, ' SPECIAL COMMUNICATION I I. Pulmonary Edema* Physiologic Approaches to Management

Transcription

1 SPECIAL COMMUNICATION Pulmonary Edema* Physiologic Approaches to Management Norman C. Staub, M.D. FLUID AND PRoTEIN ExCHANGE IN THE LUNG The blood vessels of the lung's microcirculation tend to allow a small amount of fluid to leak from the circulating blood into the lung's perimicrovascular (interstitial) tissue. Normally, this fluid is efficiently removed by the pulmonary lymphatic system so that the. fluid does not accumulate within the alveolar walls or flood the air spaces. 1 The filtration of the fluid across the microvascular endothelium is governed by the same principles as is flow lengthwise along the vessels (Fig la), namely: where Q is the net transvascular flow of fluid ( filtration), K is the endothelial conductance of fluid ( filtration coefficient), P is the hydrostatic pressure in the microvascular lumen ( mv) and in the perimicrovascular interstitial fluid ( pmv), u is the reflection coefficient which determines the "effective" difference in transvascular protein osmotic pressure, and., is the protein osmotic pressure in the plasma ( mv) Flow = Conductance X Driving Pressure The special consideration that makes filtration of fluid different from intravascular flow is that the pores in the vascular endothelium, through which the fluid flows, are so small that they markedly restrict the flow of molecules of plasma protein relative to water (Fig lb). This causes a difference in the concentration of protein to develop between the plasma and the perimicrovascular (interstitial) fluid. In 1896, Ernest Starling2 3 recognized the importance of the endothelial restriction to the proteins in the plasma. He clearly saw, for the first time, that the induced difference in protein osmotic pressure was similar in magnitude to but oriented so as to oppose the transvascular hydrostatic pressure. The equation for net filtration of fluid is called Starling's equation in his honor. The equation in its modem form is: Q = K( Pmv- Ppmv) - Kv( nnv -!rpmv) ' I ' I I I, ' From the Cardiovascular Research Institute and Department of Physiology, University of California, San Francisco. Supported in part by Program Project grant HL 6285, Pulmonary Specialized Center of ResearCh grant HL 1421, and Pulmonary Vascular Disease Specialized Center of Research grant HL from the Public Health Service. Based in part on the Louis Mark Memorial Lecture presented at the 43rd Annual Scientific Assembly, American College of Chest Physicians, Las Vegas, Nev, Nov 2, Reprint requests: Dr. Staub CardiOOoacular Rsaearch In- 8titute, 1315 Mol/itt HospitQ/, San Francisco FJcUBE 1. A (top), Filtration of fluid across microvascular endothelium is governed by same principles as flow along microvessels. Model shows typical microvasculjr segment. Conductance of fluid by endothelium is represented by "pores.. in wall. B (bottom), Same microvascular segment as in Figure IA. Plasma proteins are represented as short rods. Small size of pores in endothelium restricts proteins and induces difference in concentrations of proteins 8CI'8S microvascujar endothelium. PULMONARY EDEMA 558

2 and in the perimicrovascular compartment ( pmv). Thus, there are four pressures ( a hydrostatic pair and an osmotic pair), usually referred to as "Starling's forces." One must never forget that the difference in osmotic pressure has the opposite sign from the difference in hydrostatic pressure; that is, osmotic pressure always opposes hydrostatic pressure. The filtration of fluid depends only on the sum of pressures (with proper regard for signs) and the endothelial conductance (the ease with which fluid Rows across the microvascular barrier). The principal Starling's force is the microvascular hydrostatic pressure. This force is responsible for the generation of the difference in osmotic pressure. If there were no difference in hydrostatic pressure, there would be no difference in osmotic pressure. Normally, in the lung (as in all organs), there is a net outward filtration of fluid. In my investigations, the sheep is the experimental animal of choice. In sheep the normal steady-state filtration of pulmonary fluid is 5 to 1 ml/hr, measured as the Bow of pulmonary lymph. The filtration of fluid cannot be measured directly in humans, but it is estimated to be 1 to 2 ml/hr in adults. To summarize, the filtration of fluid across the microvascular endothelium in the lung is a normal process. Therefore, pulmonary edema is an extension of this normal process, not a new phenomenon. PuLMONARY EDEMA By examining Starling's equation, we can see that there are really only two kinds of edema: ( 1) edema due to increased pressure ( cardiogenic); and ( 2) edema due to increased microvascular permeability ( noncardiogenic). Edema due to Increased Pressure Any factor that increases the sum of pressures will increase the outward filtration of fluid. In hypoproteinemia