Abstract
ABSTRACT Alcohol is the most commonly used and abused drug in the United States. The deleterious health effects of alcohol can be attributed both to its acute intoxicating effects, which result in temporary impairment of judgment and motor skills, and to its more chronic and toxic effects on the liver, pancreas, heart, and brain, all of which may result in irreversible organ damage. Although recognized for more than a century as a major risk factor for pneumonia, alcohol abuse was until recently perceived to have no significant effects on lung structure and/or function. However, within the past decade, epidemiologic studies have revealed that alcohol abuse independently increases the risk of acute respiratory distress syndrome (ARDS) two- to fourfold in patients with sepsis or trauma and may play a role in ARDS pathogenesis in as many as half of all patients with the syndrome. Although alcohol abuse alone does not cause acute lung injury, it renders the lung susceptible to dysfunction in response to the inflammatory stresses of sepsis, trauma, and other clinical conditions recognized to cause ARDS. Recent investigations in both animal models of chronic ethanol ingestion and in human subjects with a history of alcohol abuse have explored this previously unrecognized connection between alcohol and acute lung injury and have uncovered multiple derangements, which we now characterize as the “alcoholic lung.” This review summarizes the epidemiologic association between alcohol abuse and acute lung injury and the recent experimental findings that are unraveling the underlying pathophysiology.
EPIDEMIOLOGIC ASSOCIATION BETWEEN ALCOHOL ABUSE AND ACUTE LUNG INJURY
Brief History of Acute Respiratory Distress Syndrome
Acute respiratory distress syndrome (ARDS) is a severe form of edematous lung injury characterized by noncardiogenic pulmonary edema and flooding of the alveolar airspaces with proteinaceous fluid. ARDS develops in response to inflammatory stresses, including sepsis, trauma, gastric aspiration, pneumonia, and massive blood transfusions.1Originally described in 1967,2ARDS is characterized by alveolar epithelial and endothelial barrier disruption, surfactant dysfunction, and intense inflammation, which, in concert, produce profound derangements in gas exchange and severe respiratory failure. The clinical definition of the syndrome has been modified several times over the past four decades. The most commonly applied definition currently derives from an American-European consensus conference published in 1994,3in which ARDS is defined as bilateral infiltrates seen on a frontal chest radiograph, a pulmonary arterial occlusion pressure of ≤ 18 mm Hg when measured or no evidence of left atrial hypertension (ie, “noncardiogenic edema”), and severe hypoxemia with a PaO2/FIO2 (P/F) ratio of ≤ 200. The P/F ratio is the arterial oxygen tension (PaO2) in mm Hg over the fractional inspired oxygen concentration (FiO2) expressed as a decimal. In parallel, some patients are given the closely related clinical diagnosis of acute lung injury, or ALI. ALI is defined by the same radiographic manifestations and lack of evidence of left atrial hypertension but with less severe derangements in gas exchange (P/F ratio ≤ 300). Thus, the only difference between ARDS and ALI is the severity of hypoxemia. Therefore, a representative patient with ARDS may have a PaO2 of 60 mm Hg on an FIO2 of 0.5 (50%), producing a P/F ratio of 120 (60/0.5), whereas another patient with a similar clinical picture but a PaO2 of 75 mm Hg on an FIO2 of 0.3 (30%) would be designated as having ALI because the P/F ratio was 250 (75/0.3). This distinction is arbitrary, and whether a patient meets the criteria for ARDS or ALI may depend on factors such as mechanical ventilator strategy as much as on the severity of lung injury, and, in reality, the overall prognosis is the same for both groups. These criteria for ARDS and ALI have been widely adopted and used in clinical studies over the past decade, in large part to standardize the definition for multicenter clinical trials and epidemiologic studies. Although a great deal has been learned about the underlying pathophysiology of this syndrome in the past four decades, our treatment remains essentially supportive with broad-based intensive care unit (ICU) resuscitation, the foundation of which is mechanical ventilatory support. Unfortunately, despite aggressive supportive care and mechanical ventilation, the mortality rate for ARDS remains unacceptably high at 40 to 60%.1
Although it has long been recognized that ARDS evolves in response to one of a relatively small group of defined insults (sepsis, severe pneumonia, trauma, pancreatitis, massive gastric aspiration, and massive blood transfusions), it was unknown why a given patient with one of these acute insults developed ARDS, whereas another patient who appeared to be at equal risk did not. In the early 1980s, two studies identified patients assumed to be at risk of the development of ARDS and determined their actual incidence of ARDS to be between 7% and 34%, respectively.4,5In both studies, several diverse diagnoses, such as sepsis, trauma, and aspiration of gastric contents, were prospectively identified that were associated with a higher incidence of ARDS. However, no individual diagnosis was associated with an incidence of ARDS of greater than 40%. Multiple studies over several decades failed to identify independent factors that could explain the variable incidence of ARDS among seemingly comparable patients in terms of the severity of their acute predisposing illness. An important clue to understanding which patients are at greater risk of developing ARDS has been uncovered in recent epidemiologic studies demonstrating a link between alcohol abuse and ALI. These exciting clinical observations have prompted new investigations into the complex pathophysiologic effects of ethanol ingestion on lung structure and function that comprise the focus of this review.
