Pulmonary Edema

.11 IOdl"1 Dr- 01 cardiOlUlmonn mediCine .=

~§.~

- t

Pulmonary Edema* ~Ioglc

App

Mechanisms and New

to Therapy

Gordon R. Bernard, M.Dt, F.C.C.E; and Kenneth L. Brigham, M.D. Since the 6rst report by Ashbaugh and Petty in 1967 categorizing patients with acute respiratory failure and labelling the process the adult respiratory distress syndrome (ARDS~ this illness has heen recognized with increaiing frequency. Adult respiratory distress syndrome now afHicts more than 150,000 people a year and the mortality remains in the 70 percent range in spite of significant advances in critical care medicine. As a consequence, research in the area of acute lung injury has intensi6ed resulting in the elucidation of many of the pathophysiologic mechanisms operative in the process. As

long as the mortality ofARDS remains high, research ofthis kind wiD be needed especiaDy with regard to clinical trials of pharmacologic agents which have appeared promising in the laboratory. An aggressive approach, possibly utilizing multiple drug regimens, seems justi6ed considering the morbidity and mortality with which are dealing. Until success is achieved, supportive intensive care with diligent attention to oxygen and ventilator management, infection control, and 8uid therapy wiD be the mainstay of treatment forARDS.

In simplest terms, pulmonary edema is the accumulation of abnormally large amounts of interstitial and alveolar fluid. Contrary to what was thought only a few years ago, the lung interstitium is not cCdry" even in the normal state. Instead, the dynamics ofHuid Hux are such that there is a constant net transfer of fluid from the microvasculature to the interstitium which returns to the circulation by way of the lymphatics. The lymphatics can readily transport the lymph produced under normal conditions, and in fact, can handle a several fold increase in flow without resultant pulmonary edema. The functional interrelationship of the heart and lungs is such that dysfunction of one can significantly affect the othe!: If the left ventricle fails or the mitral valve becomes stenotic causing pulmonary microvascular pressure to rise, lymph flow will increase and hydrostatic edema may result. This category ofpulmonary edema remains the most common. Conversely, the severely destroyed lungs of a patient with ARDS may provide such resistance to blood flow that cardiac output is impaired. Furthermore, therapeutic interventions which may improve respiratory function such as positive end-expiratory pressure (PEEP) can be detrimental to cardiac function. The effect of heart failure on the lung was the subject of an excellent recent review by Murray. 1

Finally, increased lymph production can result from acute lung injury even when microvascular pressures are normal. This condition is thought to result from greatly increased permeability of the pulmonary microvasculature. This condition is referred to as permeability pulmonary edema, noncardiac pulmonary edema or if severe, the adult respiratory distress syndrome (ARDS). Whether it is due to increased sensitivity to diagnosis, improved survival due to better emergency care such that patients at risk for development of acute lung injury live longer, or some other reason, ARDS has become a common illness. This review will focus on this form ofpulmonary edema especially with regard to its pathophysiology and management.

*From the Pulmonary Circulation Center, De~ent

ofMedicine, Vanderbilt University School of Medicine, Nashville, TN. Supported by NHLBI Grant No. 19153 and The Bernard Werthan, Sr. Fund for Pulmonary Research. lleprynt requut8: Dr. Bernard, Room B-1317 Medical Center North, Vanderbilt Univernty, Nashville 37232

CLINICAL PRESENTATION

While a precise widely accepted definition ofARDS still does not exist (tOr reasons to be discussed), there is general agreement on several points. Derangement in gas exchange must be sufficient to produce dyspnea on room air breathing and the hypoxemia must be relatively resistant to oxygen administration. Any precise value reqtiired for POt or shunt would surely be arbitrary. Most patients with ARDS have increased resistance to lung inflation as measured by total thoracic compliance at the bedside. Chest roentgenograms are a routine part of the surveillance of the critically-ill patient for several reasons but pertinent to this discussion, the chest roentgenogram serves as a time-tested if not quantitative means of lung water estimation. Generally, bilateral diffuse infiltrates which are interstitial or alveolar and consistent with pulmonary edema are reqtiired for diagnosis. I Lung water can be measured more objectively through the use of double indicator dilution technology. The most common method employing this technique uses thermal and green dye signals comparing the pulmonary mean transit time of each indicator in Pulmonary Edema (Bernard, BrIgham)

