Higher versus lower positive end-expiratory pressure in patients without acute respiratory distress syndrome: a meta-analysis of randomized controlled trials

Table 1 Grades of recommendation, assessment, development, and evaluation summary of findings table

Certainty assessment								Summary of findings
Participants (studies)	Risk of bias		Inconsistency	Indirectness	Imprecision	Other considerations	Overall certainty of evidence	Higher PEEP (N/total or N)		Lower PEEP (N/total or N)	Relative effect (RR, MD or SMD, 95% CI) of higher PEEP
Hospital mortality
1502 (9 RCTs)	Very serious^a	Not serious		Not serious	Not serious	None	⨁⨁◯◯ LOW	274/760 (36.1%)	259/742 (34.9%)		1.02 (0.89 to 1.16)
PaO₂/FiO₂
1444 (8 RCTs)	Serious^b		Serious^c	Not serious	Not serious	Strong association	⨁⨁⨁◯ MODERATE	732		712	50.46 mmHg (33.93 to 66.99)
Compliance
189 (3 RCTs)	Very serious^d		Not serious	Not serious	Not serious	None	⨁⨁◯◯ LOW	101		88	8.46 mL/cmH₂O (3.11 to 13.82)
Hypoxemia
1320 (5 RCTs)	Very serious^e		not Serious	Not serious	Not serious	None	⨁⨁◯◯ LOW	121/667 (18.4%)		168/653 (25.7%)	0.60 (0.40 to 0.92)
Barotrauma
1372 (7 RCTs)	Very serious^f		Not serious	Not serious	Not serious	None	⨁⨁◯◯ LOW	38/697 (5.5%)		52/675 (7.7%)	0.78 (0.55 to 1.11)
ARDS
1315 (6 RCTs)	Very serious^g		Not serious	Not serious	Not serious	None	⨁⨁◯◯ LOW	28/672 (4.2%)		54/643 (8.4%)	0.50 (0.32 to 0.78)
Duration of ventilation
1510 (10 RCTs)	Very serious^h		Seriousⁱ	Not serious	Not serious	Publication bias strongly suspected^j	⨁◯◯◯ VERY LOW	771		739	− 0.03 (− 0.27 to 0.21)
Hospital stay
1245 (5 RCTs)	Serious^k		Not serious	Not serious	Not serious	None	⨁⨁⨁◯ MODERATE	629		616	− 0.02 (− 0.69 to 0.66)

PEEP, positive end-expiratory pressure; N, number of patients; RR: risk ratio; MD, mean difference; SMD: standardized mean difference; CI, confidence interval; RCT, randomized controlled trials; PaO₂/FiO₂, arterial partial pressure of oxygen to fraction of inspired oxygen ratio; ARDS, acute respiratory distress syndrome. “Other considerations” include publication bias, large effect, plausible confounding, and dose response gradient
^aFive studies were at high risk of bias; four studies were at moderate risk of bias
^bSix studies were at moderate risk of bias; two studies were at high risk of bias
^cModerate heterogeneity (I² = 59%) present across studies
^dThree studies were at high risk of bias
^eThree studies were at high risk of bias; two studies were at moderate risk of bias
^fFive studies were at high risk of bias; two studies were at moderate risk of bias
^gFour studies were at high risk of bias; two studies were at moderate risk of bias
^hSix studies were at high risk of bias; four studies were at moderate risk of bias
ⁱModerate heterogeneity (I² = 65%) present across studies
^jVisual inspection of the funnel plot suggests that publication bias cannot be excluded
^kThree studies were at moderate risk of bias; two studies were at high risk of bias

ISSN: 1364-8535