Intraoperative mechanical ventilation is a major component of general anesthesia. The extent to which various intraoperative tidal volumes and positive end-expiratory pressures (PEEP) effect on postoperative hypoxia and lung injury remains unclear. We hypothesized that adults having orthopedic surgery, ventilation using different tidal volumes and PEEP levels affect the oxygenation within first hour in the postoperative care unit.
We conducted a two-by-two factorial crossover cluster trial at the Cleveland Clinic Main Campus. We enrolled patients having orthopedic surgery with general anesthesia who were assigned to factorial clusters with tidal volumes of 6 or 10 ml/kg of predicted body weight and to PEEP of 5 or 8 cm H2O in 1-week clusters. The primary outcome was the effect of tidal volume or PEEP on time-weighted average peripheral oxygen saturation measured by pulse oximetry divided by the fraction of inspired oxygen (Spo2/Fio2 ratio) during the initial postoperative hour.
We enrolled 2,860 patients who had general anesthesia for orthopedic surgery from September 2018 through October 2020. The interaction between tidal volume and PEEP was not significant (P = 0.565). The mean ± SD time-weighted average of Spo2/Fio2 ratio was 353 ± 47 and not different in patients assigned to high and low tidal volume (estimated effect, 3.5%; 97.5% CI, –0.4% to 7.3%; P = 0.042), for those assigned to high and low PEEP (estimated effect, –0.2%; 97.5% CI, –4.0% to 3.6%; P = 0.906). We did not find significant difference in ward Spo2/Fio2 ratio, pulmonary complications, and duration of hospitalization among patients assigned to various tidal volumes and PEEP levels.
Among adults having major orthopedic surgery, postoperative oxygenation is similar, with tidal volumes between 6 and 10 ml/kg and PEEP between 5 and 8 cm H2O. Our results suggest that any combination of tidal volumes between 6 and 10 ml/kg and PEEP between 5 versus 8 ml cm H2O can be used safely for orthopedic surgery.
Low tidal volume and high positive end-expiratory pressures (PEEP) are preferable in critical care patients, but it remains unclear whether they are beneficial in surgical patients.
A total of 2,860 orthopedic surgical patients having general anesthesia were assigned in a 2 x 2 factorial cluster trial to 6 versus 10 ml/kg tidal volume and to 5 versus 8 cm H2O PEEP.
There was no interaction between VT and PEEP. The primary outcome, the Spo2/Fio2 ratio, was similar in each tidal volume and PEEP group. Secondary outcomes including postoperative oxygenation, duration of hospitalization, and composite pulmonary complications also did not differ significantly.
Tidal volumes between 6 and 10 ml/kg and PEEP between 5 and 8 cm H2O are similar with respect to pulmonary outcomes.
Annually, 313 million surgical procedures are performed worldwide,1 and many experience potentially preventable postoperative complications.2 Among the most common are postoperative hypoxemia and pulmonary complications.3,4 Sun et al.,5 in a retrospective analysis, reported that hypoxemia was common and prolonged in patients recovering from major noncardiac surgery, with a fifth having at least 10 min/h with oxygen saturation measured by pulse oximetry (Spo2) less than 90%. Postoperative hypoxemia can be caused by atelectasis,6 ventilator-induced lung injury, ventilation/perfusion mismatch,6 and pulmonary edema.4 Hypoxemia, even without other respiratory complications, is associated with prolonged hospitalization, intensive care unit admissions, mortality, and increased cost of care.