Table 3-1ime Between Onaet and Development of AlIDS for 83 Patienta*

Table l-AlIDS-Related DiBorden

Direct Lung Injury

Fat embolism syndrome Pulmonary contusion Radiation therapy Oxygen toxicity Pulmonary infection Drugs Paraquat Heroin Salicylates Inhalational injury Smoke Drowning Gastric aspiration

TIme Interval (hr)

Percentage Developing ARDS

0-24

62 20 7 11

24-48

48-72 >72 *From Fowler et al. 3

I ndirect, Mixed or Poorly Understood Mechanisms Sepsis Multisystem trauma Disseminated intravascular coagulation Pancreatitis Idiopathic

order to calculate extravascular lung water. This technique is limited in that it is somewhat invasive and it may underestimate lung water in conditions where pulmonary perfusion is unreliable (eg, pulmonary embolism). There is also some question as to whether the measurement of lung water has significance with regard to ARDS severity or prognosis as will be discussed later in more detail.

Microvascular Permeability Although most would agree that one ofthe fundamental defects in ARDS is increased microvascular permeability, this abnormality is not used as a criterion for the diagnosis of the syndrome because there is no clearly established method for measuring it. Instead, a patient is assumed to have increased microvascular permeability if he has pulmonary edema in the absence of increased hydrostatic pressures, eg, heart failure. The advent of the Swan-Ganz Bowdirected pulmonary artery catheter has greatly enhanced our ability to make accurate estimations of pulmonary microvascular pressures by allowing direct measurement of pulmonary arterial pressure and to estimate left atrial pressure by balloon wedge. Generally, a wedge pressure of less than 18 cm HtO would tend to rule out hydrostatic pulmonary edema. Various research tools do exist which purport the ability to measure microvascular permeability. Multiple tracers are used, most of which are radiolabeled looking for rates of appearance from blood into lung or lung into blood, but despite the number and variety of techniques, there remains no gold standard by which these methods may be compared. Finally, the above described defects should occur in a clinical setting known to be associated with ARDS (Thble 1).

be predicted with varying degrees of accuracy by closely defining and examining certain risk factors. A study by Pepe et all revealed that the major independent risk factors for the development of ARDS in hospitalized patients were sepsis syndrome, aspiration of gastric contents, pulmonary contusion, multiple emergency transfusions, and others (1llble 2). The percentage of patients at risk who actually developed the syndrome ranged from 21 to 36 with even higher rates when more than one risk factor was present in the same patient. The encouraging aspect of studies of this type is that the ability to predict ARDS allows for consideration of early intervention. A lag time exists between the pulmonary insult and full-blown ARDS which varies from a few minutes to several days with precise time of onset being very difficult to pinpoint in certain cases. Work by Fowler and others3 revealed that 82 percent of patients develop ARDS within 48 hours of the insult, while a small number did not develop respiratory failure for more than 72 hours (11lble 3). Having this information allows a more precise approach to patients at risk, especially with regard to need and duration of intensive care precautions, as well as allowing the better design of clinical trials using interventions intended to prevent ARDS from occurring in patients at risk. In animal models of ARDS, Significant hypoxemia occurs before the accumulation ofsignificant pulmonary edema suggesting that the two are not necessarily related. Indeed, when ARDS patients were examined for alveolar-arterial oxygen difference and lung water, there was no correlation between the degree of derangement in the two (Fig 1). It is also clear that derangements in gas exchange occur during sepsis in the absence of chest roentgenographic changes which may occur later or not at all. On the other hand, interstitial Buid is thought to accumulate in the spaces around the distal bronchioles early in pulmonary edema resulting in decreased crosssectional area available for gas exchange. This phenomenon in and of itselfwould probably not result in serious de&ciency in gas exchange if hypoxic vasoconstriction were intact, but animal models of sepsis and ARDS suggest that it is nol Instead, poorly ventilated areas of lung continue to be perfused resulting in Significant arterial desaturation. Gas exchange difficulties become even more pronounced when Buid accumulation becomes severe enough to cause 0.15

Clinical Course Adult respiratory distress syndrome has been associated with a multitude of disorders and Table 1 lists some of the more common associations. Several recent studies have made it clear that ARDS can

Table I-Incidence of AlIDS After No.