the decrease in the osmotic pressure of plasma protein has been reported to increase the flow of lymph from the lung. Increasing the volume of extracellular fluid. by the administration of large amounts of electrolyte solutions will increase microvascular hydrostatic pressure, as well as decrease the osmotic pressure of plasma protein. Another example may be the unilateral edema sometimes seen after rapid reexpansion of an atelectatic lung or lobe. The cause of edema in this instance would be primarily due to a decrease in the perimicrovascular hydrostatic pressure. Of course, the most common clinical form of edema and the one easiest to study experimentally is that due to increased microvascular pressure, such 58 NORMAN C. STAUB Lymph o. 1.! Plasma tw.raoe MlcrOVCIICular 2 Pressure (em H2) LuiiQ Lymph Flow (ml/hr) lncl..eft Atrial Pre11ure ~~- L ~ ~-==~---- 6/tJIJu//n I I I I I Hours F'IGVBE 2. Time course for effect of increased pulmonary microvascular pressure on flow of pulmonary lymph and on ratio of protein level of lymph over that of plasma in unanesthetized sheep. After two hours of baseline study, left abial pressure was elevated by inflating implanted balloon (modised from Erdmann et al ). as seen in congestive heart failure.' When microvascular hydrostatic pressure rises abruptly, as shown in Figure 2,' the following events take place. The Bow of lymph from the lung rises and eventually reaches a new steady-state level. The interstitial concentration of protein and, consequently, the perimicrovascular osmotic pressure decrease because, with normal integrity of the endothelial barrier, leakage of protein is' markedly restricted relative to the Bow of water. The edema fluid due to increased pressure is characterized by a low concentration of protein relative to that in plasma (Table 1).1.s The decreasing perimicrovascular concentration of protein (as demonstrated at lower left in Fig 3) is part of an important negative-feedback system. AJJ hydrostatic pressure rises, the perimicrovascular osmotic pressure decreases. Thus, the difference in osmotic pressure between the plasma and the interstitial fluid increases. Since the difference in osmotic Tole 1---l.eeel. o/ Prolein In Pluma aad Flaid /rom ~ In Aellle PulmofUII:1' Edema In M- Level of Protein, gm/1 ml Condition Plasma Fluid from Airways Edema due to increased pressure Edema due to increased permeability *Data from Staub1 and Brigham et al.t

3 , ' I o ' ' ~ NORMAL,,_, HIGH PRESSURE b 6 8 INCREASED PERM EABILITY pressure acts as a negative hydrostatic pressure, it opposes the rising microvascular pressure. Experimentally, in sheep, I have found that the negativefeedback mechanism is 5 percent effective over a wide range of hydrostatic pressures. This means that for a given rise in the microvascular pressure, about half of it is nullified due to the increasing difference in osmotic pressure. 4 From this evidence, we see that the lung with normal endothelial integrity has at least the following three important safety factors helping to prevent pulmonary edema:u.7 ( 1) the lymphatic system; ( 2) low endothelial conductance for fluid; and ( 3) very low perimicrovascular hydrostatic pressure. The first factor, the mechanism of lymphatic clearance, is capable of removing filtered fluid at a rate sufficient to keep the content of pulmonary water reasonably normal. In sheep, I have recorded flows of lymph in excess of 5 ml/hr, that is, five to ten times the baseline rate. With respect to the second safety factor, the endothelial barrier itself (because of its relatively low conductivity for fluid), restricts the filtration of fluid under normal driving pressures. As a third safety factor, the endothelial barrier markedly restricts the flow of protein, thereby pro- ducing the difference in protein osmotic pressure, which opposes hydrostatic pressure. There may be a fourth safety factor, namely, a rise in the perimicrovascular hydrostatic pressure. 8 Edema due to Increased Microvascular Permeability When the endothelial barrier is injured, it be- FicUBE 3. Changes at level of microvascular endothelium in two principal types of pulmonary edema. Diagram at lower left shows that in edema due to high pressure, integrity of microvascular barrier is normal, as represented by same size and number of pores; however, increased driving pressure increases filtration of 6uid and washes out protein from perimicrovascular interstitium, thus lowering concentration of protein there. Diagram at lower right shows that in edema due to increased permeability, integrity of endothelial barrier is compromised. This is represented by increase in size and number of pores. For same hydrostatic driving pressure, flow of lymph is greatly increased, and perimicrovascular concentration of protein is high. comes less restrictive to fluid and protein; that is, vascular conductance increases. One can imagine that the microvascular endothelium looks something