Positive end-expiratory pressure and emergence preoxygenation after bariatric surgery: effect on postoperative oxygenation

Abstract

BACKGROUND

Positive end-expiratory pressure (PEEP) is important to increase lung volume and counteract airway closure during anaesthesia, especially in obese patients. However, maintaining PEEP during emergence preoxygenation might increase postoperative atelectasis by allowing susceptible lung areas to be filled with highly absorbable oxygen that gets entrapped when small airways collapse due to the sudden loss of PEEP at extubation.

OBJECTIVE

This study aimed to test the hypothesis that withdrawing PEEP just before emergence preoxygenation would better maintain postoperative oxygenation.

DESIGN

Prospective, randomised controlled trial.

SETTING

Single centre secondary hospital in Sweden between December 2019 and January 2023.

PATIENTS

A total of 60 patients, with body mass index between 35 and 50 kg m⁻², undergoing laparoscopic bariatric surgery.

INTERVENTION

Intraoperative ventilation was the same for all patients with a fixed PEEP of 12 or 14 cmH₂O depending on body mass index. No recruitment manoeuvres were used. After surgery, patients were allocated to maintained PEEP or zero PEEP during emergence preoxygenation.

MAIN OUTCOME MEASURES

The primary outcome was change in oxygenation from before awakening to 45 min postoperatively as measured by estimated venous admixture calculated from arterial blood gases.

RESULTS

Both groups had impaired oxygenation postoperatively; in the group with PEEP maintained during awakening, estimated venous admixture increased by mean 9.1%, and for the group with zero PEEP during awakening, estimated venous admixture increased by mean 10.6%, difference −1.5% (95% confidence interval −4.6 to 1.7%), P = 0.354. Throughout anaesthesia, both groups exhibited low driving pressures and superior oxygenation compared with the awake state.

CONCLUSIONS

Withdrawing PEEP before emergence preoxygenation, did not alter early postoperative oxygenation in obese patients undergoing laparoscopic bariatric surgery. Intraoperative oxygenation was excellent despite using fixed PEEP and no recruitment manoeuvres, but deteriorated after extubation, indicating a need for future studies aimed at improving the emergence procedure.

CLINICAL TRIAL NUMBER AND REGISTRY

www.clinicaltrials.gov, NCT 04150276.

KEY POINTS

• Obese patients undergoing bariatric surgery have impaired oxygenation after awakening from anaesthesia.

• Withdrawing positive end expiratory pressure (PEEP) before emergence preoxygenation, aimed at avoiding high oxygen concentrations in lung areas otherwise susceptible to atelectasis formation after the loss of PEEP, did not hinder impaired postoperative oxygenation.

• Intraoperative oxygenation and respiratory mechanics were excellent with fixed PEEP and no recruitment manoeuvres, suggesting that time-consuming PEEP titration may not always be necessary in obese patients.

• Regardless of any chosen intraoperative ventilation strategy, the study results demonstrate that much of the impaired oxygenation seen postoperatively arises during awakening.

Introduction

Patients with obesity have an increased risk of developing atelectasis and impaired arterial oxygenation during and after general anaesthesia.^1,2 The presumed link between atelectasis and other postoperative pulmonary complications has resulted in numerous studies in obese and normal weight patients trying to optimise positive end-expiratory pressure (PEEP) and intraoperative ventilation. The overall goal is to keep the alveoli open and gas exchange maintained while also avoiding overdistention. Unfortunately, it seems that the favourable respiratory function achieved intraoperatively is lost after awakening from anaesthesia, since ‘protective ventilation’ strategies have not consistently shown improvements in postoperative patient outcomes.^3,4

PEEP increases the end-expiratory lung volume and prevents airway closure by having its dominant effect in dependent lung regions.^5–7 During the emergence phase of anaesthesia, which includes preoxygenation with a high inspired oxygen fraction (FIO₂) before extubation (emergence preoxygenation), maintained PEEP facilitates the entrance of oxygen in these dependent lung regions. At the time of extubation, PEEP is discontinued, resulting in decreased lung volume and closure of small airways, distal to which the highly absorbable oxygen is entrapped. Thus, regardless of any intraoperative ventilation strategy, the fundamental prerequisites for atelectasis formation and impaired oxygenation are, just as after induction, present immediately after awakening. This is especially relevant for patients with obesity, who are at increased risk of reduced lung volume and airway closure due to compression from obese tissue acting on the thoracic cavity and the diaphragm.^8,9

