Rationale: The stress index (SI) is an airway pressure-time profile (Paw-t) derived parameter measured at the airway opening and intended to detect tidal recruitment (TR) and overdistension of the lung during constant flow mechanical ventilation at the bedside. This potential has not been tested in experimental conditions other than the saline lavage model. Because both TR and overdistention can occur simultaneously, we compared the performance of SI and quantitative computed tomography (CT), in a unilateral pleural effusion (PLEF) model. Methods: Thirteen ml/kg (PLEF-Mod) and 26 ml/kg (PLEF-Large) of an isotonic radioopaque solution were instilled into the right hemithorax of five anesthetized pigs mechanically ventilated under controlled conditions with tidal volume 9 ml/kg. Low (1 cmH20, P1) and moderate (10 cmH20, P10) PEEP were applied under three conditions: baseline (No PLEF, BSL), PLEF-Mod and PLEF-Large. Thus, six stages were studied: BSL-P1, BSL-P10, PLEF-Mod-P1, PLEF-Mod-P10, PLEF-Large-P1 and PLEF-Large-P10. Whole-lung CT scans were acquired during end-inspiration and end-expiration at each stage. Static tidal compliance of the respiratory system (Cst) and dynamic Paw-t waveforms were recorded during tidal breathing. Four to eight supradiaphragmatic slices were analyzed. TR was calculated as the difference ( ƒ ¢) in collapsed lung volume (-100 to 200 HU voxels) at end expiration and end-inspiration, indexed to the corresponding ƒ ¢ gas volume. Tidal overinflation (TO) was calculated as the ƒ ¢ overinflated lung volume (-1000 to -900 HU voxels) at end-inspiration and end-expiration indexed to the corresponding ƒ ¢ gas volume. The inspiratory Paw-t plot was fit to an exponential equation, as originally described. The overall SI was computed as the average of the exponent in 54-96 breaths per stage. The same analysis was performed for the first (SI-1) and second half (SI-2) the inspiratory Paw-t curves. RESULTS: Values are presented as mean +SD in table 1. Except for BSL-P1, the SI was always >1 suggesting overdistension preponderance, while TR increased remarkably with PLEF addition. PEEP increase always augmented the SI and reduced TR, but effects on Cst were variable. PEEP increase worsened Cst at BSL suggesting overdistension. In contrast, TO was always minimal. Except for BSL-P1, SI-2 was always greater than SI-1. CONCLUSIONS: 1) The SI detects small amounts of TR healthy animals ventilated with low PEEP, but can be misleading under multi-compartmental conditions with heterogeneous collapse and extensive TR. 2) Inconsistency of the SI partition within the breath suggests diversity of TR and TO events throughout insufflation.