Comments
Abstract
Rationale: In normal lungs, local changes in pleural pressure (Ppl) are generalized over the whole pleural surface. However, in a patient with injured lungs, we observed (using electrical impedance tomography) a pendelluft phenomenon (movement of air within the lung from nondependent to dependent regions without change in tidal volume) that was caused by spontaneous breathing during mechanical ventilation.
Objectives: To test the hypotheses that in injured lungs negative Ppl generated by diaphragm contraction has localized effects (in dependent regions) that are not uniformly transmitted, and that such localized changes in Ppl cause pendelluft.
Methods: We used electrical impedance tomography and dynamic computed tomography (CT) to analyze regional inflation in anesthetized pigs with lung injury. Changes in local Ppl were measured in nondependent versus dependent regions using intrabronchial balloon catheters. The airway pressure needed to achieve comparable dependent lung inflation during paralysis versus spontaneous breathing was estimated.
Measurements and Main Results: In all animals, spontaneous breathing caused pendelluft during early inflation, which was associated with more negative local Ppl in dependent regions versus nondependent regions (−13.0 ± 4.0 vs. −6.4 ± 3.8 cm H2O; P < 0.05). Dynamic CT confirmed pendelluft, which occurred despite limitation of tidal volume to less than 6 ml/kg. Comparable inflation of dependent lung during paralysis required almost threefold greater driving pressure (and tidal volume) versus spontaneous breathing (28.0 ± 0.5 vs. 10.3 ± 0.6 cm H2O, P < 0.01; 14.8 ± 4.6 vs. 5.8 ± 1.6 ml/kg, P < 0.05).
Conclusions: Spontaneous breathing effort during mechanical ventilation causes unsuspected overstretch of dependent lung during early inflation (associated with reciprocal deflation of nondependent lung). Even when not increasing tidal volume, strong spontaneous effort may potentially enhance lung damage.
When spontaneous breathing occurs during mechanical ventilation, negative pleural pressure adds to and increases transpulmonary pressure; because of fluid-like behavior in normal lungs, expansion occurs across all areas of the pleural surface and deeper inside the lung. In this way, adding spontaneous breaths augments ventilation and gas exchange and is traditionally encouraged in mechanically ventilated patients.
This study demonstrates that in the injured lung, local negative pleural pressure generated by diaphragmatic contraction is not uniformly transmitted, but is concentrated in dependent lung (corresponding to a vertical pressure gradient in negative pleural pressure, from nondependent to dependent regions). In the presence of lung injury, this causes a pendelluft phenomenon, with shift of air from nondependent to dependent lung regions, without changes in tidal volume. Thus, during lung-protective ventilation with strictly limited tidal volume the presence of strong inspiratory effort can result in a hidden, local overstretch of the dependent lung.
The normal lung is traditionally considered to be a continuous elastic system, exhibiting fluid-like behavior (1, 2), such that distending pressure applied to a local region of the pleura becomes generalized over the whole lung (pleural) surface (3–7). This uniform distribution of forces, known as interdependence, seems to be a consequence of the special arrangement of fibers inside the lung interstitium and the free transmission of negative pleural pressures (Ppls) between lobes encased within a common chest wall (7–9). Spontaneous breathing effort exposes the lung to negative Ppl generated by diaphragmatic contraction: if the patient undergoes mechanical ventilation, then positive airway pressure (Paw) generated by the ventilator is also present. Thus, a diaphragmatic contraction uniformly lowers Ppl by the same amount at all points on the lung surface, creating a uniform increase in transpulmonary pressure (PL) (3–7): the combination of these pressures, the PL gradient (PL = Paw − Ppl), is the net pressure driving inflation of the lung (10). Therefore, spontaneous breathing, by making Ppl more negative, uniformly augments PL for any given Paw.
Thus, spontaneous breathing is traditionally encouraged in patients receiving mechanical ventilation (11, 12) because it is thought to provide lung expansion at lower levels of Paw, a strategy that results in better local (especially dependent) lung aeration, thereby enhancing gas exchange and potentially improving hemodynamics (13).
Recently, techniques such as electrical impedance tomography (EIT) that allow visualization of regional lung inflation in real-time have become available (14). We report here a patient with acute respiratory distress syndrome (ARDS) receiving mechanical ventilation in whom pendelluft (i.e., movement of gas within the lung without a change in tidal volume) during early inflation was caused by spontaneous breathing effort. The transient intrapulmonary gas flow corresponded to inflation of dependent lung with concomitant deflation of nondependent lung.
