Airway Closure during Surgical Pneumoperitoneum in Obese Patients

Airway closure is a lack of communication between proximal airways and alveoli due to airway collapse. When this phenomenon is present, tidal inflation starts only after a critical airway opening pressure is overcome and airway pressure at end expiration does not reflect alveolar pressure, with some alveoli still inflated with higher pressure not transmitted to airway opening.¹ 

Airway closure has been suggested as a mechanism driving expiratory flow limitation in mechanically ventilated patients with asthma and chronic obstructive pulmonary disease.^2,3  Recently, it has been systematically defined and its occurrence described in up to 20% of patients with acute respiratory distress syndrome.^1,4  These patients show a pronounced inflection point in the inspiratory limb of the respiratory system pressure-volume curve obtained during low-flow inflation. Below this pressure, the compliance of the system equals compliance of an occluded respiratory circuit, and the increase in airway pressure does not yield any change in esophageal pressure (P_ES), suggesting that no changes in lung volume are occurring. This allows the hypothesis that below this “critical point” (i.e., the airway opening pressure), no gas is inflated in the lung because of an obstacle distal to the respiratory circuit but proximal to alveoli, which was hence defined airway closure.¹  A case report recently confirmed that this mechanical pattern may be dependent on airway collapse, as electrical impedance tomography showed that, effectively, any increase in airway pressure below airway opening pressure is not accompanied by lung inflation.⁵  During general anesthesia, a similar phenomenon has been reported in obese patients undergoing abdominal surgery but was interpreted as an airway pressure threshold to generate positive transpulmonary pressure.^6,7 

Among the procedures of a clinical trial in obese patients undergoing robot-assisted laparoscopic surgery in the Trendelenburg position, we measured end-expiratory lung volume, respiratory, and lung mechanics before and after surgical pneumoperitoneum induction. During this study, we noticed that some patients exhibited a pressure-volume curve profile compatible with airway closure, with some of them having an airway opening pressure that reached high values after pneumoperitoneum was instituted in Trendelenburg position.

We hypothesized that these patients could have a peculiar physiologic behavior during general anesthesia, paralysis, and mechanical ventilation. We therefore conducted a study to report the occurrence of airway closure in obese patients undergoing robotic gynecological surgery, comparing their respiratory mechanics with those of matched control subjects who were enrolled in the same trial but did not show any airway closure phenomenon.

This analysis was conducted on patients enrolled in the ongoing randomized study “Intraoperative protective ventilation for obese patients undergoing gynecological laparoscopic surgery,” which is aimed at assessing the effects of a composite strategy providing low tidal volumes (V_T), higher positive end-expiratory pressure (PEEP), and scheduled recruitment maneuvers on postoperative oxygenation. The study was approved by the local ethics committee and is conducted in accordance with the Declaration of Helsinki. All patients provided written informed consent to participate in the trial and data analysis; the study protocol was registered at clinicaltrials.gov (NCT03157479).

All adult morbidly obese patients (body mass index greater than 35 kg/m²) scheduled for gynecological laparoscopic and/or robot-assisted surgery in the Trendelenburg position were considered eligible for inclusion in this trial. Main exclusion criteria were clinical history or signs of chronic heart failure, history of neuromuscular disease, history of thoracic surgery, pregnancy, and chronic respiratory failure requiring any kind of domiciliary respiratory support.

Among patients enrolled in the trial, we identified those who showed airway closure during respiratory mechanics assessment after intubation in the supine position. Airway closure was diagnosed on the low-flow inflation pressure-volume curve of the respiratory system conducted after prolonged exhalation, according to the criteria proposed by Chen et al.1 Airway closure was diagnosed when the pressure-volume curve exhibited a pronounced lower inflection point and the compliance (ratio of inflated volume to pressure change) of its initial flat part equaled the one of an occluded respiratory circuit, which was measured by the ventilator during the pre-use self-test and always ranged between 1.5 and 2.5 ml/cm H₂O (figs. 1 and 2). In the current study, we compared affected subjects with a control group of patients who had been enrolled in the same study, who did not show any airway closure phenomenon and were matched to affected subjects in a 1:1 ratio, according to the body mass index (±2 kg/m²). If in the control group more than one patient met this criterion, the one with body mass index closest to the case subject was included.

