Individualized Positive End-expiratory Pressure on Postoperative Atelectasis in Patients with Obesity: A Randomized Controlled Clinical Trial

This is a single-center, randomized controlled trial in patients undergoing laparoscopic bariatric surgery. The trial was registered before patient enrollment at clinicaltrials.gov (NCT04169607; uniform resource locator: https://clinicaltrials.gov/ct2/show/NCT04169607; accessed May 20, 2023; principal investigator: Yuan Han; date of registration: November 20, 2019). The trial was approved by the Ethics Committee of the Affiliated Hospital of Xuzhou Medical University (Xuzhou, China; ID: XYFY2019-KL168-01). All participants signed informed consent before enrollment. The study was conducted following the Declaration of Helsinki. Trial reporting complied with the CONsolidated Standards of Reporting Trials guidelines.

Adult inpatients scheduled to undergo laparoscopic bariatric surgery were eligible. Recruitment was conducted by the site investigator on the day of admission. Subjects with one of more of the following conditions were excluded: American Society of Anesthesiologists (Schaumburg, Illinois) physical status IV or greater, lung bullae, continuing smoking within 1 week before surgery, chronic obstructive pulmonary disease requiring oxygen, congestive heart failure (New York Heart Association [New York, New York] classification III or greater), thoracic surgery history, and planned transfer to intensive care unit after surgery. Enrolled patients were randomized at a 1:1 ratio to undergo surgery using either dynamic compliance–guided PEEP (2 cm H₂O above the PEEP with the highest dynamic compliance) or a fixed PEEP at 8 cm H₂O.

The randomization sequence was generated using SPSS 23 for Windows (SPSS Inc., USA) by a staff member who was not involved in this trial otherwise. Concealment was conducted using sealed, opaque envelopes. The envelope with the group allocation for each patient was opened at the completion of PEEP titration. The patients, the radiologist, and the investigators who assessed the outcomes (except the intraoperative oxygen saturation measured by pulse oximetry (Spo₂), respiratory mechanical parameters, and hemodynamic parameters) were blinded to group allocation; all other personnel were aware of the grouping.

Patients received routine IV rapid sequence anesthesia induction using sufentanil (0.5 μg/kg lean weight) and etomidate (0.3 mg/kg total body weight). Cis-atracurium (0.2 mg/kg total body weight) is given to facilitate tracheal intubation. After intubation, transversus abdominis plane block was performed under ultrasound guidance. Combined IV and inhaled anesthesia was maintained with propofol, remifentanil, sevoflurane, and cis-atracurium until the end of surgery. Routine intraoperative monitoring included Spo₂, electrocardiogram, invasive arterial blood pressure, nasopharyngeal temperature, and muscle relaxant monitoring (train-of-four).

Mechanical ventilation was conducted in the volume- controlled ventilation mode with a tidal volume of 8 ml/kg predicted body weight (male 50 + 0.91 × [height (cm) – 152.4], female 45.5 + 0.91 × [height (cm)– 152.4]) using an A7 anesthesia machine (Mindray Biomedical Electronics, China). Other parameters included PEEP of 8 cm H₂O, inspiratory oxygen fraction (Fio₂) of 1.0, inspiratory-to- expiratory ratio of 1:2, 20% inspiratory pause, and respiratory rate of 12 breaths/min. After randomization, the respiratory rate was adjusted to maintain end-tidal carbon dioxide partial pressure at 35 to 45 mmHg.