Regional deflation induced by diaphragmatic contraction has not been previously described. Although lowered regional compliance and/or increased airway resistance could explain why dependent lung might inflate less than nondependent lung regions (15), all regions would inflate or deflate simultaneously, albeit at different rates. For a superimposed spontaneous breath to inflate one region and (simultaneously) deflate another, the occurrence of large differences in the changes of local Ppl is necessary.
We used a laboratory model (anesthetized pigs with lung injury) to determine whether the pendelluft induced by spontaneous breathing that we observed in our patient was reproducible. We also sought to determine the mechanisms involved, using dynamic imaging and by measuring Ppl changes in the affected lung regions.
We hypothesized that, in contrast to the uniform distribution of pressures exhibited in normal lung, negative Ppl generated by diaphragmatic contraction might not be transmitted all over the lung surface in the setting of lung injury, but rather would be transiently localized to dependent lung, and that such locally concentrated Ppl would cause pendelluft and expansion of the local dependent lung. Some of the results of these studies have been previously reported in the form of an abstract during 2013 American Thoracic Society International Conference (16).
This study was approved by the ethics committee for clinical studies (patient report) and the ethics committee for experimental studies (animal study) at Faculdade de Medicina da Universidade de São Paulo. Detailed clinical data and experimental methods are described in the online supplement.
A 40-year-old man developed postoperative hypoxemia while receiving mechanical ventilation in the intensive care unit after coronary artery bypass graft surgery. As part of comprehensive monitoring, EIT was used to visualize the regional distribution of inflation. EIT data were recorded using the Enlight impedance tomography monitor (Dixtal, São Paulo, Brazil), placed on the perimeter defining a cross-sectional plane of the thorax at the level of the fifth to sixth intercostal space (14, 17). Regional distribution of ventilation was analyzed with the subdivision of the thorax in four zones (ventrodorsal) in the presence or absence of spontaneous breathing effort.
Seven Landrace pigs were anesthetized with midazolam, ketamine, and fentanyl, and tracheotomized. An esophageal balloon catheter (SmartCath; Bicore, Irvine, CA) was inserted to measure the esophageal pressure (Pes) and its position validated as described previously (18). After establishing lung injury, animals were ventilated with assisted pressure-controlled ventilation at driving pressure of 5–15 cm H2O to obtain a tidal volume of 6 ml/kg, respiratory rate of 20–30 breaths per minute, and pressure trigger of −2 cm H2O. Positive end-expiratory pressure was set at “open-lung positive end-expiratory pressure” −10 cm H2O. All animals preserved spontaneous breathing effort, with titrating sedatives. EIT data were recorded at the level of the sixth intercostal space and analyzed with the same subdivision as for the patient thorax (as above).
Local changes in Ppl were measured in nondependent versus dependent regions (of five animals) by balloon catheter occlusion (Pulmonx, Inc., Redwood, CA) of subsegmental bronchi via a fiberoptic bronchoscope as follows: nondependent region, left B; dependent region, left lower lobe beyond D4. The pressure swings in the occluded subsegments were used as surrogates for the relative changes in local Ppl, as described previously (19). Simultaneous pressure recording of Ppl and Pes swings were performed, while preserving spontaneous effort.
Ppl was directly measured using a balloon-tipped catheter in five pigs. The catheter was surgically inserted between the dependent parts of the lung (dorsal aspect of the diaphragmatic surface) and the diaphragm. We simultaneously measured Ppl and Pes while preserving spontaneous effort.
To estimate the magnitude of local lung inflation caused by spontaneous effort, we determined the pressures required to inflate a region of dependent lung with positive pressure (i.e., ventilator, airway driving pressure only) versus spontaneous effort (i.e., ventilator + diaphragmatic contraction, airway driving pressure + the change in Ppl) in five pigs. Then, after measuring the local (dependent) impedance change (delta Z) during assisted ventilation (representing local tidal strain) at a fixed ventilator driving pressure of 10 cmH2O, we paralyzed the animal and commenced controlled ventilation with progressive increments in driving pressure, until the same delta Z developed in the dependent lung.
For the validation of regional distribution of ventilation and pendelluft detected by EIT, dynamic computed tomography (CT) scans were performed in one representative pig to evaluate the absolute movement of air within thick slices close to the diaphragm, and in additional thick slices above the carina. Dynamic CT scans were performed over 10-second intervals without interrupting mechanical ventilation, using the same ventilator settings as recorded in EIT. For each slice, we assessed the regional ventilation at nondependent one-third versus dependent one-third of the ventrodorsal chest diameter.
Intergroup differences were evaluated using one-way repeated measurement analysis of variance followed by Dunnett test. Two-condition comparisons (Ppl measurements) were done by unpaired t test. Statistical significance was considered where P less than 0.05.