like that shown schematically in Figure 3 at lower right, where there are not only more but slightly larger pores. The increased permeability to protein is represented by the relatively high concentration of protein in the perimicrovascular fluid. Clinically, edema due to increased permeability tends to form more rapidly than edema due to high pressure, and the fluid has concentrations of protein that are close to those in circulating plasma (Table 1 ). Figure 4 is a theoretic comparison of the filtration of pulmonary fluid as a function of microvascular hydrostatic pressure in the two kinds of edema. One expects that for the same driving pressure, filtration will be greater when the endothelium is injured. Not only that, but since molecules of protein leak through the endothelium more readily, the difference in osmotic pressure across the endothelium (negative-feedback regulation) is reduced. In effect, there is nothing to oppose rising microvascular hydrostatic pressur~. This explains why the slope of the line of increased permeability is so steep. It is clear that two of the major pulmonary safety factors against edema are impaired when the endothelium is injured. An important concept to keep in mind is that experimental evidence now favors only slight changes in the structure of the endothelial intercellular junction as the basis for the altered integrity of the barrier. Large increases in the conductance of fluid and the permeability to proteins occur for very small changes in the size of the endothelial pores. 7 PULMONARY EDEMA 581

4 Fluid Filtration Rote o~--~ Microvosculor Hvdrostotic Pressure FIGURE 4. Theoretic effect of integrity of microvascu1ar endothelial barrier on balance of pulmonary fluid. In presence of increased microvascular permeability, rate of filtration of fluid is increased at any given hydrostatic pressure (as shown by vertical distance between two large points on graph); however, in addition, because of impairment of barrier's restriction to flow of protein, rate of filtration of fluid increases much more rapidly with rising hydrostatic pressure in increased permeability. The old notion that permeability injury was due to wholesale desbuction of endothelium turns out to be incorrect. In fact, if such desbuction were necessary, there would be little we could hope for in terms Concentration Protein 1'1 Albumin ( lymph/plasmaf> Globulin AiF Emboli of specific treatment.5 An important advance in our approach to understanding increased permeability is that we now have several models in animals to study. We are able to produce mild to moderate increases in pulmonary microvascular permeability by the use of gramnegative sepsis (bacteremia due to Pseudomonas), 7 uneven pulmonary arterial obsbuction (multiple pulmonary emboli),8 and intravenous infusion of histamine. 9 Figure 5 shows an example of this type of pulmonary injury. In this type a very mild injury due to air emboli was produced. The important things to note are that at the time of the highest flows of lymph, the concentration of protein in the lymph remained high; that is, leakage of protein from the microvascular bed increased in proportion to the flow of lymph. It is also important to note that even after the air emboli were stopped, the vascular injury persisted for several hours; however, it was reversible, and eventually, full recovery occurred. PHYSIOLOGIC APPROACHES TO TREATMENT Endothelial Injury There is no specific treatment that will reverse the injury to the endothelium in all forms of edema due to increased permeability; however, there are some promising leads, and intense effort is being directed toward the finding of such a magic substance. The basic thesis upon which finding such a substance depends is that loosening of the intercellular june- Pulmonary Vascular Resistance ( em HzO) I/ min Lun9 Lymph Flow (mil h) Hours FIGURE 5. Time course for effect of moderate pulmonary arterial air embolization on pulmonary microvascular permeability. As flow of pulmonary lymph increases, concentration of protein in pulmonary lymph (as measured by ratio of protein level in lymph over that in plasma) remains elevated, indicating increase in flow of protein in parallel with flow of fluid. This is in sharp contrast with increased microvascu1ar pressure alone shown in Figure 2. Even after air embolization has been discontinued, permeability injury lasts for several hours, as indicated by persistent increased flow of lymph; however, injury is reversible within several hours. 562 NORMAN C. STAUB Narmol Increased Permeobllitv FICVBE 6. Model of intercellular junctions as final common pathway for microvascu1ar protein permeability. Analogy to slightly defective zipper allows for reversible change in "size of pores." At right, junction with increased permeability is only slightly different from normal junction at left, yet difference is enough to account for large increases in conductance of fluid and protein.

