Ventilator-induced lung injury (VILI) is an inflammatory process that can be attenuated by lung protective ventilation strategies. Our objectives to further investigate the pathogenesis of ALI and VILI and the mechanism of lung protection in these syndromes were: 1) to determine if plasma measurements of soluble TNF receptor I (sTNFRI) and II (sTNFRII) would predict the development of ALI and mortality in a small single center trial; 2) to test the predictive value of these markers and of TNF-α in a larger, broader group of patients with ALI; 3) to test the hypothesis that low tidal volume ventilation (LTVV) would be associated with a decrease in plasma levels of TNF-α, sTNFRI, and sTNFRII. In the single center study, sTNFRI and II levels were higher in patients at risk for and with ALI, but they did not predict the development of the syndrome. In the multicenter trial sTNFRI and II were strongly associated with mortality (OR 5.76/1 log10 increment in receptor level; 95% CI 2.63–12.6 and OR 2.58; 95% CI 1.05–6.31, respectively) and morbidity measured as fewer nonpulmonary organ failure-free and ventilator-free days. The LTVV strategy was associated with an attenuation of plasma sTNFRI levels. In vitro, stimulated A549 cells release sTNFRI but not sTNRFII. In conclusion, plasma levels of sTNFRI and II can serve as biomarkers for morbidity and mortality in patients with ALI. Furthermore, LTVV is associated with a specific decrease in sTNFRI levels. This suggests that one beneficial effect of LTVV may be to attenuate alveolar epithelial injury.
- ventilator-induced lung injury
- acute respiratory distress syndrome
- low tidal volume ventilation
- TNF-α, sTNFRI, sTNFRII
- tumor necrosis factor
- soluble tumor necrosis factor receptor
the national heart, lung and blood institute (NHLBI) Acute Respiratory Distress Syndrome (ARDS) network multicenter randomized controlled trial of 6 vs. 12 ml/kg predicted body weight demonstrated that, in patients with acute lung injury (ALI) and ARDS, mortality was significantly reduced from 40 to 31% in the 6 ml/kg group (2). Ventilator-free days and organ failure-free days were also significantly reduced in the low tidal volume group. Previous studies have suggested that ventilator-induced lung injury (VILI) is an inflammatory process that can be attenuated by lung protective ventilation strategies (8, 9, 25). Understanding the mechanisms that contribute to lung protection could increase insight into the pathogenesis of ALI and VILI and contribute to the development of additional therapeutic interventions.
The role of pro- and anti-inflammatory cytokines in the pathogenesis of ALI and VILI has been the focus of intense investigation. On the basis of several experimental studies, the proinflammatory cytokine, TNF-α, is an important early mediator of ALI (18, 27, 30). Despite the experimental data, plasma, bronchoalveolar lavage (BAL), and pulmonary edema fluid levels of TNF-α have not been consistently correlated with clinical outcomes in patients at risk for or with ALI (12, 17, 21, 24). This result may be due in part to the fact that these studies did not consider the interplay between TNF-α and the two cell surface receptors that mediate its inflammatory effects: TNFRI and TNFRII. In response to inflammatory stimuli including TNF-α, soluble TNF receptors sTNFRI and sTNFRII are shed from the cell surface (5, 14). Elevated circulating levels of these receptors are often considered to be markers of a proinflammatory state, although their mechanism of action is actually anti-inflammatory. These soluble receptors bind to circulating TNF-α and compete with TNF for cell surface receptor binding. Thus the reported lack of correlation of TNF-α levels or activity with clinical outcomes may be related to sTNFR that have downregulated the activity of TNF-α. Alternatively, levels of sTNFR themselves may be a better indication of the biological impact of TNF-α.
TNFR are increased in the circulation of patients with trauma (10, 11, 13, 26) and patients who have undergone coronary bypass graft surgery (7). Although these studies were small, each suggested that there is an association between sTNFR in the plasma and morbidity and mortality. The relationship between circulating sTNFR levels and the development of ARDS or outcome from ARDS is not known. Interestingly, in one small, randomized, single center trial of 44 patients with ARDS, Ranieri et al. (25) found that a lung protective ventilatory strategy was associated with a decrease of both TNF-α and both TNFR in the plasma and BAL of patients treated with a lung protective ventilatory strategy compared with the patients treated with conventional ventilation.