During controlled mechanical ventilation with muscle paralysis (i.e., no spontaneous effort), we observed simultaneous inflation of different lung regions, although the rates of inflation differed across the lung (Figure 1, right; see Video E1 in the online supplement). In contrast, when spontaneous effort was present (i.e., without neuromuscular paralysis) we observed two key differences (Figure 1, left; see Video E2). First, in the initial stages of the breath (until peak inspiratory flow was reached, after completion of the “triggering phase”) spontaneous effort inflated dependent lung regions to a greater extent (∼1.5–2 fold) than with controlled ventilation (i.e., in the presence of muscle paralysis). Second, the early inflation in the dependent region was accompanied by concomitant (transient) deflation in the nondependent region, suggesting the existence of pendelluft (i.e., movement of alveolar air from one region of lung to another, without significant gain in tidal volume). Importantly, aside from deflections in the pressure-time tracing, no differences were noted with conventional respiratory monitoring (e.g., tidal volume, minute ventilation, plateau pressure) during spontaneous breathing efforts as compared with controlled ventilation under muscle paralysis (Table 1).
Figure 1. Electrical impedance tomography (EIT) waveforms in a patient with acute respiratory distress syndrome, spontaneous versus ventilator breaths. The patient was ventilated with assist pressure-controlled ventilation (IP, 5 cm H2O; f, 12 min−1; PEEP, 13 cm H2O; triggering threshold, −2 cm H2O). EIT data were recorded with spontaneous breathing (left) and without spontaneous breathing (right). The EIT image was divided into four zones, each covering 25% of the ventrodorsal diameter (zones 1–4). During controlled ventilation (under muscle paralysis), simultaneous inflation of each of the different lung regions was observed, although at differing inflation rates. In contrast, when spontaneous efforts were present, two observations were noted. First, in the initial stages of the breath spontaneous effort caused inflation of dependent lung regions (red in zone 4), which was greater with controlled breaths. Second, the early inflation in the dependent region was accompanied by concomitant (transient) deflation of nondependent region (red in zones 1 and 2), indicating movement of gas from nondependent to dependent lung regions: a pendelluft phenomenon. f = respiratory frequency; IP = inspiratory pressure; PEEP = positive end-expiratory pressure.
[More] [Minimize]| Age, yr | 40 | |
| Sex | Male | |
| Height, cm | 172 | |
| Weight, kg | 98 | |
| Surgery | Coronary artery bypass graft surgery with two internal thoracic arteries and two saphenous veins | |
| Extracorporeal circulation time, min | 101 | |
| Amount of lost blood, ml | 450 | |
| Blood transfusion | No | |
| pH | 7.28 | |
| PaO2/FiO2, mm Hg | 132 | |
| PaCO2, mm Hg | 46.4 | |
| Base excess | −7.8 | |
| Lactate, mg/dl | 107 | |
| Spontaneous Breathing | Muscle Paralysis | |
| Plateau airway pressure, cm H2O | 19.9 | 17.7 |
| Positive end-expiratory pressure, cm H2O | 13.9 | 13.0 |
| Tidal volume, ml | 532 | 477 |
| Tidal volume, ml/kg | 7.8 | 7.0 |
| Respiratory rate, breaths/min | 15 | 12 |
| Minute ventilation, L/min | 8.0 | 5.7 |
| Peak flow, L/s | 0.8 | 0.6 |
Using EIT, in injured lungs, we demonstrated that in the presence of neuromuscular paralysis there was heterogeneous but simultaneous inflation in the dependent and in the nondependent lung regions (Figure 2, right; see Video E3), as observed in our patient.
Figure 2. Electrical impedance tomography (EIT) waveforms in experimental lung injury, spontaneous versus ventilator breaths. In an anesthetized pig model of acute lung injury assist pressure-controlled ventilation (IP, 15 cm H2O; f, 25 min−1; PEEP = 13 cm H2O; triggering threshold, −2 cm H2O) was used. The EIT image was divided into four zones, each covering 25% of the ventrodorsal diameter (zones 1–4). During controlled ventilation (under muscle paralysis), simultaneous inflation of each of the different lung regions was observed, although at different inflation rates. In contrast, when spontaneous efforts were present, two observations were noted. First, in the initial stages of the breath, spontaneous efforts caused inflation of dependent lung regions (red in zones 3 and 4), which was greater with controlled breaths. Second, the early inflation in the dependent region was accompanied by concomitant (transient) deflation of nondependent region (red in zone 1), indicating movement of gas from nondependent to dependent lung regions; because this was not associated with alterations in tidal volume it indicates a pendelluft phenomenon. This finding was always present during spontaneous breathing efforts in all animals with experimental lung injury. f = respiratory frequency; IP = inspiratory pressure; PEEP = positive end-expiratory pressure.
