Fluid Management in ARDS: Wet or Dry?
By David J. Pierson, MD
The mortality rate among patients who develop the acute respiratory distress syndrome (ARDS) appears to have declined since about 1990.1 Because there has been no convincing demonstration that any mode or ventilatory approach makes a difference in survival (or in any other outcome measure), one must conclude that this reduction in mortality must be due to improvements in general supportive care. One important aspect of such supportive therapy is fluid administration.
In this area, two competing approaches have emerged, one designed to maximize tissue oxygen delivery, which generally requires increased fluid administration in comparison to traditional management, and the other aimed at decreasing extravascular lung water, which often involves administering diuretics and restricting fluids. These two approaches seem mutually incompatible, and the debate over which one is superior has often been heated, both in the literature and at the bedside. Although the argument cannot be settled unequivocally, some clarification of the issues may help clinicians to sort through the competing claims and to select a rational approach for managing a given patient with ARDS.
Physiological Considerations
The physiological principles underlying the "wet vs. dry" controversy are those governing tissue oxygen delivery and the passage of fluid from capillaries into surrounding tissues. Systemic oxygen delivery (DO2) is the product of cardiac output and arterial oxygen content, the latter being determined primarily by oxyhemoglobin saturation and the quantity of hemoglobin present. Arterial PO2, the measurement most often used in gauging the adequacy of oxygenation, is pertinent mainly with respect to the oxyhemoglobin saturation with which it is associated, since dissolved oxygen contributes little to DO2 in clinical situations.
The net movement of fluid across a capillary wall into the surrounding interstitial space is determined by the capillary filtration coefficient, capillary hydrostatic pressure, interstitial hydrostatic pressure, and the osmotic pressures in the capillary and interstitium, according to Starling’s law of fluid exchange. In ARDS, endothelial injury increases the propensity of fluid to leak out of the capillary into the interstitium, which tends to occur earlier and in greater magnitude than is the case when lung injury is absent; the higher the hydrostatic gradient from capillary lumen to surrounding tissue, the more fluid will leak into the pulmonary interstitium. However, total lung water in ARDS includes both interstitial and alveolar fluid, and accumulation of fluid in the alveolar space is also influenced by factors such as interstitial compliance and lung epithelial function.2
The "Wet" Approach: Maximize Tissue Oxygen Delivery
Patients with severe trauma, septic shock, and other acute life-threatening illnesses demonstrate a spectrum of systemic responses, which correlate with survival and other outcomes. A quarter-century ago, Shoemaker et al observed that patients with shock who had the most vigorous hyperdynamic response (e.g., the highest cardiac output and systemic DO2) were also more likely to survive the illness than those who did not manifest this response.3 A natural extension of this observation was the hypothesis that achieving the "supranormal" hemodynamic values of survivors in all patients with similar illness, particularly early in the course of their illness, would improve survival and other outcomes.4 The management approach based on this hypothesis, which was subsequently used in clinical studies in a variety of patient groups, included attempts to achieve a cardiac index at least 50% above normal, a blood volume 500 mL above normal, a normal or slightly elevated systemic DO2, and an oxygen consumption at least 25% above normal for the patient.4
Although a series of studies from Shoemaker’s group,4 most recently in 1995,5 found that a variety of patients treated by this aggressive, "wet" approach had significantly improved outcomes compared to patients who were not so vigorously managed, methodological concerns and other factors have prevented wide acceptance of these results at face value.2,6-9 Other investigators have only been able to demonstrate improved outcomes using this "wet" approach after the treatment groups were redefined retrospectively.10,11
Several recent studies of an aggressive fluid management approach have been unable to duplicate the positive results of Shoemaker’s group. Hayes et al compared median survivor values for oxygen transport, primarily achieved using dobutamine infusion, with normal hemodynamic values as management goals in critically ill patients who had been optimally fluid resuscitated prior to commencement of the experimental therapy.12 No advantage was found for the more aggressive management approach, and, in fact, there was a trend toward worsened outcomes among the dobutamine-treated patients.