All patients, regardless of the group assignment, received a titration trial immediately after intubation. An example of the titration trial and results (inverted U-shaped relationship between dynamic compliance and PEEP) is shown in figure 1. Before the start of the titration trial, patients were switched to a 20° to 25° reverse Trendelenburg and 20° to 25° right-leaning supine position (identical to the position during surgery). The titration trial was initiated by a recruitment maneuver, with the following ventilator setting: pressure-controlled ventilation mode, 10 cm H₂O PEEP, 25 cm H₂O inspiratory pressure, 6 breaths/min respiratory rate, 1:2 inspiratory-to-expiratory ratio. The PEEP and inspiratory pressure were then increased in steps of 5 cm H₂O per 30 s until the final step of 25 cm H₂O PEEP and 40 cm H₂O inspiratory pressure. The driving pressure was maintained at 15 cm H₂O throughout the increase. Then the titration of dynamic compliance–guided PEEP was performed by decreasing the PEEP in steps of 2 cm H₂O every 30 s until the final step of 5 cm H₂O in volume-controlled ventilation mode, with the same ventilatory parameters except PEEP as those at the beginning of mechanical ventilation. The dynamic compliance was calculated using tidal volume (V_T)/(peak pressure - PEEP). To avoid potential underestimation of optimal PEEP value due to pneumoperitoneum, dynamic compliance–guided PEEP was set as 2 cm H₂O above the PEEP with the maximum dynamic compliance.

After the titration trial, Fio₂ was lowered to 0.5 to mitigate absorptive atelectasis. A second recruitment maneuver was performed using the same protocol as the first recruitment maneuver. All patients received 250 ml IV fluid and 40 μg phenylephrine in a bolus before the first recruitment maneuver to prevent hypotension (mean arterial pressure less than 65 mmHg). If hypotension still occurred, an additional bolus of phenylephrine (40 μg) was used.

Upon completion of the titration trial, all patients (regardless of the group assignment) received dynamic compliance–guided PEEP for 1 min for respiratory mechanical and hemodynamic parameters recording. Then investigators opened the envelopes containing the group assignment information. Surgical incision and pneumoperitoneum insufflation proceeded under dynamic compliance–guided PEEP for patients in the dynamic compliance–guided PEEP group, and under 8 cm H₂O for patients in the PEEP of 8 cm H₂O group. PEEP value was maintained throughout the surgery after randomization using the volume- controlled ventilation mode. Intra-abdominal pressure was maintained at 14 mmHg during surgery.

After exsufflation of pneumoperitoneum, the ventilation mode was switched to pressure control, and the patients were switched to a supine position. V_T was maintained by adjusting inspiratory pressure. Extubation was performed in the operating room, and then patients were transferred to postanesthesia care unit. After returning to the ward, supplemental oxygen was provided through a venturi mask at a flow rate of 3 l/min until 8 am next day.

The timepoints of measurements are shown in Table E1 in the Supplemental Digital Content (https://links.lww.com/ALN/D153). Arterial blood gas analysis was conducted before preoxygenation, after intubation, 1 h after insufflation of pneumoperitoneum, before extubation, and 30 min after extubation. Respiratory mechanical parameters and hemodynamic parameters were measured after intubation, titration, and randomization, 10 min and 1 h after insufflation of pneumoperitoneum. Respiratory mechanical parameters consisted of peak pressure, plateau pressure, driving pressure (plateau pressure - PEEP), dynamic compliance [V_T / (peak pressure – PEEP)], and static compliance [V_T / (plateau pressure – PEEP)]. Hemodynamic parameters included heart rate and mean arterial pressure.

Spo₂ was monitored continuously until 8 am the next day using a B650 monitor (GE Healthcare, Finland) in the operating room and a portable pulse oximeter in the ward. Postoperative hypoxemia was assessed based on the percentage of time with a Spo₂ less than 92%.¹² 

Incidence and severity of postoperative pulmonary complications were collected on postoperative days 0, 1, 2, and 7, as previously described (Table E2 in Supplemental Digital Content, https://links.lww.com/ALN/D153).¹⁷  The mechanical ventilation duration, intraoperative vasopressor dosage, and IV fluids were also recorded.

Chest computed tomography scans were conducted at 60 to 90 min after extubation and completed at end- expiration by instructing patients to exhale normally and hold their breath at functional residual capacity. The computed tomography scan was set with a scan quality reference of 120 kVp and 100 mAs by Somatom Definition (Siemens Medical Systems, Germany). Image reconstruction was performed using a slice thickness of 5 mm. The Digital Imaging and Communications in Medicine images were analyzed using Advantage Workstation 4.6 (GE Healthcare, USA).

