|Home | About | Journals | Submit | Contact Us | Français|
Acute respiratory distress syndrome (ARDS) is characterised by lung inflammation with severe hypoxia, which usually develops over 4-48 hours and persists for days or weeks. The main causes of ARDS are infections, aspiration of gastric contents, and trauma. Between a third and a half of people with ARDS die from the disease, but mortality depends on the underlying cause. Some survivors have long-term respiratory or cognitive problems.
We conducted a systematic review and aimed to answer the following clinical question: What are the effects of interventions in adults with acute respiratory distress syndrome? We searched: Medline, Embase, The Cochrane Library and other important databases up to August 2006 (BMJ Clinical Evidence reviews are updated periodically, please check our website for the most up-to-date version of this review). We included harms alerts from relevant organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA).
We found 21 systematic reviews, RCTs, or observational studies that met our inclusion criteria.
In this systematic review we present information relating to the effectiveness and safety of the following interventions: corticosteroids, low tidal volume mechanical ventilation, nitric oxide, prone position, and protective ventilation.
Acute respiratory distress syndrome (ARDS) is characterised by lung inflammation with severe hypoxia, which usually develops over 4-48 hours and persists for days or weeks.
Low tidal volume ventilation, at 6 mL/kg of predicted body weight, reduces mortality compared with high tidal volume ventilation, but can lead to respiratory acidosis.
People with ARDS may remain hypoxic despite mechanical ventilation. Nursing in the prone position may improve oxygenation, but has not been shown to reduce mortality, and can increase sedation and facial oedema.
We found insufficient evidence to draw reliable conclusions on the effects of corticosteroids on mortality or reversal of ARDS.
Nitric oxide has not been shown to improve survival or duration of ventilation, or hospital admission, compared with placebo. It may modestly improve oxygenation in the short term but the improvement is not sustained.
Acute respiratory distress syndrome (ARDS), originally described by Ashbaugh et al in 1967, is a clinical syndrome that represents the severe end of the spectrum of acute lung injury (ALI). In 1994, the American-European Consensus Conference on ARDS made the following recommendations. Widespread acceptance of these definitions by clinicians and researchers has improved standardisation of clinical research. Acute lung injury: a syndrome of acute and persistent inflammatory disease of the lungs characterised by three clinical features: 1) bilateral pulmonary infiltrates on the chest radiograph; 2) a ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2 ) of less than 300; 3) absence of clinical evidence of left atrial hypertension (if measured, the pulmonary capillary wedge pressure is ≤ 18 mm Hg). Acute respiratory distress syndrome: The definition of ARDS is the same as that of ALI, except that the hypoxia is severe a PaO2/FiO2 ratio of 200 mm Hg or less. The distinction between ALI and ARDS is arbitrary, because the severity of hypoxia does not correlate reliably with the extent of the underlying pathology, and does not influence predictably clinical course or survival. ARDS is an acute disorder, which typically develops over 4-48 hours and persists for days to weeks. Subacute or chronic lung diseases, such as sarcoidosis and idiopathic pulmonary fibrosis, are excluded from the definition of ARDS. The early pathological features of ARDS are generally described as diffuse alveolar damage. Recognition of diffuse alveolar damage requires histological examination of the lung tissue, which is not necessary to make a clinical diagnosis. Population: For the purpose of this review, we have defined ARDS as including people with ALI and ARDS. It therefore includes adults with ALI and ARDS from any cause and with any level of severity. Neonates and children less than 12 years of age have been excluded.
Ten to fifteen per cent of all people admitted to an intensive care unit, and up to 20% of mechanically ventilated people meet the criteria for ARDS. The incidence of ALI in the USA (17-64/100 000 person years) seems to be higher than in Europe, Australia, and other developed countries (17-34/100 000 person years). One prospective, population based cohort study (1113 people in Washington State, aged over 15 years) found the crude incidence of ALI to be 78.9/100 000 person years, and the age adjusted incidence to be 86.2/100 000 person years. An annual incidence of 15.5 cases a year or 5.9 cases/100 000 people a year was reported in a recent epidemiological study from Iceland. An observational cohort reported that in Shanghai, China, of 5320 adults admitted to intensive care units in 1 year, 108 (2%) had symptoms that met with ARDS criteria.
ARDS encompasses many distinct disorders that share common clinical and pathophysiological features. More than 60 causes of ARDS have been identified. Although the list of possible causes is long, most episodes of ARDS are associated with a few common causes or predisposing conditions, either individually or in combination. These include sepsis, aspiration of gastric contents, infectious pneumonia, severe trauma, surface burns, lung contusion, fat embolism syndrome, massive blood transfusion, lung and bone marrow transplantation, drugs, acute pancreatitis, near drowning, cardiopulmonary bypass, and neurogenic pulmonary oedema. The incidence of ALI in a large cohort of people with subarachnoid haemorrhage has been reported to be 27% (170/620 people; 95% CI, 24% to 31%). One or more of these predisposing conditions are often evident at the onset of ALI. When ARDS occurs in the absence of common risk factors such as trauma, sepsis, or aspiration, an effort should be made to identify a specific cause for lung injury. In such cases, a systematic review of the events that immediately preceded the onset of ARDS is normally undertaken to identify the predisposing factors.
