Limits on VA/Q distributions from analysis of experimental inert gas elimination.

Distributions of ventilation-perfusion ratios (VA/Q) actually present in the lung cannot be exactly recovered using current inert gas elimination methods, principally because of the limited number of gases and errors in their measurement. The amount of information that can be gained is studied here. When six gases are used the results are difficult to visualize so that a graphical analysis is first given for only two gases. Methods are proposed for 1) the probabilistic description of retentions compatible with a set of measured values of all six gases, and 2) the placing of limits on the associated VA/Q distributions. It is found that the variability among distributions with compatible retentions depends greatly on the particular set of data. Distributions consisting of a single or of two separated narrow modes can be identified reliably.