Published ahead of print on June 16, 2004, doi:10.1164/rccm.200310-1434OC
© 2004 American Thoracic Society doi: 10.1164/rccm.200310-1434OC
Significance of Von Willebrand Factor in Septic and Nonseptic Patients with Acute Lung InjuryDivision of Allergy, Pulmonary and Critical Care Medicine, Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee; Division of Pulmonary and Critical Care Medicine and Division of Occupational and Environmental Medicine, Department of Medicine, and Department of Anesthesia and Cardiovascular Research Institute, University of California, San Francisco; Pulmonary/Critical Care Unit and ARDS Network Clinical Coordinating Center, Department of Medicine, Massachusetts General Hospital, Massachusetts; and Division of Pulmonary and Critical Care Medicine, Department of Medicine, Fletcher Allen Health Care, University of Vermont, Burlington, Vermont Correspondence and requests for reprints should be addressed to Lorraine B. Ware, M.D., Division of Allergy, Pulmonary and Critical Care Medicine, T1217 MCN, Vanderbilt University School of Medicine, Nashville, TN 37232–2650. E-mail: lorraine.ware{at}vanderbilt.edu
Systemic endothelial activation and injury are important causes of multiorgan system failure. We hypothesized that plasma levels of von Willebrand factor (VWF), a marker of endothelial activation and injury, would be associated with clinical outcomes in acute lung injury (ALI) and acute respiratory distress syndrome (ARDS). In 559 patients with ALI/ARDS enrolled in the National Heart, Lung, and Blood Institute ARDS Network trial of two VT strategies, plasma VWF levels were measured at randomization (mean 350 ± 265% of normal control plasma) and Day 3 (344 ± 207%). Baseline VWF levels were similar in patients with and without sepsis, and were significantly higher in nonsurvivors (435 ± 333%) versus survivors (306 ± 209%) even when controlling for severity of illness, sepsis, and ventilator strategy (increased odds ratio of death of 1.6 per SD size increase in VWF; 95% confidence interval, 1.4–2.1). Higher VWF levels were also significantly associated with fewer organ failure–free days. Ventilator strategy had no effect on VWF levels. In conclusion, the degree of endothelial activation and injury is strongly associated with outcomes in ALI/ARDS, regardless of the presence or absence of sepsis, and is not modulated by a protective ventilatory strategy. To improve outcomes further, new treatment strategies targeted at the endothelium should be investigated.
Key Words: acute respiratory distress syndrome • sepsis • von Willebrand factor Systemic endothelial activation and injury are important causes of multiorgan system failure in patients with sepsis (1). Unlike sepsis, acute lung injury and the acute respiratory distress syndrome (ALI/ARDS) are traditionally considered to be primarily pulmonary disorders. However, ALI/ARDS may have major systemic manifestations and may occur in the setting of a systemic inflammatory insult such as sepsis (2) or multiple blood transfusions (3). Multiorgan system failure is a common feature of ALI/ARDS from any cause, and death may often be attributed to nonpulmonary causes rather than to primary respiratory causes (4, 5). Thus, it may be more appropriate to characterize ALI/ARDS as a systemic disease. Furthermore, although pulmonary endothelial injury and activation are clearly present in patients with ALI/ARDS (6, 7), systemic endothelial activation and injury may also be a major pathogenetic feature. Von Willebrand Factor antigen (VWF) is a macromolecular antigen that is produced predominantly by endothelial cells and to a lesser extent by platelets. In the setting of endothelial activation or injury, VWF is released from preformed stores into the circulation (8, 9). VWF has been investigated as a biological marker of endothelial injury in patients both at risk for and with established ALI/ARDS (10–17). In one study of patients with sepsis, plasma VWF levels had predictive value for the development of ARDS (12). However, other studies of patients at risk for ALI/ARDS from a variety of causes did not confirm this association (11, 13, 14, 16). In one small single-center study of 51 patients with early