Alcohol Abuse and Susceptibility to Lung Injury
Until recently, there was little evidence that chronic ethanol ingestion had significant effects on the pulmonary parenchyma. The lack of a clinical syndrome such as alcoholic pneumonitis or alcoholic pneumopathy, analogous to other organ-specific disease states caused by excessive ethanol ingestion, suggested that other than an increased susceptibility to pneumonia, the lung was relatively spared from the consequences of chronic alcohol abuse. However, in 1996, a seminal study demonstrated that chronic alcohol abuse independently increased the incidence of ARDS in critically ill patients at risk of developing the syndrome.6This study analyzed the hospital courses of 351 critically ill patients admitted to an urban county hospital with a known risk factor for ARDS, such as sepsis, trauma, gastric aspiration, or multiple blood transfusions. The incidence of ARDS in patients with known alcohol abuse was compared with the incidence of ARDS in those patients without alcohol abuse. Patients with a known history of alcohol abuse had a 43% chance of developing ARDS compared with a 22% chance in patients who did not have a diagnosis of alcohol abuse. Within the largest subgroup of patients, those with sepsis as a risk factor for ARDS, the incidence in the patients with known alcohol abuse was 52% compared with only 20% in the patients without a known history of alcohol abuse. Multivariate analysis determined that a history of alcohol abuse increased the risk of ARDS independently of factors such as the severity of the illness, liver disease, or other factors that can be associated with alcohol abuse. A subsequent multicenter, prospective evaluation of 220 patients with septic shock confirmed an association between alcohol and ARDS.7In that study, the incidence of ARDS was 70% (46/66 patients) in septic patients with a history of alcohol abuse compared with only 31% (47/154) in septic patients without a history of alcohol abuse. After adjusting for a variety of factors, including the source of the infection and the severity of the illness, the relative risk of developing ARDS attributed to alcohol abuse was 3.7 (95% confidence interval 1.83-7.71). Remarkably, in both the original and the follow-up study by Moss and colleagues, half of the patients who developed ARDS had a significant history of alcohol abuse.6,7
Taken together, these two studies demonstrate a previously unrecognized association between alcohol abuse and ARDS.6,7Other investigators have confirmed these findings, as Licker and colleagues independently determined that prior alcohol abuse approximately doubles the risk of ARDS in patients who undergo thoracic surgery for lung cancer.8Although the epidemiologic association between alcohol abuse and ARDS is becoming generally accepted, not all published studies substantiate this relationship. In 2000, Iribarren and colleagues reported their retrospective cohort analysis of 121,012 health plan subscribers and identified by chart review 56 patients with ARDS.9By analyzing the alcohol consumption history provided by the patient during an enrolment examination, these investigators stated that they found no relationship between alcohol intake and ARDS. However, there was no assessment of pattern of alcohol consumption contemporaneous with the episodes of ARDS, meaning that a patient may have been classified as a nonalcoholic (or, conversely, an alcoholic) on entry into the health plan but subsequently developed alcohol abuse (or, conversely, quit drinking). Such misclassifications would obscure the association identified independently by Moss and Licker and their colleagues. In contrast, the two studies by Moss and colleagues prospectively analyzed a combined total of 571 patients who were at risk of ARDS, 195 of whom developed the syndrome.6,7Overall, the two studies by Moss and colleagues,6,7along with the study by Licker and colleagues,8strongly support an association between alcohol abuse and ARDS.
