Acute respiratory distress syndrome (ARDS) was initially thought to be a lethal double pneumonia and was identified as a syndrome by Ashbaugh et al. (1967). Unfortunately, in the 50 years since ARDS was identified, only a few treatments have been used, with the mainstay being supportive in the form of mechanical ventilation (Slutsky and Ranieri, 2013). However, mechanical ventilation constrained to the limitations of a heterogeneously injured lung can cause unintended tissue damage, referred to as ventilator-induced lung injury (VILI), which can significantly increase mortality (∼40%) as compared to lung protective ventilation (∼31%) (Acute Respiratory Distress Syndrome Network, 2000). Initial randomized controlled trials (RCTs) attempting to reduce VILI by lowering tidal volume (Vt) failed to reduce mortality (Stewart et al., 1998; Brower et al., 1999). It was not until the ARDS Network (ARDSnet) conducted the seminal ARMA study, published in 2000, that a reduction in mortality was shown (Acute Respiratory Distress Syndrome Network, 2000). However, most (Phua et al., 2009; Villar et al., 2011; Caser et al., 2014; Bellani et al., 2016; Laffey et al., 2016; Villar et al., 2016; Maca et al., 2017; Raymondos et al., 2017; Rezoagli et al., 2017; Fan et al., 2018; McNicholas et al., 2018; Pham et al., 2019; Shen et al., 2019) but not all (Brun-Buisson et al., 2004; Fan et al., 2005; Putensen et al., 2009; Petrucci and De Feo, 2013; Shen et al., 2019) of the recent statistical- and meta-analyses have shown that ARDS mortality has not been reduced below the 31% “gold standard” of the 2000 ARMA study but rather remains unacceptably high at ∼40% (Figure 1). Despite these disappointing results, the low-Vt ARDSnet method is still recommended as the standard-of-care protective ventilation strategy for ARDS patients (Fan et al., 2017, 2018; Papazian et al., 2019).

Since outcome data for ARDS patients has not improved for almost 20 years, it is imperative to (1) ascertain the mechanisms of dynamic alveolar and alveolar duct volume change during mechanical ventilation (elastic, viscous, or viscoelastic), (2) characterize ARDS-induced changes in alveolar mechanics (i.e., the dynamic change in alveolar size and shape during mechanical ventilation) (Grune et al., 2019) that drive VILI-induced tissue damage, (3) identify the role of airway pressure and the duration at both inspiration and expiration on alveolar mechanics in the acutely injured lung (Kollisch-Singule et al., 2014a, 2018), and (4) use this knowledge to develop novel ventilation strategies to better reduce VILI and protect the lung. Although pulmonary inflammation (biotrauma) also plays a critical role in ARDS and VILI pathogenesis, the focus of this review will be the mechanical injury to tissue caused during ventilation.

The current concept is that ARDS causes heterogeneous injury in three general lung compartments that are layered by gravity (Figure 2). The first compartment, or the non-dependent area, contains a small number of normal compliant alveoli that remain inflated at end-expiration—functional residual capacity (FRC)—and is referred to as the “baby lung” (Gattinoni and Pesenti, 2005). The second compartment consists of alveoli in the dependent areas that are collapsed and/or edema filled. The third compartment consists of alveoli that remain in the transition zone between healthy and unstable, due to loss of surfactant function, and that open and collapse with every breath.

Consequently, the ARDSnet low-Vt and plateau pressure (Pplat) strategy is constrained to ventilating this heterogeneous lung tissue without causing VILI using a three-tiered approach: (1) protect the baby lung by not overdistending the compliant tissue that is open at FRC, (2) rest the dependent collapsed and edema-filled tissue by keeping it out of the ventilatory cycle, and (3) stabilize the tissue in between by applying positive end-expiratory pressure (PEEP), usually adjusted by oxygenation (Figure 2) (Acute Respiratory Distress Syndrome Network, 2000; Del Sorbo et al., 2017).

