This article requires a subscription to view the full text. If you have a subscription you may use the login form below to view the article. Access to this article can also be purchased.
It has been known for decades that shock and sepsis can cause a syndrome of acute respiratory failure with characteristics of non-cardiogenic pulmonary edema. Over the years, this syndrome has been given a number of names, including congestive atelectasis, traumatic wet lung, and shock lung. In 1967 the modern counterpart to this syndrome was described and subsequently called the “acute respiratory distress syndrome” (ARDS). This syndrome results from lung injury and inflammation. As with inflammation elsewhere, ARDS is accompanied by many cellular and molecular processes, some of them specific to the syndrome, others perpetuating the syndrome, and others inactivating the by-products of inflammation. Since no specific clinical sign or diagnostic test has yet been described that identifies ARDS, its diagnosis is based on a constellation of clinical, hemodynamic, and oxygenation criteria. Current ARDS treatment is mainly supportive, since these patients frequently have coexisting conditions. Although in 1994 a new standard ARDS definition was accepted, that definition failed to standardize the measurement of the oxygenation defect and does not recognize different severities of pulmonary dysfunction. Based on current evidence there is a need for a better definition and classification system that could help us to identify ARDS patients who would be most responsive to supportive therapies and those unlikely to benefit because of the severity of their disease process. This paper examines our current understanding of ARDS and discusses why the current definition may not be the most appropriate for research and clinical practice.
- acute respiratory distress syndrome
- acute lung injury
- positive end-expiratory pressure
- lung inflammation
In August 1967, Ashbaugh et al1 published an article in the British journal The Lancet in which they described for the first time a syndrome that they termed the “acute respiratory distress syndrome” (ARDS). They studied a cohort of 272 patients who were receiving respiratory support, and in that cohort they identified 12 patients with a syndrome that was similar to the infant respiratory distress syndrome. The respiratory distress was defined as the presence of tachypnea, hypoxemia, decreased respiratory-system compliance, and bilateral pulmonary infiltrates on chest radiograph. Ashbaugh et al indicated that respiratory support consisted of oxygen therapy via nasal prongs or face mask (in 5 patients) and mechanical ventilation (in 7 patients). The mortality rate was 58%, and pathology examination found that the non-survivors' lungs were heavy, had atelectasis, interstitial and alveolar edema, and hyaline membranes. Since that time, the hallmarks of this syndrome have included:
A risk factor for the development of ARDS (eg, sepsis, trauma, pancreatitis)
Severe hypoxemia with a relatively high FIO2
Decreased pulmonary compliance
Bilateral pulmonary infiltrates
No clinical evidence of cardiogenic pulmonary edema
Although this condition had been known for over a century and there were published data from the Second World War,2,3 it was not until the landmark paper by Ashbaugh et al that interest in ARDS increased. In the subsequent 40 years, very few acronyms have received as much attention in critical care medicine.
Pathophysiology and Histopathology of ARDS
ARDS is caused by an insult to the alveolar-capillary membrane that results in increased permeability and subsequent interstitial and alveolar edema. The mechanisms by which a wide variety of insults can lead to this syndrome are not clear. Acute lung injury (ALI) includes injury to both the pulmonary capillary endothelium and the alveolar epithelium.4 Independent of the clinical disorders associated with ARDS (Table 1), it is useful to think of the pathogenesis of ARDS as a result of 2 different pathways: a direct insult on lung cells, and an indirect insult as a result of an acute systemic inflammatory response. Despite our improved understanding of the role of cellular and humoral components of the inflammatory responses in the lung, we still do not understand the precise sequence of events leading to lung damage.
The typical histopathological features of ARDS are known as diffuse alveolar damage.5 The early phase of ALI (exudative phase) is characterized by leakage of protein-rich edema fluid into the lung and inflammatory cellular alveolar infiltrates. During this phase a cytokine storm and an array of inflammatory mediators are released into the interstitium and alveolar space, perpetuating inflammation and promoting the development of atelectasis and structural damage to the lung architecture. In addition, damage to the alveolar-capillary barrier enhances the difficulty in removing the excess of extravascular lung fluid. Clinically, this initial phase is manifested as marked hypoxemia and reduced pulmonary compliance. Eventually these changes evolve to a fibroproliferative phase in which capillary thrombosis, lung fibrosis, and neovascularization take place. Most ARDS non-survivors die during this phase, despite aggressive ventilatory support with high FIO2 and PEEP.