Sepsis syndrome Aspiration of gastric contents Pulmonary contusion Multiple emergency transfusions Multiple major fractures

5/13 7/23 5/29 4/17 1/12

*From Pepe et al.

2

0.60 EX'RAVASCULAR LUNG WAtER ._.

--- -- - --

•

•

0.30 0.15

38 30 17 24

8

•

0.45

•

TLC

Specific CUnical ConditionB*

Clinical Condition

r=0.25 p:NS

0

•

• ••

• ••

•

l---...L

150

225

!

300

• • • !

375

450

525

600

A- a DO 2 ( tor r)

FIGURE 1. Extravascular lung water (normalized to predicted total lung capacity) as a function ofalveolar to arterial oxygen difference in patients with ARDS. (From Brigham et al. 4) CHEST I 89 I 4 I APRIL, 1986

595

o

PItIsmtJ

o

o

0

0

0

PROTEIN

OSMOTIC

PRESSURE

mechanism. Mechanical ventilation is frequently required because ofthe tremendous work ofbreathing encountered by these patients. This is due first to markedly decreased lung compliance with inBation pressures often in excess of 50 cmH.O needed to deliver a tidal breath being common, and second, to the very large physiologic deadspace encountered with minute volumes in excess of 30 Umin required to maintain a nonnal Peo. in some cases.

(em HzO)

PATHOPHYSIOLOGY

20

:so

40

'em N.n,

PULMONARY MICROVASCULAR PRESSURE FICURE 2. Steady-state plasma and lymph protein osmotic pressures as a function ofmicrovascular pressure in a sheep during experimental elevation ofleft atrial pressure. There is a progressive decrease in protein osmotic pressure with increasing left atrial pressure. (From Erdmann et al, with permission ofthe American Heart Association. 6) alveolar flooding. 5 This condition can result in right-to-Ieft shunting that is refractory to administration of oxygen. The patient with fully developed ARDS is usually severely dyspneic and at least initially hyperventilating with hypoxemia undoubtedly in part responsible for both. More dimcult to explain is why most patients continue to be dyspneic and hyperventilate even after administration ofoxygen sufficient to correct arterial desaturation. Stimulation of J-receptors in the pulmonary interstitium as a consequence of edema or disturbance of normal length-tension relationships, eg, inappropriately decreased lung expansion for a given amount of negative pleural pressure generation is a possible

Although acute lung injury investigations have been conducted in several in vivo and in vitro models, probably none has been studied more than the sheep lung lymph fistula model. The advantage ofthis model is that through use ofchronic indwelling catheters, the sheep can be studied unanesthetized with the output of the lymph cannula giving continuous information regarding the quality and quantity of pulmonary interstitial fluid formation. The sheep model has also been used to explore the effects of congestive heart failure simulated by inHation ofa balloon positioned in the mitral valve orifice so to elevate left atrial pressure. In this situation, the Starling forces appear to operate as predicted. Erdmann and others, 6 however, have shown that an additional factor must be taken into consideration; that is, that elevated hydrostatic pressure produces a sieving effect in normal vessels such that interstitial Huid protein concentration begins to fall almost immediately under such conditions (Fig 2), lowering interstitial oncotic pressure thereby acting as a protective mechanism tending to reduce lymph flow 75

mean! S.E.M.

50 N-6 DYNAMIC LUNG COMPLIANCE

(cm~20

mean t S.E M.

