Bongard Frederic , Darryl Sue. Diagnosis and Treatment Critical Care

Подождите немного. Документ загружается.



RESPIRATORY FAILURE

297

and lymphocytes and their cytokines, prostaglandins,

leukotrienes, platelets, coagulation factors, adhesion mole-

cules, and immunoglobulins, as well as exogenous substances

such as endotoxin and other products of bacteria and

fungi. Endogenous cell products have received the most

attention—some as mediators of injury, such as oxygen rad-

icals and proteolytic and elastolytic enzymes, but others as

amplifiers of inflammation and injury, such as interleukins,

platelet-activating factor, complement, and other substances

that are chemotactic, bronchoreactive, or vasoreactive. These

may be active in the early, middle, or late phases of lung

injury. A role for injury from oxygen radicals is supported by

the finding of reduced alveolar fluid glutathione in patients

with ARDS. The coagulation system has been suggested by

some investigators as having a central role in lung injury, per-

haps by linking intravascular events to direct injury to the

endothelium and by activation of inflammatory sequences.

Elevated plasma levels of tumor necrosis factor (TNF) are

found in some patients with ARDS but also in those with

sepsis and other systemic disorders. A potent cytokine, TNF

has a variety of systemic effects, some of which could cause

or potentiate lung damage. The finding of several elevated

cytokines in ARDS suggests the possibility of common regu-

latory factors being involved. One factor, NF-κB, regulates

production of TNF, interleukin 1 (IL-1), IL-6, and IL-8. This

hypothesis is attractive because of the frequent association of

IL-6, TNF, and IL-8 with lung injury. More recently, ARDS

has been linked to toll-like receptors (TLRs), which respond

to a variety of substances to trigger a vast cytokine response.

TLRs are responsible for innate immunity; this could

explain the common finding of ALI in response to the range

of inciting factors.

A key role for polymorphonuclear leukocytes in lung

injury is supported by finding neutrophils in large numbers

in the lungs of ARDS patients. Furthermore, neutrophils are

primed to release potentially toxic substances from their

granules, neutrophil chemotactic factors and activators are

increased, and in some animal models, lung injury is attenu-

ated after neutrophil depletion. For example, neutrophil-

activating protein/interleukin-8 (NAP-1/IL-8) has been

found in high concentrations in alveolar fluid, and there was

a correlation with the number of neutrophils. High concen-

trations of NAP-1/IL-8 also were associated with poor clinical

outcome. On the other hand, neutropenic cancer patients

may develop ARDS indistinguishable from that observed in

nonneutropenic patients, and diffuse alveolar damage in

some animal models does not require the presence of neu-

trophils. Levels of both cytokines and modulators of cytokine

function are highly variable in ARDS, and it is clear that

cytokines taken individually or as patterns of response are not

able to predict development or prognosis of ARDS. In paral-

lel with the diversity of clinical conditions associated with dif-

fuse alveolar damage and ALI, it is highly likely that different

conditions in different patients explain why consistent find-

ings cannot be identified. While this makes a single common

causative mechanism unlikely, this hypothesis helps to

explain why so many conditions can result in very similar his-

tologic and physiologic features. Nevertheless, there is now

ample evidence that persistent elevation of inflammatory

cytokines in blood or alveolar fluid is associated with poor

outcome in ARDS in all forms of this disorder.

Patients may develop secondary bacterial or fungal pneu-

monia during the course of ARDS, further confusing the pic-

ture. Administration of high concentrations of inspired oxygen

contributes to lung injury, and high airway pressure and rela-

tively high tidal volume during mechanical ventilation are

closely linked to worsening pulmonary edema and fibrosis. On

the other hand, higher oxygen requirements and airway pres-

sures may simply indicate more severe underlying disease.

B. Noncardiogenic Pulmonary Edema—Normal lungs are

kept very dry to permit efficient gas exchange, and the struc-

ture and activity of the lungs maintain only a small amount

of fluid in the lungs. Normal lungs have tight junctions

between alveolar epithelial cells, an extensive lymphatic sys-

tem, low hydrostatic pressure in the pulmonary capillaries,

and other mechanisms to avoid pulmonary edema. Thus,

lung injury from any number of insults can promote pul-

monary edema by damaging these mechanisms.

Pulmonary edema is a major clinical manifestation of

ARDS, and the pulmonary edema fluid contains a high con-

centration of protein. This is in marked contrast to pul-

monary edema owing to elevated pulmonary venous

pressure (hydrostatic pulmonary edema) or to decreased

plasma albumin concentration (hypo-oncotic pulmonary

edema), in which the edema fluid is a low-protein transu-

date. ARDS also has been called exudative or noncardiogenic

pulmonary edema, reflecting the increased permeability of

the injured lung to water, solute, and protein. Exudative pul-

monary edema forms in the absence of elevated pulmonary

artery wedge pressure, and the ratio of edema fluid protein to

plasma protein is high. Edema fluid accumulates both in the

pulmonary interstitium and in the alveoli, and because of

potential fluid pathways, lung lymphatics and bronchovascu-

lar spaces (surrounding the bronchioles, bronchi, and pul-

monary arteries) may become engorged. Pulmonary edema

removal by the pulmonary circulation and lymphatics,

including active transport of solute and water, is severely

impaired because of the ALI. In a minority of ARDS patients,

the pulmonary epithelium is able to resolve pulmonary

edema during the first 12 hours. This probably reflects rela-

tively preserved epithelial cells that might increase solute and

water transport in response to β-adrenergic agonists. These

and other drugs are currently being studied.

