Lung volumes and lung volume recruitment in ARDS: a comparison between supine and prone position

Patients

Patients were considered eligible for the study if they met the Berlin definition criteria for ARDS [22] and had an indication for PP in accordance with our department’s protocol (PaO₂/FiO₂ ratio of < 150 mm Hg and FiO₂ of ≥ 0.6 with PEEP of at least 5 cm H₂O). We recommend to use protective ventilation with individualized low tidal volume (Vt) and moderate PEEP levels. Essentially, PEEP is titrated according to the gas exchange (Sat O₂, measured by pulse oxymeter, around 95%) with end-inspiratory plateau airway pressure (Pplat) not higher than 28 cm H₂O and without hemodynamic instability (mean arterial pressure above 65 mm Hg and no need for fluid replacement). Our detailed ventilatory strategy is included in Additional file 1. Hence, all our patients had been turned in PP before inclusion in the study. To be included, patients had to present an improvement in gas exchange (FiO₂ ≤ 0.6 and PEEP ≤ 12 cm H₂O) in SP in order to avoid severe hypoxemia because of the derecruitment (induced by PEEP withdrawal and ventilation at ZEEP) during the measurement of FRC. Exclusion criteria were: age < 18 years, tracheostomy, pregnancy, major trauma, barotrauma (presence of extra-alveolar air during mechanical ventilation as assessed by daily chest X ray) and hemodynamic instability (systolic blood pressure < 80 or > 160 mm Hg, heart rate < 50 bpm or > 130 bpm or changes in ± 20% from baseline).

All patients were under continuous sedation and analgesia with intravenous perfusion of midazolam and/or propofol and opioids. During the study period, all patients received neuromuscular blocking agents.

Protocol

The following data were collected: age, height, simplified acute physiology score III at admission, ARDS etiology, days of mechanical ventilation, intensive care unit outcomes, respiratory rate, Vt, PEEP, peak airway pressure, Pplat and arterial blood gases. Respiratory variables were recorded directly from the ventilator.

All patients were ventilated in volume control ventilation using the same ventilator model (Engström Carestation ICU ventilator, General Electric, Madison, WI, USA).

To avoid hypoxemia, defined as oxygen saturation ≤ 88% measured through pulse oximetry, we increased the FiO₂ to 0.8 1 h before starting the protocol.

Measurements

Baseline ventilatory and hemodynamic parameters were collected before the protocol to measure lung volumes. The same procedures were carried out in SP and PP and are outlined below (see also Fig. 1):

1. 1.

   Measurement of end-expiratory lung volume (EELV): EELV is the resting end-expiratory lung volume measured at baseline PEEP.

    
2. 2.

   Removal of PEEP and continuation of mechanical ventilation at ZEEP. This derecruitment maneuver is mandatory to conduct the following step 3, and it is the reason to increase the FiO₂ to 0.8 immediately before starting the protocol (i.e., to avoid hypoxemia).

    
3. 3.

   Measurement of functional residual capacity (FRC): FRC is the resting lung volume measured at ZEEP.

    
4. 4.

   Measurement of the Vt, delivered from ZEEP, that generated a Pplat equal to the basal PEEP. This step (see Fig. 1) is mandatory to allow a proper estimation of the PEEP-induced lung volume recruitment [19, 20, 23–25].

    

Once step 4 was completed, the same PEEP that was used at baseline was resumed.

Measurements at ZEEP (FRC and Vt delivered from ZEEP, that generated a Pplat equal to the basal PEEP) included a lung derecruitment maneuver (PEEP removal) that can produce hypoxemia. For the purpose of our investigation, we defined hypoxemia as oxygen saturation ≤ 88% measured through pulse oximetry.

Lung volumes (EELV and FRC) were measured twice using the nitrogen washout/washin technique available in Engström Carestation ICU ventilator as previously described [24, 26]. Washout/washin technique is a multiple breath maneuver that with a modification of 0.1 in FiO₂ calculates the residual nitrogen in the lung (assuming there is not exchange of nitrogen) by continuous measurements of oxygen and carbon dioxide. The ventilator was carefully calibrated before the measurements according to the manufacturer’s specifications. We obtained four values for each lung volume. The mean of the four values was used. As previously suggested [27], patients were excluded if the differences between the four values were more than 20% (cutoff determined by the manufacturer).