Ten computed tomography sections (the most cranial and caudal ones, and eight evenly spaced sections in between) were used to calculate the amount of atelectasis for each patient. The validity of this method has been established based on good agreement of differently aerated compartments between the extrapolated results from 10 sections and those from all computed tomography sections by a previous study.¹⁸  The lung area was delineated, and major pulmonary vessels (short diameter 3 mm or greater) were excluded manually. Lung aeration compartments were calculated as a percentage of the total lung tissue volume using the following Hounsfield unit thresholds: nonaerated (–100 to + 100 Hounsfield units), poorly aerated (−500 to −101 Hounsfield units), normally aerated (−900 to −501 Hounsfield units), and hyperaerated (−1,000 to −901 Hounsfield units).¹⁹  Total lung volume (V Lung; including gas and tissue volume) was calculated used the formula²⁰ 

where N is the number of slices, t is the slice thickness, f is the distance between slices (feed), and V_i is the lung total volume in the ith slice. The following formula was used to calculate the tissue volume of each lung aeration category²¹ :

Initially, the primary outcome was the difference of atelectasis between two chest computed tomography scans performed preoperatively and at 30 to 60 min after extubation. Due to the COVID-19 outbreak soon after the trial commencement, preoperative computed tomography was no longer possible due to a strict mandate from the government to minimize preoperative stay and the risk of COVID-19 transmission within hospitals. The change from 30 to 60 min to 60 to 90 min after extubation was due to staff shortage and work overload during the COVID-19 pandemic. At this point, the primary outcome was changed to the amount of postoperative atelectasis in postoperative computed tomography images, expressed as the percentage of nonaerated tissue in total lung tissue volume (amount of atelectasis = nonaerated tissue volume/total lung tissue volume * 100%).

Secondary outcomes included PEEP value, the ratio of Pao₂ to Fio₂, Spo₂, respiratory mechanical parameters, hemodynamic parameters, intraoperative vasopressor dosage, fluids volume, and postoperative pulmonary complications. Postoperative pulmonary complications are defined in detail in Table E2 in Supplemental Digital Content (https://links.lww.com/ALN/D153).

The sample size was calculated based on the assumption of 3% group difference in the amount of postoperative atelectasis, with a SD of 3% in both groups. The 3% margin was arbitrarily set following previously used criteria.⁶  The 3% SD was based on the results of a previous study.³  Other assumptions included α = 0.05, power = 0.85, and dropout rate = 10%. The calculation using a two-tailed Student’s t test yielded 40 patients (20 in each group).

Normality of continuous variables was examined using the Shapiro–Wilk test. Normally distributed variables, including the primary outcome of atelectasis, were analyzed using a Student’s t test and presented as mean ± SD. Variables not following normal distribution were analyzed using a Mann–Whitney U test and presented as the median (interquartile range). Categorical variables were analyzed using the chi-square test or Fisher exact test, as appropriate. For correlation between two variables, the Spearman rank correlation test was used.

To analyze repeated measure outcomes, including respiratory mechanics, hemodynamic parameters, Pao₂/Fio₂, and Spo₂, a linear mixed model with fixed effects for group, timepoint and interaction of timepoint and group, random intercept at the level of participants, and an unstructured covariance matrix was used for calculation of significant differences between groups. Pairwise comparisons were run at each timepoint if statistically significant interaction existed. No correction was applied for multiple comparisons of secondary outcomes. All statistical tests were two-sided and conducted using SPSS 23 for Windows (SPSS Inc., USA). Statistical significance was set at P < 0.05. No imputation of missing values was performed due to the small amount of missing data.

