Clinical studies have found that patients with acute respiratory distress syndrome (ARDS) account for approximately 5% of all hospitalized patients subjected to mechanical ventilation.1 Most studies have shown that patients with mild ARDS (PaO2/FiO2 200–300) represent only 25% of the cases, while the remaining 75% correspond to patients with moderate or severe ARDS.1,2 Acute respiratory distress syndrome is associated to an in-hospital mortality rate of approximately 40%.2,3 The mortality rate in turn is conditioned to the severity of the oxygenation defect. In the Berlin definition clinical trial, the mortality rate was found to be 27% in patients with mild ARDS, 32% in cases of moderate ARDS, and 45% in those with severe ARDS.4 Although the deterioration of oxygenation is a mortality risk factor in ARDS, patients generally die as a consequence of multiorgan failure, and only a minority (13–19%) die as a result of refractory hypoxemia.3,5 Although the mortality rate has decreased in the last few decades, probably due to the adoption of protective ventilation strategies (low tidal volume, optimum PEEP level, and limitation of plateau pressure), which lessen ventilator-associated lung injury,6,7 the fatality statistics have remained stable in recent years,2,3,5 without additional reductions in patients of this kind. It is therefore essential to establish other strategies or treatments that may result in further significant reductions in patient mortality.7

Mechanical ventilation in the prone position has been used for several decades in patients with ARDS with the purpose of improving oxygenation.8 It is currently well recognized that the prone position is associated to important improvements in the oxygenation indices when compared with the supine position. Furthermore, different studies in animals and humans have found that the prone position can reduce ventilator-associated lung injury.9 A number of clinical trials have been made with the aim of extrapolating these results to the clinical setting,10–15 though none of them have clearly demonstrated a positive impact upon patient survival.16 Likewise, the meta-analyses and systematic reviews published to date only suggest a tendency to reduce mortality in patients with severely impaired oxygenation.5,9,17–23 However, in the time elapsed from the first clinical trial10 to the most recent study,16 there have been important changes in ventilation and pronation strategies, as described in the latest study by Guerin et al.,16 where notorious benefit in terms of survival was documented among patients in the prone position – with an absolute decrease in mortality risk of 37%.

In the light of the fact that few strategies have had an impact upon survival in patients with ARDS, it is important that ventilation in the prone position has reappeared with strong results. However, because of the changes in pronation strategy and in the patient inclusion criteria in the most recent studies, we considered it necessary to determine the true impact of pronation, and to define which patients can benefit from it. A meta-analysis was therefore conducted with the primary objective of determining whether ventilation in the prone position reduces mortality in patients with ARDS compared with traditional ventilation in the supine position. As a secondary objective we described the groups of patients that show a positive impact upon survival and the parameters to be applied during the use of this ventilation strategy (timing of the start, daily duration, associated management strategies, etc.). Lastly, based on the results obtained, we aimed to offer a series of evidence-based recommendations on the use of ventilation in the prone position in patients with ARDS.

A total of 371 literature references were identified: PubMed n=134, EMBASE n=160, Cochrane Library n=62 and LILACS n=15. We discarded 124 due to duplication. Twenty-two references were fully reviewed, and of these 7 met the study inclusion criteria (Fig. 1).

Figure 1.

Study screening process. ON: nitric oxide.

(0.26MB).

Prone position and mortalityGlobal mortality

A total of 456 events were recorded in the prone position group (41.9%), versus 483 in the supine position group (46.8%), with an OR of 0.76 (95%CI: 0.54–1.06; p=0.11; I2 63%) showing a tendency in favor of the prone position group–though statistical significance was not reached (Fig. 2). Since the studies had different follow-up periods, and we used the final report on events at the end of this period for the global result, the findings were stratified and evaluated in different time periods: after 28 days (OR: 0.73; 95%CI: 0.41–1.32; p=0.3; I2 85%), after 90 days (OR: 0.64; 95%CI: 0.29–1.40; p=0.26; I2 84%) and after 180 days (OR: 0.97; 95%CI: 0.67–1.40; p=0.26; I2 28%), together with mortality in intensive care (OR: 0.86; 95%CI: 0.61–1.22; p=0.4; I2 31%)–no significant differences being found in each of the groups.

