Dynamic hyperinflation and intrinsic positive end-expiratory pressure in ARDS patients

ARDS patients enrolled in seven previous studies [20‐26] were analyzed. The institutional review board of each hospital approved each study, and written consent was obtained according to the regulations applicable in each institution.

Patients were retrospectively classified according to the Berlin criteria [27] for five studies [20, 22‐25] and according to the American European Consensus Conference for two studies [21, 26]. The median time between diagnosis and enrollment was 2 days [1‐5].

Exclusion criteria were age less than 16 years, pregnancy, asthma or chronic obstructive pulmonary disease (COPD) reported in the patient’s medical history, and hemodynamic instability.

Patients were classified according to the presence or absence of intrinsic PEEP at the end of an expiratory pause (3–5 s) when any total PEEP higher than an external PEEP of 5 cmH₂O was detected.

Patients were supine, orally intubated, sedated, and paralyzed, with a bolus of vecuronium [23, 24, 26] or rocuronium every 20 min [22] for the duration of the study, and thereafter according to clinical judgment. Patients were mechanically ventilated in volume control mode, and all measurements were recorded in volume control mode.

The tidal volume was set at 6–8 mL/kg of predicted body weight with a respiratory rate to maintain an arterial pH of 7.30–7.45. The inspiratory time was set at 33% of the total respiratory time using a square-wave inspiratory flow.

A recruitment maneuver was done to standardize lung volume history, in pressure-controlled ventilation at PEEP 5 cmH₂O, with a plateau pressure of 45 cmH₂O, I:E 1:1, and respiratory rate 10 breaths/min for 2 min [18]. Then, from a minimum of 20 min to a maximum of 60 min after the recruitment maneuver, the same tidal volume, respiratory rate, and oxygen fraction were applied and a PEEP trial (at PEEP 5 and 15 cmH₂O) was done. At both PEEP levels, respiratory mechanics were measured, and blood gas analyses were done after 20 min. The external PEEP and total PEEP were measured at the end of the expiration of a regular breath and at the end of an expiratory pause (3–5 s). Intrinsic PEEP was defined as the total PEEP minus the external PEEP [4, 28].

Respiratory mechanics data were acquired with dedicated transducers. Flow at airway opening was measured with a heated pneumotachograph (Fleish 2, Lausanne, Switzerland). Airway pressure was measured proximally to the endotracheal tube with a pressure transducer (MPX 2010 DP. Motorola, Solna, Sweden). In all patients, a heat and moisture exchanger (DARTM, Convidien™, Adult-Pediatric) was placed between the endotracheal tube and the pneumotacograph.

A radio-opaque esophageal balloon (SmathCath Bicore, USA) was positioned in the lower third of the esophagus 35–40 cm from the nose. Esophageal pressure was measured by connecting the balloon, inflated with 1.0–1.5 mL of air, to a pressure transducer [29]. All traces were sampled at 100 Hz and processed on a dedicated data acquisition system (Colligo and Computo).

During inspiratory and expiratory pauses, the static airway and esophageal pressures were measured. The resulting physiological variables were computed according to standard formulas (see Additional file 1).

Arterial blood gases were analyzed. The alveolar dead space was computed measuring the end-tidal expired partial pressure of carbon dioxide with a CO₂SMO monitor; the physiological dead space was computed according to the Enghoff modification of Bohr’s equation, measuring the mixed expired partial pressure of carbon dioxide with a CO₂SMO monitor (Novametrix, Wallingford, UK) [30]. (see Additional file 1).

After the PEEP trial, patients were moved to the radiology department and two whole-lung computer tomography (CT) scans were taken, after a recruitment maneuver. During an end-expiratory pause at 5 cmH₂O of PEEP and an inspiratory pause at 45 cmH₂O of plateau pressure. Lung CT scans were taken using the following parameters: 110 mAs, tube voltage 120 kV, rotation time 0.5 s, collimation 128 × 0.6 mm, reconstruction matrix 512 × 512. In each CT slice, lung profiles were outlined manually and analyzed with a dedicated software package (Soft-E-Film, www.​softefilm.​eu). The total lung gas volume, weights, and amounts of the different compartments (not inflated, poorly inflated, well inflated, and over-inflated) were calculated as previously described [31].

Lung recruitability was computed as the ratio of the difference between not-inflated tissue at 5 cmH₂O of PEEP and that at 45 cmH₂O of plateau pressure to the total lung tissue at 5 cmH₂O of PEEP [21].