Mortality: Survival for people with ARDS has improved remarkably in recent years, and cohort studies have found mortality to range from 34% to 58%. Mortality varies with the cause; however, by far the most common cause of death is multiorgan system failure rather than acute respiratory failure. In a prospective cohort study (207 people at risk of developing ARDS, of which 47 developed ARDS during the trial), only 16% of deaths were considered to have been caused by irreversible respiratory failure. Most deaths in the first 3 days of being diagnosed with ARDS could be attributed to the underlying illness or injury. Most late deaths (after 3 days, 16/22 [72.7%]) were related to sepsis syndrome. One prospective cohort study (902 mechanically ventilated people with ALI) found that an age of 70 years or younger significantly increased the proportion of people who survived at 28 days (74.6% aged ≤ 70 years v 50.3% aged ≥ 71 years; p < 0.001). In a recent observational study (2004), the overall intensive care unit mortality was 10.3%. In-hospital mortality was 68.5%, and 90 day mortality was 70.4% in people with ARDS, and accounted for 13.5% of the overall intensive care unit mortality. Lung function and morbidity: One cohort study of 16 long term survivors of severe ARDS (lung injury score ≥ 2.5) found that only mild abnormalities in pulmonary function (and often none) were observed. Restrictive and obstructive ventilatory defects (each noted in 4/16 [25%] people) were observed in ARDS survivors treated with low or conventional tidal volumes. One cohort study of 109 people found no significant difference between various ventilatory strategies and long term abnormalities in pulmonary function or health related quality of life. However, it did find an association between abnormal pulmonary function and decreased quality of life at 1 year follow up. One retrospective cohort study (41 people with ARDS) found that duration of mechanical ventilation and severity of ARDS were important determinants of persistent symptoms 1 year after recovery. Better lung function was observed when no illness was acquired during the intensive care unit stay, and with rapid resolution of multiple organ failure (e.g. pneumonia during ARDS: 7/41 [17.1%] people with long term impairment v 2/41 [4.9%] with no long term impairment; significance assessment not performed). Persistent disability 1 year after discharge from the intensive care unit in survivors of ARDS is secondary to extrapulmonary conditions, most importantly muscle wasting and weakness. Cognitive morbidity: One cohort study (55 people 1 year after ARDS) found that 17/55 (30.1%) exhibited generalised cognitive decline and 43/55 (78.2%) had all, or at least one, of the following: impaired memory, attention, concentration, and decreased mental processing speed. These deficits may be related to hypoxaemia, drug toxicity, or complications of critical illness. To date, no association between different ventilatory strategies and long term neurological outcomes has been found.
Goals of treatment of people with ARDS are identification and treatment of the underlying clinical disorder and optimal supportive care. In many people with ARDS, the insult that caused lung injury, such as aspiration or multiple transfusions, cannot be treated except to prevent recurrences. Supportive care consists of appropriate ventilator management, prevention of infections, multiorgan failure, and complications of critical care.
Short term: Mortality; length of intensive care unit and hospital stay; ventilator free days (defined as days alive and free from mechanical ventilation); adverse events associated with mechanical ventilation (barotrauma, haemodynamic dysfunction); adverse events associated with low tidal volume ventilation (severe acidosis, central nervous system dysfunction). Long term: Survival to discharge from hospital; survival at 28 days; quality of life/functional outcomes (people discharged home or to an institution, or measured using validated methods such as health related quality of life or the Medical Outcome Study 36 Item Short Form Health Survey.
BMJ Clinical Evidence search and appraisal August 2006. The following databases were used to identify studies for this review: Medline 1966 to August 2006, Embase 1980 to August 2006, and The Cochrane Library and Cochrane Central Register of Controlled Clinical Trials Issue 3, 2006. Additional searches were carried out using these websites: NHS Centre for Reviews and Dissemination (CRD) — for Database of Abstracts of Reviews of Effects (DARE) and Health Technology Assessment (HTA), Turning Research into Practice (TRIP), and National Institute for Health and Clinical Excellence (NICE). Abstracts of the studies retrieved were assessed independently by two information specialists using predetermined criteria to identify relevant studies. Study design criteria for inclusion in this review were: published systematic reviews, RCTs, prospective clinical trials with a control group (non-randomised); case control studies, and prospective and retrospective cohort studies in any language. Studies had to contain 20 or more individuals, and RCTs had to be at least single blinded, unless blinding was impossible, with at least 80% of the population followed up. There was no minimum length of follow up required to include studies. We excluded all studies described as "open", "open label", or not blinded, unless blinding was impossible. Lower quality evidence was only included in the review when RCT evidence was found to be unavailable for the outcomes of interest. Studies where the outcomes did not include any from the above list were excluded. In addition we use a regular surveillance protocol to capture harms alerts from organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA), which are added to the review as required.
The information contained in this publication is intended for medical professionals. Categories presented in Clinical Evidence indicate a judgement about the strength of the evidence available to our contributors prior to publication and the relevant importance of benefit and harms. We rely on our contributors to confirm the accuracy of the information presented and to adhere to describe accepted practices. Readers should be aware that professionals in the field may have different opinions. Because of this and regular advances in medical research we strongly recommend that readers' independently verify specified treatments and drugs including manufacturers' guidance. Also, the categories do not indicate whether a particular treatment is generally appropriate or whether it is suitable for a particular individual. Ultimately it is the readers' responsibility to make their own professional judgements, so to appropriately advise and treat their patients.To the fullest extent permitted by law, BMJ Publishing Group Limited and its editors are not responsible for any losses, injury or damage caused to any person or property (including under contract, by negligence, products liability or otherwise) whether they be direct or indirect, special, incidental or consequential, resulting from the application of the information in this publication.