N:6

50 MEAN PRESSURE ( cmH20)

25

)

25 left atrium

0

0

12 N-6 LUNG RESISTANCE (em H2O) Lisee

9

CARDIAC OUTPUT (°1. baseline)

6 :3

0

ROOM AIR ALVEOLARARTERIAL OXYGEN GRADIENT (torr)

LUNG LYMPH FLOW (m1/15 min)

BASELINE

I

n

PRETERMINAL

PHASES OF THE ENDOTOXIN REACTION

FICURE 3. Mean pulmonary artery and left atrial pressures, cardiac output, lymph flow, and lymph-to-plasma protein concentration ratio in six sheep that died in respiratory failure after endotoxin infusion. (From Esbenshade et aI.')

•

-!

1.00fN'6

0.75

0.50

l~r6 O'lN'6

50

25

~I

-10

PROTEIN 7 CONCENTRATION O. (lymph Iplasma) 0.6

0.5

~ I

BASELINE

I

I

I

U

,

PRETERMINAL

PHASES OF THE ENDOTOXIN REACTION

FIGURE 4. Mean changes in lung mechanics and gas exchange over time in six sheep that died in respiratory failure after endotoxin infusion. (From Esbenshade et ale') Pulmonary Edema (Bernard, Bttgham)

towards normal. There is a limit to how low interstitial protein concentration can go so that this mechanism is not inexhaustible. . Esbenshade et al7 have shown that by giving high doses of purified Escherichia coli endotoxin intravenously to the sheep, a condition very closely resembling human ARDS is obtained (Fig 3 and 4). Consistent with the high lymph Hows produced, the chest x-ray films taken ofthese animals revealed diffuse interstitial and alveolar edema. The current commonly used models employ the use of lower doses of E coli endotoxin which produce a much less severe but qualitatively similar response. Interestingly, intravenous infusion of glass bead microemboli, microthrombi, air emboli, live Pseudomonas bacteria, and other agents can produce remarkably similar responses. These models collectively have proved exceedingly useful in elucidating the pathophysiologic mechanisms of human ARDS. Increased Permeability

Lymph continuously enters the pulmonary interstitium from the microvasculature in small quantities, and this movement of Huid has been described by Starling as being governed by the drop in hydrostatic pressure from the capillary to the interstitium tending to cause Huid transudation as offset by the differences in oncotic pressure between the two tending to draw Huid back into circulation. The permeability of the microvascular barrier must also be taken into consideration, but under normal circumstances, this is assumed to be relatively constant. Therefore, in the usual circumstance, the amount of lymph produced is governed by the balance ofoncotic and hydrostatic pressures. Interstitial and intravascular oncotic pressures, as well as intravascular hydrostatic pressure, have been the subject of many investigations. Until recently, interstitial hydrostatic pressure has received little attention, possibly because it is so difficult to measure. It is quite possible that changes in interstitial Huid pressure in certain illnesses such as ARDS may be enough to have a significant impact on Huid exchange. In addition, it is now known that the total surface area of exchange vessels also determines quantity of lymph production. Surface area may be affected by such factors as cardiac output, exercise, left atrial pressure elevation, and by such humoral mediators as histamine and prostaglandin 12 and E1• Adult respiratory distress syndrome is considered to result from increased microvascular permeability within the confines of Starlings la\\'. Howeve~ the condition is likely to be much more complex. Light and electron microscopy oflung sections from patients with severe respiratory failure often show a severely deranged capillary bed if it is not indeed necrotic.

