Optimal positive end-expiratory pressure obtained with titration of a fraction of inspiratory oxygen: a randomized controlled clinical trial

Background Optimal intraoperative positive end-expiratory pressure (PEEP) improves patient outcomes. Pulse oximetry has been used to determine the lung opening and closing pressures. Therefore, we hypothesized that intraoperative optimal PEEP obtained by titrating inspiratory oxygen fraction (FiO2) guided with pulse oximetry could improve perioperative oxygenation. Methods Forty-six males undergoing elective robotic-assisted laparoscopic prostatectomy were randomly assigned to either the optimal PEEP group (group O; n=23) or the fixed PEEP of 5 cmH2O group (group C; n=23). Optimal PEEP, defined as the PEEP with the lowest FiO2 or 0.21 to maintain SpO2 greater than or equal to 95%, was obtained in both groups after placing the patients in the Trendelenburg position and conducting intraperitoneal insufflation. Optimal PEEP was maintained for patients in group O. A PEEP of 5 cmH2O intraoperatively was maintained for patients in group C. Both groups were extubated in a semisitting position once the extubation criteria were met. The primary outcome was the arterial oxygen partial pressure (PaO2) divided by the inspiratory oxygen fraction (FiO2) prior to extubation. The secondary outcome was the incidence of postoperative hypoxemia (SpO2 less than 92% on room air after extubation) in the postanesthesia care unit (PACU). Results The median optimal PEEP was 16 cmH2O (IQR 12–18). The PaO2/FiO2 prior to extubation was significantly higher in group O than in group C (77.0±4.9 kPa vs. 60.6±5.9 kPa; P=0.04). PaO2/FiO2 was also significantly higher in group O 30 minutes after extubation (57.6±1.9 vs. 46.6±1.8 kPa; P=0.01). The incidence of hypoxemia on room air in the PACU was significantly lower in group O than in group C (4.3% vs. 30.4%; P=0.02). Conclusions Intraoperative optimal PEEP can be achieved by a titration of FiO2 guided with SpO2. Maintaining intraoperative optimal PEEP improves intraoperative oxygenation and reduces the incidence of postoperative hypoxemia. Trial Registration The study was prospectively registered on September 10, 2021, in the Chinese Clinical Trial Registry (identifier: ChiCTR2100051010).


Abstract Background
Optimal intraoperative positive end expiratory pressure (PEEP) improves patient outcomes. The pulseoximetry has been used to determine the lung opening and closing pressures. Therefore, we hypothesized that intraoperative optimal PEEP obtained by titrating inspiratory oxygen fraction (FiO 2 ) guided with pulse-oximetry could improve perioperative oxygenation.

Methods
Forty-six males undergoing elective robotic assisted laparoscopic prostatectomy were randomly assigned to either optimal PEEP (Group O, n=23) or control with xed PEEP of 5 cmH 2 O (Group C, n=23). Optimal PEEP, de ned as the PEEP with lowest FiO 2 or 0.21 to maintain SpO 2 ≥ 95%, was obtained in both groups after placing the patients in Trendelenburg position and peritoneal insu ation. Patients in Group O maintained the optimal PEEP and in Group C maintained PEEP of 5cmH 2 O intraoperatively. Both groups were extubated in a sitting position once the extubation criteria met. The primary outcome was the partial arterial oxygen pressure (PaO 2 )/inspiratory oxygen fraction (FiO 2 ) prior to extubation. Secondary outcome was the incidence of postoperative hypoxemia (SpO 2 92% on room-air after extubation) in postoperative care unit.