The hallmark of ARDS is a heterogeneous lung injury encompassing normal, collapsed, edematous, and unstable tissues (Figure 2). This pathology alters pulmonary microanatomy and dynamic alveolar inflation physiology, generating three basic VILI mechanisms: volutrauma (overdistension of airways), atelectrauma (R/D of alveoli), and biotrauma (inflammation) (Thompson et al., 2017). From an engineering perspective, volutrauma is caused by excessive static strain and atelectrauma by excessive dynamic strain (Seah et al., 2011; Protti et al., 2013a, b, 2014). In this review we do not discuss biotrauma but rather focus on the unintentional mechanical damage to the pulmonary parenchyma caused during mechanical ventilation.

Mechanical ventilation of the acutely injured lung with altered alveolar opening and collapse time constants can cause tissue stress-failure by the following mechanisms: (1) alveolar R/D-induced excessive stress on the epithelial cells as the alveolar walls in apposition peel apart (Bilek et al., 2003), (2) stress-focusing in areas of open alveoli adjacent to collapsed or edema fill alveoli (Gattinoni et al., 2012; Cressoni et al., 2014; Retamal et al., 2014), and (3) collapsed tissue stretching and overdistending the shared walls of patent alveoli, as alveolar walls are interconnected (Figure 4) (Nieman et al., 2017b). Theoretically, if it were possible to minimize or prevent all of the above VILI mechanisms, ARDS associated mortality would be significantly reduced.

In normal pigs ventilated for 54 h, Protti et al. (2013b) examined the impact of high static strain and several levels of dynamic strain. They demonstrated that high static strain, using elevated PEEP and minimal Vt, caused little damage, but when PEEP was reduced, producing a high dynamic strain, it caused pulmonary edema and death. In subsequent work they showed that high static strain did not merely act as a dam to prevent edema formation by altering the Starling forces (i.e., increasing the pulmonary interstitial pressure) (Effros and Parker, 2009) but rather preserved the integrity of the alveolar-capillary membrane (Figure 3, Endothelial Leakage) (Protti et al., 2013a). These findings were supported by Jain et al. (2017) using a porcine heterogeneous lung injury model. They subjected two study groups to high static strain (Pplat = 40 cm H₂O) that was believed to be more than sufficient to cause volutrauma-induced VILI (Acute Respiratory Distress Syndrome Network, 2000; Sahetya et al., 2017). Following heterogeneous lung injury, the animals in the second group were also subjected to high dynamic strain. High static strain did not injure normal open tissue (baby lung), nor did it exacerbate acutely injured tissue. However, combining high static strain with high dynamic strain caused significant damage to both of these tissues (Figure 5). Further support for the contention that volutrauma of normal lung tissue is not a primary VILI mechanism comes from the Bates group, who showed that 4 h of mechanical ventilation in mice with high static strain was not associated with lung injury, but when it was combined with high dynamic strain, it caused VILI-induced tissue damage (Seah et al., 2011).

Protti et al. (2014) showed that a large dynamic strain (atelectrauma) is much more harmful to the normal lung than a large static strain (volutrauma), and when combined, the two work additively or synergistically to greatly accelerate tissue damage (Seah et al., 2011; Hamlington et al., 2016; Ruhl et al., 2019). In addition, the heterogeneous injury caused by ARDS establishes many areas of stress-focus between the open tissue and collapsed or unstable tissue and have been shown to double the stress and strain calculated for the entire lung (Cressoni et al., 2014). The impact that areas of stress-focus have on the forces generated on alveolar walls during ventilation was first described by Mead et al. (1970) and was more recently analyzed by Makiyama et al. (2014) using computer-simulated alveolar walls. This latter group demonstrated that a Pplat at the upper end of the “safe” level (30 cm H₂O) could result in a local stress-focusing in individual alveolar walls of up to 48 cm H₂O and could concentrate stress in an individual alveolar wall as much as 16-fold (Figure 6).

In summary, ARDS results in a heterogeneous loss of surfactant function (Figure 3, Surfactant Deactivation) and airway flooding (Figure 3, Alveolar Edema) that cause the lung to become time and pressure dependent, which means that the lung will collapse in a relatively short time at atmospheric pressure and will require a longer period of time to open even at high inflation pressures (Takahashi et al., 2015). Mechanical ventilation can exacerbate the initial ARDS-induced inflammatory injury (Figure 3, Endothelial Leakage, Surfactant Deactivation, and Alveolar Edema) by generating excessive stress and strain on alveolar and alveolar duct walls resulting from a collapsing and reopening of alveoli and heterogeneous areas of stress-focusing in open tissue adjacent to collapsed or edema-filled tissue (Figure 4).

