Show Summary Details
Page of

Acute Respiratory Distress Syndrome (ARDS) 

Acute Respiratory Distress Syndrome (ARDS)
Acute Respiratory Distress Syndrome (ARDS)

John W. Kreit

Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2020. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

date: 03 July 2020


ARDS is a syndrome. In order to study its epidemiology and clinical course and enroll appropriate patients in clinical trials, it is essential to have a validated and widely accepted definition. During the past 50 years, as we have learned more about ARDS, its defining characteristics have been modified several times. In 2011, an international panel of experts convened in Berlin, Germany, in an attempt to more accurately define the syndrome and stratify its severity. After likely spending several months trying to come up with just the right name, the Berlin Definition of ARDS was published in 2012 (Table 12.1).

Table 12.1 The Berlin Definition of the Acute Respiratory Distress Syndrome


Within one week of a known clinical insult or new or worsening respiratory symptoms

Chest Imaging (Chest radiograph or CT)

Bilateral opacities not fully explained by effusions, lobar/lung collapse, or nodules

Origin of Edema

Respiratory failure not fully explained by cardiac failure or fluid overload

The consensus conference also defined three levels of severity based on the PaO2/FIO2 (P/F) ratio (Table 12.2).

Table 12.2 ARDS Severity


200 mmHg < PaO2/FIO2 ≤ 300 mmHg with PEEP or CPAP ≥ 5 cmH2O


100 mmHg < PaO2/FIO2 ≤ 200 mmHg with PEEP or CPAP ≥ 5 cmH2O


PaO2/FIO2 ≤ 100 mmHg with PEEP or CPAP ≥ 5 cmH2O

Precipitating Factors

Table 12.3 shows that some recognized triggers for ARDS cause direct lung injury, while others are extra-thoracic or systemic disorders.

Table 12.3 Common Risk Factors for ARDS

Direct Lung Injury

Indirect Lung Injury



Aspiration of gastric contents


Pulmonary contusion

Acute pancreatitis

Near drowning

Multiple trauma

Inhalation injury

Drug overdose

Fat embolism

Transfusion of blood products


ARDS is caused by acute lung injury that leads to abnormal pulmonary capillary permeability, alveolar flooding, and loss of surfactant. Studies using computed tomography (CT) imaging have identified three types of alveoli in patients with ARDS. Some are filled with edema fluid, others are collapsed (atelectatic), and some appear to be normal. In a supine patient, CT typically shows that the dependent, dorsal regions of the lungs are densely consolidated and airless, while the non-dependent, ventral areas are air-filled. Atelectatic alveoli populate the transition zone between these two regions.

Not surprisingly, during mechanical ventilation, most of the delivered gas goes to the open, relatively normal alveoli. The remainder enters previously atelectatic alveoli that are opened or “recruited” by the pressure applied during the mechanical breath. None of the delivered volume gets to the densely consolidated areas of the lungs. So, in effect, ARDS patients have very small lungs, and this is a key pathophysiological feature.

Respiratory Mechanics

Lung volume can increase only when sufficient pressure is supplied to overcome the elastic recoil of the respiratory system. Recall that elastic recoil or “stiffness” is inversely related to compliance (C), which is the change in lung volume (Δ‎V) divided by the required change in transmural pressure (Δ‎PTM).


In patients with ARDS, lung and respiratory system compliance are reduced because extra pressure is needed to recruit atelectatic alveoli and because even modest volume changes over-distend the relatively small number of gas-filled alveoli.

Gas Exchange

One of the major defining features of ARDS is arterial hypoxemia that is resistant or refractory to high inspired oxygen concentrations. This is caused by the perfusion of fluid-filled and atelectatic alveoli, which produces a right-to-left intra-pulmonary shunt. As the proportion of the cardiac output passing through unventilated alveoli (the shunt fraction) rises, increases in FIO2 have less and less effect on PaO2, and the P/F ratio falls.

The lack of ventilation to some alveoli means that other alveoli must receive too much. At the same time, the high pressure within these alveoli compresses the alveolar capillaries and reduces or stops blood flow. Over-ventilated and under- or non-perfused alveoli generate alveolar dead space, which increases the ratio of physiologic dead space to tidal volume (VD/VT). As VD/VT increases, alveolar volume and alveolar ventilation fall, and less CO2 is excreted by the lungs. The relationship between VD/VT, minute ventilation (V˙E), and CO2 excretion (V˙eCO2) can be expressed as:

V˙eCO2V˙Ex (1VD/VT)

This shows that V˙E must increase with VD/VT in order to maintain the same rate of CO2 excretion and keep PaCO2 constant. The high VD/VT explains why patients with ARDS have an abnormally elevated V˙E requirement and why low tidal volume ventilation (see later in this chapter) is often accompanied by hypercapnia.