[More] [Minimize]Also in injured lungs, spontaneous breathing effort resulted in initial inflation of dependent lung with concomitant deflation in nondependent lung (Figure 2, left; see Video E4). Synchronous inflation and deflation occurred without significant difference in exhaled tidal volume between spontaneous effort and muscle paralysis (5.9 ± 2.1 vs. 5.2 ± 2.0 ml/kg; P = 0.13) (Table 2), suggesting transfer of gas within the lung, in this case from nondependent (deflating) to dependent (inflating) regions. This observation is consistent with pendelluft, as observed in our patient. The inflation of dependent lung (zones 3 and 4) and simultaneous deflation of nondependent lung (zone 1) extended well beyond the triggering phase and lasted until esophageal pressure reached its most negative levels (Figure 2, red). Such pendelluft-type ventilation was observed in all lung-injured pigs. Continuous recording of EIT and Ppl was made during the transition from spontaneous effort to complete muscle paralysis (after a bolus of succinylcholine chloride) (Figure 3); this demonstrated that the extent of the pendelluft was proportional to the intensity of the respiratory effort.
| Respiratory Variables at Electrical Impedance Tomography Recording (n = 7) | |||
|---|---|---|---|
| Body weight, kg | 31.2 ± 4.1 | ||
| pH | 7.26 ± 0.11 | ||
| PaO2/FiO2, mm Hg | 79.4 ± 61.2 | ||
| PaCO2, mm Hg | 46.2 ± 14.7 | ||
| Base excess | −5.0 ± 1.8 | ||
| Injured Lung | |||
| Normal Lung (before lung injury) | Spontaneous Breathing | Muscle Paralysis | |
| Plateau airway pressure, cm H2O | 21.2 ± 1.4 | 20.1 ± 3.6 | 18.8 ± 4.3 |
| Positive end-expiratory pressure, cm H2O | 10.3 ± 0.4* | 7.3 ± 2.9 | 7.3 ± 2.9 |
| Mean airway pressure, cm H2O | 14.1 ± 0.9 | 14.2 ± 3.3 | 11.3 ± 3.6 |
| Change in esophageal pressure, cm H2O | 3.4 ± 0.9 | −5.1 ± 3.5* | 2.4 ± 0.6 |
| Peak transpulmonary pressure, cm H2O | 6.0 ± 2.3 | 9.6 ± 4.5 | 9.2 ± 5.8 |
| Tidal volume, ml/kg | 9.7 ± 1.3* | 5.9 ± 2.1 | 5.2 ± 2.0 |
| Dynamic compliance of respiratory system, ml/cm H2O | 27.9 ± 3.6* | 14.4 ± 3.7 | 14.1 ± 3.5 |
| Respiratory rate, breaths/min | 22.6 ± 2.7 | 47.4 ± 11.8* | 25.3 ± 4.2 |
| Peak flow, L/s | 0.73 ± 0.08 | 0.69 ± 0.28 | 0.51 ± 0.17 |
| Lung Regions | |||
| Nondependent Lung | Dependent Lung | ||
| Airway occlusion manometry (n = 5) | |||
| Change in esophageal pressure, cm H2O | −6.5 ± 3.1 | −6.8 ± 2.9 | |
| Change in pleural pressure, cm H2O | −6.4 ± 3.8 | −13.0 ± 4.0† | |
| Direct measurement (n = 5) | |||
| Change in esophageal pressure, cm H2O | — | −7.1 ± 2.1 | |
| Change in pleural pressure, cm H2O | — | −14.9 ± 3.1‡ | |
Figure 3. Electrical impedance tomography waveforms in transition from spontaneous breathing to muscle paralysis. After injection of succinylcholine chloride, negative deflections in esophageal pressure were reduced (yellow) as paralysis developed. As spontaneous breathing effort weakened, the degree of pendelluft became smaller (green).
[More] [Minimize]In contrast, in normal lungs (i.e., before lung injury), we observed simultaneous inflation of different lung regions during spontaneous breathing (see Figure E1), consistent with fluid-like behavior of the lung.
Representative fast dynamic CT acquisitions in an experimental animal corroborated the EIT data, confirming the induction of pendelluft by spontaneous breathing (Figure 4; see Figure E2B). Fast dynamic CT acquisitions also revealed that spontaneous breathing effort caused early local deflation of the nondependent regions in already aerated lung during inspiration (Figure 4, green). Simultaneously, transient local recruitment of dependent regions was observed because of pendelluft (i.e., tidal recruitment) (Figure 4, red). No pendelluft was observed during controlled mechanical ventilation (see Figure E2A). The CT studies also showed that, when present, the intraslice movement of air (from nondependent to dependent zones) was most pronounced in lung slices close to the diaphragm (see Figure E2B, red) than in sections located further cephalad from the diaphragm.