Unfortunately, the morbidity and mortality in alcoholic patients with ARDS are compounded by their increased risk of critical illnesses that predispose them to ARDS, as well as their increased risk of other complications associated with ARDS. For example, alcohol abuse is well recognized as an important risk factor for bacterial pneumonias,10and alcoholic patients with pneumonia are more likely to be infected with either serious gram-negative pathogens, such as Klebsiella pneumoniae,11or to develop bacteremia and shock from even typical pathogens, most notably Streptococcus pneumoniae.12In addition, alcohol abuse is a major risk factor for serious trauma requiring admission to specialized trauma centers.13-15Further, alcohol abuse is widely recognized as a major risk factor for aspiration of gastric contents and severe upper gastrointestinal bleeding from gastritis and esophageal varices that requires massive blood transfusions. Therefore, chronic alcohol abuse increases the risk of developing the acute insults, specifically sepsis, trauma, pneumonia, pancreatitis, aspiration, and massive blood transfusions, that can lead to ARDS. In addition, alcohol abuse may worsen the outcome in patients who develop ARDS, even if they do not succumb to the respiratory failure per se. In the more recent epidemiologic study discussed above,7the effects of chronic alcohol abuse on other organ systems were assessed by comparing the daily aggregate Sequential Organ Failure Assessment score in alcoholic and nonalcoholic patients. Based on this standardized scoring system to grade multiple organ dysfunction in critically ill patients, alcoholic patients with ARDS also had more severe nonpulmonary organ dysfunction when compared with nonalcoholics with ARDS.7In parallel, at least two studies implicate alcohol abuse in the frequency and severity of ventilator-associated pneumonia (VAP) in trauma patients.16,17Therefore, chronic alcohol abuse also increases the risk of multiorgan failure and VAP, which may further amplify the morbidity and mortality associated with ARDS, although mortality rates were not increased in ARDS patients with a history of alcohol abuse.7Taken together, these observations argue that chronic alcohol abuse sequentially amplifies one's risk for ARDS and its complications. Figure 1 depicts this stepwise risk in schematic form. Specifically, alcoholics are at increased risk of serious illnesses that lead to ARDS and then are independently at increased risk of ARDS and subsequent multiorgan failure. For example, an otherwise healthy 30-year-old alcoholic is more likely to have a car accident with serious trauma. Unfortunately, he or she is then more than twice as likely to develop ARDS than a comparably injured nonalcoholic trauma victim. Finally, the alcoholic who develops ARDS is more likely to have multiorgan failure and appears to be at increased risk of VAP as well.
CONSEQUENCES OF ALCOHOL ABUSE
Alcohol Abuse: Definition and Magnitude of the Problem
Alcohol is the most frequently abused drug in the world.18In the United States, 50% of the population regularly consumes alcohol, and in 2002, nearly 18 million American adults met the clinical diagnostic criteria for alcohol abuse or alcohol dependence.19The annual financial impact of alcohol abuse on our society has been estimated at $185 billion.20Recent advances in our understanding of the impact of chronic alcohol ingestion on the lung and other nontraditional targets of alcohol-induced organ injury suggest that these estimates of the impact of alcohol abuse on our society, albeit staggering, may underestimate the true costs of these disorders. Although avoidance or limitation of alcohol consumption provides an obvious strategy to prevent the medical complications of alcohol use, the prevalence of alcohol use and abuse in our country indicates that advances in the understanding and management of the detrimental effects of alcohol intake are indicated.
The most recent Diagnostic and Statistical Manual of Mental Disorders IV definition of substance abuse is “a maladaptive pattern of substance use leading to clinically significant impairment or distress,” whereas substance dependence is essentially substance abuse with additional features, including tolerance and withdrawal phenomenon.21Therefore, abuse is characterized by use of a substance with psychoactive properties either in socially inappropriate ways or in spite of serious consequences, such as disruption of one's personal or professional life. In contrast, dependence implies a state in which sudden withdrawal of the substance produces significant biologic consequences. Although there is considerable overlap between abuse and dependence, the important point in this context is that one may suffer significant biologic consequences of alcohol abuse even if one does not exhibit features of dependence, such as delirium tremens or other manifestations of a withdrawal syndrome. Further, it is important to note that alcohol abuse is not defined by the quantity of ethanol consumed but rather by the harmful consequences of alcohol consumption. Therefore, in epidemiologic studies on alcohol abuse and its health effects, such as the second study by Moss and colleagues,7alcohol abuse is assessed by validated questionnaires that quantitate these harmful consequences. Although beyond the scope of this discussion, an excellent review of the application of questionnaires that identify harmful effects of alcohol abuse was recently published.22