Mechanical Ventilation

Ventilatory support for the patient with ARDS has two major goals:

  • Maintain adequate oxygenation

  • Minimize ventilator-induced lung injury

Adequate Oxygenation

If you think about it, “adequate oxygenation” is a really vague term. What does it mean? In patients with ARDS, most physicians adjust FIO2 and PEEP, and, if necessary, use ancillary therapies to keep the arterial hemoglobin saturation (SaO2) in the range of 91–93%. Given the sigmoid shape of the oxygen–hemoglobin dissociation curve, this goal is reasonable because it prevents small decrements in gas exchange from causing a precipitous fall in SaO2. Although it’s easy to continuously monitor SaO2 using pulse oximetry (SpO2), is this really the best measure of “oxygenation”?

Remember from Chapter 9 that hemoglobin saturation is only one factor that determines the oxygen content of arterial blood (CaO2). If the very small volume of dissolved O2 is neglected, CaO2 (in ml of O2 per dl of blood) is the product of the volume of O2 carried by one gram (g) of fully saturated hemoglobin (1.34 ml/g), the hemoglobin concentration (Hb; g/dl), and the fractional hemoglobin saturation (SaO2/100).


But there’s something even more important than O2 content, and that’s the rate at which O2 is delivered to the tissues and organs of the body. Oxygen delivery (D˙O2) is expressed in terms of ml of O2 per minute and is calculated by multiplying the arterial O2 content by the cardiac output (CO).


Substituting Equation 12.3 into Equation 12.4 gives us:


Note that the components of O2 content have been multiplied by a factor of 10 to convert ml/dl to ml/L. Now, when we multiply by CO (in L/min), DO2 will be in units of ml/min.

It’s logical (although unproven) that ensuring adequate O2 delivery is a more important and appropriate goal than simply maintaining the SaO2 on the top portion of the dissociation curve. Focusing on O2 delivery also helps us to remember that hemoglobin saturation is relatively unimportant. Since Equation 12.5 uses the fractional saturation (SaO2/100), you can see that a small change in hemoglobin concentration or cardiac output affects O2 delivery more than a larger change in SaO2. For example, SaO2 would have to increase from 80% to 90% to produce the same change in O2 delivery as an increase in hemoglobin concentration from 8 to 9 g/dl or an increase in CO from 6.0 to 6.75 L/min. This is especially important because PEEP often reduces venous return, LV preload, and cardiac output. Increasing SaO2 from 80% to 90% by raising PEEP from 10 to 15 cmH2O might make you feel good, but it’s unlikely to help your patient if it also causes a significant fall in cardiac output and systemic oxygen delivery.

The main obstacle to monitoring O2 delivery, of course, is that it’s not nearly as convenient as simply watching the continuous SpO2 display. It requires the repeated measurement of hemoglobin concentration and cardiac output—and then there’s that equation. Measuring hemoglobin concentration obviously isn’t a problem, and cardiac output can now be assessed using a number of minimally invasive or noninvasive techniques. But here’s the thing. You don’t really need to calculate O2 delivery. All you need to do is follow cardiac output or stroke volume as you increase PEEP. If there’s a big drop, try to correct it with fluid boluses. As we’ll discuss, PEEP does more than reduce shunt fraction and increase PaO2 and SaO2. It also plays an important role in limiting lung injury from mechanical ventilation, so some tradeoff between O2 delivery and lung protection is often necessary.

A bigger problem is that we currently have no accurate way of determining whether O2 delivery is “adequate.” That’s why I’ve intentionally used such an inexact term. Furthermore, since blood flow and metabolic rate vary considerably throughout the body, adequate O2 delivery differs from one organ or tissue to another. Unfortunately, the best we can do is watch for signs of total body hypoxia, such as impaired organ function, elevation of the serum lactate concentration, and reduced central venous (Scv¯O2) or mixed venous (Sv¯O2) hemoglobin saturation. When these signs are present, you should consider augmenting O2 delivery by transfusing red cells, increasing cardiac output, or both.

Ventilator-Induced Lung Injury

In the latter part of the 20th century, studies in animals and humans provided convincing evidence that mechanical ventilation itself can cause lung injury and worsen or perpetuate ARDS. This so-called ventilator-induced lung injury (VILI) was shown to result from both alveolar over-distention during mechanical inflation and the repeated recruitment and de-recruitment of terminal respiratory units during each respiratory cycle. This led to the concept of “protective lung ventilation,” in which tidal volume and PEEP are adjusted to avoid alveolar over-distention while maintaining the recruitment achieved during positive pressure breaths.

Tidal Volume

Low tidal volume ventilation has been the standard of care since 2000, when a prospective, randomized trial by the Acute Respiratory Distress Syndrome Network (ARDSnet) showed that patients receiving a tidal volume ≤6 ml/kg of ideal body weight (IBW) that maintained a plateau pressure ≤30 cmH2O had significantly lower mortality than those receiving a tidal volume of 12 ml/kg IBW.