Conclusions
Intraoperative optimal PEEP can be achieved by titration of FiO 2 guided with SpO 2 . Maintaining intraoperative optimal PEEP improves intraoperative oxygenation and reduces the incidence of postoperative hypoxemia.
Trial registration : Chinese Clinical Trial Registry identi er: ChiCTR2100051010. Prospectively registered on 10 September, 2021 Background Optimal intraoperative positive end expiratory pressure (PEEP) has been demonstrated to improve patient outcomes [1,2]. However, the optimal PEEP is not only very different among individuals, but individual's optimal PEEP is affected by positioning, muscle paralysis, and several other factors [3,4]. The common application of a xed PEEP often leads to either lung overin ation or atelectasis. Therefore, optimal PEEP should be individualized and adjusted dynamically according to patients' needs [5]. Several techniques have been used to determine the optimal PEEP[6-10]. For example, electrical impedance tomography (EIT) can be performed at the bedside [5,11,12]. However, the application of this technique requires special training, increases the workload of the care team, and the cost-e ciency of this procedure remains to be determined. Chest CT is the gold standard technique for the assessment of lung in ation [13]. However, it is not feasible for use at the bedside, it exposes patients to X-rays, and its costeffectiveness is not favorable. Transpulmonary pressure is another alternative that can be used at the bedside and is potentially cost-effective [14]. However, it requires special training and additional equipment in order to measure transpulmonary pressure. Lung opening or closing pressure can be used to assess and calculate intrapulmonary shunts [15]. Normally, the physiologic shunt is set at approximately 5%; if arterial blood oxygen saturation is >97% on room air, the intrapulmonary shunt is estimated to be <7% [16]. Therefore, this method can be used to assess the fraction of intrapulmonary shunts and to subsequently estimate the optimal PEEP [17,18].
Recently, Ferrando et al. reported that optimal PEEP can be obtained via titration of PEEP by administering a minimal fraction of inspiratory oxygen (FiO 2 ) with the guidance of pulse oximetry (SpO 2 ) and measurements of transpulmonary pressure in anesthetized patients [19]. The authors found that the optimal PEEP values obtained using the two methods were comparable. Another study demonstrated that SpO 2 could be used to determine the individualized lung opening and closing pressures in patients undergoing anaesthesia and mechanical ventilation [20]. We hypothesized that optimal PEEP could be obtained by titration of intraoperative PEEP levels and FiO 2 with SpO 2 guidance. Our secondary hypothesis was that maintenance of intraoperative optimal PEEP derived via this method improves intraoperative oxygenation and reduces the incidence of postoperative hypoxemia. We tested our hypothesis in patients undergoing robotic-assisted laparoscopic prostatectomy (RALP).

Ethics
This single-centre, two-arm, parallel, randomized controlled study was approved JZ). The study was conducted from 6 May 2021 until 10 October 2021. All patients were approached by the principal investigator and after presentation of the study purposes, written informed consent was obtained before inclusion. All methods were carried out in accordance with Declaration of Helsinki.

Inclusion and exclusion criteria
Between 6 May 2021 and 10 October 2021, adult patients aged 18 years or older who were scheduled for elective robotic-assisted laparoscopic prostatectomy under general anaesthesia and who presented with ASA physical status of I-III were recruited for this study. Patients with acute or chronic respiratory disorders, including chronic obstructive pulmonary disease (COPD), asthma, pulmonary hypertension, neuromuscular disease, and/or preoperative SpO 2 <95% on room air were excluded. The patient enrollment process is illustrated in Fig.1.

Anaesthesia management
Patient's general demographic and medical characteristics were abstracted from medical records; the characteristics investigated herein included sex, age, body mass index (BMI), predicted body weight (PBW), ASA classi cation, medical history, and preoperative SpO 2 on room air. Intravenous access was established upon arrival at the operating room. Routine monitoring for general anaesthesia was performed, including ECG, noninvasive blood pressure, SpO 2 , capnography, and temperature. A radial arterial line was established in order to continuously measure arterial blood pressure and an intermittent blood draw was conducted for blood gas analysis. Patients were pre-oxygenated as usual at an O 2 ow rate of 8 l min -1 until their expiratory oxygen concentration reached 80% or higher. Anaesthetic induction was conducted with intravenous targeted control infusion (TCI) 4 μg mL -1 of propofol (Marsh mode), 0.3 μg kg -1 of sufentanil, and 0.6 mg kg -1 of rocuronium [21]. A 7.0 size tracheal tube was inserted, and correct placement was con rmed with auscultation and the presence of bilateral equal breath sounds. General anaesthesia was maintained with continuous TCI infusion of 3-4 μg mL -1 propofol and 1-2 ng mL -1 remifentanil (Minto mode) as well as intermittent administration of rocuronium in order to maintain adequate muscle paralysis.

