Meta-analysis: Ventilation Strategies and Outcomes of the Acute Respiratory Distress Syndrome and Acute Lung Injury
C. Putensen,N. Theuerkauf,Jrg Zinserling,H. Wrigge,P. Pelosi
DOI: https://doi.org/10.7326/0003-4819-151-8-200910200-00011
IF: 39.2
2009-10-20
Annals of Internal Medicine
Abstract:Context Ventilation strategies to protect the lungs of patients with the acute respiratory distress syndrome (ARDS) include low tidal volume, limited airway pressures, and medium to high positive end-expiratory pressure (PEEP), but the adoption of these strategies has been slow in some clinical settings. Contribution This review of randomized trial evidence for low tidal volume and high PEEP ventilation on mortality of patients with ARDS or acute lung injury found that trials were limited in number but showed mortality benefits with lower versus higher tidal volume. High PEEP did not improve mortality in unselected patients but may help patients with life-threatening hypoxemia despite other interventions. Implication Lower tidal volume ventilation strategies should be used for patients with ARDS or acute lung injury. The Editors The acute respiratory distress syndrome (ARDS) is clinically characterized by sudden onset, severe hypoxemia, radiographic evidence of bilateral pulmonary infiltration, and absence of left heart failure (13). Acute lung injury is a subset of ARDS with less severe impairment in oxygenation. Despite apparent improvement in management and outcome of ARDS, the mortality rate in persons with the disease remains high, ranging from 35% to 65% (4). Although mechanical ventilation provides essential life support, it can worsen lung injury (5). Computed tomography images of patients with ARDS show nonhomogeneous distribution of pulmonary aeration. Normally aerated lung regions are relatively small but, when they receive the largest part of tidal volume (Vt) (6, 7), may be exposed to excessive alveolar wall tension and stress because of overdistention (8, 9). Atelectatic lung regions are prone to cyclic recruitment and derecruitment, leading to shear stress in adjacent aerated and nonaerated alveoli (1012). Ventilator-induced lung injury is caused by excessive stress or strain to lung tissues that occurs during mechanical ventilation and aggravates inflammation and diffuse alveolar damage (5, 13). Lung-protective ventilation strategies include ventilation with low Vt and limited airway pressure to reduce ventilator-induced lung injury from overdistention while allowing hypercapnia and medium to high positive end-expiratory pressure (PEEP) to keep alveoli open throughout the ventilator cycle (14). Hypercapnia and acidosis may increase intracranial pressure, induce pulmonary hypertension, depress myocardial contractility, decrease renal blood flow, and release endogenous catecholamines (15). In addition, prevention of cyclic derecruitment with higher PEEP may contribute to overdistention of normally aerated alveoli, counterbalancing the benefits from low Vt and limited airway pressure ventilation cycles (14). The effect of different lung-protective ventilatory strategies in patients with acute lung injury or ARDS has been investigated in randomized, controlled trials (RCTs) testing higher versus lower Vt ventilation at similar PEEP (1619), higher versus lower PEEP strategies during low Vt ventilation (2022), and lower Vt and PEEP titrated greater than the lower inflection point of the individual pressure volume curve versus higher Vt and lower PEEP (23, 24). Results were partially conflicting because of differences in study design and number of enrolled patients. This may explain why most critically ill patients are still ventilated with high Vt at lower or even no PEEP (4, 25). Our objective was to determine whether the different lung-protective ventilatory strategies improve outcome in critically ill adults with acute lung injury or ARDS. Methods Data Sources and Searches We aimed to identify all RCTs assessing the efficacy and outcomes of lower Vt ventilation, higher PEEP application, or a combination of both in adults with acute lung injury or ARDS. The electronic search strategy applied standard filters for identification of RCTs. We searched the Cochrane Central Register of Controlled Trials, MEDLINE (from inception to March 2009), and EMBASE (from inception to March 2009). Our search included the following keywords: acute lung injury, ALI, adult respiratory distress syndrome, ARDS, protective ventilation, lung protective ventilation strategy, pressure-limited ventilation, tidal volume, positive end-expiratory pressure, PEEP, and random. We did not apply language restrictions. In addition to the electronic search, we checked out cross-references from original articles and reviews. Selection of Studies We restricted the analysis to RCTs to guarantee control of selection bias. We did not include study designs containing inadequately adjusted planned co-interventions and quasi-randomized or crossover trials. We considered RCTs that reported mortality as a predefined end point and compared lower versus higher Vt ventilation, lower versus higher PEEP