“High intraoperative oxygen administration has no clinically significant benefit on the incidence of postoperative nausea and vomiting.”
Oxygen delivery is a universal component of hospital care, although its ideal titration is still surprisingly elusive. Hypoxemia has well known deleterious effects on cellular and organ function and overall health. The main purpose of oxygen administration during perioperative care is to prevent or treat hypoxemia. Oxygen administration resulting in blood oxygen content in excess of physiologic needs (i.e., hyperoxemia) has been proposed to reduce the incidence of various complications, including postoperative nausea and vomiting. The first report supporting intraoperative hyperoxemia for postoperative nausea and vomiting prevention was published more than 20 yr ago.1 Since then, the same authors and others have been investigating this possible use of intraoperative hyperoxemia in various patient populations with inconsistent results.2,3 More recently, concerns of harm related to hyperoxemia have been suggested.4
In this issue of Anesthesiology, Markwei et al.3 examine the effect of an intraoperative inspired oxygen fraction (Fio2) of 80% versus 30% to reduce the incidence of postoperative nausea and vomiting in a secondary analysis of a previously published randomized controlled trial.5 The authors included 4,001 patients having 5,057 colorectal surgeries at a single hospital in their secondary analysis. Each surgical encounter was randomly assigned in a cluster-crossover fashion to intraoperative Fio2 of 80% (n = 2,554) or 30% (n = 2,503) for the duration of their procedure. Operating rooms at the single site alternated between 80% and 30% Fio2 at 2-week intervals for 39 months. The protocol allowed clinicians to increase inspired oxygen concentrations to maintain an oxygen saturation measured by pulse oximetry of at least 95%. Postoperative nausea and vomiting was assessed by postanesthesia care unit nurses every 15 min for the first 2 h postoperatively, and ward nurses assessed nausea and/or vomiting every 4 h throughout hospitalization. Nearly every patient (97% in the Fio2 80% group and 98% in the Fio2 30% group) received at least one intraoperative antiemetic.
The authors found no difference in the incidence of postoperative nausea and vomiting between the Fio2 80% and Fio2 30% groups (33% vs. 33%, respectively; relative risk, 1.04; 95% CI, 0.96 to 1.12). Based on these findings, we can conclude that any clinical differences in postoperative nausea and vomiting between the Fio2 groups are insignificant. There were also no significant differences in the secondary outcomes between Fio2 groups, including the median number of antiemetic doses administered within the first 24 h (0 vs. 0; mean difference, 0.003; 95% CI, –0.04 to 0.05) and postoperative nausea and vomiting severity during the first 2 h postoperatively (88% “none to minimal” in both groups). It is important to note that the parent trial had an alternating intervention design in which the Fio2 provided to all patients receiving care in the designated operating-room suites was randomized to either 30% (as tolerated) or 80% for periods of 2 weeks.5 Oxygen delivery was not randomized on a per-patient or even per-period basis, and therefore, it may not account for differences in patient-level factors at randomization. The population sampled may not be representative of broader populations, and the open label with lack of blinding design might also introduce some possible bias. We do not know the number of antiemetic agents administered, specific antiemetic agents received, or the timing of antiemetic agents in relation to surgical start and end times. These details could be important because the effects of antiemetic prophylaxis for postoperative nausea and vomiting are additive.6 Further, the timing of antiemetic administration is an essential component of efficacy and varies significantly between agents of different classes.6 It is also true that differences in these care details would likely be evenly distributed in the approximately 4,000 patients of the original cluster-crossover trial.