Data were not normally distributed, and non-parametric methods were applied. When not otherwise specified, categorical data are reported as frequencies and percentages, while continuous data are presented as medians [interquartile range]. Baseline characteristics of the patients with and without intrinsic PEEP were compared by the Mann-Whitney rank sum test.

Two-way repeated measures analysis of variance (ANOVA) followed by all pairwise multiple comparison procedures (Holm-Sidak method) was applied to investigate the effects of intrinsic and external PEEP on respiratory mechanics, CT data, and gas exchanges.

The intrinsic PEEP significantly on raising PEEP from 5 to 15 cmH₂O (1.1 [1.0–2.3] vs 0.6 [0.0–1.0] cmH₂O; p < 0.001).

Among these patients with intrinsic PEEP at 5 cmH₂O, when external PEEP was raised, 89% presented a decrease in intrinsic PEEP with a median of − 1 [− 2.00 to − 0.96], 7% showed an increase (0.6 [0.17–1.25] cmH₂O), and only 4% had no change. Eight of the patients without intrinsic PEEP (6%) developed a small amount of intrinsic PEEP while the remaining 94% did not (Fig. 1).

At both 5 and 15 cmH₂O of PEEP, the applied tidal volume was significantly lower in patients with intrinsic PEEP (480 [430–540] vs 520 [445–600] mL at 5 cmH₂O; 480 [430–540] vs 510 [430–590] at 15 cmH₂O) while the respiratory rate was significantly higher (18 [15–20] vs 15 [13–19] bpm at 5 cmH₂O; 18 [15–20] vs 15 [13–19] bpm at 15 cmH₂O) (Table 2).

Lung compliance was significantly higher in patients with intrinsic PEEP (p = 0.004) while the airway plateau pressure and driving pressure were not different (Table 2 and Additional file 1: Table S1).

Expiratory time was lower than 3 time constants in 56 patients (26%) out of the whole population; this happened in 30 (34%) patients with and 26 (20%) without intrinsic PEEP. The expiratory time was lower in patients with intrinsic PEEP at both PEEP levels while total airway resistance and compliance of the respiratory system were not different in patients with and without intrinsic PEEP (Table 2).

Table 3 shows the CT data at 5 and 45 cmH₂O of airway pressure. The total lung gas volume was no different in the two groups at the two pressures. In both groups, the percentages of not inflated and poorly inflated tissue were similarly reduced while the well-inflated tissue increased from 5 to 45 cmH₂O of airway pressure.

Lung recruitability was not different among patients with and without intrinsic PEEP (15 [0–32] % vs 22 [0–36] %).

Pulmonary gas exchange data are presented in Table 4. PaO₂ was similar between patients with and without intrinsic PEEP at 5 cmH₂O (71 [64–83] mmHg vs 73 [63–93]) and similarly increased at 15 cmH₂O (94 [75–124] mmHg vs 99 [80–123] mmHg). PaCO₂ and physiological and alveolar dead space did not differ between the two groups (Table 4 and Additional file 1: Table S2).

Regarding respiratory mechanics, only lung compliance was significantly different but not to any clinically relevant extent. In patients without muscle activity, deflation of the respiratory system is completely passive and the entire force for exhalation derives from the elastic potential energy stored in the respiratory system during the inspiration [36]. Taking into account the equation of motion of the respiratory system, the time required for complete exhalation depends on the product of total resistance and compliance, which is the time constant of the respiratory system [37]. Assuming a homogeneous lung compartment model with only one time constant, where the expiratory time equals the product of resistance for compliance, the volume that will remain in the respiratory system should be approximately 37% of the inspired volume [36]. Three time constants are required to reach a 95% volume exhalation [38, 39]. Thus, for a given applied tidal volume, if there is a significant increase in compliance or resistance, it will take longer to reach the resting volume. ARDS patients, characterized by reduced compliance [1], should be theoretically protected from developing dynamic hyperinflation [28, 39].

On the other hand, an increase in resistance can promote the development of dynamic hyperinflation. In the present study, the product of compliance and resistance was not different (i.e., time constant) in the two groups, suggesting the respiratory mechanics have no role in promoting intrinsic PEEP.

Morbidly obese patients compared with normal-weight patients, especially in a supine position and during general anesthesia, can present a combination of factors, such as expiratory flow limitation and a larger amount of peripheral airway closure volume [40, 41], that can result in intrinsic PEEP. Grieco et al., with a low-flow P-V curve technique, showed that 22% of obese patients under general anesthesia presented airway closure with a median airway opening pressure of 9 cmH₂O [42].

Similar data were reported in ARDS patients by Yonis et al. who found a significantly higher airway opening pressure in patients with a higher body mass index [18].