C. Chronic Lung Injury—Diffuse alveolar damage seen in

ARDS may follow several courses, including resolving

entirely with little or no evidence of chronic damage after

weeks or months. However, other patients develop mild to

severe pulmonary fibrosis. One of the most interesting find-

ings in ARDS is evidence of very early deposition of type III

collagen (procollagen III peptide in alveolar fluid) in the

lung, sometimes within 24 hours of the onset of diffuse



CHAPTER 12

298

alveolar damage. Evidence for early fibrosis has been associ-

ated with poor prognosis, stressing the potentially inappro-

priate role of remodeling and repair in late lung injury. These

findings have challenged the time course of ALI, with irre-

versible fibrosis occurring much sooner than in earlier pro-

posed models. Attraction and activation of fibroblasts may

be mediated by various substances such as platelet-activating

factor that are increased in blood and pulmonary edema

fluid. Oxygen toxicity may play a role, especially if patients

require very high inspired oxygen to treat hypoxemia.

Another potential contributor to chronic lung damage is rec-

ognized to be overdistention of the lungs by high tidal vol-

ume, high PEEP, or high airway pressure.

In chronic lung injury, variable amounts of collagen are

laid down into the alveolar and interstitial spaces with distor-

tion and disruption of the normal lung parenchyma, result-

ing in restrictive lung disease, reduced exercise capacity, and

hypoxemia. Histologic findings may be indistinguishable

from idiopathic pulmonary fibrosis. Particularly sensitive

findings are the Pa

and P(

–a)

during exercise. Recent

data show that survivors with more severe early ARDS had

worse late-stage pulmonary function than those with less

severe acute disease. Lung function as assessed by spirometry,

total lung capacity, and diffusing capacity for carbon monox-

ide averaged about 80% of predicted in these survivors.

The greatest degree of improvement was seen in the first

3 months, and there was little additional improvement

between 6 months and 1 year.

The relationship of collagen deposition, severity of lung

injury, and outcome provides some possibilities for interven-

tion. These include potential inhibition of chemotactic or

activating factors for fibroblasts, removal of some forms of

collagen during tissue repair phases, and limitation of lung

injury by controlling inflammation, oxygen toxicity, and

ventilator-induced lung injury.

D. Physiologic Manifestations

1. Refractory hypoxemia—Hypoxemia in ARDS is due to

right-to-left shunt and

Q mismatching resulting from

atelectasis and filling of alveolar spaces with edema fluid.

Q mismatching also may result from nonuniform

changes in airway resistance, decreases in regional lung

compliance, and primary and secondary alterations of

lung blood flow. Hypoxemia usually is severe and not cor-

rected even when the patient is given high concentrations

of inspired O

,termed refractory hypoxemia (Figure 12–7).

As part of the definition of ARDS, the ratio of Pa

to F

is less than 200.

As shown by CT scanning, the majority of lung in ARDS

is completely airless, with small proportions either normally

inflated or collapsed with potential for recruitment. These

findings suggest that right-to-left shunting plays the major

role in refractory hypoxemia, whereas the use of PEEP walks

a fine line between recruitment of atelectatic lung and

overdistention of normal lung.

2. Altered static lung mechanics—Lung compliance is

severely decreased and airway resistance mildly increased in

ARDS. Decreased lung compliance results from a combina-

tion of interstitial pulmonary edema, collapse of lung units,

airway obstruction, and inactivation of alveolar surfactant.

Ineffective surfactant in ARDS may be due to reduced pool

sizes, alteration in surfactant proteins, altered metabolism,

or inactivation by plasma proteins, oxygen radicals, or

phospholipases exuding into the alveolar spaces. In later

stages, lung compliance is reduced because of accumulation

of collagen.

Studies correlating regional radiographic lung volume

change and inflation pressure show that disease involvement

in ARDS is much less uniform than formerly thought. This

finding has resulted in a major change in understanding of

ARDS. Most lung regions are extensively involved and com-

pletely airless, some participate variably in gas exchange, and

other uninvolved regions accept the bulk of ventilation.

These latter regions have normal specific lung compliance,

indicating that the primary cause of decreased overall lung



Figure 12–7. Lines showing Pa

versus F

for vari-

ous constant-shunt fractions (

) between 0% and

30%. Superimposed are typical responses in patients with

ARDS (lines a and b). Line b shows severe hypoxemia

with a response suggesting severe right-to-left shunt as

the primary mechanism of hypoxemia. Line a demon-

strates severe hypoxemia but with Pa

increasing more

rapidly with increasing F

than with pure right-to-left

shunt. The mechanism of hypoxemia is probably severe

ventilation-perfusion mismatching.



RESPIRATORY FAILURE

299

compliance is overdistention of these small uninvolved

regions of the lung rather than uniform stiffening of the

entire lung. This finding has major implications for how

patients with ARDS should be mechanically ventilated,

notably the use of a low-tidal-volume strategy.

The pressure-volume (PV) curve of the lungs in ARDS is

shifted downward and rightward (Figure 12–8). The lungs

require greater pressure to inflate, and the work of breathing

is increased. An increase in lung compliance may indicate

improvement of the disease or recruitment of atelectatic

lung, especially with the application of PEEP. Some investiga-

tors have noticed that the respiratory system compliance

curve, which considers both lung and chest wall mechanics,

is altered in ARDS. Patients with nonpulmonary causes of

ARDS (eg, abdominal sepsis, trauma, or postoperative) had

greater response to PEEP, suggesting that the “pulmonary

ARDS” patients (mostly those with pneumonia) had more

severe and less recruitable lung consolidation, whereas the

nonpulmonary ARDS patients had more atelectasis.

The shape of the PV curve has been stressed by some cli-

nicians. The curve has been divided into sections, including

a flat initial increase in volume with increasing pressure, a

lower inflection point after which compliance (slope)

increases, an upper inflection point, and then another region

of low compliance (flat slope). The regions of the PV curve

may have implications for adjusting PEEP (see below) and

limiting tidal volume.

3. Increased airway resistance—Increased airway resist-

ance is described in ARDS patients, probably owing to edema

in the bronchovascular spaces surrounding the bronchi, but

there may be inflammatory mediators that induce bron-

choconstriction. Another cause may be the normal increase

in airway resistance in areas of decreased pulmonary perfu-

sion in response to ventilation-perfusion mismatching. The

increased airway resistance, as much as sixfold compared

with normal individuals, contributes to higher airway pres-

sure and work of breathing. In one study, increased resistance

correlated positively with both peak airway pressure and the

severity of gas-exchange abnormality.

