Table 1 Characteristics of the included studies
From: Arterial hyperoxia and mortality in critically ill patients: a systematic review and meta-analysis
Study | Design | Country | Hyperoxia, definition | Comparator group | Outcome measure reported | ||
|---|---|---|---|---|---|---|---|
PaO 2/ABG | Time of assessment | Cutoff value | |||||
Mechanically ventilated ICU patients | |||||||
de Jonge et al. 2008 [16] | Retrospective cohort, multicenter | Netherlands | Worst PaO2 | First 24 hours | ≥120 mmHg (upper quintile) | PaO2 between 66 and 80 mmHg | In-hospital mortality |
Eastwood et al. 2012 [20] | Retrospective cohort, multicenter | Australia, New Zealand | Worst PaO2 | First 24 hours | >120 mmHg for unadjusted analysis; ≥305 mmHg (upper decile) for adjusted analysis | PaO2 <120 mmHg for unadjusted analysis; PaO2 between 75 and 85 mmHg for adjusted analysis | In-hospital mortality |
Suzuki et al. 2013 [2] | Prospective observational cohort, single-center | Australia | Time-weighted SpO2 | Whole period of mechanical ventilation | Time-weighted SpO2 >98% | Not exposed to hyperoxia | In-hospital mortality |
Suzuki et al. 2014 [32] | Prospective before-after, single-center | Australia | Conventional period: oxygenation goals at the discretion of the attending physicians | Conservative period: SpO2 between 90 and 92% | 28-day mortality | ||
Post-cardiac arrest patients | |||||||
Bellomo et al. 2011 [21] | Retrospective cohort, multicenter | Australia, New Zealand | Worst PaO2 | First 24 hours | ≥300 mmHg | Normoxia | In-hospital mortality |
Ihle et al. 2013 [31] | Retrospective cohort, multicenter | Australia | Worst PaO2 | First 24 hours | ≥300 mmHg | Normoxia | In-hospital mortality |
Janz et al. 2012 [35] | Retrospective analysis of a prospective cohort study, single-center | USA | Highest PaO2 | First 24 hours | ≥300 mmHga | Not exposed to hyperoxia | In-hospital mortality |
Kilgannon et al. 2010 [17] | Retrospective cohort, multicenter | USA | First PaO2 | First 24 hours | ≥300 mmHg | Not exposed to hyperoxia | In-hospital mortality |
Lee et al. 2014 [36] | Retrospective cohort, single-center | Korea | Mean PaO2 | From return of spontaneous circulation to the end of therapeutic hypothermia | ≥156.7 mmHg (upper quartile) | PaO2 between 116 and 134.9 mmHg (second quartile) | In-hospital mortality |
Nelskyla et al. 2013 [37] | Retrospective analysis of a prospective cohort study, single-center | Australia | Highest PaO2 | First 24 hours | ≥300 mmHg | Not exposed to hyperoxia | In-hospital mortality |
Stroke patients | |||||||
Rincon (a) et al. 2014 [19] | Retrospective cohort, multicenter | USA | First PaO2 | First 24 hours | ≥300 mmHg | Normoxia | In-hospital mortality |
Young et al. 2012 [23] | Retrospective cohort, multicenter | Australia and New Zealand | Worst PaO2 | First 24 hours | >341 mmHg (upper decile) | Normoxia (PaO2 >69 and <341 mmHg, 2nd to 9th deciles) | In-hospital mortality |
Traumatic brain injury | |||||||
Asher et al. 2013 [33] | Retrospective cohort, single-center | USA | Highest PaO2 | First 72 hours | ≥200 mmHg | Not exposed to hyperoxia | In-hospital mortality |
Brenner et al. 2012 [18] | Retrospective cohort, multicenter | USA | Mean PaO2 | First 24 hours | >200 mmHg | Normoxia | In-hospital mortality |
Davis et al. 2009 [34] | Retrospective cohort, multicenter | USA | First PaO2 | On arrival | >487 mmHg | Not exposed to hyperoxia | In-hospital mortality |
Raj et al. 2013 [22] | Retrospective cohort, multicenter | Finland | Worst PaO2 | First 24 hours | >100 mmHg | Normoxia | In-hospital mortality |
Rincon (b) et al. 2014 [38] | Retrospective cohort multicenter | USA | First PaO2 | First 24 hours | ≥300 mmHg | Normoxia | In-hospital mortality |