Any decrease in tidal volume must raise VD/VT, which further impairs CO2 excretion and increases the minute ventilation needed to maintain a given PaCO2 (Equation 12.2). That’s why large drops in tidal volume should be avoided. Instead, there should be a stepwise reduction in tidal volume until the target of 6 ml/kg IBW is reached. At each step, arterial PCO2 and pH are measured and respiratory rate (and minute ventilation) is increased as needed. Even with this stepwise approach, respiratory rate often cannot be increased sufficiently to maintain PaCO2 within the normal range. Fortunately, this so-called permissive hypercapnia is usually well-tolerated. If necessary, intermittent boluses of IV sodium bicarbonate can be used to keep the arterial pH above 7.20.

Positive End-Expiratory Pressure

Since the initial description of ARDS in 1967, PEEP has been an essential part of management. By increasing end-expiratory lung volume, PEEP opens or “recruits” collapsed alveoli, thereby reducing shunt fraction and increasing PaO2 and SaO2. Although PEEP was mentioned in the previous section on “oxygenation,” I will focus on it here because of its potential to improve lung compliance and reduce the cyclical opening and closing of alveoli that leads to lung injury.

Unfortunately, PEEP can also have two important detrimental effects. First, PEEP may reduce venous return, LV preload and stroke volume, and cardiac output (Chapter 3). This means that PEEP may improve PaO2, SaO2, and O2 content while decreasing oxygen delivery to the tissues. Second, PEEP-induced elevation of end-expiratory pressure and volume may worsen lung injury by contributing to alveolar over-distention during a mechanical breath.

Three large, prospective trials that randomized ARDS patients to high or low PEEP were unable to show a difference in survival rates. This could mean that the level of PEEP really doesn’t matter. Alternatively, it could mean that this “one size fits all” approach ignores the need to select PEEP based on the characteristics, extent, and duration of each patient’s lung disease. Since the volume of “recruitable” lung has been shown to vary widely among patients, I think the second possibility is much more likely.

Unfortunately, there is no well-accepted method of identifying the level of PEEP that optimizes gas exchange while minimizing cyclical alveolar recruitment-de-recruitment and over-distention. This elusive ideal pressure is often referred to as “best PEEP.” Attempts to individualize PEEP have focused primarily on techniques that assess respiratory mechanics, gas exchange, or both. The most commonly used methods are listed in Table 12.4. It’s important to recognize that the impact of most of these techniques on any meaningful patient outcome has never been studied.

Table 12.4 Methods of Selecting “Best PEEP”

Gas exchange

PEEP is set at the lowest level needed to maintain adequate SpO2.

PEEP is selected to maximize SpO2 or systemic O2 delivery.

PEEP is chosen to maximize CO2 excretion or minimize alveolar or physiologic dead space.

Respiratory mechanics

PEEP is chosen to maximize respiratory system compliance.

PEEP is set to maintain the “stress index” between 0.9 and 1.1.

PEEP is set just above the lower inflection point of the pressure–volume curve.


PEEP is chosen from a table listing combinations of FIO2 and PEEP.

Before I describe each approach, I want to review two important concepts: recruitment maneuvers and PEEP titration. A recruitment maneuver is performed by applying high levels of positive pressure for a short period of time. For example, pressure-control breaths might be used to provide a PEEP of 30 cmH2O for two minutes, while driving pressure is set to maintain the previously set tidal volume. The goal is to maximize alveolar recruitment, which can hopefully be maintained on a much lower PEEP setting. To this end, recruitment maneuvers are often followed by decremental PEEP titration (Figure 12.1). At each PEEP level, respiratory mechanics, gas exchange, or both, are assessed, and the results are used to select best PEEP. Alternatively, best PEEP can be based on measurements during stepwise increases in PEEP (incremental titration) without a recruitment maneuver. It’s important to recognize, though, that both animal and human studies have shown that, at the same PEEP level, mechanics and gas exchange are much better following a recruitment maneuver.

Figure 12.1 The airway pressure (PAW) gradient of pressure control breaths during a recruitment maneuver (RM) and subsequent decremental PEEP titration. The top and bottom of each block represents end-inspiratory and end-expiratory pressure, respectively, at each PEEP level.

Figure 12.1 The airway pressure (PAW) gradient of pressure control breaths during a recruitment maneuver (RM) and subsequent decremental PEEP titration. The top and bottom of each block represents end-inspiratory and end-expiratory pressure, respectively, at each PEEP level.

Methods Assessing Gas Exchange

Least PEEP 

By far, the most common way of setting PEEP is to simply titrate it up or down based on the FIO2 and SpO2. If FIO2 is high, PEEP is increased until a minimum SpO2 (often 90–93%) is reached. As the patient improves, PEEP is reduced before, after, or with FIO2 in order to keep SpO2 within this same range. This is often referred to as the “least PEEP” method, because the lowest possible PEEP is used to achieve a predetermined SpO2.

There are several advantages to this approach. It’s easy to do, PEEP is repeatedly assessed and adjusted throughout the course of the disease, and there is a reasonable correlation between SpO2 and alveolar recruitment. On the other hand, PEEP is not directly linked with alveolar recruitment because the selected level is dependent on the FIO2. For example, best PEEP will be very different depending on whether a patient is receiving an FIO2 of 1.0 or 0.5. Furthermore, there is no mechanism for detecting alveolar recruitment-de-recruitment or over-distention.