E. Multiple-Organ-System Failure—Although often viewed

as a primary lung disorder, ARDS is clearly associated with

multiple-organ-system dysfunction and failure. Subtle evi-

dence of organ-system dysfunction is very common in both

survivors and nonsurvivors of ARDS, and renal failure, liver

failure, CNS failure, heart failure, thrombocytopenia, and GI

bleeding and malabsorption contribute to mortality and mor-

bidity. In fact, of those who die with ARDS, respiratory failure

has been estimated to be the primary cause in as few as 16%.

Sepsis and nonrespiratory failure accounted for the remainder.

The cause of multiple-organ-system failure in ARDS

remains poorly understood. There are three major theories.

First, some investigators believe that the same systemic

process that damages the lungs injures other organs as well;

ARDS is simply the most obvious and earliest manifestation.

For example, one study found that ARDS patients had

increased urinary myoglobin and β

-microglobulin during

development of pulmonary edema, and there was a correla-

tion between pulmonary edema and the amount of these

proteins in the urine. This mechanism of multiple-organ-

system failure may be particularly likely in patients with sep-

sis and shock. Circulating factors such as endogenous

cytokines or endotoxin are likely mediators of multiple-

organ-system failure in this hypothesis.

Another hypothesis is that hypoxemia and inadequate O

delivery are the causes of multiple-organ-system failure.

Some studies of ARDS patients suggest that tissue oxygen

consumption depends on systemic oxygen delivery even

when oxygen delivery is normal or elevated. Thus a small

decrease in oxygen delivery can cause oxygen consumption

to fall, resulting in tissue hypoxia and potential organ dam-

age. Studies undertaken to test the hypothesis that increasing

oxygen delivery will raise oxygen consumption, however,

have not shown improved outcomes.

A final consideration in multiple-organ-system failure is

the effects of therapy of ARDS. Positive-pressure ventilation

and PEEP are important parts of the supportive care of

ARDS, but these can have adverse effects on cardiac output

and oxygen delivery. Invasive monitoring, artificial airways,

and other devices increase risk of infection and sepsis. This

theory is attractive because mortality in ARDS, while still



Figure 12–8. Hypothetical respiratory system pressure-

volume curves for a patient with ARDS showing a flatter

than normal relationship (decreased respiratory system

compliance, C

= V

/ΔP

). With addition of PEEP, a shift to

a more compliant curve may occur such that C

= V

(ΔP

− PEEP) increases. The change in compliance may

represent recruitment of poorly ventilated or nonventi-

lated lung units with application of PEEP and may be

correlated with improved oxygenation and gas exchange.



CHAPTER 12

300

high, is less often related to respiratory failure than to non-

respiratory organ dysfunction and sepsis. Recent clinical

studies provide circumstantial support for this theory. For

example, use of a low-tidal-volume strategy reduces mortal-

ity in ARDS patients by close to 25%. Because there was no

difference in the degree of obvious barotrauma, a beneficial

effect on nonrespiratory organ system function might be one

explanation.

Clinical Features

A. Predisposing Conditions—More than 100 clinical situa-

tions have been associated with development of ARDS.

Prospective studies have helped to identify the most com-

mon conditions, and 60–80% of cases of ARDS can be

accounted for by sepsis, trauma, diffuse pulmonary infec-

tion, and aspiration of gastric contents, with other condi-

tions being much less frequent (Table 12–18).

The overall attack rate of ARDS is relatively low. In one

study, only 8% of patients at risk with any of the nine most

frequently associated predisposing conditions developed

ARDS, although the incidences ranged from 6–50% for indi-

vidual risks. In those with multiple predisposing conditions,

the incidence of ARDS increases substantially, with an aver-

age incidence of 42% for two or more risks for ARDS com-

pared with 19% for those with a single risk in one

study—although 41% of patients with sepsis alone devel-

oped ARDS.

1. Sepsis—Sepsis is present in as many as 50% of ARDS

patients. In medical patients, sepsis is the most common

ARDS association; in trauma patients, sepsis is the most

common association in ARDS developing 48 hours or more

after admission. A distinction sometimes can be made

between pneumonia leading to ARDS and infection from a

nonlung site leading to ARDS. In severe pneumonia, gas-

exchange abnormalities and chest x-ray features of ARDS are

sometimes seen prior to or in the absence of other character-

istics of systemic infection, suggesting that much of the lung

injury is due to the infectious agent and local response to

infection within the lung. Pathogens include bacteria,

mycobacteria, fungi, viruses, Pneumocystis jerovici, and rick-

ettsiae. On the other hand, a nonpulmonary primary site of

infection is often identified in ARDS patients, including the

GI tract, urinary tract, heart, or soft tissues. Although gram-

negative enteric bacilli are encountered most often, ARDS

can result from systemic infection with any bacterial, viral, or

fungal pathogen. Sepsis is of particular interest because it

appears that circulating mediators, including interleukins

and other responses to infection, are largely responsible for

ARDS, hypotension, and multiple-organ-system failure, in

addition to the microbial organisms themselves. Sepsis com-

plicated by ARDS has a higher mortality, especially if the

source of infection cannot be identified and treated readily.

2. Aspiration of gastric contents—Aspiration of gastric

contents is defined as an observed aspiration during intuba-

tion or witnessed vomiting in a patient with impaired airway

protective mechanisms. Although acidic gastric fluid is

thought to be the primary cause of lung injury and ARDS,

studies support a contributory role for bacteria, partially

digested food particles, gastric enzymes, and other noxious

substances. Neutralization of gastric acid prior to aspiration

does not prevent or moderate ALI. Other syndromes of aspi-

ration, including necrotizing pleuropulmonary infection and

lung abscess, can lead to ARDS if there is severe lung injury

response or sepsis. Aspiration of gastric contents is particu-

larly common in patients of advanced age, those who have

neurologic diseases resulting in paralysis or impaired swal-

lowing, or those who have advanced organ-system failure. In

other ICU patients, including surgical and obstetric patients,

aspiration potential is increased by sedatives, muscle relax-

ants, general anesthesia, and local anesthesia to the pharynx

and larynx; during endotracheal intubation; during enteral

feeding; and in diabetics or others with impaired GI motility.

