Objective: To determine whether positive end-expiratory pressure (PEEP) and prone position present a synergistic effect on oxygenation and if the effect of PEEP is related to computed tomography scan lung characteristic.
Design: Prospective randomized study.
Setting: French medical intensive care unit.
Patients: Twenty-five patients with acute respiratory distress syndrome.
Interventions: After a computed tomography scan was obtained, measurements were performed in all patients at four different PEEP levels (0, 5, 10, and 15 cm H2O) applied in random order in both supine and prone positions.
Measurements and Main Results: Analysis of variance showed that PEEP (p < .001) and prone position (p < .001) improved oxygenation, whereas the type of infiltrates did not influence oxygenation. PEEP and prone position presented an additive effect on oxygenation. Patients presenting diffuse infiltrates exhibited an increase of Pao2/Fio2 related to PEEP whatever the position, whereas patients presenting localized infiltrates did not have improved oxygenation status when PEEP was increased in both positions. Prone position (p < .001) and PEEP (p < .001) reduced the true pulmonary shunt. Analysis of variance showed that prone position (p < .001) and PEEP (p < .001) reduced the true pulmonary shunt. The decrease of the shunt related to PEEP was more pronounced in patients presenting diffuse infiltrates. A lower inflection point was identified in 22 patients (88%) in both supine and prone positions. There was no difference in mean lower inflection point value between the supine and the prone positions (8.8 ± 2.7 cm H2O vs. 8.4 ± 3.4 cm H2O, respectively).
Conclusions: PEEP and prone positioning present additive effects. The prone position, not PEEP, improves oxygenation in patients with acute respiratory distress syndrome with localized infiltrates.
Positive end-expiratory pressure (PEEP) usually is used to restore oxygenation by various mechanisms that include maintaining alveolar units open after they have been recruited, reversal of alveolar flooding, and alteration in regional blood flow and hemodynamics. However, not all patients with acute respiratory distress syndrome (ARDS) respond to PEEP. Patients with primary ARDS and those with localized infiltrates by computed tomography (CT) are much less responsive to PEEP than are those whose lungs are diffusely injured by an extrapulmonary process (e.g., sepsis) (1, 2). CT scan has demonstrated that PEEP-induced aeration of collapsed lung regions was markedly influenced by the lung morphology feature at zero end-expiratory pressure (ZEEP) (2, 3). In patients with diffuse infiltrates, PEEP-induced aeration has been shown to be homogeneously distributed along the cephalocaudal axis (2, 3). In contrast, in patients with lobar and patchy infiltrates whose upper lobes remained partially aerated, PEEP-induced aeration of collapsed lung regions decreased from the lung apex to the diaphragm. PEEP is able to modify pulmonary blood flow anterior to posterior distribution. For example, when PEEP overdistends alveolar units situated in nondependent parts of the lungs, it redirects blood flow toward dependent injured lung areas and therefore can deteriorate oxygenation. Prone positioning decreases the vertical gradients of pleural and transpulmonary pressures, tending to reopen the previously dependent dorsal and juxtadiaphragmatic regions. Because pulmonary blood flow remains distributed primarily to the dorsal quadrants, better ventilation/perfusion matching often is accomplished (4). It was hypothesized (2) that in ARDS patients with localized infiltrates, although PEEP does not modify oxygenation while increasing PEEP level in the supine position, PEEP could improve oxygenation in the prone position (PP). Our main objective was to evaluate if PEEP and PP presented a synergistic effect on oxygenation and to investigate whether the oxygenation response to the level of PEEP differs in magnitude according to the lung morphology (localized infiltrates or diffuse infiltrates) and the position (supine or prone).