Two RCTs found that low tidal volume ventilation decreased mortality at 28 and 180 days compared with high tidal volume ventilation. One RCT also found that low tidal volume ventilation increased ventilator free days. The target low tidal volume was 6 mL/kg of predicted body weight. One RCT found no significant difference between low tidal volume and high tidal volume ventilation in the duration of hospital stay. Low tidal volume ventilation may lead to respiratory acidosis, which may require treatment with either high respiratory rates, sodium bicarbonate infusion, or both.
We found three systematic reviews and one observational study comparing the effects of low tidal volume mechanical ventilation versus high tidal volume mechanical ventilation in acute respiratory distress syndrome (ARDS). The first systematic review (search date 2003, 3 RCTs) defined low tidal volume mechanical ventilation as a tidal volume of 7 ml/kg or less, and high tidal volume mechanical ventilation as a tidal volume in the range of 10–15 ml/kg ideal body weight. The review found that, compared with high tidal volume mechanical ventilation, low tidal volume mechanical ventilation reduced mortality at 28 days and 180 days, but the reduction was not statistically significant at 180 days (critically ill people aged ≥ 16 years with acute lung injury (ALI) or ARDS; mortality at 28 days: 3 RCTs, 1030 people; 142/519 [27.3%] with low tidal volume v 189/511 [37.0%] with high tidal volume; RR 0.74, 95% CI 0.61 to 0.88, P = 0.0008; mortality at 180 days: 4 RCTs, 1086 people; 189/547 [34.6%] with low tidal volume v 227/539 [42.1%] with high tidal volume; RR 0.84, 95% CI 0.68 to 1.05). It is difficult to interpret these combined results because of the differences between included trials in clinical parameters, such as different lengths of follow up and relatively higher plateau pressures in the control arms in two of the included trials. The review conducted a subgroup analysis comparing the effects of different plateau pressures. It found that overall mortality was significantly lower with low tidal volume when a higher plateau pressure (mean pressure > 31 cm H2O) was applied in the control arm (2 RCTs, 914 people; 146/461 [31.7%] with low tidal volume v 187/453 [41.2%] with high tidal volume; RR 0.76, 95% CI 0.64 to 0.91). It was not significantly lower when a lower plateau pressure (mean pressure ≤ 31 cm H2O) was used in the control arm (3 RCTs, 288 people; 70/144 [48.6%] with low tidal volume v 62/144 [43.0%] with high tidal volume; RR 1.13, 95% CI 0.88 to 1.45). The three individual RCTs are tabulated in table 1 . The second systematic review (search date 2006, 5 controlled trials, 1202 people with ALI or ARDS) compared protective ventilation with control on 28 day mortality. The review found that protective ventilation significantly improved mortality at 28 days compared with control ventilation (OR 0.71, 95% CI 0.56 to 0.91; P = 0.006, no absolute figures reported). The third systematic review (search date 2006, 7 trials, 1435 people) evaluated whether low ventilation reduced mortality in ARDS. The review found that low tidal volume ventilation significantly improved mortality compared with traditional ventilation (36% with low tidal volume ventilation v 43% with traditional ventilation; OR 0.75, 95% CI 0.61 to 0.93, P < 0.01). The observational study (3147 people admitted to 1 of the included 198 European intensive care units) evaluated the use of higher tidal volumes than those applied in the ARDS Network (ARDSnet) study (> 7.4 mL/kg of predicted body weight). The review found that use of tidal volumes higher than those used by the ARDSnet study was an independent risk factor for mortality.
The three reviews and observational study gave no information on adverse effects. The first RCT identified by the review found no significant difference between lower and higher tidal volumes in barotrauma between days 1 and 28 (10% with lower tidal volumes v 11% with higher tidal volumes; P = 0.43) It acknowledged that pneumothorax (the most common manifestation of barotrauma) is not a sensitive or specific marker of stretch induced injury with the tidal volumes used in the study. The RCT also excluded people with elevated intracranial pressure and with sickle cell haemoglobin, because hypercapnia and acidosis could have adverse effects in these conditions. A slightly lower level of oxygenation was observed in the low tidal volume group over the first few days; however, participants did not develop significant degrees of hypercapnia, and did not require excessive neuromuscular paralysis. The second RCT identified by the review found that, compared with high tidal volume ventilation, low tidal volume ventilation increased the requirement for paralytic agents and dialysis for renal failure (paralytic agents: 23/60 [38%] with low tidal volume v 13/60 [22%] with high tidal volume; p = 0.05; dialysis for renal failure: 13/60 [22%] with low tidal volume v 5/60 [8%] with high tidal volume; P = 0.04). The RCT also found that, compared with high tidal volume ventilation, low tidal volume ventilation significantly increased the proportion of people with permissive hypercapnia, as well as its severity and length of duration (permissive hypercapnia defined as an arterial carbon dioxide tension > 50 mm Hg: 31/62 [50%] with low tidal volume ventilation v 17/60 [28%] with high tidal volume ventilation; P = 0.009; mean PaCO2 : 54.4 mm Hg with low tidal volume ventilation v 45.7 mm Hg with high tidal volume ventilation; P = 0.002; mean length of hypercapnia: 146 hours with low tidal volume ventilation v 25 hours with high tidal volume ventilation; P = 0.017). One retrospective study (111 people) examined the effects of mechanical ventilation with a tidal volume of 6 mL/kg compared with 