In a sizable number of patients, aspiration of gastric contents

can only be presumed because predisposing factors are

absent and there is no observed aspiration event.

3. Trauma—Trauma is a common predisposing condition to

ARDS, but the precise mechanism is uncertain. Direct

trauma to the thoracic wall may result in lung contusion,

with hemorrhage into the lung causing abnormal gas

exchange, atelectasis, and further lung injury.

Nonthoracic trauma is also associated with an increased

incidence of ARDS. In some of these patients, lung injury

Table 12–18. Some predisposing conditions associated

with ARDS.

Infection

Pneumonia: bacteria, fungi, viruses,

Pneumocystis jerovici

Nonpulmonary:

Sepsis from gram-negative bacilli, staphylococci,

other gram-positive cocci,

Candida

Aspiration of gastric contents

Trauma

Thoracic: lung contusion

Nonthoracic

Hemorrhagic shock

Head trauma

Burns

Blunt abdominal trauma and pancreatitis

Orthopedic: fat embolism syndrome, severe fracture

Other conditions

Drugs: opiate or salicylate overdose

Pancreatitis

Toxic: smoke or gas inhalation

Amniotic fluid embolism

Central nervous system pulmonary edema

Near-drowning

Multiple transfusions of blood and blood products

Collagen-vascular disease, including vasculitis and pulmonary

hemorrhage



RESPIRATORY FAILURE

301

results from fat embolism syndrome, a disorder seen in frac-

tures of long bones, or pancreatitis from blunt or sharp

abdominal trauma. In others, the risk of ARDS correlates

with hypotension and shock and with the amount of blood

and blood products transfused, suggesting that the nature and

extent of trauma may have something to do with the develop-

ment of ARDS. Hypotension and shock release inflammatory

cytokines, and tissue damage liberates a variety of products

that could result in lung injury. Although the number of

blood transfusions administered seems to correlate with the

development of ARDS, the requirement for transfusions is

usually closely tied to the severity of trauma. Any patient,

however, who receives blood products has a risk of lung

injury owing to transfusion-related acute lung injury

(TRALI). While other mechanisms are present, TRALI is pri-

marily thought to be mediated by antibodies in donor plasma

reacting with recipient leukocytes. Blood products from mul-

tiparous donors have an increased risk of causing TRALI.

4. Risk modifiers—Cigarette smoking is associated with

increased likelihood of permeability pulmonary edema and

alveolar hemorrhage. With similar acute risk for ARDS, a

higher proportion of chronic alcoholics will develop ARDS.

Although risk balancing is difficult, elderly patients seem to

be at somewhat higher risk of developing ARDS, and out-

come has been reported to be poorer. Some studies have

shown that ARDS mortality is higher in men.

B. Symptoms and Signs—Patients have severe dyspnea and

respiratory distress. Findings include features of hypoxemia,

such as cyanosis, tachycardia, and tachypnea, but rales and

wheezes are the only features on chest examination, and

these are often surprisingly mild. Although the lungs are

filled with fluid, sputum production is rare, except in those

with bacterial pneumonia. Evidence of the underlying prob-

lem leading to ARDS may be found, including fever,

hypotension, trauma, or findings of organ-system dysfunc-

tion. Features of congestive heart failure are notably absent.

Early in ARDS, symptoms and signs are limited to the lungs.

If multiple-organ-system failure develops, then features of

hepatic, renal, or CNS failure may become evident.

C. Laboratory Findings—A key feature of ARDS is refrac-

tory hypoxemia. Pa

is severely reduced even when the

patient is given supplemental oxygen. Even 100% O

may not

raise Pa

above 60–100 mm Hg. Arterial pH may be high,

normal, or low depending on the success of the patient in

maintaining Pa

with severe lung disease and the presence

of hypotension and metabolic acidosis.

Other laboratory findings reflect the clinical condition

leading to ARDS and the multiple-organ-system dysfunction

seen as a consequence of this disorder. Renal and hepatic fail-

ure and electrolyte disturbances are frequent complications.

D. Imaging Studies—At disease onset, the chest x-ray shows

diffuse bilateral infiltrates consistent with pulmonary edema.

The infiltrates range from patchy reticular to dense consoli-

dation. Unless there is coexisting heart disease, cardiomegaly

is absent, and there is a lack of the central perihilar promi-

nence of edema seen in congestive heart failure. In fact, some

have pointed out that ARDS is associated with predomi-

nantly peripheral distribution of infiltrates. Nevertheless,

distinguishing cardiogenic from noncardiogenic pulmonary

edema is never perfect. In patients with severe pneumonia

leading to ARDS, there may be focal densities in addition to

diffuse pulmonary edema. In later stages of ARDS, the origi-

nal dense opacification may change to a pattern of reticular

densities consistent with the proliferative and fibrotic stages

of lung injury. While chest x-rays usually suggest diffuse

involvement, marked nonhomogeneity of lung involvement

is seen on chest CT scans. Chest x-rays are essential in mon-

itoring patients for complications of ARDS, including baro-

trauma. Early evidence of air leaking into the lung

interstitium sometimes may be found as linear low-density

streaks of air surrounding bronchovascular bundles.

Occasionally, this finding is seen as a rounded lucent area

surrounding a pulmonary artery and bronchus. Air subse-

quently may track inward to the mediastinum (pneumome-

diastinum) or into the pleural space (pneumothorax).