5 tions between endothelial cells is the final common pathway in all forms of edema due to increased permeability (Fig 6). I emphasize that I am talking about mild to moderate injury and not complete devastation of the alveolar walls, for which I do not anticipate that there will ever be a specific treatment. Some substances will affect endothelial permeability in specific types of edema; for example, any H-1 antihistamine will immediately reverse the increased permeability due to infusion of histamine.9 An exciting new lead is our chance observation that prostaglandin Et, in addition to being a pulmonary vasodilator, partially reverses the increase in permeability due to multiple pulmonary emboli, Prostaglandin Et appears to be acting by a mechanism that is independent of its hemodynamic action. Whether the action of prostaglandin Et is specific for this type of injury or is more generally useful is not known.1 General supportive therapy is important because many endothelial injuries are limited in severity. Regeneration and repair may occur, given sufficient time. In the context of our experimental injuries, the recovery time is of the order of several hours (Fig 5) up to two or three days. 7 Our assessment of the role of pharmacologic doses of corticosteroids (for example, methylprednisolone at 3 mg/kg of body weight) is confined to the supportive role of the therapy, in general, and suppression of inflammation. There is some evidence suggesting that the corticosteroids may aid in preventing increases in permeability. There is little evidence that therapy with corticosteroids is useful in restoring injured endothelium to normal. Driving Pressures Both for edema due to high pressure and edema due to increased permeability, the usual therapy consists of various attempts to alter Starling's forces. There is no known way to specifically alter the perimicrovascular interstitial protein osmotic pressure ( 71'pmv). Theoretically, raising the perimicrovascular hydrostatic pressure ( Ppmv) by means of positivepressure breathing ought to decrease the filtration of fluid, but there has been no firm scientific evidence for this effect. In fact, there are scattered suggestions that in the normal lung, at least, positive-pressure breathing increases the amount of water in the pulmonary tissue. Woolverton et al11 have recently completed a study of the effect of therapy with continous positivepressure breathing at 1 em H2 in sheep. Figure 7 summarizes the results. It was found that both under normal conditions and with increased left atrial Lung Lymph Flow (ml/h) BL PPB INCREASED Left Atrial Pressure FIGURE 7. Steady-state flow of pulmonary lymph under baseline ( BL) conditions and during breathing with 1 em H2 of continuous positive airway pressure in unanesthetized sheep. Regardless of whether left atrial pressure is normal or increased, effects of positive-pressure breathing ( PPB) on steady-state flow of pulmonary lymph are insignificant Reason is that increase in permicrovascular fluid hydrostatic pressure caused by increased pressure breathing is counteracted by almost identical rise in microvascular hydrostatic pressure. pressure, the steady-state How of pulmonary lymph was not significantly affected. There were transient effects at the onset and completion of pressure breathing. The explanation is that the rising alveolar pressure is counterbalanced by a rising microvascular hydrostatic pressure due to the increase in pleural pressure and to direct compression of the microvessels of the alveolar wall by the increased perimicrovascular pressure. As for trying to affect the plasma protein osmotic pressure ( 71'mv), this is of value in edema due to high pressure where the endothelial barrier is normal. It is certainly physiologically reasonable to raise the plasma concentration of protein if it is abnormally low; however, it has not been demonstrated that increasing plasma concentrations of protein above the normal range is useful. The coupling of osmotic therapy with administration of diuretic drugs is necessary in order to remove the reabsorbed water from the circulation and to prevent a rise in general vascular hydrostatic pressure. In edema due to altered permeability, I have already pointed out that the microvascular membrane is more permeable to protein and that the fluid contains high concentrations of protein. Thus, there is no physiologic reason for using macromolecular therapy in the presence of an injured microvascular barrier. (8) PULMONARY EDEMA 563