12 mL/kg predicted body weight on haemodynamics, vasopressor use, fluid balance, diuretics, sedation, and neuromuscular blockade within 48 hours in people with ALI/ARDS. Compared with 12 mL/kg predicted body weight, treatment with a tidal volume of 6 mL/kg predicted body weight had no adverse effects on haemodynamics. There were also no differences in the need for supportive treatments, including vasopressor drugs, intravenous fluids, or diuretic drugs. In the 12 mL/kg tidal volume group, 44% of people received dopamine and 17% received phenylephrine (P = 0.84); whereas, in 6 mL/kg tidal volume group 47% were treated with dopamine and 9% with phenylephrine (P = 0.38). In addition, there were no differences in body weight, urine output, or fluid balance. Finally, there was no difference in the need for sedation or neuromuscular blockade between the two tidal volume protocols. On day 1, 16% of the 12 mL/kg group required neuromuscular blockade compared with 7% of the 6 mL/kg group (P = 0.22). On day 2, 13% of the 12 mL/kg group required neuromuscular blockade compared with 4% of the 6 mL/kg group (P = 0.15). A secondary analysis of an RCT (61 people with ALI) compared the doses and duration of sedatives and opioid analgesics in people receiving low versus traditional tidal volumes. In 33 people randomised to the lower tidal volume arm (6 mL/kg of predicted body weight) and 28 people randomised to the higher tidal volume arm (12 mL/kg of predicted body weight), there were no significant differences in the percentage of study days people received sedatives, opioids, or neuromuscular blockade. People received neuromuscular blockade an average of 5% of study days in the lower tidal volume group compared with 13% of study days in the higher tidal volume group (difference –8%, 95% CI –20% to +3%; P = 0.16). Furthermore, there were no significant differences in the proportion of people receiving benzodiazepines, propofol, haloperidol, and opioids on days 1, 2, 3, and 7 of mechanical ventilation. There were no differences in the doses of benzodiazepines and opioids on those days. Sedatives were used (including benzodiazepines and propofol) an average of 81% of study days in the lower tidal volume group and 92% of study days in the higher tidal volume group (difference –11%, 95% CI –24% to +3%; P = 0.13). Likewise, opioid analgesia was used an average of 85% of study days in the lower tidal volume group compared with 83% of study days in the higher tidal volume group (difference 2%, 95% CI –14% to +18%; P = 0.85).
ARDS is associated with lung regions marked by decreased respiratory system compliance caused by atelectasis, alveolar flooding, and increased surface tension at air–fluid interfaces. Other lung regions will have normal compliance and aeration, and in other intermediate regions alveolar collapse and flooding can be reversed. Traditional tidal volumes of 10–15 mL/kg result in elevated airway pressures and overdistention of the less affected lung regions, which exacerbate or perpetuate the lung injury. Ventilation with small tidal volumes and limited airway pressures can reduce ventilator associated lung injury from overdistention. In a person with ALI/ARDS requiring mechanical ventilation, the goal must be to provide adequate oxygenation without engendering morbidity from oxygen toxicity, haemodynamic compromise, barotraumas, and alveolar overdistention. Traditional approaches to mechanical ventilation have involved the use of tidal volumes of 10–15 mL/kg body weight. These large tidal volumes are known to cause stretch induced lung injury and release of inflammatory mediators, and perpetuate the cycle of inflammation and injury in people with ALI and ARDS. Unfortunately, despite published evidence supporting ventilation with low tidal volumes, such ventilatory strategy remains underused. A recent observational cohort study (88 people with ARDS) found that 39% had ventilation with tidal volumes of 7.5 mL/kg predicted body weight or less on day 2, 49% on day 4, and 56% on day 7. In contrast, 49% of people had ventilation with tidal volumes of more than 8.5 mL/kg predicted body weight on day 2 of ALI, 30% on day 4, and 24% on day 7. The use of low tidal volumes was significantly associated with clinical parameters indicative of worse disease severity compared with other tidal volumes, including low values for arterial oxygen tension (PaO2 ; P = 0.01), ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2 ; P = 0.08), and static compliance of the respiratory system (P = 0.006).
There is strong evidence of benefit of mechanical ventilation utilising a low tidal volume. The ARDS Network trial recommended mechanical ventilation at 6 ml/kg ideal body weight in those with ALI/ARDS. A common consequence of a low tidal volume ventilatory strategy is development of hypercapnia and respiratory acidosis which is tolerated (permissive hypercapnia). As long as adequate oxygenation is achieved, hypercapnia is an acceptable adverse effect of controlled ventilation. Contraindications to permissive hypercapnia include predisposition to increased cranial pressure (intracerebral bleeding, brain tumour, fulminant hepatic failure), and haemodynamic instability. Sedation administration should not be considered a barrier to implementing a lung protective ventilation strategy.
Two additional RCTs found no significant difference between low tidal volume and high tidal volume mechanical ventilation in mortality and duration of mechanical ventilation. However, both of these studies had a sample size too small to detect a treatment difference. In addition, the differences in intervention (tidal volumes and plateau pressures) between the two groups were modest, and therefore not likely to show a benefit, particularly in view of the low sample size (see table 2 ).