Few studies have investigated the effects of PEEP during the emergence procedure, and its possible influence on early postoperative respiratory function. In one recent study on normal weight patients undergoing non-abdominal surgery, the effect of withdrawing PEEP just before emergence preoxygenation was investigated; postoperative atelectasis formation was limited but similar compared to a group with maintained PEEP.¹⁰ However, obese patients need higher PEEP levels to maintain lung volume¹¹ and could benefit more from withdrawing PEEP when the oxygen concentration is still low. Theoretically, sufficient nitrogen in the dependent lung regions would stabilise the alveoli, reducing the risk of atelectasis and thereby minimising impaired oxygenation shortly after extubation.

We designed this study to test the hypothesis that withdrawing PEEP just before emergence preoxygenation, to avoid high oxygen concentrations in susceptible parts of the lung likely exposed to airway closure after loss of PEEP, would better maintain postoperative oxygenation in obese patients undergoing laparoscopic bariatric surgery.

Materials and methods

Study design

This was a single-centre, randomised controlled superiority trial with a two-arm parallel design. The study protocol was approved by The Swedish Ethical Review Authority, Box 2110, 750 02 Uppsala, Sweden (Chairperson Håkan Julius) on 3 December 2019 (reference number 2019-04860). It was registered prior to patient enrolment at clinicaltrials.gov (NCT 04150276, Principal investigator: Erland Östberg, Date of registration: 4 November 2019). The study was conducted in accordance with the Declaration of Helsinki and the reporting complied with the CONsolidated Standards of Reporting Trials (CONSORT) guidelines. All patients gave written informed consent to participate in the study.

Patients

Adult patients were screened for eligibility if they had a body mass index (BMI) >35 kg m⁻² and were scheduled for laparoscopic bariatric surgery at Västmanland Hospital Västerås. The preoperative exclusion criteria were age ≥60 years, BMI ≥50 kg m⁻², peripheral oxygen saturation (SpO₂) <94% breathing room air, significant gastro-oesophageal reflux disease, coexisting chronic pulmonary disease or asthma, heart failure, ischaemic heart disease, active smoking, or smoking cessation <9 months ago. We also excluded patients with obstructive sleep apnoea syndrome with a prescribed continuous positive airway pressure machine and patients with language barriers.

Anaesthesia and monitoring

No patient received sedating premedication. A radial artery arterial line was used for invasive blood pressure monitoring and arterial blood sampling. Monitoring also included 3-lead electrocardiography and pulse oximetry (Masimo Rad-5, Masimo Corporation, USA).

Patients in both groups were treated identically until starting the awakening procedure. The patients were placed supine with a head-up tilt of 20° to 30°.¹² Preoxygenation was undertaken with an FIO₂ of 1.0 for 3 min, using the anaesthesia machine (Datex-Ohmeda S/5 Avance, GE Healthcare, USA) set to deliver a continuous positive airway pressure (CPAP) of 10 cmH₂O.^13–15 Propofol and remifentanil target-controlled infusions (Injectomat TIVA Agilia, Fresenius Kabi AB, Sweden) were used to induce and maintain anaesthesia. All patients received rocuronium 0.5 mg kg⁻¹ ideal body weight (IBW) to facilitate tracheal intubation. After loss of spontaneous breathing, the anaesthesia ventilator setting was changed and pressure-controlled face mask ventilation with an FIO₂ of 1.0 was commenced, maintaining a PEEP of 10 cmH₂O,¹³ and with an initial inspiratory pressure of 5 cmH₂O. An oropharyngeal airway was used if necessary. The inspiratory pressure was adjusted to each patient's respiratory compliance to deliver a tidal volume of 5 to 6 ml kg⁻¹ IBW until 2 min had elapsed since administration of neuromuscular blockade. After tracheal intubation, mechanical ventilation was performed in a volume-controlled mode with tidal volume 7 ml kg⁻¹ IBW, respiratory rate 10, inspiratory to expiratory ratio 1 : 2, and PEEP 12 or 14 cmH₂O with the higher setting in case the BMI was ≥43 kg m⁻². After securing the airway, the initial ventilator settings also included a high fresh gas flow of 12 l min⁻¹ and FIO₂ 0.21.¹⁶ As soon as the end-tidal oxygen concentration reached 25%, the FIO₂ was set to 0.30 to 0.35 with a fresh gas flow of 1 l min⁻¹. The respiratory rate was adjusted to maintain end-tidal carbon dioxide at approximately 5 kPa. No routine recruitment manoeuvres were used.¹⁷ Any patient with SpO₂ <92% despite increasing FIO₂ to 0.5, suggesting the need for an alveolar recruitment manoeuvre, was to be excluded. Intraoperative mean arterial blood pressure was maintained above 65 mmHg using phenylephrine infusion and intermittent doses of i.v. ephedrine if needed. Surgery was performed in the reverse Trendelenburg position with an intra-abdominal pressure of 15 mmHg.