Traditional Targets and Mechanisms of Alcohol-Mediated Toxicity
Although this review focuses on the biologic consequences of alcohol abuse on the lung, these effects are inextricably linked to the larger problem of alcohol effects on other organ systems (Figure 2). The best-known toxic effects of alcohol abuse relate to its effects on the brain and the liver, with cirrhosis representing the paradigm of alcohol-induced tissue damage. In addition to its acute intoxicant effects, alcohol can produce irreversible brain damage, including cortical atrophy and dementia.23Beyond the liver and brain, alcoholic pancreatitis, cardiomyopathy, and peripheral neuropathy constitute other well-recognized organ targets of alcohol-mediated toxicity. Alcohol abuse also contributes to cancer of the upper aerodigestive tract24and breast.25Finally, alcohol abuse impairs host immunity. This is particularly true within the respiratory tract, where alveolar macrophage dysfunction has been implicated in the observed susceptibility to a range of pulmonary infections in alcoholic patients.10-12
Ethanol metabolism occurs primarily, although not exclusively, within the liver through the action of alcohol dehydrogenase and inducible cytochrome P-450 enzymes.26Alcohol dehydrogenase is a cytosolic enzyme with multiple isoforms that vary in binding affinity for ethanol. Only the liver and the gastric mucosa express the high-affinity isoform; therefore, ethanol metabolism by alcohol dehydrogenase in tissues other than the liver and the stomach under normal conditions may be limited.26Ethanol is also metabolized in microsomes via the cytochrome P-450 system, including the CYP1 complex.26This enzyme complex has a lower affinity for ethanol than does the hepatic alcohol dehydrogenase enzyme and therefore may not contribute significantly to ethanol metabolism when alcohol consumption is modest or infrequent. However, in the context of chronic heavy consumption or abuse, when ethanol levels are high, the P-450 enzymes are induced, and a significant percentage of ingested ethanol will be metabolized through this system. The first metabolite of ethanol metabolism, either through alcohol dehydrogenase or cytochrome P-450 complexes, is acetaldehyde. Acetaldehyde causes lipid peroxidation in isolated perfused livers,26the formation of protein adducts, and alterations in mitochondrial electron transport, resulting in enhanced superoxide generation.27Microsomal alcohol metabolism also stimulates the production of oxygen radicals,26further supporting the role of oxidative mechanisms in alcohol-mediated tissue injury. These are but a few of the putative mechanisms by which metabolism of this relatively simple two-carbon alcohol produces tissue damage. Although beyond the scope of this review, the field of alcohol metabolism and our understanding of related tissue effects continue to evolve.
MECHANISMS OF ALCOHOL-INDUCED SUSCEPTIBILITY TO LUNG INJURY
Alcohol-Induced Susceptibility to Lung Injury Can Be Reproduced in an Animal Model
The epidemiologic observations identifying alcohol abuse as an independent risk factor for ARDS prompted studies directed at the mechanisms underlying this association. An animal model of alcohol-mediated susceptibility to lung injury was developed for this purpose, permitting the investigation of the effects of chronic ethanol ingestion, independently of other comorbid factors, on lung function. This model also permitted the control of potential confounding variables, such as malnutrition, the severity of the illness, and liver disease, which could potentially modulate susceptibility to lung injury. In this model, rats were fed ethanol in the drinking water at a concentration of 20% for 3 to 5 weeks.28Lungs isolated from ethanol-fed rats were not inherently edematous during ex vivo perfusion in an isolated organ chamber but developed more edema than lungs isolated from control-fed rats when subjected to endotoxemia prior to lung isolation and perfusion.28Therefore, the lung was intrinsically susceptible to inflammatory injury following chronic ethanol ingestion. Although this study was limited by its employment of an ex vivo organ preparation, this approach eliminated the potentially confounding effects of systemic responses to endotoxin and the possibility that the acute edematous lung injury could be attributed to hemodynamic derangements. In parallel, alveolar type II epithelial cells isolated from rats that ingested ethanol had decreased ability to synthesize and secrete surfactant phospholipid and were more susceptible to oxidant-induced cell death during exposure to hydrogen peroxide.28Subsequent studies using standardized isocaloric liquid diets have confirmed that ethanol ingestion alone, in the absence of malnutrition, smoking, and other health factors commonly associated with alcohol abuse, renders the lung susceptible to acute edematous injury. In particular, chronic ethanol ingestion increased hypoxemic respiratory failure and alveolar protein leak in rats in an experimental model of sepsis in vivo, demonstrating that the aforementioned experimental findings were not an artifact of in vitro or ex vivo model systems.29Taken together, these studies provide strong experimental evidence that chronic ethanol ingestion alone can render the lung susceptible to acute edematous lung injury, consistent with the epidemiologic observations that alcohol abuse predisposes patients to ARDS following an acute inflammatory insult such as sepsis or trauma.
