Introduction
Obese patients with COVID-19 are at greater risk of requiring mechanical ventilation and developing ARDS.1 Obesity is characterized by an increased pleural pressure due to higher thoraco-abdominal loading, which reduces end-expiratory lung volume, and causes atelectasis and airway compression.2 Consequently, it is usually recommended to use high PEEP to improve gas exchange and avoid atelectrauma.2 However, inadequately high PEEP may cause overdistention and lung injury together with increased dead space, reduced cardiac output, and ultimately decreased oxygen delivery.3 Whereas clinicians often target oxygenation for PEEP titration, identifying more relevant bedside tools to guide PEEP selection may help to decide whether high PEEP might be deleterious. We assessed whether selecting subjects based on response in airway driving pressure (ΔP) when setting high PEEP would indicate potential physiological benefits that would otherwise not be recognized by the oxygenation response. The study was performed in a relatively homogeneous population of obese subjects with COVID-19 ARDS.
Methods
We conducted a physiological study including consecutive, sedated, and curarized obese (body mass index [BMI] > 30 kg/m2) subjects with moderate-severe COVID-19 ARDS ventilated in volume control with tidal volume (VT) of 6 mL/kg of predicted body weight to keep plateau pressure (Pplat) < 30 cm H2O and ΔP < 15 cm H2O, with 0.3 s of end-inspiratory pause and breathing frequency to achieve pH 7.20–7.45. We evaluated the respiratory system mechanics using a specific device (FluxMed, MBMed, Buenos Aires, Argentina). We inserted an esophageal balloon catheter (VA-A-008, MBMed) filled with 0.5 mL of air that correct position was confirmed accordingly.4 Expired CO2 was measured with a monitor previously validated (FluxMed, MBMed) that combines a mainstream Capnostat (Capnostat 5, Philips, Amsterdam, the Netherlands) and a specific software. After a stepwise recruitment maneuver (maximum airway pressure = 40 cm H2O), we assessed variables related to oxygenation, dead space, respiratory mechanics, and hemodynamics at 2 PEEP levels: PEEPlow = 5 cm H2O or at airway opening pressure and PEEPhigh = PEEPlow + 10 cm H2O; 2-s end-expiratory occlusions were performed to rule out auto-PEEP. Airway opening pressure was assessed as previously described.5 To minimize fluctuations in lung volumes and measure recruitability, PEEPhigh was always the first condition evaluated. Each PEEP phase lasted 10 min, which has been shown to be enough time to assess changes in gas exchange.3-6 The ΔP was calculated as Pplat − PEEPtotal using 0.3 s end-inspiratory pause. We divided subjects into groups based on changes in ΔP, considering as responders those in whom ΔP was lower at PEEPhigh and non-responders otherwise. The study protocol was approved by the institutional review board (IRB code #16/2020) and registered in ClinicalTrials.gov (NCT04486729).
Continuous variables were reported as mean (SD) or median [25th–75th percentile] according to normality of data checked through Shapiro-Wilk test. Categorical variables were reported as frequencies and percentages. Variables between PEEP levels were compared using t test or Wilcoxon test, without P value correction for multiple comparisons. Correlations were performed with Spearman rank correlation. Results with 2-tailed P ≤ .05 were considered statistically significant. Data were analyzed with R software (R Foundation for Statistical Computing, Vienna, Austria).
Results
We included 15 subjects (55.6 [9.4] y old; BMI 36.6 [31.6–40.3] kg/m2). Baseline characteristics are shown in Table 1. All subjects were evaluated at PEEPlow = 5 cm H2O, except for one individual in the responder group (airway opening pressure = 12 cm H2O). In 8 subjects, ΔP decreased (responders), whereas in 7 subjects ΔP increased (non-responders) at PEEPhigh (Fig. 1-A).
Effects of PEEP on respiratory and lung mechanics and gas exchange according to changes in airway driving pressure (ΔP). A: Subjects were divided as responders and non-responders based on the reduction or increase of ΔP at high PEEP. B: The changes in ΔP with higher PEEP levels were very similar to the changes in driving transpulmonary pressure (ΔPL). C: Despite the elastance-derived end-inspiratory transpulmonary pressure (PLinsp-ED) augmented in both groups, non-responders showed a more pronounced increase in total lung stress, whose values exceeded safety limits in almost all participants. D: A good correlation between changes in PLinsp-ED and change in ΔP was observed. E: The changes in oxygenation showed significant improvement only in responders and correlated overall with change in ΔP. E and F: Wilcoxon sum-rank tests were used for comparison in PLinsp-ED within responders and PaO2/FIO2 within non-responders between 2 PEEP levels. T tests were used for all other comparisons.
Baseline Characteristics of Subjects
The effect of PEEP on ΔP was not explained by changes in chest wall compliance and was entirely explained by a comparable response in transpulmonary ΔP (Fig. 1-B). The BMI significantly correlated with change in ΔP (R2 = 0.37, P = .02); that is, the higher the BMI, the greater the ΔP reduction at PEEPhigh. The elastance-derived end-inspiratory transpulmonary pressure (PLinsp-ED) increased in all subjects at PEEPhigh, but non-responders showed a significantly higher increment, and the increase in PLinsp-ED was proportional to change in ΔP (Fig. 1-C; 1-D). The mean (SD) end-expiratory transpulmonary pressure at PEEPhigh increased in both groups (non-responders: PEEPhigh 4.4 [3.4] cm H2O vs PEEPlow −2.1 [3] cm H2O, P < .001, t test; responders: PEEPhigh 2.4 [3.1] cm H2O vs PEEPlow −4.3 [3.7] cm H2O, P < .001, t test).