Awakening from anaesthesia

After completion of surgery and exsufflation of capnoperitoneum, sufficient recovery from neuromuscular blockade was ensured in all patients by a train-of-four ratio of at least 0.9. In the control group, PEEP was maintained during awakening and emergence preoxygenation with FIO₂ 1.0. In the intervention group, before starting emergence preoxygenation, the PEEP level was set to zero (ZEEP, zero positive end-expiratory pressure) while still using the maintenance level of FIO₂ of 0.30 to 0.35. After allowing 2 min for the lungs to adapt to the lower end expiratory pressure, the FIO₂ was increased to 1.0 as for the control group. The upper airways were cleared of mucus through an oropharyngeal airway, but no suction was undertaken in the orotracheal tube. Extubation took place in the operating theatre when patients were awake and breathing spontaneously. All patients received a face mask with FIO₂ 1.0 for a few breaths while the attending anaesthetist ensured a patent airway after extubation. Thereafter all patients breathed room air. The study protocol is summarised in Fig. 1.

Fig. 1.

Schematic diagram of the study protocol.

CPAP, continuous positive airway pressure; FIO₂, fraction of inspired oxygen; IBW, ideal body weight; PEEP, positive end-expiratory pressure; TV, tidal volume; ZEEP, zero positive end-expiratory pressure.

Postoperative care

Pulse oximetry was measured continuously in all patients. Supplemental oxygen 2 l min⁻¹ through a nasal cannula was initiated for any patient with SpO₂ <94% on route to or upon arrival at the postanaesthesia care unit (PACU). Patients were maintained supine with a head-up tilt of 20° to 30° until sampling the fourth (last) blood gas.

The postoperative analgesia regimen consisted of intraoperatively administered i.v. paracetamol, betamethasone, clonidine, parecoxib, and morphine. The attending PACU nurse administered small doses of additional analgesia, either i.v. morphine or alfentanil, at their discretion.

Arterial blood gases

In total, four arterial blood gases were sampled, all with the patients lying supine and with a head-up tilt of 20° to 30°. The first sample was collected while the patients were still awake and breathing room air; the second blood gas was taken 10 min after intubation during maintenance FIO₂ of 0.30 to 0.35; the third blood gas, which was the baseline sample for the primary outcome, was collected after surgery and exsufflation of capnoperitoneum, just before starting emergence preoxygenation and the awakening procedure; the fourth, comparative postoperative blood gas was sampled 45 min after extubation. The postoperative sampling was done with all patients breathing room air to ensure an exact FIO₂, hence enabling an accurate assessment of oxygenation. Thus, any supplemental oxygen was turned off 10 min before this time point. The study protocol allowed a patient's SpO₂ to fall to 88%, in which case the blood gas was collected before restarting supplemental oxygen. Any air bubbles were carefully removed from the blood gas syringe. All blood gas samples were analysed within a few minutes using the Radiometer ABL800 Flex (Radiometer Medical, Denmark).

