INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Acute respiratory distress syndrome (ARDS) in large part arises as a consequence of overwhelming inflammation. The evidence for this comes, firstly, from dynamic studies in vivo with radiolabeling techniques that show increased localization of neutrophils within the lung in ARDS (1). In addition, in established ARDS significantly elevated levels of the proinflammatory cytokines, interleukin-1 (IL-1), tumor necrosis factor (TNF-) (2), as well as products of inflammatory cell activation, including the protease enzymes, neutrophil elastase (3), and collagenase (4), are found within the alveolar air spaces. We have previously reported a highly significant association between elevated levels of the circulating protease enzyme elastase and alveolar interleukin-8 (IL-8) and the subsequent development of clinical lung injury and ARDS (5, 6). As a result of this overwhelming inflammatory response, endothelial activation and injury and consequent disruption of the alveolar/capillary interface occurs, producing leakage of a protein-rich fluid into the alveolar air spaces and the clinical presentation of ARDS.

Ferritin is a 480-kD iron-storage protein that sequesters iron in the ferric (Fe³⁺) state. Ferritin may have important antioxidant implications because free iron facilitates the formation of highly toxic hydroxyl radicals from superoxide anion (O₂^·) and hydrogen peroxide (H₂O₂) through the Haber-Weiss reaction (7):

Although iron bound to ferritin is generally unavailable for participation in the Haber-Weiss reaction, O₂^· and acidosis, conditions commonly encountered in disorders at risk for ARDS, may release iron from ferritin. Thus, ferritin-derived iron may worsen oxidative damage in critically ill patients, contributing to the pathologic derangements seen in ARDS. Furthermore, ferritin synthesis may be increased by proinflammatory cytokines such as IL-8, IL-6, and TNF-, which are increased and presumably contributing to the pathogenesis of ARDS (6, 8, 9). Clearly, elevated ferritin levels could be caused by oxidative stress, proinflammatory cytokines, and the degree of tissue injury seen in these critically ill patients.

In the present investigation, we sought to extend our initial findings of increased serum ferritin levels in at-risk patients who progress to ARDS. Our previous study population was a heterogeneous group of patients who were at-risk because of medical (predominantly sepsis) predisposing insults (10). Whether ARDS complicating such different initial insults shares a similar pathogenetic pathway is an important concern. Furthermore, the duration of illness prior to presentation to medical care in septic patients is often difficult to determine. Thus, it is possible and, indeed, probable that the inflammatory cascade has been activated for some time before presentation in most septic patients. By comparison, investigating trauma patients allows the identification of a well- defined homogenous at-risk patient group in whom the timing of the initiating event, namely, the trauma incident, is known. In this study, we have analyzed the initial circulating ferritin levels in blood samples taken in the emergency room in patients with multiple trauma and investigated whether these initial levels in this well-defined patient group correlated with subsequent clinical disease severity and/or circulating products of endothelial activation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We enrolled 42 consecutive patients with severe multiple trauma who were admitted to the Accident and Emergency Department of the Royal Infirmary Edinburgh, Scotland over a 2-yr period. Informed consent was obtained from the patients' relatives, or guardians, and the study was approved by the Lothian Health Board Ethics Committee. The severity of the trauma was measured using both anatomic (injury severity score: ISS) and physiologic (revised trauma score: RTS) methods (11). Severe trauma was defined as an ISS  15 (12). ARDS was defined using the European/American Consensus definition (13). This definition explains ARDS as having a Pa_O2/FI_O2  200 mm Hg and bilateral pulmonary infiltrates on chest radiograph. Multiple organ failure was defined using a modified scoring system as used previously (14).

Patient samples were obtained at the time of initial venous blood sampling for clinical indications in the Accident and Emergency Department. Five milliliters were collected and centrifuged at 1,000 × g for 10 min. The serum was aspirated and stored at  70° C. Serum ferritin was assayed by a standard commercially available double antibody ¹²⁵I radioimmunoassay (Diagnostic Products Corporation, Los Angeles, CA) using a rabbit antihuman ferritin antibody and a goat antirabbit antibody (10).

Because the serum ferritin values were not normally distributed, all values are presented as medians with their ranges in parentheses. Nonparametric analyses, specifically the Mann-Whitney test or Spearman's rank correlation when appropriate, were used for intergroup comparisons. Significance differences between the two groups was defined as a p value of < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient group details are outlined in Table 1. Forty-two patients were studied, 25 male and 17 female. Using the European/American Consensus definition, 16 (38%) patients subsequently developed ARDS, 11 men and five women. The median time from injury to blood sampling was 90 min (range, 45 to 240). The overall mortality rate was 21% (nine of 42) with a rate of 25% (four of 16) in the ARDS group and 19% (five of 26) in the non-ARDS group.

