Mechanism of Action

In the atelectatic lung, PEEP increases FRC and lung compliance by increasing transpulmonary pressure (P_A − P_pl) at the end of expiration. This effect is beneficial only in people who already have abnormally low FRC and low lung compliance. In such individuals, PEEP may prevent alveolar collapse at the end of expiration and at sufficiently high levels may help to reopen collapsed alveoli.⁵ The reopening of previously collapsed alveoli reduces intrapulmonary shunt and improves PaO₂. In addition, PEEP improves lung compliance (i.e., less pressure is required to inflate the lung to a given volume) by restoring the FRC of the diseased lung; that is, PEEP places the tidal volume on a steeper portion of the lung compliance curve (Figure 14-2), requiring less pressure to achieve a given tidal volume. As collapsed alveoli are reopened, the lung’s compliance curve shifts back toward its normal position.

Basic Concepts of Hyperinflation Injury

Few subjects have been made more complicated than the cause of hyperinflation lung injury. In thinking about what causes overdistention of the lung, during mechanical ventilation, one must keep in mind that changes in lung volume cannot be caused by any factor other than changes in alveolar distending pressure i.e., transpulmonary pressure (see Chapter 3). That is, if transpulmonary pressure does not change, volume cannot change. One must be clear about cause and effect: Pressure changes are the cause of lung volume changes; the lung cannot somehow gain volume independent of a pressure change. The separation of pressure changes from volume changes during ventilation, whether spontaneous or mechanical, is conceptually incorrect. Lung inflation, and thus stretch of the alveolar epithelium, can occur only when the transpulmonary pressure gradient (P_A − P_pl) increases. Several references in the medical literature erroneously imply that high volume produces alveolar trauma, not high pressure—as though pressure and volume changes are independent of one another.

A study sometimes cited in support of the idea that pressure and volume have separate effects on the lung is one in which rats (presumably with similar lung compliance) were ventilated with (1) high airway pressures and high tidal volumes; (2) low airway pressures and similarly high tidal volumes—accomplished with a negative pressure ventilator that mimicked spontaneous breathing mechanics; and (3) high airway pressures and low tidal volumes—accomplished by strapping the rats’ chest walls with rubber bands.⁷ The results showed that only the lungs exposed to high volumes sustained injury, as was evidenced by the presence of high-permeability pulmonary edema.⁸ That is, the rats in group 3, exposed to high airway pressures but low tidal volumes (because their chest walls could not expand), did not develop edema and were presumed to be uninjured.

These findings have been interpreted to mean that lung volume, not pressure, is the causative agent that produces lung injury. Although it is true that lung volume must increase to produce alveolar overdistention, it cannot do so without an increase in distending pressure, which should not be confused with the absolute alveolar pressure. The conclusion that volume is more important than pressure in producing lung injury is simplistic; it overlooks the inseparable relationship between transpulmonary pressures and lung volumes. The results of the rat study are not surprising when one realizes that transpulmonary pressure determined the alveolar volume and stretch in each ventilation method. As shown in Figure 14-5, ventilation with positive pressure (method 1) and with negative pressure (method 2) are the same with respect to their effects on alveolar epithelial stretch because their transpulmonary pressures were the same and thus produced the same tidal volumes. In ventilation method 3, the strapped chest walls caused alveolar and intrapleural pressures to increase together, keeping transpulmonary pressures (and thus tidal volumes and alveolar stretch) relatively small; therefore, this lung sustained no injury.

Figure 14-5 oversimplifies the lung’s extremely complex architecture; that is, alveoli do not behave like individual balloons; they have shared walls. An intricate framework of connective tissue binds adjacent alveoli together such that they are interdependent. The mechanism of lung inflation does not involve balloon-like expansion and contraction of alveoli in the normal lung (see Chapter 3). However, regardless of inflation mechanism, lung volume changes can occur only as a result of transpulmonary pressure changes.

