Effects of Clinical Trials on Mechanical Ventilation Practice

Abstract & Commentary

By David J. Pierson, MD, Editor, Professor, Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington, Seattle, is Editor for Critical Care Alert.

Synopsis: Among 1675 adult patients who were mechanically ventilated for more than 12 hours in 107 ICUs, noninvasive ventilation was used more often, smaller tidal volumes were used in patients with ARDS, and spontaneous breathing trials were used more often to predict successful extubation, in comparison with findings 6 years previously in a similar patient cohort from these same ICUs.

Source: Esteban A, et al. Evolution of mechanical ventilation in response to clinical research. Am J Respir Crit Care Med 2008;177:170-177.

In 2002, esteban et al reported the results of an ambitious study, enrolling a prospective cohort of consecutive adult patients in 361 ICUs in 20 countries who received mechanical ventilation for more than 12 hours during the month of March 1998.1 Although the primary purpose of the study was to determine mortality among patients requiring mechanical ventilation, and to look for variations according to geography, demographics, and ventilatory strategy, the investigators also documented how the patients were ventilated.

Of the 5183 ventilated patients in the 1998 study (33% of all patients admitted to the study ICUs during the study month), the majority (roughly two-thirds, across all diagnoses) were managed on volume assist-control ventilation, while about 15% received synchronized intermittent mandatory ventilation (SIMV), with or without added pressure support. Among the 231 patients with the acute respiratory distress syndrome (ARDS), 10% received pressure control ventilation, whereas of 522 patients ventilated because of exacerbations of chronic obstructive pulmonary disease (COPD), this mode was used in 4%. In COPD patients who received mechanical ventilation, noninvasive positive-pressure ventilation (NPPV) was used 17% of the time, while this proportion was 4% in cases of hypoxemic acute respiratory failure.

Seventy percent of all ventilated patients in 1998 underwent attempts at weaning, and these attempts consisted of once-daily spontaneous breathing trials (SBTs, in 78% of patients), multiple SBTs per day (in 14%), gradual reduction in pressure support (in 21%), and gradual reduction in SIMV, either alone (in 8%) or in combination with pressure support (in 22%). For SBTs, a T-tube was used in 52% of patients, continuous positive airway pressure in 19%, and pressure support in 28%. About 12% of patients received a tracheostomy, and 28-day mortality for the entire cohort was 31%.

Because a number of important changes have occurred both in the evidence base supporting mechanical ventilation and in prevailing practice standards since the 1998 study1 was carried out, Esteban and colleagues in the so-called VENTILA Group undertook a second, similarly large-scale, international cohort study of ventilated patients, to assess the evolution of clinical practice during the interval. Their ambitious investigation had two parts: a descriptive study of how and why mechanical ventilation was carried out in 349 adult ICUs in 23 countries during the month of March 2004; and a comparison of various aspects of ventilatory management, in 2004 vs 1998, in the 107 ICUs from 18 countries that participated in both studies. For the second part, the authors carried out a systematic review of studies dealing with mechanical ventilation in acute respiratory failure that were published prior to the appearance of their earlier paper, and also studies appearing between 1998 and 2004. Using the results of this literature search, and before examining the data from their 2004 patient cohort, the authors generated 11 hypotheses about how the practice of mechanical ventilation might have changed as a result of the research findings. They then tested each of these hypotheses using data from the 1998 and 2004 patients studied in the same set of ICUs. The hypotheses, shown in the form of expected changes in each of 11 aspects of ventilator management, are listed in Table 1, along with the actual changes that were found.

Table 1

Altogether, the 2004 study included 4968 patients ventilated in the 349 ICUs during the one month observation period. This represents 25% of all the patients admitted to those ICUs during that period. As in 1998, their mean age was 59 years, and 60% were men. A medical problem was the reason for admission in 59% of the patients; it was COPD in 267 patients (5%) and ARDS in 148 (3%), both of these being only about half as many as in the earlier cohort. Postoperative mechanical ventilation (in 21% of patients), pneumonia or sepsis (20%), and coma (19%) were the most common admission diagnoses. Mortality in the ICU and in the hospital were 31% and 37%, respectively, essentially the same as in the 1998 cohort.

In the 2004 cohort, 1675 patients were managed in the 107 ICUs that had also participated in the 1998 study. As predicted (see Table 1), NPPV was used substantially more often in 2004 than in 1998, although neither the need for intubation nor hospital mortality changed for such patients. For patients with ARDS, tidal volumes used during the first week decreased, and values exceeding 10 mL/kg (in 8% vs 30% of patients) were less frequently encountered. Higher levels of positive end-expiratory pressure (PEEP) were used during the first week (>10 cm H2O in 40% of ARDS patients as compared with 28%), although the amount of PEEP used in the entire patient cohort was little changed. Volume assist-control remained the mode most often used; the use of pressure control ventilation did not increase. Prone positioning was used less often in the second study. With respect to weaning, SBTs were used more often prior to extubation, and the use of a T-piece remained the most common method for doing this. However, there was a dramatic decrease in the use of SIMV during weaning, along with a concomitant increase in the use of pressure support. The use and timing of tracheostomy did not change.


As a result of their comprehensive literature review, Esteban et al proposed 3 general areas in which the practice of mechanical ventilation might have changed since 1998: reducing the need for invasive mechanical ventilation, reducing the duration of invasive mechanical ventilation, and improving its safety. To examine these 3 areas they generated 11 specific hypotheses: increased use of NPPV, for both COPD exacerbations and acute hypoxemic respiratory failure; increased use of SBTs to assess readiness for extubation, with more use of pressure support and less use of SIMV during the weaning process; the use of smaller tidal volumes and higher PEEP levels for patients with ARDS; and no change in the use of pressure control ventilation, prone positioning, or tracheostomy. Of these 11 practice-change hypotheses (see Table 1), 10 were borne out in the authors' comparisons of patients managed in the same ICUs in 1998 and 2004.

Of importance, there were 3 areas in which the observed practice change was concordant with the results of randomized controlled trials demonstrating decreased mortality and other improved patient-relevant outcomes: using NPPV for COPD exacerbations, using smaller tidal volumes in ARDS, and using SBTs to shorten the duration of intubation by identifying patients ready for weaning. Clearly, the practice of mechanical ventilation appears to be evolving in the right direction, and this study's results suggest that randomized trial results have advanced such practice internationally. However, Esteban et al found no differences in ICU length of stay, or in ICU or hospital mortality, between 1998 and 2004 among either patients treated with NPPV or patients with ARDS. Although the study was designed primarily to compare practice patterns in the two cohorts, and not to demonstrate outcome differences in comparable patient populations, this finding is disappointing.


  1. Esteban A, et al., for the Mechanical Ventilation International Study Group. Characteristics and outcomes in adult patients receiving mechanical ventilation. JAMA 2002;287:345-355.