Applying the Starling equation alone to such a situation would seem to be an over simplification of the complicated situation in ARDS. Mediators of Injury The search for the mediator(s) responsible for acute lung injury (especially septic injury) has been long, and the list of proposed agents continues to gro\\'. Again, it would probably be simplistic to think that any one mediator or one group of mediators could explain all of the aspects of ARDS. Indeed, several humoral mediators have been able to cause an increase in protein rich lymph flow in sheep including serotonin, bradykinin, histamine, proteolytic enzymes, activation products ofcomplement, leukotrienes C. and D., and oxygen-free radicals whether chemically generated or granulocyte derived. Other prostaglandin products of arachidonic acid such as PGH1 and thromboxane are potent pulmonary vasoconstrictors and can cause an increase in pulmonary lymph How (Table 4). The protein content of lymph produced by these and therefOre, is thought to be mediators is lo~ secondary to increases in hydrostatic pressure rather than an increase in permeability. Similary, PG~, PGB2, PGD, PGE 2, and PGF2 do not appear to alter vascular permeability. Thus, products of the cyclooxygenase pathway of arachidonic acid, although hemodynamically potent, do not appear to increase microvascular permeability. Not surprisingly antihistamines, oxygen radical scavengers, corticosteroids, and other blocking agents have been shown to limit lymph How increases in animal experiments. 8 In septic patients, the granulocyte tends to disappear from circulation and lodge in pulmonary capillaries, and upon examination by electron microscopy, are found to be in various stages of activation suggesting that proteolytic enzymes and toxic oxygen species are being released into the microvascular milieu. H this is so (and granulocytes are easily stimulated to produce Table 4-Selective Producta tf ArtJClaidonic Acid and Their Pulmonary SflBtem Effects

Products of Cyclooxygenase Pathway PGH,-Pulmonary hypertension, increased lymph flow (hydrostatic

effect)

hypertension, increased lymph flow (hydrostatic effect) Bronchoconstriction Platelet aggregation ? Leukostasis PGIt (Prostacyclin) -Vasodilation -Platelet dysaggregation -? Bronchodilation Products of Lipoxygenase Pathway Leukotriene B4 and C4-Chemotaxis -Bronchoconstriction -Increased pulmonary vascular permeability Leukotriene D4 - Pulmonary hypertension -Bronchoconstriction

T~-Pulmonary

CHEST I 89 I 4 I APRIL, 1988

587

these compounds in vitro), then the granulocyte as well as related cells must be considered to be major offenders during sepsis, their antimicrobial role notwithstanding. 9 Hypoxemia, the Role of the Ainooys As described above, one of the striking, if not fundamental, aspects of ARDS is the decrease in lung compliance. In the sheep model, there is a clear decrease in dynamic compliance early in the endotoxin reaction accompanied by a sharp increase in airway resistance which appears to be due to reversible bronchospasm since these findings tend to revert toward normal later in the reaction even though lymph flow is still elevated. It is unknown how much a part reversible bronchospasm plays in human sepsis and ARDS, but clearly bronchial edema and bronchial debris could contribute to this abnormality. Static compliance is also clearly abnormal in ARDS patients and this is likely a result of decreased lung distensibility due to inflammation, edema, and scarring. As previously described, these abnormalities combined with loss of hypoxic vasoconstriction are the cause of significant ventilation-perfusion mismatch which results in severe hypoxemia and a marked increase in physiologic deadspace. Nonsteroidal anti-inflammatory agents have been shown to reduce the degree of hypoxemia (Fig 5a) and bronchoconstriction in the sheep ARDS model (Fig 5b).10 Nonsteroidal antiinflammatory agents can also nearly eliminate thromboxane production in sheep (a prostaglandin thought to be an important mediator ofbronchoconstriction) as well as prostacyclin (PGIJ, which is a potent pulmo60

(forr)

30

20

CUNICAL STRATEGY

Prophylaxis With the identification ofrisk factors having a proven high degree of association with ARDS, considerable attention has been given toward approaching the patient at this stage with specific interventional therapeutic maneuvers. It is generally held that if any successful specific therapy for ARDS exists, then it will likely be most effective if given very early in the process and ideally before ARDS is evident. The notion that positive end expiratory pressure benefits gas exchange in ARDS patients has been known for over ten years. As a matter of fact, if there is one therapy for ARDS which might be considered specific, this would have to be it at the present state of our knowledge. Understandably, it was the first prophylactic therapeutic measure to be widely tested in both animals and man. Earlier studies suggested some benefit, but a well-designed recent study by Pepe et alll suggests that prophylactic PEEP for patients at risk for ARDS is of no long-term benefit. 11 Corticosteroid prophylaxis for ARDS remains 0.100

40

l\ AaPo2

nary vasodilator thought to be responsible for inhibition of hypoxic vasoconstriction. Both of these compounds are markedly elevated in sepsis and septic shock. Patients who recover from ARDS frequently maintain an increased bronchomotor reactivity as assessed by methacholine challenge. The mechanism responsible for this phenomenon is not known and neither is it known whether this response is present at the onset of acute lung injury.