Highest SpO2 or D˙O2

A variant of the least PEEP approach is to perform incremental or decremental PEEP titration and identify the level that maximizes SpO2 or systemic oxygen delivery. As discussed previously, I believe that oxygen delivery is a much more appropriate target. Unfortunately, calculating oxygen delivery at multiple PEEP levels is time-consuming and requires repeated measurements of cardiac output. Titrating PEEP based on SpO2 takes relatively little time, but like the least PEEP approach, best PEEP depends on FIO2, and there is no way to assess for PEEP-induced VILI.

Highest CO2 Excretion and Lowest Dead Space 

Alternatively, ventilation can be assessed instead of oxygenation. Volume capnography continuously displays exhaled volume versus the expired CO2 fractional concentration (FCO2) or partial pressure (PCO2) (Chapter 9). These data can be used to determine the volume of CO2 exhaled per breath and per minute, the volume of the physiologic and alveolar dead space, and several dead-space-to-volume ratios. Alveolar recruitment augments CO2 excretion by increasing the surface area of the gas–blood interface and reduces alveolar and physiologic dead space by diverting gas from previously over-ventilated alveoli. Alveolar over-distension, on the other hand, leads to excessive alveolar volume and reduced capillary blood flow, which generates high V˙/Q˙ regions, increases alveolar and physiologic dead space, and reduces CO2 excretion. Accordingly, best PEEP is the level associated with the highest CO2 excretion and the lowest alveolar and physiologic dead space (Figure 12.2).

Figure 12.2 CO2 excretion (V˙eCO2) and the dead space to tidal volume ratio (VD/VT) during decremental PEEP titration. In this example, best PEEP is 12 cmH2O because it corresponds to the highest V˙e CO2 and the lowest VD/VT.

Figure 12.2 CO2 excretion (V˙eCO2) and the dead space to tidal volume ratio (VD/VT) during decremental PEEP titration. In this example, best PEEP is 12 cmH2O because it corresponds to the highest V˙e CO2 and the lowest VD/VT.

Methods Assessing Respiratory Mechanics

All three techniques listed in Table 12.4 are based on the concept that alveolar recruitment increases lung compliance by distributing the tidal volume to more alveoli, and that compliance falls as the lungs are stretched by alveolar over-distention.

Highest Compliance 

One method of determining best PEEP is to calculate respiratory system compliance from measurements of plateau and end-expiratory pressure (Chapter 9) during stepwise PEEP titration. The level that produces the highest compliance is best PEEP (Figure 12.3). Alternatively, dynamic compliance (CDYN) can be used. “Dynamic compliance” is an oxymoron (think “jumbo shrimp”) because compliance can only be measured under the static condition of zero gas flow. Nevertheless, this term has been around for decades and is unlikely to go away. It is calculated by dividing the delivered tidal volume (VT) by the difference between the peak (PPEAK) and the total end-expiratory pressure (PEEPT) during a passive, volume-set breath.


Figure 12.3 Respiratory system (static) compliance (CRS) or dynamic compliance (CDYN) during decremental PEEP titration. In this example, best PEEP is 16 cmH2O because it corresponds to the highest compliance.

Figure 12.3 Respiratory system (static) compliance (CRS) or dynamic compliance (CDYN) during decremental PEEP titration. In this example, best PEEP is 16 cmH2O because it corresponds to the highest compliance.

The advantage of substituting dynamic for true “static” compliance is that it can be measured without performing an end-inspiratory pause, and many ventilators automatically calculate it on a breath-by-breath basis.

A large, randomized trial recently reported higher patient mortality when PEEP was adjusted to maximize compliance. Although several explanations have been proposed, this method of choosing PEEP cannot currently be recommended.

Optimal Stress Index 

During a mechanical breath, the pressure needed to balance the elastic recoil of the respiratory system increases with lung volume, and the slope of pressure versus volume equals the elastance (the reciprocal of compliance) of the respiratory system. If inspiratory flow is constant, time can be substituted for volume, and the pressure required to overcome viscous forces can be assumed to be constant. This means that the slope of airway pressure (PAW) versus time during each mechanical breath approximates respiratory system elastance (Figure 12.4). If elastance is constant throughout inspiration, the slope of this relationship will also be constant (Figure 12.4A). If elastance increases (compliance falls), progressively more pressure is needed to balance elastic recoil, and the plot of PAW vs. time will be convex (Figure 12.4B). If elastance falls (compliance increases), the curve will be concave (Figure 12.4C). Studies in both ARDS patients and animal models suggest that increasing elastance reflects alveolar over-distention (too much PEEP), whereas falling elastance reflects ongoing alveolar recruitment during inspiration and presumed de-recruitment during expiration (insufficient PEEP).

Figure 12.4 During a mechanical breath with constant inspiratory flow, the slope of the airway pressure (PAW)–time relationship equals the elastance of the respiratory system. If elastance remains constant (A), the slope is also constant. If the slope increases (B), elastance increases (compliance falls). If the slope decreases (C), elastance falls (compliance increases).

