Definition and epidemiology
ARDS (previously ‘shock lung’) is not a single entity but represents the severe end of a spectrum of acute lung injury due to many different insults. Manifests as acute and persistent lung inflammation with increased vascular permeability. Most commonly seen on the ITU where about 10–20% of such patients will have ARDS depending on the definition. The 2012 ‘Berlin Definition’ of ARDS requires:
• Respiratory symptoms within 1 week of known clinical insult
• Bilateral opacities consistent with pulmonary oedema on CXR or CT
• Respiratory failure must not be fully explained by cardiac failure or fluid overload. Echo may be required to exclude hydrostatic pulmonary oedema
• Oxygenation impairment, defined by PaO2/FiO2 ratio at least <300mmHg (40kPa) despite positive end-expiratory pressure (PEEP) ≥5cmH2O:
• Mild ARDS—PaO2/FiO2 >200mmHg (27kPa)
• Moderate ARDS—PaO2/FiO2 >100mmHg (13kPa)
• Severe ARDS—PaO2/FiO2 ≤100mmHg (13kPa).
Other ARDS grading scores exist, e.g. Murray lung injury score (based on plain CXR findings, oxygenation, PEEP level, and respiratory system compliance; see p. [link]).
Inflammatory damage to the alveoli, either by locally produced pro-inflammatory mediators or remotely produced and arriving via the pulmonary artery. Changes in pulmonary capillary permeability allow fluid and protein leakage into the alveolar spaces with pulmonary infiltrates. The alveolar surfactant is diluted with loss of its stabilizing effect, resulting in diffuse alveolar collapse and stiff lungs. This leads to:
• Gross impairment of V/Q matching with shunting, causing arterial hypoxia and very large A–a gradients. There are usually enough remaining functioning alveoli such that hyperventilation maintains CO2 clearance; thus, hypercapnia is infrequently a problem
• PHT will develop 2° to the hypoxia, but this may be helpful (aids V/Q matching), rather than deleterious
• Reduced compliance (stiff lungs) due to loss of functioning alveoli (alveolar collapse, filled with fluid and protein) and hyperinflation of remaining alveoli to their limits of distension.
There are many causes
of pro-inflammatory mediator release sufficient to cause ARDS, and there may be more than one present. Common causes, in order of prevalence:
• Sepsis/pneumonia (2° risk factors for developing ARDS, when septic, alcoholism, and cigarette smoking)
• Gastric aspiration (even if on a PPI, indicating that a low pH is not the only damaging component)
• Trauma/burns (via sepsis, lung trauma, smoke inhalation, fat emboli, and possibly direct effects of large amounts of necrotic tissue).
• Acute pancreatitis
• Transfusion-related acute lung injury (TRALI), caused by any blood product (possibly due to HLA/white blood cell antibodies, commoner with older blood products, >6U); usually occurs within a few hours of transfusion. No specific therapy or evidence of steroid response
• Transplanted lung—worse if the lung poorly preserved
• Post-bone marrow transplant as bone marrow recovers
• Drug overdose, e.g. tricyclic antidepressants, opiates, cocaine, aspirin
• Near drowning
• Following upper airway obstruction; mechanism unclear.
The course of ARDS is fairly characteristic. Phase 1 is the early period of diffuse alveolar damage and hypoxaemia with pulmonary infiltration. Phase 2 develops after a week or so as the pulmonary infiltrates resolve and, on histology, seems to be associated with an increase in type II pneumocytes (surfactant producers), myofibroblasts, and early collagen formation. Phase 3 occurs in some. This is a fibrotic stage that leaves the lung with cysts, deranged micro-architecture, and much fibrosis on histology.
ARDS should be considered in any patient with a predisposing risk factor who develops severe hypoxaemia, stiff lungs, and a widespread diffuse pulmonary infiltrate. Approximately 1–2 days following the clinical presentation of the precipitating cause (sepsis, aspiration, etc.), there is rapidly worsening dyspnoea (± a dry cough) and hypoxaemia, requiring rapidly escalating amounts of supplemental O2 up to 100% via a non-rebreathe system (see pp. [link]–[link]). Coarse crackles in the chest. Intubation and ventilation are nearly always required, although initiating CPAP via a face mask at 5–10cmH2O with 100% O2 can improve oxygenation temporarily.
There are no specific tests that allow a confident diagnosis, and exclusion of other more specifically treatable diagnoses is required. The cause for the ARDS needs to be established and prevented from continuing or recurring, if possible. The CXR or CT shows diffuse alveolar infiltrates and air bronchograms, similar in appearance to cardiogenic pulmonary oedema or diffuse pulmonary haemorrhage.
• LVF (may be excluded on clinical grounds, echo, or, less commonly, checking pulmonary capillary wedge pressure <18mmHg)
• Diffuse alveolar haemorrhage (e.g. Goodpasture’s, Wegener’s, and SLE; clues will include a drop in Hb, blood in the airways and pulmonary secretions, and other clinical features of one of these disorders)
• Idiopathic acute eosinophilic pneumonia
• Cancer and lymphangitis carcinomatosis.