Figure 2.

Global mortality and bias risk. Mortality was evaluated up until the end of follow-up in each study. The bias risk of each study was scored as high (−), intermediate (?) or low (+).

(0.38MB).

Mortality and protective ventilation

On examining the association between mortality and the administered tidal volume, stratification into two groups was made according to whether a low tidal volume was used as part of a protective ventilation strategy, or a high tidal volume was administered (the latter appearing to be related to the development of ventilator-associated lung injury). Four studies11,14–16 used a tidal volume 8ml/kg ideal weight, showing a decrease in mortality risk of 36% (OR: 0.58; 95%CI: 0.38–0.87; p=0.009; I2 33%). This finding was not obtained when using a tidal volume >8ml/kg ideal weight (OR: 1.01; 95%CI: 0.77–1.32; p=0.94; I2 18%) (Fig. 3).

Figure 3.

Mortality and protective ventilation. The prone position exerted a protective effect in patients ventilated with a tidal volume <8ml/kg ideal weight.

(0.32MB).

Mortality and number of hours a day in the prone position

All the studies reported and analyzed data regarding the duration of pronation. In the four most recent publications11,12,14,16 the number of daily hours was increased (18h on average)–this resulting in a significant decrease in event risk in favor of the group of patients placed in the prone position for more than 12h (OR: 0.6; 95%CI: 0.43–0.83; p=0.002; I2 27%) (Fig. 4).

Figure 4.

Mortality and hours a day in the prone position. Significant differences were found in favor of the group placed in the prone position during more than 12h a day.

(0.33MB).

Mortality and start of pronation

In the same way that the number of hours in the prone position appears to be important, so does the timing of the start of patient placement in the prone position. In effect, greater benefit was observed when the patients were placed in the prone position within the first 48h after the start of mechanical ventilation, with an OR of 0.49 (95%CI: 0.35–0.68, p=0.0001; I2 0%) (Fig. 5).

Figure 5.

Mortality and timing of the start of pronation. Pronation within the first 48h was associated to a decrease in mortality risk.

(0.32MB).

Mortality and severity of hypoxemia

The studies were stratified according to the severity of hypoxemia: moderate (PaO2/FiO2 100–200) or severe (PaO2/FiO2<100). Taccone et al. stratified the population into these two groups; their data were therefore taken separately for analysis. The study published by Voggenreiter et al. reported patients with PaO2/FiO2>200. Five studies10–14 documented patients with moderate hypoxemia, while two studies14,16 reported patients with severe hypoxemia. On performing the meta-analysis, the group with severely impaired oxygenation showed clear benefit with the prone position (OR: 0.51; 95%CI: 0.36–1.25; p=0.0001; I2 0%) (Fig. 6).

Figure 6.

Mortality and severity of hypoxemia. Taccone et al. stratified the patients into individuals with moderate or severe hypoxemia. The data of each group were taken separately to the effects of analysis.

(0.37MB).

Prone position, stay in intensive care and days on mechanical ventilation

Four studies reported the stay in intensive care,11,12,14,16 and 5 studies11,13–16 recorded the days on mechanical ventilation. Analysis was discarded in both cases,14 since different measurement units were used. The study published by Guerin et al.16 reported the results in subgroups of survivors and non-survivors. No differences were found in the results between the two studied groups (stay in intensive care: difference of means −0.05; 95%CI: −2.98–2.89; p=0.00001; I2 95%, and days on mechanical ventilation: difference of means −1.19; 95%CI: −2.74–0.35; p=0.00001; I2 91%) (Fig. 7).

Figure 7.