ALI/ARDS, plasma VWF levels were independently associated with mortality (17). However, other studies have yielded conflicting results. For example, Sabharwal and colleagues did not find an association of VWF levels with death in 22 patients with ARDS (14). In addition to yielding conflicting results, none of the prior studies has been adequately powered to examine differences in VWF levels in established ALI/ARDS in relationship to the cause of lung injury or the mode of mechanical ventilation. Based on the hypothesis that ALI/ARDS is a systemic disorder with endothelial activation and injury in both the systemic and pulmonary circulation, we hypothesized that plasma levels of VWF would be elevated in patients with ALI/ARDS and would be associated with important clinical outcomes, including mortality. Because mechanical ventilation with higher VTs or pressures has been associated with lung endothelial injury in animal models (18–23), we also hypothesized that a lower VT strategy would be associated with lower plasma levels of VWF. To test these hypotheses, we measured VWF levels in plasma obtained from 559 patients enrolled in the National Institutes of Health ARDS Network multicenter randomized trial of a 6- versus 12-ml/kg VT ventilatory strategy (24). Analysis of VWF levels in this large, well-characterized group of patients allowed us to determine whether the degree of endothelial activation and injury, as measured by VWF levels, is associated with clinical outcomes in patients with ALI/ARDS and whether VWF levels were altered by a lung protective ventilatory strategy that used a low VT (6-ml/kg predicted body weight) and a plateau pressure limit (< 30 cm H2O). Some of the results of these studies have been previously reported in the form of an abstract (25).
Patients Patients were selected from the 861 patients in the National Heart, Lung, and Blood Institute ARDS Network randomized controlled trial (24) of two ventilatory strategies: VT of 6- versus 12-ml/kg predicted body weight. Five hundred fifty-nine patients were selected, based solely on the availability of baseline plasma samples. Patients who were included in this study and those not included (n = 302) had similar Acute Physiology and Chronic Health Evaluation III (APACHE III) scores, age, sex, creatinine, platelet count, direct versus indirect lung injury, PaO2/FIO2 ratios and a similar percentage were randomized to 6 versus 12 ml/kg. The protocol was approved by the institutional review board at each hospital, and informed consent was obtained from all patients or their surrogates except at one hospital where this requirement was waived.
Clinical Trial Procedures
Clinical Data
Measurement of VWF
Statistical Analysis
Patient Characteristics Baseline demographic and clinical characteristics are summarized in Table 1. There were 274 patients (49%) randomized to 6 ml/kg tidal volume and 285 patients (51%) randomized to 12 ml/kg (Table 2).
Baseline and Day 3 Plasma VWF Levels At baseline, mean plasma VWF was 350 ± 265% of control plasma (mean ± SD, n = 559). At Day 3, mean plasma VWF was similar (344 ± 207%) (n = 487). Because sepsis is a common cause of both ALI/ARDS and endothelial activation and injury, baseline VWF levels were compared between patients with and without sepsis and were not different (p = 0.82) (Figure 1A). In contrast, baseline VWF levels were lower in patients with trauma as the underlying cause of ALI/ARDS compared with those with other causes of ALI/ARDS (p = 0.0005) (Figure 1B). Baseline VWF levels were also modestly lower in patients with indirect lung injury compared with direct lung injury (mean 322 ± 269% vs. 377 ± 260%, respectively, p = 0.015). There was no relationship between treatment with lisofylline and VWF levels at baseline (p = 0.18) or Day 3 (p = 0.37). To evaluate the possibility of a contribution of platelets to levels of circulating VWF, VWF levels were compared between patients with and without thrombocytopenia, an indirect measure of platelet activation. VWF levels were not significantly different in patients with thrombocytopenia (platelet count of less than 50,000 per mm3) compared with those without thrombocytopenia (374 ± 222% vs. 348 ± 220%, p = 0.45).