Figure 4. Regional distribution of inflation-deflation in lung injury, dynamic computed tomography. When spontaneous breathing effort was present, the computed tomography studies showed an intraslice movement of air (from nondependent to dependent zones) in lung slices close to the diaphragm (deflation from 1 to 2 in nondependent and concomitant inflation from 3 to 4 in dependent regions). This is consistent with pendelluft observed in electrical impedance tomography. Importantly, pendelluft caused tidal recruitment of dependent regions (red) by concomitant deflating nondependent regions (green).
[More] [Minimize]Spontaneous effort was associated with a large vertical gradient of Ppl swings, from nondependent to dependent regions. The swings in Ppl were larger in dependent regions, as compared with nondependent regions (−13.0 ± 4.0 vs. −6.4 ± 3.8 cm H2O; P < 0.05) (Table 2). The changes in Pes were similar to the changes in Ppl observed in nondependent lung regions, but smaller than the Ppl changes in dependent regions, indicating that ΔPes significantly underestimated ΔPpl in dependent regions. Furthermore, with spontaneous effort, directly measured swings in Ppl were significantly larger in dependent lung than swings in Pes (−14.9 ± 3.1 vs. −7.1 ± 2.1 cm H2O; P < 0.01) (Table 2), corroborating the indirect measurements with intrabronchial balloon catheter. A sample of Ppl waveform in dependent regions versus a simultaneous change in Pes is illustrated in Figure 5.
Figure 5. Pleural pressure during breathing effort, regional versus esophageal. As a spontaneous breathing effort is made, the negative change (reduction) in regional pleural pressure in the dependent lung was twofold that measured using esophageal manometry. In addition, the regional pleural pressure in the dependent lung became negative more rapidly than the esophageal pressure. Pes = esophageal pressure; Ppl = pleural pressure.
[More] [Minimize]To achieve comparable tidal inflation of dependent lung regions, passive inflation (positive pressure only) of the lung required a threefold increase in airway driving pressures when compared with airway driving pressures used during spontaneous effort (28.0 ± 0.5 vs. 10.3 ± 0.6 cm H2O; P < 0.001). In contrast, similar levels of local delta transpulmonary pressures were observed during spontaneous versus controlled ventilation (peak ΔPL: 22.3 ± 3.9 vs. 24.9 ± 1.7 cm H2O; P = 0.14) when achieving comparable inflation of the dependent lung (Figure 6, upper).
Figure 6. Local dependent lung inflation, controlled versus spontaneous breath. (Top) The driving pressure required to produce a comparable local delta-transpulmonary pressure (delta Z, local tidal strain) in the dependent lung during controlled breathing (under neuromuscular paralysis) was threefold that developed during spontaneous breathing. †P < 0.05 compared with local delta Z in SB; ¶P < 0.05 compared with local delta PL in SB. (Bottom) Tidal volume was maintained at less than 6 ml/kg during spontaneous breathing. However, a tidal volume of almost 15 ml/kg was required during controlled breaths (muscle paralysis) to obtain the same degree of local inflation in the dependent lung as that achieved in spontaneous breaths. Thus, spontaneous breathing caused unsuspected local overstretching of the dependent lung because of pendelluft, even when ventilator tidal volume was limited to less than 6 ml/kg. *P < 0.05 compared with global delta Z in SB; #P < 0.05 compared with Vt in SB. Delta Z = tidal strain; Delta-transpulmonary pressure = (driving pressure − change in pleural pressure); EIT = electrical impedance tomography; SB = spontaneous breathing.
[More] [Minimize]During spontaneous effort, the large tidal excursion in the dependent lung (as indicated by EIT) was associated with no changes in the (small; 5.8 ± 1.6 ml/kg) global tidal volume. During controlled ventilation, to achieve comparable inflation of the dependent lung, a significantly larger global tidal volume (14.8 ± 4.6 ml/kg; P < 0.05) (Figure 6, lower) was required. This was related to the presence of pendelluft during spontaneous breaths, which transiently inflated the dependent lung, but transiently deflated the nondependent lung.
This is an early description of a pendelluft phenomenon in a patient with lung injury, consisting of shifts of alveolar air from nondependent to dependent lung regions without a change in tidal volume, and caused by spontaneous breathing effort in the presence of mechanical ventilation.
This study has two important implications. First, lung-protective ventilation strategy (i.e., the limitation of plateau pressure and tidal volume) cannot prevent unsuspected local overstretch of the dependent lung because of pendelluft when moderate to strong spontaneous effort is preserved during mechanical ventilation. Second, the phenomenon occurs because, in contrast to normal lung, the injured lung does not exhibit fluid-like behavior, and in this case the transmission of local changes in Ppl is heterogeneous: a large vertical pressure gradient in Ppl swings. In this context, a single spatial measurement (e.g., esophageal pressure monitoring) does not represent an overall change in Ppl.