The Typical ARDS Patient
There is no typical ARDS patient. There are more than 50 recognized conditions associated with the development of this syndrome. The risk of developing ARDS depends on the predisposing clinical condition (ie, some events are more likely to progress to ARDS than others) but it also increases with the number of predisposing factors. Sepsis, bacterial pneumonia, multiple trauma, and aspiration pneumonia are the most common predisposing factors, accounting altogether for more than 70% of cases.4 Overall mortality from ARDS has not decreased substantially since the publication of the 1967 report, and the current survival rate approximates 45% in all major epidemiological series.6,7 Sepsis-related ARDS has a higher overall disease severity, poorer recovery from lung injury, and higher mortality than non-sepsis-related ARDS.8 As part of the therapy for the underlying disease, patients with ARDS invariably require mechanical ventilation to decrease the work of breathing and to improve oxygen transport. An improvement in oxygenation can be obtained in many ARDS patients by an increase in PEEP, a strategy that was originally suggested by Ashbaugh et al.1
Since it is difficult to measure changes in capillary and alveolar permeability at the bedside, diagnosis of ARDS is based on a combination of clinical, oxygenation, hemodynamic, and radiographic criteria. These criteria allow the inclusion of a highly heterogeneous group of critically ill patients, since various types of lung injury can lead to a similar pulmonary response.4 Although there is a general agreement on the overall criteria on which to base a definition of ARDS (ie, severe hypoxemia, marked decreased of pulmonary compliance), the specific values of these variables and conditions of measurement vary greatly among clinicians and scientists. Thus, the original description of ARDS has proved to be incapable of identifying a uniform group of patients. Several of the patients in the original report would not be classified as ARDS today, since fluid overload was an important etiological factor in those patients. Some investigators have questioned whether ARDS is a distinct entity,9 an issue addressed by one of the original authors of the 1967 article.10 Others have suggested that ARDS should not be considered a separate syndrome, but should be seen as part of the multiple system organ dysfunction syndrome.11
From a clinical point of view, one can argue that a strict definition of ARDS may not be required, since current management for ARDS is supportive. However, from a therapeutic point of view, the need for a more precise definition is probably necessary, since the effects on outcome of certain ventilatory and adjunctive techniques could depend on the degree of lung severity.12–14 In terms of prognosis, several investigators have examined whether various oxygenation and/or lung mechanics variables help predict outcome. In any case, there would certainly be some utility in having a standard definition so that data in the literature can easily be compared. From the perspective of ARDS research, there is a very strong argument for having a standard definition: it would help standardize experimental and clinical studies of ARDS natural history, incidence, pathophysiology, treatment, and outcomes. It also would help in the comparison of data among various clinical studies and centers. Thus, a precise definition is clearly important to accurate identification and quantification of various aspects of the underlying pathophysiology and to identify the best therapeutic approach.
A good example of the problems inherent to a definition for ARDS is the wide disparity in the literature on the incidence of ARDS. The most common figure cited for the incidence of ARDS is 75 cases per 100,000 population per year. This is based on an American Lung Program Task Force of the National Heart and Lung Institute in 1972, which was an internal report that suggested that were about 150,000 cases per year of ARDS in the United States, a value similar to the number of all new cases of cancer. In 1988, Webster and colleagues, in England, estimated an incidence of 4.5 cases per 100,000 population,15 and in 1989, Villar et al, in Spain,16 calculated the incidence at 3.5 cases per 100,000. In an attempt to overcome some of these problems, Murray et al17 proposed an expanded definition of ARDS, which takes into account various pathophysiological features of the syndrome. The definition uses a “lung injury score” to characterize the acute pulmonary damage, by considering 4 components: chest radiograph, degree of hypoxemia (PaO2/FIO2), PEEP, and (when available) respiratory-system compliance (Table 2). The lung injury score is obtained by dividing the total score by the number of components that were used. A score of 0 indicates no lung injury, a score of 1–2.5 indicates mild to moderate lung injury, and a score > 2.5 indicates severe lung injury or ARDS. However, the lung injury score is not specific for ARDS and has not been validated, since it is not clear whether patients with identical lung injury scores have similar degrees of lung injury.18 Furthermore, a patient with a major component of cardiogenic edema may be misidentified as having ARDS, and a postoperative patient with moderate atelectasis and mild fluid overload may fulfill the lung injury score criteria of ARDS.