Ventilation in the prone position, stay in intensive care and days on mechanical ventilation. No significant differences were found between the groups of patients.

(0.41MB).

Prone position and adverse effects

Pressure ulcers were the most frequent adverse events (34%), followed by ventilator-associated pneumonia (21.4%), orotracheal tube obstruction (14.6%), accidental extubation (10.9%), venous access loss (10.9%), pneumothorax (5.8%), and displacement of the orotracheal tube (3.7%).

The prone position was associated to a significantly increased risk of orotracheal tube obstruction (OR: 2.19; 95%CI: 1.55–3.09; p<0.0001; I2 0%) and the development of pressure ulcers (OR: 1.53; 95%CI: 1.21–1.94; p=0.0003; I2 0%). No differences were observed in relation to the rest of the described events (Fig. 8).

Figure 8.

Ventilation in the prone position and adverse effects. A direct relationship was found between pronation and the risk of pressure ulcers and orotracheal tube obstruction.

(1.02MB).

Assessment of publication bias

Visual inspection of the funnel plots revealed no evidence of publication bias (Fig. 9).

Figure 9.

Funnel plot. Visual inspection reveals no selection bias.

(0.06MB).

The primary objective of our study was to examine the impact of mechanical ventilation in the prone position upon mortality in patients with ARDS. Global assessment of the results revealed a nonsignificant tendency to reduce mortality risk in favor of the prone position group. However, on individually analyzing the studies, we found that the most recent trials11,12,14,16 incorporated a number of changes in both the inclusion criteria and in the prone position protocol, based on the analysis of the probable causes underlying the discouraging results recorded in the earlier studies.10,13,15 A first change was the inclusion of more severely compromised patients, with PaO2/FiO2<200, while a second change was the prolongation of pronation time to over 16 consecutive hours a day. A third change was the utilization of protective ventilation strategies with a tidal volume of <8ml/kg ideal weight, the maintenance of a plateau pressure of <30cmH2O, and titration of the optimum PEEP level according to the recommendations of the ARDS network. On examining the studies linearly over time, we observed a tendency to gradually favor the prone position, culminating in the study published by Guerin et al.,16 where a strong reduction in mortality risk was recorded in favor of pronation (HR 0.44; 95%CI: 0.29–0.67; p=0.001 after 90 days). These findings are consistent with those of two recent meta-analyses. The first study was carried out by Beitler et al.,34 who conducted a meta-analysis of individual patient data, including the PROSEVA study,16 assessing mortality after 60 days, with the identification of a nonsignificant tendency in favor of pronation (RR: 0.83; 95%CI: 0.68–1.02; p=0.073). The second study was published by Lee et al.,35 based on broader inclusion criteria and a larger number of studies, and identified a global decrease in mortality risk, though with a confidence interval very close to one (OR: 0.77; 95%CI: 0.59–0.99; p=0.039).

However, interesting data were obtained on stratifying the results by subgroups, supporting our theory regarding the evolution and optimization of the prone position ventilation strategy based on plausible physiology and clear clinical characteristics. Firstly, the utilization of a low tidal volume (<8ml/kg ideal weight) in patients with ARDS became generalized after publication of the ARDS network study6, which revealed a decrease in mortality risk probably related to the generation of lesser mechanical stress upon the alveolar membrane by preventing overdistension and improving alveolar stability.36 This, being associated to recruitment capacity and homogenization of the distribution of ventilation, flow and airway pressures attributed to pronation,37 probably also produce an additive effect in the prevention and reduction of ventilator-associated lung injury. Thus, on evaluating the subgroup of patients in which a tidal volume of <8ml/kg ideal weight was used,11,14–16 a significant decrease in mortality risk was noted in comparison with the group receiving a greater tidal volume10,13,15–these findings possibly being attributable to the decrease in ventilator-associated lung injury. Beitler et al.34 and Lee et al.35 also recorded a significant decrease in mortality risk in those studies that used low tidal volumes (RR: 0.66; 95%CI: 0.5–0.86; p=0.002 and OR: 0.62; 95%CI: 0.48–0.69; p=0.015, respectively). Beitler et al.34 in turn performed a metaregression analysis that revealed a dose–response relationship between the mean basal tidal volume and the relative risk of death after 60 days during pronation. Specifically, a decrease in mean basal tidal volume of 1ml/kg ideal weight was seen to be associated to a 16.7% mortality risk.34

