Acute lung injury: how to stabilize a broken lung

It has been theorized that ARDS causes heterogeneous lung injury with both stress-risers (i.e. collapsed or edema-filled alveoli directly adjacent to patent alveoli) and alveolar instability [8] secondary to surfactant deactivation [4]. This heterogeneous injury causes excessive alveolar strain during tidal ventilation resulting in a secondary VILI, which may lead to placing the patient with ARDS on extracorporeal membrane oxygenation (ECMO). Once on ECMO, the clinician may decide to initiate “lung rest”, a concept [9, 10] somewhat analogous to putting a cast on the lung until it heals. The problem, however, with this approach is that significant lung pathology can occur due to lung collapse alone [4, 11, 12] and the lung will eventually have to be reopened so the patient can be weaned off ECMO. Thus, resting the acutely injured lung in a collapsed state would be like casting a fractured arm without properly aligning the bones first. In addition, chronic collapse may lead to irreversible collapse induration, a form of serve alveolar fibrosis [13‐15]. The lung is designed to function optimally only when fully inflated and thus the ideal situation would be to cast the open lung and let it heal at its biologically natural volume. If the open lung could be “rested” using a time-controlled adaptive ventilation (TCAV) protocol we would have the best of both worlds; VILI would be eliminated and all the negative ramifications of ECMO would be avoided.

There is a tetrad of pathology associated with ARDS: (1) increased pulmonary capillary permeability; (2) surfactant deactivation; (3) alveolar flooding with edema; and (4) altered alveolar mechanics with a dynamic change in alveolar size and shape with each breath (Fig. 1) [16]. The combined impact of this pathology is a significant loss of functional residual capacity (FRC) with heterogeneous lung collapse and instability with closing volume greater than FRC-producing tidal opening and closing of airspaces [17]. Although the pathophysiology of VILI is complex [18‐21], in part due to the complexity of alveolar microanatomy, we postulate that ventilating the unstable lung at low lung volume is the core mechanism of VILI-induced tissue damage. In addition, raising tidal volume without stabilizing FRC would result in a greater number of lung units that were previously protected from mechanical ventilation with permissive atelectasis to now engage in VILI-inducing recruitment-derecruitment (R/D)-induced tissue damage [10].

Studies have shown that elevated airway pressure with levels known to cause VILI is relatively benign if the lung is not allowed to fall significantly below FRC [22‐24]. The majority of these studies used positive end-expiratory pressure (PEEP) to prevent lung collapse during expiration. Maintaining adequate FRC has also been shown to be protective in the normal lung. Pigs mechanically ventilated for 54 h at total lung capacity (TLC) with very high strain (global strain = 2.5, near TLC) do not develop ARDS as long as PEEP is sufficient to prevent lung collapse at end-expiration. Pigs with normal lungs ventilated at the same high strain (2.5) but without PEEP, allowing the lung to collapse at expiration, develop severe ARDS with a high mortality rate (Fig. 2) [22]. High ventilation driving pressure (DP) in humans, measured by dividing lung compliance (Cstat) into tidal volume (Vt) (DP = Vt/Cstat), correlates with an increase in ARDS mortality [25]. Decreasing Vt can lower DP, but that would lead to further heterogeneous lung collapse. The other solution would be to increase lung compliance, which can be accomplished by recruiting the lung. In a novel heterogeneous porcine lung injury model, Jain et al. showed that peak airway pressures of 40 cmH₂O did not injure normal lung tissue or exacerbate damage to the acutely injured tissue as long as lung volume was maintained during expiration (Fig. 3) [23]. Thus, as long as the normal or heterogeneously injured lung is not allowed to collapse at expiration, VILI will be prevented, even with very high airway pressures and static strain.

