Determination of PAO2 by the “ideal” alveolar gas concept and by miget in ventilated animals

The Ideal Alveolar Gas Concept calculates the alveolar partial pressure of oxygen (PAO₂) using arterial (PaCO₂) instead of alveolar (PACO₂) partial pressures of carbon dioxide in the alveolar gas equation. We compared the effects of shunt on PAO₂ calculated by the “ideal” gas concept with a calculation using the multiple inert gas elimination technique (MIGET). We tested the hypothesis that shunt affects PACO₂ and introduces significant errors in the estimation of PAO₂ in two experimental porcine models. First, in a lung-lavage model (n = 10), shunt changes were induced by applying different levels of positive end-expiratory pressure. PaCO₂ median 71 mmHg, (IQR 21 mmHg) was higher than MIGET PACO₂ 52(20) mmHg (p < 0.001). Ideal PAO₂ 654(33) mmHg was lower than MIGET PAO₂ 670(19) mmHg (p < 0.001). Bias in PAO₂ was −23 mmHg with limits of agreement between 19 to −65 mmHg. Second, in a one-lung ventilation model (n = 10), changes in shunt were performed decreasing cardiac output by inferior vena cava balloon inflations (OLV_CB), inhibiting hypoxic pulmonary constriction with nitroprusiate (OLV_NPS) and increasing cardiac output with dobutamine (OLV_DBT). Baseline PaCO₂ was 55(10) mmHg and shunt 9(19)%. Cardiac output reduction did not affect PaCO₂ 57(9) mmHg (p = 0.19) or shunt 9(15)% (p = 0.570). Inhibiting hypoxic pulmonary vasoconstriction increased PaCO₂ [62(13) mmHg; p = 0.083] and shunt [26(20)%; p = 0.020]. Cardiac output increase resulted in higher PaCO₂ [67(11) mmHg; p = 0.004] and shunt [30(29)%; p = 0.012]. The assumption of the ideal gas concept that PaCO₂ is not affected by shunt is incorrect introducing significant estimation errors in the alveolar gas equation.