Primary and secondary outcomes

The primary outcome was change in oxygenation from before awakening to after extubation. We used the estimated venous admixture (EVA), expressed as a percentage of cardiac output, as the principal measure of oxygenation, which has been recommended as the preferable measurement of oxygenation in the absence of a pulmonary artery catheter for the collection of mixed venous blood.¹⁸ EVA is calculated from the arterial blood gas and takes into account not only FIO₂ and the arterial oxygen partial pressure (P_aO₂) but also the arterial carbon dioxide partial pressure (P_aCO₂), the blood pH level, and the haemoglobin level. The calculation was made with the assumption that the arteriovenous oxygen content difference was 40 ml min⁻¹. The P_aO₂/FIO₂ ratio was also reported.

Respiratory mechanics and haemodynamics were recorded throughout anaesthesia. Secondary outcomes included the number of patients in each group in need of postoperative supplemental oxygen to maintain SpO₂ ≥94% and other postoperative variables that could potentially affect the primary outcome: coughing during emergence,¹⁹ administered i.v. opioids, respiratory rate, and postoperative pain.

Randomisation

An independent statistician used the procedure PROC PLAN in SAS version 9.4 (SAS Institute Inc., USA) to perform the randomisation. The 1 : 1 treatment allocation sequence was generated with the use of permuted blocks with random block sizes of eight and ten, and stratified by BMI, creating two strata (<43 and 43−49.9 kg m⁻²). The attending anaesthetist used a web-based system (www.randomize.net) to allocate the patients to their respective groups, the PEEP group or the ZEEP group, just before the start of awakening. Arterial blood gas sampling was performed by the same anaesthetist, who was thus aware of grouping. Patients and post-anaesthetic staff were blinded to group allocation.

Statistical analysis

Based on data from a previous study²⁰ we expected that EVA in the control group would increase by mean 5.5 percentage units with a SD of 6 percentage units. It was deemed clinically important if the intervention group underwent awakening from anaesthesia with essentially preserved oxygenation. We thus hypothesised a mean increase in EVA of 1 percentage unit for the intervention group and considered the corresponding difference of 4.5 percentage units between groups as the minimally clinically important difference. With a two-sided alpha at 0.05 and a power of 0.8, we calculated that the number of patients needed for the study would be 56. Four additional patients, two in each group, were added to allow for dropouts.

Continuous variables that were normally distributed, including the primary outcome change in EVA, were analysed using Student's t-test and presented as mean ± SD. Evaluation of normality was done using visual judgement of histograms. The Mann-Whitney U-test was used to compare non-normally distributed continuous variables. Categorical variables were analysed using the χ² test, or Fisher's exact test for variables with cell counts less than five in the contingency table. Comparisons within each group were analysed using the Wilcoxon signed-rank test.

As a sensitivity analysis, the primary outcome was also analysed using ordinary linear regression. Analyses were performed as crude, including only the randomisation stratification variable BMI and treatment group indicator (ZEEP vs. PEEP), and adjusted additionally, including age, sex and ASA physical status classification. The fit of the regression model to data was examined using residual scatter plots and qq-plots to evaluate normality assumptions.

For all tests, a two-sided P value of less than or equal to 0.05 was considered statistically significant. Statistical analyses were done in R studio using R statistical programming language version 2022.07.2 (R Core Team, R Foundation for Statistical Computing, Vienna.)

Results

The study completion was delayed because of the COVID-19 pandemic. Between December 2019 and January 2023, we assessed 146 patients for eligibility; 86 were excluded for reasons outlined in Fig. 2. The 60 included patients all received their allocated treatment, but for the primary outcome, only 59 patients could be analysed due to a missing baseline blood gas in one patient (Fig. 2).

Fig. 2.

Consort diagram of the study.

CPAP, continuous positive airway pressure; PEEP, positive end-expiratory pressure; ZEEP, zero positive end-expiratory pressure.

Preoperative patient characteristics were similar in the two groups, except that more men were allocated to receive PEEP during awakening (Table 1).

Table 1.

