What every intensivist should know about acute respiratory distress syndrome and diffuse alveolar damage

Acute respiratory distress syndrome is a challenging entity for the intensivist. The pathological hallmark of the acute phase is diffuse alveolar damage, which is present in approximately half of living patients with acute respiratory distress syndrome. It is clear that respiratory support for acute respiratory distress syndrome has gradually been improving over recent decades. However, it is also evident that these procedures are beneficial, as they reduce lung injury and keep the patient alive. This could be interpreted as a time-gaining strategy until the trigger or causal or risk factor improves, the inflammatory storm decreases and the lung heals. However, all except two pharmacological treatments (neuromuscular blockers and steroids) were unable to improve the acute respiratory distress syndrome outcome. The hypothesis that pharmacological negative results may be explained by the histological heterogeneity of acute respiratory distress syndrome has been supported by the recent demonstration that acute respiratory distress syndrome with diffuse alveolar damage constitutes a specific clinical-pathological entity. Given that diffuse alveolar damage is a pathological diagnosis and that open lung biopsy (the most common technique to obtain lung tissue) has several side effects, it is necessary to develop surrogate biomarkers for diffuse alveolar damage. The aim of this narrative review is to address the following three topics related to acute respiratory distress syndrome: (a) the relationship between acute respiratory distress syndrome and diffuse alveolar damage, (b) how diffuse alveolar damage could be surrogated in the clinical setting and (c) how enrichment in diffuse alveolar damage may improve the results of pharmacological clinical trials tried out on patients with acute respiratory distress syndrome.

Acute respiratory distress syndrome is a challenging entity for the intensivist. The pathological hallmark of the acute phase is diffuse alveolar damage, which is present in approximately half of living patients with acute respiratory distress syndrome. It is clear that respiratory support for acute respiratory distress syndrome has gradually been improving over recent decades. However, it is also evident that these procedures are beneficial, as they reduce lung injury and keep the patient alive. This could be interpreted as a time-gaining strategy until the trigger or causal or risk factor improves, the inflammatory storm decreases and the lung heals. However, all except two pharmacological treatments (neuromuscular blockers and steroids) were unable to improve the acute respiratory distress syndrome outcome. The hypothesis that pharmacological negative results may be explained by the histological heterogeneity of acute respiratory distress syndrome has been supported by the recent demonstration that acute respiratory distress syndrome with diffuse alveolar damage constitutes a specific clinical-pathological entity. Given that diffuse alveolar damage is a pathological diagnosis and that open lung biopsy (the most common technique to obtain lung tissue) has several side effects, it is necessary to develop surrogate biomarkers for diffuse alveolar damage. The aim of this narrative review is to address the following three topics related to acute respiratory distress syndrome: (a) the relationship between acute respiratory distress syndrome and diffuse alveolar damage, (b) how diffuse alveolar damage could be surrogated in the clinical setting and (c) how enrichment in diffuse alveolar damage may improve the results of pharmacological clinical trials tried out on patients with acute respiratory distress syndrome. (e.g., low tidal volume or low pressure plateau) where the most likely main effect is to avoid lung injury associated with mechanical ventilation. Except for early paralyzation and likely steroids, all pharmacological treatments tried on patients with ARDS were unable to demonstrate a relevant effect. (3,4) Diffuse alveolar damage (DAD) is considered the histological hallmark for the acute phase of ARDS. (5) It has been well known for many years that DAD is present in only half of autopsies from patients with ARDS. (6,7) However, the recent demonstration that the same proportion occurs in living patients, (8) as well as the effect that DAD exerts over ARDS outcome, shine a new light on this entity. (9)(10)(11) The aim of this narrative review is to address three topics about ARDS. First, we address the relationship between ARDS and DAD. Second, we analyze how DAD could be surrogated in the clinical setting. Finally, we address how enrichment in DAD may improve the results of clinical trials tried out on ARDS patients.

What is the relationship between acute respiratory distress syndrome and diffuse alveolar damage?
