This article is accompanied by the following Invited Commentary:
Aveline C. Intra-operative high inspired oxygen during open abdominal surgery and postoperative pulmonary complications: From physiology to individualised strategies. Eur J Anaesthesiol 2019; 36:317–319.
The WHO and the United States Centers for Disease Control and Prevention (CDC) recently recommended administration of 80% inspired oxygen fraction (FiO2) during and immediately after surgeries performed with general anaesthesia and endotracheal intubation.1,2 The recommendation was based on some data suggesting that intra-operative high FiO2 reduces the incidence of surgical site infections. Many anaesthesia clinicians are reluctant to provide higher concentrations of oxygen because of potential adverse effects on the lungs. Most randomised trials utilising intra-operative high FiO2 evaluated its effect on various wound-healing related complications, and many do not even report pulmonary outcomes.3–7 Neither the WHO nor the CDC recommendations fully address these potentially harmful effects. Furthermore, a recent large registry analysis of adults having noncardiothoracic surgery reported a dose-dependent association between increased intra-operative FiO2 and the incidence of postoperative pulmonary complications (PPC), as well as 30-day mortality.8
Respiratory complications after anaesthesia are common and have reported incidences ranging from 4% to about 40%, depending on the definition and the baseline risk.9,10 Although oxygen administration is a key component in the treatment of hypoxaemic patients, its effects on the lungs of normoxaemic patients remains controversial. Even short exposure to high oxygen concentration promotes absorption atelectasis.11 Reduced mucociliary function and cellular damage from reactive oxygen species increase the risk of inflammatory and infectious complications.12 In critical care settings excessive exposure to oxygen is associated with worsening acute respiratory distress syndrome and ventilator-associated pneumonia.13–15
We therefore evaluated the effect of intra-operative high FiO2 on postoperative respiratory outcomes in patients who participated in a recent large trial. Specifically, we tested the primary hypothesis that 80% inspired intra-operative oxygen decreases the postoperative ratio of arterial saturation to fraction of inspired oxygen (SpO2/FiO2) compared with 30% inspired oxygen. And secondarily, we tested the hypothesis that 80% FiO2 increases a composite of PPC.
Study setup and patients
Adult patients scheduled for surgery in a surgical suite dedicated to colorectal surgeries in Cleveland Clinic's Main Campus were enrolled into an alternating cohort quality improvement project, in which the target FiO2 was altered every 2 weeks between 30 and 80%, for a total of 39 months from 2013 to 2016. This resulted in a large cohort of more than 8000 patients assigned to intra-operative administration of either 30 or 80% oxygen that were well balanced on more than 50 potential confounders, as described in the main study analysis of wound-related complications.16 The rest of the intra-operative ventilation management, as well as the postoperative FiO2, were not controlled and were left to the discretion of caregivers. In addition, clinicians were instructed to provide sufficient intra-operative oxygen to maintain patients’ saturation at or above 95%, and were allowed to use FiO2 as high as 100% during induction of and emergence from anaesthesia. Cases lasting less than 2 h, cases with missing intra-operative FiO2 data, re-operations during same hospitalisation, and cases with missing FiO2 or SpO2 values in the postanaesthesia care unit were excluded from the current analysis.
SpO2 data were collected from the patients’ electronic medical records, and FiO2 in the postanaesthesia care unit was calculated from the reported oxygen enrichment device used and the oxygen flow, based on a conversion table (Supplemental Digital Content 1, https://links.lww.com/EJA/A193). We assumed FiO2 remained unchanged until the next recording. Demographic, surgical and anaesthetic data as well as postoperative complications were also extracted from the electronic medical records.
For the primary outcome, we used the lowest SpO2/FiO2 ratio recorded in the postanaesthesia care unit for each patient. The secondary outcome was a composite of respiratory complications; pulmonary infection; pneumonia; respiratory failure or distress; bronchospasm; tracheitis or bronchitis; pulmonary collapse or atelectasis; emphysema; pulmonary oedema; pneumothorax or air leak; pleural effusion; acute respiratory distress syndrome; exacerbation of chronic obstructive pulmonary disease or asthma; continuous invasive mechanical ventilation; transfusion-related acute lung injury and pulmonary embolism or infarction. These are listed and defined in Supplemental Digital Content 2, https://links.lww.com/EJA/A193, by diagnosis codes or events documented in the medical record from admission to the postanaesthesia care unit until hospital discharge.