In this study our first objective was to determine whether plasma measurements of sTNFRI and sTNFRII levels would predict the development of ALI and mortality in a small single center trial. The second objective was to test the predictive value of these biological markers and of TNF-α in a larger and broader group of patients with ALI. The third objective was to test the hypothesis that low tidal volume ventilation would be associated with a decrease in plasma levels of TNF-α and sTNFRI and sTNFRII. The fourth objective was to determine whether alveolar epithelial cells are a potential source of TNFR.
Study Design and Patient Selection
Single center study.
The hypothesis that sTNFRI and II levels would predict the development of ARDS in at-risk patients and mortality among patients with established ARDS was initially tested in a single center prospective study. In a single medical intensive care unit, patients with a defined risk factor for ARDS were identified and enrolled within 24 h of meeting the at-risk criteria. The defined clinical risk factors included sepsis, aspiration, hypertransfusion, and pancreatitis. The detailed criteria for these clinical risk factors have been previously reported (15, 16, 19).
ARDS was defined according to the criteria proposed by the American-European Consensus Conference (4). The patients had to have a PaO2/FiO2 ratio <200, a chest radiograph that showed bilateral infiltrates consistent with pulmonary edema, and no clinical evidence of increased left atrial pressure.
Informed consent was obtained for all subjects before enrollment. Peripheral blood was obtained within 24 h of meeting either at-risk or ARDS criteria. On the day of enrollment, plasma samples were obtained from 95 subjects: 10 healthy controls, 35 subjects with established ARDS, and 50 subjects at risk for ARDS. Those subjects at risk were followed, and 11 of them subsequently developed ARDS. A blood sample was obtained from each of those subjects when they had ARDS. This study was approved by the Institutional Review Board at the University of Colorado.
To further evaluate the hypotheses that levels of sTNFRI and II are associated with mortality from ALI/ARDS and that that low tidal volume ventilation would attenuate sTNFRI and II, we measured plasma levels of TNF-α, sTNFRI, and sTNFRII in a larger patient population comprising a subset of patients enrolled in the NHLBI ARDS Clinical Trials Network multicenter randomized controlled trial of 6 vs. 12 ml/kg tidal volume (2). Two other clinical trials, in which ketoconazole or lisofylline was compared with placebo in a factorial design, were conducted simultaneously. A detailed clinical protocol has been previously reported (2). In brief, intubated patients with ALI were randomized to receive 6 or 12 ml/kg tidal volume predicted body weight ventilation (2). At the time of enrollment, clinical data including acute physiology and chronic health evaluation (APACHE) III and clinical risk factor were determined. At baseline and subsequently on days 1, 2, 3, 4, 7, 14, 21, and 28, additional clinical variables including laboratory data, ventilator parameters, and vasopressor use were documented. The primary outcome measurement was mortality before discharge home with unassisted breathing. The patients were followed for 180 days or until they had been discharged home with unassisted breathing. For this study the secondary outcomes were ventilator-free days and organ failure-free days (1–3, 6). Ventilator-free days were defined as the number of days of unassisted breathing from days 1–28. Unassisted breathing had to continue for >48 consecutive hours. We calculated organ failure-free days by subtracting the number of days with organ failure from either 28 days or the number of days until death, whichever was fewer.
Blood samples were obtained at the time of enrollment (day 0), day 1, and day 3 of the protocol. As described in previous publications (1–3), the protocols were reviewed by an NHLBI Protocol Review Committee and by each participating institution's Institutional Review Board. With the exception of one hospital where the requirement for consent was waived, informed consent was obtained from each patient or his/her surrogate.
Plasma samples from the NHLBI ARDS clinical network trial represent a valuable, limited resource. Accordingly, initial measurements of TNF-α, sTNFRI, and sTNFRII were made in the plasma collected at day 0 (time of enrollment) and day 3 from 377 consecutively enrolled patients. At this point, the distribution of values was examined. Because TNF-α levels were detectable only in a minority of patients (9%), we did not measure TNF-α in subsequent patients. Because sTNFRI and sTNFRII levels were similar (r = 0.90, P < 0.001), we selected sTNFRI, which is the predominate TNF-α receptor in pulmonary epithelium (5, 14, 23), to measure in the subsequent 185 consecutively enrolled patients. There were no differences in age, APACHE III, PaO2/FiO2, tidal volume group assignment, vasopressor use, platelet count, or creatinine between this group of 185 consecutively enrolled patients and the first 377 patients (P > 0.10 in all cases).