Association of VWF Levels with Outcomes Both baseline plasma VWF levels (Figure 2A) and Day 3 plasma VWF levels (Figure 2B) were significantly associated with hospital mortality. Using multivariate analysis to control for ventilator protocol, a higher baseline VWF level was also independently associated with a higher risk of death, a reduction in ventilator-free days, and a reduction in organ failure–free days (Table 3). Controlling for ventilator strategy, APACHE III score, PaO2/FIO2 ratio, creatinine, platelet count, presence or absence of sepsis, and participation in the lisofylline trial, higher baseline VWF levels were independently associated with a higher risk of death (odds ratio, 1.6 per SD size increase in VWF; 95% confidence interval, 1.4–2.1) and a significant reduction in ventilator-free days (mean reduction of 1.1 days per SD size increase in VWF) and organ failure–free days (mean reduction of 1.4 days per SD size increase in VWF). When we restricted the analysis to subjects who had both baseline and Day 3 VWF values (n = 487), VWF remained associated with a greater risk of death (odds ratio, 1.7 per SD sized increment; 95% confidence interval, 1.3 to 2.1), fewer ventilator-free days (–1.7 days, p = 0.002), and fewer organ failure–free days (–2.2 days, p < 0.0001).
Association of VWF of 450% or More with Clinical Outcomes A plasma VWF level of 450% or more has previously been reported to be associated with a greater risk of developing ALI/ARDS in patients with sepsis (12) and has been associated with higher mortality in patients with established ALI/ARDS (17). In this study, we prospectively tested the value of this previously suggested cutoff. Hospital mortality to Day 180 was higher in patients with a baseline VWF of 450% or more (50% vs. 30%, p < 0.001) (Figure 3). Using multivariate analysis to control for ventilator protocol, a baseline VWF level of 450% or more was independently associated with a higher risk of death, a reduction in ventilator-free days, and a reduction in organ failure–free days (Table 3). Controlling for ventilator strategy and the other covariates, a baseline VWF level of 450% or more was independently associated with a near doubling of the risk of death (odds ratio, 1.9; confidence interval, 1.2–2.9).
Effect of Ventilator Strategy on Plasma VWF Levels There was no association between ventilator group and change in VWF levels during the first 3 study days (p = 0.96 by analysis of covariance) (Table 2). There was also no impact of the 6-ml/kg strategy on VWF levels, controlling for covariates (p = 0.84).
In this large, multicenter study of 559 patients with ALI/ARDS, higher plasma levels of VWF, a marker of endothelial activation and injury, were independently associated with adverse outcomes, including mortality, duration of unassisted ventilation, and organ failures. Also, a plasma VWF level of greater than 450%, a previously established threshold (12, 17), was associated with increased mortality. These findings lend support to the hypothesis that systemic and pulmonary vascular injury in ALI/ARDS may be major contributors to the pathogenesis of this disorder. In addition, there were several novel and unexpected findings. First, VWF levels were not different between patients with or without sepsis, and second, a lower VT ventilatory strategy was not associated with a reduction in plasma VWF levels. Our findings can be compared with previous reports of VWF levels in various disease processes (Table 4). As stated in the introduction, there have been several prior reports of VWF levels in patients both at risk for and with ALI/ARDS (10–17). In these studies, the mean reported plasma VWF ranged from 269 to 632% in patients at risk for ARDS and from 375 to 632% in patients with established ARDS. There is not complete concurrence in these studies with regard to the prognostic value of VWF, either in association with the development of ARDS or with outcomes in established ARDS. However, the largest prior study of established ARDS, a single-center study of 51 patients with ARDS from our group, also found that plasma VWF levels were independently associated with clinical outcomes (17). This study was not adequately powered to examine the association between cause of lung injury or ventilator strategy and VWF levels. Elevated levels of VWF have been reported in a wide array of vascular diseases, including ischemic heart disease (30), peripheral vascular disease (30), hypertension (30), diabetes (31), stroke (32), thrombotic thrombocytopenic purpura (33), and the vasculitides associated with collagen vascular diseases (34). Although levels are typically less than two times normal, these findings indicate that VWF is not specific for ALI/ARDS but rather is a marker of vascular injury or activation from diverse causes.