Figure 12.4 During a mechanical breath with constant inspiratory flow, the slope of the airway pressure (PAW)–time relationship equals the elastance of the respiratory system. If elastance remains constant (A), the slope is also constant. If the slope increases (B), elastance increases (compliance falls). If the slope decreases (C), elastance falls (compliance increases).

Although a significant change in elastance can sometimes be determined visually, the degree of convexity or concavity has also been quantified by fitting the PAW–time curve to a power equation containing three variables (a, b, and c) and the inspiratory time (TI):

PAW= aTIb+ c

In this equation, the coefficient b is referred to as the “stress index” and reflects the slope of the PAW–time curve. If b = 1, the curve is straight, indicating no change in elastance during the mechanical breath. If b > 1, elastance increases, and if b < 1, elastance falls during inspiration. At least one ventilator manufacturer has incorporated software to constantly calculate the stress index. Using this method, PEEP may be adjusted up or down to maintain the stress index between 0.9 and 1.1.

Lower Inflection Point 

In many ARDS patients, the plot of volume versus pressure, when measured in the absence of gas flow, has a characteristic S-shaped appearance (Figure 12.5). Since the slope of this curve equals compliance, some investigators have equated the abrupt increase in slope at the “lower inflection point” (LIP) with the pressure needed to recruit all available alveoli (i.e., best PEEP). Using the same reasoning, flattening of the curve at the “upper inflection point” (UIP) indicates alveolar over-distention and suggests an excessively high tidal volume.

Figure 12.5 A plot of lung volume versus airway pressure (PAW) in the absence of gas flow. The slope of the curve equals respiratory system compliance. The lower inflection point (LIP) may indicate the pressure needed for optimal alveolar recruitment (best PEEP). The upper inflection point (UIP) may identify optimal end-inspiratory pressure.

Figure 12.5 A plot of lung volume versus airway pressure (PAW) in the absence of gas flow. The slope of the curve equals respiratory system compliance. The lower inflection point (LIP) may indicate the pressure needed for optimal alveolar recruitment (best PEEP). The upper inflection point (UIP) may identify optimal end-inspiratory pressure.

Since it’s very difficult and time-consuming to plot a static volume–pressure curve in patients with ARDS, a “real-time” curve generated during a passive mechanical breath with slow, constant flow can be used instead. This “quasi-static” technique (another oxymoron) minimizes the pressure needed to overcome viscous forces and simulates a static volume–pressure curve.

This method of determining best PEEP has a number of major drawbacks that have limited its use. First, accurate determination of the LIP requires that measurements be performed on zero PEEP. This is potentially hazardous in patients with ARDS, especially if repeated measurements are performed. Second, even when performed correctly, there is no discernable LIP in a significant portion of patients. Finally, some studies suggest that alveolar recruitment continues well beyond the LIP.

The “Table” Method

I have included the “table” method approach to PEEP selection because many trainees mistakenly think that it’s evidence-based. I’m talking about the tables pairing levels of FIO2 and PEEP that have been published with a number of randomized ARDS trials. Most are identical to the one used in the ARDSnet low tidal volume trial (Table 12.5). Since the investigators of these studies wanted to determine the effect of a specific intervention, it was essential that all other management be identical for both patient groups. This required a number of treatment protocols, including one linking FIO2 and PEEP. These tables were developed by the investigators to standardize care. They are not an evidence-based method for selecting the appropriate level of PEEP.

Table 12.5 Permitted FIO2–PEEP Combinations





































Putting It All Together

My recommendations for initial ventilator settings in patients with ARDS are shown in Table 12.6.

Table 12.6 Initial Ventilator Settings for Patients with ARDS



Breath type

VC or aPC

Tidal volume

≤6 ml/kg IBW*

Respiratory rate





5.0 cmH2O

* As needed to keep PPLAT ≤ 30 cmH2O

CMV = continuous mandatory ventilation; VC = volume control; aPC = adaptive pressure control; IBW = ideal body weight

I use the CMV mode and volume-set breaths because this combination provides a guaranteed minute ventilation while allowing the patient to easily increase it if necessary. Both volume control (VC) and adaptive pressure control (aPC) breaths are acceptable. Pressure control (PC) breaths should be used with caution because tidal volume can vary with changes in compliance, resistance, and patient effort. If you choose VC breaths, increase inspiratory flow as needed to optimize patient comfort and ventilator synchrony (Chapter 11). If aPC is used, monitor the inspiratory pressure gradient to detect excessive patient work of breathing (Chapter 11, Figure 11.6).

Always start with an FIO2 of 1.0, because this maximizes your ability to quickly correct arterial hypoxemia. If possible, FIO2 can be reduced later based on PaO2 and SpO2.