This trial was conducted during a period from December 16, 2019, to September 30, 2020. Patient flow through the trial is shown in figure 2. A total of 40 patients (28 ± 7 yr of age; 25 females) were randomized (20 in each group). In the analysis of the titration data in the entire sample of 40 patients, the PEEP with maximum dynamic compliance was 15 cm H₂O (interquartile range, 13 to 17; range, 8 to 19). The observation suggested a weak positive correlation between body mass index and individualized PEEP with maximum dynamic compliance (Spearman correlation coefficient, r = 0.36), and the scatterplot showing their relationship is presented in figure 3. Baseline and intraoperative characteristics were generally comparable between the two groups (table 1).

As the linear mixed model exhibited statistical significance for group by time interaction in most of the repeatedly measured respiratory mechanics, we reported the results of group comparisons of these variables at each timepoint (fig. 4 and table E3 in the Supplemental Digital Content, https://links.lww.com/ALN/D153). Respiratory mechanics were similar between the two groups during titration. The dynamic compliance–guided PEEP group had higher mean dynamic compliance, static compliance, peak pressure, plateau pressure, and driving pressure throughout the anesthesia than the PEEP of 8 cm H₂O group (e.g., after randomization, 10 min after pneumoperitoneum, 1 h after pneumoperitoneum, and immediately before extubation; all P < 0.001).

The analysis of the primary outcome was conducted using the data of 38 patients (19 in each group) who completed the postoperative computed tomography scan. The postoperative atelectasis was 13.1 ± 5.3% in the PEEP of 8 cm H₂O group versus 9.5 ± 4.3% in the dynamic compliance–guided PEEP group (difference, 3.7%; 95% CI, 0.5 to 6.8%; P = 0.025; table 2; fig. 5). The group difference was 4.5% (95% CI, 0.9 to 8.0%; P = 0.015) in the right lung, and 3.6% (95% CI, –0.3 to 7.5%; P = 0.069) in the left lung. The amount of poorly aerated, normally aerated, and hyperaerated tissue in postoperative computed tomography did not differ significantly between the two groups (table 2).

Arterial blood analysis was not available in two patients (one in each group) before extubation and one patient in the dynamic compliance–guided PEEP group at 30 min after extubation. Pao₂/Fio₂ did not differ between the two groups either before intubation (P = 0.536) or after intubation (P = 0.211). Pao₂/Fio₂ was significantly higher in the dynamic compliance–guided PEEP group at 1 h after pneumoperitoneum (P = 0.001) and before extubation (P = 0.017; fig. 5). Pao₂/Fio₂ at 30 min after extubation did not differ between the two groups (P = 0.183). Percentage of the time with hypoxemia (Spo₂ less than 92%) until 8 am the day after surgery was 1.1% (interquartile range, 0.1 to 8.2%) in the PEEP of 8 cm H₂O group and 0.6% (interquartile range, 0.1 to 3.0%) in the dynamic compliance–guided PEEP group (difference, 0.20%; 95% CI, –0.3 to 4.1; P = 0.407). The rate of hypotension during the recruitment maneuver and the dosage of phenylephrine throughout the surgery did not differ between the two groups (table 1). Persistent hypotension was not observed in either group.

Four patients in each group (20% for both) developed postoperative pulmonary complications within 7 days after surgery. No grade 3 or higher postoperative pulmonary complications were observed.

The results from this trial demonstrated that in patients undergoing laparoscopic bariatric surgery, postoperative atelectasis was lower under dynamic compliance–guided PEEP versus PEEP of 8 cm H₂O. The dynamic compliance–guided PEEP group had higher Pao₂/Fio₂ during but not after surgery.