Sepsis is associated with widespread endothelial activation and injury (1, 35, 36), a process that likely contributes to the high frequency of organ failure in patients with sepsis (37). One might hypothesize that evidence of endothelial activation and injury in ALI/ARDS occurs predominantly in those patients with ALI/ARDS caused by sepsis. However, our findings do not support this conclusion. Indeed, plasma VWF levels were not significantly different in those patients with ALI/ARDS from sepsis compared with those with ALI/ARDS from other causes. This finding suggests that ALI might be an independent cause of systemic endothelial activation and injury, a hypothesis that deserves further study. This finding may help to explain why mortality in patients with ALI/ARDS is often due to multisystem organ failure rather than just a primary respiratory cause (4, 5). Although there were only 54 patients in this study with trauma as the underlying cause of ALI/ARDS, these patients clearly had lower levels of plasma VWF than patients with ALI/ARDS from other causes. This finding is in agreement with prior studies. Moss and colleagues reported that markers of endothelial cell activation, including VWF, were significantly lower in patients who were at risk for ALI/ARDS from trauma than those at risk from other clinical disorders (38). Thus, the pathogenesis of ALI/ARDS associated with trauma may be different than with other causes and may be associated with less endothelial activation and injury. This could, in part, explain why Treggiari and colleagues (39) found that a diagnosis of ALI was not associated with increased mortality in patients with severe trauma. In contrast to our hypothesis, plasma VWF levels were not modulated by the ventilator strategy during the first 3 days of treatment. This finding is somewhat counterintuitive, as both a lower VT and lower VWF levels were independently associated with better clinical outcomes. However, mortality in the low VT arm of the ARDS Network trial was 31%, indicating that there are factors other than ventilator-associated lung injury that contribute to ARDS mortality. The effect of the ventilator strategy on specific markers of endothelial activation and injury has not previously been investigated clinically. One group reported that bronchoalveolar lavage fluid from patients with ARDS could cause in vitro endothelial cell cytotoxicity, but the significance of these findings to the in vivo lung is unclear (40). In experimental studies, ventilation of normal lungs with very high VTs can lead to pulmonary endothelial injury, as manifested by protein rich pulmonary edema (18–22). In one recent study of acid-induced lung injury in rats (23), ventilation with 6-ml/kg VT compared with 12-ml/kg VT led to a markedly lower plasma VWF levels and less lung endothelial injury at 6 hours. However, in this experimental study, the animals were randomized to the different ventilation strategies after 2 hours of injury, whereas in the clinical study, randomization could occur up to 36 hours after the onset of ALI/ARDS. Therefore, lung endothelial injury may have already reached a plateau in clinical ALI/ARDS. Another explanation for the difference between the clinical and experimental findings is that in clinical ARDS plasma VWF levels are associated more with the degree of systemic endothelial injury rather than simply with pulmonary endothelial injury. In support of this hypothesis, Muller and colleagues reported that VWF staining in the lung vascular bed was not different between normal human lungs and lungs from patients who died from ARDS associated with Gram-negative sepsis (41). A third explanation is that the 6-ml/kg ventilation strategy may primarily protect the lung epithelium and thereby influence proinflammatory mechanisms generated within the airspace compartment. This hypothesis is consistent with our recently reported finding that levels of circulating surfactant protein D, an alveolar epithelial type II cell marker, were attenuated by low VT ventilation in the ARDS Network trial of 6-ml/kg versus 12-ml/kg ventilation (42). Circulating interleukin-6 levels were also lower in 6-ml/kg arm (24), a finding that might reflect decreased intra-alveolar inflammation, as has been reported in experimental models (22, 43).