It is important to note that parenchymal overdistension following acute lung injury is a regional phenomenon and occurs only in open alveoli and alveolar ducts that are adjacent to the unstable or collapsed tissue (Figure 7, PEEP 16, PEEP 5, APRV 10%); it does not occur in homogeneously inflated acutely injured lung tissue (Figure 7, APRV 75%) (Nieman et al., 2017b). It has been shown that combining high pressure with alveolar instability greatly exacerbates tissue tearing in a rich-get-richer fashion (i.e., the larger the initial tear in the epithelial membrane the more that this tear will be expanded by increased airway pressure) (Figure 8) (Hamlington et al., 2016; Ruhl et al., 2019).

The ARDSnet method is a logical physiologically based lung ventilation strategy within the constraints of protecting a heterogeneously injured lung. The better strategy would be to eliminate these constraints by opening and stabilizing the acutely injured lung or, better yet, by applying a protective ventilation strategy early and never letting the lung collapse (Satalin et al., 2018). Gattinoni and Pesenti stressed that in a patient with ARDS the Vt should not be set by body weight (Vt/kg ratio) but rather by the size of the remaining normal open tissue at FRC: the baby lung (Vt/baby lung volume ratio) (Gattinoni and Pesenti, 2005). They hypothesized that the baby lung was the small amount of open tissue at end-expiration (i.e., FRC) surrounded by a large volume of collapsed tissue with very low compliance. Since the specific elastance [E_spec = transpulmonary pressure (Ptp)/Vt x baby lung volume, which is the airway pressure at which expiratory lung volume or FRC or baby lung volume doubles in size (i.e., when Vt/baby lung volume = 1)] if the baby lung is postulated to be normal, there would be a greater potential for overdistension in normal lung tissue because of the high compliance at any given level of static strain (Gattinoni and Pesenti, 2005). Both the potential of excessive stress (Ptp) to cause volutrauma and of excessive strain (Vt/end-expiratory lung volume) to cause atelectrauma are linked with the following equation:

Gattinoni and Pesenti further concluded that the above equation suggests the baby lung must be treated gently, using low Vt, low Ptp, and proning so as not to cause volutrauma. Gently ventilating the lung should be protective, within the constraints of a heterogeneously injured lung (i.e., high E_spec and low baby lung volume).

The current strategy of lung protection is designed to minimize normal tissue overdistention (low Vt) and to stabilize (PEEP) injured tissue. Although the current hypothesis is that the primary mechanism of VILI is overdistension of the baby lung (Acute Respiratory Distress Syndrome Network, 2000), recent studies have demonstrated that high Ptp does not result in injury to normal lung tissue (Seah et al., 2011; Protti et al., 2013a, b, 2014; Jain et al., 2017). Since much of the lung remains collapsed and unstable using the ARDSnet method, there are numerous areas of stress-focus, as PEEP adjusted by changes in oxygenation may not be adequate to prevent R/D (Baumgardner et al., 2002; Boehme et al., 2015). Even with low Vt, tissue injury may occur regionally, since overdistension is not global but occurs in the microenvironment. Specifically, alveolar and alveolar duct overdistension occurs in open tissue surrounding a collapsed area of stress-focus or adjacent to areas of alveolar instability (Figures 4, 6, 7) (Mead et al., 1970; Cereda et al., 2011, 2013, 2016a; Kollisch-Singule et al., 2014b, 2018; Makiyama et al., 2014; Nieman et al., 2017b; Ruhl et al., 2019). Indeed, it has been shown that atelectasis, not high Vt, causes overdistension

in the adjacent patent alveoli and that alveolar size actually decreases with increased airway pressure (i.e., PEEP) if lung tissue is recruited by a mechanism of gas redistribution (Cereda et al., 2011, 2013, 2016a). It has also been shown that gas redistribution in the microenvironment is not only pressure dependent but also time dependent (i.e., the longer the pressure is applied, the better the gas redistribution) (Figure 7) (Kollisch-Singule et al., 2014b).