Three RCTs found no significant difference between higher and lower positive end expiratory pressure (PEEP) strategies in overall mortality, although two of the RCTs found that protective ventilation improved other outcomes, including weaning rates. One RCT found no significant difference between high PEEP and lower PEEP ventilation strategies in people managed with low tidal volume ventilation. There is consensus that PEEP is effective in people with acute respiratory distress syndrome. Protective ventilation uses PEEP to keep the alveoli open throughout the entire respiratory cycle, but current evidence does not support the routine application of a high PEEP strategy in people with acute lung injury/acute respiratory distress syndrome.
We found three RCTs investigating a ventilatory strategy using protective ventilation (positive end expiratory pressure; PEEP). The first RCT (28 people with early acute respiratory distress syndrome [ARDS]) compared protective ventilation versus conventional ventilation (protective ventilation: consisting of maintenance of end expiratory pressures above the lower inflection point of the static pressure–volume curve, tidal volume < 6 ml/kg, peak pressures < 40 cm H2 O, permissive hypercapnia, and stepwise utilisation of pressure limited modes; conventional ventilation: consisting of conventional volume cycled ventilation, tidal volume 12 ml/kg, minimum PEEP guided by FiO2 and haemodynamics, and normal PaCO2 levels). It found that, compared with conventional ventilation, protective ventilation significantly improved the ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2 ratio; p < 0.0001), lung compliance (P = 0.0018), and weaning rate (P = 0.014). However, mortality was similar in both groups (overall mortality: 5/15 [53.8%] with protective ventilation v 7/13 [33.3%] with mechanical ventilation; P = 0.45). The second RCT (53 people, 28 people with early ARDS taken from the first RCT) compared conventional versus protective mechanical ventilation (conventional ventilation: lowest PEEP for acceptable oxygenation and a tidal volume of 12 mL/kg body weight; protective ventilation: PEEP above the lower inflection point on the static pressure–volume curve and a tidal volume < 6 ml/kg body weight). It found that, compared with conventional ventilation, protective ventilation reduced mortality at 28 days and at hospital discharge, although this was not significant at discharge (mortality at 28 days: 11/29 [38%] with protective ventilation v 17/24 [71%] with conventional ventilation; P < 0.001; mortality at hospital discharge: 13/29 [45%] with protective ventilation v 17/24 [71%] with conventional ventilation; P = 0.37). It found that, compared with conventional ventilation, protective ventilation significantly increased the proportion of people successfully weaned from mechanical ventilation (66% with protective ventilation v 29% with conventional ventilation; P = 0.005). The third RCT (ARDS Network trial, 549 people with acute lung injury and ARDS) compared mechanical ventilation with lower versus higher PEEP levels (mean PEEP values at day 4: 8.3 cm H2 O with lower PEEP v 13.2 cm H2 O with higher PEEP; P < 0.001). It found no significant difference between lower and higher PEEP in mortality before discharge or ventilator free days at day 28 (mortality before hospital discharge: 24.9% with lower PEEP v 27.5% with higher PEEP; P = 0.48; mean ventilator free days at day 28: 14.5 days with lower PEEP v 13.8 days with higher PEEP group; P = 0.50).
There is consensus that PEEP is effective in people with ARDS. Ideal PEEP will help to achieve adequate oxygenation and decrease the requirement for high fractions of inspiratory oxygen without causing any of the harmful effects of PEEP. However, the level of PEEP needed to confer maximum benefit with minimum complications has only recently been studied. In the ARDS Network trial, higher PEEP produced better oxygenation and lung compliance, but no benefit to survival, time on ventilator, or non-pulmonary organ dysfunction. Although sufficient PEEP is essential in ventilation management of people with ARDS, this level varies from person to person.
Management guidelines for mechanical ventilation in sepsis-induced ARDS/acute lung injury have been formulated under the auspices of the Surviving Sepsis Campaign, an international effort to increase awareness and improve outcome in severe sepsis. A minimum amount of PEEP should be set to prevent lung collapse at end expiration, and may be guided by FiO2 requirement or measurement of thoracopulmonary compliance. Low tidal volume ventilation and limitation of end inspiratory plateau pressure is important, and may be facilitated by permissive hypercapnia. In people with severe ARDS, prone positioning should be considered. The ARDS Network trial found no survival advantage in using higher PEEP levels independently of the tidal volume strategy. High levels of PEEP may also lead to overdistention, barotrauma, decreased venous return, and impaired oxygen delivery. After institution of low tidal volume mechanical ventilation (6 ml/kg), PEEP should be gradually increased to lower FiO2 to less than 60% to maintain PaO2 above 60 mm Hg. There is no evidence that high levels of PEEP (> 10 cm H2O) can be routinely recommended. However, severely hypoxaemic people may require a higher PEEP to recruit atelectatic alveolar units in order to decrease intrapulmonary shunt. Titrating PEEP to maximise lung compliance (which is monitored) and ensuring that lungs are being ventilated well above the lower inflection point could be tried in people whose survival is threatened because of hypoxaemia.
