The rise and fall of β-agonists in the treatment of ARDS

The acute respiratory distress syndrome (ARDS) is a severe inflammatory condition of the lung, which can be triggered by a number of different pulmonary and extra-pulmonary insults [1]. The characteristic pathological changes of ARDS include an exudative phase, with the accumulation of fluid within the lung, the release of pro-inflammatory cytokines and infiltration of inflammatory cells, especially neutrophils, into the lung parenchyma. Damage to the alveolar epithelium and pulmonary capillary endothelium occur and patients develop the characteristic histological appearance of diffuse alveolar damage [1]. This manifests clinically as non-cardiogenic pulmonary edema, which reduces lung compliance and impairs gas exchange.


Introduction
Th e acute respiratory distress syndrome (ARDS) is a severe infl ammatory condition of the lung, which can be triggered by a number of diff erent pulmonary and extrapulmonary insults [1]. Th e characteristic pathological changes of ARDS include an exudative phase, with the accumulation of fl uid within the lung, the release of proinfl ammatory cytokines and infi ltration of infl ammatory cells, especially neutrophils, into the lung parenchyma. Damage to the alveolar epithelium and pulmonary capillary endothelium occur and patients develop the characteristic histological appearance of diff use alveolar damage [1]. Th is manifests clinically as non-cardiogenic pulmonary edema, which reduces lung compliance and impairs gas exchange.
Pharmacological interventions to date have had limited success in improving outcomes [2]. Improvements to supportive care (protective ventilation [3] and conservative fl uid management [4]) are thought to have contributed to the improved outcomes observed in recent years [5]. β-adrenoceptor agonists (β-agonists) are well established in the treatment of airfl ow obstruction. In addition to actions as bronchodilators, they have anti-infl ammatory properties, promote the clearance of alveolar fl uid, and promote epithelial and endothelial repair [6]. Th e scientifi c rationale for a potential role in the treatment of ARDS is summarized in Figure 1. Th e clinical eff ectiveness of β-agonists has been the subject of clinical trials spanning the last 25 years. Despite early studies showing promise, two large scale randomized controlled trials have recently been terminated on the basis of futility and concerns about safety. In this review, we will outline the pre-clinical evidence for β-agonists and discuss the results of recent clinical trials.

The β-adrenoceptor in the lung
Th e β-adrenoceptor is a transmembrane G-proteincoupled receptor linked to adenylate cyclase (AC). Activa tion of the β-adrenoceptor stimulates an increase in the production of cAMP from adenosine triphosphate by this enzyme [7]. Th ere are three distinct β-adreno ceptor subtypes: β1, β2 and β3, with diff erent distribu tions, eff ects and genetics. β1 receptors are primarily present within the heart, and β3 receptors present principally in adipocytes, but also found on lung endothelial cells. β2 receptors are the most important pulmonary adrenoceptor subtype, present in increasing numbers with each generation of airway branching; greatest amounts are, therefore, present in the distal airways and alveoli where they are expressed on the surface of alveolar type I and type II cells [8].

β-agonists improve alveolar fl uid clearance
Th e presence of non-cardiogenic pulmonary edema is central to the pathophysiology and outcome of ARDS [9]. Th e most well studied mechanism for the clearance of alveolar fl uid is the active transport of ions across the alveolar epithelium, creating an osmotic gradient for the subsequent movement of fl uid. Th ere is good evidence that transported sodium ions are the main driver for this process, entering the alveolar cell through amiloridesensitive Na + channel (ENaC) or other cationic channels on the apical alveolar cell surface, and actively transported out by Na + -K + -ATPase on the basal surface [10]. Th e role of chloride ions is less well characterized; although they must follow sodium ions to maintain electro-neutrality, the pathway through which they move The rise and fall of β-agonists in the treatment of ARDS is as yet unidentifi ed. Until recently, alveolar type II cells were thought to be responsible for the majority of ion transport. Sodium and chloride channels have recently been found on the more numerous alveolar type I cells, which may indicate a signifi cant functional role. A contribution to this process may also be made by the distal airway epithelium.
β-agonists up-regulate the transport of both sodium and chloride ions through the increase in intracellular cAMP caused by β-adrenoceptor stimulation. A number of mechanisms have been proposed by which ion transport is increased by raised cAMP levels, including a greater sodium channel open probability, changes in the phosphorylation of the Na + -K + -ATPase α-subunit, greater delivery of ENaC and Na + -K + -ATPase, and increased chloride transport by the cystic fi brosis transmembrane conductance regulator [6,10].
A higher rate of alveolar fl uid clearance following βagonist administration has been demonstrated in a number of experimental animal models [6], as well as the ex vivo human lung [11]. Additionally, over-expression of the epithelial β-adrenoceptor induced a higher rate of lung edema clearance in a rat lung injury model, increasing sensitivity to endogenous catecholamines [12].