TNF-α and sTNFRI and II Measurements
Measurements were made by standard sandwich ELISA according to established protocols. The matched antibody pairs (capture and detection) for TNF-α were obtained from Biosciences Pharmingen (San Diego, CA), for TNFRI from R & D Systems (Minneapolis, MN), and for TNFRII from Biosource International (Camarillo, CA). The protein standards came with the pairs.
sTNFRI and II Release From Human Alveolar Epithelial-Like Cells
Human alveolar epithelial-like cells (A549 cells; ATCC, Manassas, VA) were grown in MEM with 10% FBS in 5% CO2. Cells were plated in 24-well plates at a density of 5 × 104 cells/well. At 72 h when cells reached confluence the medium was replaced with serum-free MEM containing 20 ng/ml of cytomix, a mixture of equal amounts of three proinflammatory cytokines (TNF-α, IL-1β, IFN-γ; R&D Systems). Cytomix has been used in our prior clinical and experimental studies as a standard method to challenge the alveolar epithelium with proinflammatory stimuli that are clinically relevant (20, 29). Each condition was assayed in triplicate. After 6 h, conditioned medium was removed and stored at −20° until measurement of sTNFR1 and sTNFRII with commercially available ELISA assays (R&D Systems).
The analysis was conducted using SAS 8.2 (Cary, NC). Because TNF-α, TNFRI, and TNFRII are not normally distributed, values were expressed as the median with 25th–75th interquartile range.
The Kruskal-Wallis test was used to compare TNFRI and TNFRII levels among four groups in the single center study: normal subjects, patients at risk for ARDS who never developed ARDS, patients at risk for ARDS who developed the syndrome, and patients with established ARDS. We conducted pairwise comparisons using Tukey's test on rank-transformed data.
The bivariate relationship between TNFRI and TNFRII levels and mortality was evaluated by the Wilcoxon test. Because TNF-α was often undetectable, we examined the proportion of those with detectable levels among survivors and nonsurvivors using the χ2-test.
In the multicenter trial, we examined the association between baseline TNFR levels and morbidity indexes (ventilator-free days and organ failure-free days) by linear regression analysis, controlling for mechanical ventilation strategy. On the basis of model diagnostics, TNFRI and TNRFII were log10 transformed. Multivariate linear regression was used to control for ventilation strategy and markers of illness severity: APACHE III score, PaO2/FiO2, creatinine, platelet count, and vasopressor use. The impact of TNFR levels on mortality was examined in logistic regression analysis in analogous fashion.
Analysis of covariance was used to evaluate whether mechanical ventilation strategy was associated with the change of TNFR levels over time. For TNF-α, logistic regression analysis was used to examine the impact of ventilator group on the proportion with detectable levels over time. Neither ketoconazole nor lisofylline changed clinical outcomes (1, 3) so they were not considered as confounding variables.
Single Center Study
sTNFR I and sTNFR II levels were significantly higher in subjects at risk for ARDS than in normal subjects (Table 1). sTNFRI and II levels were also significantly higher in subjects with ARDS than in normal subjects (Table 1). There was no statistical difference in sTNFRI and II levels between subjects at risk for and with established ARDS (Table 1). Among the 11 subjects who had blood drawn both when they were at risk and when they developed ARDS, there were no differences in the sTNFR I or II measurements between the two time points. Thus, although sTNFRI and II levels were significantly increased in subjects at risk for and with ARDS, they did not predict the development of ARDS.
There was a suggestion that sTNFRI and II levels were higher among subjects at risk for ARDS and with established ARDS who ultimately died (Table 2), although this difference was not statistically significant. Because this small study was underpowered to evaluate the relationship with mortality, measurements were then carried out in plasma from subjects enrolled in a large multicenter trial.
At the time of enrollment in the clinical trial, the median plasma TNF-α level was 0 [25th–75th interquartile range (0–0)]. TNF-α was detectable in only 34 subjects (9%). The median levels (25th–75th interquartile range) of sTNFRI and sTNFRII were 8,511 pg/ml (5,525–15,699) and 3,131 pg/ml (2,086–5,790), respectively.
Plasma levels of TNF-α were not different at baseline or at day 3 for those subjects who did and did not survive. The proportion of subjects at baseline with detectable TNF-α levels was similar among those who lived (9.5%) vs. those who died (8.2%).