PEEP can initially be set at 5 cmH2O, but best PEEP should be determined as quickly as possible. Although none of the methods listed in Table 12.4 have been shown to improve patient morbidity or mortality, a few are more evidence-based than the others. In particular, several groups of investigators have quantified the extent of alveolar recruitment and over-distention in ARDS patients by performing CT after a recruitment maneuver and at each level during decremental PEEP titration. In these studies, the level of PEEP that maximized CO2 excretion and minimized dead space correlated with optimal alveolar recruitment.

I recommend the following approach:

  • If possible, measure baseline cardiac output.

  • Perform a recruitment maneuver for 2–3 minutes using 30 cmH2O PEEP and a pressure-control breath set to deliver a tidal volume ≤6 ml/kg IBW.

  • Return to volume-set breaths with a PEEP of 24 cmH2O. After 10 minutes, measure CO2 excretion, VD/VT or both.

  • Reduce PEEP in steps of 4 cmH2O and repeat measurements after about 10 minutes at each level. The decremental PEEP titration is stopped when CO2 excretion falls, when VD/VT increases, or when a PEEP level of 8 cmH2O has been reached.

  • Perform a second recruitment maneuver, then resume volume-set ventilation with PEEP set at the level corresponding to the highest CO2 excretion and/or lowest VD/VT (best PEEP).

  • Measure cardiac output. If there has been a significant fall, attempt to restore it to baseline with volume boluses.

  • Reduce FIO2 until SpO2 is in the 90–92% range.

  • Repeat this process once a day.

Ancillary Therapies

In most patients with ARDS, adequate oxygenation and lung protection can be maintained with low tidal volume ventilation using volume-set breaths, high FIO2, and best PEEP. Some patients, though, require ancillary therapies, and these can be divided into ventilatory and non-ventilatory types (Table 12.7). All of the listed therapies may increase PaO2 and SaO2, and experimental data suggest that several may improve patient survival.

Table 12.7 Ancillary Therapies for ARDS



High I:E ventilation

  • Conventional breaths

  • Bi-level ventilation

  • Airway pressure release ventilation

Prone positioning

Neuromuscular blockade

Inhaled vasodilators

Extracorporeal membrane oxygenation

Ventilatory Therapies

I will briefly review these techniques for the sake of completeness. However, since none have been shown in improve patient outcome and all have the potential to reduce systemic oxygen delivery and worsen ventilator-induced lung injury, I don’t recommend them.

High I:E Ventilation

Recall that the ratio of inspiratory to expiratory time (I:E) is normally much less than one. That is, the duration of inspiration is usually a fraction of the time between breaths. As the I:E ratio increases, so does the mean alveolar pressure (MAP), which is the alveolar pressure averaged over the entire respiratory cycle. In general, an increase in MAP improves PaO2 and SaO2. An I:E ratio greater than 1.0 is referred to as inverse ratio ventilation (IRV).

Most often, the I:E ratio is increased by prolonging the set inspiratory time (Figure 12.6). Alternatively, on ventilators that require a set peak flow during VC breaths, an end-inspiratory pause can be added to produce the desired ratio.

Figure 12.6 Plots of airway pressure (PAW) vs. time during pressure control (A) and volume control (B) breaths with an I:E ratio of 1:2 and 2:1. The shaded areas equal mean alveolar pressure (MAP), which increases with I:E ratio and PEEP.

Figure 12.6 Plots of airway pressure (PAW) vs. time during pressure control (A) and volume control (B) breaths with an I:E ratio of 1:2 and 2:1. The shaded areas equal mean alveolar pressure (MAP), which increases with I:E ratio and PEEP.

Another option is to use bi-level ventilation, which is available on most ventilators. As discussed in Chapter 5, the bi-level mode alternates between a high (PH) and a low (PL) clinician-selected airway pressure, and patients may trigger spontaneous breaths at both levels (see Figure 5.3). The clinician also selects the duration of PH and PL, which determines the I:E ratio. Airway pressure release ventilation (APRV) is a variant of bi-level ventilation that uses a high PH:PL ratio.

By increasing MAP, high I:E ratios, like PEEP, increase pleural pressure, which can reduce venous return, cardiac output, and blood pressure. These effects become especially prominent when insufficient expiratory time causes dynamic hyperinflation and intrinsic PEEP. That’s why it’s essential that cardiac output be monitored and both the beneficial and detrimental effects of high I:E ventilation be considered before abandoning more conventional forms of mechanical ventilation.

Non-Ventilatory Therapies

Prone Positioning


It has long been recognized that turning patients with ARDS from the supine to the prone position usually increases PaO2 and SaO2. This was initially thought to result simply from the effect of gravity on lung perfusion. That is, the position change was assumed to reduce the shunt fraction by decreasing blood flow to the unventilated, fluid-filled or atelectatic alveoli in the dorsal, previously dependent regions and increasing flow to ventral, well-ventilated, and newly dependent alveoli. As usual, it turns out that things are a bit more complicated.

Studies in ARDS patients and in animal models of acute lung injury have revealed two key results of prone positioning. First, CT scans show that, within a short period of time, consolidation and atelectasis move from the dorsal to the now-dependent, ventral portions of the lungs. Second, in the prone position, most of the pulmonary blood flow is still directed to the non-dependent, dorsal lung regions. Taken together, this means that prone positioning improves arterial oxygenation by increasing the ventilation of persistently well-perfused dorsal lung regions.