No new evidence
Prone position can improve oxygenation, but this benefit must be carefully weighed against the lack of any good evidence of benefit or mortality, and uncommon but potentially serious harms. One systematic review found that prone positioning improved oxygenation in 69% of people with acute respiratory distress syndrome. However, the review and one subsequent RCT found no difference in mortality at 10 days and at 6 months between supine and prone positioning. One small controlled clinical trial found that both prone positioning and positive end expiratory pressure improved oxygenation compared with supine positioning alone. Subgroup analysis found that only prone positioning improved oxygenation in those with localised infiltrates, compared with supine positioning or positive end expiratory pressure. Adverse effects of prone positioning include increased sedation, facial oedema, and accidental extubation. Spinal instability is an absolute contraindication to prone positioning. Relative contraindications include haemodynamic and cardiac instability, and recent thoracic or abdominal surgery.
We found one systematic review, two subsequent RCTs, and one randomised prospective trial. The systematic review (search date 1998, 297 people, 14 prospective cohort studies, 3 RCTs) compared prone positioning with usual care in the supine position. The systematic review did not report the RCT data separately. It noted that the timing from the onset of respiratory failure to when participants were first positioned prone, and the frequency of the prone position, varied between studies (length of time in prone position: 30 minutes to 42 hours). It found that 148/213 (69.5%) of people had an improved ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2 ) ratio of more than 20 mm Hg, or 20% of the baseline within 2 hours, when moved from supine to prone. However, it found no significant difference between supine and prone positioning in mortality (figures not reported; significance assessment not performed). The first subsequent RCT (304 people, 94% with acute respiratory distress syndrome [ARDS] and 6% with acute lung injury [ALI]) compared prone position with supine position. It found no significant difference between prone position and supine position in mortality at 10 days, at time of discharge from the intensive care unit, or at 6 months (mortality at 10 days: 32/152 [21.1%] with prone position v 38/152 [25.0%] with supine position; RR 0.84, 95% CI 0.56 to 1.27; mortality at intensive care unit discharge: 77/152 [50.7%] with prone position v 73/152 [48.0%] with supine position; RR 1.05, 95% CI 0.84 to 1.32; mortality at 6 months: 95/152 [62.5%] with prone position v 89/152 [58.6%] with supine position; RR 1.06, 95% CI 0.88 to 1.28). The second subsequent RCT (130 people with ARDS, ALI, or respiratory failure) compared supine position ventilation with prone position ventilation (continuous prone ventilation treatment for 20 hours/day). The RCT found no significant difference in intensive care mortality with supine position ventilation compared with prone position ventilation (35/60 [58%] with supine position v 37/76 [43%] with prone position; P = 0.12). The prone position for ventilation had a higher Simplified Acute Physiology Score II at inclusion. Multivariate analysis found that Simplified Acute Physiology Score II at inclusion (OR 1.07, CI not reported; P < 0.001), number of days elapsed between ARDS diagnosis and inclusion (OR 2.83, CI not reported; P < 0.001), and randomisation to supine position (OR 2.53, CI not reported; P = 0.03) were independent risk factors for mortality. The prospective randomised trial (40 people with ALI or ARDS) compared the effect of prone position ventilation (at least 8 hours and a maximum of 23 hours/day) with supine position ventilation on the duration of mechanical ventilation. The study found that the duration of ventilatory support did not differ significantly between groups (30 ± 17 days with prone position v 33 ± 23 days with supine position; no significance assessment performed). Death and deterioration of gas exchange were seen (PaO2/FiO2 ratio) for 41 ± 29 days with prone position and 61 ± 35 days with supine position (P = 0.06). Oxygenation (PaO2/FiO2 ratio) improved significantly over the first 4 days of treatment with the prone position compared with the supine position (P = 0.03). The prevalence of ARDS after ALI (P = 0.03) and of pneumonia (P = 0.048) were also reduced in the prone position compared with the supine position.
We found one controlled clinical trial (25 people with ARDS; computed tomography scan used to identify those with localised infiltrates or diffuse infiltrates), which compared the effect on oxygenation of prone position versus supine position in the presence of varying levels of additional positive end expiratory pressure (PEEP). Oxygenation measurements were taken at four PEEP levels (0, 5, 10, and 15 cm H2O), applied in a random order in both positions. It found that, compared with the supine position, the prone position significantly improved oxygenation, defined as an increased PaO2/FiO2 ratio (mean: 86 with supine v 152 with prone at zero PEEP; p = 0.002; overall results presented graphically; p < 0.001). PEEP independently improved oxygenation compared with supine positioning (p < 0.001). A subgroup analysis found that, although both PEEP and the prone position significantly improved oxygenation in people with diffuse infiltrates compared with baseline measures (P < 0.001), only the prone position improved oxygenation in people with localised infiltrates (results presented graphically; significance assessment not performed).