Differential Diagnosis

Cardiogenic pulmonary edema is the most important dis-

order to be distinguished from ARDS (Table 12–19)

because treatment is often different. This may be particu-

larly difficult when ARDS is seen in conjunction with fluid

overload or concomitantly with congestive heart failure.

Septic shock may confound this distinction because circu-

lating endotoxin or cytokines may exert myocardial

depressant activity.

Table 12–19. Distinguishing cardiogenic from noncardio-

genic pulmonary edema.

Cardiogenic Noncardiogenic (ARDS)

Prior history of heart disease Absence of heart disease

Third heart sound No third heart sound

Cardiomegaly Normal-sized heart

Central distribution of infiltrates Peripheral distribution of infiltrates

Widening of vascular pedicle

(increased width of mediastinum

at level of azygos vein)

Normal width of vascular pedicle

Elevated pulmonary artery wedge

pressure

Normal or low pulmonary artery

wedge pressure

Positive fluid balance Negative fluid balance



CHAPTER 12

302

Treatment

Treatment of ARDS centers on management of severe hypox-

emia, correction of the underlying disease that led to ARDS,

and supportive care to prevent complications. Four major

treatment principles have evolved. First, almost all types of

therapy shown to benefit patients with ARDS—including oxy-

gen, PEEP, and positive-pressure ventilation—have potentially

severe adverse effects. Second, although ARDS is often consid-

ered primarily respiratory failure, multiple nonpulmonary

organ-system failure and infection are the major causes of

death in ARDS. Third, careful management of mechanical

ventilation, especially tidal volume, is associated with fewer

complications and is the only treatment demonstrated to

improve survival. Finally, prognosis is especially poor if the

underlying process is not identified or is poorly treated.

A. Oxygen—Treatment of hypoxemia in ARDS is begun

almost always using 100% oxygen (F

= 1.0), and the con-

centration of O

is reduced with the goal of maintaining a

greater than 60 mm Hg (arterial O

saturation about

90%). Pa

increases only slightly with administration of

increasing concentrations of inspired oxygen (refractory

hypoxemia), even when 100% O

is given, indicating right-

to-left shunt or severe

Q mismatching. A very few patients

can be managed using a nonrebreathing oxygen mask, but

most patients will be given oxygen via mechanical ventilation

because they are unable to tolerate the increased work of

breathing without mechanical support. The typical response

to administration of O

in ARDS is shown in Figure 12–7,

and examining the changes in venous admixture as O

and

other therapy are given can be helpful. The F

should be

lowered as soon as possible to less than 0.5 to reduce the risk

of lung damage from oxygen toxicity. In most patients, low-

ering F

is helped by using PEEP or other mechanical ven-

tilation methods to improve lung gas exchange. Because both

PEEP and high F

both have the potential for complica-

tions, a compromise between high F

and high PEEP often

must be chosen.

B. Positive End-Expiratory Pressure—PEEP includes both

positive end-expiratory pressure provided with mechanical

ventilation and continuous positive airway pressure (CPAP)

given to spontaneously breathing patients. During exhala-

tion without PEEP, alveolar pressure is higher than atmos-

pheric pressure, providing a pressure gradient, until

equilibrium is reached. When PEEP is applied, a pneumatic

valve terminates exhalation when the pressure in the system

decreases to a value set by the clinician. This PEEP is termed

extrinsic PEEP. The effective PEEP is the sum of extrinsic and

intrinsic PEEP (PEEPi).

1. Mechanism—The mechanism of action of PEEP is not

known, but PEEP probably works by counteracting the ten-

dency of alveoli to collapse in the face of pulmonary edema,

low lung volume, and loss of surfactant. Current understand-

ing of ARDS suggests that a majority of lung (although highly

variable) is completely atelectatic, whereas smaller propor-

tions are normally aerated or partially collapsed. PEEP

“recruits” some proportion of partially collapsed areas,

improving of gas exchange to these areas of low

Q matching.

PEEP does not decrease the rate of pulmonary edema forma-

tion nor speed the rate of water reabsorption. It is also unlikely

that PEEP is able to open completely collapsed alveoli. Some

investigators have found that PEEP may affect the distribution

of pulmonary artery blood flow away from poorly ventilated

areas and toward better-ventilated regions, resulting in

improved arterial oxygenation. Finally, some studies have indi-

cated that PEEP may have a beneficial effect on the amount or

nature of lung injury.

If partially collapsed alveoli are recruited by PEEP to par-

ticipate in gas exchange, each tidal volume will be delivered

into a larger number of lung units. Lung compliance should

increase as PEEP as added, and the increase in compliance

should parallel improvement in arterial oxygenation (see

Figure 12–8). On the other hand, if PEEP simply distended

alveoli that are already participating in tidal gas exchange, then

lung compliance would remain constant or, if the lung units

become overdistended, lung compliance would decrease. An

upward shift in the position of the pressure-volume curve

indicates recruitment of lung units; movement along the orig-

inal pressure-volume curve suggests that no recruitment

occurred and that gas exchange will not be improved.

2. PEEP and the PV curve—In patients with early acute-

phase ARDS, the pressure-volume curve (Figure 12–9) is flat

at lung volumes near the end-expiratory volume. In this

region, compliance is abnormally low. As airway pressure is

raised—for example, during a tidal volume breath given with

positive-pressure ventilation—many patients will demon-

strate a region of steeper slope or higher compliance. In the-

ory, this region of higher compliance can only be explained by

recruitment of additional lung units. These units must have

been completely collapsed previously, but open when the air-

way pressure exceeds the units’ critical opening pressures. The

changeover point between low-compliance and higher-

compliance regions on the PV curve has been called the lower

inflection point. In theory, PEEP should be set just above the

lower inflection point to indicate recruitment.

Some investigators also have recommended giving suffi-

cient PEEP to prevent the patient breathing in the region cross-

ing the lower inflection point. At lung volumes near the

end-expiratory volume, alveoli are highly prone to collapse. As

the tidal volume is administered, alveoli are repeatedly exposed

to large shear stress as they change from completely closed to

completely open. This may lead to damage in the small airways

and bronchioles leading to the atelectatic lung units. Lung

injury may be moderated if sufficient PEEP is given to raise the

end-expiratory lung volume high enough to prevent cyclic lung

unit collapse (above the lower inflection point).