Preoperative patient characteristics

	PEEP group (n = 31)	ZEEP group (n = 29)
Age (years)	39 ± 9	35 ± 10
Sex
Male	9 (29)	1 (3)
Female	22 (71)	28 (97)
Height (cm)	169 ± 7	166 ± 8
Weight (kg)	113 ± 15	109 ± 12
Body mass index (kg m−2)	39 ± 3.3	40 ± 2.8
35–42.9	27 (87)	25 (86)
43–49.9	4 (13)	4 (14)
Haemoglobin (g l−1)	140 ± 11.1	135 ± 9.6
ASA physical status classification
2	20 (65)	18 (62)
3	11 (35)	11 (38)
Type of surgery
Gastric bypass	25 (81)	23 (79)
Gastric sleeve	6 (19)	6 (21)

Data are presented as mean ± SD or number (%).

Intraoperative ventilatory and haemodynamic measurements were similar between the groups (Table 2).

Table 2.

Intraoperative ventilation, haemodynamics, and procedure data

	PEEP group (n = 31)	ZEEP group (n = 29)
Tidal volume (ml kg−1 of ideal body weight)	7.1 ± 0.2	7.1 ± 0.2
PEEP level
12 cmH2O	27 (87)	25 (86)
14 cmH2O	4 (13)	4 (14)
Peak pressure (cmH2O)
After intubation	21 ± 1.4	22 ± 1.6
Mid surgery	25 ± 1.8	25 ± 1.7
Before awakening	23 ± 1.8	23 ± 1.3
Plateau pressure (cmH2O)
After intubation	19 ± 1.4	20 ± 1.7
Mid surgery	23 ± 1.6	23 ± 1.7
Before awakening	20 ± 1.6	21 ± 1.4
Driving pressure (cmH2O)
After intubation	7 ± 1.3	8 ± 1.5
Mid surgery	11 ± 1.6	10 ± 1.7
Before awakening	8 ± 1.5	8 ± 1.3
Static respiratory system compliance (ml cmH2O−1)
After intubation	59 ± 11	53 ± 13
Mid surgery	43 ± 9	39 ± 10
Before awakening	53 ± 11	48 ± 9
Respiratory rate (breaths min−1)
After intubation	10 ± 0.8	10 ± 0.8
Mid surgery	12 ± 1.5	12 ± 1.9
Before awakening	14 ± 1.6	14 ± 2.0
End tidal CO2 (kPa)
After intubation	5.2 ± 0.3	5.2 ± 0.4
Mid surgery	5.3 ± 0.3	5.4 ± 0.3
Before awakening	5.8 ± 0.4	5.7 ± 0.4
Mean arterial pressure (mmHg)
After intubation	82 ± 11	80 ± 10
Mid surgery	84 ± 13	83 ± 14
Before awakening	82 ± 16	79 ± 10
Total i.v. phenylephrine (mg)	1.5 ± 0.7	1.7 ± 1.1
Total i.v. ephedrine (mg)	2.1 ± 3.7	4.1 ± 7.2
Duration of anaesthesia (min)	109 ± 28	120 ± 29
Duration of surgery (min)	66 ± 18	73 ± 18

Data are presented as mean ± SD or number (%).

Before awakening is after completed surgery and after capnoperitoneum.

IBW, ideal body weight; PEEP, positive end-expiratory pressure.

No patient needed increased FIO₂ above the intended 0.30-0.35 to maintain intraoperative SpO₂; thus, none fulfilled the intraoperative exclusion criteria of needing a recruitment manoeuvre. The two groups had similar oxygenation at the end of surgery before randomisation and awakening (Table 3).

Table 3.