In contrast, sTNFRI and sTNFRII were strongly associated with mortality (Tables 3 and 4). Baseline levels of both sTNFRI and sTNFRII were significantly associated with a higher risk of death, if one controls for indicators of acute illness severity [odds ratio (OR) 5.76 per 1 log10 increment in receptor level; 95% confidence interval (CI) 2.63–12.6 and odds ratio (OR) 2.58; 95% CI 1.05–6.31, respectively] (Tables 3 and 4). Furthermore, higher baseline levels of sTNFRI and II were also associated with greater morbidity as indicated by fewer nonpulmonary organ failure-free days and fewer ventilator-free days (Table 4).
Impact of Low Tidal Volume
The low tidal volume strategy was associated with a decrease in the median values of plasma sTNFRI levels by 84 pg/ml during the first 3 days, whereas median sTNFRI levels increased by 114 pg/ml in the 12 ml/kg ventilator group (P = 0.037). The overall reduction in sTNFRI in the low tidal volume group was 8.3% (CI 0.5–15.5%) compared with the 12 ml/kg strategy (Table 5 and Fig. 1). In contrast, sTNFRII levels were not statistically different between the two tidal volume strategies over time (P = 0.56, Table 5).
To evaluate the alveolar epithelium as a potential source of the TNFRI, we measured baseline and stimulated release of TNFRI and II from human alveolar epithelial-like (A549) cells. There was no detectable release of sTNFR II from these cells. However, A549 cells released sTNFRI at baseline, and the release was significantly increased following exposure to cytomix that included equal amounts of TNF-α, IL-1β, and IFN-γ (Fig. 2).
In the small single center study, our findings suggested a possible association between soluble plasma TNFRI and TNFRII levels and mortality. In the large multicenter trial of patients with ALI, we found that sTNFRI and II are independently related to mortality and other important clinical outcomes. In addition, the 6 ml/kg mechanical ventilator strategy attenuated plasma levels of sTNFRI, suggesting that this strategy may be associated with a reduction in inflammatory lung injury. The finding that TNFRI but not TNFRII is released from alveolar epithelial cells could suggest that the reduction in sTNFRI with low tidal volume ventilation may, in part, reflect an attenuation of alveolar epithelial injury.
TNF-α was undetectable in the majority of patients with ALI. Even when detectable, TNF-α had no prognostic value for clinical outcomes in patients with ALI. These data suggest that plasma TNF-α is not a valuable biomarker in ALI.
Although TNFRI and II are released simultaneously from myriad cell types, epithelial cells predominantly express TNFRI (14, 28). Immunohistochemistry studies have previously shown that TNFRI is the predominant TNFR in human and murine lungs. Additional studies demonstrated that TNF induces the upregulation of mRNA for TNFRI but not TNFRII in epithelial cell lines (23, 28). Our study extends these findings by demonstrating that a mixture of TNF-α, IL-1β, and IFN-γ, which are known to be present in the lungs of patients with ARDS, stimulated the release of sTNFRI but not sTNFRII from alveolar epithelial-like cells (A549).
The association of morbidity and mortality in ALI patients with elevations of both sTNFRI and sTNFRII does not allow for the site or source of their release to be identified. Cells other than epithelial cells do release sTNFR. For example, neutrophils release both TNFRI and II (22). However, the specific decrease in sTNFRI with lung protective ventilation suggests that this ventilatory strategy may be associated with a decrease in pulmonary inflammation and, perhaps, specifically, with less lung epithelial injury. This is a particularly intriguing result since one recent experimental study reported that a lung protective ventilatory strategy has a major effect on attenuating alveolar epithelial injury measured at molecular, biochemical, morphological, and physiological levels (9). In addition, we have found that the rise in plasma surfactant protein D, a product of alveolar type II cells, is significantly attenuated by low tidal volume ventilation (6). Taken in aggregate, these studies implicate a role for the alveolar epithelium in VILI although additional studies are needed to clarify the mechanism of injury.
In conclusion, the results of this study indicate that plasma levels of sTNFRI and II can serve as biomarkers for morbidity and mortality in patients with ALI. Furthermore, the study confirms that low tidal volume ventilation is associated with a decrease in inflammation as measured by sTNFRI levels. The specific decrease in sTNFRI suggests that one beneficial effect of low tidal volume ventilation may be on the lung epithelium.