According to the Berlin definition, (5) ARDS is a clinical construct composed by (i) the presence of at least one risk factor associated to (ii) acute hypoxemia not fully explained by cardiac failure or fluid overload and (iii) bilateral infiltration on radiology. On the other hand, the American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias (12,13) defined two histological, indistinguishable patterns: the acute interstitial pneumonia (AIP) and the DAD. The former term, AIP, is reserved for cases of unknown causes, and the latter term, DAD, is for patients with ARDS. In other words, both terms exhibit the same pathological pattern but differ in the clinical context in which they are diagnosed. The aforementioned consensus defined DAD (or AIP) by the presence of key histological features (diffuse distribution, uniform temporal appearance, alveolar septal thickening due to organizing fibrosis, usually diffuse airspace organization may be patchy or diffuse, hyaline membranes) and pertinent negative findings (lack of granulomas, necrosis, or abscesses, lack of infectious agents, no viral inclusions and negative results with special stains for organisms, lack of prominent eosinophils and neutrophils and negative cultures).
Although it is not unanimously accepted, (14,15) the Berlin definition considered DAD as the hallmark for the acute phase of ARDS. (5) This discrepancy may be explained by (i) the fact that a high proportion of the knowledge related to the ARDS pathology has been derived from autopsy studies, (ii) the effect of DAD on the ARDS outcome was unknown and (iii) what occurred in patients with mild ARDS was not described. (3,15,16) In addition, the complexity of diagnosing DAD in patients with ARDS (see below) creates a great challenge for its study. (14) Despite all of these difficulties, recently, several advances have been reported in understanding the relationship between ARDS and DAD. First, it was demonstrated that approximately half of living patients with ARDS present DAD in the pathological analysis of lung tissue obtained with an open lung biopsy. (8) The other half showed one among a number of heterogeneous diseases (Figure 1), some of them with a specific treatment in the case of being diagnosed (e.g., pneumonia, pulmonary embolism or carcinomatous lymphangitis). Second, the effect of DAD on ARDS outcome was demonstrated in post-mortem and living patients. Lorente et al. (11) analyzed 150 autopsies from patients with ARDS and found that the presence of DAD was associated with a lower age, lower ratio of partial oxygen pressure and inspiratory fraction (Pa02/ Fi02), and lower respiratory dynamic compliance, as well as a higher punctuation in the sequential organ failure score (SOFA) scale.
Of paramount importance was the fact that the cause of death was associated with the histological finding (in patients without DAD, refractory shock was the main cause of death in 55% and refractory hypoxemia in 5%; in contrast, in patients with DAD, refractory shock was the main cause of death in 29% and refractory hypoxemia in 25%). Similar differences were found between patients with ARDS and DAD versus ARDS with histological pneumonia. Cardinal-Fernández et al. (8) analyzed 350 living patients with ARDS and open lung biopsy. They found that, although no differences were observed in the severity of the patients with and without DAD (Pa0 2 /Fi0 2 and SOFA punctuation were similar on the day that the ARDS diagnosis and open lung biopsy were performed), mortality in patients with DAD was almost double than in patients without DAD (OR 1.81; IC95% 1.14 -2.86). Kao et al. (10) found that DAD is an independent risk factor for hospital mortality in living patients with ARDS (OR 3.55; IC95% 1.38 -9.12). Finally, given that pneumonia (viral and bacterial) is the second most frequent histological finding in patients with ARDS ( Figure 1), it has been postulated that DAD and histological pneumonia could be considered together, with the aim to increase the correlation between clinical and pathological findings. (17) Although inconclusive, several facts argue against this proposal: (a) from a pathological point of view, DAD and pneumonia constitute two different entities that may exist independently from each other, (b) the microbiological rate of isolation differs in both entities, (18) (c) the clinical evolution and the cause of death are different (11) and (d) the mortality rate is also different. (10) However, all of these differences do not exclude the possibility that some physiopathological pathways may be present in both entities and might explain why some pharmacological treatments may improve both conditions (see below).

How can diffuse alveolar damage be diagnosed?