Adverse effects are uncommon but potentially serious during prone positioning in people with ARDS. The total number of prone cycles (from supine to prone and back again) in the review was 746. It found that prone positioning was associated with haemodynamic instability, inadvertent extubation, desaturation, endotracheal tube obstruction, dislodgement of a central venous catheter, and dislodgement of a femoral haemodialysis catheter (haemodynamic instability: 8 events, 1.1% per prone cycle; inadvertent extubation: 3 events, 0.4% per prone cycle; desaturation: 2 events, 0.3% per prone cycle; endotracheal tube obstruction: 1 event, 0.1% per prone cycle; dislodgement of central venous catheter: 1 event, 0.1% per prone cycle; dislodgement of femoral haemodialysis catheter: 1 event, 0.1% per prone cycle). Significance assessments were not performed for any of these comparisons. In the subsequent RCT, there was no significant difference between the prone and supine positions in the number of pressure sores, new or worsening pressure sores, tracheal tube displacement, loss of venous access, or displacement of thoracostomy (number of pressure sores: 22.5% with supine position v 24.0% with prone position; p = 0.78; new or worsening pressure sores during 10 day study period: 27.5% with supine position v 36.0% with prone position; p = 0.13; tracheal tube displacement: 9.9% with supine position v 7.9% with prone position; p = 0.68; loss of venous access: 9.2% with supine position v 5.3% with prone position; P = 0.27; displacement of a thoracotomy tube: 0.7% with supine position v 3.9% with prone position; p = 0.12; absolute figures for all outcomes not reported). Adverse effects associated with prone positioning included an increased need for sedation and muscle relaxants (55.2%), airways obstruction (39.3%), and facial oedema (29.8%). The second subsequent RCT reported a total of 718 turning cycles. Although a total of 28 complications were reported, most were rapidly reversible. These included oedema (facial, limbs, thorax), conjunctival haemorrhage, pressure sores, and accidental dislodgement of lines and tubes. The prospective trial gave no information on adverse effects.
The first subsequent RCT (162 people) performed a subgroup analysis not originally part of the study design in people with at least one of three high risk characteristics: low Pao2/Fio2 ratio, high Simplified Acute Physiology Score, and high tidal volume (79 with supine position and 83 with prone position; 111 with one characteristic and 51 with 2 or 3). It found that, compared with the supine position, prone positioning significantly decreased the proportion of people who had died at 10 days (40.0% with supine position v 20.5% with prone position; RR 0.54, 95% CI 0.32 to 0.90). These differences in mortality did not persist beyond discharge from the intensive care unit. Because the mortality benefit was evident only on this subgroup analysis, further studies are required to validate the results.
Despite mechanical ventilation — the primary treatment used in ARDS to improve arterial oxygenation — a significant number of people remain hypoxaemic. Prone position ventilation may help in 60–70% of people. Because not everyone will respond, a brief test of the prone position is recommended to assess responsiveness. The review recommended that an increase in PaO2 within the first 60 minutes after prone positioning predicts continued improvement for several hours. The optimal duration of this treatment, and the repeat benefit of successive trials, are not currently known.
There is insufficient evidence to assess corticosteroids in people with acute respiratory distress syndrome (ARDS). One RCT found no significant difference between methylprednisolone and placebo in mortality or reversal of ARDS at day 45. However, another weak RCT found that methylprednisolone reduced mortality in the intensive care unit, and increased the proportion of people discharged at day 10. Although RCT evidence is lacking, corticosteroids are sometimes used in people with persistent ARDS.
We found four RCTs. Two RCTs found different results for methylprednisolone compared with placebo. The first RCT (99 people with acute respiratory distress syndrome [ARDS] and refractory hypoxaemia, diffuse bilateral infiltration on chest radiography, and absence of congestive heart failure by pulmonary catheterisation) compared methylprednisolone with placebo. It found no significant difference between methylprednisolone and placebo in mortality or reversal of ARDS at day 45 (mortality: 30/50 [60%] with methylprednisolone v 31/49 [63%] with placebo; P = 0.74; reversal of ARDS: 18/50 [35%] with methylprednisolone v 19/49 [39%] with placebo; P = 0.77). The second small RCT (24 people with ARDS who failed to improve after the seventh day of respiratory failure) compared prolonged methylprednisolone with placebo. Sixteen people received methylprednisolone and eight people received placebo. Methylprednisolone was given at 2 mg/kg daily initially and was then tapered over the next 32 days. It found that, compared with placebo, methylprednisolone significantly and substantially reduced mortality in the intensive care unit, mean lung injury score, and multiorgan dysfunction syndrome score. In addition, methylprednisolone significantly increased the proportion of people discharged and extubated at day 10, and the mean ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2 ratio; see table 3 ). The small number of enrolled subjects, and crossover of half the placebo-treated participants to the corticosteroids group, make these results difficult to interpret. Post hoc analysis of a placebo controlled RCT (300 people with septic shock with or without early ARDS) evaluated low doses of corticosteroids (7 day treatment with hydrocortisone 50 mg every 6 hours and 9 alpha fludrocortisone 50 µg once a day) in septic shock. The post hoc analysis of the people with septic shock and ARDS found that low dose corticosteroids significantly increased median time to death, and significantly reduced mortality at day 28 (median time to death: 16 days with corticosteroids v 12 days with placebo; HR 0.58, 95% CI 0.39 to 0.85, P = 0.05; mortality at day 28: 49/85 [58%] deaths with steroids v 62/92 [67%] deaths with placebo; OR 0.48, 95% CI 0.23 to 0.98, P = 0.04). The fourth RCT (180 people with ARDS of at least 7 days' duration) compared methylprednisolone versus placebo. The RCT found that methylprednisolone was associated with significantly increased 60 and 180 day mortality among people enrolled at least 14 days after the onset of ARDS compared with placebo (60 day mortality: 35% with methylprednisolone v 8% with placebo; P = 0.02; 180 day mortality: 44% with methylprednisolone v 12% with placebo; P = 0.01). Despite a lack of mortality benefit, compared with placebo, methylprednisolone significantly increased the number of ventilator free (P < 0.001) and shock free days (P = 0.03), improved oxygenation, respiratory system compliance, and blood pressure, with fewer days of vasopressor treatment during the first 28 days (no further data reported).