The authors should be commended for performing an additional rigorous systematic review and meta-analysis following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses7 recommendations. This additional analysis included 10 randomized controlled trials representing 6,749 surgical encounters, with the current secondary analysis contributing 75% of the surgeries included in the meta-analysis. This meta-analysis by Markwei et al.,3 the largest published meta-analysis of intraoperative oxygen administration strategies for postoperative nausea and vomiting prevention, showed some publication bias as demonstrated by asymmetry of the funnel plot in appendix 2. The overall risk ratio was weighted toward the three largest trials, including the current study,3 and those by McKeen et al.8 and Turan et al.9 The primary outcome, postoperative nausea and vomiting, was not statistically different between the high- and low-Fio2 groups (relative risk, 0.97; 95% CI, 0.86 to 1.08). The authors also did not find evidence that surgery type influenced the effect of intraoperative Fio2 on postoperative nausea and vomiting. Neither abdominal surgery (relative risk, 0.87; 95% CI, 0.72 to 1.05) nor nonabdominal surgery (relative risk, 1.06; 95% CI, 0.87 to 1.28) showed differing rates of postoperative nausea and vomiting between Fio2 groups.
The results from this meta-analysis are clear: high intraoperative oxygen administration has no clinically significant benefit on the incidence of postoperative nausea and vomiting. Intraoperative hyperoxemia has failed to demonstrate any benefit for other disease processes, including the prevention of surgical site infection, and has detrimental impact on others, including atelectasis and myocardial injury.5,10 The presumed benefits of intraoperative hyperoxemia for postoperative nausea and vomiting prevention were contentious and lacked a solid mechanistic foundation other than theoretically preventing subclinical gastrointestinal tract hypoxia. The potential risks of hyperoxemia have been increasingly understood, but its clinical impact is also yet to be confirmed. We do know 2 h of exposure to 80% Fio2 increases reactive oxygen species generation, initiating an inflammatory cascade that impairs wound healing.11 Higher intraoperative Fio2 increases reactive oxygen species formation and oxidative stress, increasing the risk of end-organ injury.12 Hyperoxia inhibits hypoxic pulmonary vasoconstriction, thus worsening ventilation or perfusion mismatch while promoting atelectasis and systemic vasoconstriction.13 Each 10% increase in median intraoperative Fio2 is associated with an increased risk of myocardial injury during noncardiac surgery.14 The potential increased morbidity while administering intraoperative hyperoxemia seems to clearly outweigh the benefits. Frankly, there are no benefits of oxygen administration other than preventing hypoxemia.
One-third of surgical patients continue to experience postoperative nausea or vomiting even with prophylactic antiemetic treatment, so it is essential to remain open-minded to emerging results from rigorously conducted studies of novel treatments. However, the results and meta-analysis reported by Markwei et al.3 do not support the use of intraoperative hyperoxemia to reduce postoperative nausea and vomiting in any surgical population. Given the potential harm from hyperoxemia, evidence to date suggests that perioperative clinicians should avoid unnecessary use of intraoperative oxygen and target normoxemia whenever possible. As a reference, normoxemia has been recently defined as an oxygen saturation of 90 to 96% in critically ill trauma patients,15 and this may be considered a reasonable target for most surgical patients. Last but not least, these findings1,3 suggest that a persistent, rigorous, and multifaceted research approach should be encouraged to confirm initial research results and elucidate their mechanistic foundation. As we know, many well intentioned observations later become disputed or surface as ineffective or harmful. The article by Markwei et al.3 strongly suggests that high oxygen administration during surgery does not reduce the risk of postoperative nausea and vomiting. For now, intraoperative oxygen administration should be titrated to prevent or treat hypoxemia, nothing else.
Supported by National Institutes of Health (Bethesda, Maryland) awards T32GM135169 (to Dr. Douin) and UH3HL140177 (to Dr. Fernandez-Bustamante).
Dr. Fernandez-Bustamante reports additional research funding unrelated to this work from the U.S. Department of Defense (Washington, D.C.), Merck (Rahway, New Jersey; postoperative pulmonary function after neuromuscular blockade reversal), and the University of Colorado Institute for Healthcare Quality, Safety and Efficiency (Aurora, Colorado). Dr. Douin is not supported by, nor maintains any financial interest in, any commercial activity that may be associated with the topic of this article.