The anti-infl ammatory eff ects of β-agonists
Th ere is a well recognized interaction between the sympathetic nervous system and the immune response, and β-adrenoceptors are present on many of the cells that have relevance in the development of ARDS, including macrophages, T-lymphocytes and neutrophils. ARDS is usually associated with an intense alveolar neutrophilia. Neutrophils migrate into the lungs, within hours of the causative insult, and their numbers correlate with severity of disease, and with mortality when persistently high numbers are seen [13]. β-agonists decrease neutrophil-related infl ammation, and this may therefore confer a therapeutic benefi t [6,14]. β-agonists reduce neutrophil adhesion to bronchial epithelial and endothelial cells [15], which may contribute to the reduced neutrophil accumulation within the alveolar space. βagonists also decrease neutrophil production of cytotoxic reactive oxygen species (ROS) in vitro. In a human model of acute lung injury (ALI), pre-treatment of health volunteers with inhaled salmeterol prior to lipopolysaccharide (LPS) inhalation attenuated neutrophil infi ltration into the lung, myeloperoxidase (MPO) release and tumor necrosis factor (TNF)-α production [16].
Pro-infl ammatory stimulation in the pathogenesis of ARDS is mediated within individual cells through the transcription factor, nuclear factor-κB (NF-κB). Th is cytosolic protein relocates to the nucleus under the infl uence of pro-infl ammatory cytokines, where it binds to specifi c regions of DNA and in so doing increases the transcription of various infl ammatory gene products. To allow movement into the nucleus, the endogenous inhibitor of this protein; inhibitor-κB (I-κB), must dissociate from the protein. During tissue infl ammation cytosolic levels of I-κB decrease and increased activity of NF-κB is seen. Th is decrease in I-κB levels is reversed by the action of β-agonists. In the setting of LPS stimulation of human mononuclear cells, β-agonists act in a protein kinase A and cAMP dependent manner to increase cytosolic I-κB and thereby inhibit the pro-infl ammatory action of NF-κB [17].
Th e interaction between β-agonists and the infl ammatory response is, however, not straightforward. Although in the presence of infl ammation, β-agonists have antiinfl ammatory properties, in the absence of pro-infl ammatory stimulation they themselves can be pro-infl ammatory. In vitro, unstimulated macrophages exposed to β-agonists, increase their production of pro-infl ammatory cytokines, such as interleukin (IL)-1β and IL-6 [18]. Increases in infl ammatory mediator concentrations are also seen in β-agonist stimulation of skeletal muscle, fi broblasts and adipocytes, through β-adrenoceptor and non-β-adrenoceptor pathways.