Ethanol Produces Subclinical but Significant Alveolar Epithelial Barrier Dysfunction
The alveolar epithelium under normal conditions is extremely tight and maintains a dynamic air-liquid interface such that the alveolar airspaces remain aerated or relatively “dry” despite their immersion within a dense capillary network that continuously routes the entire blood supply just microns away. This equilibrium depends on many functions, including the production and dispersion of surfactant along the air-liquid interface, tight intercellular junctions that sharply limit the passive transit of interstitial fluid into the alveoli, and active transport of sodium (and therefore water) by the epithelium out of the alveolar space. This dynamic barrier becomes disrupted as part of the pathophysiology of ARDS, and the cardinal features of the syndrome, that is, noncardiogenic pulmonary edema and surfactant dysfunction, reflect widespread alveolar epithelial injury. Therefore, the clinical observations that alcohol abuse renders patients more susceptible to ARDS suggested that chronic ethanol ingestion might alter alveolar epithelial function.
Studies of alveolar epithelial barrier function both in vivo and in isolated alveolar epithelial cells from ethanol-fed rats cultured in vitro reveal that chronic ethanol ingestion alters the physical barrier properties within the alveolus. Specifically, alveolar epithelial cells isolated from ethanol-fed rats and cultured for 6 to 8 days in the absence of ethanol had a persistent defect in epithelial monolayer formation and were four to five times more permeable than comparable monolayers formed by cultured epithelial cells from control-fed rats.30In parallel, alveolar epithelial permeability in vivo was likewise increased approximately fivefold in ethanol-fed rats, as reflected by flux of radiolabeled albumin.30In contrast, alveolar epithelial cells from ethanol-fed rats had increased expression of apical sodium channels when analyzed by patch clamp techniques in vitro,30suggesting that transcellular active sodium transport is up-regulated to counteract the increased paracellular leak in the alcoholic lung. This alteration in the dynamic equilibrium of the alveolar epithelial barrier, that is, increased sodium and water transport in the face of increased permeability, may explain why alcohol abuse alone, in the absence of an acute inflammatory stress, does not cause pulmonary edema. Specifically, although epithelial permeability is increased, a compensatory increase in sodium and water transport maintains a relatively normal air-liquid interface in the alveolar space (Figure 3).
However, the experimental model illustrates that the overall capacity to clear a saline challenge in the airway is diminished in the ethanol-fed rats, even in the absence of an inflammatory challenge, such as endotoxemia.30These findings indicate that the alcoholic alveolar epithelium, although apparently “normal” in the otherwise healthy animal or human, is actually dysfunctional, even though it may be able to maintain a relatively normal air-liquid interface through enhanced sodium and water clearance. Unfortunately, these compensatory mechanisms become overwhelmed when the lung is confronted with an inflammatory challenge, resulting in greater susceptibility to disruption and subsequent flooding of the alveolar space with proteinaceous fluid.
Although there are undoubtedly multiple mechanisms by which chronic ethanol ingestion alters the permeability of the alveolar epithelium, one important candidate mechanism has recently been identified. Experimental evidence suggests a role for transforming growth factor β1 (TGF-β1) as a mediator of ALI.31TGF-β1 is a pluripotent cytokine with multiple potential effects on tissue injury and repair during lung injury, including disruption of epithelial integrity in experimental models.31,32In the chronic ethanol-fed rat model, latent or inactive TGF-β1 protein expression was increased twofold in lung tissue.33There was no evidence of release and/or activation of TGF-β1 into the alveolar airspace during ethanol ingestion alone,33again consistent with the experimental and clinical observations that in the absence of an acute stress, the alcoholic lung is not edematous. However, when compared with rats fed control diets, ethanol-fed rats released approximately five times more activated TGF-β1 into the airspace during endotoxemia.33Further, bronchoalveolar lavage fluid from endotoxemic ethanol-fed rats was capable of inducing a significant permeability defect in intact alveolar epithelial monolayers derived from control-fed rats, and this permeability defect was completely inhibited by neutralizing antibodies to TGF-β1.33Taken together, these experimental findings indicate that chronic ethanol ingestion increases the expression of TGF-β1 in the lung, which can then be released and activated in the alveolar space during acute inflammatory stress, where it can directly cause epithelial permeability.