Based on the previously described evidence, it appears necessary to recognize the subgroup of patients with ARDS and DAD with the aim to define a clinical-pathological entity, (3,11,16,19) to increase the correlation between clinical and histological findings and to develop personalized pharmacological treatments (see below). (20)(21)(22) Currently, the only model to estimate the probability of presenting DAD in patients with ARDS has been developed and validated in autopsies, and its accuracy is just moderate (area under receive operative curve 0.74, IC95% 0.65 -0.82). (11) Likewise, the most frequent procedure to diagnose the DAD is performing an open lung biopsy, which is a risky procedure reserved for centers with demonstrated experience. Open lung biopsy is only recommended in two scenarios: (a) when there is high suspicion of curable etiology, less invasive procedures (e.g., bronchoalveolar lavage, blood samples and CT scan) are inconclusive and the risk of empirical therapy is too high and/or (b) when it is considered necessary to identify the fibro-proliferative phase (towards the end of the first week of evolution) to prescribe steroids. (15,23,24) The problem of diagnosing the gold standard is common to numerous diseases (e.g., myocardial infarction, neurodegenerative diseases and osteoporosis) and can be resolved using surrogate biomarkers. A biomarker is a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes or biological responses to a therapeutic intervention." (25) A surrogate endpoint is "a biomarker that is intended to substitute for a clinical endpoint. A surrogate endpoint is expected to predict clinical benefit (or harm) based on epidemiologic, therapeutic, pathophysiologic, or other scientific evidence." (25)  The most common types of biomarkers are based on measuring clinical parameters or molecules. Imaging techniques have also been successfully used as surrogate biomarkers. In recent years, the combination of structural (e.g., computer tomography or nuclear magnetic resonance) with functional (e.g., positron emission tomography) imaging techniques has determined the appearance of a new kind of biomarker called functional imaging, which allows for the understanding of how physiological (or physiopathological) processes occur in a specific structure of the body.
A surrogate biomarker for DAD should have particular characteristics such as: (a) high accuracy for the diagnosis of DAD as well as ruling out any other diseases that may mimic the ARDS (this statement determines that the discovery and validation of a surrogate biomarker for DAD has to be performed using pathological findings); (b) high precision (the result can not vary if the same sample is analyzed several times using the same technique and the same laboratory conditions); (c) reflect the stage of the DAD evolution; (d) correlate the amount of parenchyma with DAD and (e) with the response of a specific treatment for DAD.
Finally, each kind of biomarker presents specific requirements. For example, if it is a molecule, it should be (a) present in minimally invasive samples (e.g., blood, urine or bronchio-alveolar lavage); (b) simple (e.g., a unique molecule with different levels of cut-off); (c) measurable with laboratory equipment available in average hospitals; (d) able to allow results to be obtained in a brief period of time; and (e) easily interpreted by physicians at the bedside. In addition, if it is a causal factor for DAD, it is more relevant because it could also be considered a therapeutic target. If the biomarker is an imaging technique, it should (a) be able to be performed with minimal displacement of the patient; (b) in a short period of time; (c) allow for the maintenance of all treatment and monitoring and (d) avoid the use of contrast that could harm the patient.
At this moment, N-terminal-peptide type III procollagen (NT-PCP-III) appears to be the most plausible surrogate biomarker for the fibro-proliferative phase of patients with ARDS. Forel et al. (26) conducted an elegant study which included 32 consecutive patients presenting non-resolving, moderate or severe ARDS and open lung biopsy. In the study, they assessed the NT-PCP-III in serum and bronchioalveolar lavage as a surrogate biomarker of the fibro-proliferative phase in patients with ARDS. They found that the NT-PCP-III, measured 3 days (median) before the open lung biopsy, was higher in patients with ARDS with fibro-proliferation than in ARDS without fibro-proliferation (area under ROC was 0.90 [95%CI 0.80 -1.00] for bronchioalveolar lavage and 0.75 [95%CI 0.57 -0.92] for serum).

Why have almost all pharmacological treatments tried on acute respiratory distress syndrome failed?