β-agonist eff ects on alveolar epithelial and endothelial repair
Th e alveolar capillary barrier is comprised of the capillary endothelium, the interstitial space including the basement membrane and the extracellular matrix, and the alveolar epithelium. During ALI there is signifi cant damage to all three structures. Th e clinical importance of this is highlighted by the fi ndings that markers of endothelial damage (von Willebrand factor [vWF]) and epithelial cell damage (KL-6 or receptor for advanced glycation end products [RAGE]) are elevated in those who die from ARDS. Effi cient alveolar epithelial repair is therefore important for ARDS patients' recovery. Several diff erent lines of research suggest that βagonists may reduce endothelial damage or enhance repair in models of lung injury. Evidence for endothelial protection comes from the fi ndings that in experimental acid-induced lung injury, β-agonists and cAMP donors signifi cantly reduced lung endothelial cell permeability. In vitro thrombin-induced pulmonary endothelial permeability is reduced by β-agonists by directly maintaining actin fi laments and the shape of endothelial cells [19]. Ischemia-reperfusion injury of the pulmonary endothelium is attenuated by agents that increase intracellular cAMP such as β-agonists, a feature that is of specifi c interest in the prevention of damage to the lung in nonheart beating organ donation, but is also relevant to all conditions that involve injury to the alveolar-capillary membrane [20].
Evidence for a role of β-agonists in epithelial repair comes from in vitro studies showing that β agonists stimulate the closure of mechanically induced wounds of epithelial monolayers by increasing cAMP and activating protein kinase A (PKA). Incubating bronchoalveolar lavage (BAL) fl uid from patients with ARDS who had received treatment with intravenous salbutamol enhanced in vitro epithelial wound repair [21].

Early studies
Th e key elements of each trial are summarized in Table 1. Th e fi rst clinical trial of β-agonists took place over 25 years ago. Basran et al. conducted an open label trial which examined the eff ect of a 7 μg/kg infusion of terbuta line on lung vascular permeability in 10 ventilated patients with ARDS [22]. Lung vascular permeability was measured by recording the pulmonary accumulation of radio-labelled transferrin. Th ere was no overall change in lung vascular permeability; however, the fi ve patients who demonstrated a reduction in plasma protein accumu lation index survived as opposed to the fi ve patients who showed no improvement (or worsening) of plasma protein accumulation index.
Subsequent trials focused on the eff ects of β-agonist therapy on pulmonary mechanics. Th e increase in cAMP levels caused by β-adrenoceptor stimulation promotes bronchial smooth muscle relaxation which results in bronchodilation, reducing airways resistance and the pressures required for adequate mechanical ventilation. Pesenti et al. studied the eff ect of an intravenous infusion of salbutamol in seven mechanically ventilated, paralyzed patients with ARDS [23]. Inspiratory resistance (maximal [total] and minimum [ohmic airway resistance]) and respiratory system compliance were measured at diff erent levels of positive end-expiratory pressure (PEEP) before and 30 minutes after a continuous intravenous infu sion of 15 μg/min salbutamol. Treatment with salbutamol reduced maximum and minimum inspiratory resistances (from 6.48 ± 2.56 to 4.67 ± 1.74 and from 4.06 ± 2.12 to 2.07 ± 0.95 cmH 2 O/l/sec, respectively) but had no eff ect on eff ective additional resistance or respiratory system compliance.
Wright et al. conducted the fi rst randomized controlled crossover trial of β-agonists in the treatment of ARDS [24]. In this study, eight patients with ARDS were random ized to receive 1 ml of 0.5% metaproterenol solution mixed with 3 ml of normal saline solution, or 4 ml solution of normal saline solution (placebo). Six hours later, treatment arms were crossed over and the opposite regime administered to the same patient. Lung mechanics, shunt fraction, dead space and oxygenation were measured at baseline, 30, 60 and 120 minutes. Aerosolized metaproterenol promptly reduced peak and plateau airway pressure and airway resistance whereas dynamic compliance increased. Th e eff ects persisted over the 2 hours of the study. Total compliance also increased as did oxygenation but the changes did not reach statistical signifi cance. Th ere was no eff ect on minute ventilation, pulmonary shunt fraction or deadspace.
In a subsequent cohort study, the eff ect of 1 mg nebulized salbutamol was examined in 11 patients with ARDS [25]. Compared to baseline, nebulized salbutamol was associated with modest reduction in peak and plateau airway pressures (4.9 ± 0.8 cmH 2 O and 2.1 ± 0.6 cmH 2 O, respectively), intrinsic PEEP (1.9 ± 0.5 cmH 2 O) and minimal respiratory resistance (1.9 ± 0.3 cmH 2 O/l/s). Additional resistance, static compliance, oxygenation, heart rate and blood pressure did not change.

