How Can We Explain the Lack of Improvement in Outcome in ARDS?
The positive results reported in RCTs may not be as efficient when applied to all nonselected patients with ARDS in an ICU. It has been found that mortality in ARDS was lower in RCTs than in observational studies that are closer to real life [51]. Indeed, patients included in RCTs are expressly selected, and those with major comorbidities such as hematological malignancies, cirrhosis, and chronic cardiac or respiratory disease are usually excluded. Among the three studies having compared two levels of PEEP, only one of them has provided a flow chart study. In this study, only 22% of the patients assessed for eligibility were included and randomized (768 of 3429 patients) [15].
In a RCT, after excluding patients with exclusion criteria, a high proportion of patients potentially eligible are not enrolled. The outcome of ARDS patients enrolled in a recent RCT has been compared to that of patients who met inclusion criteria but who were not enrolled in the study due to various reasons such as no consent, physician refusal (24%), missed randomization window, etc. [56]. The patients who were included in the study had lower mortality than those who were potentially eligible but not enrolled, suggesting that enrollment in clinical trials may be associated with improved outcomes. The better outcome reported in patients included in RCTs may be due to optimal management including standardized lung protective ventilation and application of other effective therapies. Indeed, in real- life situations, as reported in the recent LUNG SAFE survey, 35% of patients with ARDS were ventilated with a tidal volume above 8 mL/kg of predicted body weight. The mean PEEP level was only 8 ± 3 cmH2O in moderate and 10 ± 4 cmH2O in severe ARDS. Prone positioning was used in only 16% of the patients with severe ARDS. Therefore, in 2014, patients with ARDS were receiving excessively high tidal volumes and excessively low levels of PEEP in “real life,” while plateau pressure was measured in only 40% of the cases.