Study protocol
The study protocol is summarized in Fig.2. After tracheal intubation, mechanical ventilation was conducted with pressure-regulated volume-controlled ventilation using an operating room ventilator (Flow-I, Maquet Inc., Heidelberg, Germany). The ventilation was set at a tidal volume 6 mL kg -1 , was initiated with an FiO 2 of 1.0 to 0.21, a PEEP of 18 cmH 2 O, and a respiratory rate of 12-15 beats min -1 in order to keep the end-tidal CO 2 partial pressure between 35-45 mmHg. After placement in the Trendelenburg position and peritoneal insu ation, all patients received the rst recruitment maneuver (RM1) of 40 cmH 2 O for 15 seconds followed by PEEP at 18 cmH 2 O, similar to a previous study demonstrating that the maximal optimal PEEP was not greater than 18 cmH 2 O 3 . If the peak inspiratory pressure was >40 cmH 2 O at a PEEP of 18 cmH 2 O, the participant's study would be terminated.
The PEEP titration process is shown in Supplementary Fig.1. If the SpO 2 was at 95-96% with a FiO 2 of 0.21 and a PEEP of 18 cmH 2 O, the optimal PEEP was 18 cmH 2 O; this was kept constant throughout the procedure until extubation. If the SpO 2 was greater than 96%, the PEEP was reduced by 2 cmH 2 O stepwise, with each step lasting for 5 min until SpO 2 dropped below 95%. Then, PEEP was increased up to 18 cmH 2 O in reverse order in the same stepwise manner until the intended SpO 2 was reached and remained at a steady saturation of 95-96%. At a PEEP of 18 cmH 2 O, if the SpO 2 was lower than 95%, the FiO 2 was incrementally increased by 0.05 per step; each step lasted for 5 minutes in order to achieve an SpO 2 of 95-96%. If PEEP was increased to 18 cmH 2 O and FiO 2 was measured at 1.0 (while the SpO 2 remained lower than 95%), the study was terminated. The PEEP level at the minimal FiO 2 necessary to maintain a SpO 2 of 95-96% was considered the optimal PEEP. Once the optimal PEEP was achieved, patients randomized to Group C received a PEEP of 5 cmH 2 O intraoperatively or were maintained within Group O, thus maintaining optimal PEEP until extubation. Patients in both groups were extubated in the post anaesthesia care unit (PACU) in the sitting position once they met the criteria for extubation according to the judgment of their medical care team.
For both groups, intraoperative pulmonary dynamic compliance (Cdyn), PEEP, FiO 2 (i.e., real-time FiO 2 obtained from the gas analyzer within the anaesthesia machine), driving pressure, and plateau pressure were recorded continuously. Intermittent blood gas analysis was performed in order to verify the accuracy of the SpO 2 readings and to calculate the alveolar-arterial gradient [P(A-a)O 2 ], while the respiratory rate was adjusted in order to maintain the PaCO 2 in the range of 35-45 mmHg.
In the PACU, vital signs and arterial blood gas analysis were recorded at 5, 10, and 30 minutes after extubation, and supplementary O 2 was provided to the patients via nasal cannula if the SpO 2 was below 92%.

