End-tidal CO2 monitors are used to estimate arterial CO2 pressure (PaCO2), but appropriate use of this noninvasive method of assessing blood gases is unclear. In patients with lung disease, the end-tidal CO2 pressure (PETCO2) can differ from PaCO2 because of ventilation-perfusion (VA/Q) mismatching, and changes in PETCO2 may be seen with corresponding increase, decrease, or no change in PaCO2 depending on what happens to VA/Q mismatching. We compared the difference between PETCO2 and PaCO2 in 17 patients undergoing mechanical ventilation. Large differences were found between PaCO2 and PETCO2 in individual patients; P(a-et)CO2 correlated closely with VD/VT. Our studies confirm that PetCO2 is a poor estimate of PaCO2 in patients with respiratory failure. However, the P(a-et)CO2 may be the most appropriate use for end-tidal PCO2 monitoring. In addition, we found that the end-tidal CO2 monitor may be easily adapted for expedient measurement of VD/VT.