Given that severe hypoxemia is the hallmark of ARDS, it should be crucial to assess the severity of ARDS, for predicting the development and evolution in any given patient, and for assessing the response to treatment. In order to better characterize the severity of lung damage, in 1994 an American-European Consensus Conference on ARDS19 defined ALI and ARDS as follows:
Acute and sudden onset of severe respiratory distress
Bilateral infiltrates on frontal chest radiograph
Absence of left-atrial hypertension (pulmonary capillary wedge pressure < 18 mm Hg or no clinical signs of left-ventricular failure)
Severe hypoxemia (assessed via PaO2/FIO2)
According to these guidelines, ALI exists when PaO2/FIO2 is ≤ 300 mm Hg regardless of the PEEP or FIO2 level, and ARDS exists when PaO2/FIO2 is ≤ 200 mm Hg regardless of PEEP or FIO2 level (Table 3). Although that definition formalized the criteria for ARDS diagnosis and is simple to apply in the clinical setting, it has been challenged in several studies.20–23 For example, all patients included in published papers start off with terrible oxygenation and there is little room for stratifying and separating the patients if there is no further re-evaluation of the hypoxemia. In a secondary analysis of 56 patients who met the American-European Consensus Conference criteria for ARDS, Villar et al20 found that PaO2 response to PEEP allowed the separation of patients into 2 groups with different severities and outcomes. In that cohort, a patient could fit the ARDS criteria when the PaO2 was measured with zero PEEP, but not when measured at a PEEP of 5 or 10 cm H2O. These findings illustrate the major problems in trying to compare the findings from various clinical trials of ventilation strategies,24–29 since patients with very different levels of lung dysfunction and disease may have been enrolled. Furthermore, none of those studies24–29 used the same ARDS definition nor evaluated the same ventilation approach. Diversity in ARDS definitions is apparent in a large number of studies16,20,24–33 (Table 4).
In a retrospective analysis of 74 patients with ARDS who entered into the placebo arm of a pharmacologic study,33 the investigators found that PaO2/FIO2 was identical at baseline in patients who died and in those who survived their ARDS. However, over the subsequent few days there was a separation in the oxygenation status of the survivors and non-survivors. One can argue that the increasing difference in oxygenation in all of the studies as ARDS progressed is not surprising, since one would expect that the sicker patients would develop worse hypoxemia or would certainly not quickly improve their hypoxemia. In 1999, Villar et al20 proposed the need for different guidelines, based on a specific, standard method of evaluating oxygenation status: a proposal that has been advocated by other authors.20,21 In order to determine the impact of various PEEP and FIO2 levels on the classification of patients meeting the American-European Consensus Conference definition for ARDS, Villar et al23 evaluated the impact of standard ventilation settings applied on the day the patients met American-European Consensus Conference ARDS criteria and 24 hours later. They studied 170 patients and found that only 58% of them fulfilled ARDS criteria when evaluated on PEEP of ≥ 10 cm H2O and FIO2 of ≥ 0.5 at 24 hours. The ICU mortality of those patients was 46%. By contrast, 32% of patients were classified as having ALI, and their mortality was 20%. In addition, 10% of patients had a PaO2/FIO2 > 300 mm Hg and were simply categorized as having acute respiratory failure; their ICU mortality was 6%. That study demonstrated the large variability in the severity of lung damage in patients who meet the American-European Consensus Conference definition of ARDS and the strong correlation between oxygenation impairment at 24 hours after ARDS onset and ICU outcome.
Biochemical Diagnosis of ARDS
In patients with acute coronary syndromes, the working diagnosis is based on the presence of acute chest pain accompanied by an abnormal electrocardiogram and the biomarker troponin. Troponin is the biomarker for detection of heart injury and the basis for risk stratification and therapeutic interventions in patients with coronary artery disease. By contrast, there is a lack of a pathognomonic laboratory or clinical feature in ARDS patients. There are no data that link a particular PaO2/FIO2 to predictable structural changes in the alveolar-capillary membrane, most likely because ARDS represents a common pathway of diverse events and disease entities. Also, current guidelines for ARDS management do not follow a strict stratification, as seen in patients with coronary artery diseases. Villar et al34 postulated that stratification of respiratory and ventilatory variables at the onset of ARDS could help identify and select (for clinical trials) patients with different risks of death. They evaluated data from 220 patients included in 2 multicenter trials of ARDS patients ventilated with a lung-protective strategy. Using demographic, pulmonary, and ventilatory data collected at ARDS onset, they derived and validated a simple prediction model based on a stratification of variable values into low, intermediate, and high tertiles. They found that tertile distribution for age, plateau airway pressure, and PaO2/FIO2 at ARDS onset identified subgroups with markedly different mortalities.