Role of Glutathione in Alcohol-Induced Susceptibility to Lung Injury
A fundamental aspect of the “alcoholic lung” in both experimental models and clinical studies is evidence of chronic oxidative stress and depletion of the critical antioxidant glutathione within the alveolar airspace. Parallel to the aforementioned effects on alveolar epithelial function, ethanol ingestion significantly decreases levels of glutathione in the type II alveolar epithelial cells by more than 90% and in the alveolar epithelial lung fluid of rats that ingested ethanol by approximately 80%.28This novel finding of ethanol-induced glutathione depletion in the lung was consistent with evidence in models of ethanol-induced hepatic dysfunction that glutathione depletion played an important role in ethanol-mediated tissue dysfunction. The importance of ethanol-mediated glutathione depletion in altered lung cell function was substantiated by the demonstration that isolated lungs from rats whose ethanol-water mixture was supplemented with glutathione precursors were less susceptible to endotoxin-mediated edema formation.28Multiple subsequent studies have demonstrated that glutathione supplementation of the experimental diet prevents ethanol-mediated defects in alveolar epithelial function.29,30,34-37These experimental observations were made even more exciting by the recognition that otherwise healthy subjects with a history of alcohol abuse have dramatically decreased levels of glutathione in their lung lavage fluid when compared with nonalcoholic control subjects.38Otherwise healthy subjects with a history of alcohol abuse but normal nutritional indices and no clinical evidence of lung or liver disease had dramatically decreased levels of glutathione in their lung lavage fluid when compared with nonalcoholic control subjects.38In that study, as in the animal model, the glutathione concentrations were corrected for dilution during the lavage procedure and therefore represent the actual concentrations within the alveolar epithelial lining fluid in vivo. Further, the relative degree of glutathione deficiency in the epithelial lining fluid in human subjects with chronic alcohol abuse compared with nonalcoholic subjects was virtually identical to that observed in the experimental model28and in some cases approximated only 10% of the glutathione levels in control subjects. This previously unrecognized deficiency in a critical antioxidant in the lungs of alcoholics not only parallels and validates the animal model but also reflects a significant vulnerability to oxidative stress in the lungs of these individuals that is consistent with their increased susceptibility to ARDS in response to acute oxidative stresses such as sepsis or trauma.
Further investigation provided additional insights into ethanol-mediated glutathione deficiency and consequent susceptibility to lung injury. Although earlier studies suggested that the lung does not metabolize significant amounts of ethanol through alcohol dehydrogenase, metabolism through the cytochrome P-450 system in the lung is significant.39Local ethanol metabolism within the lung may be sufficient to exert significant oxidative stress because ethanol-fed animals and alcoholic humans have increased levels of oxidized glutathione, not simply glutathione depletion, in their lung lavage fluid.28,38These findings suggest that chronic oxidative stress leads to glutathione consumption and depletion, which then leaves the alveolar space relatively unprotected and vulnerable to acute oxidative stresses, such as sepsis or trauma. Within cells, there are distinct mitochondrial and cytosolic pools of glutathione. Although chronic ethanol ingestion depletes both the mitochondrial and cytosolic glutathione pools, it appears that the mitochondrial levels may be more critically involved in the pathophysiology of ethanol-mediated alveolar epithelial dysfunction.30,34,36,37Specifically, supplementing the diets of ethanol-fed rats with N-acetylcysteine (NAC), the only clinically available glutathione precursor, prevented alcohol-mediated depletion of the alveolar epithelial cell cytosolic glutathione pool. In contrast, NAC failed to prevent alcohol-mediated depletion of the mitochondrial glutathione pool in alveolar epithelial cells and failed to prevent alcohol-mediated derangements in surfactant synthesis and enhanced susceptibility to oxidant-induced apoptosis.30,34,36,37However, supplementing the diets of ethanol-fed rats with procysteine, a glutathione precursor that prevented depletion of both mitochondrial and cytosolic glutathione pools, abrogated alcohol-induced derangements in alveolar epithelial cell function in vitro.30,34,36,37Importantly, procysteine supplementation also eliminated ethanol-mediated susceptibility to surfactant dysfunction and hypoxemia in an experimental model of sepsis-induced lung injury in vivo.29Although it is unknown why procysteine restores and/or maintains both the cytosolic and mitochondrial glutathione pools during chronic ethanol ingestion, these and related studies nonetheless implicate mitochondrial glutathione depletion as a fundamental feature of ethanol-induced lung dysfunction. These findings parallel multiple studies implicating mitochondrial glutathione depletion in ethanol-induced liver injury40-42and strengthen the evidence that mitochondrial dysfunction is a common mechanism by which alcohol abuse leads to tissue injury.
Potential Therapies for Alcohol-Associated Lung Injury
Although there is now strong clinical evidence that chronic alcohol abuse significantly increases the risk of ARDS and consequent ICU-related illnesses, such as VAP, currently, there are no specific therapies that have been tested in this vulnerable population. Although numerous pharmacologic interventions for the treatment of ARDS have been investigated, none have proven effective in limiting the morbidity and/or mortality from the syndrome. The challenge in identifying effective therapies in patients with alcohol abuse who are even more vulnerable to developing ARDS would appear to be daunting. However, studies investigating the mechanisms underlying this association are providing new insights into the pathophysiology of ARDS that may ultimately lead to the design of novel therapies.