This study was approved by the Cleveland Clinic Institutional Review Board (IRB #12-891, Cleveland, Ohio, United States, executive director Daniel Beyer) on 29 August 2012 and was registered in clinicaltrials.gov (NCT01777568). This posthoc analysis was also approved by the Cleveland Clinic institutional review board (IRB #17-824, Cleveland, Ohio, United States, executive director Bridget Howard) on 26 June 2017.
We descriptively compared the 30 and 80% oxygen groups on demographic, baseline and procedural variables using standard descriptive statistics and absolute standardised difference (ASD). The ASD was calculated as absolute difference in means or proportions divided by the pooled SD.
To assess the effect of supplemental oxygen, we compared the lowest SpO2/FiO2 ratio between two treatment groups using Wilcoxon rank-sum test. Median difference with its confidence interval (CI) was reported from Hodges–Lehmann estimator. Secondarily, the relative risk (RR) of PPC comparing 30 to 80% oxygen groups was assessed using a generalised linear model with log link. The significance criterion was 0.05 for both primary and secondary analyses.
We expected that each group would have 2500 eligible patients who would meet all inclusion/exclusion criteria of this study. Based on literature and a preliminary query of our peri-operative database, we assumed that the mean of lowest SpO2/FiO2 was 330 with a SD of 200. The estimated power was thus 0.94 for a difference of 20 U in the SpO2/FiO2 ratio and 0.76 for a difference of 15 U, with an alpha level of 0.05.
A total of 8097 colorectal surgeries performed at Cleveland Clinic Main Campus from 28 January 2013 to 11 March 2016 were retrieved. After excluding 3041 nonqualifying cases, 5056 surgeries were analysed, with 2486 surgeries allocated to 30% oxygen and 2570 allocated to 80% oxygen. Patients’ flow and reasons for exclusion are described in Fig. 1.
Patient characteristics by group are summarised in Table 1, and their surgical and anaesthetic information are presented in Table 2. The 30 and 80% oxygen groups were well balanced on all demographic, baseline and procedural variables (ASD < 0.10). Therefore, adjustment for potential confounders was unnecessary. The median intra-operative time-weighted average inspired oxygen concentration was 43% (IQR 38 to 54%) in patients assigned to 30% or the lowest tolerated concentration, and 81% (IQR 78 to 82%) in those assigned to 80% oxygen. FiO2 in the postanaesthesia care unit was kept at a level of 27% by low oxygen flow through a nasal cannula in 77% of patients (median time-weighted-average FiO2 [IQR] 27% [27%, 27%]).
There was no significant difference in the lowest SpO2/FiO2 ratio in the postanaesthesia care unit, comparing the high and low FiO2 groups (Fig. 2). The estimated median difference was 0 (95% CI: 0, 0; P = 0.91, Table 3). The primary outcome did not change over the course of the study. Hyperoxia also had no effect on the composite outcome of PPC. The overall observed incidence of pulmonary complications was 16.3% (406/2486) in patients assigned to 30% inspired oxygen and 17.6% (451/2570) in those assigned to 80% oxygen. The estimated RR was 1.07 (95% CI: 0.95, 1.21), P = 0.25. The incidence of the composite and its components is shown in Table 3, including the RR (95% CI) for components with incidence of 1% or higher. In-hospital mortality rate was similar in both groups. Twelve patients (0.48%) in 30% oxygen group and 12 (0.47%) in 80% oxygen group died in hospital after surgery.
High intra-operative FiO2 did not reduce the average minimal SpO2/FiO2 ratio in the postoperative care unit. Furthermore, supplemental oxygen did not worsen the risk of pulmonary complications. Since the underlying trial included more than 5000 patients, we had 94% power to detect changes as small as 20 U in the SpO2/FiO2 ratio. It is thus unlikely that 80% intra-operative oxygen substantively worsens postoperative oxygenation in our study population of adults having colorectal surgery.
Our finding is consistent with the smaller PROXI trial which randomised 1400 patients to 30 or 80% oxygen intra-operatively and for 2 postoperative hours.17 There was no difference in the incidence of atelectasis, pneumonia, or respiratory failure in the two groups. A meta-analysis by Hovaguimian et al.18 reported no difference in the incidence of postoperative atelectasis and partial arterial oxygen pressure to FiO2 (PaO2/FiO2) ratio ratio in patients receiving high or normal intra-operative FiO2.