The analysis of computed tomography revealed that patients in the dynamic compliance–guided PEEP group developed atelectasis accounting for 9.5% of the total lung tissue volume, less than the 13.1% in the PEEP of 8 cm H₂O group at 60 to 90 min after extubation. However, the two groups did not differ in Pao₂/Fio₂ at 30 min after extubation. Such a discrepancy indicates intraoperative lung recruitment may not necessarily translate into high postoperative lung aeration, as previously suggested by Lagier et al.²³  The lack of improvement in postoperative oxygenation in our trial could be due to the fact that postoperative atelectasis differed between the two groups by only 3.6%. However, the similar discrepancy between the reduced postoperative atelectasis and no improvement in postoperative oxygenation has also been reported in several previous studies,^6,8  suggesting the existence of other underlying mechanisms. One potential mechanism is that morphological appearances of atelectasis (such as computed tomography) occur later than physiologic changes (such as Pao₂/Fio₂). Small airways tend to close in the early period after extubation and impair gas exchange in distal alveoli in the absence of postoperative lung stabilization strategies. However, the morphological performance of collapsed alveoli as revealed by computed tomography requires complete absorption of the trapped air, which is a slow process at low inspiratory oxygen fraction.²⁴  Accordingly, the open lung effects of individualized PEEP may partly remain in the early extubation period.

We observed a reduced amount of postoperative atelectasis in the dynamic compliance–guided PEEP group in the right but not in the left lung. In reference to previous studies of intrapulmonary gas distribution of individual lungs,²⁵  this finding is not surprising since surgery was conducted in a partial right-leaning lateral decubitus position in all patients in this trial. Since the effects of individualized PEEP on lung ventilation in the lateral decubitus position has not been explored in other studies, this finding cannot be taken as a definitive conclusion. It is important to note, however, that since the lateral decubitus angle was small in our trial, this conclusion cannot be extended to a completely lateral decubitus position.

The median optimal PEEP in this trial was 15 cm H₂O with a range of 8 to 19 cm H₂O in patients with an average body mass index of approximately 40 kg/m². Such a relatively high PEEP may raise concerns about adverse events and reduce the enthusiasm to use individualized PEEP in routine practice.^26–28  Several previous studies have indicated that high levels of PEEP (10 cm H₂O or greater) are necessary to improve pulmonary function in patients with morbid obesity, pneumoperitoneum, and Trendelenburg position.^6,29,30  A study by Spadaro et al. showed that in patients with a body mass index around 25 kg/m², 5 cm H₂O PEEP was required in nonpneumoperitoneum state in the supine position to significantly reduce the intrapulmonary shunt rate, whereas 10 cm H₂O PEEP was required with pneumoperitoneum.²⁹  A study by Tharp et al. calculated optimal PEEP as the originally set PEEP minus transpulmonary pressure, and showed that in patients with a BMI greater than 40 kg/m², optimal PEEP was 16.8 and 21.5 cm H₂O before and after insufflation of pneumoperitoneum, respectively.³⁰  Considering the mean body mass index of 40 kg/m² in our study, dynamic compliance–guided PEEP with a median of 17 cm H₂O is within the reasonable range, in our opinion.

As mentioned in the introduction, the goal of optimal PEEP is to produce the best compromise of atelectasis and alveolar hyperdistention.⁶  Excessive PEEP leads to alveolar hyperdistension and increased elastic resistance, which in turn counteracts or even exceeds the reduction in airway resistance. Therefore, during the stepwise decrease of PEEP from 25 to 5 cm H₂O in this trial, the relationship between dynamic compliance and PEEP followed an inverted U-shaped pattern (initial increase and then decrease). The PEEP value with maximum dynamic compliance was selected as the optimal PEEP to balance small airway closure versus alveolar hyperdistension. Pneumoperitoneum may impact optimal PEEP value,³⁰  but repeating the titration procedure after insufflation/exsufflation of pneumoperitoneum is not pragmatic. Considering increased need for higher PEEP with pneumoperitoneum, 2 cm H₂O was added to the final dynamic compliance–guided PEEP for use during the entire surgery. Another consideration is the change of body position from a reverse Trendelenburg to supine position upon exsufflation of pneumoperitoneum, which in turn tends to cancel out the effects of exsufflation. Thus, we did not remove the additional PEEP value of 2 cm H₂O after exsufflation of pneumoperitoneum.