-"..,.;::~~

2·

'l'

~

"',

~--...

l..····· ...

.

0.080

..

a..........•

.__ 0 - ' - 9 - -9 ...

.&.

Cdyn (Literlcm H2 0) OD60 0.040

.0

0.020 0 .....0~+~~--~2-~3~ENOOTOXW

TIME (HOURS)

. x

....

·_.-o-·~6_

.,..

~ ~ ~ --0--' __0_.0

. .·.·f······· ·~· · · · 1· ·

\

\

...

•• ~

w

•

/*t

'..,

:

.....• ..:,

9~9~9

/1*

1

. :

1*

2*t

..... 4 -~ O~"'O-"""""--""2--...013~-"""4--"'5

FIGURE Sa (left). Effect of nonsteroidal antiinflammatory drug meclofenamate on endotoxin·induced changes in &P(A-a)OIl in sheep. Endotoxin alone is represented by (0---0), meclofenamate alone is represented by (l:::r---f).), and the combination of the two by (0---0). Note that the sheep pretreated with meclofenamate did not become hypoxemic when given endotoxin, suggesting that cyclooxygenase products of arachidonic acid metabolism are at least in part responsible for the derangements in gas exchange. The asterisk indicates a value that is significantly different from meclofenamate alone and plus indicates a Significant difference from meclofenamate plus endotoxin. (From Snapper et al.lO) b (right). Effect of meclofenamate on the endotoxin induced changes in dynamic compliance. Data are expressed as ± SE. Symbols are the same as in Figure Sa. (From Snapper et al.1O)

588

Pulmonary Edema (Bernard, BrIgham)

unproven and controversial. Early successes in animal studies gave cause for optimism; howevet; these were mainly pretreatment studies in endotoxin models, and human studies thus far have been either uninterpretable or generally negative. This does not mean the issue is dead. Instead, it is quite possibie that certain subpopulations of patients at risk for ARDS may benefit (eg, patients with multiple long bone fractures). Oxygen and Positive End-Expiratory Pressure (PEEP) Supplemental oxygen is required by all patients with ARDS. The extent to which the patient's condition is made worse by oxygen (eg, oxygen toxicity or absorption atelectasis) is not currently known. Although use of PEEP in ARDS has never been proven to improve mortality rates, the technique is almost universally employed because ofthe sometimes dramatic responses achieved in. some patients. 12 PEEP exerts its effect by keeping alveoli open that would otherwise collapse during expiration and which are difficult to re-expand due to the loss of surfactant and structural derangements. With fewer collapsed alveoli, there can be a marked reduction in shunt. Unfortunately, the situation is not so straightforward, and PEEP has now emerged as a double-edged sword in that the improvements in gas exchange may be offset by significant adverse hemodynamic changes. It is clear that an improvement in Pa02 through use of PEEP cannot be taken alone as evidence of benefit. Instead, documentation of improvement in oxygen delivery is essential in guiding PEEP therapy. The upper limits of safe and effective PEEP utilization are still controversial, but levels in excess of 20 to 25 cm H 20 are not commonly used and are less well studied than lower levels. When all best efforts at decreasing shunt and Flo2 to acceptable levels fails, strong consideration must be given to other maneuvers which can improve oxygenation (Table 5). Mechanical Ventilation The immense increase in work of breathing that occurs in ARDS as alluded to earlier is due to lungs stiff from edema, inHammatory reaction, and fibrosis, as well as to resistance to airflow from bronchial edema, debris, and possibly bronchospasm. It is difficult if not