In contrast, a recent large registry-based analysis reported a dose–response association between intra-operative FiO2 and incidence of pulmonary complications.8 Of note, the overall incidence of pulmonary complications in that registry analysis was about 4%, compared with 16% in our patients. Complications were presumably more common in our patients because they were sicker. For example, 70% of our patients had American Society of Anesthesiology physical status scores at least 3, compared with less than 30% in the report by Staehr-rye et al.8 But a more serious question is why Staehr-rye et al. report an association between inspired oxygen concentration and pulmonary complications whereas we saw no such effect. Unadjusted confounding in the registry analysis seems the most likely explanation. For example, many clinicians tend to give supplemental oxygen to sicker patients, a practice that potentially introduces substantial bias into retrospective analyses.
Our study differed in that we prospectively allocated patients to the high and low oxygen administration groups. This method proved effective in creating remarkably comparable groups across more than 50 baseline and intra-operative potential confounding factors. In addition, the outcomes in our study were collected from the nursing vital signs reports as well as from diagnosis codes, which were probably both indifferent, if not completely blinded, to the study allocation. We therefore were able to effectively minimise selection and measurement biases, along with confounding, that are always dangers in retrospective analyses. Since we were not powered for subgroup analyses, our results might not generalise well to specific patient populations, who may need specific ventilatory strategies. But considering the large sample size and the remarkable balance on all tested potential confounders, the risk for subgroup imbalance is low.
The underlying trial used a novel alternating intervention design rather than conventional individual-patient randomisation. The advantage of this approach is that investigators can rapidly and inexpensively enrol large numbers of patients (our alternating intervention trial included nearly as many patients as all 20 previous randomised trials). We note though that treatment allocation was not randomised on an individual basis; in fact, even the 2-week exposure blocks were not randomised. Instead, exposure periods simply alternated. In this respect, the approach resembles cluster randomisation, with clusters distributed in time rather than in space.
A limitation of the underlying trial is that we did not routinely obtain arterial blood samples; instead we used the SpO2/FiO2 ratio as a surrogate measure of oxygenation for the more common PaO2/FiO2 ratio. However, the SpO2/FiO2 ratio is considered a reliable diagnostic tool for acute respiratory distress syndrome and can replace the PaO2/FiO2 ratio in the respiratory part of the Sequential Organ Failure Assessment score.19,20 SpO2 has also been shown to be an effective assessment of early respiratory insufficiency, either as a component of a SpO2/FiO2 ratio or as a variable in the Ellis or Rice equations to calculate an actual PaO2/FiO2 ratio.21,22 An advantage of the SpO2/FiO2 ratio is that it is a continuous measure, rather than being restricted to times at which arterial blood might be sampled. Another consequence of the design is that aside from FiO2, we did not control for any other ventilatory parameter, such as the use or level of positive end-expiratory pressure, or recruitment manoeuvres. It is possible that anaesthesia providers used these tools differentially between the two groups, to improve oxygenation in the low FiO2 group.
We assessed oxygenation only during the initial hours in the postanaesthesia care unit; delayed effects of hyperoxia on oxygenation would thus be overlooked in our analysis. However, complications were considered throughout hospitalisation. The two most common components of the composite secondary outcome were chronic obstructive pulmonary disease or asthma exacerbation and bronchospasm, each defined either by diagnosis codes or by treatment. These indirect diagnoses were surely less accurate than specific pulmonary function measures such as expired volume in one second. However, it seems highly unlikely that our postoperative outcome measures were biased by exposure.
The WHO and Center for Disease Control recommendations to routinely expose patients to intra-operative and postoperative high FiO2 have been questioned.23–26 Much of the concern relates to unconvincing evidence that supplemental oxygen reduces surgical site infection. But others have argued that the recommendations ignored potential deleterious pulmonary consequences. Although we only tested the intra-operative part of the proposed intervention, our results suggest that intra-operative administration of 80% inspired oxygen does not worsen postoperative oxygenation or promote pulmonary complications.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: supported by internal funding only. BC received a fellowship grant from the American Physicians Fellowship for Medicine in Israel. ER received a grant from Instituto Salud Carlos III (BA17/00032).
Conflicts of interest: none.