The degree of alveolar recruitment in the prone position also requires consideration. Acute respiratory distress syndrome is characterized by disruption of the alveolar-capillary barrier, with increased permeability of the latter and the production of flooding and alveolar edema furthermore associated to the depletion of surfactant–thus giving rise to alveolar instability and collapse.38 Pulmonary involvement is heterogeneous, with well aerated lung regions that participate in gas exchange, and other regions that have collapsed as a result of the overpressure exerted by the interstitial edema and alveolar flooding39,40–these mechanisms explaining the decrease in lung volume in these patients.40 Pronation makes it possible to recruit these affected zones,37 redistributing and homogenizing ventilation,30 reducing the intrapulmonary shunt effect and improving oxygenation, ventilation and lung mechanics.41 However, the degree of recruitment depends on factors such as the severity of lung involvement,25 the pronation time39 and the time elapsed from lung injury to patient placement in the prone position.19

Although the prone position can effectively increase oxygenation when used several days after onset of the disease,10 its application during the early stages was found to offer betters results. During these early stages, all the conditions favoring the effectiveness of pronation are present, such as alveolar edema, reversible collapse and the absence of pulmonary structural alterations.37 In these stages, the reduction in ventilator-associated lung injury risk probably exceeds that obtained when pronation is used in later stages of ARDS, when the damage has already been caused.42 This has been clearly demonstrated in our study, where the patients placed in the prone position within the first 48h of the evolution of the disease showed a clear protective effect referred to mortality risk.

In clinical practice, the severity of ARDS has been rated according to the PaO2/FiO2 ratio,4 though in disorders as complex as this syndrome, the PaO2/FiO2 ratio depends on the PEEP level and FiO2 administered, as well as on the prescribed treatments and/or interventions, the comorbidities, and the innate compensatory mechanisms of the disease.43 Despite the presence of these variables, the results obtained allow us to clearly establish that the prone position is indicated in patients with severely impaired oxygenation, as has already been demonstrated in other studies.5,17–19,21–23 Furthermore, prolongation of the prone position to over 12 consecutive hours a day (18h on average) in patients with severe ARDS is a highly recommended strategy.11,12,14,16 It is important to note that in patients with mild ARDS, pronation has not been found to offer clinical advantages and is therefore not advised.15 In cases of moderate ARDS, the clinical recommendation is likewise not clear, though the results of the post hoc analysis of a meta-analysis9 revealed a certain tendency to benefit patients with PaO2/FiO2<140. Consequently, when taken in combination with the results of the PROSEVA study,16 these data allow us to consider the use of this strategy in this patient subgroup.37,44

The prone position offers clinical benefits such as improved oxygenation, by optimizing lung recruitment and the ventilation–perfusion ratio, and probably also prevents and reduces ventilator-associated lung injury by homogenizing the stress and strain upon the lung parenchyma, resulting in a decrease in mortality risk.

Based on the results obtained, the prone position can be recommended in patients with severe hypoxemia (PaO2/FiO2<100), associated to a low tidal volume (<8ml/kg ideal weight), during a period of over 16h a day, and starting early during the course of the disease (<48h). These consequently would be the indications and associated strategies to be included in pronation protocols.

Pronation requires no special equipment, but should be carried out by trained personnel and adopting the required safety measures in order to avoid associated complications.

The authors declare that no financial support has been received for this study from any government or private organism.

The authors declare that they have no conflicts of interest.