There are some limitations to this study. First, because of patient deaths, there were 58 fewer patients who had VWF levels measured at Day 3 than at baseline. However, it is unlikely that this dropout was a confounding factor. Because VWF levels were higher in patients that died, the dropout of patients dying before Day 3 would actually tend to reduce the difference observed between survivors and those that died. Furthermore, when we restricted the analysis to subjects who had both baseline and Day 3 VWF levels, VWF remained associated with a greater risk of death. A second limitation is that the precise sources of VWF cannot be determined from this study. The only cell types known to release VWF are endothelial cells and platelets. The pulmonary endothelium is rich in VWF, and histopathologic studies have shown that VWF is present in hyaline membranes, suggesting localized release by the pulmonary endothelium (6). The finding of elevated pulmonary dead space in patients with ALI/ARDS (44) also suggests widespread loss of alveolar–capillary gas exchange surface in these patients, which may, in part, reflect pulmonary endothelial injury. However, it seems likely that circulating VWF in ALI/ARDS patients also originates from the systemic endothelium. In healthy volunteers, both endotoxin and tumor necrosis factor- Although our findings indicate that there is increased release of VWF into the circulation in patients with ALI/ARDS, the high levels of VWF may also reflect decreased clearance. The clearance of VWF from the circulation is poorly understood. A recent study in a mouse model suggests that the liver may be the primary site of clearance, but human studies are lacking (49). In normal individuals, large VWF multimers are cleaved to smaller multimers in the plasma by a specific VWF-cleaving protease, known as ADAMTS-13 (50, 51). Low levels of ADAMTS-13 are a hallmark of thrombotic thrombocytopenic purpura (52) but have also been reported in a variety of patients with critical illness, including patients with decompensated liver cirrhosis and patients with acute inflammatory conditions including respiratory infections (53). Decreased levels of ADAMTS-13 in the setting of ALI might contribute to the elevated plasma VWF levels, but it is not known whether cleavage by ADAMTS-13 is necessary for clearance from the plasma in humans. In conclusion, in a large, multicenter, randomized trial of a protective ventilatory strategy for ALI/ARDS, elevated plasma levels of VWF were independently associated with more systemic organ failure and mortality, suggesting that the degree of pulmonary and systemic endothelial activation and injury may be an important determinant of clinical outcomes in patients with clinical lung injury. To improve outcomes further, investigation of new treatment strategies for ALI/ARDS should include systemic and pulmonary endothelial injury as potential targets. NATIONAL INSTITUTES OF HEALTH NATIONAL HEART, LUNG, AND BLOOD INSTITUTE ARDS NETWORK Network Participants: Cleveland Clinic Foundation, Herbert P. Wiedemann, M.D.,* Alejandro C. Arroliga, M.D., Charles J. Fisher, Jr., M.D., John J Komara, Jr., B.A., R.R.T., Patricia Periz-Trepichio, B.S., R.R.T.; Denver Health Medical Center, Polly E. Parsons, M.D., Denver VA Medical Center, Carolyn Welsh, M.D.; Duke University Medical Center, William J. Fulkerson, Jr., M.D.,* Neil MacIntyre, M.D., Lee Mallatratt, R.N., Mark Sebastian, M.D., John Davies, R.R.T., Elizabeth Van Dyne, R.N., Joseph Govert, M.D.; Johns Hopkins Bayview