One potential therapy suggested by the studies discussed above is glutathione replacement therapy. NAC is a glutathione precursor that is commonly used as an antidote for acetaminophen overdose. In addition to augmenting glutathione synthesis, NAC has oxygen radical scavenging properties,43and NAC therapy has been examined in several clinical studies of ARDS patients with modest but encouraging results. In one trial of 66 patients with ARDS, NAC therapy was associated with a small increase in lung compliance compared with the control group but did not affect the survival rate.44In another trial of 61 patients, NAC therapy was associated with improved oxygenation and a decreased need for ventilatory support in patients with mild to moderate ALI but did not improve the survival rate.45Finally, a study in 30 ARDS patients showed that NAC therapy was associated with improvements in pulmonary hemodynamics and gas exchange.46However, even though clinical trials in ARDS patients have shown only modest efficacy from NAC therapy, patients with long-standing alcohol abuse may have chronic glutathione deficiency, including depleted mitochondrial pools, which cannot be overcome with NAC alone. Therefore, although there is circumstantial evidence that lung glutathione deficiency may be important in the pathogenesis and/or severity of selected patients with ARDS, it is possible that glutathione deficiency is a critical variable only in specific risk groups, such as alcoholics, and that early treatment is necessary to improve alveolar epithelial function prior to the development of acute alveolar damage. However, studies in which glutathione replacement is combined with other drugs and targeted early in the course of critical illnesses of patients with alcohol abuse could potentially improve their outcomes.
Another potential therapy is granulocyte-macrophage colony-stimulating factor (GM-CSF). GM-CSF is a 23 kDa glycosylated monomeric peptide that is secreted by multiple cell types, including the alveolar epithelial type II cell.47It was first identified in mouse lung cell-conditioned medium and was named for its ability to stimulate the growth of granulocytes and macrophages from cultured hematopoietic progenitor cells. The cloning of this protein permitted a variety of in vitro and in vivo studies to characterize its functions, and, subsequently, it was found to stimulate the production of eosinophils, erythrocytes, megakaryocytes, and dendritic cells in addition to granulocytes and macrophages. GM-CSF has been widely used clinically to improve bone marrow recovery following chemotherapy. However, targeted deletion of the GM-CSF gene in mice surprisingly had no effect on the hematopoietic system but rather produced an unexpected lung-specific phenotype that was essentially identical to pulmonary alveolar proteinosis (PAP).48PAP is characterized by alveolar macrophage immune dysfunction and impaired surfactant phospholipid recycling, leading to opportunistic infections and accumulation of surfactant phospholipid and protein (“proteinosis”) in the alveolar airspace. Although the alcoholic lung is not as severely affected as the lung in PAP, the functional defects in the alveolar macrophage in these two conditions are similar. Consistent with this, a recent study showed that recombinant GM-CSF delivered via the upper airway restored alveolar epithelial barrier function and fluid transport in ethanol-fed rats, even during endotoxemia.49Importantly, although that study showed that GM-CSF treatment decreased endotoxin-mediated lung injury even in control-fed rats, the magnitude of the efficacious response was clearly greater in the ethanol-fed rats. Interestingly, an earlier phase II clinical trial of 18 patients with septic shock demonstrated that patients who received recombinant GM-CSF treatment (n = 10) appeared to have less severe lung injury than placebo-treated patients (n = 8).50Further, alveolar macrophages from septic patients given recombinant GM-CSF treatment had improved function in vitro, including respiratory burst, when compared with macrophages from placebo-treated septic patients.50Although the investigators in that study did not evaluate patients for alcohol abuse, they nevertheless demonstrated that GM-CSF treatment might limit the incidence and/or severity of ARDS in patients with septic shock, and the aforementioned experimental study suggests that alcoholic patients might derive the greatest benefit from this novel therapy.