Recent studies
A retrospective chart review of 86 adult patients with ALI found that patients with ALI who also received high dose nebulized salbutamol (2.5-6.4 mg/day) had significantly more days alive and free of ALI (n = 22, 12.2 [4.4] days) compared with the group receiving ≤ 2.4 mg/day (n = 64, 7.6 [1.9] days). Th ere were no diff erences in nonpulmonary organ failure or hospital mortality rates (48% vs 50%) [26]. After adjustment for diff erences in case mix between the groups, high dose salbutamol remained independently associated with the number of days alive and free of ALI in a multivariate model.
Th e β-agonist Lung Injury Trial (BALTI-1) [27] was a phase II prospective randomized, double blind, placebocontrolled and the fi rst study in humans to evaluate the eff ect of β-agonists on lung water. Th is single center study randomized 40 adult patients with ARDS to an intravenous infusion of salbutamol 15 μg/kg/hr for 7 days and serially recorded the eff ect on extravascular lung water (EVLW). Th e study demonstrated that salbutamol significantly reduced lung water at day 7 (mean [SD] 9.2 [6] vs. 13.2 [3] ml/kg, p = 0.038), alveolar capillary permeability

Table 1. Summary of clinical trials of β-agonists in patients with acute respiratory distress syndrome (ARDS)
First author Basran [22] Pesenti [23] Wright [24] Morina [25] Manocha [26] Perkins [27] Licker [28] Matthay [31] Smith [33] Publication  [21] and plateau airway pressures compared to placebo and showed a trend towards reduced lung injury score (LIS). Th ere were no diff erences in alveolar neutrophil sequestration or infl ammatory cytokines [14]. Th e study protocol included a dose titration algorithm in the event of new onset tachycardia or arrhythmia. Nineteen patients received intravenous salbutamol for a total of 2,148 hours. Five patients in the salbutamol arm developed new onset of supraventricular arrhythmia compared to two patients in the placebo group (p = 0.2). No patients sustained serious ventricular arrhythmias. Heart rates were higher in the salbutamol group (average diff erence at day 1 of 11 beats per min [bpm]) although this was not statistically signifi cant. Th ere were no substantial diff erences in electrolyte or lactate concentrations between salbutamol and placebo arms. Licker et al. examined the use of nebulized salbutamol compared to ipratropium bromide in a randomized controlled crossover trial in 21 patients following lung resection who were at risk of, but had not developed ALI [28]. Th ese authors found a signifi cant reduction in EVLW, pulmonary vascular permeability index and an improvement in PaO 2 /FiO 2 ratio and increased cardiac output on the fi rst postoperative day in the salbutamol arm. Th e authors concluded that aerosolized salbutamol accelerates the resolution of lung edema, improves blood oxygenation, and stimulates cardiovascular function after lung resection in high-risk patients. Although EVLW has been shown to be an independent predictor of outcome in ARDS [29], its use as a trial outcome has been criticized as decreases in lung perfusion because of worsening disease and non-linear thermal wash-out present in hetero geneously diseased lung may be erroneously interpreted as an improvement in lung edema [30].
Th e AlbuteroL for the Treatment of ALI (ALTA) study was a multicenter, randomized controlled trial of nebulized salbutamol in patients with ALI run by ARDSnet investigators [31]. Th is double blind, placebo controlled trial allocated patients to receive either salbutamol 5 mg every four hours or saline placebo for up to 10 days. Th e primary outcome for the trial was ventilator-free days. Th e trial started in August 2007 and set out to recruit up to 1000 patients with ALI. Th e trial was terminated by the Data Monitoring and Safety Board on the grounds of futility after 282 patients had been enrolled. Th ere were no signifi cant diff erences in the number of ventilator-free days between salbutamol and placebo arms (14.4 vs 16.6 days, 95% confi dence interval [CI] diff erence -4.7 to 0.3 days, p = 0.087). Th ere was no diff erence in hospital mortality rates (23.0 and 17.7%; 95% CI diff erence -4.0 to 14.7%, p = 0.30). Th ere were no diff erences in plateau airway pressure, minute ventilation or oxygenation index between groups. Although heart rates were signifi cantly higher in the salbutamol arm (average 4 bpm), there was no diff erence in rates of new atrial fi brillation (10% in both groups) or other cardiac arrhythmias. Th ere were no diff erences in IL-6 and IL-8 at baseline or on day 3.
Th e BALTI-2 trial [32,33] sought to extend the results of BALTI-1 and determine whether treatment with intravenous salbutamol (15 μg/kg/h) early in the course of ARDS would improve clinical outcomes. Th e trial commenced in December 2006 with the target of recruiting 1334 patients but was terminated in March 2010 due to safety concerns after the second interim analysis showed a signifi cant (p = 0.02) adverse eff ect of salbutamol on 28-day mortality and the 99.8% CI excluded a benefi t for salbutamol of the size anticipated in the protocol. Th e fi nal analysis confi rmed that intravenous salbutamol increased 28-day mortality (salbutamol group 34.2% [55/161], placebo group 23.3% [38/163]; risk ratio 1·47, 95% CI 1.03-2.08). Th e increase in mortality was associated with a reduction in the number of ventilator-free days (mean diff erence -2.7 days, 95% CI -4.7 to -0.7 days) and organ failure-free days (mean diff erence -2.3 days, 95% CI -4.5 to -0.1 days). Th e risks of developing a tachycardia, new arrhythmia, or lactic acidosis severe enough to warrant stopping the study drug were sub stantially higher in the salbutamol group (23 [14.3%] versus 2 [1.2%]; risk ratio 11·71, 95% CI 2.81 to 48.88).