Animal studies have shown that the OLA can protect the lung from VILI even if ΔP is not reduced (Tojo et al., 2018). Recruiting the lung with PEEP has been shown to reduce tissue damage secondary to spontaneous breathing (SB) by two mechanisms: (1) the intensity of the SB is reduced via neuromechanical uncoupling, and (2) the reduced amount of atelectatic tissue decreases the volume of stress-focus areas (Morais et al., 2018). Other studies using animal models have also shown the physiologic and pathologic benefit of opening the acutely injured lung (Faridy et al., 1966; Webb and Tierney, 1974; Dreyfuss et al., 1988; Muscedere et al., 1994; Nakazawa et al., 2007; Nieman et al., 2015; Magalhaes et al., 2018). Additionally, animal studies have shown that to successfully implement the OLA, the RM and subsequent PEEP level must be applied properly, or the approach may actually increase lung damage. Farias et al. (2005) showed that an RM increased lung pathology if PEEP was not set sufficiently high to prevent recollapse of the newly opened tissue, a finding that is supported by direct in vivo observation of subpleural alveoli (Halter et al., 2003). In light of this evidence, discarding the OLA does not seem logical, since it is unclear whether the RCTs testing the OLA actually opened and stabilized the lung. A better strategy might be to identify ventilation strategies that are most likely to accomplish the goals of the OLA (Sahetya and Brower, 2017).

Surfactant deactivation and edema flooding with acute lung injury cause alveolar collapse; the result is very “sticky” and does not reopen easily (Figure 3) (Crotti et al., 2001; Gattinoni et al., 2003). This tissue has a very long opening time constant, and thus it will take an extended period of time at any given pressure to recruit these surfactant deficient and edematous tissues. Once the collapsed alveoli are recruited, the opposite problem develops; the newly opened alveoli have a very brief collapse time constant (Neumann et al., 1998a, b, 2000). Thus, alveoli collapse quickly (≤0.5 s) once a critical collapse airway pressure is reached (i.e., collapse is time and pressure dependent) (Kollisch-Singule et al., 2014a, b).

To effectively implement the OLA, the entire mechanical breath pattern (MB_P), including all airway volumes, flows, pressure, rates, and the times during which they apply during inspiration and expiration, must be analyzed (Kollisch-Singule et al., 2014a). The current OLA ventilation strategies do not consider the pathophysiological changes that occur in alveolar opening and collapse time constants. Attempting to recruit large volumes of collapsed lung with a single RM over a very brief period of time is often not physiologically possible depending on the degree of surfactant damage and edema. A more effective strategy may be a ventilation method that “nudges” the lung open over an extended period of time (6–24 h) using a high mean airway pressure held for most of each breath. Likewise, the very fast alveolar collapse time constants are not considered when setting PEEP with expiratory durations in the 2- to 3-s range. A more effective strategy may be to use a ventilator method with a very brief (≤0.5 s) expiratory time while maintaining a level of time-controlled PEEP (TC-PEEP). Greatly limiting the expiratory duration (≤0.5 s) keeps the lung from emptying completely, maintaining a TC-PEEP that is related to expiratory duration and the collapse time constant of the lung. Thus, for any given collapse time constant, the shorter the expiratory time, the higher the TC-PEEP and vice versa. In addition, since alveolar collapse is viscoelastic in nature (Figure 9), the very short expiratory duration would work synergistically with TC-PEEP to prevent alveolar collapse – alveoli simply would not have sufficient time to derecruit (Nieman et al., 2019). Analysis of the change in alveolar opening and collapse time constants with acute lung injury suggests that the time component of the MB_P (Kollisch-Singule et al., 2014a) can be used to improve the ability to recruit and stabilize acutely injured lung tissue, which is necessary to successfully implement the OLA.