All patients enrolled in the trial underwent pulmonary function testing at hospital admission. During surgery, patients received total intravenous anesthesia with full paralysis according to a standard protocol that included the following: propofol 2 mg/kg for induction of anesthesia and 6 to 8 mg · kg^–1 · h^–1 for maintenance, titrated to keep a bispectral index value between 35 and 45%; fentanyl 8 mcg/kg in divided doses; rocuronium bromide 0.6 to 0.8 mg/kg for induction and then administered to ensure a profound neuromuscular block, as assessed by train-of-four monitoring and absence of inspiratory swings on esophageal pressure tracing; and 3 to 5 ml · kg^–1 · h^–1 of intravenous balanced crystalloids administered throughout the whole surgical procedure. Fluid boluses or amine (i.e., norepinephrine, ephedrine, dobutamine) administration were allowed if deemed necessary by the attending anesthesiologist to maintain cardiac output and hemodynamic parameters within the physiologic range.

All patients were intubated and connected to a gas-compressed–based mechanical ventilator equipped with a tool for lung volume measurement (Carescape R860, General Electrics HealthCare, USA) and received volume-controlled ventilation with V_T and PEEP set according to the assigned protocol. In the protective ventilation arm, V_T was set at 6 ml/kg of predicted body weight (calculated as 45.5 + 0.91·[height  – 152.4]), PEEP 10 cm H₂O and scheduled recruitment maneuvers were performed; in the standard ventilation arm V_T was set at 10 ml/kg predicted body weight and PEEP 5 cm H₂O. In both groups, inspiratory flow was set at 1 l/s, and inspiratory pause at 0.3 s. Fraction of inspired oxygen (Fio₂) was titrated to obtain an oxygen saturation measured by pulse oximetry greater than or equal to 94%. Respiratory rate was set to maintain end-tidal carbon dioxide was between 30 and 40 mmHg for the whole duration of the study.

After intubation, a polyfunctional nasogastric tube provided with an esophageal balloon (Nutrivent, Sidam, Italy) was placed and secured in all patients to measure esophageal pressure and estimate transpulmonary pressure. The signal was validated according to the most recent recommendations both before and after pneumoperitoneum induction and Trendelenburg position.^8–10 

At the end of surgery, all patients were extubated and received oxygen therapy with a 40% Fio₂ VenturiMask (FIAB, Florence, Italy) in the semirecumbent position. After 1 h and on day 1 after surgery, patients were reassessed with arterial blood gas analysis and clinical evaluation. All patients underwent postoperative pulmonary function testing on day 2 after surgery.

End-expiratory lung volume, static respiratory mechanics, and pressure-volume curves during low-flow inflation (5 l/min) after 6 to 8 s of prolonged exhalation were recorded after intubation in the supine position and 15 min after pneumoperitoneum in the Trendelenburg position. End-expiratory lung volume was measured through a modified technique during nitrogen wash-out (20% Fio₂ increase) and wash-in (20% Fio₂ decrease), as previously described.^11,12  Wash-out and wash-in data were averaged automatically and considered valid if the difference between the two was less than 20% (cut-off determined by the manufacturer).

A Fleisch-type pneumotacograph (n.2, Metabo, Switzerland) and a pressure transducer measured (sample rate = 200 Hz) flow and airway pressure, which were transmitted to an analog-digital converter together with esophageal pressure. Tidal volume was calculated by digital integration of the expiratory flow signal. All the signals were recorded and reviewed offline through a dedicated software (ICU Lab, Kleistek, Italy). Airway plateau pressure (P_PLAT) and end-inspiratory esophageal pressure (P_ESend-insp) were measured during 0.5-s end-inspiratory occlusion, total PEEP (PEEP_TOT) and end-expiratory esophageal pressure (P_ESend-exp) during end-expiratory hold.

End-inspiratory and end-expiratory transpulmonary pressures were computed as airway pressure minus esophageal pressure (P_Lend-insp and P_Lend-exp, respectively). Airway driving pressure (ΔP) was calculated as P_PLAT-PEEP_TOT and transpulmonary driving pressure (ΔP_L) as P_Lend-insp-P_Lend-exp. Respiratory system and lung static compliance (C_RS and C_L, respectively) were defined as V_T/ΔP and V_T/ΔP_L, respectively, while static chest wall compliance (C_CW) was computed as V_T/(P_ESend-insp-P_ESend-exp).¹³  End-inspiratory transpulmonary pressure was also calculated with the elastance-derived method (P_Lel-der) as P_PLAT·(ΔP_L/ΔP).¹⁴  Dynamic ΔP and C_RS were calculated from set PEEP and P_PLAT measured after the 0.3-s inspiratory pause. These are commonly displayed by ventilators during ventilation and were used as surrogates for respiratory mechanics assessment or clinical interventions in several studies on this topic.^15–20 