Animal studies have also shown that prone positioning reduces the injurious effects of mechanical ventilation. Although there are several possible reasons for this, the most likely is that pleural pressure (PPL) is much more uniform in the prone than in the supine position. Recall from Chapter 3 that lung transmural pressure (PlTM) is the difference between the pressure inside (alveolar pressure; PALV) and outside (pleural pressure; PPL) the alveoli.


In the supine position, PPL is much greater (less negative or even positive) over the dorsal surfaces than over the ventral surfaces of the lungs. Because PALV is fairly uniform, PlTM and alveolar volume progressively increase from the dorsal to the ventral portions of the lungs. This predisposes to alveolar collapse in dorsal lung regions and over-distention in ventral regions. During mechanical ventilation, cyclical opening and closing of collapsed lung units and inflation of already distended alveoli lead to lung injury.

In the prone position, there is little difference in PPL and PlTM between the non-dependent and dependent regions of the lungs, and this causes alveolar distending pressures and volumes to be much more uniform. This, in turn, is thought to reduce alveolar over-distention and cyclical recruitment and de-recruitment.

Clinical Evidence

Despite evidence that it reduces ventilator-induced lung injury, it has been difficult to prove that prone positioning, like low tidal volume ventilation, reduces mortality. Since 2001, there have been five large, prospective trials in which patients with ARDS have been randomized to supine or intermittent prone positioning (Table 12.8).

Table 12.8 Randomized Trials Comparing Supine and Prone Positioning













Mean PaO2/FIO2






Average duration of prone positioning

7 hours for 5 days

9 hours for 4 days

17 hours for 10 days

18 hours for 8 days

17 hours for 4 days


6 months

3 months

Hospital discharge

6 months

3 months

  • Mortality:

  • prone vs. supine

62.5% vs. 58.6%

43.3% vs. 42.2%

50% vs. 60%

47% vs. 52.3%

23.6% vs. 41%

In the first four trials, prone positioning did not confer a survival benefit, although several meta-analyses suggested that the most severely ill patients were most likely to benefit. In the fifth (and probably final) trial, published in 2013, 466 patients with moderate or severe ARDS (defined as PaO2/FIO2 < 150 with PEEP ≥ 5 cmH2O and FIO2 ≥ 0.6) were randomized to prone-positioning for at least 16 hours per day or to be left in the supine position. Daily prone positioning was continued until day 28 or until there was significant, predefined improvement in gas exchange while in the supine position. Three-month mortality was 23.6% in the prone group and 41.0% in the supine group (P < 0.001). This dramatic effect, and the failure of previous trials to demonstrate it, was attributed primarily to greater severity of illness and the long periods of daily proning. It is worth noting, however, that these factors were very similar in the two negative studies that preceded it. Furthermore, the investigators found that the mortality benefit of prone positioning did not vary with ARDS severity. Finally, the generalizability of these results must be questioned, since only 1,434 out of 3,449 potential subjects with ARDS were screened for the study, and of these, only 474 were randomized.

Neuromuscular Blockade


Pharmacological paralysis can significantly increase the PaO2 and SaO2 of patients with severe ARDS. Although several mechanisms have been proposed, the associated increase in the saturation of mixed venous blood (Sv¯O2) is probably the most important. Let’s examine this mechanism.

Sv¯O2 is directly proportional to systemic O2 delivery (D˙O2) and inversely related to total body O2 consumption (V˙O2).


If more O2 is delivered to the tissues but consumption doesn’t change, there is more O2 “left over,” so the mixed venous blood contains more O2, andSv¯O2 rises. If O2 delivery doesn’t change but less is used, the result is the same: mixed venous O2 content and Sv¯O2 increase. Neuromuscular blockade increases Sv¯O2 by preventing respiratory and skeletal muscle contraction, thereby reducing O2 consumption.

Now look at Figure 12.7. In the presence of a large shunt fraction, SaO2 is the weighted average of the saturation of blood leaving ventilated alveoli and the saturation of mixed venous blood that passes unchanged through the lungs. Therefore, any increase in Sv¯O2 must increase SaO2 and arterial O2 content.

Figure 12.7 In patients with ARDS, SaO2 is the weighted average of the saturation of pulmonary venous blood (SPVO2) passing through ventilated and unventilated lung regions (depicted as bypassing the lungs). Here, the shunt fraction is 30%, and SaO2 increases from 85% to 88% when the mixed venous saturation (SvO2) increases from 50% to 60%.

Figure 12.7 In patients with ARDS, SaO2 is the weighted average of the saturation of pulmonary venous blood (SPVO2) passing through ventilated and unventilated lung regions (depicted as bypassing the lungs). Here, the shunt fraction is 30%, and SaO2 increases from 85% to 88% when the mixed venous saturation (SvO2) increases from 50% to 60%.