In addition to the time component of MB_P, the dynamic physiology of alveolar R/D must also be understood to design a mechanical breath for the OLA. Alveoli and alveolar ducts inflate and deflate as a viscoelastic system (Denny and Schroter, 2000; Escolar and Escolar, 2004; Farias et al., 2005; Carvalho and Zin, 2011; Suki and Bates, 2011; Nieman et al., 2017b). The model most used to analyze viscoelastic behavior is the spring and dashpot (Figure 9). The most important thing to know about a viscoelastic system, in relation to lung opening and collapse, is that there is a time lag from when the force (inspiratory pressure) is applied until alveoli begins to open and a lag between when force is removed (expiratory pressure) and when alveoli begin to collapse. Thus, the longer the inspiratory time, the more lung tissue that will be recruited, and the shorter the expiratory time, the less lung tissue that will be allowed to recollapse (Figure 9). A collapsed airway inflates after the opening threshold pressure is reached, and the pressure then propagates down the airway, inflating more airways and alveoli. This process progresses in an avalanche manner with power-law distributions of both the size of and intervals between avalanches (Suki et al., 1994; Alencar et al., 2002). It has been postulated that as the lung opens, the increase in parenchymal tethering of airways (Broche et al., 2017) and alveolar interdependence improves lung function as a power-law function (Nieman et al., 2019). These new perspectives inform the quest for novel protective ventilation strategies. Indeed, our work in translational animal models and a meta-analysis of data on surgical intensive care unit (SICU) patients has shown that our time-controlled adaptive ventilation (TCAV) method, using airway pressure release ventilation (APRV) mode, is highly effective at keeping the lung open and stable, significantly reducing morbidity in translational animal models and reducing the ARDS incidence and mortality rates of SICU patients at high risk of developing ARDS (Roy et al., 2012; Andrews et al., 2013; Emr et al., 2013; Roy S.K. et al., 2013; Kollisch-Singule et al., 2014a, b, 2015a,b, 2017; Nieman et al., 2015, 2017a,b, 2018; Smith et al., 2015; Jain et al., 2016, 2017; Satalin et al., 2018; Silva et al., 2018; Mahajan et al., 2019).

Others have shown the importance of ventilation time on lung mechanics. Saddy et al. (2013) calculated a pressure-time product (PTP)/breath as the integral of the change in esophageal pressure over time. They found that, when using biphasic positive airway pressure, PTP significantly increased when the rate of time-cycled control breaths was at 50 breaths/min as compared to when it was 100 or 75 breaths/min. The energetics of ventilation also contain components of ventilator time. Power is defined as work per unit time and is thus equal to pressure x (volume/time) (Marini, 2018). Most of the energy applied to the lung during inflation is accounted for in elastic storage and airway resistance. It is postulated that the damage due to power is caused by the energy that is dissipated and unrecovered during exhalation. In addition, it is not just the power but also the changes in the microenvironment that result in tissue damage. Regional alveolar instability can develop during ARDS and can greatly amplify energy dissipation locally. Regional instability acts as a stress-focus, and once a microstrain threshold is reached, the unrecovered energy will cause a rapid progression of lung tissue injury in a power-law fashion (Hamlington et al., 2018).

It was recently shown in ARDS patients that low Vt and Pplat do not correlate well with reduced patient mortality, whereas low ΔP correlated strongly with improved survival (Amato et al., 2015). The critical lesson is that it is not the settings dialed into the ventilator (i.e., Vt, Pplat, and so on) that are key to improved survival, but rather the impact of these settings on lung physiology measured as a change in ΔP (Kollisch-Singule et al., 2014a, b, 2015b).

Using the ARDSnet method, the Vt was set at 6 cc/kg, the Pplat was set at < 30 cm H₂O, and PEEP was set on a sliding oxygenation scale; the mechanism of improved mortality was assumed to be due to these parameters. The Amato study clearly demonstrated that the mechanism of increased survival was not these desired setting values, because none correlated with outcome (Amato et al., 2015). Lower ΔP, on the other hand, strongly correlated with reduced mortality. Respiratory system compliance (C_RS) was used to calculate driving pressure (ΔP = Vt/C_RS), which was shown to decrease in patients who survived, reflecting a desirable change in lung physiology caused by recruitment.