Oxygenation and primary outcome

	PEEP group (n = 31)	ZEEP group (n = 29)	Difference in means 95% CI	P value
FIO2
Preoperative	0.21	0.21
After intubation	0.31 ± 0.01	0.31 ± 0.01
Before awakening	0.31 ± 0.01	0.31 ± 0.02
Postoperative	0.21	0.21
SpO2 (%)
Preoperative	98.1 ± 1.6	97.8 ± 1.5
After intubation	98.6 ± 1.7	98.6 ± 1.4
Before awakening	98.2 ± 1.4	98.8 ± 0.9	−0.6 (−1.1 to 0.1)	0.082
Postoperative	94.9 ± 2.3a	95.4 ± 2.0b	−0.5 (−1.7 to 0.6)	0.319
PaO2 (kPa)
Preoperative	11.2 ± 1.0	11.4 ± 1.5
After intubation	17.6 ± 3.5	18.8 ± 2.8
Before awakening	18.3 ± 3.5	19.1 ± 3.2b	−0.8 (−2.6 to 0.9)	0.336
Postoperative	10.2 ± 1.3	10.1 ± 1.3b	0.1 (−0.6 to 0.8)	0.797
Estimated venous admixture (%)
Preoperative	9.9 ± 3.8	9.4 ± 5.2
After intubation	6.9 ± 5.1	5.3 ± 3.5
Before awakening	5.7 ± 5.2	4.4 ± 4.3b	1.3 (−1.1 to 3.8)	0.288
Postoperative	14.8 ± 7.3	14.9 ± 6.5b	−0.1 (−3.7 to 3.4)	0.933
PaO2/FIO2 ratio (kPa)
Preoperative	53.5 ± 4.9	54.3 ± 7.1
After intubation	57.0 ± 10.2	60.4 ± 9.2
Before awakening	58.5 ± 10.6	61.0 ± 9.5b	−2.5 (−7.7 to 2.7)	0.346
Postoperative	48.4 ± 6.4	48.0 ± 6.0b	0.4 (−2.8 to 3.7)	0.797
Change in estimated venous admixture
Preoperative vs. before awakening	4.2 ± 5.6	4.9 ± 6.7b	−0.7 (−2.6 to 3.9)	0.687
45 min postoperatively vs. before awakeningc	9.1 ± 6.1	10.6 ± 6.0b	−1.5 (−4.6 to 1.7)	0.354
Change in PaO2/FIO2 ratio
Preoperative vs. before awakening	−5.0 ± 10.2	−6.4 ± 10.3b	1.4 (−6.8 to 3.9)	0.598
45 min postoperatively vs. before awakening	−10.1 ± 9.0	−13.0 ± 8.4b	2.9 (−1.6 to 7.4)	0.204

Data are presented as mean ± SD.

FIO₂, inspired oxygen fraction; P_aO₂, arterial oxygen partial pressure; SpO₂, peripheral oxygen saturation.

^an = 29, two missing values in the PEEP group.

^bn = 28, one missing value in the ZEEP group.

^cPrimary outcome. Before awakening is after the completion of surgery and after capnoperitoneum. The postoperative values were recorded at the time of sampling the postoperative blood gas (at room air), 45 min after extubation. Estimated venous admixture is expressed as a percentage of cardiac output. An increase represents more venous admixture and thus impaired oxygenation.

Regarding the primary outcome, both groups had impaired oxygenation postoperatively, illustrated by an increase in EVA (Fig. 3). In the group with PEEP maintained during awakening, EVA increased by mean 9.1%, and in the group with ZEEP during awakening, EVA increased by mean 10.6%, difference in means −1.5% (95% confidence interval (CI) −4.6 to 1.7%), P = 0.354. The sensitivity analysis showed a similar result and confirmed that there was no clinically significant difference between groups regarding the primary outcome even after adjustments for age, sex, BMI, ASA class and baseline EVA before awakening.

Fig. 3.

Changes in mean estimated venous admixture (EVA) as a percentage of cardiac output for patients in the two groups (a), and the corresponding changes for individual patients within each group (b).

An increase represents more venous admixture and thus impaired oxygenation. Error bars indicate standard deviation. Both groups were ventilated with positive end-expiratory pressure and treated identically during anaesthesia until before awakening.

Both groups exhibited lower mean EVA values and thus better oxygenation, during anaesthesia compared to the awake state (Fig. 3). The 59 analysed patients combined had a preoperative mean EVA of 9.6% compared to mean 5.1% before awakening, difference 4.5% (95% CI 2.9 to 6.1%), P < 0.001. When comparing the preoperative and postoperative blood gases, the two groups together had a mean EVA of 14.8% 45 min after extubation, which was 5.2% higher than preoperatively (95% CI 3.3 to 6.9%), P < 0.001. The individual changes in EVA during the study timeline are depicted in Fig. 3b.