6 This brings me to the microvascular hydrostatic pressure ( Pmv). The single most effective treatment, both for edema due to increased permeability and edema due to high pressure, is to lower the pulmonary microvascular pressure. Although this is readily understood for edema due to high pressure, the rationale is not so obvious for edema due to increased permeability. I showed in Figure 4 that when the endothelium is injured, even normal vascular hydrostatic pressure in the lung causes an increased rate of escape of fluid. It is not sufficient to say that because pulmonary vascular pressures are in the normal range, no further manipulations are necessary. The steep slope of the line of increased permeability in Figure 4 means that any reduction in vascular pressure will benebt the patient in terms of his pulmonary edema. I am fully aware that reductions in pulmonary vascular pressures below the normal range may reduce cardiac output and systemic delivery of oxygen. Nevertheless, an important goal in the management of pulmonary edema should always be to reduce hydrostatic pressure to its lowest possible value. SUMMARY The integrity of the normal endothelial barrier is responsible for two of the three major safety factors preventing pulmonary edema. This is why edema due to increased pressure is usually not as severe as edema due to increased permeability. Management ought to follow sound physiologic principles. These principles are to lower microvascular hydrostatic pressure and to provide adequate supportive therapy. Positive end-expiratory pressure may improve arterial oxygen transport, but there is no evidence that it improves the balance of pulmonary fluid in edema. Raising the microvascular protein osmotic pressure may be beneficial in edema due to increased pressure but has no demonstrated rationale in edema due to increased permeability. lb:n:jmnces 1 Staub Ne: Pulmonary edema. Physiol Rev 54: , Starling EH: On the absorption of fluids from the cormective tissue spaces. J Physiol 19: , Staub Ne: Lung fluid and solute exchange. In Staub Ne (ed): Lung Water and Solute Exchange. New York, Marcel Dekker, Inc, 1978, pp Erdmann AJ Ill, Vaughan TR Jr, Brigham Kl.. et al: Effect of increased vascular pressure on lung fluid balance in unanesthetized sheep. Circ Res 37: , Staub Ne: Pulmonary edema due to increased microvascular permeability to fluid and proteins. eire Res, to be published 6 Taylor AE, Gibson WH, Granger HJ, tit al: The interaction between intracapillary and tissue forces in the overall regulation of interstitial fluid volume. Lymphology 6:192-28, Brigham KL, Woolverton We, Blake LH, et al: Increased sheep lung vascular permeability caused by Pseudomonas bacteremia. J Clin Invest 54:792-84, Ohlruda IC, Nakahara IC, Weidner WJ, et al: Lung fluid exchange after uneven pulmonary artery obstruction in sheep. eire Res, to be published 9 Brigham ICL, Bowers RE, Owen PJ: Effect of antihistamines on the lung vascular response to histamine in unanesthetized sheep. J elin Invest 58: , Staub Ne, Ohlruda IC: PGE 1 reverses increased lung microvascular permeability during air emboli (abstract). Microvasc Res 15: , Woolverton we, Brigham KL, Staub Ne: Effect of positive pressure breathing on lung lymph flow and water content in sheep. Circ Res 42:55-557, NORMAN C. STAUB CHEST, 74: 5, NOYEMBER, 1978