The ARDSnet method assumes that the constraints of ventilating a heterogeneously collapsed lung are unavoidable, in which case the Ptp is high due to the high E_spec value and low baby lung (FRC) volume. However, if the entire lung could be fully recruited, these constraints would be eliminated, E_spec reduced, and FRC increased, resulting in a significant reduction in the applied stress (i.e., Ptp) for any given Vt. Rahaman used a stress equation similar to Eq. 1 to analyze the mathematics of VILI (Rahaman, 2017). He concluded that stress increases for any given Vt and PEEP with an increase in respiratory rate (RR). This conclusion assumes that the clinician is constrained in ventilating a heterogeneously collapsed lung such that neither FRC (i.e., the size of the baby lung) nor E_spec can be changed. When these conditions are true, the conclusion that Vt, PEEP, and RR must be kept low to reduce stress is correct. Conversely, our perspective is the better strategy is to treat the lung by reinflating the collapsed tissue, increase FRC (eliminating the baby lung), and decreasing E_spec. This combination will reduce lung stress (Ptp) during ventilation even at higher Vt, PEEP, and RR (Eq. 1).

Of course the strategy of “casting” the broken lung open until it heals would be clinically effective only with a ventilation strategy that could perform such a feat (Nieman et al., 2018). Unfortunately, the current OLA strategies have not been shown effective at opening and stabilizing the lung (Bhattacharjee et al., 2018; Cui et al., 2019; Hodgson et al., 2019; Kang et al., 2019; Zheng et al., 2019), and the recent ART trial showed an increase in mortality using the OLA (Brower et al., 2004; Meade et al., 2008; Mercat et al., 2008; Cavalcanti et al., 2017).

Our TCAV method using the APRV mode has been discussed in detail elsewhere (Habashi, 2005; Jain et al., 2016; Nieman et al., 2018, 2019; Kollisch-Singule et al., 2019). Briefly, TCAV consists of an extended (4–5 s) open valve continuous positive airway pressure (CPAP) phase with a very short (≤0.5 s) release phase. The inspiratory:expiratory ratio is ∼10:1. The open valve allows the patient to spontaneously breath (inspiration or expiration) with little resistance. Tidal volume (Vt) is not set but rather is a product of the CPAP level and lung compliance. A heterogeneously collapsed lung will have a very low Vt because the compliance will be low, but as the lung gradually recruits over time, the compliance will increase and so will the Vt. Since the Vt is set based on lung compliance, a high Vt will never be delivered to a non-compliant collapsed lung using the TCAV method. Thus, the Vt size is personalized and adaptive as the patient’s lung gets better or worse, directed by changes in lung compliance. The extended CPAP time will gradually nudge viscoelastic alveoli open over several hours until the lung is fully inflated. The newly recruited alveoli will be prevented from recollapse through the use of a very short expiratory duration (release phase). Expiratory time is very short (≤0.5 s), and thus the lung is reinflated (CPAP phase) before it has time to completely empty, maintaining a TC-PEEP. The very short expiratory time is not sufficient for viscoelastic alveoli to collapse and when combined with the TC-PEEP is highly effective at preventing alveolar derecruitment.

Our group has investigated the physiological impact of the TCAV method in both mechanistic and efficacy animal studies (Roy et al., 2012; Emr et al., 2013; Roy S.K. et al., 2013; Roy S. et al., 2013; Kollisch-Singule et al., 2014a, b, 2015a,b, 2017; Smith et al., 2015; Silva et al., 2018) and in a meta-analysis of data on SICU patients (Andrews et al., 2013). In addition, the TCAV method is the primary ventilator strategy used at the R Adam Cowley Shock Trauma Center in Baltimore, with well over 1,000,000 h of ventilator time. Below, we discuss the results from our animal data and our clinical statistical analysis as evidence for the TCAV method’s mechanisms and efficacy.