In the largest prospective study to date, Gattinoni et al randomized 762 critically ill patients in 56 intensive care units to one of three management groups, based on attempts to achieve a supranormal cardiac index, a normal mixed-venous oxygen saturation, or a normal cardiac index.13 Patients in the three groups were well-matched in terms of diagnoses, demographics, illness severity, and baseline hemodynamic and gas exchange variables. Only 45% of the patients in the cardiac-index group actually achieved the study’s hemodynamic targets; this compares with 67% of patients in the oxygen-saturation group and 94% in the control (normal cardiac index) group. No differences in mortality were observed among the treatment groups, and this finding held when individual diagnostic categories were examined separately. Outcomes were not significantly different among patients in the three treatment groups who actually achieved the management goals.
None of these studies has definitively settled the issue, and the relatively small proportion of patients in the supranormal cardiac index group of Gattinoni et al who achieved the hemodynamic management goals illustrates the difficulties encountered in multiple centers, among a diverse group of patients, in actually carrying out the "wet" management approach as it has been intended. Because it uses hemodynamic parameters requiring a pulmonary artery catheter and typically involves administration of additional crystalloid, inotropes, and other therapies, the "wet" approach may require substantially more health care resources (both material and personnel) than less aggressive management schemes.
The "Dry" Approach: Minimize Extravascular Lung Water
Lung edema develops at lower hydrostatic pressure in the face of increased permeability, as is present in ARDS, and a low plasma oncotic pressure, also present in many critically ill patients, worsens the problem. It would thus seem logical to avoid excessive fluid administration in the management of such patients. The "keep the lungs dry" approach to fluid management in patients with ARDS is based on the notion that deliberately minimizing capillary hydrostatic pressure will reduce lung edema and potentially lessen the severity and duration of clinical illness. As with the "wet" approach, a strong rationale and several positive but potentially flawed clinical studies favor the "dry" strategy.2,9,14
With the "dry" approach, a diuretic such as furosemide is administered in an attempt to reduce hydrostatic pressure in the lungs. This has been shown to reduce lung edema in a number of animal models of acute lung injury,2 although one study found that the beneficial effect seen early did not occur if diuresis was not begun until 24 hours after lung injury.15 Available clinical data tend to support the concept that accumulation of extravascular lung water is less, and resolution occurs earlier, when clinical strategies are used that succeed in reducing pulmonary capillary wedge pressure or producing a net diuresis in patients with pulmonary edema.
The "Dry" Approach: Effects of Diuresis and Fluid Restriction
In a randomized controlled trial, Mitchell et al studied the effect of diuresis and fluid restriction on extravascular lung water (EVLW) and total ventilator and ICU times in the management of 101 patients with various forms of pulmonary edema.16 Management in 52 patients was by a "dry" approach designed to minimize EVLW using an algorithmic protocol, and during the period of observation these patients had no net fluid gain. The 49 comparison patients, managed in a more standard fashion guided by pulmonary capillary wedge pressure, had a median net fluid gain of more than two liters. In addition to having less EVLW, the "dry" group spent fewer days on the ventilator (median, 9 vs 22 days) and in the ICU (median, 7 vs 16 days) than the wedge pressure group, both of these differences being statistically significant.
Mortality in the study of Mitchell et al was not different in the two study groups.16 However, the same authors separately reported a retrospective analysis of a subset of the same patients, showing that those who had gained at least 1L of fluid during the first 36 hours of therapy had a significantly higher mortality (74% vs 50%, P = 0.05), in addition to longer times on the ventilator, in the ICU, and in the hospital, as compared with patients with negative net fluid balance or a gain of less than 1L.17
There are problems with this study, one of which relates to the patients who were included. More than one-third of them had a diagnosis of congestive heart failure rather than ARDS or sepsis, and one would expect a negative fluid balance to be easier to achieve in such patients than in those with ARDS. In addition, the "dry" management approach used by the investigators was based on serial measurements of EVLW, using a thermal-indocyanine green dye double-indicator dilution method which requires a pulmonary artery catheter. Whether use of this technique would yield results of acceptable accuracy for clinical use outside the research setting is unclear.