Ongoing clinical trials
A search of the International Standard Randomized Controlled Trial Number (ISRCTN) register and Clinicaltrials.gov for trials using β-agonists in ARDS identifi ed two additional trials. Th e BALTI-prevention study is a double blind placebo controlled trial investigating whether inhalation of the long acting β-agonist salmeterol can prevent the development of ALI in patients undergoing esophagectomy (ISRCTN47481946) [34]. Th e trial completed recruitment in June 2011 and is due to report in 2012 when follow-up is complete. A second study, the Beta-agonists for Oxygenation in Lung Donors (BOLD) is testing the eff ect of nebulized salbutamol on oxygenation and lung transplantation rates in brain dead organ donors. Th is study has completed recruit ment and is due to report shortly.

Adverse cardiac eff ects
Salbutamol is known to have arrhythmia-and tachycardiainducing properties. Traditional teaching is that βadreno ceptors in the heart are of the β-1 subtype; however up to one third of the β-adrenoceptors in the atria and ventricles are β-2. Th e electrophysiological changes seen in patients treated with salbutamol are attributed to stimulation of these β-2 receptors. Th is myocardial stimu lation has the potential to result in increased myocardial oxygen demand, with detrimental eff ects on myocardial function, especially in hypoxic ARDS patients. In a study of patients with acute severe asthma, of the 129 patients given inhaled salbutamol, there was an increase in the incidence of ventricular and supraventricular ectopic beats, but none of the patients had a clinically signifi cant arrhythmia; this study excluded hypoxic patients, however [35]. In a large casecontrol study which enrolled elderly patients with chronic obstructive pulmonary disease (COPD), a dose response relationship was identifi ed between prescription of a β-agonist and subsequent development of an acute coronary syndrome [36]. Many patients with critical illness have comorbid cardiovascular disease. It is, therefore, possible that some patients experienced adverse cardiovascular events including occult cardiac ischemia. Th ese adverse eff ects may, therefore, limit any benefi t associated with alveolar fl uid balance.

