Table 3.

Diagnostic accuracy of the end-expiratory occlusion test in the including studies

	No. of patients	AUROC	95% CI	Threshold (%)a	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Monnet et al. [8]	34	0.97	0.85–1.00	5	91	100	100	84
Monnet et al. [18]	39	0.97	0.91–1.00	5	100	91	90	100
Monnet et al. [11]	54	0.95	NA	5	93	92	94	91
Silva et al. [19]	34	0.96	0.82–0.99	6	100	90	86	100
Guinot et al. [12]	42	0.78	0.63–0.89	2.3	82	71	85	66
Biais et al. [13]	41	0.91	0.81–1.00	5	100	81	83	100
Myatra et al. [14]	30	0.95	0.88–1.00	4.1	88	93	93	87
Yonis et al. [15]	33	0.65	0.46–0.84	10	33	100	100	64
Jozwiak et al. [16]	30	0.98	0.85–1.00	4	93	100	100	93
Georges et al. [17]	50	0.96	NA	9	89	95	96	87
Dépret et al. [20]	28	0.95	0.79–0.99	3	86	93	92	87
Messina et al. [21]	40	0.93	0.84–1.00	3.6	89	86	87	88
Xu et al. [22]	75	0.9	0.83–0.97	5	81	93	91	84

AUROC area under the receiver operating characteristic, CI confidence interval, NA not available, NPV negative predictive value, PPV positive predictive value

^aThreshold of increase in cardiac output induced by the test reported as providing the best compromise between sensitivity and specificity