Positive end-expiratory pressure for one-lung ventilation : a meta-analysis

Abstract

Introduction: Positive end-expiratory pressure (PEEP) is commonly used as a primary intervention for hypoxemia during one-lung ventilation (OLV). However, the effect of PEEP on oxygenation during OLV is controversial. Therefore a systematic review and meta-analysis of randomized controlled trials (RCTs) was performed to investigate the effect of PEEP during OLV. Methods: Databases including CENTRAL, CINHAL, EMBASE, MEDLINE, SCOPUS, Web of Science, KoreaMed were searched systematically, and finally 7 RCTs with 525 participants were included. The primary outcome was the arterial oxygen partial pressure to fraction of inspired oxygen ratio (P/F ratio) and the secondary outcomes were the lung compliance, dead space to tidal volume ratio (VD/VT), shunt fraction, arterial to end-tidal CO2 partial pressure gradient (Pa-ETCO2 gradient), number of desaturation events during OLV, postoperative radiologic evaluation of lungs and pulmonary function test (PFT) findings. For the primary outcome, a standardized mean difference (SMD) with 95% confidence interval (CI) was calculated and for the secondary outcomes, mean differences (MDs) with 95% CI were calculated. Relevant subgroup analyses were performed to evaluate potential sources of heterogeneity. Results: The lung compliance improved within and over 30 minutes after onset of OLV but increments were not statistically significant [SMD (95% CI) = 0.50 (-0.10, 1.10), P = 0.10, I2 = 80% within 30 minutes after OLV, SMD (95% CI) = 0.51 (-0.04, 1.06), P = 0.07, I2 = 71% over 30 minutes after onset of OLV respectively]. The shunt fraction was significantly lower in the PEEP2 group compared with the zero end-expiratory pressure (ZEEP) group over 30 minutes after onset of OLV [MD (95% CI) = -5.72 (-7.57, -3.88), P < 0.001]. VD/VT over 30 minutes after onset of OLV was lower in the PEEP group [SMD (95% CI) = -0.67 (-1.18, -0.16), P = 0.01], tendency of decrease in the PEEP group was observed within 30 minutes after onset of OLV but decrement was not statistically insignificant [SMD (95% CI) = -1.14 (-2.38, 0.10), P = 0.07]. Results on Pa-ETCO2 gradient were heterogeneous, and there were no statistically significant differences between both groups. Within 30 minutes after onset of OLV, the P/F ratio was lower in the PEEP group though it was not significant [MD (95% CI) = -26.06 (-65.64, 13.53), P = 0.20, I2 = 86%]. Over 30 minutes after onset of OLV, no difference on oxygenation was found between both groups [MD (95% CI) = 11.73 (-19.53, 43.00), P = 0.46, I2 = 73%]. Number of desaturation event which occurred during OLV was not different between two groups [OR (95% CI) = -1.01 (0.43, 2.38), P = 0.98, I2 = 0%], but the pulmonary function test findings were higher in the PEEP group at 72 hours postoperatively. Postoperative atelectasis was not different between both groups. Conclusion: This meta-analysis suggested that applying PEEP improved shunt fraction and VD/VT over 30 minutes after onset of OLV, but improvement of oxygenation was not guaranteed by PEEP. As definite advantages on oxygenation were not found, cautious approach to hypoxemia during OLV should be taken when applying PEEP to dependent lung.