As shown in figure 1, respiratory system tidal elastic pressure was defined as the total increase in alveolar pressure due to tidal volume and was calculated on the respiratory system pressure-volume curve as end-inspiratory pressure minus PEEP_TOT or, in case of airway closure, airway opening pressure (no patient exhibited a drop in the airway pressure after airway opening pressure was overcome during low-flow inflation). Lung tidal elastic pressure (i.e., the increase in transalveolar pressure due to V_T) was measured in a similar fashion on the transpulmonary pressure-volume curve. Respiratory system and lung linear compliances (linear C_RS, C_L) were calculated as the slope of the steeper segment of the respiratory system and transpulmonary pressure-volume curve, respectively.^21,22 

No statistical power calculation was conducted before the study, and the sample size was based on the available data from patients enrolled in the trial. Continuous data are expressed as median (interquartile range), and qualitative data as number of events (%).

Comparisons between patients with and without airway closure (before and after pneumoperitoneum) were performed with the Mann–Whitney test or the Fisher exact test, as appropriate. Intragroup comparisons between measurement in the supine and Trendelenburg position, before and after pneumoperitoneum, respectively, were conducted with the Wilcoxon rank sum test. Mean differences (95% CI) are displayed for most clinically significant results. Correlation between continuous variables was assessed with Spearman correlation: rho and the P value are reported. No correction was made for multiple comparisons.

Results with a two-tail P value less than 0.05 are considered statistically significant. Statistical analysis was performed with SPSS v. 20.0 (IBM Corp. Released 2011.IBMSPSS Statistics for Windows, Version 20.0, IBM Corp., USA). Manuscript figures were prepared with GraphPad Prism (GraphPad Software, USA).

Between May 2017 and June 2018, 50 patients were enrolled in the clinical trial. Airway closure with a variable degree of airway opening pressure after intubation in the supine position was diagnosed in 11 patients (22%), who were matched with 11 unaffected patients with similar body mass index who represent the control group. There were no missing data regarding the analyzed variables.

All patients received laparoscopic/robotic surgery in the Trendelenburg position (25 to 30°): surgical pneumoperitoneum was obtained by insufflation of carbon dioxide at constant pressure of 12 mmHg, which approximately corresponds to 16 cm H₂O.

Demographics and most relevant clinical characteristics are displayed in table 1 and were not different between patients with and without airway closure. Neither preoperative nor postoperative respiratory function was different between patients with and without airway closures, being within the physiologic range for most of them. No differences were found regarding preoperative or postoperative gas exchange. Intraoperative ventilator settings and hemodynamic parameters are displayed in table 2. Main results of the study are reported in table 3.

In patients with airway closure, after intubation in the supine position, pressure-volume curve of the airway pressure exhibited various degree of airway opening pressure (median value, 9 [interquartile range, 6 to 12] cm H₂O), with complete airway closure when airway pressure was below this threshold (representative ventilator tracings in fig. 1). Airway closure was confirmed on the pressure-volume curve by measured compliance between end-expiratory pressure and airway opening pressure, which corresponded to compliance of occluded respiratory circuit (2 ml/cm H₂O automatically measured by the ventilator during pre-use self-test; fig. 2, A, B, and C).

No patient in the control group developed airway closure after pneumoperitoneum induction. Conversely, in all patients with diagnosis of airway closure after intubation in the supine position, pneumoperitoneum in Trendelenburg position (25 to 30° head-down) yielded a rise in the airway opening pressure that reached a median value of 21 (interquartile range, 19 to 28) cm H₂O (representative patient in fig. 2C, left). The mean pneumoperitoneum-induced increase in airway opening pressure was 15 (95% CI, 11 to 18) cm H₂O (P < 0.001) and was consistent with pneumoperitoneum insufflation pressure (12 mmHg = 16 cm H₂O), and the mean increase in end-expiratory esophageal pressure that this generated, which was 9 (95% CI, 5 to 14) cm H₂O (P = 0.001).

Respiratory mechanics after intubation in the supine position were similar in the two groups, except for respiratory system linear compliance, which was slightly higher in patients with airway closure (table 3).

In both groups, surgical pneumoperitoneum yielded increases in inspiratory and expiratory esophageal pressure, airway plateau pressure, dynamic driving pressure of the respiratory system, respiratory system and lung static driving pressure, and tidal elastic pressure. These were caused by drops in static respiratory system, lung and chest wall compliance, and reduction in the respiratory system linear compliance (table 3).