The properly inflated lung is highly resistant to VILI due to the structural integrity of the pulmonary parenchyma known as “alveolar interdependence” (Fig. 4) [26, 27]. When alveoli and alveolar ducts collapse during expiration, this structural interdependence is lost resulting in two key pathologic changes that are primary mechanisms driving VILI: (1) lung instability with repetitive alveoli collapsing and expansion (i.e. RACE) with each breath (Fig. 1) [27‐29] and (2) areas of alveoli that are adjacent to open alveoli remain collapsed throughout ventilation. The areas where collapsed and open alveoli connect are termed “stress-risers” and cause a great deal of stress in the patent alveoli (Fig. 5) [30‐33]. Both of these VILI mechanisms are associated with a loss of expiratory lung volume and stability. Alveolar over-distention, the third mechanical VILI mechanism, only occurs in open alveoli adjacent to those that are collapsed (Fig. 1-Alveolar Edema and Fig. 5) [8]. This hypothesis is supported in a study showing that VILI does not occur with peak airway pressures of 40 cmH₂O, which is believed to cause over-distension-induced lung injury, as long as the lung is fully recruited and not allowed to collapse during expiration (Fig. 3) [23]. Thus, opening and stabilizing alveoli would prevent all three mechanical mechanisms of VILI.

The current standard of care is application of a protective ventilation strategy, such as low tidal volume (LVt) ventilation, after significant ARDS develops [34]. The progression to ARDS is often silent with normal blood gases in the presence of ARDS with unstable alveoli [35] leading clinicians to believe the lung is “fine”’ and there is no need to change to protective mechanical ventilation. Once the LVt protocol is applied, oxygenation is maintained between arterial partial pressure of oxygen (PaO₂) of 55–80 mmHg or oxygen saturation of 88–95%, using a sliding FiO₂/PEEP scale [34]. However, by the time this level of lung pathology is present, there is already considerable tissue and surfactant damage resulting in a significant loss of lung volume predisposing the lung to VILI [36]. Switching to the LVt protocol contributes to further loss of lung volume and PEEP, which is used in combination with the LVt protocol, may be ineffective at stabilizing alveoli as a method of lung recruitment [37, 38]. With ARDS mortality remaining unacceptably high and essentially unchanged for the past 18 years even when using the LVt protocol, new ventilation strategies are being sought [39‐41].

The resolution to the problem seems to be simple; all we need to do is cast the lung at risk of developing ARDS to keep it open and stable in the event of impending lung injury. However, as ARDS progresses, vascular permeability will increase and edema fluid entering the alveolus will begin to deactivate surfactant making the lung increasingly unstable (Fig. 1). Thus the cast must be adjusted as lung pathophysiology advances or diminishes, to maintain homogeneous ventilation. Therefore, the clinician must understand how to apply the proper “dose” of the preemptive mechanical breath.

Recent work has shown that alveolar strain is viscoelastic in nature [42, 43]. The important thing to understand about a viscoelastic structure such as the alveolus, is that when force (i.e. Vt) is applied there is both a fast and slow component to alveolar opening or collapse. Thus, some alveoli might recruit in the first milliseconds (fast component) of inspiration but if the inspiratory duration is extended many more alveoli will continue to recruit (slow component). Conversely, with removal of the force (i.e. exhalation), some alveoli would begin to collapse immediately in milliseconds (fast component) but if the expiratory duration were very brief, many alveoli would simply not have time to collapse (slow component) [8].

With this knowledge, the optimal way to cast a broken lung would be with an extended time at inspiration and brief time at expiration. The extended inspiratory duration in this lung casting protocol would be a continuous positive airway pressure (CPAP). The CPAP with an open exhalation valve allows the patient to spontaneously breathe on top of the CPAP with little effort, maximizing synchrony. A very brief expiratory release time from CPAP (< 0.5 s), would not be sufficient time for the lung to completely empty.

We have developed a preemptive ventilation strategy to cast the lung maintaining homogeneous ventilation using an extended time at inspiration and a brief time at expiration [44]. The components of our TCAV protocol include the ventilator mode, the settings within this mode, and the changes in lung physiology used to modify these settings as the patient’s lung gets better or worse (Fig. 6a&b). The ventilator mode must be pressure-controlled and time-cycled, with the ability to precisely and independently control machine inspiratory and expiratory times such as airway pressure release ventilation (APRV), BiLevel, Bi-Vent, BiPhasic or DuoPAP. The settings used with our TCAV protocol include an extended time at inspiration (T_High) that occupies ~ 90% of each respiratory cycle; the high pressure (P_High) set sufficiently to recruit alveoli and regain FRC (Fig. 6a); the time at expiration (T_Low) set to terminate at 75% of the peak expiratory flow rate (PEFR), which is typically ≤ 0.5 s; and the low pressure (P_Low) set at 0 cmH₂O (Fig. 6b). Although P_Low is set at 0 cmH₂O, the pressure never reaches 0 cmH₂O since the T_Low is set sufficiently brief to maintain PEEP (Fig. 6a).

Changes in lung physiology are used to adjust the settings based on assessment of the slope of the expiratory flow curve, which reflects the elastance of the respiratory system. Respiratory system resistance also determines the slope of the expiratory flow curve; however, by setting P_Low to zero we minimize the resistance for more accurate measurement of lung elastance. As ARDS progresses and respiratory system elastance increases, the expiratory flow slope decreases (Fig. 6b - normal lung 45° and ARDS lung 30°), and the T_Low is reduced to prevent these faster collapsing alveoli from de-recruiting because the ARDS lung has a faster collapse time constant [44, 45]. The TCAV protocol with the ventilator mode APRV is the same concept as the ARDSnet protocol, which includes the ventilator mode (volume-assist control), the settings within this mode (LVt < 6cc/kg, limiting plateau pressure (Pplat) < 30 cmH₂O, etc.), and the FiO₂/PEEP sliding scale used to adjust the settings.

Our TCAV protocol is best described as CPAP with a brief release. The CPAP phase has no trigger and patients can generate unassisted spontaneous breaths. The time-controlled component of our TCAV protocol is an extended time at inspiration (T_High), which is greater than the slowest time constant, gradually “nudging” the lung open. The brief time at expiration, which is set less than the fastest alveolar collapse time constant, minimizes airway closure (Fig. 6b). The CPAP or P_High is adjusted to the pathologic condition of the patient’s lung, degree of FRC, and to changes in chest wall compliance [5]. The adaptive component of our TCAV protocol uses changes in lung mechanics to guide the clinician in setting the expiratory duration or T_Low sufficiently brief and precise to prevent derecruitment of alveoli, even those with the fastest collapse time constants (Fig. 6b) [44, 45].

In summary, the steeper the slope, the “sicker” the lung (greater respiratory system elastance) and the briefer the expiratory duration needed to prevent expiratory airway closure [46]. The TCAV protocol is so effective at recruiting lung tissue that hypercapnia is not usually a problem because there is ample alveolar surface area for CO₂ exchange. Also, pulmonary vascular resistance (PVR) is lowest when lung volume is at FRC and thus PVR is not elevated even with the higher mean airway pressure generated with the TCAV protocol. Driving pressure remains low with the TCAV protocol even with relatively high Vt (10–12 cc/kg) since lung compliance remains normal in the fully inflated lung. Full lung recruitment also eliminates the stimulus for strong inspiratory efforts (i.e. lung stretch receptors, blood pH, PO₂, PCO₂ concentrations) eliminating any problem with dyssynchrony and negative pleural pressure causing pathologically high transpulmonary pressure.

Our group has used the preemptive TCAV protocol to successfully prevent the development of ARDS in patients [47] (Fig. 7) and in translational, clinically applicable, animal models (Fig. 8) [5, 48‐50]. A recent RCT using a protocol similar to TCAV reduced the duration of ICU stay and mechanical ventilation [51] (Fig. 9). Clinical settings for the TCAV protocol have been discussed in detail elsewhere [44].

Our viewpoint is that the TCAV protocol should be applied to all patients at risk of developing ARDS as soon as they are intubated. However, the TCAV protocol also works very well to open and stabilize the lungs of patients with established ARDS. Although there is not yet a RCT comparing the TCAV protocol with the ARDSnet protocol we do have strong expert experience. The TCAV protocol is the primary mode of ventilation at R. Adam Cowley Shock/Trauma Center in Baltimore, MD, USA and thus millions of hours of expert experience have been accumulated in patients with severe ARDS, including patients on extracorporeal membrane oxygenation (ECMO).

The LVt strategy is currently the standard of care because of the positive RCT showing that low (6 cc/kg) Vt versus higher (12 cc/kg) Vt significantly reduces mortality [34]. This RCT combined with evidence from other RCTs, meta-analyses, and systematic reviews have led to evidence-based medicine (EBM) guidelines strongly recommending lower Vt and plateau pressure, with moderate confidence in the effect estimates [52]. EBM has been increasingly accepted as the gold standard to direct patient care, but physicians are beginning to challenge the exclusive use of EBM to guide patient care [53‐59]. A paper published in Lancet, a journal skeptical of the validity of EMB recommendations, criticized RCTs because they focus on internal validity (effective only on patients that fit the criteria for the controlled study) and disregard the critical issues of external validity (effective on all patients) [60]. The RCT is considered the key component of the new EBM paradigm [61]. However, EBM failures can often be attributed to this emphasis on internal validity of RCTs and thus the recommendations often fail in clinical practice (external validity) [62, 63]. Fernandez et al. stated that, “The main problem with the EMB approach is the restricted and simplistic approach to scientific knowledge, which prioritizes internal validity as the major quality of the studies to be included in clinical guidelines”. Thus, EBM suggested treatment strategies may or may not be the optimal and externally validated expert experience of physicians that have used treatment strategies successfully in their ICUs should also be considered.

The recently published ART Trial RCT applied the open lung approach (OLA) using conventional ventilation strategies (low Vt, PEEP, and recruitment maneuvers (RM)) in patients with established ARDS. In this study the OLA group had an increase in mortality, suggesting the possibility that OLA may not be an effective ventilation strategy for established ARDS [64]. However, it is not known if the OLA strategy used in this study actually recruited the lung (i.e. no lung scans showing full recruitment). We have shown that the TCAV protocol is superior to controlled mechanical ventilation (CMV) with high PEEP at opening and stabilizing subpleural alveoli [65‐68]. We postulate, and our animal data [48, 50, 65‐67] and expert clinical experience supports, that the TCAV protocol is far superior to CMV with RM plus PEEP at opening the non-compliant lung in ARDS. Using this modified TCAV protocol on brain-dead patients with collapsed lungs we have increased the number of transplantable lungs by ~ 680% (unpublished observations). Further support comes from a RCT comparing a protocol similar, but not identical to, TCAV with the ARDSnet protocol in patients with ARDS [51]. This studied showed that the TCAV protocol improved oxygenation and respiratory system compliance (suggesting superior lung recruitment), decreased plateau pressure and decreased both the ICU stay and duration of mechanical duration (Fig. 9). We postulate the reason why the TCAV protocol is superior at opening the lung in ARDS is due to the viscoelastic nature of alveolar opening and collapse [8]. The extended time at inspiration will recruit lung tissue at a much lower airway pressure than is needed to open the lung with a RM, since it is not just the pressure but the time the pressure is applied that recruits lung tissue [69]. Unlike a RM that is a one-time application of high airway pressure, the TCAV protocol maintains an elevated airway pressure almost continually (except for the very brief release phase) such that alveoli are gradually “nudged” open over time.

Using our understanding of ARDS pathophysiology, mechanisms of VILI at the alveolar level and dynamic alveolar inflation and deflation, we postulate it is possible to stabilize the lung in patients at high risk of developing ARDS, similar to “casting” a broken bone for stability. The TCAV protocol uses a simple strategy of open-valve CPAP with a brief, intermittent release guided by changes in lung mechanics. Because the TCAV protocol can be applied as soon as intubation, very early in ARDS pathogenesis, it will effectively “Never give the lung a chance to collapse” [8] and by doing so eliminate most VILI pathophysiology. The TCAV protocol has been shown to prevent ARDS in a group of trauma patients at high risk of ARDS when applied early and used as the primary mode of mechanical ventilation at the R Adam Cowley Shock/Trauma in Baltimore, MA [47, 51]; a protocol similar to TCAV was shown to improve respiratory system compliance and oxygenation and reduce the duration of mechanical ventilation and the ICU stay [51].