After the procedures in SP, the patients were turned to PP and 1 h later the same procedures (from 1 to 4 above) were made in PP. This time span was based in previous data showing that after 1 h in PP gas exchange is stable in the majority of patients [28, 29]. If a patient presented with hypoxemia (oxygen saturation ≤ 88%) at any time during the protocol, the measurements were aborted and the patient was excluded.

The normal reference values for FRC (liters) in the SP were calculated according to the equation described by Ibáñez and Raurich [30], as follows: 5.48 × height—7.05 for men and 1.39 × height—0.424 for women; height units are in meters. Compliance (ml/cm H₂O) was calculated as Vt/(Pplat minus total PEEP), being total PEEP the sum of PEEP plus intrinsic PEEP. Predicted body weight was calculated as follows: 50 + 0.91(height—152.4) for men and 45.5 + 0.91(height—152.4) for women; height units are in centimeters. Driving airway pressure was calculated as the difference between Pplat and total PEEP [31].

Calculation of lung volumes and strain

Statistical analysis

Data are expressed as mean ± SD. We used Wilcoxon test to compare variables between supine and prone positions and U the Mann–Whiney test to compare early and non-early ARDS patients. A p value < 0.05 was considered statistically significant. The SPSS^® Statistics (version 20.0, Chicago, IL, USA) statistical software was used for statistical analysis.

Prone position, oxygenation and lung volumes

In ARDS patients, lung volumes at ZEEP (FRC) and at PEEP (EELV) are typically decreased [18]. Two previous studies have shown that PP significantly increases FRC in ARDS patients [15, 16]. Nevertheless, data about the changes in EELV with the change from SP to PP in ARDS patients are not consistent. Four previous studies have shown that PP increases EELV in ARDS patients as compared to SP [14–17], but another study [4] found that the change of EELV from SP to PP was not significant. These contradictory findings might be explained by differences in lung recruitability, distribution and extension of lung volume alterations, differences in chest wall compliance, the influence of abdominal weight and heart compression, the inclination from the horizontal plane and the use or not of ventral supports [3, 9, 34, 35].

In the present study, we found a 40% decrease in FRC as compared to its reference value in SP, confirming previous results [18]. We also observed that the FRC and EELV increased significantly with the change of position (18% in FRC and 17% in EELV). Santini et al. [7] performed a study in animals with normal lungs, and they found a significant increase in FRC with the change from SP to PP. The increase in resting lung volume was mainly related to a redistribution of aeration: a minor decrease in non-aerated lung tissue (3%), a major decrease in poorly aerated tissue (17%) and a major increase (20%) in well-aerated tissue. Since recruitment, as precisely measured by thoracic CT scan, refers to tissue recruitment (i.e., amount of non-inflated tissue that reinflates at a higher pressure), the decrease in poorly aerated tissue and the increase in well-aerated tissue (which contribute to the end-expiratory lung volume increase induced by PEEP) are thus considered as better gas distribution within the lung and not recruitment per se [36].

Prone position and strain

During passive mechanical ventilation, the force applied by the ventilator generates an internal tension in the fibers of the lung skeleton, called “stress,” and the elongation of these fibers from their resting position is called “strain” [2]. High values of dynamic lung strain (lung deformation caused by Vt) and static lung strain (lung deformation caused by PEEP) are associated with ventilator-induced lung injury [32, 37].