Clinical Evidence

In a multi-center, prospective trial, 340 patients with moderate to severe ARDS (PaO2/FIO2 < 150 on PEEP ≥ 5 cmH2O) were randomized to receive continuous cisatracurium or placebo for 48 hours. Ninety-day mortality was 31.6% in the cisatracurium group and 40.7% in the control group (p = 0.08). When adjusted for PaO2/FIO2, the baseline Simplified Acute Physiology Score (SAPS) II, and plateau pressure, patients receiving cisatracurium had a hazard ratio for death at 90 days of 0.68 (confidence interval [CI] 0.48–0.98) when compared with the placebo group. As noted by the study authors, the mechanisms responsible for these beneficial effects remain the subject of speculation.

Inhaled Vasodilators


Both nitric oxide and epoprostenol are potent vasodilators that can be administered through the ventilator circuit. Because they are given by inhalation, these drugs selectively dilate the pulmonary vessels in the ventilated regions of the lungs. This reduces local vascular resistance, which “steals” blood from unventilated regions, thereby reducing the shunt fraction and improving PaO2 and SaO2.

Clinical Evidence

Three moderate-sized randomized trials and several meta-analyses have shown that nitric oxide improves PaO2 and SaO2 in patients with ARDS but has no effect on mortality. Epoprostenol has been shown to be as effective as nitric oxide at improving arterial oxygenation at a fraction of the cost. No sufficiently powered trials have assessed the effect of epoprostenol on patient morbidity or mortality.

Extracorporeal Membrane Oxygenation (ECMO)


Extracorporeal support for severe, refractory hypoxemia is performed using a veno-venous circuit (Figure 12.8). Blood is removed from the inferior vena cava using a large-bore catheter inserted via the femoral vein, pumped through a membrane that allows oxygen to be added and carbon dioxide to be removed, reheated to body temperature, and returned to the superior vena cava through an internal jugular vein catheter. If flow through the circuit approaches the cardiac output, mechanical ventilation isn’t needed for gas exchange, and the ventilator can be set at a low respiratory rate, tidal volume, and FIO2. ECMO would appear to be perfectly suited for ARDS patients because it maintains adequate gas exchange while minimizing mechanical ventilation and ventilator-induced lung injury. The downside, of course, is that ECMO is invasive and associated with a number of major complications, including thrombosis and embolism, major bleeding, hemolysis, and infection.

Figure 12.8 During veno-venous ECMO, blood is usually removed from the right femoral vein and returned through the right internal jugular vein. The ECMO circuit consists of cannulae and tubing, a blood pump, an oxygenator, and a heat exchanger.

Figure 12.8 During veno-venous ECMO, blood is usually removed from the right femoral vein and returned through the right internal jugular vein. The ECMO circuit consists of cannulae and tubing, a blood pump, an oxygenator, and a heat exchanger.

Clinical Evidence

In the only relevant prospective trial, 180 patients with “severe but potentially reversible” ARDS were randomized to receive conventional therapy or to be transferred to a single ECMO center. Six-month survival without severe disability was significantly greater in the patients transferred to the ECMO center (63% vs. 47%). Unfortunately, this trial was far from conclusive, because only 75% of the transferred patients actually received ECMO and because there were significant treatment differences between the two groups. Pending the results of truly randomized trials, veno-venous-ECMO should be considered only in patients dying from tissue and organ hypoxia.

Additional Reading

Bein T, Grasso S, Moerer O, et al. The standard of care of patients with ARDS: Ventilatory settings and rescue therapies for refractory hypoxemia. Intensive Care Med. 2016;42:699–711.Find this resource:

Fan E, Del Sorbo L, Goligher EC, et al. An official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine clinical practice guideline: Mechanical ventilation in adult patients with acute respiratory distress syndrome. Am J Respir Crit Care Med. 2017;195:1253–1263.Find this resource:

Hess DR. Recruitment maneuvers and PEEP titration. Respir Care. 2015;60:1688–1704.Find this resource:

Kallet RH. Should PEEP titration be based on chest mechanics in patients with ARDS? Respir Care. 2016;61:876–890.Find this resource:

MacSweeney R, McAuley DF. Acute respiratory distress syndrome. Lancet. 2016;338: 2416–2430.Find this resource:

Marini JJ, Josephs SA, Mechlin M, Hurford WE. Should early prone positioning be a standard of care in ARDS with refractory hypoxemia? Respir Care. 2016;61:818–829.Find this resource:

Munshi L, Rubenfeld G, Wunsch H. Adjuvants to mechanical ventilation for acute respiratory distress syndrome. Intensive Care Med. 2016;42:775–778.Find this resource:

Rittayamai N, Brochard L. Recent advances in mechanical ventilation in patients with acute respiratory distress syndrome. Eur Respir Rev. 2015;24:132–140.Find this resource:

Vieillard-Baron A, Matthay M, Teboul JL, et al. Experts’ opinion on management of hemodynamics in ARDS patients: Focus on the effects of mechanical ventilation. Intensive Care Med. 2016;42:739–749.Find this resource: