Fluid Management Strategy in Acute Lung Injury
In this issue of the REVIEW, Mitchell and colleagues (1)publish data addressing an important question in critical care management - What is the appropriate fluid management for the critically ill patient, particularly the patient with adult respiratory distress syndrome (ARDS) and/or sepsis syndrome (I)? The study was rigorously carried out with a remarkable degree of cooperation for a complicated intervention study in the ICU and is well presented. It is our responsibility as readers and practitioners to decide how to interpret their data and whether (and how) to change critical care practice based on this information. These sorts of studies are difficult to design and carry out. One ofthe major limitations of any such study is that only one major hypothesis or intervention can be tested in a critical fashion when the investigators would like to be able to test several. Thus, difficult choices must be made in designing the study. In this case, the authors tested two strategies of fluid management: a management algorithm primarily based on a measurement of extra-vascular lung water (EVLW) was compared with a management strategy primarily based on a measurement of wedge pressure (WP). The patient population studied was a heterogeneous group comprised of patients who had a pulmonary artery catheter (PAC)placed for management; they were further stratified by chest radiograph and WP. The great majority of patients had pulmonary edema. Thirtyeight patients had a clinical diagnosis of congestive heart failure (CHF) and 52 had a diagnosis of ARDS with other patients having sepsis syndrome without clinical ARDS. In another recent publication by these authors, a retrospective analysis of data from the same patients reported in the current study was made (2). When analyzed by survival, the survivors had a significant fluid gain or change in EVLW but decreased WP and body weight compared with non-survivors. In that analysis, a larger fraction of patients in the surviving group also had a clinical diagnosis of CHF. It is clearly easier to achieve a negative fluid balance in patients with CHF, and survival would be expected to be much greater in CHF than in ARDS patients. Thus, the results AM REV RESPIR DIS 1992; 145:988-989
of this analysis of survival could be explained by lumping CHF with ARDS patients and showing expected associations with a given diagnostic group (negative fluid balance and good survival in patients with CHF). The current study is a more rigorous test of the management strategy and the results are not easily explained by lumping these patient categories, even though this obscures the critical question about which one would like an answer. Whether achieving negative fluid balance or using fluid restriction and diuresis benefits outcome is not a question in CHF patients but is clearly an important unanswered question in patients with ARDS. Although this study can be criticized for combining patients with CHF and ARDS, trends in the outcome variables were similar in both groups of patients, making this criticism less relevant. One of the questions raised by the study by Mitchell and coworkers is whether improved mortality is a necessary outcome to claim success from a management strategy involving critically ill patients with high mortality. It is my belief that an improvement in mortality, especially in a single center study with an understandably limited number of study subjects, is not a necessary outcome variable. It is appropriate to use physiologicendpoints or other important practical endpoints such as lengths of ICU stay or mechanical ventilator days, although these latter endpoints should be made as objective as possible with similar criteria for discontinuing mechanical ventilation in all patients. Having said that, I would be surprised if a marked difference in the length of mechanical ventilation and stay in the ICU was not associated with some sort of trend in mortality, given the assumption that the patient with ARDS on mechanical ventilation and in the ICU environment is at higher risk for infectious complications and development of sepsissyndrome than patients transferred out of the ICU. In fact, in this study, the mortality trend was in the expected direction (lower in the patients with lower ICU stays). Does one need to measure EVLW in order to achieve the goal of minimizing
o
fluid administration in the patient with increased permeability pulmonary edema while maintaining "adequate" cardiac output and tissue oxygen delivery? If EVLW is measured, how is it to be interpreted and integrated with other physiologic data? These are complicated questions. In general, I tend to avoid complicating patient care with additional measurements unless I am convinced that they help in making beneficial patient care management decisions. I, like many critical care practitioners, have not made the measurement of EVLW using a thermodilution catheter technique a routine measurement in our critical care unit. However, in both reviewingthe paper of Mitchell and colleagues plus other publications in the process of critically reviewing their paper, I am swayed toward the idea that the measurement in EVLW in selected patients, particularly those with severe sepsis syndrome and ARDS, might provide useful information in addition to hemodynamic measurements. Sibbald and his group (3-5) have added to our understanding of how to interpret this measurement in relation to other data and particularly in the differentiation of non-cardiac from cardiac pulmonary edema. Their original work and a review placing the measurement of EVLW in clinical perspective are recommended for a greater appreciation of the clinical application of this technique. In order to assure having the same results as Mitchell and coworkers, one would need to use the same method in a similar patient population and thus EVLW would be routinely measured. It is at least feasible that a similar decrease in lung water and any attendant beneficial outcomes could be accomplished without actually measuring EVLW. Whether monitoring and management strategies that did not involve this measurement would yield similar outcomes would seem to be a worthwhile topic for a clinical study, although this would probably require a multicenter study in order to have enough statistical power to draw firm conclusions. It seems intuitively most likely that the outcome differences between the two groups in the study by Mitchell and colleagues are related to a lower positive flu988
EDITORIAL
id balance in the EVLW group and that measurement of EVLW may be helpful in achieving this result, but not absolutely necessary. I am concerned that this might be misinterpreted when attempting to translate this into practice - i.e., that fluid restriction as a primary strategy might reach the same goal. It is important to note that the strategy used in the EVLW group (and in fact in the WP group) was not simply fluid restriction but fluid restriction based on careful and frequent measurements and monitoring plus use of vasopressors, vasodilators, and inotropic agents, also all based on careful monitoring. This caveat is emphasized by the authors. Thus, the strategy in the EVLW group associated with successful outcome was based on the goal of minimizing fluid administration while maintaining cardiac output and tissue oxygen delivery. Any strategy that simply restricted fluid or employed diuresis without concern for other physiologic variables cannot be expected to have the same outcome and is potentially dangerous. If reduction in lung water is a therapeutic
989
goal, it is important that this be accomplished without adverselyaffecting cardiac output and tissue perfusion. In fact, the authors speculate that their data suggest that the true importance of hemodynamic monitoring is to avoid therapeutic complications. I draw two conclusions from this study. The first conclusion is a general onethat making careful, accurate observations that provide knowledge about multiple physiologic variables and their interactions in the critically ill patient combined with a logical and consistent management approach allowing use of this information in an integrated way is likely to yield better outcomes. The second conclusion is that this study seems to confirm the logical strategy (or hypothesis) that achievingthe lowestEVLW that maintains tissue perfusion will improve patient outcome. Whether the fluid management strategy employed in the EVLW group by Mitchell and coworkers is the most practical and efficient way to achieve this goal remains an open question.
LEONARD
D.
HUDSON
M.D.
Professor of Medicine Head, Division of Pulmonary and Critical Care Medicine University of Washington Seattle, WA References 1. Mitchell JP. Improved outcome based on fluid management in critically-ill patients requiring pulmonary artery catheterization. Am Rev Respir Dis 1992; 145:990-8. 2. Schuller 0, Mitchell JP, Calandrino FS, Schuster DP. Fluid balance during pulmonary edema: Is fluid gain a marker or a cause of poor outcome? Chest 1991; 100:1068-75. 3. Sibbald WJ, Cunningham DR, Chin ON. Noncardiac or cardiac pulmonary edema? A practical approach to clinical differentiation in critically ill patients. Chest 1983; 84:452-61. 4. Sibbald WJ, Short AK, Warshawski FJ, Cunningham DO, Cheung H. Thermal dye measurements of extravascular lung water in critically ill patients: Intravascular starling forces and extravascular lung water in the adult respiratory distress syndrome. Chest 1985; 87:585-92. 5. Sibbald WJ, Warshawski FJ, Short AK, Harris J, Lefcoe MS, Holliday RL. Clinical studies of measuring extravascular lung water by the thermal dye technique in critically ill patients. Chest 1983; 83:725-31.
In this issue of the REVIEW, Mitchell and colleagues (1)publish data addressing an important question in critical care management - What is the appropriate fluid management for the critically ill patient, particularly the patient with adult respiratory distress syndrome (ARDS) and/or sepsis syndrome (I)? The study was rigorously carried out with a remarkable degree of cooperation for a complicated intervention study in the ICU and is well presented. It is our responsibility as readers and practitioners to decide how to interpret their data and whether (and how) to change critical care practice based on this information. These sorts of studies are difficult to design and carry out. One ofthe major limitations of any such study is that only one major hypothesis or intervention can be tested in a critical fashion when the investigators would like to be able to test several. Thus, difficult choices must be made in designing the study. In this case, the authors tested two strategies of fluid management: a management algorithm primarily based on a measurement of extra-vascular lung water (EVLW) was compared with a management strategy primarily based on a measurement of wedge pressure (WP). The patient population studied was a heterogeneous group comprised of patients who had a pulmonary artery catheter (PAC)placed for management; they were further stratified by chest radiograph and WP. The great majority of patients had pulmonary edema. Thirtyeight patients had a clinical diagnosis of congestive heart failure (CHF) and 52 had a diagnosis of ARDS with other patients having sepsis syndrome without clinical ARDS. In another recent publication by these authors, a retrospective analysis of data from the same patients reported in the current study was made (2). When analyzed by survival, the survivors had a significant fluid gain or change in EVLW but decreased WP and body weight compared with non-survivors. In that analysis, a larger fraction of patients in the surviving group also had a clinical diagnosis of CHF. It is clearly easier to achieve a negative fluid balance in patients with CHF, and survival would be expected to be much greater in CHF than in ARDS patients. Thus, the results AM REV RESPIR DIS 1992; 145:988-989
of this analysis of survival could be explained by lumping CHF with ARDS patients and showing expected associations with a given diagnostic group (negative fluid balance and good survival in patients with CHF). The current study is a more rigorous test of the management strategy and the results are not easily explained by lumping these patient categories, even though this obscures the critical question about which one would like an answer. Whether achieving negative fluid balance or using fluid restriction and diuresis benefits outcome is not a question in CHF patients but is clearly an important unanswered question in patients with ARDS. Although this study can be criticized for combining patients with CHF and ARDS, trends in the outcome variables were similar in both groups of patients, making this criticism less relevant. One of the questions raised by the study by Mitchell and coworkers is whether improved mortality is a necessary outcome to claim success from a management strategy involving critically ill patients with high mortality. It is my belief that an improvement in mortality, especially in a single center study with an understandably limited number of study subjects, is not a necessary outcome variable. It is appropriate to use physiologicendpoints or other important practical endpoints such as lengths of ICU stay or mechanical ventilator days, although these latter endpoints should be made as objective as possible with similar criteria for discontinuing mechanical ventilation in all patients. Having said that, I would be surprised if a marked difference in the length of mechanical ventilation and stay in the ICU was not associated with some sort of trend in mortality, given the assumption that the patient with ARDS on mechanical ventilation and in the ICU environment is at higher risk for infectious complications and development of sepsissyndrome than patients transferred out of the ICU. In fact, in this study, the mortality trend was in the expected direction (lower in the patients with lower ICU stays). Does one need to measure EVLW in order to achieve the goal of minimizing
o