There was no significant difference between groups regarding the secondary outcome, that is, the number of patients in each group needing postoperative supplemental oxygen to maintain SpO₂ ≥94% (Table 4). When breathing room air preparatory to the postoperative blood gas, one patient in the PEEP group reached the predetermined SpO₂ level of 88% after 8 min, at which time the blood gas was sampled in line with the study protocol. No clinically relevant differences were observed between groups regarding other postoperative variables that could potentially affect the primary outcome (Table 4).

Table 4.

Postoperative parameters

	PEEP group (n = 31)	ZEEP group (n = 29)	Difference in means, 95% CI	P value
Cough score (0–3)
Before extubationa
0–1	17 (63)	16 (59)		0.780
2–3	10 (37)	11 (41)
After extubation
0–1	30 (97)	26 (90)		0.346
2–3	1 (3)	3 (10)
SpO2 (%)
Leaving operating room	96.9 ± 2.8	96.6 ± 2.3	0.3 (−1.0 to 1.8)	0.617
On arrival in PACU	94.3 ± 3.4	94.5 ± 2.0	−0.2 (−1.6 to 1.2)	0.788
On room air at postoperative blood gas	94.4 ± 2.8	95.4 ± 1.9	−0.5 (−1.7 to 0.6)	0.319
Supplemental oxygenb
Leaving operating room	9 (29)	7 (24)		0.668
Any, started in PACU	18 (58)	17 (58)		0.965
Still at 35 min postoperatively in PACUc	14 (45)	16 (55)		0.438
I.v. opioids adm. before postop. blood gas
Any	17 (55)	10 (34)
Alfentanil (mg)	0.4 ± 0.2	0.6 ± 0,2	−0.2 (−0.5 to 0.1)	0.145
Morphine (mg)	5.3 ± 3.7	4.6 ± 2.8	0.7 (−2.7 to 4.2)	0.643
Vital signs at postoperative blood gas
Respiratory rate (breaths min−1)	16 ± 3	15 ± 3	0.8 (−0.6 to 2.3)	0.259
NRS score (0–10)	5 [3.0 to 7.5]	4 [3.0 to 5.0]		0.288

Data are presented as mean ± SD, median [IQR] or number (%).

NRS, numeric pain rating scale; PACU, postanaesthesia care unit; SpO₂, peripheral oxygen saturation.

^an = 27 in both groups, four missing values in the PEEP group, and two missing values in the ZEEP group.

^bSupplemental oxygen was administered by nasal cannula, 1 to 2 l min⁻¹.

^cAny supplemental oxygen at 35 min postoperatively was turned off to allow 10 min of air breathing before sampling the postoperative blood gas at 45 min.

Discussion

The main finding of this study was that withdrawing PEEP before emergence preoxygenation, aimed at avoiding high oxygen concentrations in susceptible lung areas liable to airway closure after the loss of PEEP, was not effective in preventing impaired postoperative oxygenation in obese patients undergoing laparoscopic bariatric surgery.

Based on our current understanding of the mechanisms behind atelectasis formation, the intervention group in our study was expected to develop less postoperative atelectasis and thus better oxygenation. However, several explanations may be proposed for the observed lack of difference in oxygenation between the study groups. Firstly, any difference between groups in the amount of atelectasis might have been too small to induce a measurable effect on oxygenation, because some previous studies have demonstrated differences in postoperative atelectasis without any corresponding effect on oxygenation.^21,22 Secondly, for any patient with atelectasis, the effect on oxygenation may be blunted by the impact of hypoxic pulmonary vasoconstriction when the pulmonary circulation in the awake patient is distributed away from collapsed lung areas.²³ Thirdly, some open non-atelectatic lung tissue, stabilised by nitrogen because of the intervention, may be located behind closed airways and therefore may not contribute to improved gas exchange.

The obese patients in this study had markedly impaired oxygenation at 45 min postoperatively. This is consistent with a large observational study on similarly obese patients, where impaired oxygenation was the most commonly reported postoperative pulmonary complication and was also associated with an increased length of stay.²⁴ However, despite the deterioration in oxygenation noted in the early postoperative period in our study, no patient needed more than 2 l min⁻¹ of supplemental oxygen, and the subsequent clinical course was uneventful with all patients being discharged the same day or the day after surgery.