Serum ferritin levels on admission correlated with a measurement of the degree of initial trauma injury as assessed by the ISS (p < 0.05, r = 0.37) (Figure 1). In addition, initial ferritin levels correlated with both the degree and the severity of subsequent multiple organ failure (p < 0.05, r = 0.39) and the subsequent development of ARDS (European/American Consensus definition) (p < 0.05, r = 0.27) (progression to ARDS: median = 638 ng/ml; range, 70 to 4,500 versus nonprogression to ARDS: 185 ng/ml; range 12 to 2,850) (p = 0.02) (Figure 2). Serum ferritin levels did not correlate with other markers of clinical injury, namely, lowest Pa_O2/FI_O2 ratio (p = 0.67), days requiring ventilation (p = 0.09), or mortality (p = 0.42).

View larger version (10K): [in this window] [in a new window]   Figure 1.   Initial serum ferritin levels in all at-risk patients who subsequently progressed to ARDS compared with those who did not (p < 0.05).

View larger version (11K): [in this window] [in a new window]   Figure 2.   Initial serum ferritin levels and injury severity score (ISS) (p < 0.05) (r = 0.37).

Our previously published work on ferritin (10) used strict criteria for diagnosing ARDS, including necessity for mechanical ventilation, bilateral pulmonary infiltrates on chest radiograph, pulmonary capillary wedge pressure < 18 mm Hg, static pulmonary compliance < 50 ml/cm H₂O, and alveolar-to-arterial partial pressure of oxygen ratio < 0.25 (15). In order to compare both studies, we applied this alternative definition for ARDS in the current study and found that eight (19%) patients (all male) progressed to ARDS. Furthermore, the median ferritin level in the men who progressed to ARDS in the previous study was 925 ng/ml (range, 410 to 2,000 ng/ml) in comparison to a median ferritin of 988 ng/ml (range, 70 to 4,500 ng/ml) in the current study. For those men who did not progress to ARDS, the median ferritin level was 420 ng/ml (range, 225 to 550) in the previous study in comparison to a median ferritin of 475 ng/ml (range, 80 to 2,850 ng/ml) in the present study. Applying the previous definition for ARDS to the current study, no women progressed to ARDS, so we were unable to compare the median ferritin values for ARDS in women.

We have previously established cutoff points for serum ferritin in a heterogenous group of patients at-risk for ARDS (10) and found that a ferritin level > 680 ng/ml for men and a level > 270 ng/ml for women was associated with the development of ARDS. We also reported that the development of ARDS was predicted in men with a sensitivity of 60%, a specificity of 90%, and positive and negative predictive values of 75 and 82%, respectively (10). Using ARDS diagnosis criteria similar to that in our previous work (10), for men the results revealed a sensitivity of 86%, a specificity of 59%, and positive and negative predictive values of 46 and 91%, respectively. In this study, using older definition for ARDS, no women progressed to ARDS. Consequently, we are unable to compare predictive values for ARDS with our previously published values on female patients. When we applied the most recent definition for ARDS (European/American Consensus definition) to this study, for men, ARDS was predicted with a sensitivity of 73% and a specificity of 64% and the positive and negative predictive values were 62 and 75%, respectively. For women, the sensitivity was 60%, the specificity was 92%, and the positive and negative predictive values were 75 and 95%, respectively.

In 26 patients we also measured circulating levels of products of endothelial activation, namely, the soluble receptors, soluble ICAM-1 (sICAM-1) and E-selectin (sE-selectin). Serum ferritin levels correlated significantly with both sICAM-1 (p = 0.01, r = 0.21) and sE-selectin (p = 0.04, r = 0.37) levels.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present investigation, we found that initial serum ferritin levels were associated with the subsequent development of ARDS and multiple organ failure (MOF) in well-defined patients at-risk for trauma. Initial ferritin levels also reflected trauma severity as suggested by their correlation with the injury severity score (ISS). We also found a significant association between serum ferritin levels and certain products of endothelial activation.