Standard physiologic explanations indicate that lung exhibits fluid-like behavior while inflating (1–7). This is based on the observations in normal lung, in which diaphragmatic contraction triggers local changes in Ppl that are uniformly transmitted across the entire lung surface (3–7). This principle explains, in part, why Pes is used as a surrogate of overall Ppl to optimize lung distention, an approach that may improve outcome in patients with ARDS (18, 20).
In the current study we confirmed that spontaneous effort during mechanical ventilation in healthy (uninjured) lungs resulted in almost simultaneous inflation throughout the lung. Although inflation of different regions occurred at different rates, inflation was almost simultaneous in all areas; there was no significant concomitant deflation (see Figure E1). In addition, airway occlusion tests performed in healthy lung demonstrated that changes in Pes closely matched changes in Paw (and local changes in Ppl matched changes in Paw) (see Figure E3). Thus, these observations indicate simultaneous generalization of local Ppl changes throughout normal lung, supporting the concept of fluid-like behavior in the (normal) lung (1–7).
However, the pattern of lung inflation was different in the presence of lung injury. In the injured lung pendelluft occurred. It resulted from the development of a more negative swing in Ppl in the dependent lung than in the nondependent lung. Acute lung injury involves tissue inflammation and fluid-filled alveoli in a heterogeneous distribution; thus, some areas are well aerated and others collapse or fill with fluid and inflammatory cells (15). Such dense tissue may behave less like a fluid and more as a frame of solid areas resisting shape deformation. In this setting, part of the mechanical energy generated by the diaphragmatic contraction is exerted on local lung deformation rather than being transmitted to the rest of the lung; this results in imperfect elastic anisotropic inflation. Because atelectasis in ARDS tends to be greater in the dependent and peridiaphragmatic lung regions (21), such responses might be more marked in these areas. Supporting this idea is the finding that even in normal lung, extreme levels of phrenic nerve stimulation, which results in exaggerated diaphragmatic contraction, can cause more negative change in Ppl at the diaphragmatic surface than in other parts of the lung (3, 5, 6). Thus, even the normal lung may depart from its fluid-like behavior at extreme levels of deformation. We speculate that during injury, this phenomenon is exacerbated and, even at moderate levels of muscle effort, the physical properties of the lung cannot compensate (i.e., with sufficient rapidity) for dynamic shape changes produced by diaphragmatic contraction.
An important issue of clinical relevance is that this effect is not detectable using conventional monitoring. In the current study, demonstration of the effect required dynamic imaging (i.e., EIT or dynamic CT). In addition, although esophageal monometry (Pes) is becoming a more standard clinical tool for estimation of PL or ΔPL (18, 20), this technique significantly underestimated the ΔPL in the dependent lung (Figure 5), and would therefore not detect this potentially harmful phenomenon. Furthermore, the concomitant inflation and deflation that resulted in air exchange from nondependent to dependent lung segments resulted in little net change in tidal volume. Hence, pendelluft would not be suspected from monitoring gas flow at the trachea. Finally, low tidal volume ventilation did not prevent the phenomenon. Despite limitation of tidal volume to less than 6 ml/kg, strong spontaneous effort resulted in unsuspected local overstretch of the dependent lung because of pendelluft, leading to tidal recruitment in dynamic CT acquisitions (Figure 4); indeed, matching this degree of regional overstretch during neuromuscular paralysis required an overall tidal volume of 15 ml/kg (i.e., a highly injurious lung stretch).
Overall ventilator-induced lung injury consists of important regional elements (22) and it is possible that such regional components could be exacerbated by this pendelluft effect. For example, transient overstretch of dependent region could cause or worsen injury in more atelectatic lung, causing exacerbation of tidal-recruitment as obvious in dynamic CT acquisitions (Figure 4). Indeed, a previous paper clearly disclosed that spontaneous breathing effort associated with high PL caused tidal recruitment and worsened experimental lung injury, despite the use of low tidal volume ventilation (23). In addition, cycles of transient deflation/inflation of the nondependent regions could worsen already aerated (and potentially hyperinflated) lung. Such mechanisms could explain a recent report that muscle paralysis early in the course of severe ARDS leads to a better survival, and less barotrauma, notwithstanding the use of comparable tidal volumes and plateau pressures (24–26).