In an animal model, Protti et al. [33] showed that for the same global strain, a large static strain is less harmful than a large dynamic strain. On the same vein, González-López et al. [38] found that increased strain was associated with a proinflammatory lung response in patients with acute lung injury. Moreover, Bellani et al. [39] found in patients with acute lung injury that the intensity of metabolic activity (a surrogate of inflammation) detected by positron emission tomography was correlated with regional strain. Consequently, the significant decrease in dynamic strain in PP as compared to SP could be another mechanism of protection of PP against ventilator-induced lung injury. Therefore, the measurement of lung volumes at bedside may be an important tool to deliver a more physiologically based ventilation and encourage physicians to increase the use of PP in moderately to severe ARDS patients [40].

Prone position and PEEP-induced lung volume recruitment

It is still unclear whether the PEEP-induced alveolar recruitment varies with the change from SP to PP. In an experimental study in animals with lung injury, Richard et al. [5] analyzed the variation of alveolar recruitment at PEEP 10 cm H₂O in SP and PP by means of the positron emission tomography technique. They found that in PP, PEEP-induced alveolar recruitment was not higher than in SP. Interestingly, in this study, the authors observed a redistribution of densities in PP (recruitment in dorsal regions with derecruitment in ventral regions). Cornejo et al. [6] performed another study in ARDS patients to determine the effects of PEEP and PP on alveolar recruitment. Using the CT scan technique, they found that increasing PEEP from 5 cm H₂O to 15 cm H₂O significantly increased alveolar recruitment. However, the percentage of recruitment was similar in both positions (36% in SP and 33% in PP). Using a different methodology, the data from our study are consistent with these findings, indicating that the effects of PEEP on lung volume recruitment are similar in both positions (around 17% of EELV).

A previous study by Grasso et al. [41] found that alveolar recruitment was higher in the early phase (1 ± 0.3 days of mechanical ventilation) than in the late phase of ARDS, but a subsequent study by Gattinoni et al. [42] did not find the same results. In the study of Gattinoni et al. [42], they found that the number of days of mechanical ventilation before the study was similar in patients with a lower percentage of potentially recruitable lung and those with a higher percentage (5 ± 6 vs. 6 ± 6 days, respectively, p = 0.50). In our study when we classified ARDS patients in the early phase (< 72 h) and in late phase (> 72 h) (Additional file 2: Table S2), we observed results similar to those of Gattinoni et al. [42]: no statistical differences in lung volume recruitment between the early and late phase group were detected. In our study, however, in the early phase of ARDS, lung volumes increased and strain decreased with the change from SP to PP, whereas in late phase ARDS we did not observe these findings (Additional file 2: Table S2). These differences could be related to the presence of some degree of hydrostatic pulmonary edema in the early phase of ARDS, and to the presence of fibrosis in the non-early phase of ARDS that predisposes to non-responsiveness to PP in terms of increasing in lung volumes and decreasing strain [43]. Our findings thus suggest that the survival benefit may, in part, be related to the early application of PP as it increases resting lung volumes and decreases lung strain compared to SP. It is also tempting to speculate that the lack of differences in Vrec between supine and prone, and the increase in overall lung volume in prone as compared to supine, can be explained by a decrease of poorly ventilated areas and an increase of well ventilated areas, which in turn might help to decrease lung inhomogeneity. It has been shown that the extent of lung inhomogeneities (as quantified by the amount of poorly ventilated tissue) is associated with worse outcomes in ARDS patients, possibly due to a mechanism of “stress raisers” [44].

Limitations

Like many physiological studies [4, 6, 14–17, 34], our study has a relatively low number of patients. Another limitation is that the measurement of FRC could be subject to the tolerance to PEEP removal and the FiO₂ used. However, when the study was performed, all the patients met the criteria for mild–moderate ARDS according to the Berlin definition [22]. We did not perform a multi-slice spiral lung computed tomography to measure the quantitative changes in alveolar aeration induced by PEEP and PP. Other measurements of lung mechanics (i.e., esophageal pressure and derived variables) and lung biomarkers could help to further explain the effects of PEEP and positioning in ARDS patients, but we did not do these because of lack of adequate equipment at the time of the study. Finally, to confirm the changes, it might have been useful to return the patients from PP to SP and to repeat the same procedures and measurements; this was not done, however, because most patients remained in the PP as per clinical decision after the study had been completed.