Despite the positive results obtained in these and other studies using a "dry" approach to fluid management, such findings may not signify cause-and-effect relationships between lower wedge pressures and avoidance of fluid gain during ventilator management and mortality and other outcomes in ARDS, but rather the fact that survivors may be endowed with better ability to have a lower wedge pressure and to avoid fluid gain.18 Such a conclusion is supported by a study that found that patients with increased edema clearance in hypoxemic acute respiratory failure had better survival than patients with prolonged pulmonary capillary fluid leak.19
Thus, the rationale and enthusiasm for the "keep-the-lungs-dry" approach may be analogous to those for the "maximize oxygen delivery" approach. If patients who are more likely to survive have less lung edema and can better tolerate reductions in wedge pressure and overall fluid balance than patients who are more likely to die, it seems logical to use these phenomena as goals for managing all patients. However, as with the "wet" approach to fluid management, whether such goals can actually be achieved in clinical practice and make a clinical difference in patients who would not spontaneously achieve them remains to be convincingly demonstrated.
Despite the acknowledged deficiencies in our current database on the effects of fluid management in ARDS on patient outcomes, several respected authorities lean toward the "dry" approach.2,7,18 Sznajder and Wood advocate use of a clinical strategy that attempts to maintain the lowest possible pulmonary artery wedge pressure that permits adequate oxygen delivery and organ function.18 In addition to requiring a pulmonary artery catheter, this approach includes the use of erythrocyte transfusion, dopamine, and dobutamine to support or augment systemic oxygen delivery without having to raise wedge pressure or administer crystalloid.
Adverse Effects of "Keeping The Lungs Dry"
Keeping critically ill patients "dry" is not without potential hazards. By raising mean intrathoracic pressure, positive-pressure ventilation may impede venous return to the thorax, thus reducing right ventricular preload, an effect that can reduce cardiac output and cause hypotension. In the presence of obstructive lung disease, this effect may be compounded by dynamic hyperinflation and auto-PEEP, and may lead to pulseless electrical activity and cardiac arrest. Even in the absence of obstructive lung disease, PEEP, the primary ventilatory strategy for augmenting arterial oxygenation in patients with ARDS, can cause clinically important reductions in cardiac output in the presence of relative hypovolemia. Although a low level of PEEP (e.g., 5 cm H2O) is routinely used in intubated patients in many centers, even this so-called "physiologic PEEP" can have adverse hemodynamic effects when the adequacy of cardiac filling is borderline.
Dantzker et al reported the effects of incremental additions of PEEP in 12 patients with ARDS.20 In five patients with initial wedge pressures of 4-11 mmHg who were not already on PEEP, addition of 5 cm H2O of PEEP caused a drop in cardiac output in four patients (by as much as 42%) and a fall in mixed venous PO2 in three patients; mean cardiac output went from 7.2 to 6.1 L/min on addition of PEEP in these five patients.
Recommendations
What should the clinician do in the face of these divergent approaches and the somewhat contradictory available evidence? It is important to consider the practice settings and patient populations in which the reported experience has been gained. In general, success with the "wet" approach has mainly been in postoperative and surgical sepsis patients, and success with the "dry" approach has mainly been with medical patientsthat is, patients who may require less initial fluid resuscitationand the main proponents of the two approaches have generally been surgical intensivists vs. medical intensivists. Success with both approaches has mainly been with their early application, and comparable results might not be achievable with either strategy if begun later in a patient’s clinical course.
It does seem to be true that avoiding excessive fluid administration in any clinical setting is a good idea, although the best specific management target (e.g., cardiac output, DO2, wedge pressure, overall fluid balance) in a given clinical circumstance remains unclear. It seems to me that available evidence currently favors the "dry" approach when this is clinically feasible, especially in medical patients. The goal of minimizing extravascular lung water is just as desirable in patients with trauma, burns, and surgical sepsis but may be less readily achievable because of the fluid demands of these clinical settings. However, even in the latter circumstances it would seem like a good idea to maintain the lowest filling pressures that are consistent with adequate perfusion, and efforts should be made to reduce fluid intake and avoid unnecessary fluid overload whenever possible.
Conclusion
Whichever approach is selected, the clinician’s fluid management strategy needs to be consistent with the overall management philosophy being followed in terms of ventilator management and patient monitoring. Minimizing fluid administration in patients with ARDS, such that relative hypovolemia is present, is not compatible with the use of high levels of PEEP, and neither approach, if carried out as its proponents describe, can be adhered to without a pulmonary artery catheter. Thus, the clinician may well elect to pursue a strategy somewhere between the "wet" and "dry" extremes, perhaps requiring less invasive monitoring in some patients. I think such an approach is justifiable, in view of the present lack of outcome data from properly designed studies, of sufficient size and in appropriate patient populations, to settle the current controversy.
References
1. Milberg JA, et al. Improved survival of patients with acute respiratory distress syndrome (ARDS): 1983-1993. JAMA 1995;273:306-309.
2. Schuller D, Schuster DP. Fluid management in acute respiratory distress syndrome. Curr Opin Crit Care 1996;2:1-7.
3. Shoemaker WC, et al. Physiologic patterns in surviving and nonsurviving shock patients: Use of sequential cardiorespiratory variables in defining criteria for therapeutic goals and early warning of death. Arch Surg 1973;106:630-636.
4. Shoemaker WC. Controversies in the pathophysiology and fluid management of postoperative adult respiratory distress syndrome. Surg Clin North Am 1985; 65:931-963.
5. Bishop MH, et al. Prospective, randomized trial of survivor values of cardiac index, oxygen delivery, and oxygen consumption as resuscitation endpoints in severe trauma. J Trauma 1995;38:780-787.
6. Hinds C, et al. Manipulating hemodynamics and oxygen transport in critically ill patients. N Engl J Med 1995;333:1074-1075.
7. Hudson LD. Fluid management strategy in acute lung injury. Am Rev Respir Dis 1992;145:988-989.
8. Hudson LD. New therapies for ARDS. Chest 1995; 108:79s-91s.
9. Schuster DP. The case for and against fluid restriction and occlusion pressure reduction in adult respiratory distress syndrome. New Horiz 1993;1:478-488.
10. Tuchschmidt J, et al. Elevation of cardiac output and oxygen delivery improves outcome in septic shock. Chest 1992;102:216-220.
11. Yu M, et al. Effect of maximizing oxygen delivery on morbidity and mortality rates in critically ill patients: A prospective, randomized, controlled study. Crit Care Med 1993;21:830-838.
12. Hayes MA, et al. Elevation of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med 1994;330:1717-1722.
13. Gattinoni L, et al. A trial of goal-oriented hemodynamic therapy in critically ill patients. N Engl J Med 1995; 333:1025-1032.
14. Schuster DP. Fluid management in ARDS: "Keep them dry" or does it matter? Intensive Care Med 1995; 21:101-103.
15. Rusch VW, et al. Effect of furosemide on fully established low pressure pulmonary edema. J Surg Res 1986;41:141-145.
16. Mitchell JP, et al. Improved outcome based on fluid management in critically ill patients requiring pulmonary artery catheterization. Am Rev Respir Dis 1992;145:990-998.
17. Schuller D, et al. Fluid balance during pulmonary edema. Is fluid gain a marker or a cause of poor outcome? Chest 1991;100:1068-1075.
18. Sznajder JI, et al. Beneficial effects of reducing pulmonary edema in patients with acute hypoxemic respiratory failure. Chest 1991;100:890-892.
19. Matthay MA, et al. Intact epithelial barrier function is critical for the resolution of alveolar edema in humans. Am Rev Respir Dis 1990;142:1250-1257.
20. Dantzker DR, et al. Depression of cardiac output is a mechanism of shunt reduction in the therapy of acute respiratory failure. Chest 1980;77:636-642.