Some centres advocate lung biopsy to exclude alternatives, although most reserve biopsy for cases when the differential diagnosis includes conditions for which management would be changed (e.g. fungal infection, vasculitis, COP).
The essential aspects of management are to treat the precipitating cause, provide best supportive care with adequate oxygenation, and avoid further damage from barotrauma, hyperoxia, and nosocomial infections. Mechanical ventilation with PEEP and high inflation pressures are almost always required to maintain oxygenation (SaO2 values in the low 90s are entirely adequate). There is evidence that high inflation pressures may worsen ARDS directly (micro-barotrauma); therefore, try to maintain plateau pressures ≤30cmH2O.
Many ventilation strategies have been tried to reduce the high inflation pressures that result from the stiff lungs (low compliance). For example, using low tidal volume ventilation to reduce inflation pressures (6mL/kg ideal body weight, compared with 12mL/kg) reduces mortality by 9% and increases ventilator-free days. Use of high PEEP (attempt to open collapsed alveoli) has been shown in a meta-analysis to improve oxygenation and decrease time on a ventilator, and is associated with a lower ITU mortality for patients with PaO2/FiO2 ≤200mmHg (27kPa). Reducing the minute ventilation and allowing the PaCO2 to rise (permissive hypercapnia) also reduces the inflation pressures.
Prone ventilation has been tried in an attempt to improve V/Q matching, and initial increases in PaO2 are observed. A meta-analysis suggested a possible survival advantage for those with severe hypoxia, and a recent RCT has shown a significant reduction in both 28- and 90-day mortality for patients with severe ARDS.
Several different artificial surfactants have been tried to try and improve lung compliance, although good delivery to the abnormal areas is unlikely. Although effective in animal models, the RCTs have been negative in humans.
Haemodynamic monitoring, guided by central venous catheters, has a similar efficacy to pulmonary artery catheters but are associated with a 2× reduction in catheter-related complications (mostly, arrhythmias).
Different degrees of hydration have been compared, with reduced fluid balances improving gas exchange and shortening duration of mechanical ventilation. Secondary analysis of a cohort study suggested that a negative day 4 fluid balance is associated with decreased mortality, although this was not confirmed in a randomized study.
High-dose steroids have been used, but there is evidence of harm as well as benefit, and minimal evidence of overall improved survival. Three meta-analyses failed to demonstrate significant mortality improvements, while one showed decreased mortality. Certain subgroups may do slightly better and others worse, e.g. steroids are possibly beneficial during the first 14 days but detrimental thereafter. A 2007 RCT using prolonged methylprednisolone infusion (1mg/kg/day) for early ARDS (≤72h) showed improvement in a number of outcomes and halved mortality, but this was a small study (n = 91), and patient characteristics were imbalanced in the treatment arms.
of ARDS include:
• The high ventilation pressures lead to barotrauma: pneumothorax, surgical emphysema, pneumomediastinum. Pneumothorax may be lethal but difficult to detect on a CXR in the supine patient
• Nosocomial infections occur in about half the patients, making surveillance mini-BALs important
• Myopathy associated with long-term neuromuscular blockade, high steroid doses, and poor glycaemic control
• Non-specific problems of VTE, GI haemorrhage, inadequate nutrition.
has improved over the last 20y, probably due to improvements in supportive care and ventilator strategies, rather than an ability to modify the inflammatory process and its subsequent repair. Prognosis is worse with intra-pulmonary causes. Early deaths are usually due to the precipitating condition, and later deaths to complications. Over half of the patients will survive with varying residual lung damage, although the PFTs often show only minor restrictive abnormalities (and reduced kCO), indicating the considerable capacity of the lung to recover. A prospective cohort study showed that 6-minute walking test (6MWT) distance remained decreased (at 76% predicted), even at 5y.
• The optimal level of PEEP in a particular patient is difficult to predict. Inadequate PEEP allows more atelectasis, but too high PEEP contributes to overdistension of remaining alveoli and further barotrauma when there are no more ‘recruitable’ alveoli. Ways to estimate the best PEEP are under investigation. High-frequency oscillatory ventilation (HFOV) has been around a long while, but the recent OSCILLATE and OSCAR randomized trials suggested a lack of benefit, with possible harm associated with its use. Recently, liquid ventilation with perfluorochemicals has been tried. These dense O2-carrying liquids reduce the heterogeneity of ventilation by nullifying the requirement for surfactant, thus recruiting the collapsed alveoli. There are improvements in oxygenation but no evidence yet of clinically meaningful outcomes
• Nitric oxide (NO) has been tried, with clear improvements in oxygenation but very little effect on survival. The mode of action is not clear and may be more than just vasodilatation. Inhaled prostacyclin is similarly unconvincing. Anti-inflammatory and antioxidant therapies are still very much in the experimental phase
• ‘Off the shelf’ artificial lung systems are now becoming clinically useful to buy time while the lungs recover. The Novalung Interventional Lung Assist device is an example, but such therapy is very expensive.