Intraoperative mechanical ventilation is a major component of general anesthesia. Two key ventilator settings are tidal volume and positive end-expiratory pressure (PEEP). Traditionally, high tidal volumes (10 ml/kg or greater) were used because they reduce atelectasis and improve oxygenation. However, high tidal volumes increase concentrations of proinflammatory mediators, promote pulmonary edema, and over-distend alveoli—thus promoting lung injury and hypoxia.7 In contrast, restricted tidal volumes reduce inflammation, improve breathing mechanics, and limit over-distention injury. However, low tidal volumes also promote atelectasis, which is an important cause of postoperative lung injury, and promotes pneumonia and hypoxia. It remains unclear which of these many competing effects dominates. Consequently, low tidal volume has only inconsistently been adopted for operating room use.3
There is similar ongoing debate about the optimal level of intraoperative PEEP. High PEEP reduces atelectasis and improves arterial oxygenation and respiratory system compliance, but also promotes alveolar over-distention hypotension.8 Low PEEP decreases barotrauma but may not prevent atelectasis. Low tidal volumes combined with PEEP help maintain oxygenation in patients with acute respiratory distress syndrome and acute lung injury who are ventilated for days in critical care units. However, it remains unclear whether comparable benefit results when restricted tidal volumes and high PEEP are applied for just a few intraoperative hours.9 Also unknown is how intraoperative tidal volume and PEEP interact—that is, which combination of high and low tidal volume and high and low PEEP is preferable.
The extent to which various intraoperative tidal volumes and PEEP levels effect on postoperative hypoxia and lung injury therefore remains unclear. We thus conducted a robust 2 by 2 factorial crossover cluster trial to determine the effects of tidal volumes of 6 versus 10 ml/kg of predicted body weight and PEEP of 5 versus 8 cm H2O. Our primary outcome was oxygenation within the first hour in the postoperative care unit, defined by the peripheral Spo2 divided by the fraction of inspired oxygen (Fio2) ratio, a surrogate measure of oxygenation. Our secondary outcomes were (1) time-weighted average Spo2/Fio2 ratio on surgical wards, (2) postoperative duration of hospitalization, and (3) a composite of postoperative pulmonary complications.
Materials and Methods
This is a single-center, single-blinded, multiple crossover factorial alternating cluster trial. The protocol was approved by the Cleveland Clinic (Cleveland, Ohio) Institutional Review Board, and written informed consent was waived; however, all patients were given written information about the study well before surgery and had the opportunity to opt out of the trial. There were no substantive changes to the protocol after initiation of patient enrollment. The full protocol and statistical analysis plan are available in Supplemental Digital Content 1 (http://links.lww.com/ALN/C883). The trial was exclusively funded by departmental resources, and none of the authors has a personal financial interest in this research.
Subject Selection
Patients were enrolled at the Cleveland Clinic Main Campus between September 3, 2018, and October 24, 2020. The study was restricted to a physically distinct suite of five operating rooms that are primarily used by for orthopedic surgery and normally staffed by a small group of anesthesiologists.
All patients in the designated operating rooms were nominally included in the trial. However, the protocol specified that good clinical judgment should always prevail. Clinicians thus modified tidal volume and PEEP when they deemed it necessary. Similarly, they were able to a priori exclude particular patients whom they deemed clinically unsuitable for the trial.
Randomization and Blinding
Patients having general anesthesia in the designated operating rooms were assigned to factorial clusters with tidal volumes of 6 or 10 ml/kg of predicted body weight and to PEEP of 5 or 8 cm H2O in 1-week clusters. Thus, in a given 4-week period, all four combinations of tidal volume and PEEP were each used for 1 week. Assignments sequentially alternated among the four combinations throughout enrollment. Anesthesiologists were not blinded to treatments, but patients and outcome assessors were blinded.
Study Procedures
Inspired oxygen concentration was normally 50% during surgery, but the concentration was increased as necessary to maintain oxygen saturation 95% or greater as determined by pulse oximetry. The respiratory rate was adjusted to maintain an end-tidal carbon dioxide partial pressure 35 to 45 mmHg, with a default inspired to expired ratio of 1:2. Clinicians were asked to perform a recruitment maneuver after induction of anesthesia at a Fio2 of 50% and shortly before extubation. Typically, the recruitment maneuver consisted of maintaining an airway pressure of 40 to 45 cm H2O for 40 s.10 During the postoperative period, supplemental oxygen was increased as necessary to maintain oxygen saturation 92% or greater. We use the phase 1 discharge scoring tool to assess patients for postanesthesia care unit (PACU) discharge. It includes 10 items scored from 0 to 2 about the level of consciousness, physical activity, blood pressure, heart rate, respirations, oxygen saturation status, pain, postoperative nauseas and vomits, temperature, and bleeding. The maximum score is 20, and patients need to score 18 or more to be discharged from the PACU. Specifically, the oxygen saturation status scores 2 if the patient has an oxygen saturation more than 92% on room air or on supplemental oxygen with IV patient-controlled analgesia (PCA); 1 if saturation is greater than 92% on supplemental oxygen not involving IV PCA; and 0 if saturation less than 92% on supplemental oxygen‚ but in this case‚ if there is not a treatable cause, the patient is transfered to intensive care unit.