A drug is usually defined as any chemical substance that affects the functioning of living things and organisms (e.g., bacteria, fungi, or viruses). Likewise, a drug target is "a molecular structure (chemically definable by at least a molecular mass) that will undergo a specific interaction with chemicals that we call drugs because they are administered to treat or diagnose a disease. The interaction has a connection with the clinical effect(s)." (27) On the other hand, clinical trials are a type of experiment designed to answer a specific question related to biomedical or behavioral intervention, including new treatments, protocols or medical devices. Currently, under the term "ARDS" in the international database Clinical Trials, 58 studies (drugs 41, cell therapy 7 and biological therapy 10) appeared, including 8376 patients (Tables 1 and 2). (28) For ARDS, no pharmacological treatment other than early paralyzation and prolonged steroids are routinely used at the bedside. This reality certainly demonstrates that we can identify targets and effective treatments in preclinical studies. However, we are unable to transfer the benefits to "real patients." In this context, we have to keep in mind that "clinical trials are not designed to demonstrate the effectiveness of a treatment in a random sample of the general population," (29) since drugs exert their effect on specific targets, and obviously the target has to be present in the cohort in which the drug is tried on. (21) In other words, you can only lump patients who carry the same target. If not, you have to split them into subgroups of patients that carry the same target. Using this point of view, if only half of the patients with ARDS present DAD, and if most of the targets have been identified in animal models (in which the histology was considered as the gold standard), the high number of failing pharmacological treatments applied to ARDS cannot be a surprise. (3,11,16,19,22) The term enrichment refers to the "prospective use of any patient´s characteristic to select a study population in which the detection of a drug effect (if one is present) is more likely than it would be in an unselected population." (29) Here, the great interest lies in biomarkers, Table 1 -Studies based on pharmacologic treatment attempted in patients with acute respiratory distress syndrome and registered in a clinical trial database (28)    present in minimally invasive samples, such as serum, urine or bronchoalveolar lavage, to surrogate the diagnosis of DAD. As previously mentioned, only early paralyzation (30) and prolonged steroid therapy (31) may be considered effective pharmacological treatment for severe ARDS. We hypothesize that this positive result may be related to the fact that they exert their effect over targets present in several entities that may mimic the ARDS. (32,33) For that reason, it is possible to lump these entities in a clinical trial. Specifically, in the case of early paralyzation, the targets could be (a) the reduction in lung injury arising from ventilator desynchrony, (b) the attenuation of biotrauma and (c) limited expiratory muscle function, which reduces the respiratory system collapse and derecruitment. (30) In addition, a recent experimental study suggests that neuromuscular blockers may inhibit the nicotinic pathway and induce an anti-inflammatory effect. (34) For all of the above reasons, although not definitively, the most plausible mechanism to explain the beneficial effect of early paralyzation on ARDS outcome is the attenuation of mechano-transduction related to lung injury ( Figure 2). This is non-specific to ARDS patients and may also benefit all subjects who require mechanical ventilation. (35,36) On the other hand, the effectiveness of steroids in ARDS may be explained by at least three reasons: (a) the potent down-regulation of inflammatory and fibroproliferative pathways; (b) the benefit of steroids in pneumonia (this is the second most common histological pattern in patients with ARDS); (37) and (c) other specific diseases that may mimic ARDS (e.g., acute eosinophilic pneumonia, diffuse alveolar hemorrhage from vasculitis, cryptogenic organizing pneumonia, acute hypersensitivity pneumonitis and pneumocystis jiroveci pneumonia). (38) CONCLUSION Every intensivist should know that diffuse alveolar damage is present in only half of patients with acute respiratory distress syndrome. Based on recent discoveries, diagnosing diffuse alveolar damage is not merely an academic exercise because its effects on acute respiratory distress syndrome outcome have been demonstrated. At this moment, the only way to diagnose diffuse alveolar damage is to perform an open lung biopsy. However, recently, several efforts have been performed to identify a surrogate biomarker that would allow us to diagnose diffuse alveolar damage without the risk of open lung biopsy. Currently, N-terminal-peptide type III procollagen appears to be an accurate surrogate biomarker for the fibro-proliferative phase of acute respiratory distress syndrome. In coming years, it will be of paramount importance to validate N-terminal-peptide type III procollagen in a large cohort of patients with acute respiratory distress syndrome, as well as to seek out other molecular or imaging biomarkers able to surrogate the diagnosis of diffuse alveolar damage.