Although fluids and vasopressor were given in advance and the maximum plateau pressure was at a relatively low level (40 cm H₂O), hypotension was common during the recruitment maneuver. No refractory hemodynamic instability was observed in this trial, perhaps due to the exclusion criterion of severe cardiopulmonary diseases.³¹  The safety of individualized PEEP in patients with compromised cardiopulmonary function needs further evaluation in future studies. Consistent with previous trials of individualized PEEP in both nonobese and obese patients,^3,4,6  the dosage of phenylephrine and fluid volume during anesthesia were comparable between the two groups, adding support to the hemodynamic safety of individualized PEEP.⁹  The potential reasons of good hemodynamic stability lie in adequate fluid and vasoactive drugs infusion before maneuver,³²  and preserved right ventricular function due to low pulmonary vascular resistance caused by individualized PEEP.³³ 

This trial has several limitations. First, recruitment maneuver with a plateau pressure of 40 cm H₂O may be insufficient for optimal alveolar recruitment in obese patients. A previous study suggested that a plateau pressure as high as 55 cm H₂O may not be suffiicient.²²  This relatively low level was chosen to minimize the impact on hemodynamic stability. Second, atelectasis was assessed using 10 slices instead of the whole lung, and therefore was subject to bias. Having said that, good agreement of differently aerated compartments between the extrapolated results based on 10 sections and those from all computed tomography sections has been established.¹⁸  Accordingly, this method has been widely used in studies of atelectasis.^6,11,18,20  Third, this trial was not sufficiently powered to detect a change in more clinically relevant outcomes, such as postoperative pulmonary complications or Pao₂/Fio₂. Fourth, the fixed PEEP value was 8 instead of 5 cm H₂O in our trial, which was different from previous studies in daily clinical practice.^3,6–8  A higher fixed PEEP in the PEEP of 8 cm H₂O group would tend to decrease the group difference. Under such a condition (favors the PEEP of 8 cm H₂O group), the results (reduced atelectasis) are more solid. Fifth, the average body mass index in this trial (40 kg/m²) is considerably lower than in the obese patients undergoing bariatric surgery in Western countries. Sixth, based on previous findings, optimal PEEP in morbidly obese patients was elevated by approximately 5 cm H₂O after pneumoperitoneum insufflation.³⁰ Adding 2 cm H₂O to the PEEP with maximum dynamic compliance may not be sufficient to balance the impact of pneumoperitoneum. Last, a fixed PEEP rather than static compliance-guided PEEP was used as the control. Accordingly, whether static compliance-guided individualized PEEP is superior to dynamic compliance–guided PEEP remains unknown.

Postoperative atelectasis was lower in patients undergoing laparoscopic bariatric surgery under dynamic compliance–guided PEEP versus PEEP of 8 cm H₂O. Postoperative Pao₂/Fio₂ did not differ between the two groups.

The authors thank Kehong Zhang, M.D., Ph.D., from Ivy Medical Editing (Shanghai, China) for revising the final manuscript. The authors thank the surgical team of Xiaocheng Zhu, M.D., Ph.D., from the Department of Gastroenterological Surgery, Affiliated Hospital of Xuzhou Medical University (Xuzhou, China) for excellent surgery.

This work was supported by the National Natural Science Foundation of China (Beijing, China; #82293641, and 82130033, to Jun-Li Cao), the Jiangsu Provincial Special Program of Medical Science (Nanjing, China; #BL2014029, to Jun-Li Cao), the National Natural Science Foundation of China (#82271295, to Yuan Han), the Natural Science Foundation of Shanghai (Shanghai, China; #21ZR1411300, to Yuan Han), and the Shenkang Clinical Study Foundation of Shanghai (Shanghai, China; #SHDC2020CR4061, to Yuan Han). The funders have no involvement in study design, data collection and interpretation, writing of the manuscript, and the decision to submit the manuscript for publication.

The authors declare no competing interests.

Full protocol available at: yuan.han@fdeent.org. Raw data available at: yuan.han@fdeent.org.

Table E1, E2, E3, https://links.lww.com/ALN/D153