Statistical analysis
Intraoperative PaO 2 /FiO 2 was reported as 55.7±10.9 kPa before extubation in patients undergoing RALP [4]; we assumed that there were 10 kPa differences between the two groups, with a variance of 10.9 kPa, a statistical power of 80%, and a two-sided α signi cance level of 0.05. A sample size of 18 patients in each arm was required to test our hypothesis. Considering a dropout rate of 30%, a total of 24 patients for each group (for a total of 48 patients) were enrolled; randomization was performed using a minimization randomization method as previously described [22]. Patients were strati ed by age (<65 vs. ≥65yrs) and BMI (<24 vs. ≥24 kg m -2 ) in order to test differences in age and BMI distribution. The randomization was performed via MinimPy2 software (version2.0, OSDN, Columbus, OH, USA).
Randomization was performed the day before surgery by a research team member who was blinded to the trial condition. The data were managed and analyzed by an independent researcher (PL).
Continuous variables were presented as means ± SD or medians with IQR according to whether the distribution was normal, while categorical variables were presented as counts and percentages. The χ 2test was used to compare differences in patient characteristics between the two groups. The unpaired t tests were used to compare differences in oxygen indices, driving pressure, and Cdyn at different time points. Repeated-measures analysis of variance (ANOVA) was used to compare differences in PaO 2 /FiO 2 , driving pressure, and Cdyn under mechanical ventilation prior to extubation. Differences in vital parameters, vasoactive medication dosage, and incidence of complications were tested via unpaired t tests, and the Wilcoxon Man-Whitney test was used when the data were not normally distributed.
Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS)software (version 24, IBM, Armonk, NY, USA) and GraphPad Prism 8.0 software (GraphPad Inc., San Diego, CA, USA). Statistical signi cance was set at P < 0.05.

Clinical characteristics
A total of 48 patients were initially enrolled in this study, though two patients were excluded from the study due to operation cancelation. Therefore, a total of 46 patients completed the study and underwent a nal analysis (Fig. 1). There were no statistically signi cant differences between the two groups in terms of clinical characteristics (Table 1) or perioperative data ( Table 2).

Optimal PEEP level
For all patients (i.e., including those in both groups) the median optimal PEEP was 16 cm H 2 O (interquartile range, 12-18). The FiO 2 needed to obtain the optimal PEEP was 0.21±0.03. The details of the titration process are presented in Supplementary Table 1. The time allotted to complete the titration of the optimal PEEP was half an hour or less.

Secondary outcomes
The respiratory mechanics corresponding to FiO 2 are shown in Fig.3b. There was no statistically signi cant difference in the driving pressure between the two groups (Fig.3c) Postoperative hypoxemia was de ned as postoperative hypoxemia if SpO 2 <92% was detected in room air within 30 min after extubation in the PACU. The incidence of hypoxemia was statistically signi cantly lower in Group O as compared to Group C (1/23 or 4.3% vs. 7/23 or 30.4%, P=0.02) ( Table 2). The P(A-a) O 2 in Group C (9.0±2.0 kPa) was statistically signi cantly higher than that in Group O (3.62±0.9 kPa, P=0.01). PaO 2 /FiO 2 ratios at three time points in the PACU after extubation are shown in Supplementary