impossible for the patient to maintain this level of ventilatory work for an extended period of time without assisted ventilation. Experimental evidence in laboratory animals suggests that high inHation pressures can cause diffuse microvascular injury, and extrapolation from these data to the human may not be entirely valid. High frequency ventilation appears to be able to achieve adequate minute volumes with much lower mean airway pressures, so this technique may prove useful after further investigation. After some initial enthusiasm, this mode of ventilation does not appear to be clearly superior to standard modes now available, though the final answer is not in. 13 Phannacologic Therapy Corticosteriods, usually in the form of methylprednisolone, remain the most tested group of drugs in human; animal, and in vitro studies. The rewarding experiences with the latter two models have continued to inspire clinical investigation. In spite of the lack of well-designed clinical trials demonstrating efficacy, corticosteroids are widely used clinically in the treatment ofestablished ARDS. Doses used range from 1 to several grams of methylprednisolone over a 24-hour period. Longer treatment periods seem to be associated with significant side effects in both human and animal studies. 14 An ongoing multicenter double-blind randomized trial presently in progress may heip to shed some light on the continuing controversy. Fortunately, there are several other classes of drugs under investigation presently which may have eventual human application. Like steroids, these drugs are generally thought to modulate the inHammatory response. Nonsteroidal anti-inflammatory drugs

40

SHUNT 30 FRACTION (%)

Table 5-Maneuven to Improve Chygeraation Methods of Optimizing Oxygen Delivery Reduce oxygen consumption Reduce fever phannacologically Decrease anxiety Sedation Neuromuscular paralysis if necessary Maintain a favorable oxygen-hemoglobin dissociation curve Prevent hypothermia Increase 2,3 DPG Avoid low Peol and alkalosis (metabolic or respiratory) Maintain a normal hemoglobin level

20

10

0'-----......----......- 21% 02

21% 02

+

MECLOfENAMATE 6. The effect of meclofenamate on shunt fraction in dogs with experimentally produced lobar atalectasis. The mean decrease is statistically significant P
CHEST I 89 I 4 I APRIL, 1988

689

(NSAIDs) such as ibuprofen, meclofenamate, and indomethacin show considerable promise in animal models and in vitro. The NSAIDs seem to work through a variety of mechanisms including modulation of granulocyte activity (both migration and oxidative burst potential), as well as through cyclooxygenase blocking activity which seems to restore vasomotor reactivity both systemically and in the pulmonary circulation. Restoration of vasomotor reactivity in the periphery may prove to be bene6cial in acidosis and hypotension ofseptic shock and in the lung may help to restore hypoxic vasoconstriction and thereby improve gaS exchange. Figure 6 shows beneficial effects of indomethacin on gas exchange and hypoxic vasoconstriction in a canine atalectasis model. 15 Infusion of certain prostaglandins such as PCE 1 and possibly PGE 2 may also prove to be useful as suggested by some recent animal and human studies. These agents have multiple effects, and at present, the exact mechanism of action of these agents is somewhat speculative. Although these new pharmacologic apis not sufficient proaches are quite interesting, th~re human data available to indicate their use except in a research setting. Fluid Mahagement The crystalloid-colloid controversy continues in spite of extensive investigation. Use of the pulmonary artery catheter has simplified management considerably such that at least an accurate estimation of pulmonary microvascular pressures can be obtained. The ultimate goal of fluid therapy in ARDS is to maintain an adequate cardiac output while keeping microvascular pressures as low as possible. At one time, diuretics and fluid restriction were used extensively in order to volume deplete tlte patient. There was also considerable interest in the use of central vasodilators such as nitroglycerin or nitroprusside. These agents are interesting because they allow the rapid reduction ofwedge pressure even in the absence of good renal function. The principle disadvantage is that these drugs ate capable ofpulmonary vasodilation which results in increased shunt and worsened hypoxemia. Long-Term Outcome There are certain physiologic measurements which are easily attainable in ARDS patients that can help to predict the parameters which portend a bad prognosis such as increased pulmonary artery pressure, increased pulmonary vascular resistance, and decreased static lung compliance. Interestingly, extravascular lung water and shunt (or venous admixture) are not predictors of ultimate outcome (Table 6). 16 About one third of patients with ARDS survive. They generally fall into three groups, with some achieving complete resolution, some retaining a mild degree ofpulmonary 800