fluid administration in the patient with increased permeability pulmonary edema while maintaining "adequate" cardiac output and tissue oxygen delivery? If EVLW is measured, how is it to be interpreted and integrated with other physiologic data? These are complicated questions. In general, I tend to avoid complicating patient care with additional measurements unless I am convinced that they help in making beneficial patient care management decisions. I, like many critical care practitioners, have not made the measurement of EVLW using a thermodilution catheter technique a routine measurement in our critical care unit. However, in both reviewingthe paper of Mitchell and colleagues plus other publications in the process of critically reviewing their paper, I am swayed toward the idea that the measurement in EVLW in selected patients, particularly those with severe sepsis syndrome and ARDS, might provide useful information in addition to hemodynamic measurements. Sibbald and his group (3-5) have added to our understanding of how to interpret this measurement in relation to other data and particularly in the differentiation of non-cardiac from cardiac pulmonary edema. Their original work and a review placing the measurement of EVLW in clinical perspective are recommended for a greater appreciation of the clinical application of this technique. In order to assure having the same results as Mitchell and coworkers, one would need to use the same method in a similar patient population and thus EVLW would be routinely measured. It is at least feasible that a similar decrease in lung water and any attendant beneficial outcomes could be accomplished without actually measuring EVLW. Whether monitoring and management strategies that did not involve this measurement would yield similar outcomes would seem to be a worthwhile topic for a clinical study, although this would probably require a multicenter study in order to have enough statistical power to draw firm conclusions. It seems intuitively most likely that the outcome differences between the two groups in the study by Mitchell and colleagues are related to a lower positive flu988
EDITORIAL
id balance in the EVLW group and that measurement of EVLW may be helpful in achieving this result, but not absolutely necessary. I am concerned that this might be misinterpreted when attempting to translate this into practice - i.e., that fluid restriction as a primary strategy might reach the same goal. It is important to note that the strategy used in the EVLW group (and in fact in the WP group) was not simply fluid restriction but fluid restriction based on careful and frequent measurements and monitoring plus use of vasopressors, vasodilators, and inotropic agents, also all based on careful monitoring. This caveat is emphasized by the authors. Thus, the strategy in the EVLW group associated with successful outcome was based on the goal of minimizing fluid administration while maintaining cardiac output and tissue oxygen delivery. Any strategy that simply restricted fluid or employed diuresis without concern for other physiologic variables cannot be expected to have the same outcome and is potentially dangerous. If reduction in lung water is a therapeutic
989
goal, it is important that this be accomplished without adverselyaffecting cardiac output and tissue perfusion. In fact, the authors speculate that their data suggest that the true importance of hemodynamic monitoring is to avoid therapeutic complications. I draw two conclusions from this study. The first conclusion is a general onethat making careful, accurate observations that provide knowledge about multiple physiologic variables and their interactions in the critically ill patient combined with a logical and consistent management approach allowing use of this information in an integrated way is likely to yield better outcomes. The second conclusion is that this study seems to confirm the logical strategy (or hypothesis) that achievingthe lowestEVLW that maintains tissue perfusion will improve patient outcome. Whether the fluid management strategy employed in the EVLW group by Mitchell and coworkers is the most practical and efficient way to achieve this goal remains an open question.
LEONARD
D.
HUDSON
M.D.
Professor of Medicine Head, Division of Pulmonary and Critical Care Medicine University of Washington Seattle, WA References 1. Mitchell JP. Improved outcome based on fluid management in critically-ill patients requiring pulmonary artery catheterization. Am Rev Respir Dis 1992; 145:990-8. 2. Schuller 0, Mitchell JP, Calandrino FS, Schuster DP. Fluid balance during pulmonary edema: Is fluid gain a marker or a cause of poor outcome? Chest 1991; 100:1068-75. 3. Sibbald WJ, Cunningham DR, Chin ON. Noncardiac or cardiac pulmonary edema? A practical approach to clinical differentiation in critically ill patients. Chest 1983; 84:452-61. 4. Sibbald WJ, Short AK, Warshawski FJ, Cunningham DO, Cheung H. Thermal dye measurements of extravascular lung water in critically ill patients: Intravascular starling forces and extravascular lung water in the adult respiratory distress syndrome. Chest 1985; 87:585-92. 5. Sibbald WJ, Warshawski FJ, Short AK, Harris J, Lefcoe MS, Holliday RL. Clinical studies of measuring extravascular lung water by the thermal dye technique in critically ill patients. Chest 1983; 83:725-31.