Discussion
The main ndings of this study are as follows: (1) intraoperative optimal PEEP can be achieved by titration of PEEP and FiO 2 guided by the SpO 2 readout in patients likely requiring high PEEP; (2) maintaining optimal PEEP improves intraoperative oxygenation and reduces FiO 2 to maintain normoxemia; and (3) the bene t of intraoperative optimal PEEP remains postoperatively in terms of reductions in the incidence of postoperative hypoxemia.
Our results con rmed the observations from a previous study [4] demonstrating that using equipment for routine anaesthesia care can obtain optimal PEEP. This technique has the substantial advantage of being simple to use. In this study, titration of PEEP and FiO 2 were started simultaneously as the surgery progressed. Therefore, clinicians were able to obtain individualized optimal PEEP levels without interrupting or prolonging the surgery. This technique does not require additional training or equipment such as an intra-esophageal balloon to calculate transpulmonary pressure [23] or electric impedance tomography to measure lung aeration [9,24]. All the equipment needed to obtain and maintain the optimal PEEP is readily available in any modern operating room or anaesthesia site. In addition, PEEP can be constantly reassessed and adjusted intraoperatively in order to maintain the optimal PEEP when respiratory mechanics change due to changes in the patient's position or intra-abdominal insu ation pressure. A new algorithm may be developed using a closed-loop system to assess PEEP and automatically implicate the individual's optimal PEEP using this technique.
We tested our hypothesis in patients who underwent RALP because this population is more likely to require high PEEP to minimize intraoperative atelectasis [25]. The surgery was performed within the pneumoperitoneum with an intra-abdominal pressure of approximately15 mmHg, and the patient was placed in a steep Trendelenburg position (approximately 30 degrees) intraoperatively for over 3-4 hours[26, 27]. Therefore, patients are more prone to perioperative atelectasis formation if the PEEP is not high enough to counteract the reduction in functional residual capacity [28]. However, in our institute, a PEEP of 5 cmH 2 O for patients undergoing robotic-assisted laparoscopic prostatectomy is a common practice. There are a few recommendations stating that the PEEP should be higher than 5 cmH 2 O if the patient is in the Trendelenburg position and/or if there is pneumoperitoneum [29]. A PEEP of 5 cmH 2 O seems to be lower than that reported in the literature. However, guidelines for selecting PEEP for this patient population are unavailable due to insu cient literature informing these criteria. Any xed PEEP would render some patients either below and above the optimal PEEP as the variation in optimal PEEP is large [4], and optimal PEEP is likely not a constant but rather varies depending on the patient's physiology and positioning as well as the speci c surgical intervention [30].The range of optimal PEEP observed in this study ranged between 2 and 18 cmH 2 O. We encountered a patient who was able to maintain a SpO 2 >95% with a FiO 2 of 0.21, even when the PEEP was set at 2 cmH 2 O. We also performed arterial blood gas analysis and con rmed that the SpO 2 and arterial hemoglobin oxygen saturation readouts were comparable. This indicates that, even with pneumoperitoneum and in the steep Trendelenburg position, this patient had an intrapulmonary shunt of less than 10% with a PEEP of 2 cmH 2 O [16].
A question remains as to whether the optimal PEEP we obtained was truly the optimal PEEP, since we did not have access to validation via a chest CT scan or electric impedance tomography. However, though this is a scienti cally important question, but it may not be clinically important. Speci cally, the approach employed in this study may not achieve a true PEEP (with no over-or under-PEEP). However, the oxygenation index improved by 27% (77.0/60.6 kPa) in Group O vs. Group C prior to extubation. Further studies are needed to determine the e cacy of this technique for achieving true optimal PEEP versus that obtained with EIT. Nevertheless, using this technique, we could achieve clinically relevant improvements in intraoperative oxygenation as compared with routine care. The mean FiO 2 used to achieve optimal PEEP was 0.21, the SpO 2 was 95%-96% prior to extubation, and the intrapulmonary shunt was estimated to be <10% (as compared with that of the control group). Because we chose to titrate the optimal PEEP stepwise and bidirectionally, we were unlikely to in ate the lung at the optimal PEEP level. Therefore, even though the optimal PEEP in the present study may not be a true optimal PEEP, it is likely very close to the true optimal PEEP and the difference between the two may not have clinical implications. Further studies are needed to assess the agreement of optimal PEEP obtained with the method used in this study as well as other well-established techniques, such as CT scans or EIT.
It is important to note that we found that the bene t of intraoperative optimal PEEP is sustained postoperatively. This is consistent with previous observations suggesting that intraoperatively individualized PEEP can reduce postoperative atelectasis [2]. However, a recent study [31] showed that intraoperative PEEP only improves intraoperative, but not postoperative, oxygenation. This discrepancy among studies, including our current study, maybe due to the different extubation approaches employed in the investigations. It is well known that a patient's functional residual capacity (FRC) depends on sedation level, muscle tone, and position [32]. In our institution, it is routine practice for patients to be extubated in a sitting position. We observed the bene t of intraoperative PEEP on postoperative oxygenation when all patients were extubated in the sitting position. Therefore, patients likely maintain a larger FRC (i.e., closer to the normal value) than that of patients extubated in the supine position. This notion requires further validation. However, in a report by Simon et al., the position of the patients during extubation was not stated [31]. In our study, because there were no statistically signi cant differences between the two groups in terms of the consumption of intraoperative and postoperative narcotic and residual sedation levels in the PACU, the reduction in the incidence of postoperative hypoxemia in room air in Group O was likely due to a reduction in postoperative atelectasis. Since the sample size was relatively small, we could not determine the effect of intraoperative PEEP on other outcomes, such as the incidence of reintubation and postoperative pneumonia. Nevertheless, the intrapulmonary shunt in these patients was an important factor. This is because we chose an SpO 2 of 92% or lower as the cutoff for the diagnosis of hypoxemia on room air; if we assume that the hypoxemia was due to the intrapulmonary shunt only and that no hypoxic vasoconstriction was involved, at an SpO 2 of 92%, the intrapulmonary shunt was estimated to be approximately 24% using the equation described previously [16]. Further studies should be conducted to assess the effect of intraoperative optimal PEEP on outcomes.
In addition to the substantial strengths of this investigation, this study had several limitations. First, we did not validate our observation that the optimal PEEP achieved with this technique is indeed the true optimal PEEP. Validation using EIT or transpulmonary pressure will be important for assessing the sensitivity and speci city of this method. Second, the inaccuracy of pulse oximetry for reporting hemoglobin oxygen saturation was recently determined by the FDA. However, in this study, we validated SpO 2 readings using arterial blood gas analysis. In addition, we used the same brand of oximetry in both groups of patients. Therefore, this potential inaccuracy does not affect our conclusions. Third, it is possible that the PEEP obtained in our study is above the true optimal PEEP. However, we allowed for the descending and ascending stepwise titration of FiO 2 and PEEP. Therefore, over-and under-PEEP at the level that would affect outcomes are possible, but unlikely. Though we may not achieve a perfectly individualized optimal PEEP, but in practical terms, the PEEP value achieved in our study is likely close to the true value when the intrapulmonary shunt is less than 10%.