There was no other restriction on anesthetic management. Clinicians were thus free to use any combination of drugs they wished for general anesthesia and patients, although patients who had neuraxial anesthesia were excluded from analysis. There was no restriction on peripheral nerve blocks or postoperative analgesic management.
Monitoring, Measurements, and Data Collection
All data were obtained from the Cleveland Clinic Perioperative Health Documentation System and the Cleveland Clinic Electronic Medical Records. Demographic and morphometric characteristics were recorded, including age, sex, race, weight, height, and body mass index. We also recorded factors that might increase risk of pulmonary complications including American Society of Anesthesiologists (ASA; Schaumburg, Illinois) Physical Status, preoperative comorbidities, and smoking history.
Types of surgery were characterized from International Classification of Diseases, Ninth Revision, codes using Clinical Classifications Software (Agency for Healthcare Research Quality‚ USA).
Intraoperative data and routine anesthetic variables recorded in electronic medical records include use of regional anesthesia, patient position, blood pressure, heart rate, Spo2, expired carbon dioxide partial pressure, anesthetic agent, tidal volume, PEEP, ventilation frequency, minute volume, airway pressures, inspired oxygen fraction, transfused blood products, IV fluid types and volumes given during surgery, vasoactive medication needs, and duration of surgery.
Spo2 was monitored continuously in the PACU by pulse oximetry. Fio2 was estimated from the type of device and the oxygen flow rate as described in Supplemental Digital table 1 (http://links.lww.com/ALN/C884). Both were electronically recorded at approximately every 15 min. Unreasonable values such as Spo2 less than 10%, Fio2 less than 21, and Fio2 greater than 100% were excluded.
Outcomes
Primary Outcome
The primary outcome was the time-weighted average Spo2/Fio2 ratio during the initial postoperative hour. We first calculated the Spo2/Fio2 ratio at each measurement time point during the initial postoperative hour, and then averaged all Spo2/Fio2 ratios weighted by measurement interval. Diagnosis of acute lung injury has traditionally been based on clinical findings and the Pao2/Fio2 ratio.11,12 For example, a Pao2/Fio2 ratio 300 mmHg or less characterizes acute lung injury, and a ratio 200 or less is consistent with acute respiratory distress syndrome.13 Few patients having orthopedic surgery require arterial cannulation. We thus substituted arterial oxygen saturation/Fio2, which provides good sensitivity and specificity for diagnosing lung injury. For example, Rice et al.14 found that Spo2/Fio2 ratio correlates well with a simultaneously obtained Pao2/Fio2 ratio in patients with acute lung injury and acute respiratory distress syndrome. Furthermore, the Spo2/Fio2 ratio correlates well with the Pao2/Fio2 ratio and predicts respiratory failure in critical care patients,14 pediatric patients,15,16 and emergency department patients.17
Secondary Outcomes
We defined three a priori secondary outcomes: (1) time-weighted average Spo2/Fio2 ratio on surgical wards, normally recorded at 4-h intervals until discharge; (2) postoperative duration of hospitalization; and (3) a composite of postoperative pulmonary complications defined as the presence of at least one of the International Classification of Diseases, Tenth Revision, codes listed in Supplemental Digital table 2 (http://links.lww.com/ALN/C885) that were not present at admission. Pulmonary complications included respiratory complications, respiratory failure and distress, reintubations, pulmonary edema, and atelectasis.
Statistical Methods
Analysis was restricted to nonexcluded adults 18 yr or older who had an ASA Physical Status score I to III, were scheduled for elective orthopedic surgery lasting at least 2 h, and had general anesthesia with endotracheal intubation and mechanical ventilation.