Finally, angiotensin II blockade represents another potential therapeutic target in critically ill patients with alcohol abuse. Angiotensin II is a pluripotent vasoactive peptide that is increased in patients with ARDS.51Angiotensin II is formed by the renin-angiotensin system through the sequential conversion of angiotensinogen to angiotensin I and then to angiotensin II, the latter conversion primarily by the angiotensin-converting enzyme. Chronic ethanol ingestion increases plasma levels of angiotensin II in rats,52and it has been postulated that activation of the renin-angiotensin system may explain the association between alcohol abuse and hypertension in humans.52,53Although a mechanism is not known, it has been shown that acetaldehyde, the primary metabolite of ethanol, can convert angiotensinogen to angiotensin I in rat plasma in vitro.54The biologic effects of angiotensin II depend on its interaction with specific angiotensin II receptors, and at least seven subtypes have been identified. Of the angiotensin II receptors, the type 1 receptor (AT1) and the type 2 receptor (AT2) have been best characterized. The majority of the well-known effects of angiotensin II, such as vasoconstriction, sodium retention, and tissue hypertrophy and hyperplasia, are mediated via the AT1 receptor.55,56By contrast, the AT2 receptor is present in few tissues during adulthood, whereas it is abundantly expressed during embryogenesis and in response to injury.55,57Much less is known about the role of the AT2 receptor in the postembryonic state, although it is expressed in some tissues in the adult, particularly the ovaries, the adrenal medulla, and neurons. Stimulation of the AT2 receptor inhibits cell proliferation and leads to apoptosis, actions that are directly opposed to the proliferative responses that often follow AT1 activation.57-60The net result of angiotensin II stimulation in a given context depends on the relative expression of these two functionally opposing receptor subtypes. Experimentally, chronic ethanol ingestion markedly increases the relative expression of the AT2 receptor within the alveolar epithelium and in parallel renders these cells susceptible to apoptosis when exposed to oxidative stress or proinflammatory cytokines.61Importantly, selective inhibition of the AT2 receptor completely inhibits angiotensin II- and tumor necrosis factor α-induced apoptosis in alveolar epithelial cells isolated from ethanol-fed rats.61Therefore, chronic ethanol ingestion shifts the angiotensin II receptor phenotype within the alveolar epithelium to predominantly AT2 receptor subtype expression that mediates epithelial cell apoptosis in response to inflammatory stimuli. Although selective AT2 receptor blockers have not yet been tested clinically (as opposed to AT1 receptor blockers that are in widespread use to treat a variety of cardiovascular diseases), these experimental findings suggest that selective AT2 receptor blockade could potentially prevent or at least limit alveolar epithelial cell death in the alcoholic lung during acute inflammatory stresses.
Figure 4 illustrates some of the complex pathophysiologic mechanisms by which alcohol abuse renders patients susceptible to ARDS, as well as potential therapeutic strategies that are being elucidated by recent studies. Although, to date, no therapeutic trials have been directed at the subset of patients at risk of ARDS who have underlying alcohol abuse, the experimental evidence cited above raises the possibility that such trials could be forthcoming. Importantly, because the incidence of ARDS is so high in alcoholic patients with sepsis (˜70%), a study targeted at this vulnerable population would require fewer subjects than one that included all patients with sepsis in whom the overall incidence of ARDS is only ˜30%. Further, it may well be that the therapeutic “target” would be different in these patients. Even if a therapeutic intervention decreased the incidence of ARDS in alcoholic patients to 40 to 50%, this would represent an enormous improvement in outcome for this group and could save many thousands of lives each year in the United States alone. Ironically, although patients with underlying alcohol abuse are more susceptible to ARDS, they may respond to therapies that are focused on ethanol-induced defects in lung structure and function that might be ineffective in the nonalcoholic patients. Given the enormous magnitude of the problem, including the staggering morbidity and mortality caused by alcohol-mediated lung injury, it may be appropriate to focus clinical trials of currently available agents, such as glutathione replacement, GM-CSF, or angiotensin II receptor blockers, on this high-risk population.
SUMMARY
Alcohol abuse independently and significantly increases the risk of ARDS and associated ICU-related illnesses. In fact, it is now recognized that alcohol abuse is a factor in approximately half of all cases of ARDS. Unfortunately, this means that ARDS rivals cirrhosis in terms of alcohol-related deaths in the United States every year. Even more tragically, the average age of patients with alcohol-related ARDS tends to be younger than patients with alcohol-related cirrhosis, thereby magnifying the impact. Recent and ongoing research in both experimental models of chronic ethanol ingestion and in patients with alcohol abuse is providing new insights into the pathophysiology of this recently recognized association. Although the biologic consequences of ethanol metabolism on lung structure and function are proving to be more varied and complex than had been previously appreciated, discreet pathways by which chronic ethanol ingestion renders the lung susceptible to acute inflammatory injury are being elucidated. Therefore, there is reason for optimism in this otherwise tragic context because experimental models are beginning to demonstrate that the phenotype of the alcoholic lung may be amenable to targeted therapeutic interventions.