In the control group, surgical pneumoperitoneum caused reduction in the linear compliance of the lung, which was accompanied by higher tidal lung elastic pressure. Conversely, in patients with airway closure, neither tidal lung elastic pressure nor lung linear compliance was modified by pneumoperitoneum (table 3; representative patient in fig. 2D).

In the supine position, end-expiratory lung volume was not different between groups (P = 0.511); differently, after institution of surgical pneumoperitoneum, patients with airway closure showed higher end-expiratory lung volume (P = 0.017). This occurred because in patients in the control group, but not in those with airway closure, gas insufflation for pneumoperitoneum reduced end-expiratory lung volume (P = 0.006 and P = 0.155, respectively; P = 0.045 for intergroup difference, fig. 3), with a mean decrease of 312 (95% CI, 118 to 505) ml.

As patients with airway closure experienced decreases neither in lung linear compliance nor in end-expiratory lung volume due to pneumoperitoneum, the pressure-volume curves of the transpulmonary pressure with patent airways (i.e., the transalveolar pressure) showed no changes in alveolar recruitment due to this procedure (representative patient in fig. 2D, left).

In the airway closure group, but not in the control group, static and dynamic respiratory system driving pressure overestimated respiratory system tidal elastic pressure (intergroup difference, both P = 0.007; fig. 4). Similarly, lung driving pressure was higher than lung tidal elastic pressure (P = 0.01). Correspondingly, lung dynamic, respiratory system static, and dynamic compliances were lower than the corresponding linear ones (P ≤ 0.05 for all). The entity of elastic pressure overestimation and compliance underestimation by dynamic measurements was proportional to the airway opening pressure (fig. 5).

Four patients with airway closure and six in the control group received scheduled recruitment maneuvers over the course of the trial. According to study protocol, these were performed after intubation, after pneumoperitoneum induction, and then on an hourly basis by switching to pressure-controlled ventilation (inspiratory pressure 10 cm H₂O, unchanged respiratory rate) and increasing PEEP by 5 to 10 cm H₂O every 4 to 6 breaths up to PEEP equals 30 cm H₂O. During recruitment maneuvers after pneumoperitoneum in patients with airway closure, pressure-controlled ventilation did not generate any tidal inflation until total inspiratory pressure overcame airway opening pressure, resulting in complete apnea. Conversely, in the control group, pressure-controlled ventilation always produced some tidal ventilation, irrespective of the applied pressure (fig. 6).

The results of the current study can be summarized as follows:

1. A relevant proportion of obese patients undergoing general anesthesia exhibit airway closure after intubation, with a variable degree of airway opening pressure. When pneumoperitoneum is combined with Trendelenburg position, airway opening pressure increases consistently with the rise in end-expiratory esophageal pressure and pneumoperitoneum insufflation pressure.

2. Airway closure can be suspected in the presence of total PEEP greater than set PEEP and particularly high plateau/driving pressure after pneumoperitoneum institution.

3. When pneumoperitoneum is instituted in the head-down position, airway closure increases end-expiratory alveolar pressure, maintaining transalveolar pressure at end expiration unchanged. This hampers any pneumoperitoneum-induced change in end-expiratory lung volume and lung compliance, finally preventing alveolar derecruitment.

4. With airway closure, static respiratory mechanics and dynamic measures performed by ventilators lead overestimation of lung and respiratory system driving pressures with underestimation of corresponding compliances. This effect, being proportional to the airway opening pressure, is already present after intubation in the supine position but is more pronounced as pneumoperitoneum is instituted.

5. When patients with airway closure receive pressure-controlled ventilation, no tidal volume is delivered if inspiratory pressure does not overcome airway opening pressure.

Airway closure with a variable degree of opening pressure has been documented in patients with acute respiratory distress syndrome, in human cadavers, and in cardiac arrest patients undergoing cardiopulmonary resuscitation.^1,4,23,24 

Different from what initially suggested during general anesthesia,^6,7  the opening pressure is not an airway pressure threshold to overcome pleural pressure and generate positive transpulmonary pressure. Indeed, this phenomenon has been reported in the excised lung, where transpulmonary pressure is necessarily positive,^25,26  and in our study, airway opening often occurred at a transpulmonary pressure lower than 0 cm H₂O (figs. 1, 2, and 6). Moreover, transpulmonary pressure is variable across the lung, both in patients with lung injury and (to a lesser extent) in those with healthy lungs, depending on the ventral-to-dorsal gradient in pleural pressure,¹⁴  while the airway opening we are reporting is an all-or-nothing phenomenon.

