24.4} changes in cardiac index detected by the commercial version of the Starling device (averaging time 24 s, refresh time 4 s), ΔCI_Starling-8.1 changes in cardiac index derived through raw data analysis of the Starling device (averaging time 8 s, refresh time 1 s)