To summarize, the rat study⁸ did not distinguish between the effects of pressure and volume on alveolar injury. Instead, it demonstrated that an increase in transpulmonary pressure, no matter how it is produced, generates an increase in lung volume, which, if great enough, injures the alveoli. It should be apparent that lung volume is a function of distending pressure and that volume and pressure cannot be separated in the roles they play in VILI. The idea that pressure and volume are independent of one another led to the suggestion that a mechanical ventilator can be engineered to guarantee delivery of a desired tidal volume without exceeding a preselected maximal pressure. In actuality, the microprocessor in such ventilators computes lung compliance and then determines whether or not the

selected volume can be delivered without exceeding the targeted pressure; if it cannot, the microprocessor simply adjusts the pressure target upward (within a factory preset limit) to deliver the desired volume. If still more inflation pressure is required to deliver the desired volume, the ventilator delivers only the volume that the lung can accept within the preset pressure limit (i.e., not all of the desired volume is delivered).

On the other hand, the rat study⁸ points out the important fact that high alveolar pressure is not always associated with high lung volume. That is, high alveolar pressure does not always mean distention (transpulmonary) pressure is high, as demonstrated by the ventilation of rats with rubber-banded chest walls. Only when extrathoracic forces hinder chest wall expansion or hinder diaphragmatic displacement (e.g., rib cage deformities, severe abdominal distention, morbid obesity) is it possible for high alveolar pressures to be associated with a relatively low distention pressure. (Two other ordinary examples are the Valsalva maneuver, in which a person strains to exhale against a closed glottis, and the playing of a wind instrument such as an oboe, which presents high expiratory resistance; both situations increase the alveolar pressure without increasing transpulmonary pressure, as is evidenced by the lack of lung expansion.)

In mechanically ventilated patients, the correlate of alveolar pressure is plateau pressure, measured at the endotracheal tube at end-inspiration under zero airflow conditions (see Chapter 3). In all patients except patients with restricted chest wall and diaphragmatic movement, plateau pressure measurement is a reliable surrogate for assessing overdistention of the lung. It is generally accepted that the incidence of stretch-related lung injury is minimized if inspiratory plateau pressure is limited to less than 30 cm H₂O.^9,10

This pressure may not be safe in highly compliant alveoli that are dispersed throughout the lungs, as seen in severe emphysema; such alveoli may sustain injury even if plateau pressure is less than 30 cm H₂O. That is, highly compliant alveoli may become overdistended at pressures much less than 30 cm H₂O.

Mechanisms of Lung Injury

Mechanisms of VILI fall into three general categories: (1) mechanical stretch injury from hyperinflation (barotrauma/volutrauma); (2) repetitive collapse and reinflation of alveoli (atelectrauma); and (3) toxic effects of inflammatory mediators on pulmonary tissues (biotrauma). As already indicated, mechanical overdistention of the alveoli disrupts the structural integrity of the lung, allowing air to dissect through lung tissues into the pleural space and mediastinum, producing pneumothorax and possibly dissection of air into subcutaneous tissues around the neck and upper chest (subcutaneous emphysema). In addition, overstretch of the alveoli disrupts the alveolar capillary membrane and induces the release of inflammatory mediators, both of which increase membrane permeability.9–11 The heterogeneous distribution of diseased lung regions in ARDS makes the lung especially susceptible to this type of injury because of regional compliance differences; that is, a mechanically delivered tidal volume is preferentially distributed to the most compliant alveoli, placing them at greater risk of overdistention.

