Twenty-four-hour mechanical power variation rate is associated with mortality among critically ill patients with acute respiratory failure: a retrospective cohort study

Table 3 AUC-ROC of baseline mechanical power, 24-h MP and power variation rate for ICU mortality in different populations

Population	Number of patients	ICU mortality, n (%)	APACHE II score	AUC-ROC (95% CI) for baseline MP	AUC-ROC (95% CI) for 24-h MP	AUC-ROC (95% CI) for 24-h MP variation rate
Unselected ICU patients	8597	405 (4.7)	13 (10, 17)	0.752 (0.743–0.761)	–	–
ARF at baseline with mechanical ventilation > 48 h	1925	322 (16.7)	18 (13, 23)	0.634 (0.612–0.655)	0.613 (0.591–0.635)	0.508 (0.485–0.531)*
ARF with baseline MP > 10 J/min	602	166 (27.8)	20 (15, 26)	0.595 (0.554–0.634)	0.630 (0.590–0.669)	0.584 (0.543–0.623)
ARF with baseline MP > 15 J/min	107	43 (40.2)	23 (18, 30)	0.526 (0.427–0.623)	0.683 (0.586–0.769)	0.683 (0.586–0.770)*

AUC-ROC, areas under curve of receiver operating characteristic; CI, confidence interval; ICU, intensive care unit; ARF, acute respiratory failure; MP, mechanical power
^*p < 0.05 compared with AUC-ROC of baseline MP as reference by DeLong’s test

ISSN: 1471-2466