In the current study, we show that airway closure affects a relevant proportion of obese patients with no impairment in preoperative respiratory function who undergo general anesthesia. Because no patient without airway closure after intubation developed it as a consequence of pneumoperitoneum, its occurrence possibly depends on some anatomo-functional individual characteristic of patients’ airways, such as the surface tension in the liquid–gas interface, as already suggested.^1,25,27  In patients with airway closure, conversely, airway opening pressure was systematically increased by pneumoperitoneum in the head-down position: it is likely that the increase in pleural pressure caused by abdominal gas insufflation generates a drop in the transmural pressure of the airways, which finally yields airway closure. This is consistent with the hypothesized cause for airway collapse in spontaneously breathing patients with active expiration, highlighting the close relationship between this condition and the wider concept of expiratory flow limitation.^3,28  Importantly, in our patients with airway closure, the increase in the expiratory driving pressure (i.e., plateau pressure minus set PEEP) produced by pneumoperitoneum did not result in reduction in end-expiratory lung volume, which is itself a diagnostic criterion for certifying limited expiratory flow.³ 

Other authors reported respiratory mechanics compatible with airway closure during laparoscopic surgery in obese patients, showing that neither the “critical pressure” (i.e., the airway opening pressure) nor end-expiratory esophageal pressure was modified by surgical pneumoperitoneum and suggesting that this procedure could not act as a compression force for lungs and airways at end expiration.^6,7  Differently, our findings indicate that pneumoperitoneum increases esophageal pressure and airway opening pressure: in previous case series,^6,7  patients were studied in the supine or reverse Trendelenburg position, while our study shows pneumoperitoneum effects in head-down position. Thus, the Trendelenburg position plays a crucial role in enhancing the transmission of intraabdominal force to the pressure that is superimposed on the lungs (estimated by esophageal pressure¹⁴ ), which also squeezes airways contributing to determine the pressure required to maintain airway patency. It appears reasonable to suggest that any increase in the lung superimposed pressure (i.e., obesity, Trendelenburg position, alveolar flooding in acute respiratory distress syndrome, lung edema with chest compressions during cardiopulmonary resuscitation^1,24,29 ) can enhance airway closure by determining a decrease in the functional residual capacity below the closing capacity (i.e., sum of residual and closing volume) of the airways.³⁰ 

Our results on end-expiratory lung volume demonstrate that, in the case of airway closure, pneumoperitoneum does not cause alveolar derecruitment. This confirms that the airway opening pressure does not reflect alveolar kinetics and alveolar collapse, as the “lower inflection point” concept would suggest.²⁵  In our patients, the increase in airway opening pressure caused by pneumoperitoneum, being itself proportional to pneumoperitoneum pressure, prevented alveolar collapse when this external load was applied. Essentially, the rise in airway opening pressure induced by pneumoperitoneum generated consistent changes in alveolar pressure at end expiration, which kept the transalveolar pressure unmodified and prevented any lung volume loss. Accordingly, while in control patients pneumoperitoneum yielded decreases in end-expiratory lung volume and lung linear compliance, the pneumoperitoneum-induced decrease in respiratory system linear compliance detected in patients with airway closure was due to a more rigid chest wall, with unchanged lung linear compliance and lung tidal elastic pressure.^31,32  Unrecognized increases in airway opening pressure and end-expiratory alveolar pressure may explain why obese patients are less prone to experience reduction in functional residual capacity due to pneumoperitoneum in Trendelenburg position.³³ 

Our results carry relevant clinical consequences. First, under this condition, driving pressure, measured both with occlusions and dynamically by the ventilator, relevantly overestimates tidal elastic pressure. Similarly, dynamic compliance does not accurately reflect actual respiratory system compliance.^1,13  These aspects appear relevant in the intraoperative setting, as the driving pressure has been shown to be the final mediator of the effects of ventilation settings on the rate of postoperative complications,¹⁵  and optimization of dynamic respiratory system compliance has been advocated as a tool to titrate PEEP in the setting of general anesthesia.¹⁷ 

Second, with airway closure, real alveolar pressure at end expiration is unknown but can be very high and, ultimately, close to the airway opening pressure. Hence, alveolar pressure at end expiration is independent from applied PEEP if this is set below airway opening pressure. This should be taken into account when assessing the effects on obese patients of PEEP-setting strategies, such as those currently under evaluation in ongoing randomized studies during general anesthesia.^34–40  Of note, when PEEP is set below airway opening pressure, distal airways suffer from cyclic opening and closing, and this may yield bronchiolar injury.⁴¹  Whether this may clinically jeopardize the population of surgical patients is unknown. Moreover, the high occult alveolar pressure, while preventing alveolar derecruitment, can generate detrimental hemodynamic effects,^42–44  whose mechanisms in the intraoperative setting could be misinterpreted.