Baseline Variables Balancing
Exposures were controlled but not randomly assigned. We therefore controlled for observed potential confounding variables (table 1) using the inverse probability treatment weighting method for multiple groups. Missing values for confounding variables were imputed using the chained equation, and the single imputation dataset was used in all analysis. We first fitted a multinomial logistic regression model with four group settings as the outcome variable, and all observed confounding variables in table 1 as the independent variables without interactions. We then estimated propensity scores, which are the probability of receiving treatment, for each patient from the model. After weighting each patient by the inverse of the corresponding propensity score, the success of the confounding control was assessed by pairwise comparison on potentially confounding baseline characteristics using the absolute standardized difference, defined as the absolute difference in means or proportions divided by the pooled SD. Observations in all primary and secondary analyses were weighted by the inverse of the relevant propensity score. Any confounding variables with an absolute standardized difference greater than 0.1 would be adjusted for in all analyses.
Primary Analysis
The effect of tidal volume, PEEP, and their interaction on the time-weighted average Spo2/Fio2 ratio in PACU was assessed in a linear mixed model with surgeries from the same patients as repeated measures, weighted using the inverse probability of propensity score and adjusting for unbalanced confounders as appropriate. If the interaction effect was not significant (P > 0.10), treatment effect estimates would be summarized using the mean difference comparing tidal volume of 10 versus 6 ml/kg and PEEP of 8 versus 5 cm H2O. If the interaction was significant, the effects of each intervention would be assessed within levels of the other intervention. With an overall alpha of 0.05 for the primary analysis, the significance criterion is 0.025 for each treatment effect without significant interaction (i.e., 0.05/2, Bonferroni correction). As a sensitivity analysis for the primary outcome, we included patients who were excluded based on decisions from surgeons or anesthesiologists, and then assessed the treatment effect using the same statistical method.
Post Hoc Analysis
We did post hoc analysis to explore whether the treatment effects of tidal volume and PEEP were modified by age, body mass index, smoking status, obstructive sleep apnea, or type of anesthesia.
Secondary Analysis
Secondary outcomes were restricted to inpatients and again included propensity score weighting to balance all baseline and surgery variables in table 1. Time-weighted average Spo2/Fio2 ratio while patients were on surgical wards was assessed using the same method as for the time-weighted average Spo2/Fio2 ratio in the PACU. We assumed that the Fio2 remain until updated in the medical record.
Length of hospital stay was log-transformed to meet the assumption of normality and then assessed in a linear mixed model. The ratio of geometric means with CI was reported as treatment effect. The odds ratio for pulmonary complications comparing different treatment groups was estimated from a logistic regression model after weighting by propensity score. With an overall alpha of 0.05 for all secondary analyses, the significance criterion is 0.0083 for each treatment effect without significant interaction (i.e., 0.05/6, Bonferroni correction).
Exploratory Analysis
Exploratory outcomes are presented descriptively without statistical analyses.
Sample Size and Power Estimation
Based on literature14,17,18 and a retrospective data analysis of Spo2/Fio2 levels, we assumed that the mean time-weighted average Spo2/Fio2 would be 330% with a SD of 100%. After accounting for three interim analyses and one final analysis, a maximum of 2,500 total patients (i.e., 625 for each of the four groups for assessing each main effect) were needed to provide 90% power at the 0.025 significance level for detecting main effects of 15% or more in Spo2/Fio2 ratio for the two tidal volumes and two PEEP levels. We used gamma error spending function for both type I and type II errosr. The parameter was –4 for alpha spending function to control the overall type I error at 0.025 and was –1 for beta spending function to preserve the overall power at 0.9 for multiple looks at the data for interim analyses. The significance level for the first, second, third, and final look was 0.0008, 0.003, 0.008, and 0.025, respectively.
Power Re-estimation
The observed SD of Spo2/Fio2 ratio was 49%, with our current sample size of 2,860, we have a power of more than 0.9 to detect the predefined mean difference of 15% at the significance level of 0.025 for the two tidal volumes and two PEEP levels.
