Observational data collected for patients in intensive care under mechanical ventilation.

The application of alveolar recruitment manoeuvres consists of transiently increasing alveolar pressure in order to re-open collapsed alveolar territories and to re-ventilate poorly ventilated or non-ventilated lung areas.

The correlation between the pairs of values of CVP and PVP during an alveolar recruitment manoeuvre will be evaluated by Pearson's coefficient or Spearman's coefficient if the distribution of the values is not parametric. The coefficient of determination r² will be calculated by linear regression between the 2 variables. Finally, the agreement between the two measurements will be evaluated by the Bland-Altman method: the Bland-Altman graph will be provided and the 95% agreement limits will be calculated (+/- 1.96 SD, SD being the standard deviation of the difference between the measurements).

Predictive value of the evolution of PVP during the recruitment manoeuvre: the ROC curve (receiver-operating characteristic) will be performed by logistic regression. The area under the ROC curve will be given as well as the sensitivity, specificity and negative and positive predictive values (Younden index will be used to determine the optimal threshold of PVP).