,Table 6-Early Predictors of AHDS Beveraal in Patientl with E,tabliahed AHDS· C STAT PVR (UcmH 2O) (cmH 2O/L·min) Reversal Nonreversal P-value

.045±.005 .034±.OO2 .034

1.53±.19 2.72±.33 .003

PS u

EVLW

7.2± 1.3 799± 147 12.2±2.3 745± 78 .035 NS

*From Bernard et a1. 16

insUfficiency, and some going on to severe pulmonary 6brosis. It is interesting that a subgroup of patients who develop significant fibrosis as indicated by chest roentgenogram or even lung biopsy are capable of resolving these lesions over several months. There is no well accepted marker fOr those patients who take this course vs those who get worse. REFERENCES

1 Murray JF: The lungs and heart failure. Hosp Pract 1985; 20:55-68 2 Pepe PE, Potlan In: Reus DH, Hudson LD, Carrico CJ. Clinical predictors of the adult respiratory distress syndrome. Am J Surg 1982; 144:124-30 3 Fowler AA, Hamman RF: Good TI: Benson KN, Baird M, Eberle J, et all Adult respiratory distress syndrome: risk with cotnmon predispositions. Ann Intern Med 1983; 98:593-97 4 Brigham KL, Kariman K, Harris TR, et a1. Correlation of oxygenation with vascular permeability-surface area but not with lung water in humans with acute respiratory failure and pulmonary edema. J Clin Invest 1983; 72:339-49 5 Staub NC, Nagano H, Pearce ML. Pulmonary edema in dogs, especially the sequence of fluid accumulation in lungs. J Appl Physiol 1967; 22:227-40 6 Erdmann JA, Vaughn TR, Brigham KL, Woolverton WC, Staub NC. Effects of increased vascular pressure on lung fluid balance in unanesthetized sheep. Cire Res 1975; 37:271-84 7 Esbensahde AM, Newman JH, Lams PM, Jolles H, Brigham KL. Respiratory failure aher endotoxin infusion in sheep: Lung mechanics and lung fluid balance. J Appl Physioll979; 53:967-76 8 Brigham KL. Pulmonary edema. Sem Respir Med 1983; 4:267-328 9 Faritone JC, Ward PA. Role of oxygen-derived free radicals and metabolites in leukocyte-dependent inflammatory reactions. Am , J Patholl982; 107:397-418 10 Snapper JR, Hutchison AA, Ogeltree ML, Brigham KL. Effects ofcyclooxygenase inhibitors on the alterations in lung mechanics caused by endotoxemia in the unanesthetized sheep. J Clio Invest 1983; 72:63-76 11 Pepe PE, Hudson LD, Carrico CJ. Early application of PEEP in patients at risk for the adult respiratory distress syndrome. N Eng} J Med 1984; 311(5):281-86 12 Ashbaugh DO, Petty TL, Bigelow DB, Harris TM. Continuous positive-pressure breathing (CPPB) in adult respiratory distress syndrome. JThorac Cardiovasc Surg 1969; 57:31-41 13 Hudson LD. Ventilatory management of patients with ARDS. Sem Respir Med 1981; 2:128-39 14 Flick MR, Murray JF: High-dose corticosteroid therapy in the adult respiratory distress syndrome. JAMA 1984; 251:1054-56 15 Garrett RC, Thomas HM. Meclofenamate uniformly decreases shunt fraction in dogs with lobar atalectasis. J Appl Physioll983; 54:~9

16 Bernard GR t Rinaldo J, Harris "I: Kariman K, Sibbald W, Bradley R, et aI. Early predictors of ARDS reversal in patients with established ARDS. Am Rev Respir Dis 1985; 131:143 Pulmonary Edema (Bemard, Brigham)