National Institutes of Health
National Heart, Lung, and Blood Institute ARDS Network
Network Participants: Cleveland Clinic Foundation-Herbert P. Wiedemann, MD*; Alejandro C. Arroliga, MD; Charles J. Fisher, Jr., MD; John J. Komara, Jr., BA, RRT; Patricia Periz-Trepichio, BS, RRT; Denver Health Medical Center-Polly E. Parsons, MD; Denver Veterans Affairs Medical Center-Carolyn Welsh, MD; Duke University Medical Center-William J. Fulkerson Jr., MD*; Neil MacIntyre, MD; Lee Mallatratt, RN; Mark Sebastian, MD; John Davies, RRT; Elizabeth Van Dyne, RN; Joseph Govert, MD; Johns Hopkins Bayview Medical Center-Jonathan Sevransky, MD; Stacey Murray, RRT; Johns Hopkins Hospital-Roy G. Brower, MD; David Thompson, MS, RN; Jonathan Sevransky, MD; Stacey Murray, RRT; Latter-Day Saints Hospital-Alan H. Morris, MD*; Terry Clemmer, MD; Robin Davis, RRT; James Orme, Jr., MD; Lindell Weaver, MD; Colin Grissom, MD; Frank Thomas, MD; Martin Gleich, MD (posthumous); McKay-Dee Hospital-Charles Lawton, MD; Janice D'Hulst, RRT; MetroHealth Medical Center of Cleveland-Joel R. Peerless, MD; Carolyn Smith, RN; San Francisco General Hospital Medical Center-Richard Kallet, MS, RRT; John M. Luce, MD; Thomas Jefferson University Hospital-Jonathan Gottlieb, MD; Pauline Park, MD; Aimee Girod, RN, BSN; Lisa Yannarell, RN, BSN; University of California, San Francisco-Michael A. Matthay, MD*; Mark D. Eisner, MD, MPH; John Luce, MD; Brian Daniel, RCP, RRT; University of Colorado Health Sciences Center-Edward Abraham, MD*; Fran Piedalue, RRT; Rebecca Jagusch, RN; Paul Miller, MD; Robert McIntyre, MD; Kelley E. Greene, MD; University of Maryland-Henry J. Silverman, MD*; Carl Shanholtz, MD; Wanda Corral, BSN, RN; University of Michigan-Galen B. Toews, MD*; Deborah Arnoldi, MHSA; Robert H. Bartlett, MD; Ron Dechert, RRT; Charles Watts, MD; University of Pennsylvania-Paul N. Lanken, MD*; Harry Anderson, III, MD; Barbara Finkel, MSN, RN; C. William Hanson, III, MD; University of Utah Hospital-Richard Barton, MD; Mary Mone, RN; University of Washington/Harborview Medical Center-Leonard D. Hudson, MD*; Greg Carter, RRT; Claudette Lee Cooper, RN; Annemieke Hiemstra, RN; Ronald V. Maier, MD; Kenneth P. Steinberg, MD; Utah Valley Regional Medical Center-Tracy Hill, MD; Phil Thaut, RRT; Vanderbilt University-Arthur P. Wheeler, MD*; Gordon Bernard, MD*; Brian Christman, MD; Susan Bozeman, RN; Linda Collins; Lorraine B. Ware, MD.
Clinical Coordinating Center: Massachusetts General Hospital, Harvard Medical School-David A. Schoenfeld, PhD*; B. Taylor Thompson, MD; Marek Ancukiewicz, PhD; Douglas Hayden, MA; Francine Molay, MSW; Nancy Ringwood, BSN, RN; Gail Wenzlow, MSW, MPH; Ali S. Kazeroonin, BS.
NHLBI Staff: Dorothy B. Gail, PhD; Andrea Harabin, PhD*; Pamela Lew; Myron Waclawiw, PhD.
*Steering Committee: Gordon R. Bernard, MD, Chair; Principal Investigator from each center as indicated by an asterisk above.
Data and Safety Monitoring Board: Roger G. Spragg, MD, Chair; James Boyett, PhD; Jason Kelley, MD; Kenneth Leeper, MD; Marion Gray Secundy, PhD; Arthur Slutsky, MD.
Protocol Review Committee: Joe G. N. Garcia, MD, Chair; Scott S. Emerson, MD, PhD; Susan K. Pingleton, MD; Michael D. Shasby, MD; William J. Sibbald, MD.
This work was supported by NHLBI Contracts NO1-HR-46054, 46055, 46056, 46057, 46058, 46059, 46060, 46061, 46062, 46063, and 46064. M. D. Eisner was also supported by NHLBI Grant K23 HL-04201. M. A. Matthay was supported by NHLBI Grants R01 HL-51856 and HL-51854, and P50HL-74005. L. B. Ware was supported by NHLBI Grant HL-70521.
The authors thank May Gillespie and Dr. Jay Westcott for help with the cytokine assays.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society