On the other hand, selecting an optimal PEEP value using

the lower inflection point remains controversial for several

reasons. First, some patients with ARDS have PV curves that



RESPIRATORY FAILURE

303

have no inflection point and no region of increasing compli-

ance. This has been associated with later stages of ARDS dur-

ing which lung units are poorly or not recruited even at

higher lung volume, and improvement of Pa

in response

to PEEP is generally poor. Second, some ARDS patients

whose respiratory PV curves have been partitioned into

lung and chest wall PV curves show no inflection point on

the lung PV curve but only on the chest wall PV curve. This

finding challenges the singular notion of lung recruitment

with PEEP. Next, the lower inflection point in ARDS patients

may be as high as 15–18 cm H

O, or considerably higher

than many clinicians believe is necessary. It is likely that a

more modest level of PEEP in conjunction with a low-tidal-

volume strategy will have the most physiologic and clinical

benefit. Finally, some workers have not found that the linear

segment of the PV curve is associated with constant increase in

compliance, implying that the lower inflection point is not

sharply defined.

3. Adverse effects—Adverse effects of PEEP include

reduced cardiac output that, in turn, decreases systemic oxy-

gen delivery. There have been several postulated mechanisms

of decreased cardiac output from PEEP, including, among

others, (1) decreased systemic venous return, (2) impaired

ventricular performance, (3) increased pulmonary vascular

resistance, and (4) decreased left ventricular compliance. It

is important to distinguish between decreased cardiac out-

put and decreased cardiac function, however, because

positive-pressure ventilation and PEEP can support ventric-

ular function, especially in severe pump failure and cardio-

genic shock. Decreased cardiac output lowers systemic

oxygen delivery, the product of cardiac output and arterial

oxygen content.

PEEP has been associated with barotrauma, including

pneumothorax and lung injury, and with exacerbation of

pulmonary edema, interstitial fibrosis, and inflammation.

Debate continues about whether the underlying lung disease

that led to the use of PEEP contributes to barotrauma or

whether barotrauma is solely related to PEEP. It is clear that

the degree of lung distention rather than the level of PEEP or

airway pressure is the important variable. In animal studies,

high end-inspiratory lung volumes were reached using either

low PEEP and high tidal volume or high PEEP and low tidal

volume. Both groups showed evidence of barotrauma. The

relationship of PEEP to barotrauma is clearly linked to tidal

volume, and the low-tidal-volume strategy will influence the

incidence of lung injury from PEEP.

Patients in whom PEEP is poorly tolerated have hypoten-

sion, tachycardia, and decreased cardiac output as the earli-

est manifestations. Hemodynamic intolerance may be signs

of volume depletion, severe pulmonary hypertension, or

ventricular dysfunction. A pulmonary artery catheter may be

helpful, and volume expansion or vasopressor drugs may be

necessary. In other patients, PEEP may improve Pa

only

slightly, suggesting that they have few recruitable lung units.

PEEP is generally contraindicated in very asymmetric or

localized lung disease with hypoxemia.

PEEP may paradoxically worsen Pa

in ARDS patients

with a patent foramen ovale. In 39 patients with ARDS given

PEEP of 10 cm H

O, mean P(

–a)

improved, and only 7

patients had an increase in shunt fraction, but in 7 patients

with patent foramen ovale, 6 had an increase in shunt frac-

tion and little or no reduction in P(

–a)

. Failure to

improve oxygenation with PEEP was likely due to increased

right-to-left shunting of blood through the foramen ovale,

owing to an increase in pulmonary vascular resistance medi-

ated by PEEP.

4. Application of PEEP—In patients with ARDS, many cli-

nicians prefer to give oxygen at F

1.0 initially and then to

decrease F

as long as adequate Pa

and O

saturation are

maintained. To do this, PEEP is titrated upward starting at 5

cm H

O. One strategy is to use the lowest predetermined

combination of PEEP and F

(Table 12–20), with a clear

target for Pa

and O

saturation. Higher PEEP levels have



Figure 12–9. Hypothetical PV curves shown for normal

individuals and ARDS patients. Regions on the ARDS PV

curve include (1) a region of low compliance at low lung

volume—with a lower inflection point; (2) a region with a

steeper slope showing higher compliance—with an upper

inflection point; and (3) a region with a flatter slope

(poorly compliant). In theory, PEEP might be chosen to be

above the lower inflection point to maximize recruitment

and minimize shear stress on the lungs. Tidal volume

should be adjusted to stay within region (2) to avoid

overdistention. This concept has been challenged because

of evidence that recruitment occurs throughout the PV

curve and is not restricted to the area around the lower

inflection point. See text.



CHAPTER 12

304

not shown benefit, although studies of high versus low PEEP

have shown variable results. As described earlier, to avoid the

lowest lung volume range at which additional stress-induced

lung injury may occur, a minimum PEEP of 5 cm H

O prob-

ably should be given, but the minimum PEEP defined by the

lower inflection point may be substantially higher in some

patients. Recently, it has been suggested that the level of

PEEP could best be set by measuring the percentage of

recruitable lung. This avoids excessive PEEP in “nonre-

cruitable” patients but requires CT scanning of the lungs

while inflated to different pressures.

PEEP should be adjusted incrementally, with close moni-

toring of respiratory mechanics (especially inspiratory

plateau pressure), arterial blood gases, and hemodynamic

variables. An initial PEEP of 5 cm H

O is almost always well

tolerated, but blood pressure, heart rate, and pulse oximetry

should be checked immediately before and after the addition.