Identifying the molecular mechanisms responsible for ARDS is the most important obstacle to the successful diagnosis and treatment of ARDS patients. When comparing the management of acute chest pain to the management of ARDS, the former is based on an emergency medical model of awareness of a life-threatening condition and the importance of adherence to predefined decision algorithms. No comparable awareness and emergency decision algorithms are evident for the care of ARDS patients. It is plausible that a new definition based on specific biochemical criteria of lung inflammation, rather than on clinical parameters, is likely to provide us with a better stratification and identification of a more homogeneous population of patients with ALI and ARDS.34–38 Thus, stratification of ARDS patients should be linked to 2 measures of severity: one that specifically quantifies the severity of ALI/ARDS, and another that quantifies the overall physiologic response along with comorbidities and premorbid health. Adding objectives measures, such as levels of biological markers, could facilitate recognition of ALI/ARDS. The use of simple thresholds for the diagnosis of disease processes of increasing prevalence in the general population is common. This is the case for the use of blood sugar for diabetes and hemoglobin for anemia. It appears improbable in the case of ARDS that a biomarker alone will resolve this issue. Instead, a clinical prediction model34 or a combination of such a prediction model with a biomarker would provide a better definition of ALI/ARDS (Table 5).
An appropriate biomarker for early identification, diagnosis or severity of lung injury should provide information for appropriate stratification of patients at risk for ARDS, at ALI/ARDS onset and during the evolution of the disease process. Ideally, such a biomarker should be:
Easy to measure in blood, exhaled air, or other biological sample
Affected by treatment
There have been recent efforts to identify biological markers in pulmonary edema fluid and in blood from patients with and without ARDS.35,37,39–41 It is postulated that, due to increased permeability of the alveolar-capillary barrier, these proteins leak into the circulation. Donnelly et al35 found that patients at risk for ARDS who had more interleukin-8 in bronchoalveolar lavage fluid subsequently progressed to ARDS. In 180 patients with severe sepsis, Villar et al39 found a direct correlation between lipopolysaccharide-binding protein level and severity of lung injury, which suggests that serial lipopolysaccharide-binding protein may be a clinically useful biomarker for identifying patients at risk for the worst outcomes and with the highest probability of developing sepsis-induced ARDS. Ware et al41 analyzed a combination of 8 biological markers that reflect endothelial and epithelial pulmonary injury, inflammation, and coagulation in 549 patients in the ARDS Network trial of low versus high PEEP, and, although they found that a combination of biomarkers and clinical predictors was superior to clinical predictors or biomarkers alone for predicting mortality or stratifying ALI/ARDS patients, the sensitivity and specificity for ARDS was low. Finally, Determann et al42 measured plasma Clara cell protein (CC16) in 22 patients with pneumonia and 15 controls, and found that plasma CC16 was 3 times higher in the patients with ALI/ARDS than in the patients without ALI/ARDS. A CC16 level of ≥ 18 ng/mL was diagnostic for ALI/ARDS with a sensitivity of 80% and a specificity of 92%. Furthermore, plasma CC16 was elevated 24–48 hours before ALI/ARDS was diagnosed, which suggests that CC16 may predict the development of ALI/ARDS.
The current ARDS definition was established almost 2 decades ago. Based on recent evidence, there is a need for a new definition that better characterizes ARDS. In addition to an appropriate clinical setting and radiographic infiltrates consistent with non-cardiogenic pulmonary edema, we need to measure the oxygenation defect (PaO2/FIO2) under standard ventilator settings (specific FIO2 and PEEP). In addition, it would be very useful to incorporate into the new definition specific biomarkers of lung injury.
- Correspondence: Jesús Villar MD PhD, Multidisciplinary Organ Dysfunction Evaluation Research Network, Hospital Universitario Dr Negrin Barranco de la Ballena, s/n 4th Floor, South Wing, 35010 Las Palmas de Gran Canaria, Canary Islands, Spain. E-mail: .
On behalf of the author, Robert M Kacmarek PhD RRT, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, presented a version of this paper at the 26th New Horizons Symposium, “ARDS Update,” at the 56th International Respiratory Congress of the American Association for Respiratory Care, held December 6–9, 2010, in Las Vegas, Nevada.
This research was partly supported by grants PI10/0393, CB06/06/1088 from Instituto de Salud Carlos III, Spain.
The author has disclosed no conflicts of interest.
This paper is dedicated to the memory of Thomas L Petty and Roger C Bone.
- Copyright © 2011 by Daedalus Enterprises Inc.