The present findings add meaningfully to our previous observations that ferritin levels increase and are associated with the development of ARDS in at-risk persons (10). Firstly, it is valuable when an objective biochemical observation in a dynamic disease process can be repeated in another hospital and in a different group of patients. Because our previous study consisted of a heterogenous group of at-risk patients with a diversity of predisposing factors for ARDS development, it is reassuring when clinical results are repeated in a more homogenous group of patients. Secondly, the present findings show that ferritin increases very quickly in trauma patients, suggesting that ferritin increases are part of the process rather than an antecedent condition since trauma patients were ostensibly healthy before encountering injury. This contrasts with other at-risk patients, e.g., with sepsis, who may have been ill for a considerable time before developing ARDS. Thirdly, by finding that ferritin levels are associated with the development of both ARDS and MOF, the work further links these two processes and is consistent with the premise that ARDS is the pulmonary manifestation of a systemic process.

The mechanisms responsible for the increased serum ferritin are unknown. Serum ferritin levels may increase solely as a consequence of tissue damage or cell lysis. This possibility is supported by our finding of an association between serum ferritin levels and injury severity score. Serum ferritin levels could also reflect the intensity of the enhanced systemic circulatory and intrapulmonary inflammatory response that is manifest in trauma and other patients at risk for ARDS (5, 6, 14). Because synthesis of ferritin may be induced by proinflammatory cytokines such as IL-8 and TNF-, the observed higher serum ferritin levels could reflect the exaggerated host inflammatory response. Our findings of a significant association between circulating ferritin and products of endothelial activation, namely, sICAM-1 and sE-selectin, supports this potential mechanism. In addition, elevated serum ferritin levels may also reflect an increase in ferritin synthesis as a protective response to increased oxidative stress, increasing the ability to sequester iron and thereby increasing antioxidant activity (16, 17). However, the early time point of sampling in our trauma patient group (median time, 90 min post-trauma event) would argue against this potential explanation. An additional cause of elevated serum ferritin is liver impairment. However, only two of our enrolled patients had significant hepatic impairment, and their ferritin levels were 210 and 280 ng/ml, respectively.

The biologic significance of patients "at-risk" for ARDS having increased serum ferritin levels is unknown, but several hypotheses can be advanced. Increased ferritin may serve as a source of increased free iron, which can facilitate the formation of the highly toxic hydroxyl radical (18, 19). Although the iron content of serum ferritin is usually low, iron is catalytically involved in the Haber-Weiss reaction, and, consequently, even a small increase in available iron in the correct environment could augment the production of more toxic oxygen species. Furthermore, the serum of patients with ARDS has reduced levels of transferrin (20), which sequesters iron in the ferric state, and reduced unsaturated iron binding capacity, which limits the serum iron-handling antioxidant activity. Our recent work indicates that adding ferritin to the perfusate of isolated lungs worsens neutrophil-mediated leak (J. E. Repine, personal communication). As already mentioned, the elevated serum ferritin levels could possibly be a protective response to increased oxidative stress.

Control values for serum ferritin in men are different from those in women. We previously identified cutoff values for serum ferritin in patients at-risk for ARDS and found that for men a level > 680 ng/ml and for women a level > 270 ng/ml were associated with the development of ARDS. Since our previous report, the definition for ARDS has been revised, and accordingly we applied this updated European/American Consensus definition in the current study. If we adopt our previous ARDS diagnosis criteria in the current study, we found that for men, ARDS was predicted with a sensitivity of 86% and a specificity of 59%. This is in comparison to a sensitivity of 60% and a specificity of 90% in our previous study. When we applied our previous ARDS diagnosis criteria to the women in this study, none of them progressed to ARDS, so we could not compare predictive values for ARDS with our previous values for the female subjects. Therefore, a highly predictive cutoff for serum ferritin has not been identified in this study. Furthermore, even though both the ARDS and the non-ARDS groups had significantly different ferritin levels, there was a considerable overlap in the ferritin values between both groups (Figure 1).

The ability to predict ARDS and MOF remains a desirable goal (21). The development of markers that accurately forecast or exclude the development of ARDS is crucial for limiting the number of patients needed for the efficient testing of novel interventions as well as helping in the search for potential mechanisms for the pathogenesis of ARDS. In this study, we found that initial serum ferritin levels correlated with the subsequent development of both ARDS and multiple organ failure in a group of patients at-risk for trauma. However, the lack of specific predictive values suggests that serum ferritin alone is not ideal for predicting the development of ARDS. We found a significant correlation between initial serum ferritin levels and degree of trauma, but there was no correlation with other markers of clinical injury, namely, lowest Pa_O2/FI_O2 ratio, days requiring ventilation, or mortality. Therefore, serum ferritin in conjunction with other markers could be of use in identifying individual patients at high risk of progressing to ARDS.