application, or a combination of these strategies in intubated and mechanically ventilated critically ill adults with acute lung injury or ARDS from any cause. Acute lung injury and ARDS had to be defined by the American-European Consensus Conference criteria (26) or by the Lung Injury Severity Score (27). Trials with a low Vt ventilation strategy had to use lower Vt, maximal inspiratory plateau pressure (Pei) of 30 cm H2O or less, or a combination, which resulted in Vt of 8 mL/kg of body weight or less, compared with conventional mechanical ventilation that used Vt ranging between 10 and 15 mL/kg. Regardless of the strategy used to deliver the lower Vt, the 2 study groups had to differ only for Vt and not for other variables associated with a low Vt ventilation strategy. Trials with high PEEP ventilation strategies had to use higher PEEP based on Fio 2PEEP scales, titrating PEEP to greater than the lower inflection point of the individual static or quasi-static pressure volume curve at enrollment or titrating PEEP as high as possible without increasing the maximal Pei to greater than 30 cm H2O compared with conventional mechanical ventilation that used lower PEEP based on fixed Fio 2PEEP scales or lower PEEP at higher Fio 2 to ensure adequate arterial oxygenation. We excluded studies in postoperative patients and those published only in abstract form. We contacted authors to clarify details of trials when necessary. Outcome Measures The primary outcome was mortality, evaluated at hospital discharge. Secondary outcomes included mortality at the end of the planned follow-up, barotrauma, use of rescue therapies owing to life-threatening hypoxemia, ventilator settings, and pulmonary function variables. Barotrauma was defined as any new pneumothorax, pneumomediastinum, subcutaneous emphysema, or pneumatocele after random assignment. Data Extraction and Quality Assessment Two pairs of independent reviewers performed the initial selection by screening titles and abstracts. Citations were selected for further evaluation if the studies they referred to were RCTs of lung-protective ventilatory strategies in critical ill adults or if the title or abstract did not give enough information to make an assessment. For detailed evaluation, we obtained the full text of all possibly relevant studies. Data from each study were extracted independently by the paired reviewers by using a prestandardized data abstraction form. One pair of reviewers was not informed about authors, journal, institutional affiliation, and date of publication. Data extracted from the publications were checked by another reviewer for accuracy. Quality assessment of these studies included use of randomization, reporting of allocation concealment, blinding, adequate selection and description of study population with respect to inclusion and exclusion criteria, similarity of the groups at baseline, use of a predefined treatment protocol, absence of confounders, absence of co-interventions, a priori definition of primary and secondary outcome variables, use of intention-to-treat analysis, extent of follow-up, a priori calculation of sample size, number of patients screened and included in the trial, reports on patients lost to follow-up, and planned or premature termination of the RCT. Two reviewers independently used these criteria to abstract trial quality. We resolved any disagreements by consensus in consultation with a third reviewer if needed. Data Synthesis and Analysis We studied the following comparisons: lower versus higher Vt ventilation using similar PEEP strategies, lower versus higher PEEP level during low Vt ventilation, and the combination of higher Vt and lower PEEP level versus lower Vt and higher PEEP level. Qualitative Analysis We used a narrative summary approach to describe study characteristics and variation in quality indicators among studies and to consider how these factors affect our understanding of the outcomes of the RCTs included in the Cochrane review (28, 29). Quantitative Analysis The meta-analysis was performed according to the Cochrane Collaboration guidelines (30). All statistical analyses were performed with Review Manager, version 4.2 (The Nordic Cochrane Center, Copenhagen, Denmark), the Cochrane Collaboration's software for preparing and maintaining Cochrane systematic reviews (30). The pooled effects estimates for binary variables were expressed as odds ratios with 95% CIs, whereas continuous variables were expressed as weighted mean differences with 95% CIs. We tested the difference in estimates of treatment effect between the treatment and control groups for each hypothesis by using a 2-sided z test with statistical significance considered at a P value of less than 0.05. We examined heterogeneity by using the Cochran Q and the I 2 test (31, 32). We predefined heterogeneity as low, moderate, and high, with I 2 statistics greater than 25%, 50%, and 75%, respectively (32). Meta-analysis with a random-effects model was applied with I 2 statistics greater than 25% (33). Otherwise, we performed meta-analysis by using a fixed-effects model. However, the possibil