If the patient is stable, allow 10–20 minutes before arterial

blood gases and cardiac output are determined. If desired

goals are not met, PEEP is increased using selected combina-

tions of PEEP and oxygen concentration. In some patients,

changes in Pa

may not occur until several hours after PEEP

is changed. These patients cannot be easily identified, but cli-

nicians should be aware that changes in blood gases and

hemodynamics may occur both immediately and long after a

change is made in PEEP.

The goal of PEEP is to facilitate oxygen transfer across the

lungs without impairing systemic oxygen delivery. In most

patients, this is achieved by using the lowest PEEP consistent

with adequate arterial O

saturation (>90%). For some time,

the concept of “best PEEP” has been promoted, variously

described as the PEEP level applied to an individual patient

with ARDS that results in the best balance between tissue

oxygenation and adverse effects. Most now agree that the

optimal PEEP is the least that achieves predetermined objec-

tives of patient management rather than some theoretically

ideal value. Thus PEEP should be titrated until Pa

, arterial

oxygen content, and oxygen delivery increase to acceptable

levels as long as adverse effects are minimized. In practice,

most clinicians prefer PEEP levels between 5 and 12 cm H

and rarely exceed this range because of fear of barotrauma

and decreased cardiac output. Although respiratory system

PV curves can be determined, most clinicians adjust PEEP

using a combination of response of arterial blood gases,

hypothetical maximum and minimum PEEP values, and

hemodynamic response.

C. Mechanical Ventilation

1. Low-tidal-volume strategy—The most important

development in the management of ARDS is that mechan-

ical ventilation with lower tidal volume than previously

used is associated with improved clinical outcome. This

Step 1: Calculate predicted body weight (PBW) in kg. 0.91 x (height, cm – 152.4) + 50 (for men) or 45.5 (for women).

Step 2: Set ventilatory mode (volume-cycled

assist/control) and tidal volume.

a. Initial tidal volume = 6 mL/kg PBW (if already set higher, then lower 1 mL/kg/h).

b. Measure inspiratory plateau pressure (Pplat) with 0.5 s pause every 4 hours and after every

change in PEEP or tidal volume.

c. Adjust tidal volume based on inspiratory plateau pressure. If Plat >30 cm H

O, decrease

tidal volume to 4–5 mL/kg. If Pplat <25 cm H

O and tidal volume <6 mL/kg,

increase tidal volume by 1 mL/kg.

Step 3: Adjust respiratory rate. a. Initial respiratory rate to maintain same minute ventilation.

b. Adjust to keep pH 7.30–7.45.

c. Do not exceed rate >35/min or increase rate if Pa

<25 mm Hg.

Step 4: Adjust F

and PEEP (cm H

O) to maintain

55–80 mm Hg using only these combinations

Minimize both F

and PEEP

PEEP F

PEEP

0.3–0.4 5 0.7 10, 12, 14

0.4 8 0.8 14

0.5 8, 10 0.9 16, 18

0.6 10 1.0 18, 20, 22, 24

Step 5: Manage acidosis or alkalosis as needed. pH <7.30, increase rate (see Step 3)

pH <7.30 and rate = 35, consider bicarbonate administration

pH <7.15, consider increase in tidal volume (even if limited in Step 2)

pH >7.45 and no patient triggering, decrease rate (keep >6/min)

Modified from The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal vol-

umes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301–8.

Table 12–20. Lower tidal volume strategy for ARDS.



RESPIRATORY FAILURE

305

has been termed a low-tidal-volume or lung-protective

strategy. In animals, overdistention of lung regions causes

increased epithelial permeability and a histologic picture

of diffuse alveolar damage. In a multicenter study of

ARDS, mortality decreased from 39% (12 mL/kg of pre-

dicted body weight) to 31% in a group whose tidal volume

was set at 6 mL/kg of predicted body weight. Even smaller

tidal volumes were used if inspiratory plateau pressure

remained greater than 30 cm H

O. A protocol for adjusting

the mechanical ventilation using a low-tidal-volume strat-

egy is summarized in Table 12–20. Reducing tidal volume

is associated with little or no change in the dead space:tidal

volume ratio and may result in improved O

delivery.

Because lung injury is associated with excessive inspiratory

lung volume, higher levels of PEEP may be safer if tidal

volume is limited. Studies have shown little or no adverse

effects of low tidal volume as long as PEEP is given, but

may increase as minute ventilation is decreased with

the lower tidal volume. Low-tidal-volume ventilation with

elevated Pa

has been termed permissive hypercapnia.

Acute respiratory acidosis with permissive hypercapnia

was surprisingly well tolerated in several clinical trials.

However, there may be more subtle effects of respiratory

acidosis on nonpulmonary system function, including

renal and neurologic, that are not yet appreciated.

There is not yet agreement on the mechanism of

improved clinical outcome with a low-tidal-volume strategy.

No differences in pneumothorax (explosive barotrauma) are

reported, but there have been small differences in the rate of

nonpulmonary organ dysfunction. An attractive hypothesis

is that ALI is worsened by lung overdistention, and low tidal

volume moderates this “ventilator-associated lung injury.”

For example, the levels of a variety of proinflammatory

cytokines in blood and alveolar fluid are lower in ARDS

patients treated with low tidal volumes. It is likely that the

severe regional heterogeneity of lung involvement is impor-

tant because overdistention would predominate in the most

compliant portions of the lungs (Figure 12–10).

Interestingly, there are data supporting the hypothesis that

a tidal volume of 6 mL/kg designed to reach a target of a

plateau pressure of less than 30 cm H

O may not be the final

answer. Data suggest that outcome appears to continue to

C. Nonuniform (collapsed

region and normal lung)



Figure 12–10. Schematic illustrating heterogeneity of lung injury in ARDS and effect on respiratory system compli-

ance for the same tidal volume (ΔV) in all three examples.

Normal lungs with uniformly normal compliance. A rep-

resentative PV curve is shown with a small ΔP change as the tidal volume is delivered (ΔV).

Uniformly decreased

compliance of the lungs. The flatter slope of the PV curve results in a larger ΔP required to deliver the same tidal vol-

ume (ΔV).