Important potential influences on changes to thoracic impedance during the respiratory cycle include cephalocaudal displacement of the lung during the respiratory cycle and pulmonary-cardiac blood volume change. Approximately 10-cm spans of lung were sampled with EIT; thus, the bulk of a whole lobe would be in measurement range of EIT at all stages of the respiratory cycle. In addition, dynamic, thin-slice (5 mm) CT ensured that we compared the same anatomic slices at the end of inspiration and of expiration. Notwithstanding slight cephalocaudal movement of lung with respiration, the broad sampling range used in EIT and the thin sampling slice used in dynamic CT precludes movement artifact. Finally, we carefully excluded the heart and diaphragm from regions of interest, and CT analysis was based on absolute (not relative) air content, thereby excluding artifact from changes in intrathoracic blood volume.
A potential concern is that PL during spontaneous breathing does not directly reflect transalveolar pressure (i.e., alveolar pressure = Paw – resistive pressure). However, the resistive pressure generated is negligible. We demonstrated (Figure 6, upper) that at equal local tidal volume (delta Z), the ΔPL generated during paralysis versus spontaneous breaths were similar, and thus at that level of delta Z the resistive pressure must be negligible. Note that during paralysis, PL = transalveolar pressure because we have zero flow conditions at the end inspiration.
It is important to emphasize that spontaneous breathing caused pendelluft whether triggering was with pressure triggering or flow triggering (see Figure E4). The flow-time curves indicate that pendelluft is apparent after inflow has begun, after triggering is complete. In addition, simultaneous EIT and pressure recording during transition from spontaneous effort to paralysis demonstrate that the degree of pendelluft is proportional to the intensity of spontaneous effort (Figure 3). These observations suggest that triggering is associated with, but not the mechanism of, pendelluft. Nonetheless, we anticipate that during spontaneous breathing, triggering delay or auto positive end-expiratory pressure could augment pendelluft.
Although we demonstrated extreme, transient degrees of regional stretch, additional experiments are needed to determine whether such overstretch actually caused localized lung injury. Indeed, it is possible that the phenomenon may selectively inflate dependent (atelectatic) lung regions thereby augmenting overall lung recruitment, especially when lung injury is mild, spontaneous effort is modest, and PL can be maintained at low levels (12, 23, 27).
In conclusion, this study describes a new phenomenon during mechanical ventilation whereby spontaneous breathing during mechanical ventilation causes unsuspected, transient overstretching of dependent lung regions with concurrent deflation of nondependent lung.
| 1. | Hoppin FG Jr, Green ID, Mead J. Distribution of pleural surface pressure in dogs. J Appl Physiol 1969;27:863–873. |
| 2. | Krueger JJ, Bain T, Patterson JL Jr. Elevation gradient of intrathoracic pressure. J Appl Physiol 1961;16:465–468. |
| 3. | D’Angelo E, Sant’Ambrogio G, Agostoni E. Effect of diaphragm activity or paralysis on distribution of pleural pressure. J Appl Physiol 1974;37:311–315. |
| 4. | D’Angelo E, Agostoni E. Continuous recording of pleural surface pressure at various sites. Respir Physiol 1973;19:356–368. |
| 5. | Minh VD, Friedman PJ, Kurihara N, Moser KM. Ipsilateral transpulmonary pressures during unilateral electrophrenic respiration. J Appl Physiol 1974;37:505–509. |
| 6. | Minh VD, Kurihara N, Friedman PJ, Moser KM. Reversal of the pleural pressure gradient during electrophrenic stimulation. J Appl Physiol 1974;37:496–504. |
| 7. | Zidulka A, Nadler S, Anthonisen NR. Pleural pressure with lobar obstruction in dogs. Respir Physiol 1976;26:239–248. |
| 8. | Macklem PT, Murphy B. The forces applied to the lung in health and disease. Am J Med 1974;57:371–377. |
| 9. | Minh VD, Dolan GF, Kurihara N, Friedman PJ, Konopka RG, Moser KM. Stability in lobar ventilation distribution during change in thoracic configuration. J Appl Physiol 1975;39:462–468. |
| 10. | Talmor D, Sarge T, O’Donnell CR, Ritz R, Malhotra A, Lisbon A, Loring SH. Esophageal and transpulmonary pressures in acute respiratory failure. Crit Care Med 2006;34:1389–1394. |
| 11. | Stock MC, Downs JB, Frolicher DA. Airway pressure release ventilation. Crit Care Med 1987;15:462–466. |
| 12. | Putensen C, Zech S, Wrigge H, Zinserling J, Stüber F, Von Spiegel T, Mutz N. Long-term effects of spontaneous breathing during ventilatory support in patients with acute lung injury. Am J Respir Crit Care Med 2001;164:43–49. |
| 13. | Wrigge H, Zinserling J, Neumann P, Defosse J, Magnusson A, Putensen C, Hedenstierna G. Spontaneous breathing improves lung aeration in oleic acid-induced lung injury. Anesthesiology 2003;99:376–384. |
| 14. | Borges JB, Okamoto VN, Matos GF, Caramez MP, Arantes PR, Barros F, Souza CE, Victorino JA, Kacmarek RM, Barbas CS, et al. Reversibility of lung collapse and hypoxemia in early acute respiratory distress syndrome. Am J Respir Crit Care Med 2006;174:268–278. |
| 15. | Hubmayr RD. Perspective on lung injury and recruitment: a skeptical look at the opening and collapse story. Am J Respir Crit Care Med 2002;165:1647–1653. |
| 16. | Yoshida T, Torsani V, Gomes S, Santiago RRDS, Beraldo M, Tucci MR, Zin WA, Kavanagh BP, Amato MBP. Spontaneous effort causes occult pendelluft during mechanical ventilation [abstract]. Am J Respir Crit Care Med 2013;187:A1121. |
| 17. | Costa EL, Borges JB, Melo A, Suarez-Sipmann F, Toufen C Jr, Bohm SH, Amato MB. Bedside estimation of recruitable alveolar collapse and hyperdistension by electrical impedance tomography. Intensive Care Med 2009;35:1132–1137. |
| 18. | Baydur A, Behrakis PK, Zin WA, Jaeger M, Milic-Emili J. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982;126:788–791. |
| 19. | Martin CJ, Young AC, Ishikawa K. Regional lung mechanics in pulmonary disease. J Clin Invest 1965;44:906–913. |
| 20. | Talmor D, Sarge T, Malhotra A, O’Donnell CR, Ritz R, Lisbon A, Novack V, Loring SH. Mechanical ventilation guided by esophageal pressure in acute lung injury. N Engl J Med 2008;359:2095–2104. |
| 21. | Caironi P, Cressoni M, Chiumello D, Ranieri M, Quintel M, Russo SG, Cornejo R, Bugedo G, Carlesso E, Russo R, et al. Lung opening and closing during ventilation of acute respiratory distress syndrome. Am J Respir Crit Care Med 2010;181:578–586. |
| 22. | Dreyfuss D, Saumon G. Ventilator-induced lung injury: lessons from experimental studies. Am J Respir Crit Care Med 1998;157:294–323. |
| 23. | Yoshida T, Uchiyama A, Matsuura N, Mashimo T, Fujino Y. The comparison of spontaneous breathing and muscle paralysis in two different severities of experimental lung injury. Crit Care Med 2013;41:536–545. |
| 24. | Gainnier M, Roch A, Forel JM, Thirion X, Arnal JM, Donati S, Papazian L. Effect of neuromuscular blocking agents on gas exchange in patients presenting with acute respiratory distress syndrome. Crit Care Med 2004;32:113–119. |
| 25. | Forel JM, Roch A, Marin V, Michelet P, Demory D, Blache JL, Perrin G, Gainnier M, Bongrand P, Papazian L. Neuromuscular blocking agents decrease inflammatory response in patients presenting with acute respiratory distress syndrome. Crit Care Med 2006;34:2749–2757. |
| 26. | Papazian L, Forel JM, Gacouin A, Penot-Ragon C, Perrin G, Loundou A, Jaber S, Arnal JM, Perez D, Seghboyan JM, et al.; ACURASYS Study Investigators. Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010;363:1107–1116. |
| 27. | Yoshida T, Uchiyama A, Matsuura N, Mashimo T, Fujino Y. Spontaneous breathing during lung-protective ventilation in an experimental acute lung injury model: high transpulmonary pressure associated with strong spontaneous breathing effort may worsen lung injury. Crit Care Med 2012;40:1578–1585. |
Supported by Fundação de Amparo à Pesquisa do Estado de São Paulo, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Financiadora de Estudos e Projetos, Fukuda Foundation for Medical Technology, Tokibo CO, Newport Medical Instruments Inc., and the Osaka University Scholarship for Short-term Overseas Research Activities 2012.
This study was performed at Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil.
Author Contributions: T.Y. designed the study, conducted the study, analyzed the data, and wrote the manuscript. V.T. and M.A.B. participated in the study design and conducted the study. S.G. conducted the study and organized experiments as a responsible person in an animal intensive care unit. R.R.D.S. and M.R.T. conducted the study. W.A.Z. and E.L.V.C. participated in the study design and interpreted the data of pleural pressure and esophageal pressure. B.P.K. participated in the study design and revised the manuscript. M.B.P.A. designed the study, revised the manuscript, and organized the study as overall supervisor.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.201303-0539OC on November 7, 2013
Author disclosures are available with the text of this article at www.atsjournals.org.