1. Allegranzi B, Bischoff P, de Jonge S, et al. New WHO recommendations on preoperative measures for surgical site infection prevention: an evidence-based global perspective. Lancet Infect Dis
2. Berríos-Torres SI, Umscheid CA, Bratzler DW, et al. Centers for Disease Control and Prevention guideline for the prevention of surgical site infection, 2017. JAMA Surg
3. Kurz A, Fleischmann E, Sessler DI, et al. Effects of supplemental oxygen and dexamethasone on surgical site infection: a factorial randomized trial. Br J Anaesth
4. Bickel A, Gurevits M, Vamos R, et al. Perioperative hyperoxygenation and wound site infection following surgery for acute appendicitis: a randomized, prospective, controlled trial. Arch Surg
5. Greif R, Akcą O, Horn E-P, et al. Supplemental perioperative oxygen to reduce the incidence of surgical-wound infection. N Engl J Med
6. Belda FJ, Aguilera L, García de la Asunción J, et al. Supplemental perioperative oxygen and the risk of surgical wound infection a randomized controlled trial. JAMA
7. Pryor KO, Iii TJF, Lien CA, et al. Surgical site infection and the routine use of perioperative hyperoxia in a general surgical population – a randomized controlled trial. J Am Med Assoc
8. Staehr-Rye AK, Meyhoff CS, Scheffenbichler FT, et al. High intraoperative inspiratory oxygen fraction and risk of major respiratory complications. Br J Anaesth
9. McAlister FA, Bertsch K, Man J, et al. Incidence of and risk factors for pulmonary complications after nonthoracic surgery. Am J Respir Crit Care Med
10. Hemmes SNT, Gama de Abreu M, Pelosi P, et al. PROVE Network Investigators for the Clinical Trial Network of the European Society of Anaesthesiology. High versus low positive end-expiratory pressure during general anaesthesia for open abdominal surgery (PROVHILO trial): a multicentre randomised controlled trial. Lancet
11. Edmark L, Kostova-Aherdan K, Enlund M, et al. Optimal oxygen concentration during induction of general anesthesia. Anesthesiology
12. Patel VS, Sitapara RA, Gore A, et al. High Mobility Group Box-1 mediates hyperoxia-induced impairment of Pseudomonas aeruginosa clearance and inflammatory lung injury in mice. Am J Respir Cell Mol Biol
13. Six S, Jaffal K, Ledoux G, et al. Hyperoxemia as a risk factor for ventilator-associated pneumonia. Crit Care
14. Kallet RH, Matthay MA. Hyperoxic acute lung injury. Respir Care
15. Sinclair SE, Altemeier WA, Matute-Bello G, et al. Augmented lung injury due to interaction between hyperoxia and mechanical ventilation. Crit Care Med
16. Kurz A, Kopyeva T, Suliman I, et al. Supplemental oxygen and surgical-site infections: an alternating intervention controlled trial. Br J Anaesth
17. Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA
18. Hovaguimian F, Lysakowski C, Elia N, et al. Effect of intraoperative high inspired oxygen fraction on surgical site infection, postoperative nausea and vomiting, and pulmonary function: systematic review and meta-analysis of randomized controlled trials. Anesthesiology
19. Rice TW, Wheeler AP, Bernard GR, et al. Comparison of the SpO2
Ratio and the P
ratio in patients with acute lung injury or ARDS. Chest
20. Pandharipande PP, Shintani AK, Hagerman HE, et al. Derivation and validation of SpO2
ratio to impute for P
ratio in the respiratory component of the Sequential Organ Failure Assessment score. Crit Care Med
21. Sanz F, Dean N, Dickerson J, et al. Accuracy of P
calculated from SpO2
for severity assessment in ED patients with pneumonia. Respirology
22. Festic E, Bansal V, Kor DJ, et al. US Critical Illness and Injury Trials Group: Lung Injury Prevention Study Investigators (USCIITG–LIPS). SpO2
ratio on hospital admission is an indicator of early acute respiratory distress syndrome development among patients at risk. J Intensive Care Med
23. Volk T, Peters J, Sessler DI. The WHO recommendation for 80% perioperative oxygen is poorly justified. Anaesthesist
24. Myles PS, Kurz A. Supplemental oxygen and surgical site infection: getting to the truth. Br J Anaesth
25. Hedenstierna G, Perchiazzi G, Meyhoff CS, et al. Who can make sense of the WHO guidelines to prevent surgical site infection? Anesthesiology
26. Akca O, Ball L, Belda FJ, et al. WHO needs high FiO2
? Turk J Anaesthesiol Reanim
* Barak Cohen and Kurt Ruetzler contributed equally to the article.