The repetitive collapse and reopening of alveoli (recruitment and derecruitment) during mechanical ventilation is another mechanism of VILI.2,10,12 This phenomenon has led to the term atelectrauma. Such opening and closing of alveoli subjects the lung fibroskeleton, microvasculature, and delicate juxtaalveolar tissues to mechanical sheer stress forces, especially at the junctions of regions with differing compliances and anchoring attachments.¹⁰

The term biotrauma was coined to describe the damage brought about by the release of inflammatory mediators in the lung, such as cytokines, in response to atelectrauma and hyperinflation injury.¹² It has been shown in animal models that physical stretch of the lung during mechanical ventilation greatly increases lung epithelial release of tumor necrosis factor-alpha (TNF alpha), a key cytokine that mediates the effects of sepsis.¹²

Other cytokines, such as interleukins, may also be involved. Inflammatory mediators not only can affect the lung but also can affect other body organs; if alveolar disruption allows inflammatory mediators to translocate from the lung into the systemic circulation, it is possible that they can damage other body organs. It has been shown in adult patients with ARDS that a minimal stress mechanical ventilation strategy (low tidal volumes and the use of PEEP to prevent alveolar collapse) resulted in lower levels of cytokines in the systemic circulation than an injurious ventilation strategy (high tidal volumes with no PEEP).¹² Inflammatory mediators in the systemic circulation attract neutrophils and other inflammatory cells, potentially leading to injury of other organs, such as the kidneys and liver. This phenomenon may explain the increased incidence of multisystem organ dysfunction syndrome (MODS) seen in mechanically ventilated patients with severe ARDS.¹² A more recent study confirmed that a low tidal volume ventilation strategy coupled with adequate PEEP to prevent alveolar collapse during expiration reduces circulating cytokines.¹³

Preventing Overdistention

A landmark study published by the ARDS Network in 2000 demonstrated convincingly that in ALI, mechanical ventilation with lower tidal volumes and inspiratory pressures than traditionally used significantly decreases mortality and the number of days required on the ventilator.¹¹ This study confirmed the results of earlier studies in which low volumes were used despite the resulting hypercapnia and acidosis.¹⁵ Ventilation with traditional volumes (12 mL/kg predicted body weight) was compared with lower ventilation volumes (6 mL/kg). In addition, the inspiratory plateau (alveolar) pressure in the traditional group was limited to 50 cm H₂O or less, whereas it was limited to 30 cm H₂O or less in the low-volume group. These pressure ceilings meant that the full calculated tidal volume could not be delivered if lung compliance was too low to permit its delivery; the low-volume ventilation group was most affected by this limitation as was evidenced by significantly higher PaCO₂ and lower arterial pH. Nevertheless, mortality was reduced by 22% in the low-volume group.¹¹ More recently, ARDS Network guidelines that limit plateau pressures to 30 cm H₂O or less were validated with regard to significant reduction in hospital mortality.¹⁶ Ventilation with low volumes and the limitation of plateau pressures to 30 cm H₂O or less is considered the standard of care in the mechanical ventilation of patients with ARDS.¹⁷

If the low tidal volume ventilation strategy causes hypercapnia and acidosis, it is often referred to as permissive hypercapnia.¹⁵ In other words, the adverse effects of hypercapnic acidosis that may result from low ventilation volumes are considered to be less detrimental to the patient than the alternative of overdistention lung injury. No safe upper limit for hypercapnia has been established, although some authorities believe the accompanying acidosis should not be allowed to fall below a pH of 7.20; infusion of sodium bicarbonate has been used to prevent arterial pH from falling below this level.^11,17 However, more recent evidence shows that hypercapnic acidosis may actually have a protective effect independent of lung volume.¹⁸

Preventing Repetitive Collapse and Reopening of Alveoli (Atelectrauma)