All the tests were two-tailed hypotheses testing. All the analyses are conducted in SAS 9.4 (SAS Institute Inc.‚ USA) and R 4.0 (R foundation for statistical computing‚ Austria).
Results
A total of 3,481 orthopedic surgeries with general anesthesia received assigned treatments from September 3, 2018, through October 24, 2020, and met the inclusion and exclusion criteria (fig. 1). Enrollment ceased when the target sample size was obtained. The number of patients enrolled well exceeded our sample size estimate because it was unknown how many would be excluded for various reasons. A total of 621 surgeries were excluded by surgeons and anesthesiologists, leaving 2,860 treated patients. Specific reasons for exclusion are summarized in Supplemental Digital table 3 (http://links.lww.com/ALN/C886) by treatment group. Patient characteristics, surgery information, and treatment compliance by group are summarized in table 1 and Supplemental Digital table 4 (http://links.lww.com/ALN/C887). Additional intraoperative factors and type of surgery are summarized in Supplemental Digital table 5 (http://links.lww.com/ALN/C888) and Supplemental Digital table 6 (http://links.lww.com/ALN/C889), respectively. After applying propensity score weighting, all baseline and surgery factors in table 1 were well balanced. The maximum pairwise absolute standardized difference among four groups was less than 0.10, decreased substantially from before matching.
Primary Analysis Results
The interaction between tidal volume and PEEP was not significant (P = 0.565). We therefore assessed the weighted treatment effects of tidal volume and PEEP on time-weighted average Spo2/Fio2 ratio independently. The time-weighted average of Spo2/Fio2 ratio was not different in patients assigned to high and low tidal volume, and for those assigned to high and low PEEP, as shown in tables 2 and 3 and in figure 2. We also summarized he minimum, maximum, and range of Spo2/Fio2 ratio per group (Supplemental Digital table 7, http://links.lww.com/ALN/C890). Neither tidal volume nor PEEP significantly altered the time-weighted average Spo2/Fio2 ratio in PACU. The estimated difference of time-weighted average Spo2/Fio2 ratio between tidal volume groups (6 vs. 10 ml/kg) was just 3.5% (97.5% CI, –0.4 to 7.3; P = 0.042, with <0.025 required for significance), and the estimated difference of time-weighted average Spo2/Fio2 ratio between PEEP groups (5 vs. 8 cm H2O) was only –0.2% (97.5% CI, –4.0 to 3.6; P =0.906; tables 2 and 3). In the sensitivity analysis, we included the 621 surgeries that were excluded based on decisions from a surgeon or anesthesiologist. The results were consistent with the primary analysis, with the estimated difference of time-weighted average Spo2/Fio2 ratio being 2.5% (97.5% CI, –1.0 to 6.0; P = 0.117) for tidal volume (6 vs. 10 ml/kg) and –1.2% (97.5% CI, –4.7 to 2.3; P = 0.430) for PEEP (5 vs. 8 cm H2O). Of note‚ the percentage of patients who required supplemental oxygen on the ward and the duration was not different between groups. Neither was the length of PACU stay (Supplemental Digital table 8, http://links.lww.com/ALN/C891).
Post Hoc Analysis Results
Additionally, we performed heterogeneity tests to evaluate whether the treatment effects of tidal volume and PEEP were modified by age, body mass index, smoking status, obstructive sleep apnea, or type of anesthesia. As shown in figure 3, A and B, we did not find significant evidence of heterogeneity for any factors. Age was considered as a continuous factor, and also had no significant impact on modifying the treatment effect of tidal volume and PEEP (interaction P = 0.838 for tidal volume, and P = 0.914 for PEEP).
Secondary Analyses Results
Among 2,340 inpatient surgeries, we compared the secondary outcomes among patients assigned to various tidal volumes and PEEP levels. We did not find significant differences in any of these outcomes, either by tidal volume or PEEP. The estimated difference in ward Spo2/Fio2 ratio for patients assigned to tidal volumes of 6 and 10 ml/kg was just –2.3% (99.2% CI, –6.8 to 2.2; P = 0.172).