Medical Center, Jonathan Sevransky, M.D., Stacey Murray, R.R.T.; Johns Hopkins Hospital, Roy G. Brower, M.D., David Thompson, M.S., R.N., Henry E. Fessler, M.D.; LDS Hospital, Alan H. Morris, M.D.,* Terry Clemmer, M.D., Robin Davis, R.R.T., James Orme, Jr., M.D., Lindell Weaver, M.D., Colin Grissom, M.D., Frank Thomas, M.D., Martin Gleich, M.D. (posthumous); McKay-Dee Hospital, Charles Lawton, M.D., Janice D'Hulst, R.R.T.; MetroHealth Medical Center of Cleveland, Joel R. Peerless, M.D., Carolyn Smith, R.N.; San Francisco General Hospital Medical Center, Richard Kallet, M.S., R.R.T., John M. Luce, M.D.; Thomas Jefferson University Hospital, Jonathan Gottlieb, M.D., Pauline Park, M.D., Aimee Girod, R.N., B.S.N., Lisa Yannarell, R.N., B.S.N.; University of California, San Francisco, Michael A. Matthay, M.D.,* Mark D. Eisner, M.D., M.P.H., John Luce, M.D., Brian Daniel, R.C.P., R.R.T., Thomas J. Nuckton, M.D.; University of Colorado Health Sciences Center, Edward Abraham, M.D.,* Fran Piedalue, R.R.T., Rebecca Jagusch, R.N., Paul Miller, M.D., Robert McIntyre, M.D., Kelley E. Greene, M.D.; University of Maryland, Henry J. Silverman, M.D.,* Carl Shanholtz, M.D., Wanda Corral, B.S.N., R.N., University of Michigan, Galen B. Toews, M.D.,* Deborah Arnoldi, M.H.S.A., Robert H. Bartlett, M.D., Ron Dechert, R.R.T., Charles Watts, M.D.; University of Pennsylvania, Paul N. Lanken, M.D.,* Harry Anderson, III, M.D., Barbara Finkel, M.S.N., R.N., C. William Hanson, III, M.D.; University of Utah Hospital, Richard Barton, M.D., Mary Mone, R.N.; University of Washington/Harborview Medical Center, Leonard D. Hudson, M.D.,* Greg Carter, R.R.T., Claudette Lee Cooper, R.N., Annemieke Hiemstra, R.N., Ronald V. Maier, M.D., Kenneth P. Steinberg, M.D.; Utah Valley Regional Medical Center, Tracy Hill, M.D., Phil Thaut, R.R.T.; Vanderbilt University, Arthur P. Wheeler, M.D.,* Gordon Bernard, M.D.,* Brian Christman, M.D., Susan Bozeman, R.N., Linda Collins, Teresa Swope, R.N., Lorraine B. Ware, M.D. Clinical Coordinating Center: Massachusetts General Hospital, Harvard Medical School, David A. Schoenfeld, Ph.D.,* B. Taylor Thompson, M.D., Marek Ancukiewicz, Ph.D., Douglas Hayden, M.A., Francine Molay, M.S.W., Nancy Ringwood, B.S.N., R.N., Gail Wenzlow, M.S.W., M.P.H., Ali S. Kazeroonin, B.S. NHLBI Staff: Dorothy B. Gail, Ph.D., Andrea Harabin, Ph.D.,* Pamela Lew, Myron Waclawiw, Ph.D. *Steering Committee: Gordon R. Bernard, M.D., Chair, Principal Investigator from each center as indicated by an asterisk. Data and Safety Monitoring Board: Roger G. Spragg, M.D., Chair, James Boyett, Ph.D., Jason Kelley, M.D., Kenneth Leeper, M.D., Marion Gray Secundy, Ph.D., Arthur Slutsky, M.D. Protocol Review Committee: Joe G. N. Garcia, M.D., Chair, Scott S. Emerson, M.D., Ph.D., Susan K. Pingleton, M.D., Michael D. Shasby, M.D., William J. Sibbald, M.D.
A complete list of the members of The Acute Respiratory Distress Syndrome Network appears at the end of the article. Supported by National Institutes of Health HL 70521, National Institutes of Health HL 51856, and National Heart, Lung, and Blood Institute contracts 46054–46064. Conflict of Interest Statement: L.B.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.D.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; B.T.T. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.A.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form October 21, 2003; accepted in final form June 15, 2004
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
infusion increased circulating VWF levels (