In rat VILI and hemorrhagic shock models, it was shown that ventilation for 6 h using the TCAV method was superior to volume-controlled ventilation (VCV; Vt 10 ml/kg, PEEP 0.5 cm H₂O) at preventing the development of ARDS and that lung protection was associated with stabilization of alveoli (Figure 10) (Emr et al., 2013; Roy S.K. et al., 2013). The TCAV method was also shown to be lung protective in a preterm piglet model of infant respiratory distress syndrome (Kollisch-Singule et al., 2017). These data were supported by mechanistic studies showing the ability of the TCAV method to normalize alveolar and alveolar duct microanatomy (Kollisch-Singule et al., 2014b) (Figure 7) and to reduce dynamic alveolar strain (Kollisch-Singule et al., 2014a, 2015b; Smith et al., 2015). We developed a 48-hr clinically applicable, high-fidelity, porcine peritoneal sepsis (PS) plus gut ischemia/reperfusion (I/R), multiple organ dysfunction syndrome (MODS) and ARDS model. In three studies using this clinically applicable model, we demonstrated that the TCAV method was superior to VC or the ARDSnet method at blocking progressive acute lung injury and preventing ARDS development (Figure 11) (Roy et al., 2012; Roy S. et al., 2013; Kollisch-Singule et al., 2015a). In addition, we have shown that surfactant proteins A and B are both better preserved with the TCAV as compared with the ARDSnet method, suggesting that there is sufficient lung volume change with TCAV to preserve stretch-induced surfactant release (Roy et al., 2012; Emr et al., 2013; Roy S.K. et al., 2013; Roy S. et al., 2013; Kollisch-Singule et al., 2015a). Lastly, in a statistical analysis of data from SICU patients, the TCAV method was associated with a significant reduction in ARDS incidence and mortality as compared to standard of care ventilation in 15 SICUs (Figure 12) (Andrews et al., 2013).

It is important to remember that if the acutely injured lung can be fully recruited, it will remain time and pressure dependent (i.e., will collapse quickly at atmospheric pressure) for a period of hours to days (Takahashi et al., 2015). Once the acutely injured lung is opened, the ventilator pressure/time support necessary to keep it open cannot be reduced until the lung “heals.” Gradually pulmonary edema will be reabsorbed and surfactant function restored (Vazquez de Anda et al., 2000). As more and more lung tissue is recruited, reestablishing alveolar interdependence (Figure 6) (Mead et al., 1970; Makiyama et al., 2014) and parenchymal tethering (Perun and Gaver, 1995), the lung will once again become stable at atmospheric airway pressure, and the weaning process can begin (Mead et al., 1970; Perun and Gaver, 1995; Makiyama et al., 2014).

Recent RCTs have shown mixed results in ARDS patients using APRV (but not utilizing the TCAV method). Zhou et al. (2017) randomized patients with a PaO₂/FiO₂ (P/F) ratio < 200 to either ARDSnet LVt or APRV with TCAV-like settings. The APRV group demonstrated significant decreases in length of ICU stay, tracheostomy requirement, ventilator days, and mortality. The APRV group also had significant reductions in the need for prone positioning, sedation, and neuromuscular blockade.

Lalgudi Ganesan et al. (2018) conducted an RCT in children with ARDS, who were randomized to receive either standard LVt strategy or APRV. The APRV group was shown not to be lung protective, and the trial was terminated early. We postulate the failure was due to fundamental misconceptions as to the MB_P necessary to protect the lung. A key objective in our TCAV method is to gradually nudge the lung open. In the Ganesan study, the CPAP pressure was reduced to generate a low Vt, and thus the CPAP level was likely not above the lung critical opening pressure, meaning the lung was not inflated. In the TCAV method, Vt is not set (see TCAV description above), so an increase in release volume (analogous to Vt) during the release phase and an increase in C_RS are positive signs and indictive of lung recruitment (Roy et al., 2012; Roy S. et al., 2013; Kollisch-Singule et al., 2015a). If the ventilation strategy used does not actually open the lung, the VILI mechanisms associated with heterogeneous ventilation will not be prevented (Figure 4). We postulate that this is also the reason for the failed RCTs using an RM and titrated PEEP (Brower et al., 2004; Meade et al., 2008; Mercat et al., 2008; Cavalcanti et al., 2017).