The difference for patients assigned to PEEP of 5 versus 8 cm H2O was only 1.1% (99.2% CI, –3.4 to 5.6; P = 0.522). The median length of hospital stay was 3 days (quartile 1 = 2‚ quartile 3 = 5) in each treatment group. The overall incidence of pulmonary complications was 3.1%. The odds ratio was 1.00 (99.2% CI, 0.53 to 1.87; P = 0.992) comparing low to high tidal volume, and was 0.87 (99.2% CI, 0.46 to 1.63; P = 0.553) comparing low to high PEEP.
Exploratory Analysis
Exploratory outcomes are summarized in Supplemental Digital table 8 (http://links.lww.com/ALN/C891) by treatment. Based on a limited number of events, there were fewer deaths in patients assigned to tidal volume of 10 ml/kg and PEEP of 8 cm H2O, presumably a spurious signal.
Discussion
In this factorial multiple crossover cluster trial of 2,860 adults having orthopedic surgery, intraoperative mechanical ventilation with tidal volumes of 6 versus 10 ml/kg of predicted body weight and PEEP of 5 versus 8 cm H2O did not have any clinical meaningful or statistically significant effect on the Spo2/Fio2 ratio in the PACU. There was also no interaction between tidal volume and PEEP. Secondary outcomes also did not differ significantly, including the Spo2/Fio2 ratio on surgical wards, a composite of respiratory events, and duration of hospitalization.
Our primary outcome was the Spo2/Fio2 ratio, a validated measure of oxygenation that presumably detects more subtle lung injury than overt complications. In previous studies, for example, the ratio was shown to be a reliable marker of impaired oxygenation, lung injury, and predictor for early development of acute respiratory distress syndrome and hospital mortality.14,17,18 As expected, the need for oxygen decreased over time after surgery. However, there were no statistically significant or clinically meaningful differences in the Spo2/Fio2 ratio either just after surgery or subsequently on surgical wards. Lack of difference is consistent with a previous study reporting that intraoperative tidal volumes of 6 and 10 ml/kg of predicted body weight do not cause atelectasis as assessed by computerized tomography.19
Even small amounts of PEEP undoubtedly improve oxygenation during mechanical ventilation,20 but our results indicate that there is no persistent benefit (or harm), at least over the range of 5 versus 8 cm H2O. Our findings are consistent with those of Yakaitis et al.,21 who reported with only 15 patients that intraoperative Pao2 improved with PEEP, but that improvement was not sustained in the PACU.21
The initial major trials of mechanical ventilation were conducted in critical care patients, and protective lung ventilation with low tidal volumes, a moderate level of PEEP, and recruitment maneuvers decreased lung injury and reduced morbidity and mortality. However, most ventilated critical care patients have serious pre-existing or acquired lung disease, requiring high levels of ventilatory support that often continues for days. Surgical patients differ in usually having good lung function and requiring mechanical ventilation for a matter of hours. It therefore seems unlikely that results of studies of mechanically ventilated intensive care patients will extrapolate to surgical patients.
The initial studies on intraoperative optimization of mechanical ventilation were inconclusive and inconsistent. For example, the Intraoperative Protective Ventilation trial (IMPROVE; N = 400 patients) compared protective ventilation with tidal volume of 6 to 8 ml/kg predicted body weight and PEEP of 6 to 8 cm H2O, versus tidal volume of 10 to 12 ml/kg without PEEP.22 In contrast to our findings, the researchers reported that protective lung ventilation decreased a composite of pulmonary complications (17% vs. 36%).22 The results were somewhat fragile because sample size was limited. More importantly, the reference treatment was a tidal volume 10 ml/kg or greater without PEEP, which possibly caused overdistension and volume-related trauma. Better outcomes in the protective ventilation groups may therefore have partially resulted from deleterious nonstandard ventilation in the reference group.