ARDS is now recognized to have nonuniform lung involvement, and the decreased respiratory system

compliance arises from collapse of the majority of lung units and overdistention of remaining normal units. The same

tidal volume is delivered to a smaller region of normal lung so that ΔP is large for the same ΔV. The overall compli-

ance shown in

(ΔV/ΔP) is the same as in



CHAPTER 12

306

improve as tidal volume or targeted plateau pressure

decreases. This likely would mean that some patients would

benefit from a tidal volume of less than 6 mL/kg, although

not all would require this. Further randomized trials are

needed.

2. Volume-preset ventilation—Most ARDS patients are

ventilated using conventional volume-preset (volume-

cycled) positive-pressure ventilators. The major studies using

low tidal volume for ARDS used this mode. Initial tidal vol-

ume is set at 6 mL/kg of ideal body weight, and peak inspira-

tory flow rate is generally at least 1–1.2 L/s owing to the high

demand for inspiratory flow. Adjustments in tidal volume

depend on the level of the inspiratory plateau pressure (P

plat

as shown in Table 12–20. PEEP is given as necessary for

refractory hypoxemia and in order to lower the F

using a

set of predetermined values for each variable.

Volume-preset ventilation using a high peak flow and a

descending inspiratory flow pattern may have characteristics

similar to those of pressure-controlled ventilation (PCV). In

theory, these settings may improve distribution of ventila-

tion to poorly ventilation lung regions, improving hypox-

emia with smaller increases in PEEP or F

3. Pressure-controlled ventilation—Pressure-controlled

ventilation (PCV) might seem like a very attractive option in

the management of ARDS, largely because a preset maxi-

mum positive airway pressure cannot be exceeded. However,

this feature does not automatically provide the same benefit

as a low-tidal-volume strategy unless the maximum pressure

is adjusted to also limit tidal volume. On the other hand,

PCV does have the theoretical advantage of providing maxi-

mum inspiratory flow at the beginning of the inspired

breath. In some patients, distribution of ventilation may be

enhanced, especially to the most poorly ventilated lung

regions.

A few studies have shown that PCV compared with con-

ventional volume-preset ventilation results in increased

, decreased mean and peak airway pressures, and less

impairment of cardiac output. Other studies have shown no

appreciable differences. Careful adjustment of inspiratory

time is needed to optimize tidal volume and minute ventila-

tion. PCV with pressure limited to 30–40 cm H

O might be

considered in ARDS patients with severe hypoxemia unre-

sponsive to PEEP with conventional volume-preset ventila-

tion or in those who require excessively high airway pressures

or PEEP with conventional ventilation.

Similarly, airway pressure-release ventilation (APRV; see

above) has some attractive features for ventilatory support in

ARDS. Lung recruitment might be improved because of the

higher mean airway pressure, and there may be reduced

stress-relaxation lung injury. No studies, however, have

demonstrated improved outcome in ARDS with APRV com-

pared with more conventional modes. For refractory hypox-

emia, a trial of APRV might be useful and could be used in

comparison to IRV, prone ventilation, and maximum-lung-

recruitment strategies.

4. Inverse-ratio ventilation—In ARDS patients with

refractory hypoxemia, inverse-ratio ventilation (IRV) has

been used. In contrast to conventional ventilation, inspira-

tory time is made longer than expiratory time by decreasing

the inspiratory flow rate, holding inspiration for a preset

time before allowing for exhalation, or, if a time-cycled ven-

tilator is used, directly increasing inspiratory time. How IRV

might improve oxygenation is not clear. Inspired gas may be

distributed more evenly because of the longer inspiratory

time. The shortened expiratory time, on the other hand, may

cause dynamic hyperinflation, raising end-expiratory vol-

ume and improving gas exchange in a manner similar to

intrinsic PEEP.

There have been no controlled clinical trials demonstrat-

ing improved outcome with IRV. This method should be

reserved for the rare patient with refractory hypoxemia unre-

sponsive to PEEP and oxygen therapy. Current understand-

ing of IRV suggests that there is nothing intrinsically

different about a I:E ratio greater than 1:1. Rather, the gas-

exchange effects of I:E ratio vary continually.

When initiated, an I:E ratio of 1:1 should be tried, and

blood pressure, heart rate, and pulse oximetry should be

monitored closely. If needed, I:E ratio can be further altered

to 1.5:1, 2:1, or more. It is unusual for I:E ratios of greater

than 3:1 or 4:1 to improve gas exchange.

Increased time spent during inspiration and dynamic

hyperinflation can cause severely reduced cardiac output and

hypotension. Prolonged inspiratory time may be very

uncomfortable to the patient, and sedation or muscle relax-

ants are always needed. Monitoring of IRV is complex

because I:E ratio, peak airway pressure, and PEEP do not

adequately reflect all the essential parameters and because

the pattern of inspiratory pressure and flow in IRV differs

depending on how IRV is produced. Monitoring mean air-

way pressure has been suggested, but this value does not

correlate with gas exchange or hemodynamic compromise

in IRV.

5. Other modes of mechanical ventilation—Lung-

protective strategies are the basis for several other modes of

mechanical ventilation. Carrying low tidal volume to

extreme is extracorporeal membrane oxygenation and CO

removal, in which the lungs receive no ventilation at all.

Clinical trials have been unable to demonstrate improved

clinical outcomes from this practice. High-frequency oscilla-

tion has been used for more than 20 years. With this method,

very small tidal volumes (1–2 mL/kg or less) at respiratory

rates as high as several hundred per minute are administered

by an oscillating membrane. High-frequency oscillation has

been successful in neonatal respiratory distress, but scaling

up to larger patients has not been uniformly feasible. Gas

exchange often can be maintained or improved, but no dif-

ference in clinical outcome has been demonstrated. Another

lung-protective method is tracheal gas insufflation, in which

a 4–6 L/min flow of fresh gas is provided into a small catheter

placed in the lower trachea. The effect is to reduce the apparent