Hirshberg et al. (2018) conducted an RCT in adults, and similarly to the Ganesan study, targeted a Vt at ∼6 mL/kg. The study was stopped early, in part because the Vt often exceeded 12 mL/kg, even though there were no significant differences between groups in pneumothorax, sedation, vasoactive medications, P/F ratio, or outcome (Hirshberg et al., 2018). As stated above, when the TCAV method is used, an increase in Vt is indicative of lung recruitment and is a positive effect. Thus, one of the critical protective mechanisms of the TCAV method, the ability to fully open the lung, was not incorporated into the APRV setting in either the Hirshberg or Ganesan study. The conclusion from these clinical studies should not be that “APRV is not lung protective” but rather that the Zhou (Zhou et al., 2017) and TCAV methods (Andrews et al., 2013) are superior to the Ganesan (Lalgudi Ganesan et al., 2018) and Hirshberg (Hirshberg et al., 2018) methods for lung protection. Two recent statistical reviews and meta-analyses of RCTs have shown the APRV mode is associated with a mortality benefit, improved oxygenation, and a greater number of ventilator-free days when compared with conventional ventilation strategies, without any negative hemodynamic impact or higher risk of barotrauma (Carsetti et al., 2019; Lim and Litton, 2019).

An example of the TCAV method’s effectiveness when applied to a patient on extracorporeal membrane oxygenation (ECMO) with respiratory failure is seen in Figure 13. A 35-year-old man with a history of systemic lupus erythematous, with catastrophic antiphospholipid syndrome, underwent a coronary artery bypass graft with mitral valve repair and was placed on protective pressure-control assist-control ventilation (PC-AC) with protective settings using the ARDSnet method. Chest X-rays (CXRs) on PC-AC for 21 days, 15 of those days on ECMO, showed widespread infiltrates (Figure 13A). Another 15 days on ECMO (30 days on ECMO total) and PC-AC showed minimal change in CXRs and worsening lung function assessed by P/F ratio (A: 196, B: 75) (Figure 13B). Conversion to the TCAV method for 2 days resulted in a marked improvement in CXRs and P/F ratio (275) (Figure 13C). After 11 days on TCAV, the patient had a clear CXR and normal P/F ratio (414) (Figure 13D). Figure 13D was 6 h prior to decannulation and liberation from ECMO. After 39 days on PC-AC with worsening lung function, it took only 2 days on the TCAV method to reopen and stabilize the patient’s lung, and after 11 days the patient was disconnected from ECMO, extubated, and removed from mechanical ventilation.

Current protective ventilation strategies have not significantly reduced ARDS-related mortality in the 20 years since the ARMA study, possibly because strategies are constrained to ventilating the heterogeneously collapsed and unstable lung. In addition, current protective ventilation strategies have not considered all of the physiological parameters necessary to protect the acutely injured lung. Alveoli and alveolar ducts open and collapse as a viscoelastic system such that there is a time lag between when the airway pressure is applied and alveoli open, and between when the airway pressure is removed and alveoli collapse. Loss of pulmonary surfactant function, a key ARDS pathology, amplifies these viscoelastic properties, rendering the lung time and pressure dependent, such that it takes an extended time at the higher pressure for lung tissue to open and a very brief time, even with an elevated airway pressure, for the lung to recollapse.

This knowledge suggests that to eliminate VILI the ventilation strategy must open and stabilize the lung. Current ventilation strategies attempting this OLA have not reduced mortality in clinical trials. Since alveolar volume change is viscoelastic in nature, the time at both inspiration and expiration would be critical to accomplishing the OLA goals. We have developed a time-controlled adaptive ventilation TCAV method using the APRV mode and have shown that it is highly effective at opening and stabilizing the lung, which significantly reduces VILI-induced lung damage in animal ARDS models. We have also shown that preemptive application of the TCAV method in trauma patients significantly reduces ARDS incidence and mortality. Recent RCTs using the APRV mode but not adjusted with the TCAV method have shown mixed results, suggesting that properly adjusted APRV is critical to clinical success.

GN drafted the manuscript. HA-K, MK-S, JS, SB, GT, PA, MM, LG, and NH critically revised the manuscript. All authors read and approved the final manuscript.

GN, HA-K, MK-S, JS, LG, PA, MM, and NH have lectured for Intensive Care On-line Network, Inc. (ICON). NH is the founder of ICON, of which PA and MM are employees. The authors maintain that industry had no role in the design and conduct of the study; the collection, management, analysis, or interpretation of the data; nor the preparation, review, or approval of the manuscript. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.