Two recent randomized trials support our findings, each demonstrating similar outcomes with various intraoperative ventilation strategies. For example, the PROVHILO trial (N = 900) used a tidal volume of 8 ml of predicted body weight in all patients, and compared 2 cm H2O PEEP without recruitment maneuvers to 12 cm H2O PEEP with recruitment maneuvers.23 Despite the high PEEP level, there were no differences in pulmonary complications—as in our patients. A recent trial by Karalapillai et al. (N = 1,236) compared 6 versus 10 ml predicted body weight of tidal volume. All patients had PEEP set at 5 cm H2O without recruitment maneuvers. The incidence of pulmonary complications was similar in each group, which is consistent with our findings, although we tested different tidal volumes and PEEP levels, and allowed recruitment maneuvers.24
The incidence of complications in our patients was generally lower than reported previously, presumably because our patients all had orthopedic rather than abdominal surgery. Since oxygenation was comparable with each combination of tidal volume and PEEP, as was the incidence of complications, it is unsurprising that the duration of hospitalization was also similar across our groups.
Limitations
Our trial was not randomized, with allocation instead based on sequential weeks. It seems highly unlikely that surgical scheduling over a period of years was based on tidal volume and PEEP allocations. However, previous awareness of the ventilation allocation could have led to biased exclusion of otherwise eligible patients. In fact, though, the characteristics of patients across groups were well balanced after propensity score and inverse probability treatment weighting. Inverse probability treatment weighting helps to create a synthetic sample in which the distribution of measured baseline characteristics would be independent of treatment assignment. Furthermore, we performed a sensitivity analysis that included all patients, whether or not allocated to assigned treatments; the results were consistent with the primary analysis, suggesting no substantive selection bias or confounding.
Because of our cluster design, there were slight imbalances in the number of patients in each group because fewer patients had surgery during holiday weeks. However, differences in allocation were presumably unrelated to exposures and unlikely to influence the results. Pulmonary complications were abstracted from the Cleveland Clinic registry and billing system rather than individual evaluation of patients. It is therefore likely misclassification of outcome/exposures/covariates due to International Classification of Diseases, Tenth Revision, coding, or we missed some, presumably less serious, complications. Again, there is no reason to expect detection bias.
As usual for cluster trials, the current study was not blinded, which could have contributed to measurement bias. However, our outcomes were abstracted from electronic records and seem unlikely to have been influenced by treatment allocation. There are, of course, many other factors that can cause hypoxemia after surgery, including opioids, residual muscle relaxants, fluids, and pain—factors we did not adjust for. However, there is no reason to believe that the contribution of any of these factors differed by group.
Conclusions
Among adults having orthopedic surgery, intraoperative ventilation with 6 versus 10 ml/kg tidal volume and PEEP of 5 versus 8 ml cm H2O did not significantly affect the Spo2/Fio2 ratio during the initial hour of recovery. Furthermore, the Spo2/Fio2 ratio on surgical wards was similar, as was the incidence of pulmonary complications and the duration of hospitalization. Our results suggest that combination of tidal volumes between 6 and 10 ml/kg and PEEP between 5 versus 8 ml cm H2O can be used safely for orthopedic surgery.
Research Support
Support was provided solely from institutional and/or departmental sources.
Competing Interests
The authors declare no competing interests.
Supplemental Digital Content
Supplemental Digital Content 1, Study protocol, http://links.lww.com/ALN/C883
Supplemental Digital Table 1, Estimation of FIO2, the type of device and the oxygen flow rate, http://links.lww.com/ALN/C884
Supplemental Digital Table 2, Composite of Pulmonary complications, http://links.lww.com/ALN/C885
Supplemental Digital Table 3, Reasons of surgeries excluded based on decisions from surgeons and anesthesiologist, http://links.lww.com/ALN/C886
Supplemental Digital Table 4, Compliance of treatment, http://links.lww.com/ALN/C887
Supplemental Digital Table 5, Summary of additional intraoperative factors, http://links.lww.com/ALN/C888
Supplemental Digital Table 6, Detailed surgery type, http://links.lww.com/ALN/C889
Supplemental Digital Table 7, Quantification of SpO2/FiO2 in PACU by treatment, http://links.lww.com/ALN/C890
Supplemental Digital Table 8, Summary of exploratory outcomes, http://links.lww.com/ALN/C891
References
Appendix
Ventilation-PEEP Trial Group:
Alparslan Turan, M.D., is Vice Chair for Outcomes Research and a Staff Anesthesiologist at Cleveland Clinic. Dr. Turan was a scientific advisor and provided writing and technical editing of the manuscript.