Conclusion
In conclusion, individualized optimal PEEP can be achieved with equipment available for anaesthesia by titration of PEEP and FiO 2 guided by SpO 2 . Maintaining intraoperative optimal PEEP improves intraoperative oxygenation and reduces the incidence of postoperative hypoxemia in patients likely to require high intraoperative PEEP. Since the method we used in this study to obtain optimal PEEP, this approach is practical and hopefully clinicians are willing to adopt it and improve the quality of care.

List Of Abbreviations
PEEP: positive end expiratory pressure; FiO2: inspiratory oxygen fraction; PaO 2 : partial arterial oxygen pressure; EIT: electrical impedance tomography; SPO 2 : pulse oximetry; RALP: robotic-assisted laparoscopic prostatectomy; COPD: chronic obstructive pulmonary disease; BMI: body mass index; PBW: predicted body weight; TCI: targeted control infusion; RM: recruitment maneuver; PACU: post anaesthesia care unit; Cdyn: dynamic compliance; FRC: functional residual capacity Declarations Ethics approval and consent to participate This study was approved by the Ethics Committee of the Fudan University Shanghai Cancer Center, with the ethics number IRB2010225-11. The patients provided written consent. All methods were carried out in accordance with Declaration of Helsinki.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Competing interest: Dr. Yandong Jiang is a consultant of Vyaire, which was not involved in the design, conduct or publication of this work. The other authors declare no con icts of interest.

Funding
This work was funded by a grant from the Shanghai Science and Technology Committee (No.20Y11906200). The funder was not involved in the design, conduct, or publication of this work.
Authors' contributions LLG conducted data analysis and manuscript preparation, drafted and nalized the manuscript; LY participated in study design, data analysis and manuscript preparation. LLP participated in protocol optimization, data retrieval and data analysis. YC participated in protocol optimization and manuscript preparation. YDJ participated in study design, data analysis, drafted and edited the manuscript. JZ participated study design, data collection and analysis, manuscript preparation and nalization.  Note: Data are presented as mean ± SD or numbers (%); P(A-a)O 2 : alveolar-arterial gradient; PACU: postanesthesia care unit ;pre-extubation: before extubation. Figure 1 Flowchart of patient enrollment. PEEP, positive end expiratory pressure.