In 13/15 subjects, PaO2/FIO2 was higher at PEEPhigh. However, mean oxygenation improved significantly only in responders (Fig. 1-E). The change in ΔP was correlated with the oxygenation response (Fig. 1-F), and the functional shunt estimated using central venous gases was reduced only in responders (non-responders: PEEPhigh 0.39 [0.34 to 0.45] vs PEEPlow 0.47 [0.37 to 0.56], P = .28; responders: PEEPhigh 0.3 [0.26 to 0.33] vs PEEPlow 0.44 [0.37 to 0.47], P = .006).
The change in Enghoff dead space correlated with change in ΔP (R2 = 0.26, P = .046). Finally, the PEEP level did not modify arteriovenous CO2 gradient and central venous O2 saturation in neither of the groups.
Discussion
Even though the majority of obese subjects with COVID-19 ARDS attained better oxygenation with high PEEP, they had a heterogeneous response in terms of ΔP and, therefore, risk of lung injury. Stratifying subjects according to change in airway ΔP between 2 PEEP levels allows the identification of subjects in whom using a higher PEEP strategy may provide some lung protection versus those in whom the risk of overdistention is potentially higher. Oxygenation response does not allow this individual separation.
Previous literature has argued about the utility of using high PEEP in COVID-19 ARDS. The predominant vascular involvement together with the scarce loss of lung volume has been proposed as possible causes of a poor response to high PEEP.7 In contrast, the mechanical consequences of obesity (ie, greater superimposed pressures stem from higher end-expiratory pleural pressure causing atelectasis) might determine a better response to high PEEP. In this scenario, it is likely that in obese patients with COVID-19 ARDS the response to high PEEP may depend on the balance between these opposite effects.
In obese subjects, the need for PEEP may be related to the weight of the chest wall and may differ from classical assessment of recruitment. The changes in ΔP were explained by similar changes in transpulmonary ΔP. Although the reduction in ΔP might be, at least in part, due to tidal recruitment, the end-expiratory transpulmonary pressure closest to zero (indirectly representing that atelectrauma is locally unlikely) was attained at the PEEP level that determined the lowest ΔP in both groups. Interestingly, 2 recently published studies reported better outcomes when end-expiratory transpulmonary pressure was close to zero.8-9 However, esophageal pressure is rarely used, even in severe ARDS, because of the technical challenges it represents. Based on our results, assessing the change in ΔP during a PEEP-step might be less time consuming and serves as a noninvasive first approximation about which PEEP level might be closer to the optimal end-expiratory transpulmonary pressure ranges in this population.
The changes in ΔP presented a significant correlation with BMI; that is, the higher the obesity grade, the greater the benefits of high PEEP. In line with these results, Thille et al10 recently reported a better response to noninvasive ventilation in obese-overweight compared with normal-underweight individuals. These findings suggest that the more obese patients might take greater advantages of applying high positive pressure. However, even though all our sample fulfilled obesity criteria, in non-responders, an increase in ΔP was associated with an increase in transpulmonary ΔP and in PLinsp-ED often > 20 cm H2O (markers of higher dynamic and total lung stress, respectively), suggesting that the potentially deleterious effects of high PEEP might outweigh the benefits.11 Additionally, despite better oxygenation in most non-responder (5/7 subjects) at high PEEP, the improvement in functional shunt was not significant, suggesting that lung overdistention was possibly predominant. Beloncle et al12 observed a similar response in poorly recruitable subjects with COVID-19 ARDS. Consistent with these data, Rodriguez et al13 previously showed that oxygenation often continues improving at higher PEEP even when there is increasing overdistention. Additionally, a recent study has suggested that reduced ΔP following protocolized ventilator changes was more strongly associated with lower mortality than increased PaO2/FIO2.14 Our data in a specific subpopulation give some clues about why this may be the case.
Our study has some limitations: Subject selection was based on a unique and increasingly relevant population (obese subjects with ARDS), but the study was limited by the small sample size.15 Furthermore, PEEP selection strategy was not individualized; consequently, it was not possible to explore the response to lower PEEP changes. In addition, we did not evaluate the ΔP response to PEEPhigh using lower VT ranges; and only 3 subjects had severe COVID-19 ARDS, which could have determined the low number of responders. We did not measure cardiac output, an important variable affecting oxygenation and dead-space indices. Nonetheless, indirect variables such as mixed venous oxygen saturation and arterial-to-venous CO2 difference remained unchanged with PEEPhigh. Our study was not able to assess the need to strictly maintain a Pplat < 30 cm H2O; only 2 subjects exceeded this threshold, with PLinsp-ED within target range in one but dangerously high in another. Future studies should explore the need to strictly adjust Pplat < 30 cm H2O. Finally, our study does not aim to describe an individualized PEEP selection approach; consequently, after identifying patients who might potentially benefit from higher PEEP, a complementary PEEP-titration strategy should be implemented to accurately identify the best PEEP.
In summary, even though most subjects had improved oxygenation with higher PEEP, only those in which ΔP decreased at PEEPhigh experienced a decrease in transpulmonary ΔP and a minimal increase in PLinsp-ED. Changes in ΔP during a PEEP-step may provide clinicians useful information about the benefits and risks of using high PEEP in obese patients with COVID-19 ARDS.
Footnotes
- Correspondence: Laurent Brochard MD, Keenan Research Centre and Li Ka Shing Institute, Department of Critical Care, St Michael’s Hospital, Toronto, Ontario, Canada. E-mail: Laurent.Brochard{at}unityhealth.to
The authors have disclosed no conflicts of interest.
- Copyright © 2023 by Daedalus Enterprises