Third, pneumoperitoneum in the head-down position can impair cerebral blood flow homeostasis,⁴⁵  with possible abnormal increases in intracranial pressure.⁴⁶  Whether and to what extent the high alveolar and intrathoracic pressure caused by airway closure may contribute to this is unknown and warrants further investigations.

Most important, airway closure is difficult detect in the clinical setting (i.e., low-flow inflation is needed), occurred in patients with no apparent impairment in respiratory function, and determined alarmingly high airway opening pressure values. This raises concerns regarding the safety of pressure-controlled ventilation in this setting, even though it has been proposed with encouraging results and is used worldwide.^47–50  With airway closure, use of pressure-controlled modes with inspiratory pressure lower than airway opening pressure results in complete apnea. Furthermore, if airway opening pressure varies over time, use of the same inspiratory pressure may yield relevant hypo-/hyperventilation.

Our study has limitations. First, the study was not designed to assess the presence of airway closure in morbidly obese patients undergoing general anesthesia, but rather to determine the effects of ventilatory interventions on postoperative oxygenation. Hence, our conclusions are the result of a serendipity process. This should not alter, and could even strengthen, the significance of the current investigation, especially because measurements were performed according to a prespecified protocol, patients represent a homogeneous and externally reproducible population, and airway closure was documented with different PEEP levels and tidal volume sizes.

Second, reported data are limited to 11 cases, and all patients were females, which cannot ensure the reproducibility of the findings among male patients. However, previous investigations in different settings^1,4,23  clearly showed that males may be similarly affected.

Finally, our results should be seen as a “proof of concept” study, and we acknowledge that the low number of studied patients did not allow correction for multiple comparisons: we recognize it is not possible to draw conclusions from our results on the risk factors, exact prevalence, and eventual time course of this phenomenon or on its consequences on the postoperative period. Recently, expiratory flow limitation has been shown to affect up to 30% of the patients undergoing general anesthesia and has been linked to the risk of pulmonary complications⁵¹ : to what extent development of airway closure contributes to this remains to be established.

Airway closure unpredictably affects a considerable proportion of obese patients undergoing general anesthesia. Pneumoperitoneum in the Trendelenburg position relevantly worsens this phenomenon, enormously increasing airway opening pressure and alveolar pressure at end expiration; this yields misinterpretation of respiratory mechanics and underestimation of end-expiratory alveolar pressure, which is markedly higher than measured PEEP, possibly equals airway opening pressure, and prevents alveolar derecruitment. Airway closure can be missed at the bedside, while this phenomenon generates a pressure threshold to inflate the lung that can reach high values; this raises concerns about the safety of pressure-controlled modes in this specific setting.

The authors are thankful to anesthesiologists, surgeons, nurses, scrub nurses, and all operating room personnel from their institution for the kind assistance and warm support during the study. The authors are grateful to Laurent Brochard, M.D., Lu Chen, M.D., (St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada), Salvatore Maurizio Maggiore, M.D., Ph.D., (University of Chieti, Italy), and Carlo Alberto Volta, M.D., and Savino Spadaro, M.D., (University of Ferrara, Italy) for their valuable suggestions for data analysis and interpretation.

This work was supported by Catholic University of The Sacred Heart and by Fondazione “Policlinico Universitario Agostino Gemelli IRCCS” (Rome, Italy). Dr. Grieco is supported by grants from the Italian society of Anesthesia, Analgesia, Pain and Intensive Care Medicine (SIAARTI; Rome, Italy)/Merck Sharp and Dohme (Kenilworth, New Jersey) and the European Society of Intensive Care Medicine (ESICM; Brussels, Belgium; 2017 Bernhard Drager Award for Advanced Treatment of Acute Respiratory Failure).

Dr. Grieco has received payments for travel expenses fromMaquet (Rastatt, Germany), Getinge (Getinge, Sweden), and Air Liquide (Paris, France). Dr. Antonelli has receivedpayments for board participation from Maquet, Air Liquide, and Chiesi (Parma, Italy). Drs. Antonelli and Grieco disclose an ongoing research grant received by General Electric Healthcare (Chicago, Illinois). The other authors declare no competing interests.