It stands to reason that a properly set PEEP level would prevent alveoli from collapsing at the end of expiration. Coupled with the low tidal volume strategy just outlined, the lung in ARDS would be protected from both overdistention and atelectrauma. That is, the lung would always be “open.” This open-lung approach was first advocated in 1992¹⁹; an often cited study published in 1998 confirmed the protective effect of this ventilatory approach.²⁰ In this study, the investigator set PEEP just above the lower inflection point on the pressure-volume curve (see Chapter 3) to prevent repetitive alveolar collapse and reopening. The 28-day mortality rate was significantly lower in patients with ARDS ventilated with an open-lung approach (38%) compared with a control group ventilated in the conventional manner (71%). The open-lung concept, or the use of enough PEEP to prevent alveolar collapse, and the limitation of plateau pressure to prevent overdistention—permitting hypercapnia if necessary—is now considered to be the standard of care against which all other strategies are measured in ventilating patients with ARDS (Figure 14-6).²¹ As discussed in Chapter 3, it is now questioned whether the lower inflection point on the inspiratory pressure-volume curve is an appropriate criterion for setting PEEP at a level that prevents alveolar collapse; the hysteresis of the pressure-volume curve shows that lung volume is considerably greater at a given airway pressure on the expiratory limb of the curve than it is on the inspiratory limb. Nevertheless, the lower inflection point continues to be used by some practitioners as a criterion for setting PEEP in ALI,²² although its use in this regard is controversial.

Prone Position: Improved V.A/Q.C Match and Reduced Ventilator-Induced Lung Injury

Conventional mechanical ventilation is accomplished with the patient in the semisitting supine position. However, the prone position has been shown to improve oxygenation significantly and to delay VILI.2,23–27

Generally, the anterior (ventral) chest wall is more compliant and free to move than the posterior (dorsal) chest wall. The vertebral column limits dorsal rib cage movements much more than the sternum limits ventral rib cage movements. This condition is amplified in the supine position in which movements of the rigid dorsal rib cage are further restricted by the body’s weight against the bed’s surface. For these reasons, ventral alveoli receive more ventilation than their dorsal counterparts (Figure 14-7, A). Moreover, in the supine position, the heart and the edematous ARDS lung weigh heavily on dorsal lung regions, possibly creating positive intrapleural pressure in this region. Consequently, dorsal alveoli tend to collapse. In mechanical ventilation, most of the tidal volume enters the more compliant ventral areas—both the lung and the rib cage are more compliant. Dorsal alveoli may open late in the inspiratory phase but then close again during expiration. Cyclical opening and closing of dorsal alveoli leads to lung injury, as previously described. For these reasons, VILI is most severe in dorsal lung regions of supine patients.²

In contrast to ventilation, pulmonary blood flow in the horizontal position (supine or prone) is relatively uniform from ventral to dorsal regions. Compared with the upright position, gravity has minimal effect on blood flow in the horizontal position because the ventral-to-dorsal lung distance is much less than the apex-to-base lung distance. Thus, dorsal lung units are very poorly ventilated with respect to their perfusion and are the site of the most significant shunting.

If the patient is turned to the prone position, the stiff dorsal component of the rib cage is freer to move than before, and the heart no longer compresses dorsal tissue (see Figure 14-7, B). The ventral rib cage is held against the bed surface,

decreasing its compliance and freedom of movement. More of the mechanically delivered volume is directed toward dorsal regions than before. However, the natural stiffness of the dorsal rib cage limits how much ventilation can shift in its direction, forcing some of the volume into ventral units as well. The result is a fairly even tidal volume distribution in the lung and a better match between ventilation and blood flow; shunting is reduced, and oxygenation improves. To the extent that dorsal alveoli remain open, lung compliance improves, lowering the required inflation pressures. Because of less closing and reopening of alveoli in dorsal regions, the prone position is associated with less VILI.²⁵