The first RCT found no significant difference between methylprednisolone and placebo in the proportion of people with new infections (12/16 [75%] with methylprednisolone v 6/8 [75%] with placebo; RR 1.80, 95% CI 0.86 to 3.76). The second RCT found no significant difference in infectious complications (defined as a new, unexpected, positive culture from the peritoneal fluid, cerebrospinal fluid, pleural fluid, circulating blood, sputum, or urine if not catheterised) between methylprednisolone and placebo (8/50 [16%] with methylprednisolone v 5/49 [10%] with placebo; p = 0.60). The third RCT reported no significant difference between the two treatment groups on rates of adverse events possible related to corticosteroids. The fourth RCT reported no increase in the rate of infectious complications, but corticosteroid treatment was associated with a higher rate of neuromuscular weakness. Forty-three serious infections were diagnosed in 30 people with placebo, compared with 25 serious infections in 20 people with methylprednisolone (P = 0.14). Serious adverse events of neuromyopathy were reported in nine people, all of whom were in the methylprednisolone group (P = 0.001).
Whether or not to use corticosteroids in ARDS has been controversial. When given early, in high doses, to people with acute lung injury/ARDS, corticosteroids had no impact on mortality, but increased the risk of infectious complications. Termed the fibroproliferative phase of ARDS, this stage is characterised by fever, purulent secretions, and new pulmonary infiltrates without evidence of infection.
Multiple drug treatments have been studied for people with ARDS and acute lung injury, and their role is extremely limited. Despite the vigorous inflammatory reaction in ARDS, we have found no evidence that corticosteroids have a role in early ARDS. The late phase of ARDS is characterised by an exaggerated fibroproliferative response, leading to persistent abnormalities in gas exchange. In practice, corticosteroids are sometimes used in selected cases where, despite optimal supportive therapy, ARDS is persistent (> 7 days) for late rescue treatment. If a decision is made to initiate treatment with corticosteroids, methylprednisolone is often given at 2 mg/kg daily initially, then tapered gradually over the next 32 days. People should be assessed to exclude infections at the outset, and close surveillance is required to identify subsequent infectious episodes, as people on corticosteroids may not develop a febrile response.
One systematic review and one subsequent RCT found no significant difference between nitric oxide and placebo in mortality, ventilator free days, or duration of hospital admission. One RCT identified by the review found that nitric oxide improved oxygenation compared with placebo, but the improvement was modest and not sustained.
We found one systematic review (5 RCTs, 535 people, excluding neonates in the first month of life) in people with acute hypoxaemic respiratory failure. A substantial number (about > 80%) of people in these trials had respiratory failure secondary to acute lung injury/acute respiratory distress syndrome. It found no significant difference between inhaled nitric oxide and placebo in mortality in trials without crossover (2 RCTs, RR 0.98, 95% C.I. 0.66 to 1.44; absolute figures not reported), or with crossover of treatment failures to open label (3 RCTs, RR 1.22, 95% CI 0.65 to 2.29; absolute figures not reported). One RCT identified by the review found that, compared with placebo, nitric oxide improved oxygenation in the first 24 hours after administration (mean oxygenation index in the first 24 hours: 14 [120 people] with nitric oxide v 17 [56 people] with control; p = 0.01). There was no significant difference in ventilator free days, duration of hospital stay, and intensive care stay (ventilator free days [alive and extubated at 30 days]: 80 people with nitric oxide v 148 people with placebo; WMD –1.37, 95% CI –3.62 to +0.88 [post hoc analysis; no further data reported]; people in intensive care at 90 days: 1/93 [1.1%] with nitric oxide v 1/87 [1.1%] with placebo; RR 0.94, 95% CI 0.06 to 14.73; people in hospital at 90 days: 6/93 [6.5%] with nitric oxide v 4/87 [4.6%] with placebo; OR 1.42, 95% CI 0.40 to 5.07). We found one subsequent RCT (385 people with moderately severe acute lung injury [ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FiO2 ratio) ≤ 250] not owing to sepsis and without non-pulmonary organ dysfunction at randomisation, septic causes excluded), which compared low dose (5 ppm) inhaled nitric oxide with placebo over 28 days. It found no significant difference between nitric oxide and placebo in the number of days people were alive and ventilator free at 28 days, or in mortality at 28 days (mean number of days people were alive and ventilator free: 10.7 days with nitric oxide v 10.6 days with placebo; p = 0.97; mortality: 44/192 [23%] with nitric oxide v 39/193 [20%] with placebo; P = 0.54).
The review gave no information on adverse effects. The subsequent RCT found that, compared with placebo, inhaled nitric oxide increased infections; however, none was judged by blind investigators to have been related to treatment gas administration (66/192 [34.4%] with inhaled nitric oxide v 41/193 [21.2%] with placebo; significance assessment not performed). It found a similar number of cardiovascular, gastrointestinal, endocrine, haematological, metabolic, nutritional, and nervous system adverse effects with both inhaled nitric oxide and placebo (absolute figures not reported).
Inhaled nitric oxide may improve oxygenation in people with ADRS. However, this beneficial effect on oxygenation in the people given nitric oxide compared with placebo remains modest and is not sustained.
Nitric oxide use has not been associated with better survival in RCTs; therefore, its use is not recommended as routine treatment, and must be considered experimental.
No new evidence