The superior intraoperative oxygenation and low driving pressures for the 60 patients combined in this study are notable since we did not use individualised PEEP but fixed PEEP levels at 12 or 14 cmH₂O and no recruitment manoeuvres. We did use CPAP during preoxygenation followed by pressure-controlled face mask ventilation with PEEP during induction. In addition, after intubation we used an early applied PEEP and immediately also an oxygen washout manoeuvre to restore nitrogen and thus a low oxygen concentration in the alveoli. These actions should favour an open lung, although the contribution from either measure to intraoperative oxygenation or low driving pressures is uncertain. On a group level, the patients had a much lower mean EVA during anaesthesia than awake, and no patient needed an FIO₂ above 0.35 to maintain intraoperative SpO₂. This was also true for a few patients who deviated from the mean with higher intraoperative EVA values (Fig. 3b). The intraoperative oxygenation and driving pressures were comparable to those recently reported in two studies on obese patients receiving recruitment manoeuvres and individualised PEEP, titrated by electrical impedance tomography¹¹ and dynamic compliance,²² respectively. Thus, our results suggest that intraoperative ventilation may be simplified for many obese patients, at least when undergoing uncomplicated procedures such as bariatric surgery. We adhere to the proposition that for a given patient, there is a range of PEEP levels that are ‘adequate’ in a certain clinical scenario,²⁵ rather than one specific PEEP level that must be individually titrated. Nevertheless, the individual EVA values (Fig. 3b) show that there are certainly patients who may benefit from recruitment manoeuvres and more meticulous PEEP titration also in this setting. The most rational strategy might be to apply a fixed but adequate PEEP to most patients, while using individualised PEEP for those diagnosed with oxygenation problems or high driving pressures.

Regardless of any chosen intraoperative strategy, our results demonstrate that much of the impaired oxygenation seen postoperatively arises during awakening. This is further illustrated in a very recent study by Li et al., where an intervention group treated with the most up to date intraoperative protective ventilation strategy still developed large amounts of postoperative atelectasis.²² A strategy of lowering the FIO₂ during emergence may have an effect on postoperative oxygenation but requires FIO₂ levels that will drastically impair oxygen reserve and safe apnoea time.^20,26

Altogether, these findings suggest that the emergence phase of anaesthesia should be the focus of future research. A reasonable objective for any anaesthetic procedure should be that patients after awakening have the same oxygenating capacity as before induction. A difficult but important task for future studies is to find a way to perform the awakening procedure in high-risk patients with maintained safety for any extubation-related airway problems, without making the lungs susceptible to early postoperative atelectasis.

Limitations

Our study has limitations. Firstly, although there is a strong correlation between intrapulmonary shunt induced by atelectasis and impaired oxygenation during anesthesia,²⁷ several other factors may affect oxygenation and therefore lower the sensitivity of ‘oxygenation’ as a surrogate outcome measure for alveolar collapse after awakening and extubation. For example, the effect of simultaneously acting areas of low ventilation to perfusion ratios on oxygenation will be maximised during air-breathing and cannot, without sophisticated methods, be distinguished from a pure shunt.²⁸ Secondly, computed tomography scanning remains the accepted standard to quantify pulmonary atelectasis and would have been preferred to ultimately exclude any differences in postoperative atelectasis between the two groups. Thirdly, although oxygenation was good in both groups during anaesthesia without recruitment manoeuvres, a recruitment manoeuvre before starting the awakening procedure might have been preferable from a scientific perspective to ensure that baseline atelectasis was small and similar between groups. However, any effect of such atelectasis on oxygenation should be accounted for by the choice of change in oxygenation from before awakening to after awakening as the primary outcome.

Conclusion

Withdrawing PEEP before emergence preoxygenation did not alter early postoperative oxygenation in obese patients undergoing laparoscopic bariatric surgery. Intraoperative oxygenation and respiratory mechanics were excellent despite using fixed PEEP and no recruitment manoeuvres; however, oxygenation deteriorated after extubation, indicating a need for future studies aimed at improving the emergence procedure.