Prone positioning is generally combined with the other protective strategies already outlined, including low tidal volumes to prevent overdistention and PEEP adjusted to prevent alveolar recollapse. It is reasonable to believe that if prone ventilation is effective in improving oxygenation, it should also result in better carbon dioxide elimination, presumably through better ventilation of dorsal regions. It was found that a reduced PaCO₂ in response to prone ventilation was predictive of an improved survival rate; that is, patients in whom prone ventilation was associated with decreased PaCO₂ had a lower mortality rate than patients in whom PaCO₂ remained the same.²⁸ Another predictor of response to prone positioning may be the distribution of disease in the ARDS lung.²⁹ ARDS is categorized as having one of three distribution patterns: (1) a lobar or segmental distribution; (2) a “patchy” distribution, in which lobar and segmental distribution coexist with areas not limited by anatomical boundaries; and (3) a “diffuse” distribution, in which disease is widespread and uniform throughout the lungs. Patients with lobar and patchy distributions are more likely to respond positively to prone positioning; that is, redistribution of densities is not as likely to occur in patients with diffuse, widespread patterns.²⁹

A major meta-analysis of prone ventilation in ARDS did not show improved long-term patient survival, although oxygenation was significantly improved; however, there was a trend for lower mortality in the most severely ill patients when prone ventilation was used.²⁷ Longer duration in the prone position was associated with better outcomes; some authorities recommend maintaining the prone position for 12 to 20 hours per day.^10,27 Animal studies show reduced incidence of VILI with prone ventilation,²⁵ and it is thought that the prone position should be used sooner rather than later in the course of ARDS.^26,27 Generally, prone positioning is recommended for ARDS patients who require potentially harmful levels of F_IO₂ and plateau pressure, as long as positional changes do not place them at high risk, as might be the case in patients with spinal cord injuries.^10,17 Reversion to supine positioning at least once per day is recommended for patient washing and dressing changes; many patients with ARDS require almost continuous prone positioning for the first few days.¹⁰ Of all lung protective ventilation strategies, volume-limiting and pressure-limiting strategies are the most universally accepted and most demonstrably effective in reducing mortality in ARDS.³⁰

Pressure-Targeted versus Volume-Targeted Ventilation in Acute Respiratory Distress Syndrome

The realization that pressure-limiting ventilation strategies were associated with lower mortality in ARDS led manufacturers to develop ventilators that based the size of an inspiration on a preset pressure target rather than a preset volume target. Traditionally, intensive care ventilators were designed to cycle into expiration on delivery of a preselected volume at a preselected flow rate; in other words, these ventilators were “volume-targeted” machines. Inspiratory pressure depended on the elastic and resistive properties of the lung. If the lungs became less compliant over time, inspiratory pressure increased because the machine did not cycle into expiration until the preselected volume was delivered. In contrast, pressure-targeted machines were engineered to maintain a constant pressure at the endotracheal tube for the duration of inspiration (inspiratory pressure and inspiratory time were preselected by the clinician), regardless of lung compliance or airway resistance. Decreased lung compliance resulted in a reduced tidal volume; however, the lungs were never subjected to dangerously high pressures. Volume-targeted ventilators can be made to stop volume delivery if a pressure limit is preset; however, the ability to control pressure directly as a primary variable appealed to clinicians as a more intuitive way to protect the lung from excessive pressure.

Patients with ARDS can be effectively and safely ventilated by either type of ventilator. Low volumes can be used in volume-targeted machines, and high-pressure alarms can alert the clinician to dangerously high pressures. Likewise, desired pressure limits can be set in pressure-targeted ventilators, and low-volume alarms can alert the clinician of changing lung mechanics. However, one major difference between the two types of ventilators is the ventilator’s flow rate response to the patient’s spontaneous effort to breathe. In volume-targeted ventilators, the flow pattern is preset and does not change, whereas flow increases or decreases in response to patient demand in the pressure-targeted ventilator. By design, the pressure-targeted machine must maintain a constant pressure at the endotracheal tube for a specified time, which requires the ventilator to increase its flow output if the patient’s inspiratory effort makes the circuit pressure drop. In the final analysis, the way in which inspiratory pressure is reduced (either by limiting the tidal volume in volume-targeted ventilation or by directly limiting inspiratory pressure in pressure-targeted ventilation) has no effect on mortality.³¹ The important element influencing mortality is simply the use of low pressures and volumes.