Validation of a two-compartment model of ventilation/perfusion distribution

Ventilation (V˙A) to perfusion (Q˙) heterogeneity (V˙A/Q˙) analyses by a two-compartment lung model (2C), utilizing routine gas exchange measurements and a computer solution to account for O2 and CO2 measurements, were compared with multiple inert gas elimination technique (MIGET) analyses and a multi-compartment (MC) model. The 2C and MC estimates of V˙A/Q˙ mismatch were obtained in 10 healthy subjects, 43 patients having chronic obstructive pulmonary disease (COPD) and in 14 dog experiments where hemodynamics and acid–base status were manipulated with gas mixtures, fluid loading and tilt-table stressors. MIGET comparisons with 2C were made on 6 patients and 32 measurements in healthy subjects before and after exercise at normoxia and altitude hypoxia. Statistically significant correlations for logarithmic standard deviations of V˙A/Q˙ distributions (SD(V˙A/Q˙)) were obtained for all 2C comparisons, with similar values between 2C and both other methods in the 1.1–1.5 range, compatible with mild to moderate COPD. 2C tended to overestimate MC and MIGET values at low and underestimate them at high SD(V˙A/Q˙) values. SD(V˙A/Q˙) weighted by Q˙ agreed better with MC and MIGET estimates in the normal range, whereas SD(V˙A/Q˙) weighted by V˙A was closer to MC at higher values because the V˙A-weighted SD(V˙A/Q˙) is related to blood-to-gas PCO2 differences that are elevated in disease, thereby allowing better discrimination. The 2C model accurately described functional V˙A/Q˙ characteristics in 26 normal and bronchoconstricted dogs during non-steady state rebreathing and could be used to quantify the effect of reduced O2 diffusing capacity in diseased lungs. These comparisons indicate that 2C adequately describes V˙A/Q˙ mismatch and can be useful in clinical or experimental situations where other techniques are not feasible.