Vasodilatation
Salbutamol is known to induce vasodilatation, an eff ect which precedes its bronchodilatory properties. Th is phenomenon and the increase in cardiac output cause an increase in ventilation/perfusion mis-match in patients when salbutamol is administered by the intravenous route. In a canine model of ALI, treatment with intravenous terbutaline increased cardiac index, aggravating capillary-alveolar macromolecular leakage [37]. A third possibility is the downregulation of β-agonist receptors (tachyphylaxis) during sustained treatment with βagonists. Although most experimental studies suggest this is a minor factor it has been described in some models [38].

Biochemical eff ects and lactic acidosis
Biochemical eff ects of salbutamol are known to include the induction of hypokalemia and hypomagnesemia. Th e BALTI-1 trial showed no signifi cant diff erence in plasma electrolyte concentrations between placebo treated and salbutamol treated groups. Th e risks of developing a tachycardia, new arrhythmia, or lactic acidosis, severe enough to warrant stopping the study drug, were substantially higher in the salbutamol group than in the placebo group in the BALTI-2 trial [33].
Lactic acidosis is also a recognized side eff ect of intravenous and nebulized β 2 agonists [39]. Th is eff ect is probably mediated by β 2 -adrenoreceptors and is thought to be due to an increase in skeletal muscle glycogenolysis and a subsequent rise in peripheral lactate production. Splanchnic glucose production and lactate extraction are also increased, as a result of increases in hepatic glycogenolysis and gluconeogenesis. Acidosis does not develop until the bicarbonate buff ering system is saturated, and this usually does not occur until lactate concentrations exceed 5 mmol/l [40]. In the BALTI-1 study there was no signifi cant diff erence in the incidence of lactic acidosis between the placebo and treatment arms of the study, and no patients had their infusion discontinued because of lactic acidosis. It should be noted however that this study was not statistically powered to detect such a diff erence in this adverse event.

Dosage and route of administration
Th e nebulized route for administration of β-agonists is better tolerated with fewer adverse side eff ects than the intravenous route. However, a limitation of the nebulized route is the lack of certainty that the drug reaches the required site of action, particularly in the setting of ALI which is characterized by injured, fl uid fi lled alveoli. Although the ALTA investigators had preliminary evidence that nebulized salbutamol achieved drug concen trations of 10 -6 M in undiluted pulmonary edema fl uid [41], there remains uncertainty as to whether the drug reached the required site of action.
Th e dose of salbutamol used in the BALTI-2 trial (15 μg/kg/h) was selected after an early dose ranging study identifi ed it to be the maximum dose that critically ill patients could receive without an increase in ventricular, atrial tachycardia or ectopy. Th is dose was used in the BALTI-1 study and achieved plasma levels of salbutamol (10 -6 M) [14] which are associated with a 100% increase in basal alveolar fl uid clearance in animal models of ARDS. Th is dose is at the higher end of the manufacturer's recommended dosing regimen. It is possible, therefore, that a lower dose of salbutamol might have been better tolerated so the conclusions from the BALTI studies only relate to the dose given.

Conclusion
Over three decades of intense research activity has examined the potential role that β-agonists could play in the treatment of ARDS. Pre-clinical trials suggested that these drugs could accelerate alveolar fl uid clearance, may have benefi cial immunomodulatory eff ects and may reduce alveolar-epithelial permeability. A small phase 2 randomized controlled trial demon strated proof of concept by showing that a sustained infusion of intravenous salbutamol reduced EVLW in patients with ARDS. Despite this preliminary evidence, the early promise has not held up to robust testing in the context of multicenter clinical trials. Th e ALTA trial was terminated on the grounds of futility after it became clear that salbutamol did not aff ect ventilator-free days. Th e BALTI-2 trial was terminated for similar reasons alongside concerns about safety and tolerability. Together these fi ndings suggest that the routine adminis tration of β-agonists as a treatment for ARDS should be avoided.