Wael Ali Sakr Esa, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at Cleveland Clinic. Dr. Ali Sakr Esa served as a scientific advisor and critical reviewer.
Eva Rivas, M.D., is a Staff Anesthesiologist at Hospital Clinic of Barcelona, Universidad de Barcelona, Barcelona, Spain. Dr. Rivas collected data, provided technical writing, and critical revision of the manuscript.
Jiayi Wang, M.D., is an Anesthesiologist at Shanghai Ninth People’s Hospital, Shanghai, China. Dr. Wang analyzed and interpreted data and contributed to writing.
Omer Bakal, M.D., is a Research Fellow in the Department of Outcomes Research at Cleveland Clinic. Dr. Bakal collected data and provided and cared for study patients.
Samantha Stamper, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic. Dr. Stamper served as a scientific advisor and assisted with technical editing of the manuscript.
Ehab Farag, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology with a secondary appointment in the Department of Outcomes Research at the Cleveland Clinic. Dr. Farag provided general supervision of the research group and assisted with proofreading of the manuscript.
Kamal Maheshwari, M.D., M.P.H., is a Staff Anesthesiologist in the Department of General Anesthesiology with a secondary appointment in Outcomes Research at the Cleveland Clinic. Dr. Maheshwari served as a scientific advisor and assisted with technical proofreading of the manuscript.
Guangmei Mao, M.H., is a senior biostatistician in the Department of Quantitative Health Sciences with a secondary appointment in the Department of Outcomes Research at the Cleveland Clinic. Dr. Mao served as a scientific advisor and managed technical editing and data analytics.
Kurt Ruetzler, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology with a secondary appointment in Outcomes Research at the Cleveland Clinic. Dr. Ruetzler served as a scientific advisor and critical reviewer.
Daniel I. Sessler, M.D., is the Michael Cudahy Professor and Chair, Department of Outcomes Research at the Cleveland Clinic. Dr. Sessler served as a scientific advisor and critical reviewer.
Marcelo Gama de Abreu, M.D., has dual appointments in the Department of Intensive Care and Resuscitation and Outcomes Research at the Cleveland Clinic. Dr. Gama de Abreu provided technical writing.
Metabel Markwei, Sc.M., is a medical student at the Cleveland Clinic Lerner College of Medicine. M. Markwei assisted with editing of the manuscript.
Robert Helfand, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic. Dr. Helfand served as a scientific advisor, and with general supervision of the research group.
Allen Kuhel, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic. Dr. Kuhel critically reviewed the study proposal.
Barak Cohen, M.D., is Staff in the Division of Anesthesiology, Intensive Care, and Pain Management, Tel Aviv Medical Center, Tel Aviv University, Tel Aviv, Israel. Dr. Cohen provided writing assistance.
Loran Mounir Soliman, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic. Dr. Mounir Soliman served as a scientific advisor.
Harsha Nair, M.D., is an Associate Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic. Dr. Nair provided proofreading.
Michael Ritchey, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic.
Sree Kolli, M.D., is a Staff Anesthesiologist in the Department of General Anesthesiology at the Cleveland Clinic. Dr. Kolli served as a scientific advisor and proofreader.
Syed Raza, M.S., worked as medical associate with the Department of Outcomes Research, Cleveland Clinic, and collected analytical data.
Additional Contributors: We appreciate substantial contributions from the following:
Shelby L. Farkas, C.R.N.A., Cleveland Clinic, provided and cared for study patients.
Sarah A. Lamarca, C.R.N.A., Cleveland Clinic, provided and cared for study patients.
Sarah Ceska, C.R.N.A., Cleveland Clinic, provided and cared for study patients.
Ryan Linsalata, C.R.N.A., Cleveland Clinic, provided and cared for study patients.
Caitlin Sullivan, C.R.N.A., Cleveland Clinic, provided and cared for study patients.