Prone position in mechanically-ventilated patients

The prone position, initially introduced into respiratory therapy to improve the drainage of secretions, has been suggested to improve oxygenation in anaesthetized and paralysed subjects. Here we report on the results obtained using the prone position in 17 normal subjects, 10 obese subjects and 16 patients with acute respiratory failure (ARF). In normal subjects, the prone position did not alter respiratory system compliance (Crs,st, 80.9 +/- 16.6 versus 75.9 +/- 13.2 mL.cmH2O-1), while it did improve arterial oxygen tension (Pa,O2, 21.3 +/- 4.9 versus 26.5 +/- 2.1 kPa, p < 0.01), this improvement being paralleled by an increase in functional residual capacity (FRC, 1.935 +/- 0.576 versus 2.921 +/- 0.681 L, p < 0.01). In obese subjects, the FRC (0.894 +/- 0.327 versus 1.980 +/- 0.856 L, p < 0.01) and the lung compliance (CL,st, 91.4 +/- 55.2 versus 109.6 +/- 52.4 mL.cmH2O-1, p < 0.01) increased in the prone position, while the chest wall compliance (Ccw,st, 199.5 +/- 58.7 versus 160.5 +/- 45.4 mL.cmH2O-1, p < 0.01) decreased and the Crs,st remained substantially unaffected (55.4 +/- 9.6 versus 59.6 +/- 12.1 mL.cmH2O-1). In these subjects, the improvements in FRC and CL,st were always paralleled by an increase in Pa,O2 (17.3 +/- 4.1 versus 24.1 +/- 3.7 kPa, p < 0.01). In ARF patients, no difference between the supine and prone position was found either in FRC (1.17 +/- 0.41 versus 1.29 +/- 0.57 L), or in Crs,st (38.4 +/- 13.7 versus 35.9 +/- 10.7 mL.cmH2O-1) or CL,st (52.4 +/- 23.3 versus 53.9 +/- 23.6 mL.cmH2O-1) despite a significant reduction in Ccw,st being observed (204.8 +/- 97.4 versus 135.9 +/- 52.5, p < 0.01). In this group of patients, the use of the prone position resulted in a significant increase in Pa,O2 (13.7 +/- 3.2 versus 17.2 +/- 4.4 kPa, p < 0.05), being the oxygenation improvement induced by the prone position relative to the baseline Ccw,st, according to the relationship: Pa,O2 = -32.4 + (0.24 x Ccw,st); r = 0.82, p < 0.01). The mechanisms which can be hypothesized to justify the modifications reported above are discussed.