As expected, in the present study, the patients with intrinsic PEEP had a significantly higher body mass index, even if lower than 30 kg/m², than patients without intrinsic PEEP. However, since in the present study we did not have a P-V curve loop, we could not assess the possible presence of airway closure which has been reported between 25 and 50% of ARDS patients [17, 19].

In mechanically ventilated patients, the increased expiratory airway resistance caused by the endotracheal tube and by the expiratory valve of the mechanical ventilator can promote dynamic hyperinflation [43]. In this study, although we did not record the sizes of the endotracheal tubes employed, we can assume there were no differences between the two groups because the percentages of males and females were similar. Our policy is to use larger endotracheal tubes in males than in females.

As regards the expiratory valve, we did not use the same mechanical ventilators in all patients; however, our latest-generation intensive care mechanical ventilators are able to limit any expiratory resistance as much as possible.

Under passive conditions, the short expiratory time, the high respiratory rate, and the minute ventilation promote dynamic hyperinflation [2, 44]. In ARDS patients, low tidal volume, which reduces stress and strain, is widely recommended in order to reduce ventilator-induced lung injury (VILI) [35]. However, to avoid respiratory acidosis, the respiratory rate is commonly raised to maintain adequate minute ventilation and carbon dioxide clearance. Richard et al. found that raising the respiratory rate from 17 to 30 breaths/min significantly increased intrinsic PEEP, from 1.3 to 3.9 cmH₂O, maintaining the same low tidal volume (6 mL/kg of predicted body weight) [14]. However, in our study, the respiratory rate and minute ventilation were increased simultaneously, and they can promote dynamic hyperinflation [44], although our intrinsic PEEP was lower. For a similar minute ventilation (12 L/min), obtained with a different combination of low/high tidal volume and low/high respiratory rate, De Durante et al. found a progressive increase in intrinsic PEEP from 1.4 to 5.8 cmH₂O with an increase in respiratory rate from 14 to 34 breaths/min [15].

In line with previous data, we found that patients with intrinsic PEEP were ventilated with higher respiratory rate and lower expiratory time than those without. Furthermore, the amount of intrinsic PEEP found in our study should approximately correspond to what was reported by Richard et al. for similar respiratory rates [14].

Whenever an external PEEP is changed, the resulting new intrinsic PEEP is computed as the difference between the total PEEP measured during an expiratory pause and the applied by the ventilator [5, 28]. In ARDS patients, the PEEP is commonly used not only to improve gas exchange [45] but also to reduce lung inhomogeneity [20] and optimize recruitability [46, 47]. In a mixed population of COPD patients and patients with acute respiratory failure, when there was expiratory flow limitation, the higher external PEEP significantly reduced the intrinsic PEEP, which was “absorbed” in the total PEEP. In contrast, in patients without expiratory flow limitation, the intrinsic PEEP did not change and the total PEEP could be increased [12, 48].

Many studies have proposed raising the PEEP to detect any expiratory flow limitation: an external PEEP higher than the intrinsic PEEP can reduce the expiratory flow on the flow-volume loop [49, 50]. In the present study, on raising the applied PEEP from 5 to 15 cmH₂O, intrinsic PEEP was significantly reduced in 89% of the patients and increased only in 4%; this might be due to airway closure and flow limitation at PEEP 5 and an airway opening pressure between 5 and 15 cmH₂O that absorbs intrinsic PEEP. However, this value is clinically not important, suggesting expiratory flow limitation has no substantial role. In addition, the response in oxygenation was no different among patients with or without intrinsic PEEP. In fact, as Koutsoukou et al. showed arterial oxygenation significantly increased with the rise in external PEEP independently from the amount of intrinsic PEEP [12].

The two groups presented similar amounts of not consolidated tissue, lung gas volume, and lung recruitability, while the PaO₂/FiO₂ ratio was worse and the mortality rate was slightly higher in patients with intrinsic PEEP. Although in ARDS patients PaO₂/FiO₂ is commonly used in clinical practice to assess the severity and the response to ventilation and therapy, it is not a “research” indicator of the impairment of oxygenation (i.e., alveolar shunt). In fact, several extrapulmonary factors such as the concentration of hemoglobin, cardiac output, and the arterial-venous oxygen content can significantly affect PaO₂/FiO₂, independently from the severity of lung disease.

Limitations of this study are as follows: (1) the absence of data on airway resistance, because we only computed the total resistance including the endotracheal tubes; (2) the absence of any diagnosis of the expiratory flow limitation on the flow-volume loop or measuring the end-expiratory lung volume; and (3) the lack of measurements of intrinsic PEEP at zero PEEP.